Conceptual Issues

In Environmental Archaeology

EDITED BY
JOHN L BINTLIFF
DONALD A. DAVIDSON
ERIC G. GRANT

Edinburgh
UNIVERSITY PRESS
) Edinburgh University Press 1988
22 George Square, Edinburgh

Set in Times Roman and

printed in Great Britain by
J. W. Arrowsmith Ltd, Bristol

British Library Cataloguing

in Publication Data
Conceptual issues in environmental
archaeology
1. Environmental archaeology
I. Bintliff, J. II. Davidson, D.
III. Grant, E.
930.1 CC81
ISBN 0 85224 545 9 (Cased)
ISBN 0 85224 546 7 (Pbk)
 Contents

Preface p. ix
Conceptual issues in environmental
archaeology: an overview
Robert McC. Adams

Section One: Conceptual Issues at the

Site and Microregion Scale
Introduction 19
Donald A. Davidson
The role of soils in the interpretation of
archaeological sites in northern Quebec 25
T. R. Moore and D. Denton
Environmental reconstruction at Cul A'
Bhaile, Isle of Jura 38
G. Whittington
The contributions of shells to site
interpretations: approaches to shell
material from Franchthi cave 49
Margaret R. Deith and Judith C.
Shackleton
Thoughts on archaeology, geography and
scales 59
P. J. R. Modderman
The Tenta catchment: resource exploitation
in Neolithic Cyprus 64
Malcolm Wagstaff
On-site and off-site evidence for changes in
subsistence economy: Iron Age and
Romano-British. West Wales 72
H. C. Mytum
Some quantitative approaches to site
territory and land use from the surface
record 82
C. F. Gaffney and V. L. Gaffney

Section Two: From Micro-to-Macro-

region
Introduction 93
John Bintliff
Defining the economic space of a typical
Iron Age hillfort: Rodik (Yugoslavia), a
case study 95
Bozidar Slapsak
The reconstruction of environment,
irrigation and development of settlement
on the Hbur in North-East Syria 108
Peter J. Ergenzinger, Wolfgang Frey,
Hartmut Khne and Harald Krschner
Site patterning: separating environmental,
cultural and preservation factors 129
John Bintliff
An integrative model for the use of different
landscapes in Linearbandkeramik times 145
A. J. Kalis and A. Zimmermann
Prehistoric agriculture in the New Guinea
Highlands: problems in defining the
altitudinal limits to growth 153
Tim Bayliss-Smith
Trial by dure: a review of historical-
geographical concepts relevant to the
archaeology of settlement on Corsica and
Sardinia 161
James Lewthwaite

Section Three: Conceptual Issues Associated

with Environmental and Population Change
Introduction 189
Eric Grant
Erosional equilibria under grazing 193
J. B. Thornes
Environmental thresholds, catastrophe theory
and landscape sensitivity: their relevance to
the impact of man on valley alluviation 211
P. J. Burrin and R. G. Scaife
Cultural and environmental change during the
Palaeolithic: research methods combining
geomorphology with prehistory 233
F. Hours and L. Copeland
Palynology and human ecology: theory and
method for the study of prehistoric
agriculture 246
Stanton W. Green
Smut, scab and pox: disease and environ-
mental archaeology 273
Don Brothwell
Interaction between the environment and
social subsystems 278
W. Groenman-van Waateringe
Foraging theory: mathematical modelling of
socioecological change 283
K. V. Boyle and C. D. Wright
Putting pressures on population: social
alternatives to Malthus and Boserup 291
John Chapman
Index 311
 Preface

The papers in this book evolved from a symposium which was held
at the University of Oxford, 22-25 February 1985, on the topic
"Conceptual Issues in Environmental Archaeology." We took the
initiative in organising this meeting and subsequently editing this
book since it is our view that there was a pressing need for examining
the conceptual issues associated with the integration of environmental
research and archaeology. The papers in this book are extremely
diverse in terms of topic and approach, but we feel that their overall
effect is to illustrate the many issues associated with environmental
analysis for archaeology. Through the provision of an overview
chapter by Robert McC. Adams and section introductions written
by ourselves, we have made every attempt to highlight the major
themes. Our hope is that archaeologists will be encouraged to think
more carefully about how environmental research can be planned
and executed in close liaison with archaeological objectives. Equally,
environmental specialists need to be reminded that their role is not
just the provision of palaeoenvironmental data, but rather to con-
tribute to interpretation of human societies in the past.
We are very grateful to all the contributors who made this book
possible. Particular thanks are expressed to Robert McC. Adams who
not only played a key role at the meeting, but also wrote the overview
chapter. We acknowledge various forms of help from the Universities
of Bradford, Leicester, Stirling and Strathclyde and Middlesex Poly-
technic. We are particularly grateful to Sylvia Dransfield who typed
the papers and to Alan Clough who solved the floppy disc translation
problem. Myra Shackley was a co-organiser of the meeting and we
are sorry that she was unable to participate as an editor. Finally, we
are pleased to acknowledge our collaboration with Edinburgh Uni-
versity Press and in particular we would like to thank Archie Turnbull
and Allan Woods for their support and help.
John Bintliff
Donald Davidson
Eric Grant
 Introductory Remarks: Spatial and Temporal Contexts

Robert McC. Adams

Secretary, The Smithsonian Institution

Relationships between human culture and its environmental setting

are among the oldest, most solidly grounded concerns of archaeology.
They have been a primary preoccupation of many archaeologists,
whose claims to have made major contributions to an understanding
of the complex interactive processes that are involved are strong and
well-recognized. Both the preoccupation and the claims are rein-
forced by special qualities of archaeological data that have been
greatly enhanced in recent years in precisely an environmental direc-
tion. They also reflect the unquestioned responsibility of the discipline
for the material residues of human behaviour, over spans of time
and space that are coextensive with the full sequence and distribution
of hominid species. How, then, do we justify returning yet once more
to such a familiar subject, and in particular to the substantial re-
thinking of it implied by the phrase 'conceptual issues'?
The primary justification is not to be found in what has certainly
been an impressive accumulation of detailed, local knowledge, nor
in our rapidly expanding capabilities for the recovery of new types
of evidence. More fundamental, I believe, are numerous - more
epistemological than substantive - questions that have increasingly,
and for the first time rather explicitly and self-consciously, come to
the fore. A representative, by no means exhaustive, sample of them
might include the following:
Are there discoverable regularities in human behaviour,
arising from the operation of strictly limited numbers of external
forces or prime-movers? To what extent can we justify a focus
on adaptive or economically 'rational' human behaviour alone,
stripped of culturally-imposed patterns of association, percep-
tion and meaning? Is there a valid distinction between basic
patterns of adaptation and the local complexities or micro-
dynamics of process in which these are manifested? Is there a
valid distinction between 'infrastructures' and 'superstructures',
however defined? If there are teleological traps in most 'func-
tional' explanations, what do archaeological data permit us to
substitute for them? What is the role in explanation, if any, of
2 ADAMS
historical and ethnographic analogies? How conclusively, given
the limitations of archaeological evidence, can most propositions
presently be 'tested'? Apart from anthropology, does history or
biology provide a more useful disciplinary role-model?
Such a listing can be shortened or indefinitely lengthened. But the
active controversy that continues to surround such questions surely
justifies our further attention to the theme of environmental
archaeology. While the basic terms of the debate may indeed have
become widely familiar, virtually every methodological or theoretical
advance casts them in a new light.
As one who works on synthetic problems occupying historical,
geographic, ethnographic and archaeological borderlands, I admit
to strong opinions about many of these questions. Doubtless there
are almost as many shadings of individual position as there are
practitioners. For present purposes, however, differences in stand-
point need not detain us. More crucial to what I have to say is the
proposition that there is not a large body of what can be regarded
as incontrovertible 'facts' in archaeology. Most ofthat body, further-
more, is of strictly local and descriptive, (i.e., comparatively minor)
importance. Facts we would count as significant, on the other hand,
have achieved this status only in connection with debatable and
contingent ordering principles. Although there are still many
archaeologists who will denounce this approach, it corresponds
closely to a pronounced, concurrent drift away from a positivistic
posture in the social and behavioural sciences generally. As a recent
review of dominant trends in these fields over a half-century observes,
' facts are now understood as compelling interpretive statements
reached by comparing the results of more-or-less precise measure-
ments undertaken within a theoretical scheme' (Gerstein 1986:5).
That is what I believe our debate is about: alternative theoretical
schemata.
But while focusing on theoretical or conceptual issues, we cannot
ignore what may have become almost a truism, that much of scientific
progress results from our advancing capacities for measurement.
Precision of quantitative measurement is only one of these capacities,
although an important one. Others, that I hope to show are more
relevant for this discussion, include the development of cross-cutting
and complementary units and scales of observation. Before discuss-
ing in greater detail the importance of these latter for environmental
archaeology, perhaps I can suggest their character and importance
by referring briefly to an entirely different scientific field.
Astronomy is, I believe, a more useful illustration than may be
initially apparent: Both fields deal with exceedingly large aggregates
of observational detail, all of which is remote from our closer
inspection on our own terms and for the most part not subject to
 Introductory Remarks: Spatial and Temporal Contexts 3
experimental manipulation. Moreover, the ordered constructs and
explanations sought by both fields are as uniformitarian and
evolutionary in format as the evidence will permit.
Astronomy relies almost wholly on a relatively-small number of
great observational platforms that in principle can be devoted to a
wide variety of different problems by different astronomers. This may
seem to imply a single overall path of development within the disci-
pline, consisting of successive, increasingly difficult and costly efforts
to increase mirror diameters and thus light-collecting power. And
indeed there has been a trend toward progressive increases in the
scale and cost of instrumentation, in order to increase the variety
and precision of measurements as well as to probe farther toward
the rapidly receding margins of the universe. However, the dominant
trend has not been toward instruments that are mere replications of
one another at a larger scale. Primary interest has instead been
devoted to identifying new phenomena, or aspects of phenomena -
opening what are, in effect, alternative and complementary windows
of observation by employing increasingly specialized instrumentation
(cf., e.g., Astronomy Survey Committee 1982).
Step by step, for example, the spectrum of observation has been
advanced in both directions beyond the minor portion occupied by
visible light, until it now extends from radio waves to very hard
gamma rays. Measures of the flux of elementary particles now supple-
ment those of electromagnetic radiation. Greater accuracy in calcula-
tions of astronomical distance is being sought by innovative designs
that greatly increase the observational baseline. An essentially con-
stant typology of astronomical objects and processes, founded on
the assumption of prevailingly very slow rates of change, is being
modified as the search intensifies for transient and oscillatory
phenomena that are only detected by varying observational spans
from micro-seconds to years. And increasing attention is given to
phenomena, such as "black holes" and inter-galactic matter, whose
existence cannot be directly detected at all but must be established
through complex chains of secondary inferences.
There is, of course, no direct archaeological counterpart for the
great telescopes on which astronomy relies, nor for many of the
institutional arrangements that these presuppose. However, the
underlying pattern of scientific advance in astronomy has significant
parallels with this symposium and the subject it reflects. Both are
marked, for example, by systematic efforts to extend and diversify
perspectives, relying on a conscious selection of alternate scales of
observation (a theme directly addressed in this volume by Modder-
man and Slapsak) or on phenomena whose archaeological relevance
has generally gone unrecognized (Brothwell, Boyle & Wright). Both
critically scrutinize conceptual categories or units of observation that
4 ADAMS
may no longer be optimal for current problems (cf. Chapman on the
absence of absolutes in the definition of resources or their scarcity,
Bayliss-Smith on the definition of altitudinal boundaries, and Deith
& Shackleton on the absence of a systemic shift in subsistence
behaviour accompanying the introduction of domesticates). Both
have found they can uncover rewarding new puzzles through increas-
ing precision of measurement or its statistical analysis (Green, Whit-
tington). And both have become fundamentally dependent on chains
of indirect inference in order to penetrate to latent, underlying
patterns from the limited observational data that is directly accessible.
The changing ways in which we have thought about spatial units,
both 'real' (or empirically derived) and constructed for analytical
purposes, are a particularly salient illustration of the parallelism.
Archaeological sites, to take a most basic unit, traditionally were
regarded as self-evident loci of relatively-concentrated activity of
some sort. Having no (easy) theoretical option, as the Gaffneys
observe, archaeologists assumed that sites will 'somehow rise above
any background noise and will be instantly recognizable as such.' In
that frame of thought, minimal size and duration thresholds were
non-problematic. Sites that had been destroyed or were not readily
discoverable were also of little interest, since excavation was virtually
the sole, consensually recognized end-product.
As such, sites corresponded to the traditional unit of ethnographic
observation, the community 'of any convenient size' (Leach 1965).
Both were, for some generations, generally chosen for excavation or
resident participant-observation on the basis of fairly idiosyncratic
or pragmatic considerations. They furnished the subject-matter for
monographic studies, meeting fairly standardized disciplinary and
professional expectations (reinforced by publishers and granting
agencies). Both sites and communities are now regarded more scepti-
cally, contextually, and flexibly, principally as a result of the devotion
of greater attention to patterning and interaction at the regional level.
The advent of 'site' surveys has contributed much to this develop-
ment in archaeology. It is not the case, however, that numerous sites
of varying size and composition have simply been enclosed by surveys
in analytical units of much larger dimensions, without modification
of the assumptions made about them. The process of surface recon-
naissance has instead gradually altered the significance attached to
sites, as it has exposed the arbitrariness of traditional site designa-
tions. In surveys, most site boundaries have to be drawn on the basis
of fall-offs in artifact density. Even site presence or absence has to
be linked to essentially arbitrary demarcations of size or density. The
density of sites recovered by surveys has therefore turned out - hardly
a matter of surprise - to be a dependent function of intensity of
coverage (Bintliff and Snodgrass 1985: 128), although of course a
 Introductory Remarks: Spatial and Temporal Contexts 5
very large proportion of the additional numbers of sites that system-
atic foot-coverage produces are quite small. Small and 'limited-
activity' sites then become a focus of interest in themselves (e.g.,
Ward 1978), beginning with recognizable categories like temporary
field-dwellings and plot-boundaries and ending with apocryphal 'pot-
drops' that only half-facetiously define the asymptote of the lower
tail of the distribution.
Presently it becomes apparent, as the Gaffneys' work again illus-
trates, that functionally-differentiated continua of activity are in no
sense limited to sites but extend irregularly across virtually any region
we study. Sites may be defined more narrowly to fit specific, problem-
related objectives, for example as (permanent? seasonal?) nucleated
areas of co-residence exceeding certain size and density thresholds.
Thus systematized, they provide data for regional demographic calcu-
lations, while being sufficiently tangible to justify excavation by those
who wish to 'test' such calculations. In this largely-unintended fash-
ion, sites that were originally thought of as distinctively individual
foci of interest in their own right have gradually been transformed
into operational constructs, alternately sub-classified and aggregated
according to consciously uniformitarian principles and employed as
intermediate units in an increasingly diverse range of problems.
Meanwhile, other manifestations of an ideal-typical approach to
sites continue to be found useful, perhaps even to flourish. They take
the form of'village modules' and site catchment analysis, and detailed
attention is devoted to both in the following pages. Hence any general
comments here risk superficiality. Yet the complex and conflicting
issues associated with them cannot be wholly avoided, since, in one
form or another, they are virtually at the heart of any form of
environmental analysis.
We can set aside, if only for present purposes, the difficulties and
uncertainties involved in reconstituting ancient resource availability
from its modern derivatives. In the abstract, having done so, it is
surely justifiable to systematize the collection of information about
critical subsistence and other resources. Relative proximity to settle-
ments then serves as the obvious, almost inevitable, ordering prin-
ciple. Let us grant that, in any given example, the extent of influence
of non-economic, or non-local, factors in the determination of site-
size and location is an empirical question. However, 'rather than
dismiss the fact that human groups very often do not maximize
resource exploitation, one should use site catchment analysis to assess
deviations from economically rational behaviour' (Gilman and Thor-
nes 1984: 6).
So much for the positive side of the argument. On the negative
side, we must recognize that any social reality involves variable rather
than standardized units, whether or not we are able to detect and
6 ADAMS
quantify the differences. By theorizing in terms of modules, moreover,
we drift into an acceptance of not merely ideal but equilibrium
models, resting on the - grossly unlikely - assumption that stable
relationships within and between settlement units were the norm over
long periods. Even Bintliff, after arguing cogently that 'some form
of catchment analysis' is a 'most useful analytical tool,' finds in his
own Boeotian survey that 'at no time do we seem to have a predict-
able match or semi-equilibrium of settlement and population paired
to potential resources.'
Beyond such empirical grounds for scepticism is Malcolm
Wagstaff's more theoretical rejection of 'the tyranny of economic
distance,' on the grounds that this and other forms of rational utility-
maximization must have been absent prior to capitalism. To me at
least, that position seems too sweeping, particularly since archaeo-
logical time-scales encourage us to think less in terms of conscious
strategies than of very-gradually evolving behaviour patterns whose
rationales may have remained largely unrecognized. Marxist his-
torians are among those offering powerful demonstrations that, even
over comparatively short periods of time in pre-capitalist Europe,
the major economic indicators 'functioned rationally' in relation to
one another (Bois 1984: 393).
More telling is Wagstaff's observation that behaviour has probably
always been discontinuous across 'critical thresholds' whose position
varies according to cultural and historical context. As he also insists,
site territories must always be considered within a web of functional,
hierarchical and political as well as merely spatial and economic
relationships. Thus we must recognize that forces of differentiation
were continuously at work in virtually all circumstances. Those of a
socioeconomic character may be most readily documented in the
archaeological record, but were not for that reason necessarily always
primary. In any case the crucial point is that comparative advantage,
besides being often transitory, does more than breed a larger or better
endowed individual module. It also alters the relations between that
module and others, generating conflict, exploitation and enhanced
asymmetries of power as well as indeterminancies of outcome.
Beyond differences that are fixed, or based on averages, most
threats to subsistence vary greatly in their frequency and severity
even among closely-neighbouring communities. Non-local prestige
goods have been 'shown to circulate in some areas as a presumed
fund of credit against such local calamities (Flannery 1968, Bettinger
and King 1971), but with primitive transport and low levels of surplus
the dampening effect on the most severe oscillations in subsistence
could never have been large in preindustrial settings. Hence many
forms of environmental variability have always reinforced intercom-
munity inequality, to an extent like that which can be abundantly
 Introductory Remarks: Spatial and Temporal Contexts 1
documented in late-medieval and early-modern Europe. For these
reasons, I am reluctant to see in an arrangement of dots on any
archaeological period-map a network of 'modular' villages whose
differences were essentially no more than products of their different
catchments.
I have the impression that the theme of regional resource differenti-
ation, interdependency, and subsistence complementarity has
received more sustained attention in New World archaeology than
in European prehistory (e.g., Byers, ed., 1967, Wolf, ed., 1976, San-
ders, Parsons and Santley 1979, Masuda, Shimada and Morris, eds.,
1985). Does this indicate a greater propensity for model-building, or
greater readiness to move away from an earlier preoccupation with
sites and toward patterns of macro-adaptation and interaction? Poss-
ibly, but there may be also a more prosaic explanation: much Ameri-
canist archaeological theory is formulated around the characteristics
of the New World Civilizations, which happened to occupy zones
in which great and abrupt topographic (and corresponding climatic)
differences led to great variability in subsistence resources over very
short distances.
Modules whose variability is conceived of as largely a product of
the differences in their catchments also invite attack from the opposite
direction. Within sites or communities as well as between them, illness
and death strike haphazardly. Numbers and sexes of offspring vary,
with differentiating consequences for extended families as well as
for individuals, as these variations intersect with marriage and inherit-
ance practices. Families rather than villages may have had their
more-or-less advantageous lateral ties of intermarriage, and archi-
pelagos of resource-exploitation (as Davidson notes was brought out
during the discussion). To be sure, these differentiating and destabiliz-
ing factors within individual communities may often have been
counterbalanced to some extent by customary levelling, exchange
and redistributive mechanisms. But the continuous pressures in the
opposite direction must have been strong, and the effectiveness of
such mechanisms needs to be shown rather than merely taken for
granted.
The extent of socio-economic variability, both at the intra-and
inter-community level, thus emerges as a clearly crucial question. It
is also eminently researchable with archaeological data. Hence it is
disappointing that little attention has been devoted to summarizing
what is known of this on a regional basis. For individual household
inventories, the only study focusing directly on this problem of which
I am aware deals with the lowland Maya area (Tourtellot 1983). It
disclosed variability so wide as to defy ready theoretical explanation
- or sampling strategy. The major potential deterrent to such work
in most areas, as is implied by the location of this study in a region
8 ADAMS
favored by exceptional conditions for the surface identification of
house-mounds, arises from the fact that valid comparisons are most
readily made between complete units. In most regions there probably
are few such units - even individual houses, but, a fortiori, whole
settlements - for which information of comparable completeness and
quality is available. Given the importance of the issue, further
attempts to assess such variability would be extremely valuable even
if rigorous standards of comparability could not be maintained.
Further difficulties are added to the problem of site or settlement
modules when, with the advent of 'chiefdoms,' states and civiliza-
tions, we no longer deal with grids of presumably independent
communities but with territorially extensive hierarchies of com-
munities. Settlement size, facilities, and internal division of labour
are then functions, if central place theory has any force at all, of
socioeconomic and administrative relationships to larger and smaller
members of the same hierarchies. Documentary evidence, when it
appears, introduces other and usually more transient variables: inter-
nal cohesiveness of the regime and its elites, productivity and fragility
of various forms of agricultural intensification, relative strength of
external rivals, or success at pacifying a restless, not fully sedentized
countryside. Documentary evidence also can introduce the confusing
factor of cognitive distinctions in settlement hierarchy (or the lack
of them) that fail to match the behavioural residues to which the
archaeologist's occupation is limited.
The significance of relations between small prehistoric com-
munities must of course be established on the basis of contemporary
evidence. It cannot be taken for granted merely on the basis of
analogy with later, extensive networks comprised of alternately trad-
ing and contending city-states, or with urban hierarchies in still later
imperial regimes. But why should we be more sceptical of overstated
claims in this direction than of Bintliff's downgrading of the import-
ance of long-distance trade in most pre-industrial settings? Perhaps
our views of where to strike a balance in this respect are formed by
the regional settings with whose history (or prehistory) we are most
familiar. What is striking for me is that the Near Eastern documentary
evidence, when it becomes available in quantity, indicates a vastly
greater emphasis on long-distance external relations than would be
suggested by the archaeological evidence alone.
Assur, a city-state of moderate size in northern Iraq from the third
millennium BC onward, is a case in point. With a quite-inadequate
sustaining area of its own, its caravans, at least for a time, traded
tin, textiles and precious metals with a substantial part of central
Anatolia, 800 or more kilometres distant. The tin may well have been
mined in Afghanistan, while the textiles were probably produced in
Babylonian towns of southern Iraq. Assur's primary activity is, it is
 Introductory Remarks: Spatial and Temporal Contexts 9
generally agreed, virtually invisible in archaeological remains and
known only through accompanying cuneiform texts found in its
Anatolian outposts. It was, in effect, a highly-specialized middleman
(Veenhof 1972, Larsen 1976). As such, it becomes understandable
only within a very large-scale pattern of inter-regional relations. Does
this not at least suggest what may be missing in many of our analyses
of small prehistoric communities as well?
Let me turn from spatial to temporal units. Again, they can be
regarded as a set of alternative windows providing different insights
on interactional processes involving the environment. Which window
we choose to look from is a function of the scale of the phenomena
that concern us and the questions we want to ask about them.
Periodization, as historians long preceded us in discovering, is ulti-
mately not a factual but a conceptual issue.
There is a natural tendency to demarcate archaeological periods
by abrupt environmental shifts. Such shifts frequently have the advan-
tage of providing seemingly less inconsistent and more broadly appli-
cable regional chronologies than if cultural markers or absolute
dating systems are used. There are, however, two problems. First and
more obvious, a chronological framework based on environmental
disjunctions can develop into causal explanations in which those
disjunctions are uncritically assumed to play a primary role. Secondly,
there is a tendency to assume more consistency and systematicity in
the patterning of environmental change than may be justified.
This issue is given welcome emphasis by several contributors.
Moore and Denton refer to the notorious local variability of soils, a
challenge for sampling methodologies that undermines many gen-
eralizations as to yield, carrying capacity, and density of human use.
Forest community succession, as Stanton Green reminds us, is not
infrequently a locally determined, cyclical and highly-complex
phenomenon rather than a passive reflection of climatic change.
Similar basic issues are involved in Thornes' discussion of metastable
states of soil erosion and vegetation cover under varying intensities
of grazing, and in the cultural relativem for which Chapman argues
in considering alternative models of hunter-gather carrying capacity.
In a somewhat different vein, Burrin and Scaife employ catastrophe
theory to interpret complex and extensive alluviational changes that
may be triggered by small changes of a particular variable at a
sensitive threshold.
The cumulative effect of such studies is to leave us with a heightened
sense of uncertainty about how to employ evidence for environmental
change in explanatory archaeological models. Reconstructing aggre-
gated patterns of change from small and fragmentary clues -
frequently the only kinds available - is plainly hazardous. Interpreta-
tion of the rates, and sometimes even the directions, of the processes
10 ADAMS
involved becomes a highly demanding as well as subtle study of its
own. With these uncertainties, it seems to me, should go the adoption
of a more critical and selective approach to the use of environmental
data as a basis for archaeological periodization of any kind.
In the particular case of archaeology, it is important to remember
that many of the alternative temporal windows that were referred to
earlier remain firmly shuttered and unavailable. Save with the aid of
texts, precision in the assignment of absolute dates, or even in securely
establishing relative sequences, is seldom possible. Blocks of time
covering four to eight generations, or even longer, remain our ele-
mental building units under most circumstances. As Bintliff properly
observes, chronological controls in archaeology are thus 'generally
a coarse-meshed sieve.'
Seeking to make a positive advantage out of this shortcoming,
some theorists deny the importance of what they have not, as yet,
found a way to discern in the archaeological record. Lewis Binford,
for example, detaches an implicitly static or slow-moving cultural
system from 'all the events, behaviours, or other transactional
phenomena going on at a given time,' and highlights only the former
as 'the conditioning framework within which all these events trans-
pire' (1981:201). But there can be only an artificial separation
between the category of event and the category of historically-condi-
tioned framework. Societies are continuously reproduced and trans-
formed through their - always imperfect - instantiation in an ongoing,
irregular flow of crises and events. Nothing dictates that the relatively
slow, as well as uncertainly bounded, pace of change perceptible
through the use of the lengthy temporal units is somehow more
grounded in eco-systemic realities than processes that move on
shorter time-scales.
There are more consequences of the length and uncertainty of
archaeological time-scales than merely an excessive emphasis on
gradualistic rather than punctuated change. In constructing regional
syntheses on the basis of data from many individual sites, we are,
in effect, seeking to compose a valid temporal and developmental
ordering out of hypothetical sequences of snapshots. The underlying
assumptions and hypotheses can never be made fully explicit in what
is essentially a creative process, but there is every reason to suspect
a pronounced bias toward over-determined outcomes.
David Freidel addresses this point in a provocative article on
lowland Maya political systems. What most archaeologists have
tended to reconstruct as an evolving succession of ancient settlement
patterns in that region, he suggests, might instead be a shifting,
unstable equilibrium in which heterogeneous local polities coalesced
and dissolved over time. Even if the field as a whole is no longer
involved in an explicit search for stage-like successions, there remains
 Introductory Remarks : Spatial and Temporal Contexts 11
a tendency, when there are choices, to select simple-to-complex
sequences as the most plausible. Alternatively, he invites us to view
the totality of Maya civilization 'not as a society evolving through
stages, but rather as a mosaic of societies of shifting size and composi-
tion, constantly adapting to one another It is variability mediated
by porous boundaries through which individuals and whole com-
munities are passing' (1983: 376).
Freidel's interpretation remains tentative - some would say specu-
lative. It finds an analogue in Edmund Leach's well-known work on
highland Burma (1965), save that Freidel emphasizes more strongly
the adaptive, intercommunal aspects of the oscillations in structure
and composition that he hypothesizes. These would have been more
difficult for Leach to identify since, as a social anthropologist, his
capacity to observe the interplay of action and reaction across a
whole region was strictly limited. But Leach's study did depend upon
being able to observe micro-dynamics of process in 'real time'. And
so - this is its significance for archaeologists who are limited to
'archaeological time' - does Freidel's. What alone makes such an
interpretation possible is the availability to Freidel of dated long-
count Maya monuments. Regrettably, there are not enough of them
yet known, particularly from smaller sites, to elaborate or fully
confirm the patterns of instability he posits. But the example does
indicate that the absence of assured, short-range dating alternatives
has more subtle and pervasive effects than archaeologists usually
realize.
This lesson also emerges from recent work in another region where,
surely not coincidentally, there once again happens to be a dating
system with potentialities for recording on a year-to-year basis. I
refer to the southwestern United States and dendrochronology,
readily conceding that, for various technical reasons, the chronologi-
cal framework is in some respects unrepresentative and otherwise
deficient. The new findings to which I refer, moreover, may well be
dismissed in some quarters as still too insecurely based to be useful.
Such a reaction is particularly understandable when the resultant
shift in interpretation, if it is agreed that we need to make it, will be
so abrupt as to constitute something close to a 'paradigm shift' in
the Kuhnian sense (Khn 1970). My central point, however is largely
independent of the validity of the new findings. Instead it concerns
the apparent linkage between the increased precision (or perhaps
even just the perceived potentialities for increased precision) of a
particular dating system and the widened array of hypothetical recon-
structions that archaeologists can at least plausibly introduce.
Working with distributions of dendrochronologically well-dated
sites, Fred Plog(1984) calls attention to a number of apparent hiatuses
in settlement. He finds that there are two general hiatuses throughout
12 ADAMS
the region, one in the later tenth century AD, and a longer, relatively
more pronounced, one extending from the mid-twelfth until almost
the mid-thirteenth century. With C14 dating alone, we could at
most be dimly aware of this phenomenon, and would lack the
incentive even to look for it - let alone to try to explain it. Prevailing
assumptions about continuity would instead lead us to smooth away
any apparent gaps in the sequence. At some point, his presen-
tation implies, we must recognize that the weight of the evidence no
longer justifies the reconstruction of the Southwestern prehistoric
record on the basis of an assumed maximal continuity for sedentary
occupance.
Re-studies of Southwestern survey data by Steadman Upham
(1984) reinforce this impression. A strikingly high proportion of the
sites that have been found and recorded are artefact scatters and
'limited activity' sites that cannot be assigned a clear cultural affili-
ation or date. Southwesternists find themselves in the uncomfortable
position, Upham argues, of defining their phase sequences on the
basis of a minority of material - that from nucleated sites representing
a clearly sedentary adaptation. The alternative favoured by both Plog
and Upham is that this was only one alternative within a larger set
more generally characterized by adaptive diversity. What first appears
as a striking series of local and regional abandonments can then be
interpreted instead as periodic extreme episodes within a virtually
continuous series of adaptive shifts.
Any science involves systems of constraints. Archaeologists elect
(or are accustomed, or are able) to use only certain scales or
frameworks of spatial and temporal analysis. Paralleling an interest
expressed in this volume by Modderman, I have been principally
concerned with the relationship between how we manipulate those
frameworks and our preferences for or against certain research
approaches, implicit assumptions, and judgements as to worthwhile
problems or priorities. That is a relationship of fairly general sig-
nificance. But it has special relevance when we consider the ways in
which archaeologists conceptualize the fit between their primarily
sociocultural subject matter and its environmental setting.
C. S. Holling, who has worked creatively on a variety of environ-
mental systems as well as on human-environmental interactions, has
recently offered a brief sketch of several alternative perceptions of
causality in a field of interacting behaviours (Holling 1984). Two are
particularly suggestive of consistent trends of thought in archaeology.
The first is undoubtedly closer to the one with which the greater
part of the archaeological enterprise has been traditionally - perhaps
still is - associated. Moreover, aspects of it are also to be found in
an anti-traditional context, having been incorporated (in my view
mistakenly) into the program of at least some 'new' archaeologists.
Above all, this first view is equilibrium-centered, dominated by
 Introductory Remarks: Spatial and Temporal Contexts 13
cause/effect determinism, by an overridingly quantitative emphasis,
and by a paradigm of stability, order and regulation. Particularly in
its more recent manifestations, it is also rooted in reductionism and
in the emphasis on partitive analysis leading to the formulation of
general laws on the basis of putatively-crucial tests or experiments.
While acknowledging external perturbations, it concentrates on
portraying feedback-loops that constitute a largely self-enclosed
field. Yet an equilibrium-centered view can only be maintained, as
Holling argues, by introducing explanatory variables of greater and
greater complexity, while refinements of such specifications do
nothing to assure that behaviour will fall within the same parameters
under altered conditions. With special reference to the obvious
archaeological applications of this paradigm, I would add that it
strongly favours system-maintenance over change of of any kind;
and change that, when it does occur, is predictable, gradual and
unidirectional.
The second paradigm, in my judgement, coincides more closely
with the prevailing orientation of this symposium; it also finds force-
ful support in the substance of Holling's own work. It stresses the
prevailing importance, in all of the environments with which our
species has ever had to contend, of a very nearly opposite set of
characteristics: instability, transience, indeterminacy of outcome,
reversibility - all understandable as corollaries of systems that are
open and poorly articulated. The contrast of function and dysfunction
virtually disappears, save from narrowly specified, essentially
artificial standpoints. Conflict and variance are not aberrant
symptoms of system breakdown, but common, occasionally-preva-
iling modes of operation. The challenge of human adaptation
throughout prehistory, therefore, was somehow to work out strategies
of resilience rather than stability or equilibrium, meeting 'the need
to keep options open, the need to view events in a regional rather
than a local context, and the need to emphasize heterogeneity' (Holl-
ing 1973: 21).
There may well be alternative paradigms that are more uniformly
followed or more encompassing. Having cast doubt upon ideal
models and neatly self-enclosed systematicity, I cannot plausibly
assert the congruence of all the components mentioned above, always
and exclusively, in just these two specified patterns. But I think they
at least suggest the many resonances of the work at hand, beyond
the environmental realm to which it is ostensibly devoted. From the
constructively different vantage points of its contributors, this sym-
posium grapples with some of the most general, most rewarding
questions on which archaeology - characteristically an unveiler of
new complexities, not a propounder of certainties - offers at least a
glimmer of cumulative progress.
14 ADAMS
BIBLIOGRAPHY
Astronomy Survey Committee 1982 Astronomy and astrophysics for
the 1980's, vol. 1. National Research Council. Washington,
D.C.: National Academy Press.
Bettinger, R. L. & T. F. King (1971) Interaction and political
organization: A theoretical framework for archaeology in the
Owens Valley, California. Annual Report, Archaeological
Survey. Department of Anthropology, University of
California, Los Angeles.
Binford, L. R. (1981) 'Behavioral archaeology and the "Pompei
premise",' Journal of Anthropological Research 37, 195-208.
Bintliff, J. and A. M. Snodgrass (1985) 'The Cambridge/Bradford
Boeotian Expedition: The first four years', Journal of Field
Archaeology 12, 123-61.
Bois, G. (1984) The crisis of feudalism: Economy and society in
eastern Normandy c. 1300-1550. Cambridge: Cambridge
University Press.
Byers, D. S., ed. (1967) The prehistory of the Tehuacan valley, 1.
Environment and subsistence. Austin: University of Texas
Press.
Flannery, K. V. (1968) 'The Olmec and the Valley of Oaxaca:
A model for inter-regional interaction in Formative times', in
E. P. Benson, ed., Dumbarton Oaks Conference on the Olmec,
pp. 79-110. Washington, D.C.: Dumbarton Oaks Research
Library.
Freidel, D. A. (1983) 'Political systems in lowland Yucatan:
Dynamics and structure in Maya settlement', in E. Z. Vogt
and R. M. Leventhal, eds., Prehistoric settlement patterns:
Essays in honor or Gordon R. Willey, pp. 375-86.
Albuquerque: University of New Mexico Press.
Gerstein, D. R. (1986) 'Introduction', in N. J. Smelser and D. R.
Gerstein, eds., Behavioral and social science: Fifty years of
discovery, pp. 1-17. Washington: National Academy Press.
Gilman, A. & J. B. Thornes (1984) Land-use and prehistory in
south-east Spain. London Research Series in Geography, 8.
London: Allen and Unwin.
Holling, C. S. (1973) 'Resilience and stability of ecological
systems', Annual Review of Ecology and Systematics 4, 1-23
(1984) 'Perceiving and managing complexity', International
Institute for Applied Systems Analysis, Options 1984, no. 3,
p. 16.
Khn, T. S. (1970) The Structure of Scientific Revolutions. 2nd ed.
Chicago: University of Chicago Press.
Larsen, M. T. (1976) 'The Old Assyrian city-state and its colonies',
Mesopotamia 4. Copenhagen: Akademisk Forlag.
Leach, E. R. (1965) Political systems of highland Burma. Boston:
Beacon.
Masuda, S., I. Shimada & C. Morris, eds. (1985) Andean ecology
and civilization. Tokyo: University of Tokyo Press.
Plog, F. (1984) 'Exchange, tribes, and alliances: The northern
Southwest', American Archeology 4, 217-23. [See now in
addition Plog, F. in press, 'The Sinagua and their relations',
in L. S. Cordell and G. J. Gumerman, eds., Dynamics of
Introductory Remarks: Spatial and Temporal Contexts 15
southwest prehistory. Washington: Smithsonian Institution
Press.]
Sanders, W. T., J. R. Parsons & R. S. Santley (1979) The basin of
Mexico: Ecological processes in the evolution of a civilization.
New York: Academic Press.
Tourtellot, G. (1983) 'An assessment of Classic Maya household
composition', In E. Z. Vogt and R. M. Leventhal, eds.,
Prehistoric settlement patterns'. Essays in honor of Gordon R.
Willey, pp. 55-76. Albuquerque: University of New Mexico
Press.
Upham, S. (1984) 'Adaptive diversity and Southwestern abandon-
ment, Journal of Anthropological Research 40, 235-56.
Veenhof, K. R. (1972) Aspects of Old Assyrian trade and its ter-
minology. Studia et Documenta ad lura Orientis Antiqui Per-
tinentia, 10. Leiden: E. J. Brill.
Ward, A. E., ed. (1978) Limited activity and occupation sites: A
collection of conference papers. Contributions to Anthropologi-
cal Studies, 1. Albuquerque: Center for Anthropological
Studies.
Wolf, E. R., ed. (1976) The valley of Mexico: Studies in pre-
hispanic ecology and society. Albuquerque: University of
New Mexico Press.
 Section 1
Conceptual Issues at the Site and
Microregion Scale
 Introduction

Donald A. Davidson, University of Stirling

Too often environmental archaeology is viewed as a set of specialist

techniques culled from the earth and biological sciences and applied
to archaeological investigation. The usual emphasis in environmental
archaeological research is towards a reconstruction of palaeoenviron-
ments with former diets and economies also being proposed from
the study of seeds, bones and pollen. The results of environmental
research are often presented in separate chapters, or worse still in
appendices to excavation reports. This means that environmental
analysis is perceived as being separate from the principal archaeologi-
cal objectives in excavations. Although there are many site situations
where specialist environmental analysis can be of great help in aiding
archaeological interpretation, a key issue throughout this book is the
plea for environmental research to be pursued in partnership with
archaeology.
Of course there is a long and distinguished tradition of environ-
mental investigation in archaeology. Approaches to the changing
relationships between man and his environment have been stimulated
by concepts drawn from geography, ecology and ecological
anthropology. The position is summarised by Jochim (1979: p77) as
follows:
Ecological research is thriving in archaeology. The field has
seen a proliferation of new techniques for gathering information
about past environments, an increasing research emphasis on
the nature of human interaction with these environments, and
the adoption of generalizations, concepts, and jargon from bio-
logical ecology.
In his review Jochim criticises the rather limited focus of ecological
anthropology with concern primarily given to relationships between
subsistence economies and natural environments. Instead he
emphasises the importance of a human behavioural approach in
terms of trying to understand the dynamics of sociocultural systems.
A behavioural approach is now common in other areas of archaeology
20 DAVIDSON
whereby former cultural, social or economic processes can be pro-
posed on the basis, for example, of ceramic attributes, or artefact
and site distribution.
A strong case for an ecological approach to archaeology is made
by Butzer (1982) who discusses varying ecological relationships at
different spatial and temporal scales. This leads him to provide a
synthesis of general principles governing settlement distribution by
discussing the processes of continuity and cultural change. As already
indicated, the central concept is behavioural response and this
emphasises the importance of interpreting archaeological evidence
within the context of specific societies. Butzer calls this the contextual
approach. The papers throughout this book demonstrate the applica-
tion of such a paradigm to a wide range of archaeological questions.
The first three papers discuss conceptual issues raised during the
analysis of soils, pollen and shells in and near archaeological sites.
The remaining papers in the first section are concerned with the
analysis of the relationship between sites and their local- or micro-
region.
The chemical analysis of soils from archaeological sites has been
done for many years, often with a particular emphasis on phosphorus
in order to identify former areas of occupation. The chemical loading
by anthropic processes on sites ought to be investigated with exactly
the same emphasis as other artefactual remains. Such a view is implicit
in the paper by Moore and Denton, who discuss the main issues
associated with soil-chemical properties of sites in northern Quebec.
They tackle the key issue of soil variability, which is always par-
ticularly marked with chemical properties. Through detailed sampling
and the use of control samples, they show that the statistical distribu-
tions especially of phosphorus and to lesser extents of calcium,
magnesium and pH, can be interpreted to indicate the presence of
hearths. They stress the value of such analysis since the results can
indicate activity or refuse-areas which could not otherwise be
identified.
In the latter part of their paper, Moore and Denton apply the
technique of inorganic phosphorus fractionation as adapted by Eidt
and Woods (1974), Woods (1977), and Eidt (1984). The original
method was designed to aid the identification of different land use
and settlement patterns; Moore and Denton assess the extent to
which ratios of the phosphorus fractions can be taken to indicate
relative age of hearths. Although they conclude that many difficulties
surround such an interpretation, their research on Quebec sites illus-
trates how data on soil chemistry can help with the identification of
sites and activity areas.
The analysis of soils on sites can be compared with provenancing
ceramic or lithic materials. In such research priority is given to the
 Introduction 21
determination of a wide range of chemical elements and then some
form of discriminant analysis is performed to identify groups which,
it is then suggested, to have common sources. In the analysis of soils
from archaeological sites the tendency has been to search for par-
ticular chemical indicators of specific human activities, but with the
exception of phosphorus, this has been of limited success. An alterna-
tive approach is to follow the provenancing technique of searching
for groups on the basis of a range of chemical imprints. As an
example, Quine (1987) uses cluster analysis to identify groups based
on six chemical measures and these groups are shown to correlate
with specific site formation processes. An additional advantage of
such an approach is that the effect of individual element variability
is markedly reduced.
The paper by Whittington also follows a long research tradition
in that his analysis of soils and associated pollen is directed to
environmental reconstruction. In this Scottish example from the Isle
of Jura, soil pollen was extracted from a palaeopodzol from under
an enclosure as well as from outwith the site. Discussion of the pollen
species in relation to the buried soil morphology raises interesting
points such as different phases of site exploitation, the long presence
of arable activities and the suggestion that fouled litter was added
to the land as manure. Again, results such as these demonstrate how
the overall interpretation of the site can benefit from such environ-
mental research.
A similar theme emerges from the paper by Deith and Shackleton,
who assess the extent to which shell data can aid the postulation of
behaviour patterns of former occupants of the Franchthi cave in
Greece. The changing nature of the shoreline provides a backcloth
against which the changes in species of shell are interpreted with
reference to the cultural sequence. Of particular interest are the
processes associated with the Mesolithic-Neolithic transition. The
authors support the view that this economic change evolved across
a porous frontier between hunter-gatherer and farming groups.
Although the authors are careful not to over-interpret the shell data,
their paper is a neat example of how the reconstruction of cultural
processes can be aided by environmental information.
The remaining papers in this section focus on conceptual issues
concerning the interaction of sites with their surrounding areas.
Modderman begins by stressing the close philosophical link between
geography and archaeology. Archaeologists are used to modelling
change on a range of timescales, but such an ability is far less common
in dealing with differing spatial scales. In elaborating this theme,
Modderman points out that every human activity has its own charac-
teristic spatial scale. The consequence is that spatial analysis in
archaeology needs to be carried out at relevant scales. Just as different
22 DAVIDSON
patterns can be established by looking at a microscope slide at
contrasting magnifications, so different human behavioural patterns
can be proposed on the basis of site-environment relationships at
varying scales.
The first attempts at trying to identify landscape attributes of
relevance to former societies were based on superimposing archaeo-
logical site distributions on particular types of map, e.g. a soil map.
Thus if sites seemed to be spatially associated with particular soils,
then environmental elements of behavioural significance could be
proposed. The use of site catchment analysis as originally described
by Vita-Finzi and Higgs (1970) puts such an approach on a quantita-
tive basis. The concept of economic rent is central to the methodology,
whereby it is assumed that the intensity of landscape usage decreases
with distance from settlement until a distance is reached beyond
which it is uneconomic to use the land. Detailed reviews of the
technique are given by Bailey and Davidson (1983) and Roper (1979).
Key criticisms revolve around the validity of an economic approach
and also the issue of how site catchments or territories are defined.
In the symposium at which the papers in this book were first presen-
ted, much discussion took place on the definition of such areas.
Speakers commented on the inapplicability of drawing a circular
territory or defining an area within two hours walking-distance from
a Mediterranean village. It tends to be forgotten that it is not villages
but families which exploit land at varying distances. Intermarriage
of families between villages can lead to complicated patterns of land
use and the results from site territory analysis using a simplistic
concentric approach become increasingly dubious. Although such
problems are well known, site territory analysis is widely used, to a
large extent because no better method has yet been devised. Neverthe-
less, studies using the technique have to be carried out in full aware-
ness of the assumptions and limitations.
This is very much the emphasis in the paper by Wagstaff, who
applied site territory analysis to a Neolithic site in Cyprus. In discuss-
ing the problem of defining territories he is able to draw upon the
survey which he carried out on Melos where, for example, he found
the seasonal use of land at distances in excess of 10 km. Another
key conceptual issue raised by Wagstaff is the frequent inadequacy
of site information as an aid to the interpretation of the significance
of land attributes within territories. Information is essential on
economy and available technology before land evaluation is possible.
Furthermore, an awareness of changing economies and technologies
needs to be evaluated in relation to dynamic landscapes within
territories.
The paper by Mytum illustrates the problems involved in trying
to understand the dynamics of settlement patterns from the Iron Age
 Introduction 23
to the Romano-British period in West Wales. Archaeological evidence
from individual sites as well as the setting of sites in relation to
topography and soils provides the basis for a discussion of this topic.
The latter approach is supplemented with a site territory analysis
which emphasises the distinction between defensive and farmstead
sites. Overall, the paper by Mytum demonstrates the merits of
integrating site archaeological information with a site territorial
analysis.
One unfortunate result of the site territorial approach is that sites
and landscapes are viewed as separate and different. The attitude
tends to be that human groups moved out to exploit the resources
available to them. Instead, the contextual approach reminds us that
landscapes provide biological and geological resources for use on
sites; furthermore, some of these materials may be redistributed back
onto nearby land as happens with manuring. Thus it is important to
consider landscapes as extensions of archaeological sites. The poten-
tial for generating data helpful to reconstructing cultural processes
through survey of artefact distributions is beginning to be appreciated
and this is the topic discussed by C. F. and V. L. Gaffney. They
indicate the scope of this approach by describing a survey of surface
artefacts within 2 km of a Romano-British villa complex in Berkshire.
Variations in sherd concentration are related to occupation sites,
activity areas, manuring spreads and low density localities which are
possibly the result of pasture use. The authors are then able to
compare sherd concentrations, which they interpret in terms of
agricultural intensity with catchments defined by constructing
Thiessen polygons. The spatial areas as identified by these two means
are similar and thus the potential is demonstrated for interpreting
site territories in terms of differing land uses.
This short introduction to Section 1 has tried to summarise the
main conceptual issues which arise from the papers dealing with
environmental research at the site and microregion scale. Although
the difficulties associated with interpreting environmental data from
sites as well as the problems with site territory analysis have been
highlighted, perhaps the most important issue is that environmental
evidence should be considered as integral to the archaeological
record. For example, anthropic loadings of phosphate on sites or
landscapes should be analysed with the same emphasis as other
artefactual evidence. Furthermore, the related key issue of integrating
all evidence for maximum interpretational value is emphasised.

BIBLIOGRAPHY
Bailey, G. N. & I. Davidson, (1983) Site exploitation territories
and topography: two case studies from Palaeolithic Spain',
Journal of Archaeological Science 10, 87-115.
24 DAVIDSON
Butzer, K. W. (1982) Archaeology as Human Ecology, Cambridge
University Press, Cambridge.
Eidt, R. C. & W. I. Woods, (1974) Abandoned Settlement Analysis:
Theory and Practice, Field Test Associates, Milwaukee.
Eidt, R. C. (1984) Advances in Abandoned Settlement Analysis:
Application to Prehistoric Anthrosoh in Colombia, South
America. The Center for Latin America, University of Wis-
consin, Madison.
Higgs, E. S. & C. Vita-Finzi, (1972) Prehistoric economies: a ter-
ritorial approach, 27-36 in Papers in Economic Prehistory
edited by E. S. Higgs. Cambridge University Press, Cam-
bridge.
Jochim, M. A. (1979) Breaking down the system: recent ecological
approaches in archaeology, 77-117 in M. G. Schiffer (editor),
Advances in Archaeological Method and Theory 2. Academic
Press, New York.
Quine, T. A. (1987) An evaluation of soil analysis for determining
formation processes on archaeological sites, unpublished Ph.D
Thesis, University of Strathclyde.
Roper, D. C. (1979) The method and theory of site catchment
analysis: a review, 119-140 in M. G. Schiffer (editor),
Advances in Archaeological Method and Theory 2. Academic
Press, New York.
Woods, W. I. (1977) The quantitative analysis of soil phosphate',
American Antiquity 42, 248-252.
The Role of Soils in the Interpretation of Archaeological
Sites in Northern Quebec

T.R. Moore and D. Denton

Departments of Geography and Anthropology, McGill University, Canada

The opportunities that the soil presents in the interpretation of

archaeological sites have been recognised for several decades. Human
occupance and utilisation can bring about physical disturbance in
the soil, which can be examined by macro- and micro-morphology.
The chemical properties of the soil can also be changed, especially
through the increased loading of elements such as Ca, Mg, and P,
associated with bones, refuse, fires and urine and faeces. Spatial
variations in soil chemical properties have been used in several
archaeological studies (e.g. Conway 1983, Edwards et al. 1983,
Griffith 1980, 1981, Heidenreich and Konrad 1973, Heidenreich and
Nauratil 1973, Konrad et al. 1983). One should also recognise that
'unloading' can occur, such as through the removal of organic matter
to create plaggen soils (Conry 1974) or through the scraping away
of the organic horizons before a habitation is built.
Human occupance and utilisation of the soil can also change the
chemical forms of elements, and this has been developed by Eidt
(1977, 1984). Based on the chemical forms of P in the soil, he was
able to show how patterns of land use could be deduced and how
the technique might be used to establish a chronology. In this paper,
we examine the opportunities and problems involved in using the
spatial variability of soil chemical properties and P fractionation,
with reference to archaeological sites in northern Quebec.
The area of our interest lies in the vast subarctic of Nouveau
Qubec. The hunting and trapping territories of Cre, Montagnais
and Naskapi Indians now take in most of this area, although per-
manent settlements are limited to a few villages along the James and
Hudson Bay coast, along the southern margin of the territory and at
Schefferville, one of a string of towns developed to exploit iron ore
in the Labrador Trough.
Archaeological interest in the subarctic interior is directly linked
to the hydroelectric development in the La Grande river basin which
began in the early 1970's. Salvage work has been carried out in several
areas that are now large reservoirs, both within the La Grande river
basin and the adjacent Caniapiscau basin. It was in this context that
26 MOORE AND DENTON
a study (supported by the Ministre des Affaires culturelles and the
Socit d'nergie de la Baie James) was initiated by D. Denton in
the Caniapiscau region. The goals of the Caniapiscau project since
its beginning in 1976 have been to establish a preliminary culture-
chronological sequence for the area, and to contribute to our under-
standing of the dynamics of prehistoric Indian adaptations through
the study of changing patterns of settlement and resource use (Denton
et al. 1981, 1982, 1983).
In the Caniapiscau area, the occupation sequence begins over 3,500
yr BP. Small groups of extremely nomadic people exploited caribou
and fish resources. Between about 1,500 and 2,200 yr BP, there are
very few dated remains, suggesting that at this time the area was
abandoned, or even more sparsely occupied. The florescence of
prehistoric hunting cultures begins around 1,500-1,400 yr BP. Sites
become larger and more numerous, and there is evidence of longer
occupations of these sites. Many houses become longer, frequently
including two or more hearths, and new patterns of lithic raw material
exchange develop (Denton, in prep.). The first occupation of any
intensity of the adjacent area closer to the coast of James Bay is
indicated in this period (Cerane 1984). Also, from the beginning of
this period, we feel confident in identifying the occupants as ancestors
of the modern and closely related Algonkian speaking Cre, Montag-
nais, and Naskapi.
The Caniapiscau reservoir region was deglaciated about 6,000 yr
BP (Richard et al. 1982) and the Canadian Shield rocks are covered
by a variety of glacial landforms, such as eskers and moraines. The
climate is harsh and the vegetation varies from tundra on the ridge
tops through woodlands and forests to peatlands. The soils of the
area are shallow (<1 m), very acid (pH of surface horizons generally
<4.5) and with thin accumulations of organic matter at well drained
sites. Podzolisation is the dominant process, with well developed
bleached and reddish brown horizons. Based on the soil properties,
one could expect that elements such as Ca and Mg leached from
bones, ash and refuse would be retained by the cation exchange
complex. P released from these sources could also be expected to be
retained in the soil, either as organic or inorganic complexes in the
organic horizons or as absorption by the Fe- hydroxides in the
subsoil horizons.

THE SPATIAL VARIATION IN SOIL PROPERTIES

The archaeological sites in the Caniapiscau region are generally small
and artefactual remains are limited to scatters of lithic debitage and
tools; the majority of the sites date from 1,400 to 400 yr BP. In an
attempt to aid the interpretation, the spatial variability of soil
chemical properties was examined at several sites. The aim was to
 The Role of Soils in Northern Quebec 27
identify areas of abnormal concentrations of elements such as Ca,
Mg, P and soil pH which could be associated with forms of human
occupance.
It is important to realise at the outset that soils are notoriously
variable in their properties, even within small areas. This variability
can be caused by differences in parent material, by contemporary
and historic vegetation differences, by differences in soil biota and
by differences in the soil forming processes, such as leaching, weather-
ing, erosion and deposition. Although this spatial variability is well
recognised, most attention has been directed towards variability
within large units such as soil series or landscape terrain segments,
though much of the variability that exists at this scale also occurs in
small units such as 10m2 (Beckett and Webster 1971, Wilding and
Drees 1983). To allow the ready identification of anthropic loading,
the spatial variability characteristics of soil properties should include
a normal distribution, a relatively low standard deviation (a low
coefficient of variation), a minimal skew and platykurtosis.
To examine to what extent these characteristics are met by soils
in subarctic spruce-lichen woodlands (a common vegetation associ-
ation in which the archaeological sites are found), a 25 m2 area was
sampled in two mature woodlands near Schefferville, 200 km east of
the Caniapiscau region. The areas were located away from the
influence of standing spruce trees, as this has been shown to impart
an added spatial variability (Moore 1980). There was no evidence
of human occupation or disturbance; site 2 had widely spaced trees
with a sparse shrub cover (dwarf birch and labrador tea), while site
1 had a denser cover. Both were carpeted by a thick lichen mat,
dominated by Cladina stellaris, with ericaceous plants. At 0.5 m
intervals on a 5 m grid, 10x10 cm samples were collected of the
surface organic (LFH) horizons. After air-drying and homogenising,
the 100 samples from each site were analysed for P, Ca, Mg and K
extracted by a 0.05N NH4F/0.1N HC1 solution (McKeague 1978).
This was selected because it could be expected to remove P found
in hydroxides loosely bound to Fe/, some Ca phosphates and
labile P, as well as extracting most of the cations in the potentially
mobile (exchangeable) forms. The pH of the soil in water was also
measured.
The results for these 'control' sites show that the coefficient of
variation for these 5 properties ranges from 0.27 to 1.49, with
extractable Ca the highest; pH has a very low coefficient of variation
(Table 1). With the exception of pH (which is a logarithmic transfor-
mation of [H+]), the data reveal a strong positive skew, especially
for Ca, and kurtosis varies from platykurtic variables, such as P and
Mg and K at site 2 to leptokurtic variables, such as Ca at sites 1 and
2. It may also be noted that despite the similarity in vegetation, parent
28 MOORE AND DENTON
Table 1. Soil properties at two undisturbed woodland sites near Schefferville

Property Mean St.Dev. Coeff.Var. Skew Kurtosis

Site 1 60 18 0.31 0.68** 0.34**

(n = 100) Ca 34 21 0.62 4.22** 24.97**
Mg 260 68 0.38 0.55** 0.33**
K 525 142 0.27 0.11 -0.52**
pH 3.69 0.16 0.04 -0.14 1.26**
Site 2 P 126 63 0.50 1.65** 3.00
(/! = (100) Ca 98 146 1.49 3.91** 19.17**
Mg 392 144 0.37 1.39** 3.32
K 782 263 0.34 1.72** 7.07**
PH 3.71 0.21 0.06 0.13 1.05**
Concentration in
Skew:, (Pearson, 1930) **p<0.01*p<0.05
Kurtois:/3 2 (Pearson, 1930) **p<0.01*p<0.05

material and soil morphology at the two sites, site 2 had almost
double the concentration of elements observed at site 1. This
emphasises the need to compare the spatial variability within the
background values occurring at each site, rather than using absolute
values.
These results are in accordance with the high spatial variability
encountered even within pedons, and the positively skewed distribu-
tion that many soil properties possess (Wilding and Drees 1983).
The results are also in general agreement with the data for Ca, Mg
and P presented by Konrad et al. (1983) from a series of prehistoric
sites in Maine. Their coefficients of variation for control and disturbed
sites ranged from 0.3 to 1.5, with the highest value generally for the
disturbed sites. Their data were also positively skewed, requiring a
logarithmic transformation before statistically assessing differences
between control and disturbed sites.
The high spatial variability and positive skew present problems in
the identification of anthropic loading at sites, and these problems
do not appear to have been recognised by many workers, though
efforts have been made (Proudfoot 1975, Webster 1977).
To evaluate the contribution that soil spatial variability can make
to the archaeological interpretation at these northern Quebec sites,
samples from the surface horizons at five sites were collected at 0.5 m
intervals on the grid pattern established by the undisturbed sites
discussed above; at some sites 'control' samples were also collected
from adjacent areas which appeared to be unaffected by human
disturbance. The results were analysed by histograms and simple
statistical parameters and then plotted on a map on each site. For
 The Hole of Soils in Northern Quebec 29
the graphical representation, the following categories were estab-
lished:
1 < mean
increasing 2 mean to mean +1 standard deviation
tone density 3 mean +1 standard deviation to
mean+ 2 standard deviations
4 arbitrarily selected from histograms, to split the
5 remaining values
If one assumes an approximately normal distribution, then classes
4 and 5 include the extreme upper 2.5% of the population, and thus
are candidates for areas of the site at which anthropic loading has
occurred.
The potential as well as the problems of the method are illustrated
by examination of the spatial patterns at one site, GbEl-1 (Fig. 1).
This site, excavated over 57 m2, is located on a moraine, within a
burnt spruce-lichen woodland. The majority of the lithic assemblage
is composed of quartz debitage, with smaller amounts of quartzite
and chert. The tool assemblage consists of a small number of
retouched flakes and biface fragments of these materials. The excava-
tion revealed a fairly large hearth in the southern part of the site,
based on changes in soil morphology and the presence of heat-
affected rocks. The paucity of charcoal hinders the dating of the site,
but the relatively poor preservation of the bones within the hearth
and the characteristics of the assemblage suggest that the occupation
is relatively old, perhaps 3,000 to 2,000 BP.
The histogram for the extractable P at the site reveals a strong
positive skew (Fig. 1). Many values fall into the classes 4 and 5
(>mean + 2 standard deviations) with concentrations greater than
twice the mean and greater than samples collected from outside the
site. The spatial pattern of soil P shows that these high values follow
the long axis of the hearth, with very few abnormally high concentra-
tions in other parts of the site. While this area of high P values
generally corresponds to the area of highest lithic debitage concentra-
tions (Fig. 1), the latter is oriented perpendicular to the hearth. The
high concentrations of extractable P are repeated in the subsoil
samples from beneath the hearth, compared to other subsoil samples
collected from elsewhere in the site (Table 2). The northern part of
the site has consistently low concentrations of P, suggesting that this
area was used as a workshop, rather than an area with organic
deposition.
The spatial patterns for Ca and Mg also show the highest concentra-
tions around the hearth, though the pattern is not so strongly
expressed as for P. The subsoil samples from beneath the hearth also
contain higher concentrations of Ca and Mg, compared to the non-
hearth subsoil samples (Table 2), though again the differences are
30 MOORE AND DENTON

ARTIFACTS <number/sq.m)

hearth
deposits
D o
1-100
100-400
40O- 722

EXTRACTABLE P (fig/g)

11 - 87
@ 87 - S8
188 - 289
g 289 -450
450 - WIS

400 6OO 1000

P CONCENTRATION

Figure 1. Maps of artefact and extractable phosphorus distribution and the

extractable phosphorus histogram at site GbEl-1. Each square on the
maps represents an area of 0.5 0.5 m.

.
 The Role of Soils in Northern Quebec 31
Table 2. Concentrations of extractable P, Ca, Mg and K and soil pH in subsoil
samples collected from beneath the hearth and elsewhere at site GbEl-1. Figures are
mean and standard deviation.
Sample Ca Mg K pH

Beneath hearth X 0. 186 0.151 0.051 0.035 5.16

(n =8) 0.219 0.106 0.022 0.013 0.49
Elsewhere in site X 0.018 0.037 0.025 0.039 4.84
(n=8) - 0.034 0.032 0.014 0.007 0.11

not as marked. The spatial pattern of K is not distinct, nor are the
subsoil samples different from those elsewhere at the site. The pattern
for soil pH also strongly reflects the influence of the hearth, though
there are other occasional high values which appear to be unrelated
to any other feature. Subsoil pH values are also higher beneath the
hearth than in the rest of the site.
The results for GbEl-1 are typical of the other site we have studied.
The chemical analyses are able to pick out the hearths, based on
high concentrations of extractable P and, to a lesser extent, extractable
Ca and pH, in both the surface and subsoil horizons. Based on the
relatively high concentrations of Mg in fire ash (Griffith 1981, Heiden-
reich and Nauratil 1973, Heidenreich and Konrad 1973), the ratio
of extractable Mg to Ca may be used to indicate whether the hearth
was used mainly as a source of heat, or for cooking, with the
deposition of Ca-rich bones. The soil is able to retain these features,
even after 2,000 to 3,000 years.
Apart from the hearths which, in any case, are generally obvious
from changes in soil morphology, the spatial patterns must be inter-
preted with some caution, given the high variability and the positive
skew exhibited at the two 'control' sites. In some sites the concentra-
tion of extractable P falls off within less than l m of the hearth. In
others, such as GbEl-1, the pattern is more diffuse, with high P values
spread over a wider area. However, it is not clear whether we should
interpret these differences in terms of variation in prehistoric activities
and refuse-deposition, or in terms of post-depositional natural pro-
cesses, such as freeze-thaw and nutrient cycling by vegetation which,
over time, may cause a 'smearing' of hearth patterns. Large areas of
abnormally low concentrations, such as the northern part of the
GbEl-1 site, may reflect an 'unloading' might be brought about by
an anthropogenic modification of the humus. At several sites in our
sample, areas of abnormally low concentrations occur within 1-4 m
of a hearth, suggesting the scraping away of surface soil horizons
during the preparation of the dwelling floor. At other sites, abnor-
mally high concentrations of P or Ca may indicate the occurrence
32 MOORE AND DENTON
of a door or a food processing area, for which some support can be
derived from the ethnographic data.
The main value in the spatial analyses at these sites appears to be
the confirmation of the hearth and its extent and character, and the
identification of parts of the site with abnormal values. These values
are of particular interest because they may point to activity or refuse-
areas not otherwise observable; however, archaeological support for
any given interpretation may be lacking. A comparative study of a
large sample of sites is required in order to identify clearly recurring
patterns through the variable background 'noise'.
PHOSPHORUS FRACTIONATION OF HEARTH SOILS
Phosphorus, usually as the orthophosphate forms, can occur in a
variety of different forms in soils and this topic and its archaeological
aspects has been reviewed by Proudfoot (1976). Spurred initially by
the low availability of P to plants in many soils, empirical chemical
extraction techniques have been developed to remove selectively
these different forms. Eidt (1977) has adapted the Chang and Jackson
(1957) sequence of extractants to remove P from soils in fractions
that may be of archaeological importance. His method identifies 3
main fractions:
I removed by NaCl/NaOH and Na citrate-bicarbonate;
generally the easily extractable forms, such as those soluble
in water, readily soluble forms, those loosely bound by Fe
and Al and P resorbed by CaCO3.
II removed by Na citrate-bicarbonate-dithionite; generally
P bound to Fe and Al oxides and hydrous oxides and
those forms occluded by the Fe and Al.
Ill removed by HC1; generally Ca phosphates and apatite.
Eidt (1984) has used this fractionation technique to identify land
use practices and settlement patterns in Colombia. The technique
may also establish the relative age of the site, based on the principle
that the occluded fraction II requires a long period for its formation,
so that the II : I fraction ratio could be taken as an indication of
relative age (Eidt 1977). Given the addition of P to the northern
Quebec sites mainly as Ca phosphates, such as bones and ash, one
might expect that through time the proportion of P in the Ca fraction
(III) would be reduced, with an increase in the proportion in the
labile and occluded forms (I and II), either in the surface or Fe/Al-
rich podzolic horizons.
To assess the utility of this approach, soil samples from twelve
hearths have been analysed for the Eidt I, II and III P fractions. In
addition, the samples were analysed for P extracted by hot HC1 after
ignition (approximately total P), P and Ca extracted by the
 The Role of Soils in Northern Quebec 33
NH4F/HC1 solution used above and pH in water. As examples of
the results, data for the fractionation are presented for two hearths
at each of two sites, for which the radio-carbon age has been deter-
mined from charcoal. Four bone samples collected from three sites
were analysed; they exhibited little variation, and the data are presen-
ted as the mean of the four samples (Fig. 2).
The results show the very high concentrations of P in these hearths,
particularly in the uppermost horizons, between 2 and 7% being
extracted by the I, II and III fraction sequence and similar values
removed by the hot HC1 treatment. The NH4F/HC1 treatment also
removed an average of 0.2% Ca and 1.5% P from the uppermost
horizon, much higher values than encountered in the spatial patterns

GaEl - 1 GaEk - 1

STRUCTURE I STRUCTURE 2 STRUCTURE 3 STRUCTURE 5

560 v B P 2550 v BP 660 y BP

P (%)
(BONES!

50

Si

BONES

Figure 1. The concentration and fractionation of phosphorus in soil

samples from 4 hearths, of different age. The bars indicate the
concentration of phosphorus in each sample, divided into fractions of I, II
and III (see text). The figures to the right of the bar indicate the
percentage of the phosphorus in each of the fractions. The notation 1, 2
and 3 indicate that the soil samples were taken from the upper middle and
lower horizons of the hearth. The rightmost bar indicates the phosphorus
concentration and fractionation in bone samples collected from the hearths
and other sites (mean of 4 samples).
34 MOORE AND DENTON
discussed above. The pH of the soil samples ranged from 5.2 to 5.4
in the uppermost horizon, higher than in non-hearth samples.
Although concentrations of P are smaller in the lower horizons, they
are above the values common in non-hearth subsoils.
In terms of P fractionation, most of the P appears in fraction III,
which are the Ca-phosphates, with 61 to 84% in the uppermost
horizons. In the lower horizons the pattern is more variable, with
most samples showing a less important component in fraction III.
This is in accordance with the expectation that, as the bones and ash
weather, the P becomes converted from the Ca into the Fe- and
labile forms. In all the samples, only a very small proportion (0-9%)
of the P is found in fraction II, the occluded and Fe- hydroxide
forms. This may reflect the short period since creation of the hearths,
the relative youth of the soils and the low Fe- hydroxide content
and its amorphous nature. Moore (1978) has shown that in many of
these northern soils, the subsoil extractable Fe and Al content ranges
up to 2%, much less, and of a more amorphous nature, than in
tropical soils.
It would be tempting to think that differences in the P fractionation
could be used to distinguish not only the intensity with which the
hearth was used, but also its relative age. As a hearth ages, one could
anticipate a reduction in the fraction III P and an increase in the
proportion in fractions I and II. Some evidence for this is given with
the four hearths, which have been radio-carbon dated at ages of 2550
to 560 BP. At site GaEl-1, there is little difference in the relative
proportions of P in the uppermost horizon sample. The lowest horizon
sample, however, shows a distinct difference, despite similar P con-
tents; the older hearth has 73% of its P in the I fraction, compared
to only 18% in the younger hearth. Data from the second pair of
hearths, from the GaEk-1 site with ages of 1100 and 660 yr B.P., tend
to dampen this distinction. Although 80 to 84% of the P in the
uppermost sample is in fraction III, both lowest horizon samples are
dominated (79-100%) by fraction I.
In interpreting the differences in the P fractionation, one must be
aware of the complexity of soil forming-processes and the concepts
of convergent and divergent evolution in soils. Haines-Young and
Fetch (1983) have recently discussed the role of the concept of
equifinality and multiple working hypotheses in geomorphology, and
Eidt (1984) recognises morphological convergence of hillslopes to
field types in Colombia. Convergent evolution can occur when two
soils, with different initial properties or use, evolve to produce the
same characteristics. In the case of these northern hearths, for
example, the same pattern of P fractionation may be produced if P
is added mainly as bones or mainly as the ash from a large fire.
Although the concentration and amount of P in the soil may be
 The Role of Soils in Northern Quebec 35
different, the fractionation may be similar, despite the different prac-
tices to which the soil has been subjected.
On the other hand, divergent evolution may occur. Loading a site
with bone-derived Ca and P may stimulate increased plant growth,
nutrient cycling and even a change in the type of vegetation, as occurs
on dolomite-containing soils in the Schefferville area of northern
Quebec. The change in vegetation and nutrient cycling may mean
that the P remains primarily within the organic fraction and may
mean a reduction in the intensity of podzolisation, reducing the
ability of the subsoil horizons to absorb translocated P. Compared
to the podzol with fraction I and II P at the unchanged site, one
might expect much less I and II at the site with the changed vegetation.
It is questionable whether we know enough about soil forming
processes to be able to extrapolate from P fractionation data. The
recent controversy over whether podzolisation is organic or inorganic
in origin is an example of this doubt and the fact that we have a
continuum of processes within the general process, which has been
studied intensively (e.g. Anderson et al. 1982, Buurman and van
Reeuwijk 1984).

This work, and that of many others, shows that soils do have a role
to play in archaeological interpretation, in terms both of the spatial
variations of soil properties at the site and the chemical changes
brought about by man's occupancy. Studies of northern sites undis-
turbed by human occupance has revealed the high variability and
skewed nature of the spatial distribution of soil properties, requiring
the careful identification of samples which appear to be abnormally
loaded. Phosphorus appears to be the most useful element, with Ca
of more limited value, partly because of its greater skewed distribution
in subarctic woodland soils, and its greater mobility. Although
measurements of P content and P fractionation confirm major
anthropic loading, the interpretation of prehistoric activities and age
from differences in P fractionation should be treated with some
caution.
Acknowledgement. This work has been funded by grants from La
Socit d'nergie de la Baie James and the Natural Sciences and
Engineering Research Council of Canada.
BIBLIOGRAPHY
Anderson, H. ., M. L. Berrow, V. C. Farmer, A. Hepburn, J. D.
Russell, & A. D. Walker (1982) A reassessment of podzol
forming processes, Journal of Soil Science, 33, 125-136.
Beckett, P. H. T. & R. Webster (1971) Soil variability, a review,
Soils and Fertilisers, 34, 1-15.
Buurman, P & L. P. van Reeuwijk, (1984) Proto-imogolite and the
36 MOORE AND DENTON
process of podzol formation: a critical note, Journal of Soil
Science, 35, 447-452.
Cerane, Inc. (1984) Occupations prhistoriques, historiques et
contemporaines de la rgion de Wasdimi, Rservoir de LG-2,
Baie James. Deux millnaires d'archives archologiques. Report
submitted to the Ministre des Affairs culturelles, Qubec.
Chang, S. C. & M. L. Jackson, (1957) Fractionation of soil
phosphorus, Soil Science, 84, 133-144.
Conry, M. J. (1974) Plaggen soils a review of man-made soils,
Soils and Fertilizers, 37, 319-326.
Conway, J. S. (1983) An investigation of soil phosphorus
distribution within occupation deposits from a Romano-
British hut group, Journal of Archaeological Science, 10, 117-
128.
Denton, D., M. Perdais, J-Y. Pintal, C. Rocheleau, M. Bouchard,
& P. Grgoire (1981) Investigations archologiques dans la
rgion du futur rservoir Caniapiscau, 1979. Direction
Gnrale du Patrimoine, Interventions Archologiques, No. 1.
Qubec, Ministres des Affaires culturelles.
Denton, D., L. Lafrance J-Y. Pintal, J-L. Pilon, & B. Emard
(1982) Investigations archologiques dans la rgion du futur
rservoir Caniapiscau, 1980. Direction Gnrale du
Patrimoine, Interventions Archologiques, No. 2 Qubec,
Ministres des Affaires culturelles.
Denton, D., E. Cossette, B. Hbert, N. Lafrance, & J-Y. Pintal
(1983) Recherche archologique dans la rgion du rservoir
Caniapiscau (saison 1981). Unpublished report prepared for
th Socit d'nergie de la Baie James. Montral, Qubec.
Denton, E. (in prep.) Prehistoric settlement and subsistence patterns
in the Caniapiscau area of Nouveau-Qubec. Ph.D. thesis, in
preparation. Department of Anthropology, McGill University.
Edwards, K. J., F. W. Hamond, & A. Simms (1983) The medieval
settlement of Newcastle Lyons, County Dublin: an
interdisciplinary approach, Proceedings of the Royal Irish
Academy, 83C, pp. 351-376.
Eidt, R. C. (1977) Detection and examination of anthrosols by
phosphate analysis, Science, 197, 1327-1333.
Eidt, R. C. (1984) Advances in Abandoned Settlement Analysis.
Application to Pre-historic Anthrosols in Colombia, South
America. The Center for Latin America, University of Wis-
consin, Madison.
Griffith, M. A. (1980) A pedological investigation of an archaeo-
logical site in Ontario, Canada. I. An examination of the soils
in and adjacent to a former village, Geoderma, 24, 327-336.
Griffith, M. A. (1981) A pedological investigation of an archaeo-
logical site in Ontario, Canada. II. Use of chemical data to
discriminate features of the Benson site, Geoderma, 25, 27-34.
Haines-Young, R. H. & J. R. Petch (1983) Multiple working
hypotheses: equifinality and the study of landforms, Transac-
tions, Institute of British Geographers, New Series 8, 458-466
Heidenreich, C. E. & V. A. Konrad (1973) Soil analysis at the
Robitaille site. Part II. A method useful in determining the
location of longhouse patterns, Ontario Archaeology, 20, 33-
62.
 The Role of Soils in Northern Quebec 37
Heidenreich, C. E. & S. Nauratil (1973) Soil analysis at the
Robitaille site. II. Determining the perimeter of the village,
Ontario Archaeology, 20, 25-32.
Konrad, V. A., R. Bonnischen R. & V. Clay (1983) Soil chemical
identification of ten thousand years of prehistoric human
activity areas at the Munsungen Lake thoroughfare, Maine,
Journal of Archaeological Science, 10, 13-28.
McKeague, J. A. (ed.) (1978) Manual of Soil Sampling and
Methods of Analysis. Canadian Society of Soil Science,
Ottawa.
Moore, T. R. (1978) Soil development in north-eastern Canada,
Annals Association of American Geographers, 68, 518-534.
Moore, T. R. (1980) The nutrient status of subarctic woodland
soils, Arctic and Alpine Research, 12, 147-160.
Pearson, E. S. (1930) A further development of tests for nor-
mality, Biometrika, 22, 239-249.
Proudfoot, B. (1976) The analysis and interpretation of soil phos-
phorus in archaeological contexts, pp. 93-113 in D.A. David-
son and M.L. Shackley (eds.). Geoarchaeology, Duckworth,
London.
Richard, P., A. Larouche, & M. Bourchard (1982) Age de la dg-
laciation finale et Historie postglaciaire de la vgtation dans
la partie centrale du Nouveau-Qubec. Gographie Physique et
Quaternaire, 36, 63-90.
Webster, R. (1977) Quantitative and Numerical Methods in Soil
Survey. Clarendon Press, Oxford.
Wilding, L. P. & L. R. Drees (1983) Spatial variability and
pedology, 83-140 in L. P. Wilding, N. E. Smeek and G. F.
Hall (eds.). Pedology and Soil Taxonomy. Elsevier, New York.
Environmental Reconstruction at Cul A' Bhaile, Isle of Jura

G. Whittington,
Saint Andrews University, Scotland.

The survival of sites of archaeological interest owes much to the

intensity, both in space and time, of arable agricultural activity; where
ploughing has occurred over a long period and where it has involved
a deep penetration of the soil, only the most solidly constructed
features are likely to have survived and indeed many of these will
have been destroyed by man in his attempts to expand his farmland
or in his search for readily usable stone. These effects will be most
marked in areas of fertile soil and amenable climate, characteristic
in Western Europe of lower altitude zones. In contrast the upland
areas with a greater percentage of acidic and poorly drained soil and
greater extremes of climate, both in terms of rainfall total and lower
temperatures, have, since the sub-Atlantic climatic deterioration, seen
little arable activity and hence a more complete preservation of the
features of prehistoric occupation. This is nowhere more marked
than in Scotland where the disposition of its upland areas and its
location on the western extremity of the European landmass have
led in general terms to a western zone, dominated by a maritime
climate and in which podzolic soils predominate, and an eastern
zone which climatically is more continental and where brown forest
soils are frequent. Throughout the western zone are to be found
archaeological sites, and these have attracted considerable attention
from excavators in recent years. Among the younger generation of
archaeologists is an awareness of the important contribution that can
be made by palaeoenvironmentalists to the more complete under-
standing of the functioning of communities in an excavation zone.
The request, however, for assistance in establishing environmental
and landuse features frequently faces the palaeoenvironmentalist
with severe problems. This is particularly so for the palynologist.
The sources of evidence for vegetation and landuse reconstruction
are most usually lacustrine deposits or deep accumulations of peat.
For many surviving archaeological sites such deposits are not avail-
able or, where they are, they lie at such a distance that the pollen of
the most significant economy indicator plants, such as cereals and
ruderals, are unlikely, due to the nature of their pollen dispersal
 Environmental Reconstruction at Cul A' Bhaile 39
characteristics, to be present. Such factors have led to palynologists
(e.g. Smith and Pilcher 1972, Ewan 1981, Chambers 1983) concentrat-
ing their interest upon blanket and basin peat (histosols) which is
virtually ubiquitous in the highland and western zone of Scotland.
The exploitation of this source of pollen evidence is, however, fraught
with dangers not the least of which is the unknown extent and
frequency of past peat cutting activities; in a zone which has been
virtually treeless over a very long period, the exploitation of peat for
fuel must have an extensive history. Also in many instances the onset
and development of peat accumulation postdates the structures which
the archaeologist is investigating. It is, however, that factor which
indirectly might provide the resource base which the palynologist
requires. Not only has the development of blanket peat in many cases
overgrown and thus preserved many different classes of archaeologi-
cal monument along with extensive areas of field systems but it has
also brought into existence a vast reservoir of palaeosols.
The exploitation of soil pollen spectra has been investigated by
many workers and is shown to contain many a trap for the unwary.
There is the problem of the transformation of brown forest soils into
podzols so that the soil pollen spectrum contains pollen of taxa which
belong to quite different ecological associations (e.g. Heim 1966,
Havinga 1968, Cruickshank and Cruickshank 1981). Pedogenic
processes frequently involve strong vertical translocations of mineral
and organic materials due to the movement of groundwater, and this
is further compounded by earthworm activity; the net result is an
obscuring of the stratification of pollen in the soil profile (Munaut
1967, Billard et . 1971, Guillet 1972, Iversen 1973, Havinga 1974,
Andersen 1979 and 1984). Furthermore, the differential preservation
of pollen in soils brought about by the effects of physical abrasion
and chemical and bacterial attack has been shown, as a result of
experiments conducted by many workers (Goldstein 1960, Elsik 1971,
Havinga 1971, Holloway 1981), to be an important distortion factor.
With all of these features in mind it might seem that even the general
presence of palaeosols in the western and upland areas of Scotland
is unlikely to solve the difficulty of finding a suitable resource base
by the palynologist seeking to provide environmental evidence for
the archaeologist.
Work undertaken by Dimbleby (1957,1961, 1976, see also Walker
and Taylor 1976) has shown that despite such problems much of
value can be gained from an investigation of soil pollen, and several
workers have undertaken such studies in an archaeological context
(e.g. Dimbleby and Evans 1974, Edwards and Ralston 1978, Whitting-
ton 1978, Balaam et al. 1982). Such enquiries, however, have been
mainly limited to palaeosols preserved by mounds, banks and other
structures created by man; the wider availability of palaeosols around
 40 WHITTINGTON
spcifie occupation features has not been seriously investigated. The
essence of Dimbleby's work was that despite the problems caused
by soil transformation, earthworm activity, and pollen degradation
and destruction, it was still possible to distinguish different and
coherent assemblages of pollen in the soil. This conclusion was based
upon a sampling of soils at a minimum interval of 2.5 cm. The
excavation of a field bank in the Black Moss of Achnacree, Argyllshire
allowed Whittington (1983) to test the suggestion that much closer
sampling of a palaeopodzol under a man-made feature would be
even more rewarding by providing a more detailed and refined
resolution of the palynological information. The results of that work
were encouraging and confirmed the belief that a wider and more
embracing examination of palaeosols in general might well be worth
exploring.
For five seasons (1976-80), a circular structure lying within a much
larger stone enclosure, both overwhelmed by blanket peat, at Cul a'
Bhaile on the Isle of Jura were the subject of intensive excavation
(Stevenson 1984). A strategy had to be devised for an equally intensive
palaeoenvironmental examination which, it was hoped, might reveal
the contemporary land use. The main problem to be faced, however,
was the lack of a proximate site from which a long continuous column
of material for analysis could be obtained. Investigation of the area
revealed the widespread existence of palaeopodzols beneath the
blanket peat. It was decided therefore that, in addition to the palyno-
logical investigations carried out on the apparent dwelling structure,
this would be an ideal situation in which to test the value of pollen
spectra obtained from palaeopodzols sampled at fine intervals. These
might permit not only the reconstruction of the vegetation and
landuse conditions surrounding the domestic site but also make
possible a greater understanding of the information obtained from
the domestic site itself. Therefore soil monoliths were taken from the
area both within and without the enclosure. The value of such a
project, however, had to be viewed in the light of the problems earlier
isolated as causing difficulty in the interpretation of soil pollen
spectra. First of all, it was felt that the extreme Atlantic location of
the site might negate at least two of these; the transformation of
brown earths into podzols and the mixing activity of earthworms.
The thin nature of the soil, on average about 11 cm in depth, led to
the belief that even if degradation of brown earths had been involved
here, the change had occurred at so remote a time (Romans and
Robertson 1983) that pollen from such a period would not only be
very sparse but would be virtually confined to the extreme base of
the soil profile. Secondly, it was felt that if close attention was paid
to the pedogenic processes revealed by the soil profile, and that if
these were related to any variation appearing in the pollen spectrum,

L
 Environmental Reconstruction at Cut A' Bhaile 41
a realistic interpretation might be possible of any stratigraphie
sequence of pollen that the analysis of the palaeosols might provide.
Thirdly, that if mixing of the soil by cultivation had occurred to any
depth the rudimentary nature of the cultivation (Hansen 1971) and
the tools used (Rees 1979) would have been unlikely to have led to
the destruction of the soil stratigraphy over either the whole of the
area or the whole of its depth.
Of the palaeosol monoliths obtained two will be used in discussion
here, one coming from within the enclosure and one from outside.
Fig. 1 shows the soil horizons and concentrations and general compo-
sition of the palynomorphs for the extra-enclosure monolith and Fig.
2 shows the same features for the intra-enclosure sample. The first
aspect which might be looked at is the suggestion that close sampling
of such a monolith provides valuable information which the use of
coarser sampling might miss. The intra-enclosure monolith was
sampled at 0.5 cm intervals while the extra-enclosure monolith was
sampled at 1.0 cm intervals. There are indications that the closer
sampling is even more worthwhile. An examination of the pollen
concentration shows some unexpected variations over and above the
generally encountered fall off in concentration of palynomorphs from
top to bottom of a soil profile. It is very noticeable that in the area
of the B2h horizon in Fig. 2 there are sharp variations in the concentra-
tion totals. Thus the close sampling has forced attention to be focussed

Depth of
sam| le Soil
in c -n horizons
-
-
2-
O 1
4-
AI
5- Ea
6-
7-
Palaeosol

8- B2
9-
10-
11- B2t
12-
13-
14-
S3
...
C Il

Pollen concentration Percentage composition

loo 50 0 50 100
*10 grains/gram

Figure 1. Soil horizons and palynomorph concentration and percentage

composition for the palaeosol monolith obtained from outside the stone
enclosure at Cul a' Bhaile.
42 WHITTINGTON
Depth of
samp le Soil
in c n horizons
0-
AO
o
2- AI

(C

3- Ea S.
4-
B2 I
.

5-
3
5- C Hillwash

7- A Peat
O
8-
A
9- 1
Palaeosol

10- Ea

11-
B2h
12-
Lower

13 - B2
14-
B21
15 -

16
C 11
||

Pollen concentrations Percentage composition

100 50 0 50 100

xlO grains/gram

Figure 2. Soil horizons and palynomorph concentration and percentage

composition for th palaeosol monolith obtained from inside the stone
enclosure at Cul a' Bhaile.

on this feature and the need to find reasons for its occurrence.
Foremost here must be the influence of the pedogenic process.
In podzolisation the translocation of clay down the profile and its
concentration in certain zones is a fundamental feature and, for
the palynologist, of critical importance because its action in
occupying the soil pores prevents the larger palynomorphs from
making any further progression down through the soil profile. This
is especially the case for many commonly occurring spores and
also for the pollen of cereals. This point can be pursued further by
referring to Figs. 3 and 4 which provide detailed information on
the palynomorphs encountered in the two palaeosol monoliths. In
particular, attention might be focussed on the B2t horizon in each
and on the distribution of the spores of Lycopodium, Pteridium
Figure 3. Soil horizons and detailed breakdown of the palynomorph
concentration and percentage composition for the palaeosol monolith
obtained from outside the stone enclosure at Cul a' Bhaile. Note that
concentrations for Corylus, Gramineae, Calluna and Pteridium are shown
at xlO 5 whereas all others are shown at 10".
Figure 4. Soil horizons and detailed breakdown of the palynomorph
concentration and percentage composition for the palaeosol monolith
obtained from inside the stone enclosure at Cul a' Bhaile. Note that
concentration for Gramineae and Calluna are shown at xlO 5 whereas all
others are shown at xlO 4 .
 Environmental Reconstruction at Cut A' Bhaile 45
and Polypodium. Such spores are commonly found well preserved
at the base of soil profiles and because of the relative sparseness of
other palynomorphs at these levels tend to be over-represented. Here,
however, there are unusual features which need to be accounted for.
In the extra-enclosure monolith (Fig. 3) Lycopodium is hardly rep-
resented at all, Polypodium is found throughout the profile and
Ptendium is barely present. This contrasts strongly with the situation
in the intra-enclosure monolith (Fig. 4), especially in the lower 9 cm.
At 14.5 cm of this entire monolith there occurs a B2t horizon which
has as its main characteristic the concentration of translocated clay
which forms a compact layer and leaves only a very small area of
the soil pore space unfilled. Immediately above it are strong and
numerous concentrations of Lycopodium clavatum (clubmoss) and
Ptendium aquilinum (bracken) which are highlighted by their relative
absence from the extra-enclosure monolith. This suggests that neither
of these plants played a part in the vegetation of the immediate area,
leaving a distinct possibility that their use as animal bedding was
involved in the local animal husbandry. Subsequently the fouled
litter was added to the land as manure thus introducing the spores
into the palaeosol. It is further noticeable in Fig. 4 that cerealia pollen
is found throughout the lowest 9 cm of the monolith except for the
area below the B2t horizon. Cerealia pollen is also of a size which
would prevent its passage through the B2t horizon, and thus is found
a correspondence in the distribution of an indicator of arable activity
and the two aforementioned spores. The depth of penetration of the
cerealia pollen and the spores would also suggest an arable agriculture
which persisted over a considerable period and which occurred both
inside and outside the enclosure, although being of a more intensive
nature within the enclosure.
Reference to Fig. 4 also reveals the importance of a physical
examination of the matrix from which the palynomorphs are ob-
tained. That figure shows a series of variations in pollen concentration,
the interpretation of which might lead to the formulation of some
complex theories on the low production of pollen in certain years
unless it were accompanied by information on the nature of the soil
horizons. Close sampling has revealed the existence of narrow soil
horizons which have played an important part in the overall distribu-
tion of the palynomorphs but has, just as importantly, shown that
the much greater thickness of the intra-enclosure monolith comprises
two distinct soil development periods and histories which were inter-
leaved by the formation of a peat layer and a subsequent period of
serious erosion which caused severe hillwash. This suggests at least
two distinct periods of human exploitation of the site which, as Figs.
3 and 4 show, were accompanied by cereal growing but in the second
phase with no accompanying spread of used animal bedding.
 46 WHITTINGTON
Among the interesting features revealed by the palynological work
on the dwelling structure was the manifestation of large concentra-
tions of the pollen of Vaccinium myrtillus (bilberry). It appears to
have been derived from the use of the plant as a construction material
in the repairing of the house and indicates, as does the Lycopodium
clavatum, the exploitation of vegetation from well beyond the
immediate vicinity of the dwelling because the pollen of Vaccinium
myrtillus was not encountered in the pollen spectra of any of the soil
monoliths.
The presence of cerealia pollen in the soil monoliths and the
appearance of ard marks beneath the foundations of the dwelling
structure also complement one another as pieces of evidence in the
reconstruction of the human exploitation of this site. From the
evidence in Fig. 4 it is clear that arable agriculture persisted over a
long period, being brought to an end by the deteriorating climate
which led to the renewed growth of peat and the eventual abandon-
ment of the site. Thus the ard marks are most likely to be related to
the period of arable agriculture revealed by the cerealia pollen in
the lower podzol of the two which occur within the enclosure, and
the likelihood of a further dwelling site in the area should not be
discounted especially as the palaeosol outside the enclosure also
contained cerealia pollen and is overlaid by frequent clearance cairns.
The meshing of excavation and palaeoenvironmental work could
therefore be the means of further archaeological discovery.
One of the hazards of soil pollen analysis isolated earlier was the
possibility of soil mixing which would invalidate the palynomorph
spectra that any analysis of a palaeosol might produce. The likelihood
of mixing does of course exist at Cul a' Bhaile, but three factors tend
to make this less of a problem. There are clear pollen stratification
differences and their content is not ecologically conflicting; there are
strong pollen concentration differences which are more likely to be
related to pedogenic processes than cultivation disturbance or
earthworm activity; and finally there are clear and well defined soil
horizons. These features all suggest that the palaeopodzols cannot
be dismissed merely as having preserved one homogeneous pollen
assemblage only and, furthermore, one which contains evidence
relating to a single suite of events.
Other features (see Stevenson 1984), along with those discussed
here, allow several conclusions to be drawn which could have wider
implications for workers involved in excavation and the establishment
of palaeoenvironmental conditions in upland and western Scotland.
First, that investigation of palaeosols in the vicinity of excavations
allows much more sense to be made of palynological work carried
out on the excavation site itself. Secondly, that such investigations
must be carried out by means of close sampling; that it must involve

L
 Environmental Reconstruction at Cut A' Bhaile 47
the exploitation of more than one monolith; and that excavation in
areas of acidic soils, where artefactual survival is frequently poor
and where contemporary peat deposits are often absent, will lack
important information unless palaeosol palynology is undertaken.
For this to occur, however, post-excavation financial provision would
have to be forthcoming as the vitally fine detail work is very time
consuming and exhausting. Finally, it must be emphasised that close
co-operation between the palynologist and a pedologist is desirable.

Acknowledgements. I am grateful to Dr J.A. Soulsby for helpful

discussions on the Jura soil monoliths, and to Mr C.B. Bremner and
Mr G. Sandeman who drew the diagrams.

BIBLIOGRAPHY
Andersen, S. T. (1979) Brown earth and podzol: soil genesis
illuminated by microfossil analysis, Boreas 8, 59-73.
Andersen, S. T. (1984) Stages in soil development reconstriicted
by evidence from hypha fragments, pollen and humus
contents in soil profiles, pp. 295-316 in Haworth, E.Y. &
J. W. G. Lund (eds.) Lake Sediments and Environmental
History, Leicester University Press.
Balaam, N. D., K. Smith, & G. J. Wainwright (1982) The Shaugh
Moor Project: Fourth Report Environment, Context and
Conclusions, Proc. Prehist. Soc. 48, 203-78.
Billard, ., J. Dejou, J. Guyot & J. Morizet, (1971) Analyse
pollinique des sols : un example les sols bruns acides
facis humifre d'Eymoutiers-Peyrat-le-Chateau (Haute-
Vienne) compares quelques tourbires voisines, Bull. Ass.
Fr. tude Quai. 3, 181-94.
Chambers, F. M. (1983) The palaeoecological setting of Cefn
Gwernnffrwd, a prehistoric complex in Mid Wales, Proc.
Prehist. Soc. 49, 303-16.
Cruickshank, J. G. & M. M. Cruickshank (1981) The development
of humus-iron podzol profiles, linked by radiocarbon dating
and pollen analysis to vegetation history, Oikos 36, 238-253.
Dimbleby, G. W. (1957) Pollen analysis of terrestrial soils, New
Phytol. 56, 12-28.
Dimbleby, G. W. (1961) Soil pollen analysis, J. Soil Sei. 12, 1-11.
Dimbleby, G. W. (1976) Climate, soil and man, Phi. T. Roy. Soc.
B. 275, 197-208.
Dimbleby, G. W. & J. G. Evans (1974) Pollen and land-snail
analysis of calcareous soils, J. Arch. Sei. l, 7-33.
Edwards, K. J. & I. Ralston (1978) New dating and environmental
evidence from Burghead Fort, Morayshire, Proc. Soc. Antiq.
Scot. 109, 202-9.
Elsik, W. C. (1971) Microbiological degradation of sporopollenin,
pp. 480-510, in Brooks, J., Grant, P. R., Muir, M., van Gijzel,
P., Shaw, G., (Eds.) Sporopollenin, Academic Press, London.
Ewan, L. (1981) A palynological investigation of a peat deposit near
Banchory, O'Dell Mem. Mon. 11, Aberdeen Univ., pp. 1-54.
48 WHITTINGTON
Goldstein, F. (1960) Degradation of pollen by phycomycetes,
Ecol. 41, 543-45.
Guillet, B. (1970) tude palynologique des podzols. I. La
podzolisation sur alluvions anciennes en Lorraine, Pollen et
Spores 12(1), 45-69; II. La podzolisation sur les versants sec
grseux des Basses-Vosges, Pollen et Spores 13(2), 233-53; III.
La podzolisation sur granit dans les Vosges hercyniennes de
l'tage montagnard, Pollen et Spores 13(3), 421-45.
Hansen, H. O. (1968-71) Experimental ploughing with a D0strup
ard replica, Tools and Tillage 1, 67-92.
Havinga, A. J. (1968) Some remarks on the interpretation of a
pollen diagram of a podzol profile, Acta Bot. Neer. 17(1),
1-4.
Havinga, A. J.(1971) An exprimental investigation into the decay
of pollen and spores in various soil types, pp. 446-479, in
Brooks, J., Grant, P. R., Muir., M., van Gijzel, P., Shaw, G.
(eds.) Sporopollenin, Academic Press, London.
Havinga, A. J. (1974) Problems in the interpretation of pollen
diagrams of mineral soils, Geol. Mijnbou 53(6), 449-53.
Heim, J. (1966) Analyse pollinique d'un podzol Hergenrath
(Province de Lige), Bull. Ass. Fr. tude Quai. 3, 208-16.
Holloway, R. G. (1981) Preservation and experimental diagenesis of
pollen exine, Ph.D. dissertation, Dept. Biology, Texas A & M
Univ., College Station.
Iversen, J. (1973) The development of Denmark's nature since the
last glacial, Geology of Denmark III, Dan. Geol. Unders.
Raekke 7-c, pp. 1-126.
Munaut, A. V. (1967) Recherches palaeo-ecologique en basse
moyenne Belgique, Acta Geogr. Lovan. 6, 1-196.
Rees, S. E. (1979) Agricultural implements in prehistoric and
Roman Britain, Brit. Arch. Rep., Brit. Ser. 69 (i) & (ii).
Romans, J. C. C. & L. Robertson (1983) The environment of
North Britain, pp. 55-80, in Chapman, J. C. & H. C. Mytum
(eds.) Settlement in North Britain, Brit. Arch. Rep., Brit. Ser.
118.
Smith, A. G. & J. R. Pilcher (1972) Pollen analysis and radiocar-
bon dating of deposits at Slieve Gullion, Co. Armagh, Ulster
J. Arch. 35, 17-21.
Stevenson, J. B. (1984) The excavation of a hut-circle at Cul a'
Bhaile, Jura, Proc. Soc. Antiq. Scot. 114.
Walker, M. F. & J. A. Taylor (1976) Post-neolithic vegetation
changes in the western Rhinogau Gwynedd, north-west
Wales, T. Inst. Brit. Geogs New Ser. 1(3), 323-345.
Whittington, G. (1978) A sub-peat dyke on Shurton Hill, Main-
land, Shetland, Proc. Soc. Antiq. Scot. 109, 283-291.
Whittington, G. (1983) A palynological investigation of a second
millennium BC bank-system in the Black Moss of Achnacree,
J. Arch. Sei. 10, 283-291.
 The Contribution of Shells to Site Interpretation:
Approaches to Shell Material from Franchthi Cave

Margaret R. Deith,
The Godwin Laboratory, Cambridge
Judith C. Shackleton,
Clare Hall, Cambridge
The value of palaeoenvironmental studies in the interpretation of
past behaviour from archaeological residues lies in the analysis of
landscape and history in terms of the changing options presented to
human groups (Butzer 1982). The basic assumptions made in this
paper are that, whilst the environment provides a set of limiting
factors, choices are made within that framework and, secondly, that
a socioeconomic system works as a system, so that changes in one
aspect will affect the working of the whole. In this way, the role of
a minor food resource, which is not vital to the economy, may
nevertheless be a useful way of monitoring aspects of cultural stability
or change.
The particular problem this study addresses is the transition from
a foraging to a farming way of life at Franchthi Cave, Greece. The
focus is on the use of shellfish during the transitional period. Shell
debris is consistently present throughout the Holocene deposits. A
radical change in the total economic system should affect shellfish-
gathering behaviour as well as everything else, so that analysis of
the patterning in this data set should offer insights into the system
as a whole.
Franchthi Cave, in the southern Argolid (Fig. 1), was first occupied
some time prior to 22,000 yr. BP, perhaps as early as 26-28,000 yr.
BP, and continued to be used until around 5,000 yr BP. It is the only
site so far known in Greece covering this long chronological sequence.
Survey work in the area has shown that Franchthi was the only major
site in the area during the Mesolithic and Neolithic until the final
phase of the Neolithic: 'settlement elsewhere in the southern Argolid
remained exceedingly sparse' (van Andel and Runnels 1986).

THE SHELLFISH DATA

Although the cave is not a shell midden and no part of the excavated
deposits has concentrations of shell, from about 11,000 BP onwards,
marine shells representing food refuse are consistently found. Table
1 shows the considerable variation in species present through time.
50 DEITH AND SHACKLETON

Figure 1. Map showing location of Franchthi Cave.

Table 1General Species Variation With Time in Marine Molluscs from Franchthi
Cave

Chronology and Most abundant species used as

Molluscan Zone indices of variation through time Habitat
*ca. 5,000 BP Mixed assemblage characterized by sandy/muddy sub-
IV several bivalves e.g. Cerastoderma strate; salinity: marine
glaucum, Tapes decussatus, Donax to brackish; most in
trunculus and Donacilla cornea. shallow water
ca. 6,900 BP Cerithium uu/ga/um-dominated. stones, sand, mud;
Ill Species forms 60 to 80% of assem- salinity: marine; shal-
blage. low water
ca. 8,500 BP Cyclope nerirea-dominated. Species sandy/muddy sub-
II ' forms 40 to 80% of assemblage. Rest strate; salinity: marine
mainly Cerithium vulgatum with to brackish; shallow
Cerastoderma glaucum in lower part. water
ca. 9,400 BP Patella spp, Mondonta spp, Gibbula rocky substrate; salin-
I spp, together with other species of ity: marine; shallow
gastropods form 60 to 80% of the water
assemblage

The composition of the assemblage during the transition from Upper

Mesolithic (Zone III) to Neolithic (Zone IV) is given in greater detail
in Fig. 2. Zone III is dominated by the gastropod, Cerithium vulgatum
 The Contribution of She/Is to Site Interpretation 51

100-

I I I I
140 120 100 80
FAS units
Figure 2. Molluscan assemblages from Upper Mesolithic (Zone I I I ) and
Neolithic (Zone IV) levels at Franchthi Cave.
 52 DEITH AND SHACKLETON
Bruguire, almost to the exclusion of all other species. Zone IV, by
contrast, is characterized by a number of different species, pre-
dominantly bivalves from the sandy shore. The most obvious explana-
tion for change in species collected is simple availability. With con-
tinuous post-glacial sea-level rise, change in the position of the
coastline inevitably occurred. The configuration of the coastline was
also subject to change, with concomitant effects on molluscan habi-
tats. Shoreline reconstructions demonstrate the position of the coast-
line, the types of shore and consequently the kinds of molluscan
habitats through the period under discussion (Shackleton and van
Andel 1986, Shackleton 1987). These provide a baseline for the
environmentally available options for molluscan exploitation.
Although sea level was rising throughout the period represented by
Zones III and IV, so that the position and nature of the shoreline
were changing, at no point during this time was a shellfish habitat
eradicated nor was there the introduction of a totally new one
(Shackleton 1987).
A marked feature of the Upper Mesolithic is the concentration on
gathering a single species from the range of molluscs likely to have
been available. This pattern of human prdation cannot therefore be
explained by environmental constraints and is thus to be interpreted
in socioeconomic terms. The shellfish data must now be considered
in relation to information gained from other data sets from the site.

THE NON-SHELLFISH DATA

A detailed examination of the shellfish in the context of the rest of
the data from the site is precluded by the small body of published
material currently available. It is, however, possible to make interest-
ing comparisons, even if in outline only, with the animal bone and
the botanical remains. These data sets are crucial to any interpretation
of the change from foraging to farming.
In his discussion of the faunal remains, Payne (1975) designates
the last foraging phase as Phase D, which begins with about 70%
red deer and 30% wild pig. The latter decreases relative to the red
deer towards the end of this phase.2 The first faunal phase described
as Neolithic is El, which is distinguished by sheep/goat representing
over 90% of the bones. This fits with the ceramic sequence quoted
on the same page as 'Aceramic' Neolithic (Payne 1975:129). In a
more recent paper, Payne comments that 'there is little doubt that
[the sheep and goats] were domestic even in the earliest Neolithic
levels: this is suggested by their sudden appearance, the high propor-
tion of younger animals, and the common occurrence from the earliest
(Aceramic) Neolithic deposits onward of shed sheep and goat milk
teeth, suggesting that the animals were commonly penned in the cave'
(Payne 1985:219). The second faunal phase in the Neolithic, E2, is

L
 The Contribution of Shells to Site Interpretation 53
characterized by a decrease in the proportion of sheep and goat to
about 70%. Pig, and later cattle, appear to have become significant
in the diet (Payne 1975:129).
In her thesis on the botanical remains from the cave, Hansen ties
in the introduction of the domesticated cereals, emmer wheat and
two-row hulled barley, with the inception of faunal Phase El.
Evidence of deliberate cultivation of cereals can also be seen in the
appearance of several new species of weeds common in cultivated
fields. The deliberate cultivation of cereals (botanical Zone VI)
replaced the gathering of wild oats and barley (Zone V), although
the utilization of plants such as wild pistachio, pear and almond
continued (Hansen 1980).
The data thus show a clear distinction between faunal Phase D
(with an emphasis on deer and wild pig) and Phase E, with its
overwhelming component of sheep/goat. The botanical remains
document a parallel transformation, from the gathering of wild
cereals to the cultivation of domesticated ones. The point at which
Payne divides El from E2, that is, the introduction of cattle and
domesticated pig, coincides with the unequivocal use of pottery at
the site. These data, therefore, indicate two stages in the complete
transition to a wholly Neolithic way of life.
The second stage, moreover, is marked by evidence for an
expansion of activity to the area immediately outside the cave, the
Paralia. 'Excavation ... outside and below the cave has revealed a
complex of rather substantial stone walls, the earliest of which can
be dated to the beginning of the (ceramic) Neolithic.... [These]
represent an expansion of the nuclear settlement in the cave. An area
of only slightly more than 100 m2 of this settlement has been exposed
by excavation, but we have enough to indicate that it was once
partially submerged in the Bay of Koilada' (Jacobsen 1981:309).
Since this publication, Gifford has carried out two seasons of sub-
marine coring in the bay. The cores have provided clear evidence of
Neolithic settlement in this region (Gifford 1983).
A comparison of the shellfish with the non-shellfish data shows
that the change from faunal Phase El to E2 coincides with the
boundary between molluscan Zones III and IV. However, it should
also be noted that the introduction of domesticated plants and
animals, the classic 'Neolithic package' (faunal Phase D to El,
botanical Zone V to VI), is wholly undetected in the molluscan record.

INTERPRETING THE SHELLFISH DATA

It is important at this juncture to consider the role of shellfish
collecting in foraging and farming communities. The ethnographic
literature indicates that there is a fundamental difference between
the way in which gathered wild resources, including molluscs, are
 54 DEITH AND SHACKLETON
used by hunter-gatherers and the way in which subsistence farmers
use them. In the first place, the function of such foods in the diet is
different. Among foragers, gathered foods supply basic caloric needs.
As a reliable resource, available throughout the year, shellfish can
be particularly useful when hunting expeditions have been unsuccess-
ful and also at times of year when other resources may be in short
supply (Meehan 1982). On the other hand, basic caloric needs are
normally met by cereals in subsistence farming economies, while
gathered foods such as shellfish supply essential vitamins and
minerals. They also tend to be more sharply flavoured than most
agricultural products. In general, then, for agriculturalists, such wild
foods are used as relishes rather than staples (see Scudder 1970,
Wilken 1970, Forbes 1976a, 1976b). Secondly, among subsistence
farmers, foraging activities are closely intertwined with agricultural
ones, forming part of an integrated system. Farmers gather wild foods
in the course of tending animals, visiting outlying fields or weeding.
These foraging and farming activities are seen as complementary. In
contrast with this approach, hunter-gatherers go out specifically for
the purpose of gathering wild foods. Finally, because of this difference
in approach, the social structuring of food gathering differs according
to the economic base. Among hunter-gatherer groups, the women
are responsible for gathering activities, while among farming groups
there is no division of labour by gender in this area. Whoever is close
to the resources will gather what is available.
The choice of a single species of mollusc during the Upper Meso-
lithic at Franchthi suggests a highly focused approach to shellfish
gathering, which accords well with the ethnographic data. Likewise,
the diversity of species collected during the Neolithic implies a more
opportunistic approach. Whilst the shift in the pattern of gathering
bears out the expectations generated by the ethnographic literature,
the timing appears to be offset.
Initially, it seems probable that the introduction of domestic plants
and animals to an area would have an impact sufficiently major to
alter all aspects of economic activity. Given the manner in which
shellfish resources are used by farming communities, shellfish collect-
ing could be expected to be a sensitive indicator of the advent of
farming. Yet, at Franchthi, there is no change whatever to be observed
in the shell record until the second stage of the transition to a fully
Neolithic way of life.

TOWARDS AN EXPLANATION
The molluscan record indicates stability at a time of apparent
maximum instability. The assumption has frequently been made that
the transition to the Neolithic in Greece was effected by population
replacement. Skeletal evidence from Franchthi is inconclusive on

L
 The Contribution of Shells to Site Interpretation 55
this question (Angel 1973). The idea of wholesale population replace-
ment as a mechanism for explaining the introduction of farming in
south-eastern Europe has been increasingly challenged in recent years
(e.g. by Dennell 1984, 1985). The evidence of molluscan Zone III at
Franchthi, with its emphasis on a highly discriminatory pattern of
shellfish collecting, continuing across the boundary, is incompatible
with the advent of a new people introducing domesticates to the site.
Rather, it appears that the indigenous hunter-gatherer population
was able to absorb new techniques for food procurement. Since there
are no other sites dating from this period in the southern Argolid,
lack of data precludes any attempt to explain the local mechanisms
for this first stage of the transition to farming in terms of contacts
between the group of hunter-gatherers at Franchthi and farming
communities, whether in Greece or further afield. Whatever the
precise mechanism for their introduction, however, it is possible to
argue that the annual management of sheep, goats and cultivated
cereals was not automatically as disruptive as might be expected.
Conversely, the changes that occurred at stage 2 may signify a
deeper socioeconomic transformation than is initially apparent. The
inclusion of cattle and pigs represents a more efficient use of the
total environment. It implies the existence of a fully integrated system,
in which not only were more of the natural resources, such as
woodland, exploited for feeding the animals, but also the agricultural
by-products of stubble and fallow fields could be more efficiently
used if they were grazed first by cattle and then by sheep (Halstead
1981). The introduction of cattle and pig was accompanied by an
expansion of the settled area and clear evidence for the use of pottery.
The change in shellfish collecting observed at this point may be
indicative of the manner in which it was embedded within the new
farming structure. Stage 2, then, seems to mark a critical boundary
at which the existing system broke down, unable to accommodate
all these new factors. A new pattern had emerged: the transition to
settled village life was complete.

CONCLUSIONS
The data from Franchthi, presented here, have been discussed within
the narrow confines of the site itself. The next step should be to open
up the debate to the regional level. Franchthi is not merely the only
site in the southern Argolid spanning this transitional period, but in
Greece as a whole. The only level at which comparison is possible
is that of the eastern Mediterranean. At this level, it is clear that the
two-stage transition to the Neolithic observed at Franchthi reflects
a wider eastern Mediterranean phenomenon: 'on all archaeological
sites in the eastern Mediterranean and in western Asia, there appears
to have been a rather rapid changeover in meat consumed, from wild
56 DEITH AND SHACKLETON
animals to domesticated sheep and goats, followed later by cattle
and pig' (Clason and Glutton-Brock 1982:146).
The time lag between the introduction of sheep and goats, and
cattle and pig, has also been noted in the western Mediterranean
(Geddes 1986). In this area, it has also been observed that the
appearance of cattle and pigs was followed closely by that of the
first village settlements (Lewthwaite 1986). Lewthwaite suggests that
the delay in the formation of villages 'was in fact primarily due to
the absence of such variables as cattle and pigs, vital to the efficient
operation of the farming system' (Lewthwaite 1986:112).
The hypothesis presented here, that the critical point of change at
Franchthi coincides with the introduction of cattle and pig, suggests
that the phenomenon observed in the western Mediterranean may
also be applicable further east. In relation to Franchthi, this may be
tested in the future against other data sets when they are published.
In the meantime, the highlighting of the molluscan data calls into
question the view that the abrupt introduction of domesticates rep-
resents a complete break with earlier traditions. Despite their low
economic value, shellfish may nevertheless illuminate aspects of past
behaviour to which it is difficult to gain access by more established
routes.

Acknowledgements. Whilst we take full responsibility for the views

expressed in this paper, we have very much appreciated stimulating
discussions with John Chapman, Robin Dennell, Paul Halstead and
Jim Lewthwaite. We thank Chris Bowers for drawing Fig. 2. Finally,
we are grateful to John Bintliff for the freedom he gave us to revise
the manuscript at a very late stage.

NOTES
1. Fig. 2 uses data from trench FAS. The material from this trench is representative
of the molluscan assemblage as a whole.
2. The data for Phase D are drawn from trench HI A, inside the cave. The data
Payne uses for his analysis of the Neolithic bone material are drawn from both
halves of the other deep trench within the cave, FA. The earlier phases are given
from FAN and the later ones from FAS. The excavation unit numbers (Payne
1975:129) are given below, together with the stratigraphically equivalent unit num-
bers from FAS:
El = FAN 161-151 [approximately equal to FAS units 144-136]
Aceramic Neolithic= FAN 163-150 [approximately equal to FAS units 145-139]
E2 begins at FAN 146 [approximately equal to FAS unit 135]
Ceramic Neolithic begins at FAN 148 [approximately equal to FAS units 137-

The Aceramic Neolithic deposits may be up to 30 cm. thick in FA and are thought
to represent probably not more than a few hundred years at a maximum.
 777e Contribution of Shells to Site Interpretation 57
BIBLIOGRAPHY

Andel, T. H. van & C. Runnels (1986) Five thousand years of

land use and abuse in the southern Argolid, Greece, Hesperia
55, 103-128.
Angel, J. L. (1973) Neolithic human remains pp. 277-282 in T. W.
Jacobsen Excavations in the Franchthi Cave, 1969-1971.
Part II, Hesperia 42, 253-283.
Butzer, K. W. (1982) Archaeology as Human Ecology. Cambridge,
C.U.P.
Clason, A. T. & J. Clutton-Brock (1982) The impact of domestic
animals on the vegetation during the first phases of animal
husbandry in the Mediterranean and Near East, pp. 145-148
in J. L. Bintliff and W. van Zeist (Eds.) Palaeoclimates,
Palaeoenvironments and Human Communities in the Eastern
Mediterranean Region in Later Prehistory. Oxford, BAR Inter-
national Series 133.
Dennell, R. W. (1984) The expansion of exogenous-based
economies across Europe: the Balkans and Central Europe,
pp. 93-115 in S. P. De Alley and F. J. Findlow (Eds.) Explor-
ing the Limits. Frontiers and Boundaries in Prehistory. Oxford,
BAR International Series 223.
Dennell, R. W. (1985) The hunter-gatherer/agricultural frontier in
prehistoric temperate Europe, pp. 113-139 in S. W. Green
and S. M. Perlman (Eds.) The Archaeology of Frontiers and
Boundaries. London, Academic Press.
Forbes, M. H. C. (1976a) Farming and foraging in prehistoric
Greece: a cultural ecological perspective, Annals of the New
York Academy of Science 268, 127-142.
Forbes, M. H. C. (1976b) Gathering in the Argolid: a subsistence
subsystem in a Greek agricultural community, Annals of the
New York Academy of Science 268, 251-264.
Geddes, D. S. (1986) Neolithic, Chalcolithic, and Early Bronze in
west Mediterranean Europe, American Antiquity 51, 763-778.
Gifford, J. (1983) Core sampling of a Holocene marine sedimen-
tary sequence and underlying Neolithic cultural material off
Franchthi Cave, Greece, pp. 269-281 in P. M. Masters and N.
C. Flemming (Eds.) Quaternary Coastlines and Marine
Archaeology. London, Academic Press.
Halstead, P. (1981) Counting sheep in Neolithic and Bronze Age
Greece, pp. 307-339 in I. Hodder, G. Isaac and N. Ham-
mond (Eds.) Pattern of the Past. Cambridge, C.U.P.
Hansen, J. M. (1980) The Palaeoethnobotany of Franchthi Cave,
Greece. Ph.D. dissertation, University of Minnesota.
Jacobsen, T. W. (1981) Franchthi Cave and the beginning of
settled village life in Greece, Hesperia 50, 303-319.
Lewthwaite, J. (1986) From Menton to Mondego in three steps:
application of the availability model to the transition to food
production in Occitania, Mediterranean Spain and southern
Portugal, Arqueologia 3, 95-119.
Meehan, B. (1982) Shell Bed to Shell Midden. Canberra, Aus-
tralian Institute of Aboriginal Studies.
58 DEITH AND SHACKLETON
Payne, S. (1975) Faunal changes at Franchthi Cave from 20,000
B.C. to 3,000 B.C., pp. 120-131 in A. T. Clason (Ed.)
Archaeozoological Studies. Amsterdam, North Holland.
Payne, S. (1985) Zoo-archaeology in Greece: a reader's guide, pp.
211-244 in N. C. Wilkie and W. D. E. Coulson (Eds.) Contri-
butions to Aegean Archaeology. Minnesota, University of Min-
nesota Center for Ancient Studies.
Scudder, T. (1970) Gathering among African woodland savannah
cultivators, Zambian Papers 5, 1-50.
Shackleton, J. C. (1987) Reconstructing past shorelines as an
approach to determining factors affecting shellfish collecting
in the prehistoric past, in G. N. Bailey and J. Parkington
(Eds.) The Archaeology of Prehistoric Coastlines. Cambridge,
C.U.P.
Shackleton, J. C. & Tj. H. van Andel (1986) Prehistoric shore
environments, shellfish availability, and shellfish gathering at
Franchthi, Greece, Geoarchaeology 1, 127-143.
Wilken, G. C. (1970) The ecology of gathering in a Mexican farm-
ing community, Economic Botany, 24, 286-294.
 Thoughts On Archaeology, Geography and Scales

P. J. R. Modderman,
University of Leiden

Archaeology and geography have much in common. The linkage is

perhaps stronger than supposed by most archaeologists, and it is
arguable that archaeology is closer to geography than to history. It
is not without reason that in the Dutch Academic Statute, geography
and prehistory are joined under one roof in a so-called Interfaculty.
By this means they are separated from the other sciences, which are
organised in the old faculties. With three of them, the parent-faculties,
the Interfaculty still has strong ties. They are the faculty of Arts
(history, classical and non-European archaeology), Science (earth
sciences and natural history) and Social Sciences (anthropology). To
underline the ties between the Interfaculty of Geography and Prehis-
tory and other faculties, its professors also belong to one of the
others. These regulations are partly based on the idea that geography
as well as prehistory bear important components belonging to the
three parent-faculties. On the other hand the independence of geogra-
phy and prehistory is stressed. But there is of course more behind
it, which can best be explained by comparing what both sciences are
studying.
Prehistory, to begin with, studies the results of activities carried
out by a human group in its surroundings. The evidence for these
activities is of course restricted to what has been 'materialised' and
preserved in the soil, consisting of the artefacts which enable us to
try to reconstruct human activities in the past. To distinguish human
activity from natural agencies is the first task of the archaeologist.
Geography, on the other hand, is split into physical and human
geography. The first tries to explain the recent features of the earth's
surface. The second is interested in the relationships of the human
group to its environment today. Surely one cannot do without the
other. What strikes the archaeologist is that the geographer, in his
analysis of the landscape, must separate human from natural features.
This is exactly what the excavator is doing when studying sections
and plans in the field.
Both archaeology and geography are only interested in man as a
group; the individual does not play any role. In prehistory this is a
60 P. J. R. MODDERMAN
necessity because the individual is but exemplary of the group. The
geographer realises that the decisions of one person may change the
landscape; however, such a person is restricted by the possibilities
of the community and of natural circumstances. Moreover, geography
and prehistory agree in studying human groups only in their relation-
ship with the environment. For the geographer it is a task, for the
archaeologist it is a necessity, because the data have a restriction in
themselves. Facets of the human mind like language, music and many
aspects of religion vanish completely. To geographers these aspects
play only a subordinate role.
The great difference between geography and archaeology is that
geography tries to understand and explain the recent situation,
whereas archaeology wants to reconstruct the past with recent data.
Geographers have the advantage over prehistory in that the human
group is there to be analysed, whereas past groups are difficult to
reconstruct. To the archaeologist the recent situation is the result of
the past; and this result is directed by purely natural processes and
human manipulations of the environment, as well as by natural
transformations resulting from human behaviour.
Geographers may strongly emphasize social relationships, and this
is also found in archaeology. Prehistorians use in their reconstructions
models derived from recent situations. With luck a high probability
is reached, and to this extent there may be interesting contributions
to the understanding of the past; but such reconstructions cannot be
more than fringe phenomena. The main problems for archaeologists
lie elsewhere. One may often wonder whether prehistory is more
than just the history of technology. Undoubtedly the description of
technological knowledge and its changes is emphasized strongly; but
there is more to it than that. The technology of stone, wood and
metal show what solutions man discovered and preferred in order
to improve life. Conscious and subconscious factors play a role in
the human mind in this respect. With the help of his technological
knowledge man is in search of the most pleasant way of living. In
his quest for improvements man is driven forward, and if necessary
changes are carried out.
So far we have discussed the theoretical base of what is to follow.
In the background are the ideas of the French geographers Jean
Brunhes and Vidal de la Blache. The second part of this paper will
deal with what can be done in practice with these considerations.
My archaeological colleagues can be assured that once again geogra-
phy will stand near the rostrum.
The interaction of the human group and its environment cannot
be understood without at least some knowledge of demographic
matter. In this respect the geographer has a great advantage over the
archaeologist because recent population can be studied in every
detail, whereas archaeologists have to make difficult reconstructions
 Thoughts On Archaeology. Geograpjy and Scales 61
for which palaeodemography is of great assistance. At any rate, the
possible marginal values can be mentioned. Archaeologists should
realise that they furnish the data for palaeodemographers. This means
that already at the level of an excavation close collaboration with
anthropologists is necessary. Questions about family size, the number
of inhabitants in a settlement or area, population growth etc. are
basic for forming hypotheses about cultural distributions and change.
This is not to say that such general questions can be solved by
palaeodemography, but they certainly lead to a better understanding
of what happened in prehistory.
The landscape, as the counterpoint to the interaction of man and
his environment, attracts much attention. It is not easy to obtain the
necessary data for the reconstruction of the prehistoric environment,
but attempts are greatly in vogue. One of the problems is that man
himself has spoiled so much information in the past 7,000 years. The
recent situation is not only the result of natural processes; human
beings have intruded greatly. However, reconstructions of past life
cannot do without intensive research of environmental conditions.
Reconstructions of human groups and their environment are
extraordinarily static. They are like snapshots; you do not know what
came before and afterwards. The interaction, which is changing every
second like a movie, is missing. Every living organism has its relation-
ships with the environment; everyone is in search of the best guaran-
tees for existence, which leads to a great variety of activities, and
the result is an ever changing picture. Man has a great advantage
over all other organisms in that he is so well equipped to adapt to
new situations and possibilities. He is the most flexible of all, in the
interaction of living organisms and their environment. With his
technology he seems able to cope with all the problems posed to
him, but it is clear that the environment has always laid restrictions
upon him.
An important feature of man's flexibility is his ability to move
freely in various environments. On foot a small group or an individual
can cover great distances. To us it is amazing how walking was still
so common in the 19th century. The railway has changed this pattern
greatly. Footpaths, when abandoned, are in general untraceable.
However, we must accept that pedestrians played an important role
in the contacts of sedentary groups. They are the bearers of cultural
transport, regardless of whether they carried material objects or
non-artefactual ideas.
Every human activity has its own 'space'. A hunter walks for miles
and miles, whereas the cooking of food is restricted to some square
metres. These differing distances can be measured on the surface of
the earth. A great help is to produce a map, a well known task for
archaeologists. Excavation plans, distribution maps etc. are integral
62 P. J. R. MODDERMAN
to the discipline and scales between 1:5 and 1:50 are in great use.
Which scale is chosen depends on the density of data; the more there
are, the smaller the scale will be.
In general the excavated surface is only a small part of the area
occupied by man. It is staggering to realize how many data are
assembled on an excavation, whereas just outside its borders we are
in the world of supposition as to human activities. To cope with
these difficulties we look for help from topographical, soil and geo-
logical maps. Unfortunately these maps are made for other purposes,
certainly not to serve archaeologists. Moreover, the scales on which
they are published differ greatly from those used by us and on which
the daily life of our ancestors took place. Nearest to our purpose
come topographical maps of 1:10,000 or 1:25,000. Geologists and
pedologists, however, commonly publish maps on 1:50,000. We do
realize that the latter assemble their data on 1:25,000 maps, but there
is a great gap from that to the excavation drawings of 1:20 or 1:50.
The reason for these extreme differences is again the density of the
assembled data. To geologists and pedologists there is no sense in
using a scale of 1:50. Soil survey maps, at 1:50,000 for instance, are
based on one boring every 4-8 ha. There are, of course, more detailed
maps, but they seldom reach a greater density than one boring per
ha. Translated on to a scale of 1:50 this means one datum on a sheet
of paper of 2 2 m. In reverse, the archaeologist cannot display his
data on a scale of 1:10,000. The problems with which soil scientists
and archaeologists are dealing cannot be mappped, let alone studied,
on the scale at which the other is working. Particular mention has
been made of soil and geological maps because they are widely used
to aid the reconstructions of past environments. Human traces are
hardly found on them. However, Dutch soil maps are an exception,
due to the fact that Dutch soil is like earth in a flowerpot; natural
soils are hardly found in the Netherlands.
Of course geographers also use geological and soil maps, but
topographical maps are very much in favour because they portray
the recent situation. To archaeologists it is interesting to notice what
map scales are used by planners. They try to construct the future on
the basis of the recent situation, as archaeologists try to reconstruct
the past. As an example, planning procedures in the Netherlands
will be outlined, accepting that in general there are no great differen-
ces between many countries in this respect. The Dutch planning
system distinguishes several stages. In practice every stage uses maps
on a scale which is most appropriate to its needs. National planners
prefer 1:250,000 and 1:100,000 maps, whereas municipalities work
on scales from 1:10,000 to 1:1,000. One could continue this series by
involving architects who often draw their plans at scales of 1:50 and
1:20. These are the scales familiar to archaeologists.
 Thoughts On Archaeology, Geograpjy and Scales 63
Out of the preceding discussion it must be stressed that a tremen-
dous gap exists between the scale of 1:50 used in excavations and
that of 1:50,000 soil and geological maps which are used to reconstruct
environmental conditions. Many human activities cannot be covered
by any one of these scales. From excavations we gain knowledge
about the results of many activities, but they do not give any informa-
tion about the space necessary to fabricate an artefact. Man, in
making his choice of a new settlement site, pays attention to these
factors with his technology in mind. He is in a similar situation to
the municipal planner of today, who works with maps on scales
1:10,000 to 1:,000. Archaeologists should direct their thoughts and
research in this field; it means that they must identify possibilities
for collecting data which can be best mapped on scales 1:100 to
1:2,500. In concrete form, some 30-40 observations per ha are needed.
If archaeological research is aiming at the reconstruction of human
activities resulting from the relationships between man and his
environment, then we should realize that we have to use all possible
scales. Every activity has its own level of scale. For example there
is a great difference in the occupation of the earth's surface between
a parcel of arable land and a habitation site. It can be expected that
the intensity of occupation of the field is much smaller than in the
settlement. This means that an ability to work at different scales is
necessary in our research of the space occupied by man. On the one
hand, a reconstruction of the available natural circumstances is
obligatory, on the other, the composition of the human group fixes
the interaction between them. Another factor of great importance is,
of course, the technological level at the time of the reconstruction.
Restricted by these themes, archaeologists should try to propose
hypothetical models and then test them by trying to make observa-
tions as they were perhaps made by our ancestors.
Although there are very promising studies which deal with human
activities outside settlements and burial grounds, they are still
incidental and mostly instigated by the fact that recognisable features,
such as Celtic fields or ditch systems, can be traced and mapped.
These studies should be more purposeful. A detailed soil map can
be useful, but soil research should be done with archaeology in mind,
as with biologists who are working in archaeological institutes where
their thinking is directed to archaeological questions. In the same
way aerial photography from heights of 20 to 500 m might give new
insights. Micro-surveying and phosphate sampling are other
possibilities. The key point is that human activity can be investigated
at different scales which is impossible without mapping them at these
scales.
 The Tenta Catchment: Resource Exploitation in
Neolithic Cyprus

Malcolm Wagstaff,
Southampton University

The conceptual importance of Site Territorial Analysis (STA) is now

widely recognised by archaeologists (Bailey and Davidson 1983;
Roper 1979). This paper outlines some of the problems, general and
specific, which arose in attempting to apply the concepts to a par-
ticular situation.
The site for which STA was attempted is Tenta (or strictly,
Kalavasos-Tenta). It is an aceramic Neolithic site of mid-sixth millen-
nium date (a mean of 9 Carbon 14 dates is 5530 B.C., Todd 1982a),
occupying the summit of a low hill on the western side of the Vasilikos
valley. The Vasilikos river occupies the westernmost of a group of
subparallel valleys (Figure 1) running roughly south-south-east to
the south coast of Cyprus through a fairly narrow zone of plateaux
cut across chalks and limestones of Upper Cretaceous to Miocene
date (Bagnall 1960, Pantazis 1967). The nature of the aceramic
occupation is similar to, and roughly contemporaneous with, that of
the better-known Khirokitia a few kilometres to the east (Karageor-
ghis 1968: 36-106). The site was first examined by Dikaios in 1940
and 1947, whilst excavation of the summit of the hill took place in
1976-79 under the direction of Todd (Todd 1977, 1978,1979,1982b),
with a final season in 1984.
As well as attempting to unravel and understand the cultural
deposits on the site, an early decision was made to try to assess the
resource potential of the area around the site (Wagstaff 1978). The
specific aims were, first, to establish the possible subsistence base of
the inhabitants and, second, to discover locations which would have
been suitable for the production of different types of food (par-
ticularly cereals), evidence for which could be expected from the
excavation. There was also the hope that the exploration of the
relations between potential and actual land use in the sixth millen-
nium would suggest reasons for the initial choice of settlement site.
Formulation of the objectives in this way was conditioned by the
assumption that some form of STA would be a suitable means for
discovering the possible relationships between subsistence strategies
and the natural environment. The investigators were not originally
 The Tenta Catchment 65

Figure 1. Tenta and the catchment of the Vasilikos River.

concerned with some of the other objectives since recognised for

STA by Bailey and Davidson (1983). That is, we did not seek to
establish either the function of the site or its status, nor did we
attempt to investigate the relationships (spatial and diachronic)
between sites in the area. It was assumed that these would be
studied after the completion of the excavation and the
accompanying site survey.

The attempt to apply STA raised a number of problems, some of

them subsequently isolated by Roper (1979) and Bailey and
66 WAG STA F F
Davidson (1983). Those problems of a conceptual nature can be
divided into two sets, general and specific. The general problems
arose from the decisions, first, to define a site territory ('catchment')
in terms of economic distance, and then to use off-site data to discover
patterns of potential use within it. Distance was the defining criterion
originally selected when STA was first advanced (Higgs and Vita-
Finzi 1970, 1972). Walking-time along a minimum of four orthogonal
traverses was the suggested metric, with the outer limit of the site
territory being set by empirically derived thresholds published in the
ethnographic literature. Whilst the technique takes account of the
differential effects of terrain on the ease and speed of travel, it is
time-consuming to implement and necessarily involves much interpo-
lation to delimit an actual boundary. Alternatives have been suggested
and applied. An early one was the use of linear or geodesic units
such as kilometres to define a circle of terrain exploitable from a
particular site. Again, empirically derived thresholds are employed
to set the outer limits, but no adjustment is possible to terrain or
ease of travel. Recently, Bailey and Davidson (1983) have advocated
the use of isochronic distance. This is based on Naismith's formula:
the average time needed to walk 10 km on the flat is 2 hours, but
half an hour must be added for every 300 m of elevation. The method
has the advantages of allowing for topographical variation on the
configuration of site territories at all points and, being easy to apply
to the analysis of topographical maps, of removing the need for much
time-consuming field work and more or less extensive interpolation.
Whatever the metric employed, the delimitation of boundaries by
distance relies upon two assumptions. The first is that economic
distance is a dominant constraint on resource exploitation irrespective
of time, place and culture (Binford 1983: 204). The second, which
follows from the first, is that the intensity of exploitation will decrease
with distance from an activity centre. The whole notion derives from
concepts of economic marginality formulated for capitalist
economies. It was presumably these origins which allowed Roper
(1979) to write in terms of cost/benefit ratios and Bailey and Davidson
(1983), more misleadingly perhaps, to talk in terms of a positive
trade-off between the amount of energy acquired as food and the
amount of energy expended in travel and exploitation. Unfortunately
though, the use of economic distance in this deterministic way is
fundamentally ahistorical. Evidence is needed not only that the
tyranny of economic distance has operated always and everywhere,
but also that people have at least been conscious of some notion of
ratios and energy consumption. On a priori grounds alone, doubt
may be cast on the idea of energy gained over energy expended, and
cost-benefit ratios must ideally be inferred from archaeological data
from the site about the social values of its former inhabitants. The
 The Tenta Catchment 67
critical limits for territorial exploitation are likely, in practice, to
have been determined more by drudgery, as in Chayanov's model
of peasant economy (Chayanov 1966), than some form of marginal
profits, however expressed. Whilst Chayanov recognised that the
intensity of land use would fall off with distance (Chayanov 1966:
178-79), drudgery is a subjective idea a perception which will
be determined by the values of particular cultures. The idea that
distance is not necessarily deterministic in fixing the limits of site
territories is bolstered by work done on the agricultural land use of
Melos in the 1970s. The finding there was that people actually
travelled up to 6 hours from their homes, but a mean distance of 2
hours and a median of 1.3 hours to cultivate property which they
owned (Wagstaff 1976: 18). There was no statistically significant
falling off with distance in the intensity of land use, at least as
measured by the labpur inputs of the crops grown on the most distant
plots (Wagstaff 1976: 20-23). It must be admitted, though, that
Sallade, working on cropping practice around Tokhni village in
Cyprus, did find the declines in intensity of land use with distance
predicted by von Thiinen's model (Sallade and Braun 1979).
Even if we accept the correctness of the assumption that economic
distance will define the boundary of a site territory, there is still the
question of the critical thresholds to be considered. The empirical
evidence collected by Chisholm (1962) suggested that the absolute
spatial limits of a site territory exploited by peasant agriculture would
be set at a distance of 5 km, a walking time of about an hour, from
the operational centre. Greek data seem to suggest that the outer
limit could probably be extended to perhaps twice the distance
proposed (Thompson 1963: 42-199; Wagstaff 1976: 18). The Melos
study showed further that field houses (spitakia) were used on a
seasonal basis when land was worked at or beyond the 10 km
threshold (Wagstaff 1976: 23).
Another point to make is that Chisholm's data were derived from
farming systems where the use of a plough was normal and transport
was assisted by animals. The critical limits for agricultural site ter-
ritories would be narrower in situations where transport animals were
not employed and land was prepared with digging sticks or hoes.
Lewis (1949), for instance, showed how in the territory of the village
of Tepoztlan, Mexico, the use of hoes was much more labour intensive
than plough cultivation and the work itself much more exhausting.
On the other hand, Lewis also reported that the yields were much
higher than from ploughed land, partly because of the thorough way
in which weeding was done, whilst the pressure of available time
was much reduced since bad weather inhibited work in the fields to
a far lesser extent than where draught animals were used. Bearing
this type of ethnographic information in mind, it is clearly important
68 WAGSTAFF
for the excavators to look for on-site data for the level of technology
used by the inhabitants.
The inadequacy of on-site data was the most fundamental of the
three specific problems encountered in trying to apply STA to the
Tenta area. Excavation revealed surprisingly little evidence for the
relative importance of cultivation. The recovery of seeds from culti-
vated wheat, barley and lentils (Hansen 1979) suggests that the
inhabitants of Neolithic Tenta were either engaged in farming them-
selves or had access to people who were, but the percentage of
cultivated seeds (17-20%) in the total studied in 1978 is probably
too low for larger claims to be made about the significance of
cultivation to the Tenta community. Unfortunately, other evidence
is not available to support the case for an economy based around
cultivation. Many of the seeds recovered were from wild plants.
'Surprisingly few implements for pounding or grinding have been
found' (Todd 1978: 180) and the presence of just a few does not
confirm cultivation since these implements could have been used to
process a variety of foodstuffs and other materials. No granaries have
been positively identified. It is, therefore, not at all clear that we are
dealing with a community for which cultivation was of great sig-
nificance. The use of cultivation thresholds to define a site territory
may thus be mistaken. In fact, there is some evidence to suggest that
we should think in terms of a larger territory.
Examination of the lithics recovered from Tenta showed that the
most common raw material (a grey chert) must have been brought
at least 5 km down the valley (Hordynsky and Kingsnorth 1979).
Hunting is implied by the presence of red and especially fallow deer
bones amongst the faunal evidence (Legge, 1978). Its relative import-
ance in the Tenta economy is unknown. Even so, analogy with the
practice of the Kung bushmen, much quoted in the literature, suggests
a possible distance threshold to the Tenta site territory of 10 km or
2 hours walking time (Lee 1969), whilst the use of temporary hunting
camps would extend this much further.
At this point a second specific problem becomes important.
Although it was not the objective of the STA for Tenta to study either
the status of the site or its relationships with other, possibly contem-
poraneous, sites, these questions now emerge as crucially important.
The substantial nature of the architecture at Tenta seems to indicate
year-round occupation of the site. At the moment, though, there
appears to be no other evidence to support the contention, and
seasonal occupation is at least plausible. Firm evidence on the type
and permanency of occupation is badly needed. Similarly, the inter-
pretation of the settlement at Tenta requires an understanding of its
relationships with another and probably more extensive aceramic
 The Tenta Catchment 69
Neolithic site (Mari-Mesovouni) lying on a conspicuous mesa-like
hill some 3 km to the south-south-east (Todd 1979: 285). Excavation
is required there. Finally, the attempts to put STA into operation
around Tenta soon threw up the problem of terrain as a major
constraint on the spatial patterns of resource exploitation in the area.
At the present time, travel is easiest along either the valley floor or
the edge of the hills on either side. Transverse communications are
much more difficult because of steep slopes along the valley sides,
especially on the east. Terrain would also have had other effects on
the activities of Tenta's inhabitants. It would have differentiated the
availability of various natural resources such as fruit and nuts for
gathering and animals for hunting. Terrain would also have condi-
tioned the availability of land suitable for cultivation under different
technologies and, through the quality, depth and stability of the soil
cover, its productivity. In addition, whilst it is reasonable to suppose
that the broad framework of the hills and valley have stayed stable
throughout the period of human occupation, no such claim can be
sustained for the detail of the topography. Natural processes, as well
as the complex interaction of natural and anthropogenetic processes,
have brought about many changes, some remnants of which remain
in the modern landscape. Several of these may be mentioned. The
existence of marine and fluvial terraces in the area points to higher
sea levels, some of them during the period of human occupation.
The character of the fluvial sediments in the main valley shows clearly
that the river has changed its position in detail many times, mainly
due to hydrogeomorphic processes but also in response to various
forms of human interference. Alluvial and colluvial deposits, though
largely undated, probably resulted from erosion occasioned by wood-
land clearance and the extension of cultivation. To some extent, the
adverse effects of erosion were countered by the construction of
terraces. 'Cross-channel terracing' in the tributary valleys and 'lateral
or contour terraces' on the hill slopes (Donkin 1979: 32) have retained
material carried out by hill wash and fluvial action, bringing about
a smoothing of valley and slope profiles which may have been
uncharacteristic of the sixth millennium. Dating of the terraces is
thus important, whilst reconstruction of the terrain is another task
if the Tenta site territory is to be adequately understood.

A number of conclusions may be drawn from the attempt to apply

STA to Tenta. First, the form of STA adopted assumed that economic
distance is a dominant, diachronic constraint on the relationships
between man and his environment in the spatial patterning of human
activity across the landscape. This should now be recognised as
ahistorical. If the concept is to continue in use, then the questions
of suitable thresholds must be settled. Thresholds are best set up by
70 WAGSTAFF
the economic and technological data either recovered directly by
excavation or deduced from the finds. It is absolutely imperative that
careful attention should be given to this aspect of recovery. More
thought should also be given to ways of assessing the relative import-
ance of different activities in a total economy and, where cultivation
can be documented or at least inferred, to the appropriate ways of
distinguishing plough from hoe techniques.
A second conclusion is that the assumptions of the capitalist model
of resource exploitation, with its apparent emphasis upon returns on
capital invested, must be reviewed. A labour-based theory of exploita-
tion may be more appropriate. Almost certainly this would increase
the conventional thresholds used in STA, since the usual assumptions
about marginality may not apply.
A third conclusion is that it is basically erroneous to suppose that
the territory of any one site will be unaffected by the extent of the
territory around other sites in the general area. Thus, it is important
to locate these sites and to establish their chronological, as well as
their spatial, functional and hierarchical relationships. Boundaries,
of course, were probably not set by principles of equality, any more
than by economic thresholds. Power relationships would also be
important, even if difficult to document archaeologically. *
The final conclusion is perhaps equally obvious. The Tenta
experience has shown the need to make an appraisal of the terrain
within the site territories, especially of those aspects of it which, on
the time-scale of the occupational history of the focal site or sites,
are subject to rapid change. Only by making allowance for such
changes, and locating them as precisely as possible in real space,
will it be possible to assess the full extent of the geographical
constraints upon the economic system focussed in a particular
archaeological site at a specific point in time.
BIBLIOGRAPHY
Bagnall, P. S. (1960) The Geology and Mineral Resources of the
Pano Lefkara Area. Nicosia: Geological Survey Department,
Cyprus.
Bailey, G. N. & I. Davidson (1983) Site exploitation territories
and topography: two case studies from Palaeolithic Spain,
Journal of Archaeological Science 10, 87-115.
Binford, L. R. (1983) In Pursuit of the Past: Decoding the
Archaeological Record. Thames and Hudson, London.
Chayanov, A.-V. (1966) The Theory of Peasant Economy, edited by
D. Thorner, B. Kerblay and R. E. F. Smith. Richard D. Irwin,
Inc., Homewood, Illinios.
Chisholm, M. (1962) Rural Settlement and Land Use. Hutchinson,
London.
Donkin, R. A. (1979) Agricultural Terracing in the Aboriginal New
World. Viking Publications in Anthropology 56. University of
Arizona Press, Tucson. Arizona.
 The Tenta Catchment 71
Hansen, J. (1979) Botanical studies in the Vasilikos Valley, 1978,
Journal of Field Archaeology 6, 297-98.
Higgs, E. S. & C. Vita-Finzi (1972) Prehistoric economies: a ter-
ritorial approach, pp. 27-36 in Papers in Economic Prehistory
edited by E. S. Higgs. Cambridge University Press, Cam-
bridge.
Hordynsky, L. & A. Kingsnorth (1979) Lithic reduction sequences
in the Vasilikos Valley: methods and preliminary results,
Journal of Field Archaeology 6, 287-90.
Karageorghis, V. (1968) Cyprus. Archaeologia Mundi. Nagel,
Geneva, Paris and Munich.
Lee, R. B. (1969) Kung Bushmen subsistence: an input-output
analysis, pp. 47-79 in Environment and Cultural Behaviour,
edited by A.P. Vayda. Natural History Press, Garden city
New York.
Legge, A. (1978) Faunal remains, Journal of Field Archaeology 5,
184-85.
Lewis, O. (1949) Plow culture and hoe culture - a study in con-
trasts, Rural Sociology 14, 116-27.
Pantazis, T. M. (1967) The Geology and Mineral Resources of the
Pharmakas-Kalavasos Area. Nicosia: Geological Survey
Department, Cyprus.
Roper, D. C. (1979) The method and theory of site catchment
analysis: a review, pp. 119-140 in M.G. Schiffer (Ed.),
Advances in Archaeological Method and Theory Vol. 2.
Academic Press, New York.
Sallade, J. K. & D. P. Braun (1979) Spatial organisation of
peasant agricultural subsistence territories: distance factors
and crop location, pp. 19-41 in E. Tooker (Ed.) Ethnography
by Archaeologists. The American Ethnological Society, Wash-
ington.
Todd, I. A. (1977) The Vasilikos Valley Project, 1976, Journal of
Field Archaeology 4, 375-381.
Todd, I. A. (1978) Vasilikos Valley Project: second preliminary
report 1977, Journal of Field Archaeology 5, 161-195.
Todd, I. A. (1979) Vasilikos Valley Project: third preliminary
report, 1978, Journal of Field Archaeology 6, 265-300.
Todd, I. A. (1982a) Radiocarbon dales from Kalavasos-Tenta and
Kalavasos-Ayious, Report of the Department of Antiquities,
Cyprus, 1982, Nicosia, pp. 8-11.
Todd, I. A. (1982b) Vasilikos Valley Project: fourth preliminary
report, 1979-80, Journal of Field Archaeology 9, 35-77.
Vita-Finzi, C. & E. S. Higgs (1970) Prehistoric economy in the
Mount Carmel area of Palestine : site catchment analysis,
Proceedings of the Prehistoric Society 36, 1-37.
Wagstaff, J. M. (1976) Aspects of Land Use in Melos. Department
of Geography, University of Southampton.
Wagstaff, J. M. (1978) Geographical contribution to the Vasilikos
Valley Project, 1977, Journal of Field Archaeology 5, 187-189.
On-site and Off-site Evidence for Changes in Subsistence
Economy: Iron Age and Romano-British West Wales

H. C. Mytum,
York University, England

The peninsula of West Wales consists of a central spine of the Preseli

mountains which join the western part of the central Welsh massif,
around which lies rolling countryside that leads down to the sea. On
the coast a series of cliffs is interspersed with sandy bays. The only
major rivers are the Eastern and Western Cleddau running south
from the foothills of the Preseli mountains, and the Nevern and Teifi
which flow north-west from the same upland area.
Despite considerable destruction of archaeological sites, especially
in the southern part of the region, many settlements of the Iron Age
and Romano-British period survive, both as earthworks and as crop-
marks. Survey and listing of these sites has a long and honourable
history (R. C. H. M. 1917, 1926, Savory 1951, Crossley 1963) and
limited excavation has been carried out at several settlements such
as Merryborough Camp and Knock Rath (Crossley 1964, 1965) and
Woodbarn Rath (Vyner 1982).
Extensive excavations have been conducted in a rescue context by
Wainwright (1967, 1971) and more recently by the Dyfed Archaeo-
logical Trust, particularly in the Llawhaden area. Detailed discussion
of the latter work must await the definitive publication, although
interim reports are available for the individual sites (Mytum 1982b,
Williams 1984, 1985). In addition, data from field survey of the
Eastern Cleddau valley (Mytum 1982c) and the continuing research
excavations at Castell Henllys (Mytum 1982a, 1984) can be con-
sidered, and use made of the work by the Soil Survey of England
and Wales (Rudeforth and Bradley 1972, Rudeforth 1974). Investiga-
tions are still continuing, but enough has now been done to allow
consideration of settlement change and the economic and environ-
mental factors causing and resulting from this change. Previous
general works on the region by Hogg (1972) and Williams (1978,
1979) have not considered this aspect.
During the Iron Age a series of defended settlements were construc-
ted, often on scarp edges where a tributory stream joins a larger one.
Some could be most accurately described as inland promontory forts.
Although sometimes overlooked by slightly higher ground it is clear
 Iron Age and RomansBritish West Wales 73
that such settlements were located and designed with a significant
defensive function in mind. Excavations at Castell Henllys have
shown the construction of a substantial double bank and ditch on
the one easily approached side of the fort, and on the others a smaller
bank was combined with an artificial steepening of the already
impressive scarp. The construction of such forts can be tentatively
dated to the fourth century B.C. on the basis of a few Carbon 14
dates from sites such as Holgan Camp (Williams 1984), and occupa-
tion lasted on these sites for some time. Extensive excavations at
both Castell Henllys and Walesland Rath produced evidence of
numerous roundhouses suggesting a population of perhaps 5-10
families in such sites. These forts are the largest common settlements
in West Wales but are far smaller than the traditional hillforts of the
Marches and Southern England.
At a later date, perhaps in the first century B.C., there was a
significant change in the settlement pattern. The defended settlements
were abandoned, and were replaced by a dispersed pattern of small
enclosed farmsteads scattered over a much wider area. Whereas
previously settlement had clustered along the edges of certain valleys,
there was now expansion onto the tops of the gently rolling hills.
Many of these locations were of little defensive value, and the same
is the case for other hillslope enclosures which were constructed near
to the earlier fortified sites. The detailed evidence for the exact timing
and cause of this change has yet to be set out, but it is clear that a
major settlement shift took place near the end of the first millennium
B.C. involving a change from few defended nucleated settlements to
many more smaller enclosed farmsteads for individual families. The
purpose of this paper is to examine the economic implications of
this change, and to consider some of the conceptual issues raised by
these, since many relate to other regions and periods.
There are undoubtedly important social reasons for, and effects
of the change in settlement pattern, but economic factors can also
be considered highly relevant. The on-site evidence for the subsistence
economy will be considered first, followed by that of off-site
archaeology. Finally, some of the possibilities for further testing of
hypotheses will be suggested, including ones which might more
effectively link these two sets of data.

The soils of west Wales are generally acid, and so the preservation
of faunal remains is poor. The only major exception that has been
excavated is Coygan Camp (Wainwright 1967) where considerable
quantities of bone were recovered, although the published report
does not provide the sort of detail which would now be considered
desirable. At other sites, little bone is recovered except where this
has been burnt. At Castell Henllys, burnt bone is ubiquitous in
74 MYTUM
occupation and rubbish deposits but unburnt bone only survived on
the surface of the buried soil under the large clay defensive bank
(Mytum 1982a).
As yet it is difficult to examine the pastoral element of the economy
from the excavated evidence, with only one site having significant
quantities of faunal material. It should be emphasised, however, that
even if other suitable sites are found the evidence may not be of a
kind easily open to interpretation about economic changes. The
reason for this is the difficulty in comparing assemblages of different
periods because the contexts from which they come are so varied.
For the Iron Age period at Coygan Camp when the fort was construc-
ted and in use, a total of 613 animal bones were recovered from the
whole site, in contrast to 4,732 for the Romano-British period ( Wain-
wright 1967: 191-2). There is no reason to see the later occupation
as of longer duration or of a larger group, in fact one would assume
the reverse. The Iron Age deposits consist of structural elements
associated with the defences the earthworks and gateways and
some Occupation deposits' found in some places behind the rampart.
In contrast, the material from the Romano-British phase comes from
each of the four buildings, and from a large rubbish deposit to the
rear of the west rampart (site B). The presence of actual structures
which contain a large amount of refuse accumulated during occupa-
tion and a midden which was used for the organised disposal of
further rubbish, allows us a much fuller picture of the material culture
and economic strategy in this later period. Comparisons with the
earlier phase, however, are frustrated. Discussion of a change to or
from a subsistence economy with a greater emphasis on pastoral
activity cannot be entertained on the evidence from here alone.
When the problem of cereal cultivation is considered, the evidence
available is again limited. The extensive excavations at Coygan Camp
and Walesland Rath did not include a policy of soil sieving or
flotation, and the same was true of all the earlier limited excavations
on farmstead sites. The recent Llawhaden excavations included a
substantial sampling programme though the results from this are yet
to be made available. At Pembrey Mountain, a defended site which
has been partially examined, some remains of carbonised grains were
found both from pre-rampart and later contexts (Williams 1981).
Greater numbers came from features later than the construction of
the defences, but in all cases the number of grains was low (Hillman
1981). The only coherent building identified was a square four-post
structure whch was interpreted as a granary.
The more extensive excavations at the morphologically similar site
of Castell Henllys can put this evidence in some sort of context.
Samples of carbonised grain from Iron Age features again tend to
be small, whether from post holes, gullies or Occupation layers'. The
 Iron Age and RomansBritish West Wales 75
only large sample came from a burnt layer with many fragments of
twigs and larger timber elements to the rear of the rampart. Four-post
structures are found on the site; one of these was adjacent to the
burnt layer already mentioned although a relationship between the
two has yet to be proved. At Castell Henllys it is clear that the
four-post structures appear late in the sequence of occupation, as
one cuts across several concentric roundhouse gullies. As yet there
are no four-posters early in the sequence. Evidence for underground
storage of grain in pits has not been found in West Wales.
The small size of the grain samples makes any assessment of
changes through time within the Iron Age difficult, though the struc-
tural evidence might suggest an increase in grain storage (or a different
method of storage) late in the occupation of Castell Henllys. Grain
processing may not have taken place on-site since only clean grain
was found at Pembrey Mountain (Hillman 1981) and in the samples
so far examined from Castell Henllys. The fact that the grain has
been cleaned does suggest intended human consumption rather than
cereal production for animal feed.
Artefactual evidence for cereal processing comes from querns.
These would be expected to give some index of such activity at least
within a broad time phase since broken or worn querns could be
used in flooring, revetments, or packing of post holes. It is interesting
to note, therefore, the limited number of such finds from Iron Age
contexts. Coygan Camp produced two saddle querns (Wainwright
1967: 164); Walesland Rath (Wainwright 1971), and Castell Henllys
had no querns at all. This amount of negative evidence suggests that
grain was not ground, that querns were systematically discarded
elsewhere, or that grinding was not done in the settlements. One
might assume that grain was not a major element in the economy if
is was not being ground extensively, though the importance of cereals
as a foodstuff might have increased towards the end of the Iron Age
when the four-posters became common.
The evidence for the end of the Iron Age, and the Romano-British
period when the smaller farmsteads were in use, is far more plentiful.
Some of the larger sites continue to be occupied, or more likely are
sometimes re-occupied, as at Coygan Camp and Walesland Rath,
but if so this is by the smaller family group and the old enclosure is
clearly being used out of convenience and the defences are not
refurbished. At Castell Henllys a settlement is established beyond
the old fort site, in an area called the annexe in published accounts.
Many of the Llawhaden farmstead sites have four-posters, and so
far two have been found at Castell Henllys (Mytum 1984). They were
not present, however, at Coygan Camp and Walesland Rath.
More impressive evidence for an increased importance of cereals
comes from the numbers of querns. Coygan yielded 17 saddle and
76 MYTUM
three rotary querns, and the Romano-British phase at Castell Henllys
has already yielded over 10 rotary querns. Not all sites, however, are
so rich: Walesland produced just one stone, and Drim none at all.
Other evidence for cereal production comes from the iron plough-tips
from Coygan Camp and the wooden ard-tip from Walesland Rath
(Wainwright 1971: 94; Manning 1975: 114 who re-interprets it as
such).
The carbonised grain samples from Castell Henllys also hint at an
increased amount of cereals, but once more the contexts of all finds
needs to be considered. At all the sites where Romano-British deposits
are found, there are clearer middens, dumps and refuse left in
buildings. It may be the greater amount of rubbish deposited that
gives the impression of increased arable production. There were,
after all, nearly eight times as many animal bones in the Romano-
British contexts as in the Iron Age ones at Coygan Camp, so the
ten-fold increase in quern fragments at that site may not be significant.
The same could apply to the carbonised grain at Castell Henllys
since most of the later layers analysed also contained many artefacts,
in contrast to the Iron Age ones which did not. The nature of the
contexts and the patterns of rubbish disposal (Schiffer 1976) may be
what was changing, and not the underlying economic base. Excavated
evidence has, for this reason, to be treated with some caution, and
it is with this in mind that we turn to the off-site evidence to provide
an alternative approach to the study of subsistence strategies.

Examination of the topographic location and soil associations of

sites within the Eastern Cleddau study area is in the process of
completion, but a small area of the Western Cleddau, including
within it the excavated site of Walesland Rath, can be considered
here. In microcosm this represents the type of developments that
occur in the Eastern Cleddau valley, and in most respects those of
the Nevern Valley system in which Castell Henllys is situated.
Pelcombe Brook flows west to east down a steeply cut valley to
join the Western Cleddau just north of the town of Haverfordwest.
Unlike similar valleys of Knock Brook to the north and Merlin's
Brook to the south, a number of settlements have been preserved,
though even here destruction has probably already taken its toll.
Smaller streams join Pelcombe Brook from the north and south, also
in steep-sided valleys. Between these lie gently sloping hills with flat
or slightly rounded tops. The steep valley sides have shallow
undeveloped soils, and the valley bottoms are both constricted and
poorly drained. In contrast, most of the higher ground is covered by
brown earths which are fertile, easily worked and well drained
(Rudeforth 1974). The landscape can therefore be easily divided into
two discrete zones: those of modern capability class 3 which are
 Iron Age and RomansBritish West Wales 77
suitable for intensive grazing and arable production, and those of
classes 5 and 6 which are only for use for low quality grazing. There
is very little land of intermediate class 4 in this area.
Two sites, Walesland and Pelcombe Raths, can be considered as
defensively sited, with possibly Crowhill Rath as well. The other sites
are the smaller, less defensive settlements occurring on valley sides
in the cases of East Hook Rath and Slade Farm, and on the higher
ground with Cuckoo Grove Rath. In two cases there is a definite
pairing of sites with one defensive and one less defensive site on
opposite sides of the valley. This may indicate settlement shift. East
Hook Rath for example may represent the settlement occupied in
the break in the sequence at the excavated site of Walesland Rath
to the north. Alternatively, both sites may have been in use together
since the farmstead sites each contained smaller populations than
the earlier defended sites.
The distribution of known pairings suggests that another pair may
have been destroyed around Sunnyhill where ideal topographical
situations are available. Tentative locations have been added to the
distribution. Only limited destruction of sites should be expected
along the vally scarps as these have not been intensively farmed in
periods subsequent to the settlements. In contrast, the better soils on
the flat and gently rolling hilltops have been much used, and con-
sequently destruction has been far greater. For this reason, Cuckoo
Grove Rath should be considered merely as a representative of a
once far more common settlement form to the south of Pelcombe
Brook. Aerial photography in Pembrokeshire is adding many more
sites of this kind to the distribution maps whilst altering the distribu-
tion of more defensive sites far less (James 1984).
The size of site exploitation territories belonging to the settlements
is difficult to infer. Even one-hour travelling distances would enclose
far more land than could have been intensively exploited. More
seriously, the numerous settlements which can be considered to be
roughly contemporary would suggest much smaller areas. Using the
methods outlined by Bailey and Davidson (1983), ten minute ter-
ritories have been defined (Fig. 1). These do not overlap, unlike the
twenty minute territories, and so these small areas may be considered
the main exploited area (Fig. 2b). As a test of this, the hypothetical
territories using Thiessen polygons were constructed, and these also
separate the landscape along the valley into units which are very
similar (Fig. 2a).
What can be discerned from the possible site territories?
Examination shows that because of their topographical location, the
defensive sites such as Walesland and Pelcombe Raths have heavily
dissected territories and a high proportion of poor quality land. Other
defensive locations were available along the valley sides that could
78 MYTUM

Walesland Rath Pelcomb Rath Crowhill R a t h

East Hook R a t h Slade Farm Cuckoo Grove Rath

Key

Site
10 minute isochrone
20 minute isochrone
Land capability class 3
(suitable for intensive grazing and arable)
Land capability classes 5 and 6
(suitable only tor rough grazing)

Figure 1. Possible territories of six Iron Age sites from the valley of
Pelcombe brook, a tributary of the Western Cleddau river, west Wales.

have allowed access to greater areas of brown earths, but by position-

ing the forts between closely spaced tributory streams this
opportunity was not seized. In contrast, the farmstead sites have
greater areas of brown earths immediately available, and this is
the case even with the hillslope sites which are just as close to
Pelcombe Brook as the defended ones. This might suggest a
greater reliance on arable farming at the later, smaller settlements
especially since they also supported smaller populations. When
twenty minute territories are examined, there is less difference
between the two settlement forms, suggesting that more extensive
grazing and exploitation from the earlier sites made their location
less undesirable. Such territories would overlap in places, but this
problem may have been less serious in a more pastoralist
economy.

The site territorial evidence cannot be considered conclusive in

 Iron Age and RomansBritish West Wales 79

b)

Figure 2. (a) Possible territories of Iron Age settlements in the valley of

Pelcombe Brook based on Theissen polygons. ? indicates a suitable
topographical! location for a site, but no traces have been found, (b)
Possible territories of Iron Age settlements. These are based on a 10
minute walking distance from the sites.

showing a change in emphasis in the subsistence strategy, and

neither can the on-site evidence. The relative numbers of querns
and carbonised grains at the later sites on the one hand can be
compared with bone frequencies and the relatively greater
productivity of the deposits on the other. More data may answer
some of these questions, but the fundamental problem of a
society's changing attitudes to rubbish and how it should be
disposed of can so dominate the assemblages from sites that
assessment of economic changes may be difficult. The differences
in sherd counts betwen Iron Age and Romano-British deposits
(for example at Coygan Camp, 92 and 2,378 coarse-ware
fragments respectively) is explained as a cultural change in both
80 MYTUM
the use and discard of pottery, and it is necessary for all evidence
concerned with subsistence strategy to be examined in the same
way. The real cultural and economic implications of such variations
are not always grasped, and are just taken as read. The environmental
assemblage in particular is just as culturally biased as is the artefactual
evidence, since the same cultural factors affect all disposal patterns
(Hodder 1982).
Taken superficially, the evidence from west Wales can be seen to
represent an intensification of all aspects of agriculture, which may
well be correct. The growth in the importance of cereals may have
begun even before the move from defensive to less defensive sites.
On closer and more critical examination the excavated evidence for
change is less clear, but can be supported by the off-site data. New
approaches are needed to break the deadlock in interpretation, and
several are being tried in west Wales. Careful sampling of similar
deposits on sites of various types and dates may provide a method
of examining changes in cereal production, even though those
deposits represent only a small proportion of those present on each
site. In addition, patterns of rubbish disposal are being defined to
compare with the patterning of environmental evidence. Experiments
are being conducted on the burning of animal bones, to simulate
some of the processes creating the archaeological record (Gilchrist
and Mytum 1986). Emmer wheat is already being grown experi-
mentally at Castell Henllys. Further use of the off-site data may also
be invaluable, and the larger number of sites in the Eastern Cleddau
study are sufficient for quantitative analysis and computer simulation.
The future is not bleak for the reconstruction of changes in past
subsistence strategies, but greater concern for changes in the nature
of contexts and the use of more sophisticated models of refuse discard
are essential for the selection of those trends which are related to
agricultural production, rather than those indicating rubbish disposal.
BIBLIOGRAPHY
Bailey, G. N. & I. Davidson (1983) Site Exploitation Territories
and Topography: Two Case Studies from Palaeolithic Spain,
J. Archaeol. Sei. 10, 87-115.
Crossley, D. W. (1963) List of Hill-forts and other Earthworks in
Pembrokeshire, Bull. Board Celtic Studs. 20, 171-203.
Crossley, D. W. (1964) Excavations at Maryborough Camp,
Wiston, a Pembrokeshire protected enclosure, Bull. Board
Celtic Studs. 21, 105-18.
Crossley, D. W. (1965) Excavations at Knock Rath, Clarbeston,
1962, Bull. Board Celtic Studs. 21, 264-75.
Gilchrist, R. & H. C. Mytum (1986) Experimental archaeology
and burnt animal bone from archaeological sites, Circaea 4.1,
29-38.
James, T. (1984) Aerial Reconnaissance in Dyfed, 1984, Archaeol.
in Wales 24, 12-24.
Hillman, G. (1981) Iron Age Crops, in G. Williams, 'Survey and
 Iron Age and RomansBritish West Wales 81
excavation on Pembrey Mountain', Carmarthenshire
Antiquary 17, 3-33.
Hodder, I. ed (1982) Symbolic and Structural Archaeology, Cam-
bridge University Press.
Hogg, A. H .A. (1972) Hill forts in the coastal area of Wales, in
C. Thomas (ed.) The Iron Age in the Irish Sea Province, pp.
11-24. Council Brit. Archaeol. Res. Rep. 9.
Manning, W. H. (1975) Economic influences on land use in the
military areas of the Highland Zone during the Roman
period, in J. G. Evans, S. Limbrey and H. Cleere (eds.) The
effect of man on the landscape: the Highland zone, pp. 112-16.
Council Brit. Archaeol. Res. Rep. 11.
Mytum, H. C. (1982a) Excavations at Castell Henllys, Newport,
Dyfed, Archaeol. Reps. Univs. Durham Newcastle upon Tyne
1981, pp. 14-16.
Mytum, H. C. (1982b) Excavations at Drim, Llawhaden, Dyfed,
Archaeol. Reps. Univs. Durham Newcastle upon Tyne, pp. 17-
20.
Mytum, H. C. (1982e) Survey of small enclosed settlements in the
Eastern Cleddau valley, Archaeol. Reps. Univs. Durham
Newcastle upon Tyne, pp. 21-2.
Mytum, H. C. (1984) Castell Henllys, Archaeol. in Wales 24, 45.
R. C. H. M. (Wales), 1917 Inventory of Ancient Monuments in
Carmarthenshire.
R. C. H. M. (Wales), 1925 Inventory of Ancient Monuments in
Pembrokshire.
Rudeforth, C. C. (1974) Soils in Dyfed II. Sheets SM 90/91 (Pem-
broke/ Haverfordwest). Soil Survey of England and Wales,
Soil Survey Record 24.
Rudeforth, C. C. & R. I. Bradley (1972) Soils, Land Classification
and Land Use of West and Central Pembrokeshire. Soil Survey
of England and Wales, Special Survey 6.
Savory, H. N. (1951) List of hill-forts and other earthworks in
Wales, V. Carmarthenshire, Bull. Board Celtic Studs. 16, 54-
69.
Schiffer, M. B. (1976) Behavioral Archaeology. Academic Press,
New York.
Vyner, B. (1982) Excavations at Woodbarn Rath, Wiston, 1969,
Archaeologia Cambrensis 131, 49-57.
Wainwright, G. J. (1967) Coygan Camp. A Prehistoric, Romano-
British and Dark Age Settlement in Carmarthenshire. Cam-
brian Archaeological Association.
Wainwright, G. J. (1971) The excavation of a fortified settlement
at Walesland Rath, Pembrokeshire, Britannia 2, 48-108.
Williams, G. (1978) Aspects of later prehistoric and native Roman
Carmarthenshire: Part I, Carmarthenshire Antiquary 14, 3-19.
Williams, G. (1979) Aspects of later prehistoric and native Roman
Carmarthenshire: Part II, Carmarthenshire Antiquary 15, 15-
37.
Williams, G. (1981) Survey and excavation on Pembrey Mountain,
Carmarthenshire Antiquary 17, 3-33.
Williams, G. (1984) Llawhaden, Archaeol. in Wales 24, 46-8.
Williams, G. (1985) Fighting and Farming in Iron age West Wales.
Dyfed Archaeological Trust.
Some Quantitative Approaches to Site Territory and Land
Use from the Surface Record

C. F. Gaffney and V. L. Gaffney

The implementation of rigorous collection strategies has allowed

greater quantification of most aspects of archaeological research. In
field survey this trend has had major consequences. Whereas survey
had fulfilled a site locational role, often simply with an eye towards
locating the 'best' site for excavation, the introduction of quantitative
and diachronic strategies has allowed survey to take on a positive
analytical role (Mercer 1982).
This change was inevitable and was due to the increasingly pro-
cessual nature of the questions being asked by archaeologists; ques-
tions which traditional distribution maps as produced through survey
were unable to answer. Inevitably innovative questions and tech-
niques led to novel problems. Of note is the continuing debate within
field survey regarding the scale of collection areas and, hence, the
nature of collection strategies. In short, whether to sample, how to
sample, or not to sample. After the euphoric 'discovery' of sampling
and the almost universal adoption of regional survey by field
archaeologists, it was realised that this panacea to our financial
limitations itself carried limitations (Gaffney and Tingle 1985).
Primarily, since survey scale carried behavioural implications, the
use of sampling at a regional scale is most likely to be pertinent to
behavioural processes working on a regional basis. Thus, specific
questions aimed at the resources around, or activities carried out
from single sites may be inadequately dealt with by regional
approaches. This problem has been emphasized by the growing
awareness of the importance of off-site activities (Foley 1981). Sur-
veys which function well for site location may not possess the resol-
ution to analyse relevant off-site activities. Consequently an aware-
ness that some research can only be continued at levels above the
site and below the region is beginning to pervade field survey ( C.
F. Gaffney et al. 1985). This paper is devoted to several aspects of
such archaeological analysis, specifically the problems of interpreting
behavioural land use and of boundary estimation.
In the seventies behavioural land use models were mainly derived
from geographical theory, notably site catchment analysis at the
microregional level (Higgs and Vita Finzi 1972). Whilst site catchment
 Quantitative Approaches to Site Territory 83
analysis has been justifiably criticised (C. F. Gaffney et al. 1985), it
is apparent that such a model inspired some field workers to look
beyond the problem of site definition and toward land use (e.g.
Wilkinson 1982). Whilst traditional site catchment surveys rarely
allowed the testing of the postulated land use models, the growth of
off-site survey techniques is providing an analytical format that may
test site catchment results.
A significant application of such techniques can be demonstrated
in recent attempts to define the relationship of surface artefact scat-
ters, associated with manuring, to the sites that are assumed to be
the point of origin of such material. Traditionally, research on such
agricultural practices has been fraught with problems. Even before
attempting areal survey archaeologists have found difficulty in
defining the nature of farms as an integral set of buildings. The
continuing debate concerning villa status is a case in point (Smith
1978). Once in the field the problems increase. Archaeologists,
because they have no theoretical option, assume that sites, i.e. settle-
ments, somehow rise above any background noise and will be
instantly recognisable as such. If this is found to be the case, few
surveys proceed beyond this stage. The pattern of land use or tenure
beyond the settlement is rarely attempted or is considered an impon-
derable, despite its importance to economic reconstructions. The term
'background noise' used in survey is interesting in this respect. It
appears to imply insignificance, and is usually treated as such. The
reality of the matter is that if background noise can be defined, then
its presence should not be ignored, it should be explained. Some
aspects of these problems have been investigated using information
from the Maddle Farm Project in Berkshire, England. Here data
relating to a small Romano-British villa complex were collected on
a systematic basis in a multi-stage survey design. In the first season
all material within a 2 km radius on the villa was collected and was
followed by a larger sampling design using a series of transects
(Gaffney and Tingle 1985).
In the results from the first season's work within the 2 km circle,
we can see that Roman pottery discard is apparent beyond the
confines of the three previously known sites (Fig. 1). These sites
appear as peaks, although outside these areas the distributions show
structure. From these results five descriptive classes can be suggested
for the survey patterning.
1. 'Villa': Defined by field walking, aerial photography and exca-
vation.
2. 'Settlement': as for number 1.
3. 'Activity Areas': Small discrete scatters, often with a relatively
high density of artefacts and possibly relating to non-settlement,
off-site activity foci.
84 GAFFNEY AND GAFFNEY

CONTOUR VALUE
1 - 4 00
2 - 14 00
3 - 24 00
4 - 34 00
5 - 44 00

O
MADDLE FARM PROJECT (NUMBERS) CORE AREA
1Km
A
Figure I. The number of Romano-British Sherds located by surface
survey.

4. 'Manuring Spreads': More diffuse, generally low density of

artefacts.
5. 'Low Density Areas': Whilst appearing to contain few or no
artefacts, these areas are definable and therefore have meaning
in our landscape. They might relate in some instances to areas
 Quantitative Approaches to Site Territory 85
linked to a function which does not require pottery discard or
manuring, e.g. pasture.
The structure of these groups may be checked by reference to the
distribution of sherd weight per ha (Fig. 2). This shows a marked
similarity to the numbers plot. Whilst emphasizing similarity this
check does not prove any distinction between our subjectively
ascribed groups of settlement and manuring or activity spreads.
However, an attempt at this can be made. With some provisos
concerning the variable nature of pottery fabrics (Hinchcliffe 1979)
and the extent of modern plough damage, we can model potential
differences between on- and off-site ceramic distributions and con-
sequently infer the nature of differential discard and recovery.
Agriculturally, when a field is manured with farmyard materials,
fragments of broken pottery enter the plough soil. This material is
repeatedly ploughed and abraded. On site, whilst post-settlement
deposits may also be ploughed, the material is not so finite and may
potentially suffer less than that within an off-site position, which will
receive no new deposits after manuring ceases. As ploughing damages
sealed cultural deposits the sites may act as reservoirs releasing fresh
artefacts with ploughing. Consequently off-site material should be
smaller than on-site material simply because, on average, it has been
in the system longer.
Probably the best way of testing such a hypothesis is to measure
and fabric each sherd and use 'F' or 't' tests on the results. Unfortu-
nately the number of sherds involved in this survey rendered this
impracticable. For the Maddle Farm data a more direct approach
was preferred, using the ratio:
(Wt. pot per ha)
(No. pot per ha)
The results provided by this approach and shown in Fig. 3 highligh-
ted three significant points:
1. Areas of low weights and numbers remain low.
2. There is a diffuse background which submerged not only those
areas which were earlier classified as manuring spreads, but
also those associated with activity areas.
3. On top of this diffuse background there are several foci of high
values. If the model is correct then artefacts from these areas
should be associated with settlement sites. Irregularities which
do occur can be shown to be the product of large individual
sherds.
These results argue for a slight modification of the model to suggest
that on-site interpretation is most likely in those areas above a specific
threshold in the artefact number distribution. This allows us to begin
 GAFFNEY AND GAFFNEY

CONTOUR LEVELS
1 - 50 00
2 - 60 00
3 - 70 00
4 - 80 00
5 - 90 00
MADDLE FARM PROJECT (WEIGHTS) CORE AREA
0 1Km
I I

Figure 2. The weight of Romano-British Sherds located by surface survey.

to define the potential agricultural use of low discard areas beyond

locatable sites without precluding the existence of more elusive
settlement forms.
In order to integrate this information into an economic/be-
havioural interpretation some attempt must be made at territorial
determination. Archaeologically, the basic problem in any form of
territory determination is that such boundaries as existed were never
static and were rarely marked in an archaeologically detectable
 Quantitative Approaches to Site Territory 87

CONTOUR LEVELS

MADDLE FARM PROJECT (WEIGHT/NUMBER) CORE AREA

1Km
A
Figure 3. The ratio of (weight/number) Romano-British Sherds located by
surface survey.

manner (Percival 1976, ch 6). Precise mapping consequently is

difficult. In the past, the major techniques of boundary definition
used by archaeologists at the microregional level have included:
1. Definition by natural boundaries (eg Branigan 1977) rivers
and hills have often been documented as forming boundary
edges. Unfortunately this technique has rarely risen above the
level of supposition.
2. Site Catchment Analysis has often been used, including as it
does a behavioural element through the provision of a time/dist-
ance factor (Higgs 1975). This can accommodate the role of
topography and other natural boundaries in producing an esti-
mation of a maximum behavioural territory associated with a
site.
88 GAFFNEY AND GAFFNEY
3. Thiessen Polygon analysis. The use of Thiessen polygons to
define territories differs from the first two techniques. Here the
polygons used to infer boundaries rely on the complete knowl-
edge of potentially competing communities between whom we
divide the available land. This division can be achieved using
a variety of methods and often has a statistical base (Hodder
and Orton 1976).
All of the above techniques have been used, abused and in our
opinion have generally been found wanting. Yet they all have some-
thing to offer provided they are not used uncritically. For example,
if in survey we can equate off-site discard with human activity it is
likely that we might expect a fall in discard coincident with the
decrease of activity with distance from site and proximity to borders,
much as predicted by Site Catchment Analysis (Wilkinson 1982).
An example of such a pattern may be detected within the Maddle
Farm data when we roughly equate pottery density with agricultural
intensity. Working from this concept we can produce an estimate of
the minimum probable catchment by plotting lines of least resistance
through the surface data. One must be cautious in the use of this
method, as there may be perfectly good reasons for the existence of
low density areas within a territory. The placement of pasture for
animals adjacent to settlements and indicated by low artefact densities
could be problematic if insufficient data has been collected to suggest
relative differences within discard zones. Likewise the relationship
of primary and secondary activity or settlement foci is also relevant
when we try to understand how land is used or partitioned. The
creation of individual catchments for every single site located could
be grossly misleading if the location of secondary sites is dependent
upon the primary site rather than independent of it (C. F. Gaffney
et al. 1985).
The catchments denned through artefact distribution and those
provided by Thiessen Polygon Analysis are compared in Fig. 4. The
similarities of the two are striking, the northern boundary being
significantly similar in both cases. The eastern and western boundaries
do not correlate as well. This is possibly due to the absence of site
data needed by Thiessen Polygon Analysis. Alternatively, the lack
of discarded pottery might indicate a different land use in this area,
possibly the result of distance from site. It is interesting that these
boundaries do parallel the existing natural barriers of the chalk
backslope valleys. Indeed the more extreme gradients to the east
appear to limit the spread of manuring scatters more strictly than
the gentler slopes to the west.

These results firmly indicate the validity of further research on the

nature of land use within the postulated boundaries and based upon
 Quantitative Approaches to Site Territory 89
Based on Sherd Analysis

t
0 IKm

Figure 4. A comparison of theoretically derived catchments. Thiessen

polygon analysis based on known villa position (thin line) and known
settlements (thick line). The dotted line represents boundaries suggested by
the surface evidence. In the south-west there are two possibilities (A) and
(B).

the economic practices indicated by manuring spreads. The results

from the Maddle Farm data apparently confirm the gross behavioural
parameters outlined by Site Catchment Analysis as shown by the
sampled artefact distribution around sites on the Downs, but the
detailed results also indicate the diversity of land use which is rarely
susceptible to traditional Site Catchment Analysis nor indeed large-
scale sampling survey.
Land use and boundary estimation are contentious subjects and
undoubtedly the complexity of the problems demands a pragmatic
approach rather than the setpiece explanations of earlier models (De
Atley and Findlow 1984). The increasing sophistication of field survey
techniques and the nature of survey data provides an interesting
avenue of relevant research. Field survey being independent of the
conceptually restrictive confines of the site can, if used carefully,
provide invaluable complementary evidence to the environmental
work carried out on site. From saying what crops were grown we
may eventually consider where they were grown, and begin to under-
stand the complexity of human land use.
BIBLIOGRAPHY
Branigan, K. (1977) Gatcombe Roman Villa. British
Archaeological Reports 44, Oxford.
De Atley, S. P. & F. J. Findlow (Eds.) (1984) Exploring the limits:
Frontiers and Boundaries in Prehistory. British Archaeological
Reports S224, Oxford.
Foley, R. (1981) Off Site Archaeology and Human Adaptation in
Eastern Kenya. Monographs in African Archaeology 3, British
Archaeological Reports LS97, Oxford.
Gaffney, C. F., V. L. Gaffney & M. Tingle (1985) Settlement,
Economy or Behaviour? Micro-Regional land use models and
90 GAFFNEY AND GAFFNEY
the interpretation of Surface and Artefact Patterns, pp. 95-107
in C. Haselgrove, M. Millett and I. Smith (Eds.), Archaeology
from the Ploughsoil. Sheffield University, Dept. of Prehistory.
Gaffney, V. & M. Tingle (1985) The Maddle Farm (Berks.) Project
and Micro-Regional Analysis, pp 67-73 in S. Macready and
F. H. Thompson (Eds.). Archaeological Field Survey in Britain
and Abroad. Society of Antiquaries, London.
Higgs, E. S. & C. Vita-Finzi (1972) Prehistoric economies: A ter-
ritorial approach, pp. 27-36 in E. S. Higgs (Ed.). Papers in
Economic Prehistory. Cambridge University Press.
Higgs, E. S. (Ed) (1975) Palaeoeconomy. Cambridge University
Press.
Hinchcliffe, J. (1979) Excavations within the city of Verulamium.
Hertfordshire Archaeology 7, 10-27.
Hodder, I. & C. Orton (1976) Spatial Archaeology. Cambridge
University Press.
Mercer, R. (1982) Field Survey, a Route to Research Strategies,
Scottish Archaeological Review 1, 91-97.
Percival, J. (1976) The Roman Villa. Batsford, London.
Smith, J. T. (1978) Villas as a Key to Social Structure, pp. 145-189
in M.Todd (ed) Studies in the Romano-British Villa. Leicester
University Press.
Wilknson, T. J. (1982) The definition of ancient manured zones by
means of extensive sherd sampling techniques, Journal of
Field Archaeology 9, 323-333.
 Section 2
From Micro- to Macro- Region

J
 Introduction

John Bintliff
School of Archaeological Sciences, Bradford University

The six papers that comprise this section are broadly concerned with
human activity seen at the level of the 'region' in its traditional sense,
a unit of country larger than that associated with one particular
settlement and smaller than that commonly found to be occupied by
modern nation states.
Although in some cases the significance of archaeological findspots
across such a landscape may be illuminated by contemporary texts
(Ergenzinger, Bintliff) current approaches seek to use autonomous
methods for activity analysis. Approaches widely in use include at
the initial level Catchment Analysis, followed and/or accompanied
by Thiessen Polygons (Slapsak, Bintliff). The combination of these
methods proves invaluable for asymmetric territories and for small
territories well below catchment thresholds. Even where these specific
techniques are not employed it is now normal for regional archaeo-
logical investigations to evaluate resources for human exploitation
at different levels of access to the settlements under consideration
(Kalis, Ergenzinger, Bayliss-Smith).
Many of the well-known problems with this kind of land use
evaluation crop up in these papers and give rise to varied solutions.
Comparison with modern settlement and landuse patterns, or textual
support, may be invoked; or arguments adduced from the internal
consistency of the results of analysis and complementary archaeologi-
cal data reinforcing the patterns proposed (Slapsak, Ergenzinger,
Bintliff, Kalis, Bayliss-Smith). A general optimism is expressed that
such techniques can provide invaluable provisional interpretations
of regional settlement systems, for which usually no alternative
approaches are available.
A specific problem concerns the 'missing data', parts of the original
settlement and activity system that are permanently or temporarily
unavailable for analysis due to natural or human burial or destruction
(Bintliff). Statements made from the surviving fraction will
necessarily differ in their degree of generalisation to the region as a
whole, whilst attempts at population history and inter-settlement
analysis require specific problem-orientated fieldwork to parameter-
ize if possible the 'lost portion', unless historical sources can come
to the rescue.
94 BINTLIFF
Any evaluation of landscape potential has to proceed beyond the
mere calculation of maximum population supportable in a given
landscape within a specific technology and mode(s) of exploitation
of resources (Bayliss-Smith). The range of realisation of potential
exploitation is highly variable in the historic and ethnographic record.
Although there is reason to believe that a recurrent feature of the
demography of past societies has been the agrarian cycle of popula-
tion peaks and slumps (Bintliff, Ergenzinger), the causes of any
particular level of actual landscape utilisation compared to a
hypothetical maximum must be sought in the crucial analytical zone
of culture/nature symbiosis. Bayliss-Smith offers a helpful guide in
his concept of a 'welfare filter', where it is argued that a specific
society has tried to put a preferred lifestyle into effect, with obvious
implications for the degree of exploitation of the landscape required
for its achievement. Such a cultural filter nonetheless does not invali-
date the potential of 'Neo-Malthusian' demography of the medium-
term, but in so far as responses to such pressures can vary, contem-
porary perceptions may hold the key to the historical outcome for
any particular society.
This complex relationship between human and landscape poten-
tial, and in particular the exact ways in which environmental
archaeologists are trying to 'deconstruct' the preserved remains of
that matrix of past interaction, would seem to mark a new and more
sophisticated era in regional archaeology. However, as Lewthwaite's
paper reveals, archaeologists seem in many respects to have reached
the stage French historical geography had achieved in the early part
of this century. The relative isolation of the discipline of archaeology
from important intellectual trends in related disciplines such as
History and Geography began towards the end of the last century
and has been a characteristic feature of this present century, much
to archaeology's disadvantage (on which see now my introductory
chapter in Archaeology at the Interface (Bintliff and Gaffney Eds.,
BAR Oxford, 1986)).
The spectrum of French scholarship offers an informative parallel
to our current situation (and with a similar application to problems
of regional development), from the environmental determinism of
Braudel to the constraining 'possibilism' and socio-cultural filters
espoused by Vidal de la Blache and Febvre. Lewthwaite concludes
his paper with promising insights into a new kind of regional environ-
mental archaeology once these necessary influences from geography
have been assimilated. Significantly Slapsak acknowledges a direct
debt to the culture-nature symbiosis debate in historical geography
in the creation of his 'Formation of a Cultural Region' project in
Slovenia.
Defining the Economic Space of a Typical Iron Age Hillfort:
Rodik (Yugoslavia), A Case-study

Bozidar Slapak, University of Ljubljana

The scope of this paper is to present the procedures applied in a

preliminary analysis of settlement in the Kras (Carst) region, north-
west Yugoslavia. Its aim is to formulate a hypothesis about settlement
systems and land use, and to serve as a pilot to research centred
upon the excavation of the multi-period Rodik-Ajdovscina hillfort
(Fig. 1) (Slapsak 1983a and 1986).
This research is part of the 'Geneza kulturne pokrajine' (Forming
of a Cultural Region) research programme, organized by the Depart-
ment of Archaeology, University of Ljubljana 1 . The name of the
programme, adopted in 1981, marks a certain shift in interest and

Figure 1. Location of the Rodik hillfort in Slovenia, Yugoslavia.

96 S LA PS A K
an attempt towards a systematic study of the spatial and environ-
mental context of archaeological finds. It refers to the local tradition
of anthropogeography but tends to incorporate trends in spatial
analysis initiated by the New Geography of the sixties as well as
palaeoecology and to formulate spatial phenomena in terms of sys-
tems. The overall research objective is denned so as to promote, by
archaeological means, the study of processes that formed specific
culture-traits, including landscape, settlement, economy and way of
life for historical-geographical regions. So the notion of 'region' as
a basic distinctive unit of archaeological observation is introduced,
as opposed to 'archaeological culture'. A region is viewed as a
culture-and-environment determined spatial system with a limited
choice of adaptive solutions. Historically, a part of these potential
solutions forms a genetically linked sequence of which the contem-
porary or recent system is the directly observable variance and
consequence (therefore potentially also carrying information about
the previous ones). As an object of study, the region will be chosen
deliberately, in the sense that the delimitation of the region will be
based on any kind of boundary, be it modern, historically attested,
natural, theoretically deduced or statistically generalised, as long as
the choice is guided by our explicitly stated assumptions as to its
relevance to the problem studied. One other prerequisite would be
that the methodology used be such as to enable correction of boun-
daries as a result of the study (Slapsak 1983a: 158; 1986 in press).
These general ideas are being put in operation through several
research projects of our Department, and their variety in scale, scope
and method illustrates well the range of possible units of observation
contained in the notion of a 'region'.2
The Kras project is concerned with an area of considerable his-
torical importance on the border of four distinctive geopolitical units
the Italian peninsula, the Balkans, the Danube area and the Alps,
lying on important historically attested transcontinental communica-
tion routes (Negroni-Catacchio 1976, Sasel 1974, Slapsak 1978). It
started in the seventies with a general survey of the area in the
framework of the Archaeological Survey of Slovenia (Arheoloska
topografija Slovenije), organised by the Archaeological Institute of
the Slovenian Academy of Sciences. It is a bibliographical survey,
aiming at verification of data gathered from all archaeological publi-
cations and archives published in the Archaeological Sites in Slovenia
volume (ANSI 1974), and extended through systematic collection
and verification of further data from the local inhabitants, schools,
parish archives etc., following the instructions published by Pahic
(1962) on the model of the German 'Archaeologische Landesauf-
nahme' (Tode 1924; cf. Dular 1985).
Data so gathered and verified in the field (Slapsak 1983b) permitted
a preliminary study of settlement in the region with simple correla-
 Typical Iron Age H HI fort: Rod i k (Yugoslavia) 97
tions involving environmental and locational data. Here, I will
confine myself to the problem of hillforts gradisca/ castellieri.
Hillforts are an outstanding feature of the Bronze and Iron Ages
of the region ('kastelirska' or 'castellieri' culture, e.g. Gabrovec 1983:
46; Radmilli 1974: 483). Even if we disregard earlier interest in them
(e.g. Burton 1877; for history of research cf. Rinaldi 1964), they have
been known to the specialists at least since the beginning of the
century through the publication of the Marchesetti survey (Mar-
chesetti 1903). Unfortunately, there are serious limitations to the
study of the hillfort settlement of the area. It is a karstic region of
extreme erosion activity, formation of archaeological deposits is
irregular, stratification is rare, palaeoenvironmental evidence scarce,
and technical problems posed by the building techniques and destruc-
tion patterns of massive stone fortifications are often discouraging.
Actually, there have been no modern large scale systematic excava-
tions there (for Italian research cf. Preistoria 1983; Problemi 1983;
for excavation in Istria, Bacic, 1958; cf. also Preistorija Jugoslazvije 4).
Much of the study of hillforts has therefore concentrated on the
most obvious feature, on the morphology of the fortification layouts.
Typologies resulting from these studies are usually quite unsatisfac-
tory, offering as they do a free choice of functional interpretations,
especially with complex layouts, including functional diversity of the
settlement space, addition to settlement space caused by population
pressure and development of fortification (cf. the discussion in Lonza
1977); typologies of the simple layouts produce even more irrelevant
results (cf. Suie 1975).
In order to avoid the pitfalls of simple univariate typologies we
tried to move towards widening of distinctive criteria by correlating
locational data: relative height, the space controlled and the settle-
ment surface. While these criteria certainly do not help with large
and complex hillforts which tend to form a single undifferentiated
group, with smaller ones the diagram (Fig. 2) shows a degree of
regularity in clustering which permits us to differentiate single groups
and even to venture functional interpretations: very small enclosures
with excellent overview (5), possibly over neighbouring regions, were
tentatively interpreted as 'strategic outlets'; very small enclosures on
hilltops but not necessarily controlling more than the immediate
neighbourhood and not necessarily optically communicating with
each other (3) were interpreted as 'family group settlement units';
those of variable, and often considerably larger surface but where
location as to defence and overview gave poor results (2) were
understood as 'specialized (e.g. pastoral) enclosures'; finally, those
larger than 2 ha, located so as to be able to control visually the whole
land possibly belonging to them (i.e. visually controlling the whole
pattern of land use typical of the hillfort community) and visually
to communicate with each other (4) were interpreted as 'basic
98 SLAPAK

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
settlement surface ha

Figure 2. Small hillforts grouped according to locational criteria.

hillforts'.
This picture differs markedly from that presented by V.
Karouskova-Soper who, while discussing possible functional
interpretations for the hillforts on the Italian side of the Kras
(Carso Triestino), came to the conclusion that they should all be
regarded as belonging to the same category of fortified settlements
(arguing only for a correlation between settlement surface and
territory controlled), and that geomorphology was the determining
locational factor (Karouskova-Soper 1983). We must disagree with
this rather egalitarian view of the hillfort settlement on the Karst.
The functional interpretations we propose, arbitrary as they may
be, drawing from our implicit assumptions about the possible
organisation of prehistoric communities, are founded on the
observation of empirical diversity. They will necessarily be subject
to eventual verification and change through further research (cf.
the case of Rodik, infra), but they seem to be productive
inasmuch as they honour the complexity of prehistoric settlement
and permit us to grasp a hierarchical relationship between
hillforts.
In fact, this procedure permitted us to consider the regional
settlement system by building networks of 'basic' settlements and
introducing new criteria for differentiation (magnitude, complexity
of fortification layout) among them. We experimented with two
models. The central place model produced (Slapsak 1983a, Fig. 71)
 Typical Iron Age Hillfort: Rodik (Yugoslavia) 99
as Green and Haselgrove would put it much fun in drawing
circles around places and also a quite attractive final picture. The
thing is that this picture makes sense in quite an unexpected way:
besides natural and regional boundaries, the borders of the system
coincide with certain historicaf frontiers (to the West: Sempolaj
early Roman and later, Vipava medieval and later). This does not
necessarily prove the validity of the model, but the regularities
observed are such as to demand further investigation. Just to destroy
the idyllic picture of the central place model, we also tried the gateway
community model as formulated by Hirth (1978) on some
Mesoamerican examples (Slapsak 1983a, Fig. 72). It produced a
rather complicated model with two gateways the harbour and the
pass entrance control gateway (Skocjan). The border of the system
coincides with the meteorological border: to the North and East the
snow cover makes all-year grazing impossible. Again, the result as
we understand it, was not so much to prove the validity of the system
but rather to explicate the points of possible historical economic
dynamism which must be taken into acount while studying the
prehistoric settlement of the area.
A further step in the development of the project was the choice
of scale and place for detailed systematic research. It was decided
to concentrate on a 'basic' hillfort settlement and its catchment area,
a territory on which the whole pattern of the hillfort community's
basic economic activity was supposedly taking place. Rodik-Ajdov-
soina (Figure 3) was chosen for several reasons. It is a typical 'basic'
hillfort in the core area of the supposed regional system; it has a
continuous occupation till the 5th century A.D. at least which makes
it possible for us to monitor change in a local indigenous community
under Roman rule; there are a number of subordinate settlements in
the territory, as well as traces of ancient fields in the immediate
surroundings of the hillfort; finally, diversity of environmental condi-
tions in the territory make it possible to use these data in the pre-
liminary analysis of the settlement (Slapsak 1983c).
Standard procedures for the determination of site catchment were
applied and the results compared. We tried Thiessen polygons (Fig.
4) on 'basic' hillforts, assuming that after Marchessetti's and our own
survey, the network of these hillforts (average spacing 6 km) can be
considered complete, and that the regularily spaced hillforts were
contemporary in one phase at least.3
We delimited the site catchment according to the principles
developed by Higgs and Vita-Finzi for isolated settlements. Based
on the assumption of combined economy (pastoralist/agricultural-
ist), we decided to take the 5 km circle range as the unit of observation,
additional consideration being given to the 1 hour walking distance
range4 and the 1 km distance which is the area of intense
100 SLAPSAK

I I Carst
EU Flysch
--'! Alluvium (blind valley)

Figure 3. Simplified geomorphology and geology of the Rodik-Ajdovscina

hillfort area.

exploitation from one settlement in the region today (Fig. 5).

The Thiessen polygon and the 1 hour walking distance were then
compared. While the frontiers roughly coincided on the north and
east, significant discordances were observed on the west and south-
west. On the west, the 1 hour catchment somewhat exceeded the
polygon: considering the geomorphology of the area, and the possi-
bility of visual and effective control of land from the two neighbouring
hillforts, we decided to give credit to the 1 hour walking distance
range and to make a slight correction of the polygon to the west
(Fig. 5), the more so as the new line coincided better with historical
delimitations and cadastral deliminations between villages. On the
south-west however, the polygon considerably exceeded the 1 hour
catchment, reaching (and exceeding) the 5 km range. It was observed
that the most distant of the subordinate settlements in this part of
the delimited area was located on the very border of the 1 hour
walking distance (no. 7 on Fig. 6). It was assumed therefore that this
was not coincidental and that the explanation of the discordance
 Typical Iron Age H i II fort: Rod i k (Yugoslavia) 101

SL1VJE

Figure 4. Rodik-Ajdovscina territory as defined by Thiessen polygons on

'basic' hillforts.

1 2 3 54 15

Figure 5. Rodik-Ajdovscina one hour walking distance catchment and the

proposed adjustment of the polygon to the West.
102 SLAPSAK

50 60
tf>
T

Figure 6. The settlements of the Rodik-Ajdovscina hillfort territory.

was to be sought in the logic (possibly dynamism) of the Rodik-

Ajdovscina basic hillfort settlement system.
The settlements of the Rodik-Ajdovscina hillfort territory were
then studied: single settlements were classified and correlated with
available environmental data (Fig. 7). The notion of 'ecological
units' was introduced, meaning spatial units of specific environmental

50 60

Figure 7. The settlements classified and correlated with 'ecological units'.

 Typical Iron Age H i II fort: Rodik (Yugoslavia) 103
conditions, including landuse potential. Differences in bedrock type,
causing different soil development and karstic micromorphology,
proved to be crucial in this respect. Also, this procedure enabled us
to distinguish modern settlement zones with characteristic choice of
and participation in ecological units, and subsequent typical adapta-
tions of landuse patterns (Fig. 8). Differences between them were
identified and relative advantages of the zone on the flysch/karst
juncture line (no. 2 on figure 8) were observed, diversity of resources,
rich soil, abundance of water. In relation to these data, archaeological
evidence was analysed.

50 60

Figure 8. Modern settlement ecological zones.

As a result, a choice of models of settlement could be proposed.

As no data on detailed chronological relationships between
settlements are available at this stage, initially all combinations
were assumed to be equally plausible, to be subsequently tested
against carrying capacity figures for single ecological units in
relation to the settlement surface. In this way, a number of
proposed models were ruled out and the rest accepted as the
choice of probable adaptive solutions where single models could
represent stages in a genetically linked sequence. To take an
example, the model which would explain the Rodik-Ajdovscina
hillfort as controlling the Brkini flysch only (ecological unit 7,
corresponding to the modern settlement ecological zone 3a on
figure 8) was ruled out because the carrying capacity of the unit
was estimated at just over 30 which at 3.8 ha settlement surface
104 SLAPSAK
50 60

Figure 9. Settlement model 4, hillfort 7 (Debels griza) as subordinate

specialized settlement (pastoral enclosure), efficient exploitation of the
lowland pastureland.

gives the unacceptable figure of 1,200 m2 per person. For its sub-
sistence, the Rodik-Ajdovscina hillfort obviously had to control part
of the lowland area as well. On the other hand, the lowland B type
hillfort of Debela griza (no. 7, on the border of the 1 hour walking
distance range) could plausibly be interpreted as a subordinate
specialised settlement located so as to enable more efficient control
and exploitation of the lowland pasture land (Fig. 9) 4,000 sheep/goat
units being the carrying capacity of the lowland area, we arrive at
4-5 m2 per sheep/goat in the 1-8 ha presumed pasture enclosure.
Hypotheses about the mechanisms which produced change could
now be formulated and expectations as to archaeological evidence
of them presented, to be verified in a programme of field work
including palaeoenvironmental research. We hope that we can find
the means to carry out the Rodik project and that the archaeological
evidence will prove to be of sufficient quality to permit verification.
Finally, one exciting detail of archaeological evidence may be
mentioned which indicates that what we are doing is perhaps not
completely devoid of sense. An important historical inscription was
found in the village of Materia about 100 years ago (CIL 5,698; cf.
Slapsak 1977). There is a high probability that it was found in situ
on its ancient location and this location coincides well enough with
the south-east border of the Rodik Ajdovscina hillfort territory as
defined by the Thiessen polygon (Fig. 10). The inscription is about
the border controversy between the Roman senator Caius Laecanius
 Typical Iron Age H ill fort: Rodik (Yugoslavia] 105
50 60

Figure 10. The site of the CIL 5,698 inscription.

50 60

C\J

Figure 1 1 . The territory of Rundictes in the 1st c.A.D.

106 SLAPAK
Bassus and the Rundictes, a non-latin name for a local community,
preserved in the name of the modern village of Rodik. It is highly
probable that our project is about the territory of the Rundictes of
the Materija inscription (Fig. 11).

NOTES
1. Cf. Annual research reports (including English abstracts), published by the
Scientific Institute of the Faculty of Philosophy, University of Ljubljana.
2. e.g. for the Early Middle Ages settlement in the Alpine area, Knific 1983, 1985;
for the territory of a Greek colony, Kirigin, cf. Slapsak 1986; for a neolithic settlement
in south-east Slovenia, cf. Budja, Moverna vas (unpubl.); for Ljubljansko barje
(Ljubljana Moor) Bregant et al., project in progress (unpubl.), cf. results of the
environmental research in Porocila, passim.
3. Excavation on Slivje (Stacul 1972) made it possible to distinguish Bronze Age
and Iron Age strata on castellieri and further study of pottery (cf. Cardarelli 1983)
made it clear that not all the hillforts were contemporary and that occupation on
several of them started and ended at different points of time between the Early
Bronze Age and the Iron Age. In our case, we could only use the presence of coarse
pottery with abundant calcareous inclusions as a gross criterion for the contem-
poraneity of the hillforts in the Iron Age.
4. Given the complicated geomorphology of the area and the scarcity of cultivat-
able land, we counted the walking distance for each islet of good land separately
and generalized the result, hence the rather geometrical final picture.

BIBLIOGRAPHY
Ansi (1974) Arheoloska nagdisca Slovenije. Ljubljana.
Bacic, B. (1970) Prilozi poznavanju prehistorijske gradinske
fortifikacije u Istri, in Adriatica praehistorica et antiqua,
Zbornik Grge Novaka, pp. 215-234. Zagreb.
Burton, R. F. (1877) Note sopra i castellieri o ravine preistoriche
della peninsola Istriana. Capodistria.
Cardarelli, A. (1983) Castellieri nel Carso e nell'Istria: Cronologia
degli insediamenti fra media eta del bronzo e prima eta del
ferro, in Preistoria, 1983, 87-134.
Dular, J. (1985) 'Bela krajina'. Arheoloska topografija Slovenije 1.
Ljubljana.
Gabrovec, S. (1983) Jugoistocnoalpska regija, in Praistorija
jugoslovenskih zemalja 3: Bronzano doba, pp. 23-96. Sarajevo.
Hirth, K. G. (1978) Interregional trade and the formation of
prehistoric gateway communities, American Antiquity 43, 35-
45.
Karouskova-Soper, V. (1983) The castellieri of Venezia-Giulia,
North-east Italy. BAR Int. Ser. 192. Oxford.
Knific, T. (1983) Blejski kot v zgodnjem srednjem veku. PhD thesis,
University of Ljubljana.
Lonza, B. (1977) Appunti sui castellieri dell'Instria e della provincia
di Trieste. Trieste.
Marchesetti, C. (1903) / castellieri preistorici di Trieste e della
regione Giulia. Trieste.
Negroni-Catacchio, N. (1976) Le vie dell'ambra. I passi Alpini
Orientali e l'Alto Adriatico, in 'Aquileia e l'area Alpina
Orientale'. Antiquita Altoadriatiche 9, 21-57.
 Typical Iron Age Hillfort: Rodik (Yugoslavia] 107
Pahic, S. (1962) Arheoloska topografija Slovenije. Navodila, Argo
1, 93-120.
Porocila Porocila o raziskovanju paleolita, neolita in eneolita.
Predkovinske kulture Sloveniji. 1, 1969.
Preistoria (1983) Preistoria del Caput Adriae. Exhibition catalogue.
Trieste.
Problemi (1983) Problemi storici ed archaeologici dell'Italia Nor-
dorientale e dlie regioni limitrofe dalla preistoria al medioevo.
Atti dei Civici Musei di Storia ed Arte di Trieste, Quaderno
13. Trieste.
Radmilli, A. (1974) Popoli e Civilta del Italia Anticha, I.
Rinaldi, M. L. (1964) Brevi note a proposito dei castellieri.
Archaeografo Triestino 25, 3-25.
Sasel, J. (1974) Okra, Kronika 22, 9-17.
Slapsak, B. (1977) Via derecta-translata-restituta. Ad CIL 5, 698,
Archeoloski vestnik 28, 122-128.
Slapsak, B. (1978) I valichi del Carso fra preistoria ed eta
romana, in Putevi i komunikacije u praistoriji. Matriau 6, pp.
69-76. Pristina.
Slapsak, B. (1983a) Moznosti studija poselitve v arneologiji. PhD
thesis, University of Ljubljana.
Slapsak, B. (1983b) The Kras (Carst) survey, in Keller, D.R.
Rupp, D. W., Archaeologica Survey in the Mediterranean area.
BAR Int. Ser. 155, 201-202.
Slapsak, B. (1983c) Rodik-Ajdovscina, in Problemi 1983, 225-227.
Slapsak, B. (1985) Rodik. Exhibition catalogue. Trieste-Trst 1985.
Slapsak, B. (1986) Settlement studies. BAR Int. Ser.
Stacul, G. (1972) 'II castelliere C. Marchesetti presso Slivia nel
Carso Triestino', Rivista di Scienze Preistoriche 27, 145-161.
Suic, M. (1974) Anticki grad na Istocnom Jadranu. Zagreb.
Tode, A. (1924) 'Archaeologische Landesaufnahme', Vorgeschich-
tliches Jahrbuch 3, 10-41.
The Reconstruction of Environment, Irrigation and
Development of Settlement on the Habur in North-East
Syria

Peter J. Ergenzinger, Wolfgang Frey, Hartmut Khne and

Harald Krschner, Berlin
Geographically, this paper will focus on the Lower Hbr between
the modern provincial capital of Hassaka in the north and the junction
with the Euphrates east of Der az Zor and near modern Buseira
Roman Circesium in the south.
The history of archaeological research in the Hbr area may be
briefly summarized as follows: apart from the early excavations at
Tell Halaf (1) and Tell Fecheriya (2) near Ras el G Ain on the Upper
Hbr, the mounds along the Upper Hbr are unexplored and have
never been systematically surveyed. This is important in view of
recent construction activities here for irrigation purposes by the
Syrian Government. In the fertile area of the so-called Habur triangle
archaeological activity is being carried out at a number of sites: Tell
Brak (3) and Tell Barri (4) on the Gggag, Tell Leilan (5) in the
north-east and very recently at Tell Hamidi (6) on the Gggag and
Tell Mozan (7); a number of systematic surveys along the Gggag
and mainly in the eastern part of the triangle have been undertaken
in recent years. In the past excavations have been carried out at Tell
Chagar Bazar (8), Tell Ailun (9) and Tell Aqab (10).
In contrast to this fairly active picture, the Lower Hbr has
seen only a brief excavation at Tell cAgg (11) in the middle of
the last century by H. A. Layard and a survey (12) at the beginning
of this century. A number of scholars like E. Sachau, M. v. Oppen-
heim, A. Musil, A. Poidebard and M. E. L. Mallowan travelled
through the area but no lasting projects were initiated. This changed
only in 1975 when a systematic survey was begun by the Tbinger
Atlas des Vorderen Orients which was completed in 1977 (13). As a
result, excavation was undertaken at Tell Sb Hamad on the eastern
bank of the Habur in 1978 and continued annually since 1980 (14).
This has been the first systematic excavation in the area; under the
direction of the present writer it is sponsored by the Free University
of Berlin, and carried out in cooperation with the University of
Tbingen and the National Museum of Der az Zor. Under the
direction of A. Mahmoud the excavation in Tell c Agg/Sadikanni
was taken up again in 1982 and continued in 1984 (15). Most recently
 Reconstruction of Environment in North-East Syria 109

Tall Ailun .JSVSSS*

Tall Aqab' Qami
f "^ A Tall Lelan
Ras al ftin
_, Tall Sogar Hamidi
Halaf A Fecheriye Bazar ATal[ Barrj
i Rqdd
Tall Tamer Tall Brak
U
^ %D, mWtZ^F**
1
Saba' ,i/riiii"\<M" 1 1 ' ' . 'iv ' .*
^skur( ^,u|| [ J';;;; ' "'";';;;;,!^!***
Oaba\
f Tall gaga/
6- V,Sadikanni
a

all Seh Hamad/

)ur-katlimmu
Suwarei

Deraz-Zor

Bsra

Mayadln

Nordost- Syrien
Lage der im Text
genannten Orte
A antiker Ort
moderner Ort
Satzpfannen

0 10 20 30 i 50 km
l Entwurf. H.Khne Zeichnung U Schede

Figure 1. Map of the Hbr-area in Northeast-Syria with the sites

mentioned in the text; drawn by U. Schede.
110 ERGENZINGER, FREY, KHNE AND KRSCHNER
a number of excavations have been started in the area of the yabr
dam area, between Hassaka and the mouth of Wadi Ramel, about
30 km south of the city. One of the aims of the Tbinger Atlas survey
had been to reconstruct the development of settlement, especially
during the historical periods of the second and first millennia B.C.,
with respect to the relatively substantial amount of information
provided by the archives of Mari and the Assyrian campaign-reports
(16). It now appears, especially in the light of the most recent results
of the interdisciplinary research at and around Tell Sh Hamad, that
this original aim may be very satisfactorily fulfilled.

PRESENT ECOLOGICAL SITUATION

The ybr is the largest tributary of the Euphrates; it carries water
provided by the karst springs throughout the year. According to an
investigation by the Food and Agriculture Organization of the United
Nations Syria should be divided into five agricultural zones, 'based
on intensity and reliability of rainfall' (17). The yibur region,
between the Turkish border and the Euphrates, falls within zones Ib
through zone 5, thus indicating the strong shift of conditions taking
place in a fairly small area about 350 km in a north-south direction.
Zone Ib is characterized by 'rainfall of over 300 to 600 mm where
two good wheat crops can be obtained each three seasons and a
summer crops of pulses can be grown in rotation with winter cereals'.
This applies only to a small area in the north around Qamishli.
Agricultural zone 4 consists of 'marginal lands with rainfall of 200
to 250 mm and does not drop below 200 mm in 50 percent of the
seasons; lands in this zone are utilized for growing barley or grazing'.
The southern limit of this zone corresponds roughly with the 200
mm isohyet, which crosses the Hbr not far north of Saddada. South
of this line we find agricultural zone 5 which is 'the desert
region ... which has very low rainfall and supports no crops at all.
The only use of this region is grazing'. The line demarcates what is
known as 'dry land border'. Any agriculture south of this line has
to be supported by irrigation. This is the region in which Tell Sh
Hamad is situated, about 60 km south of this agronomische Trocken-
grenze.
But even north of this line crops depend not only on the amount
of rainfall but more so on its timing. In agricultural zone 3 rainfall
remains around 250 mm in only about 50% of the seasons; in
agricultural zone 2 the rainfall ranges between 250 and 300 mm in
66% of the seasons. In other words, it remains fairly risky to grow
crops since the schedule of the rainfall is not reliable even in areas
with 250 to 300 mm precipitation (18). This applies to the rest of the
area between Qamishli in the north and the region north of Saddada.
It was dramatically demonstrated in the winter of 1983/84 when the
 Reconstruction of Environment in North-East Syria 111
rain fell not in December through to February but in April and May
(19). The logical conclusion is, that it is in any case safer to rely on
irrigation for the whole region; risks can be diminished and more
crops can be cultivated. This is also apparently the conclusion the
Syrian Government has arrived at. The latter is planning and already
constructing an irrigation system along the whole length of the Hbr,
including some water reservoirs north-west of Hassaka and a barrage
about 30 km south of Hassaka (20). Any irrigation system for the
Gggag is useless because the Turks have already built a dam and
cut off the water from the Syrians.
A brief comment should be made on soils. Most of the area is
characterized by gypsiferous soils. According to detailed studies in
the neighbouring Balib Valley, 'the soil fertility decreases with the
increase of gypsum content and the decrease of clay and other
colloids' (21). The capability of the soil to retain water depends on
the clay composition; 'good' soil means that it is capable of retaining
water, crops can be grown in zones of unreliable rainfall. In the
steppe area south of the agronomische Trockengrenze there are only
patches of soils of such quality; the Bedouin grow crops there
although it is illegal. For the area around Tell Sh Hamad/Dur-
katlimmu this means that there is no compensation for the low
precipitation so that the area is completely dependent on irrigation
(22).
The above remarks show that it is more profitable for the whole
region of the Habur, not only for the remote southern or lower part,
to rely on an irrigation system wherever possible rather than just on
rain-fed agriculture.

RESULTS OF RECENT RESEARCH IN AND AROUND

THE TELL SH HAMAD
Archaeology. (Fig. 2) A survey of the settlement area prior to excava-
tion and during the first two seasons showed that the settlement
covers more than 120 ha. It is divided into five parts: the Citadel,
Lower City I in the east, Lower City II in the north, Suburban Area
I north of the Lower City II and Suburban Area II east of Lower
City II. According to surface sherds this large area was occupied
during the Iron Age II or Neo-Assyrian period (roughly between 900
and 600 B.C.) (23).
Previously, in the Middle and Late Bronze Age, and after this
period in Parthian-Roman times the settlement appears to have been
restricted to the Citadel and Lower City I. In the Early Bronze Age
the settlement seems to have been concentrated on the Citadel mound
itself and the same applies to the Late Chalcolithic period. The older
levels, i.e. prior to the Late Bronze Age, have not yet been excavated
but are known from pottery of the surface and of the destruction fill.
112 ERGENZINGER, FREY, KHNE AND KRSCHNER

Figure 2. Tell h Hamad, topographical divisions of the settlement

ground and surroundings; drawn after the map 1:25.000 of the Khabour
Region by the Arab Republic of Syria 1960, Major Projects Administration,
by G. Neuber.
 Reconstruction of Environment in North-East Syria 113
On the Citadel mound a Middle Assyrian archive of 550 texts and
fragments was discovered which provided the Assyrian name of the
town, Dur-katlimmu (24). The archive consists mainly of administra-
tive and economic texts, but includes also a number of royal letters
of the kings Salmanassar I and Tukulti Minurta I (25). The archive
was found in what appears to be the administrative and economic
wing of a large building, most probably the palace of the governor
mentioned in the texts (26). Far-reaching trade connections are also
indicated by the archive. These prove that the city of Dur-katlimmu
had become an important administrative and economic centre in the
Middle Assyrian period of the 13th century B.C.
In the north-east corner of Lower City II a large building (27) was
unearthed which seems to be a fortress or the like. Next to it and in
the area around it, vaulted constructions have been excavated which
appear to have served as storage and distribution devices for water
(28). Both can clearly be dated by pottery and small finds to the
eighth century B.C.
In 1984 a freshwater canal (29) with a descent to the west was
discovered next to the town wall indicating that fresh water was
supplied from outside the city, but from the land not the river side.
This points to the existence of an irrigation canal at that time, a
branch of which was probably led to and along the town wall of
Dur-katlimmu (30). In the central part of Lower City II a new
excavation area was opened in 1984. Here a building was found, the
walls of which are decorated with paintings depicting a garden; a
painted inscription reads 'garden house', perhaps indicating the
function of the building as a summer residence of the Neo-Assyrian
governor (31). The style of the paintings can be dated to the late
eighth or seventh century B.C. For the time being, it appears that
the importance of the settlement grew in the Middle Assyrian period
when Dur-katlimmu became the seat of a governor, guarding the
south-western flank of the empire.
In the Neo-Assyrian period the settlement was considerably
enlarged by the foundation of the Lower City II. Dur-katlimmu then
became a military base in the hinterland. We have to assume that,
as a consequence, the population rose considerably. According to
conventional standards one would have to estimate at least 10,000
people, given a settlement area of nearly 100 ha within the walls.
This implies that the Assyrians must have had the necessary supplies
to feed this rising population. As was shown above, the geographical
position of Tell Sh Hamad is not very favourable for a centre of
this size, and the often quoted crossing point of trade routes does
not apply either. The modern population amounts to about 3,500
inhabitants in villages north and south of the ancient settlement area,
and according to a socioeconomic study by Hans Hopfinger from
114 ERGENZINGER, FREY, KHNE AND KRSCHNER
the University of Erlangen this population has already reached the
limits of the available resources for its subsistence. What then were
the means by which the Assyrians were able to maintain Dur-
katlimmu as a provincial centre?
Botany. The investigation of the vegetation around Tell Set Hamad
has shown that owing to abiotic factors xerophytic formations are
dominant, which, due to the composition of their stands and their
diffuse distribution, can only be interpreted as remnants of a former
naturally existing vegetation. This natural steppe vegetation, domi-
nated by Artemisia herba-alba, is almost completely degraded but
must have existed at least in the hill zone between the Habur valley
and the Wadi cAgig. The remnants of this formation show that there
is sufficient rainfall in this hill zone for dwarf shrubs. The substitute
communities of the Poa sinaica-'Kurzrasen' on the hills east of the
Hbr and the Chenopodiaceae-nch steppe formations in more humid
depressions and wadis present today are indicators of a heavily
overgrazed former Artemisia herba-alba steppe.
Large areas in the vicinity of Tell Sb Hamad are characterized
by Prosopis farcta (Prosipedetea farctae halo-segetalia, Zohary 1973)
and indicate that in these areas agriculture based on irrigation existed
in the past. The extension of this formation proves that an older
irrigation system must have existed as the present one does not reach
the outlying parts.
Remnants of Populus euphratica, Tamarix jordanis and Salix safsaf
along the Habur river indicate that in the Hbr valley alluvial forests
(Populetea euphratica, Zohary 1973)32 must have existed which are
completely devastated today and either bushlike or absent altogether.
These alluvial forests with their quickly growing trees were used for
building and construction activities. Remnants of woodlands and
open stands of trees which need more precipitation therefore and
are indicators of a more humid ch'mate have only been found on the
slopes of the Gabal cAbd al-Aziz. They consist of Pistacia khinjuk
and seem to reach their southernmost distribution in these mountains.
Pistacia atlantica, which is found in the Taurus area (north of Gabal
c
Ab al-Azz) and Quercus species, reported from the Gabal Singar
(east of Gabal c Ab al-Aziz) are completely missing here.
The remnants of vegetation in the Gaz ira can be explained as the
result of a regressive succession which is released by direct and
indirect anthropozoogenetic influences (Fig. 3) (33). No indications
were found that the character of the vegetation may have been
different in the Late Holocene (within the last 4,000 years). This is
in accordance with the results of van Zeist and Bottema34, which
were obtained by pollen analysis in north-west Syria (Ghab). The
analysis of macroscopic wood-remains from the excavation has
shown that there is some evidence of coniferous wood. It is interesting
 Reconstruction of Environment in North-East Syria 115

ANTIKER KANALVERLAUF
bei SABA' SKUR

Figure 3. Area of Sabac Skur in the northern part of the Lower Hbr
with the presumable head region of the Neo-Assyrian canal near the
junction of Hadi c Amar and the Hbur at Tell Kerma. Topographical map
drawn after the map 'Khabour Region' 1:25.000 of the Syrian Arab
Republic, 1960, by A. Neuber.

to note that the majority of these remains were discovered in the

Neo-Assyrian building in the north-east corner of Lower City II; on
the other hand, a similar concentration of wood remains of deciduous
species was collected in the Middle Assyrian building on the citadel
mound. While the first evidence clearly points to imported building
material from the Taurus, the latter may prove the existence of fairly
rich alluvial forests at the bottom of the tlbr valley and their
exploitation for local building activities.
Taking into account the actual remnants of vegetation, our knowl-
edge of the regressive succession and the preliminary results of
archaeology and geomorphology the following conclusion can be
given. Remnants of forests or woodlands which might indicate a
previously greater distribution of xero-euxjnian steppe forests were
not found, except in the Gabal cAb al-Aziz. The analyzed remains
116 ERGENZINGER, FREY, KHNE AND KRSCHNER,
of wood from the excavation are mostly from species dominating
the alluvial forests, as they still exist in a devastated stage along the
ysbur. The remnants of an Artemisia herba-alba steppe on the hills
east of Tell Seb Hamad show that there could also have been sufficient
rainfall for non-irrigated arable land, but the area of Tell Seh Hamad
lies in a border region, allowing only sporadically non-irrigated
cultivation in a sequence of ca. 3-5 years.
At present all available data suggests that there have been no major
changes of climate since Assyrian times, i.e. since the middle of the
second millennium B.C., and that therefore there have been no
fundamental changes in the vegetation. That is why there was no
essential and long-term move of the agronomische Trockengrenze to
the south and we may assume that in the area of Dur-katlimmu
rain-fed, large-scale agriculture could not be carried out. Whether
this proves to be correct or whether one might expect some minor
changes of climate which will be revealed by pollen analysis carried
out in cooperation with scientists from the University of Groningen
in 1985, is still to be seen.
Geomorphology. Rivers are changeable; channel pattern, bed form
and sediments respond to changes in water and sediment discharge.
River metamorphosis is also, therefore, a function of environmental
conditions in the river basin.
The Hbr river between Hasseke and Buseira shows all the
patterns of meander development. The low-water channel is
entrenched 1.2 to 3 m below the floodplain. The high-water deposits
are mainly silts, the bar deposits are sands and pebbles, the typical
channel sediments are pebbles, sands and cobbles. The older meander
generations, the flat ox-bow lakes and swampy areas on the floodplain
have a similar geometry to the active meander belts. This is a strong
indication that there has been no important change in the river regime
during the cast century.
Close to Seh Hamad there is a 'tannur' an oven for baking flat
bread which is identical with the present-day tannurs in the upper
part of the floodplain accumulations. Since there is no possibility of
permanent settlement on the floodplain, this former habitat was
visited by Bedouins periodically over a period of time. The age of
the upper 1.3 m of flood silts above the tannur is estimated at 800
years (14C analyses are underway).
At many locations there is a sandy loamy layer with some pebbles
eroded into the meander banks at the base of the river sediments.
Associated with these sediments are ceramics from Assyrian and
Roman times. Thus, present dating relies on archaeological material,
but there are many results indicating the accumulation of almost 2
m of silty sediments across the entire floodplain of the H.br during
the last 2,000 years. The state of the Hbr during Assyrian times is
 Reconstruction of Environment in North-East Syria 117
still unknown, as the bank erosion of the Hbr is not deep enough.
More special investigations are necessary. However, all preliminary
results indicate that the Hbr has had stable environmental condi-
tions and no channel metamorphosis during the last two millennia.
Hydrology. During the geomorphological investigations ancient
irrigation canals were discovered on both banks of the Hbr. The
main canals in the area of Seb Hamad (35) have astonishing
dimensions; the tops of the side walls are almost 6 metres apart and
the total depth is normally 1 to 2 m. The canals were cut through all
sediments of the so-called Pars series, through Quaternary terraces
and, locally, even through low escarpments. Constructional problems
always arose when tributary wadis had to be crossed. Since we
discovered no other traces we suspect that dams of local sediments
were piled up.
By mapping it was possible to link the canals on both sides of the
river with the canals mapped by van Liere (36) 1955 in the region
of Saddade Tell Kerma. The diversion dam or dams of the canals
must have been situated in the area of Tell Kerma or Sab a Skur,
Fig. 3 about 30 km south of Hassaka and about 100 km north of Tell
Sh Hamad. In this area the rjbr has to cross the layers of volcanic
material of the ancient volcano of Kaukab. There are several rapids
and these are favourable natural conditions for the construction of
a diversion dam. Below these rapids the fclbr never erodes hard
rocks. However, no traces of diversion dams have been found. Traces
of the canals can be mapped all the way down from Tell Kerma to
the mouth of the ybr, a distance of over 170 km. In the neighbour-
hood of Tell cAgg there is a larger branch of the main canal; a
complete system of secondary and tertiary canals exists only in a
wadi east of Saddada. At Tell Sh Hamad a trench was dug through
the canal in 1984. The sherds collected on the surface provide the
time range for the existence of the canals; from the Late Bronze Age
(Middle Assyrian) to Early Islamic times. Admittedly the dating
evidence of this material is not very strong, but it is the only material
available, and we know of no written tradition on the canal. Among
the Bedouin there is an oral tradition: the eastern canal is called
Daurin, the western canal is called H c ma. In fact, it was the local
farmers who led us to the tunnelled part of the canal near Tell
Bouderi. Other farmers told us a two-line verse roughly translatable
as: 'Which is better, the H c ma or the Daurin? My heart is with the
good Daurin'.
THE DEVELOPMENT OF SETTLEMENT IN RELATION TO THE
IRRIGATION CANAL
Once the canal had been built it must have had a strong influence
on the development of settlement, on the economic and political
118 ERGENZINGER, FREY, KHNE AND KRSCHNER
situation, and on social life. Mapping the settlements of the different
periods in combination with the canal reveals the following situations:
The distribution of settlements in the Early Bronze Age, (Fig. 4)
is concentrated in the northern part of the Lower Hbr. Centres lie
in the steppe as well as along the river, but south of the line Tell
Matiaha Tell Ahmar only villages are found. In the head region
of the canal a number of smaller settlements are lined up which are
almost exclusively settled during the Early Bronze Age (Rad Saqra,
Tell Kerma, Tell Ragai, Tell Gabi on the east bank and Tell Mul-
lamatar, Tell Gudde on the west bank). This linear arrangement
only becomes intelligible if one assumes that a canal had built been
already. But it certainly did not exceed the above mentioned line,
otherwise centres like Tell Kndig and Tell Masnaqa would have to
be expected farther south as well. This points strongly to one or more
local irrigation systems; it may not be accidental that such a local
system was found in the area of the volcanic ridges which favour
diversion dams (see above).
In contrast to this fairly clear picture the Middle Bronze Age (Fig.
5) does not lend itself to an easy interpretation. The archaeological
material and situation does not indicate any true centre anywhere,
but rather suggests rural communities along the river. The linear
arrangement of Early Bronze Age settlements beginning with Tell
Rad Saqra has disappeared and with it a possible local irrigation
system. It remains hypothetical to see if such a system could have
started with Tell Tnnir and connected all the settlements on the
east bank down to Tell Maqbara cAgg. If one takes the cuneiform
sources into account, Tabatum and Qattunan (Tell Tban and Tell
Fadgami or Tell Asamsni) should be considered centres, and the
archaeological evidence could be interpreted in favour of this
hypothesis. In the case of Saggaratum, though, if it is to be identified
with Tell Abu Ha'yit it is definitely not a centre but a military fortress.
The Mari-texts frequently mention canals but those seem to be
situated along the Euphrates, opposite Terqua, and in the area of
the junction of the Hbr with the Euphrates. Without going into
details, it seems that the political situation at the time was not very
favourable for the construction of a regional canal system along the
Hbr.
The picture changes completely in the Late Bronze Age (Fig. 6).
There is only one place which might be called a Mitannian centre:
Tell Bderi. The others, Tell Hassaka, Tell Tban, Tell Fadgami, Tell
Asamsani and Tell Set Hamad are Middle Assyrian centres (37).
Not a single settlement is found south of Tell Sb Hamad/Dur-
katlimmu. Considering the new political significance of Dur-
katlimmu, its ecological situation, and the scanty dating evidence of
the canal, one has to postulate that the canal had then been construe-
 Hb r -Survey 1975/77TAVO / <* SSSS3*
Siedlungen der Frhen / a. % K"*""
Bronzezeit
{_ MUU r > f W u
1 C . 2 - 2 0 . . o d ouschliniih FB2

Figure 4. Settlements of the Early Bronze Age on the Lower Habur in

relation to the irrigation canals; Survey of the 'Tbinger Atlas des
Vorderen Orients' in 1975 and 1977.
 Hbur Survey 1975/77 TAVO
Siedlungen der altbabylonischen
Zeit ( e. 200O - 15SO v. Chr. )

Figure 5. Settlements of the Middle Bronze Age on the Lower Habr in

relation to the irrigation canals; Survey of the 'Tbinger Atlas des
Vorderen Orients' in 1975 and 1977.
 Hbur Survey 1975/77 TAVO
Siedlungen der mitannischen
und mittelassyrische Zeit
( c. 1550 - 1200 v. Chr. )

Figure 6. Settlements of the Late Bronze on the Lower Hbr in relation

to the irrigation canals; Survey of the 'Tbinger Atlas des Vorderen
Orients' in 1975 and 1977.
122 ERGENZINGER, FREY, KHNE AND KRSCHNER
ted down to Dur-katlimmu. This, in turn, strengthened the centre
mentioned above. There is a technical reason why the construction
of the canal should have stopped at Dur-katlimmu: just south ofthat
city there is a Pliocene ridge which is difficult to cross. The fact that
there are no settlements south of Dur-katlimmu underlines this con-
clusion as well as the military significance of the place: it was the
last bastion protecting the south-west corner of the Middle Assyrian
empire, and this may have been the initial reason for the Assyrians
to choose this place for a stronghold. It remains somewhat enigmatic
why there should be so few villages or smaller settlements; on the
other hand there seems to be a good reason for this. The time to
establish agriculture on the basis of irrigation was too short since
the political power of the Middle Assyrian empire dwindled with the
death of Tukulti-Ninurta I.
In the Iron Age, there can be no doubt that the canal must have
existed over the full length. It gave rise to the most important
centres, lined up with distances of about 40 km, starting with Tell
Hassaka/Margarisi in the north, followed by Tell cvAgg/Sadikanni
and Tell Fadgami/Qatni and ending with Tell Sb Hamad/Dur-
katlimmu (Fig. 7). This corresponds well with the written evidence
that the Assyrians built a 'harran sarri', a royal road, probably along
the eastern bank of the Hbr, at least down to Dur-katlimmu (38).
Along with the centres, the villages and the population increased
mainly in the south where there had been none before; it also appears
that some fortresses were founded in between them (Tell Adla and
Tell Abu Hamda).
The network of centres became closer in the Parthian-Roman and
Early Byzantine period (Fig. 8), followed, almost on the same basis,
by the Islamic period (Fig. 9). Most probably the Mongols destroyed
both long distance canals, and there was no political power afterwards
to restore them. The canals were used thereafter only locally and
periodically, and the nomads were able to take over most of the area.
Travel reports of the last two centuries show that all urban life on
the Hbr had vanished, despite its former prosperity.
The traces of the canals mapped on Figs. 4 to 9 suggest the following
conclusions: because of the uniformity of their appearance and
technical data it seems that they were built all at one time, as part
of one major project. The most probable period at which this was
carried out seems to have been the Neo-Assyrian. It is quite possible
that a part of the canals, mainly the section on the east bank connect-
ing the centres of Tell Tban, Tell Fadgami and Tell Set Hamad,
had been already constructed along the same alignment in Middle
Assyrian times, mainly to provide the newly established provincial
centre of Dur-katlimmu with adequate water supplies for its growing
population and agriculture.
 Hbr-Survey 1975/77 TAVO
Siedlungen der neuassyrischen
Zeit ( cm. 1100 - 600 M. Chr. )

Figure l. Settlements of the Iron Age on the Lower Habur in relation to

the irrigation canals; Survey of the 'Tbinger Atlas des Vorderen Orients'
in 1975 and 1977.
 r^^stliife-j A

ybur-Survey 1975/77 TAVO

Siedlungen der parthisch rmischen
und frhbyzantinischen Zeit ss sUmPf s
Ic. 200 v.Chr. - TOO n.Chr.) r-=9ddinel

Figure 8. Settlements of the Parthian-Roman and early Byzantine period

on the Lower Hbr in relation to the irrigation canals; Survey of the
'Tbinger Atlas des Vorderen Orients' in 1975 and 1977.

_
 1i c P-~
/

-^^
i^^-
x_
Hbr-Survey 1975/77 TAVO
Siedlungen der islamischen Zeit /- wad, .*-|"
( C. 7OO - 14OO n. Chr. ) ^ Sumpf i^Solzsee
A drfliche Siedlungen.ousschlieniich in pa Stdtische
" spdlontiker und/oder islomischet- Zeil twwohnl l Zentren

0 5 10 15 20hm
ENTWURF H KHNE GEZ A NEUBER FEB 19S

Figure 9. Settlements of the Islamic period on the Lower Habr in relation

to the irrigationcanals; Survey of the'Tubinger Atlas des Vorderen Orients'
in 1975 and 1977.
126 ERGENZINGER, FREY, KHNE AND KRSCHNER
As mentioned above, the dating evidence is scanty so far. It is also
surprising that a major activity like this should not have been men-
tioned either in the Middle Assyrian archive of Dur-katlimmu, in
which one misses any mention of a vocabulary in connection with
irrigation canals, or in the Neo-Assyrian annals or other documents,
in which as far as we can see there is again no mention of a
canal along the rjbr.
It also seems possible that older local systems may have existed
along a similar alignment, the traces of which have disappeared in
the course of the construction of the Neo-Assyrian canals. The most
plausible example may be the situation in the Early Bronze Age (see
above). On the other hand, the conditions of the river are not
favourable for an easy diversion of water and hence for several local
systems.
A fuller account of our observations in relation to the settlements
surveyed in 1975 and 1977 will be given in a forthcoming publication.
Finally it may be of interest to note that the alignment of the newly
planned irrigation system will be very similar to the Neo-Assyrian
one so that the traces of it, now still visible, will disappear in the
near future.

NOTES
1. M. Freiherr von Oppenheim, Der Tell Halaf, Leipzig 1931; von Oppenheim
(ed.) Der Tell Halaf I-V.
2. 'Soundings at Tell Fakhariyah'; OIP 79, 1958; A. Moortgat, Archologische
Forschungen der Max Freiherr von Oppenheim-Stiftung im nordlichen Mesopotamien
1955, Kln 1957; A. Moortgat, Archologische Forschungen der M. von Oppenheim-
Stiftung im nordlichen Mesopotamien 1956, Kln 1959, 1-12.
3. most recently: D. Oates, 'Excavations at Tell Brak 1983-84', Iraq 47 (1985)
159-201.
4. P. E. Pecorella/M. Salvini, 'Tell Bari/Kahat' in H. Khne, 'Ausgrabungsttig-
keit in Syrien', Afo 31 (1984) 114-118.
5. H. Weiss, 'Tell Leilan 1982', in H. Khne, Ausgrabungsttigkeit in Syrien, Afo
31 (1984) 156-160
6. S. Eichler, Tell Al-Hamidiya l, Vorbericht 1984, Freiburg 1985.
7. near Amouda; joint expedition by the University of California and the Univer-
sity of Rome under the direction of G. Buccelati.
8. M. E. L. Mallowan, 'Excavation at Brak and Chagar Bazar', in Iraq 9 (1947),
81-89; see also Iraq 3 and 4.
9. A. Moortgat, op. cit. (1959) 13.
10. T. E. Davison and Trevor Watkins, 'Two Seasons of Excavation at Tell Aqab
in the Jezireh, N.E. Syria', Iraq 43 (1981) 1-18.
11. A. H. Layard, Niniveh and Babylon, London 1853, 275-284.
12. E. Sarre and E. Herzfeld, Reise in Euphrat-und Tigrisgebiet I, 1911, 171-195.
13. H. Khne,'Zur historischen Geographie am Unteren Hbr', A/o25 (1974-77)
249-255; W. Rollig and H. Khne, 'The Lower Hbr', Les Annales Archologiques
Arabes Syrienne, 27-28, 77-78, 115-140; H. Khne, 'Zur historischen Geographie am
Unteren Hbr (2), Afo 26, (1978-79) 181-195; W. Rollig and H. Khne, 'The Lower
Hbr (2)', Les Annales Archologiques Arabes Syrienne 33:2, (1983) 187-198.
 Reconstruction of Environment in North-East Syria 127
14. see most recently: H. Khne, 'Tall Seh Hamad/Dur-katlimmu 1981-1983 and
Tall Seh Hamad/Dur-katlimmu 1984', in H. Khne, 'Ausgrabungsttigkeit in Syrien
(3)', /31 (1984) 166-178; H. Khne, 'Tall Seh Hamad/Dur-katlimmu, the assyrian
provincial capital in the mohafazat Der az Zor' (forthcoming), in 'Proceedings of
the International Symposium on the History and Archaeology of the Region of Deir
az-Zor, October 2-6, 1983', Les Annales Archologiques Arabes Syriennes.
15. As c ad Mahmoud, 'Neo-Assyrian Sculptures from Sadikanni (Tell c Agaga)',
in Assur IV, 1983, 67-70; A. Mahmoud, R. Bernbeck, H. Khne, P. Pfalzner, W.
Rollig, 'Die Ausgrabung auf dem Teil cAgaga/Sadikanni 1982', Damas Hzener
Mitteilungen 3 (forthcoming).
16. H. Khne, 'Zur Rekonstruktion der Feldzge Adad-Nirari II., Tukulti-Ninurta
II, und Assurnasirpal II, im Hbr-Gebiet', Baghdader Mitteilungen 11 (1980) 44-70.
17. Regional Study on rainfed agriculture and Agro-climatic inventory of eleven
countries in the Near East Region, Food and Agriculture Organization of the United
Nations, Near East Regional Office/Land and Water Development Division, Rome
1982, 85-97; see also W.J. van Liere, Classification and Rational Utilization of Soils,
Food and Agriculture Organizations of the United Nations, Report No. 2075, Rome
1965, 17 ff., Fig. 1.
18. David and Joan Oates, 'Early Irrigation agriculture in Mesopotamia', in
Problems in Economic and Social Archaeology, London 1976, 111.
19. The International Center for Agricultural Research in the Dry Areas
(ICARDA) Annual Report 1984, XV-XX.
20. H. Hopfinger, 'Ein neues Staudamm-und Bewsserungs Grossprojekt nor-
dostsyrischen Khabur: Grundzge und Probleme seiner Planung', in Geographische
Zeitschrift 72 (1984) 189-195.
21. 'Land Evaluation Criteria for Irrigation', World Soil Resources Report, 50,
Food and Agriculture Organization of the United Nations, Rome 1979, 160-184.
22. Similar observations were made by D. Oates for the area of Assur which lies
approximately on the same latitude as Durkatlimmu: D. Oates, 'The Development
of Assyrian Towns and Cities', in Man, settlement and urbanism, London 1972, 800.
23. For an aerial photograph of the ancient settlement plan see H. Khne, op.
cit. (1984) Fig. 62.
24. W. Rllig, 'Dur-katlimmu', Orientalia 47, 1978, 419-430.
25. W. Rllig, 'Preliminary Remarks on the Middie Assyrian Archive from Tall
Sh Hamad/Dur-katlimmu', in 'Proceedings of the International Symposium on the
History and Archaeology of the Region of Deir az-Zor, October 2-6, 1983', Les
Annales Archologiques Arabes Syriennes, (forthcoming).
26. H. Khne, op. cit. (1984), Fig. 52.
27. ibid, Fig. 58.
28. ibid, Fig. 61.
29. ibid, Fig. 68.
30. see the depression along the eastern town wall on the aerial photograph, ibid,
Rg. 62.
31. ibid 171, Fig. 63-65; for a recent discussion of gardens see D.J. Wiseman,
'Palace and Temple Gardens in the Ancient Near East', in Monarchies and Socio-
Religious Traditions in the Ancient Near East (ed. Prince Takahito Mikasa), Wies-
baden 1984, 36-43.
32. M. Zohary (1973), Geobotanical Foundations in the Middle East 2; see also
H. Pabot (1956) Rapport au Gouvernement de Syrie sur l cologie vgtale et ses
applications, Food and Agriculture Organization (FAO) Rapport No. 663, Rome
(FAO 57/7/4125).
33. H. Krschner (1983), Vegetationsanalytische Untersuchungen an Halophyten-
Zentralanatoliens (Trkei). Tbinger Atlas Vorderer Orient, Reihe A. (Naturwiss).
Supplementary vol. 11, 88p., Wiesbaden; H. Krschner (1984) Der stliche Orta
128 ERGENZINGER, FREY, KHNE AND KRSCHNER
Toroslar (Mittlerer Taurus) und angrenzende Gebiete. Eine formationskundliche Dar-
stellung der Vegetation Sdost-Anatoliens. Tbinger Atlas Vorderer Orient, Reihe A
(Naturwiss). 15, 160p., Wiesbaden.
34. W. van Zeist and S. Bottema (1982), 'Vegetational history of the Eastern-
Mediterranean and the Near East during the last 20,000 years', in Bintliff, J.L. and
H. v. Zeist, Palaeoclimates, Palaeoenvironments and Human Communities in the
Eastern Mediterranean Region in Later Prehistory, pp. 277-321. B.A.R. Oxford.
35. H. Khne, op cit. (1984), Fig. 69.
36. W. J. van Liere and J. Lauffray, 'Nouvelle Prospection Archologique dans
la Haute Jezireh Syrienne', Les Annales Archologiques de Syrie 4-5, (1954-55) 129
ff.
37. Khaled Nashef, 'Babylonien und Assyrien in der zweiten Hlfte des zweiten
Jahrtausends v.Chr.', Karte Bill 7 Tbinger Atlas des Vorderen Orients, Wiesbaden
1983; K. Nashef, Die Orts-und Gewassernamen der mittelbabylonischen und mit-
telassyrischen Zeit, Repertoire Gographique des Textes Cuniformes 5, Supp. rd. to
TAVO Series B Nr. 7/5, Wiesbaden 1982.
38. K. Kessler, Untersuchungen zur historischen Topographie Nordmesopotamiens,
Supp. vol. to TAVO Series B Nr. 26, Wiesbaden 1980, 227-229.
 Site Patterning : Separating Environmental, Cultural
and Preservation Factors

John Bintliff,
School of Archaeological Sciences, Bradford University,

Although site patterning is still not infrequently interpreted in terms

of waves of migration or trade routes, archaeological chronology is
generally a coarse-meshed sieve and such a phenomenon on the
landscape is usually to be seen as a longer-term adaptation to local
environmental conditions by settled, or at least territorially-confined
human communities.
A persistent example of the 'invasion map' interpretation is the
distribution of Anglo-Saxon cemeteries in England (cf. Whitehouse
1975, map 12b), despite the period of several centuries of burial
conflated in such maps, and the neglect shown towards economic
considerations behind settlement patterns. A recent example of the
'trade route' interpretation was a well-publicised suggestion that the
well-known Bronze Age rock-carvings of Yorkshire's desolate Ilkley
Moor marked a prehistoric trunk-road running across Europe to
Italy. A subsequent analysis of the distribution of these stones,
especially the 'cup-and-ring' stones, by Claire Hoare (1984) in rela-
tion to Bronze Age activity on the Moor and the latter's environmental
history, demonstrated that most stones concentrate in a limited zone
arguably of contemporary Bronze Age farming intake, with a lesser
concentration on probable grazing land. The extensive zones least
amenable to either kind of land use are almost devoid of rock art,
but interestingly contain a network of ritual stone circles (cf. Figs. 1
and 2). A purely local interpretation of the distribution (but not
content) of the rock art seems in order. Similarly the 'trade route'
model used to explain the siting of Iron Age oppida in Germany
should probably yield to an agriculturally-based territorial model (cf.
Bintliff, 1984: 222, Fig. 6).
The role of long-distance trade, and industry, in Pre-Industrial
communities has been and continues to be seriously overrated, and
we must accept that the vast majority of activity traces across the
landscape reflect the everyday behaviour of local populations intent
upon the food quest, whether purely for subsistence or partly also
for exchange and taxes (cf. Finley 1973, Bintliff 1981; Bintliff 1984,
Chs. 3 and 5; Garnsey et al. 1983).
130 BINTLIFF

Figure 1. Ilkley Moor, West Yorkshire (England): physical divisions (after

Hoare 1984).

Numbers repr*enl cup

and ngeal*gorles1-9
' t show Clusters

Ritual monuments
R8 Ring barrow
SC Stone circle

Figure 2. Distribution of Bronze Age monuments in north central area of

Ilkley Moor (Heber Moss to Green Crag Slack). (After Hoare 1984).

In many past societies we are fortunate enough to possess patterns

of domestic sites farms, villages or larger settlements but in
some periods, particularly in prehistory, such sites remain poorly
known, and archaeologists have had recourse to surrogates to recon-
struct human activity across the physical landscape. In Yorkshire,
 Site Patterning 131
for example, no reasonable picture of Neolithic domestic sites is yet
available, but the comparison of alternative indicators of human
activity, such as burial mounds, ritual sites, and especially polished
stone axes (Figs. 3 and 4), offers a cumulative argument for human
occupance being concentrated in particular zones of Yorkshire. These
zones, such as the Yorkshire Wolds, the Hambledon and Howardian
Hills, and the drier edges of the Vale of York, are precisely those
dominated by the main expanses in the region of light, fertile soils
most sought after by early farmers. For a similar use of stone axe

Fig 3
GROUP VI AXES I Langdale I

Figure 3. Distribution of Group VI Neolithic axe finds in Yorkshire (after

L. Keen and J. Radley, Pro. Preh. Soc., 35, 1969, figure 4).

Fig 4
GROUP VII AXES IGraigLwyd]

Figure 4. Distribution of Group VII Neolithic axe finds in Yorkshire (after

L. Keen and J. Radley, Procs. Preh. Soc., 35, 1969, figure 5).
132 BINTLIFF
distributions as population indicators one can look to Neolithic
Denmark (Randsborg 1975), and for the use of burial sites as settle-
ment and territory indicators one may cite the familiar work of
Renfrew (1973) on British megaliths, or parallel work on Greek
Bronze Age tombs (Bintliff 1978). But whether by direct or indirect
methods, we can generally create a distribution map of past human
activity across specific landscapes providing the data-base for specu-
lation concerning Man-Landscape interactions.
Still the most useful analytical tool for the initial investigation of
such distributions is some form of Catchment Analysis. Particularly
when a putative domestic site stands in isolation (or its surrogate,
cf. supra), either because of a genuinely original wide-spacing
between contemporary habitations, or because field research has yet
to recover the original neighbouring communities, the mapping of
potential resources at increasing distances or access time from the
occupation site is always illuminating in suggesting what the local
economic preferences might have been, and what constraints under-
lay the development of specific settlements.
The initial theorists of its application to archaeology, Claudio
Vita-Finzi and Eric Higgs (1970), limited the obvious relevance of
the technique to pre- or proto-historic societies, considering it likely
that more complex, state societies would possess a less predictable
attitude to site location. They, and other early practitioners of Catch-
ment Analysis, also made no systematic attempt to modify the theory,
with its set 5-10 km or 1-2 hour radii, to cope with site networks
where nearest neighbours were often vastly overlapping in catch-
ments. Bailey and Davidson's recent evaluation of Catchment Analy-
sis (1983) includes an application to Upper Palaeolithic Cantabria
where this problem of overlapping territories recurs and is not to my
mind satisfactorily resolved by the device of grouping clusters of
sites into a single catchment centred on the most significant site of
each group (cf. their Fig. 2).
However, during the later 1970's essential modifications to Catch-
ment Analysis were made, particularly in the context of high cultures
and civilisations, of which the simplest and probably most useful
was the construction of unweighted Thiessen polygons over a site
network. Each site's location was then analysed in terms of radial
resources to the limits of the Thiessen cells, which were often notably
asymmetric in shape and much more varied in radius than the
heuristic Catchment Analysis guidelines. A key early example was
an application to tell-settlements in Bulgaria (Dennell and Webley
1975), and the method has been used both on Aegean Bronze Age
sites (Bintliff 1977) and more recently on Bronze and Iron Age
settlement distributions in various parts of temperate Europe (Bintliff
1984, Chs. 3 and 5).
 Site Patterning 133
A further area in which catchment principles have had to be
modified concerns the functional interdependence of sites in a settle-
ment network. We are familiar with the handling of sites related in
a yearly-round, or annual territory, for particular hunter-gatherer
groups (cf. Jochim 1976), but hitherto far less attention has been
paid to seasonal or temporary sites used by farming societies as
satellites to home-base farms, villages or towns. This is largely due
to difficulties in separating such sites from permanent domestic sites
(as with the example of the Polish Lausitz Culture, cf. Bintliff 1984:
187ff and 221, Fig. 5). Then there is the typical problem in complex
societies of distinguishing between sites that though permanent settle-
ments, are operating at different levels of a settlement hierarchy.
Objective criteria from excavated material, historic sources, or field
survey, need to be established as discriminants.
In 1976 Prof. Barry Cunliffe published a Thiessen-polygon analysis
of Iron Age hillforts in an area of southern England. If one surroun-
ded each site with a 5 km radius catchment there appeared to be
problems of irrational resource apportionment. But when we see
these sites not as the equivalent of settlement units of farmers,
exploiting their resources radially, but as 'central places' to which
farmers brought produce from numerous farmsteads within the terri-
tory of each hillfort, we can shift to a different kind of catchment
(Fig. 5), an inverse one in which access constraints are those affecting
rural peasants visiting their community focus. The 'best-fit' placing
of a constant territorial module within each Thiessen cell, still at
5 km radius but regardless of the actual hillfort position, now shows
a far better fit with the distribution, and no longer reflects a recurrent
territory farmed from one site, but a recurrent territory controlled
from or focussed on one site, an area of land and doubtless a unit
of dependent population. The similarity in cell size highlighted by
the module circle suggests the crystallisation of social and political
units of comparable size reminiscent of Renfrew's model of 'peer-
polity interaction' (Renfrew 1982: 286ff). The asymmetry of the cells
to several of the hillfort locations can clearly be tolerated provided
that the furthest distances from outlying farms do not regularly exceed
certain limits. Research by historical geographers on rural foci such
as market towns provides numerous well-documented examples of
recurrent servicing modules where there commonly applies a thresh-
old beyond which the peasant is unable to make a day-return to the
focus. Indeed, subsequent research by Cunliffe's team on the rural
hinterland of one of these hillforts, Danebury, has revealed a network
of small Celtic settlements (cf. Cunliffe 1983, Fig. 65).
Let us assume that we have been able to make reasonable progress
n all these points with our chosen site distribution, i.e. clarifying
th
e permanence of sites, their approximate territories, their role in
134 BINTLIFF

ST CATHERINES
HILL

Circles of 5km radius

Figure 5. Thiessen Polygon analysis of early Iron Age hill forts in

Hampshire, with 5 km radius circles 'best-fitted' to each cell (modified after
Cunliffe, from Bintliff 1984, p. 217).

any possible annual round or settlement hierarchy. We are also, as

noted earlier, confining ourselves in this discussion to what is assured
to be the normal situation in Pre-Industrial landscapes, one where
the main function of most sites was that of exploiting local food
sources, with the next most frequent site type being 'central places'
servicing rural populations and likewise, therefore, intimately depen-
dent on the local distribution of food sources (cf. Bintliff 1981).
We will now look for correlations of our site network with geology,
soil groups, topography, altitude and vegetation. Striking examples
of such applications include Jacobi's correlation of Mesolithic sites
in Northern England with an early Holocene treeline hunting
environment between 365 and 490m asl (Jacobi et al. 1976, Fig. 1),
or Sielmann's studies of the influence of climatic and pedological
variables on the agricultural settlement of temperate Europe by the
Linear Pottery culture (Sielmann 1972, 1976). If a settlement
hierarchy exists or an annual round, we may try to reconstruct the
mutual feedback between centre and political satellites, or the
scheduling of resources, respectively. Finally, we may take the total
site network and make suggestions about population size and density,
 Site Patterning 135
and if we are lucky enough to detect changes in distributions for the
same landscape over time, to suggest reasons for and implications
of perceived demographic fluctuations, especially if the use of resour-
ces is visibly being altered. In my own model of the settlement
correlates of agrarian cycles in European regional landscapes (Bintliff
1982), I argued for a major mutual feedback between changing use
of resources and political development.

Having set forth the procedures that I and many other scholars
operate in regional research, I would now like to undermine this
whole approach by calling for a radical reassessment of the data
base we have been manipulating, i.e. the original pattern of sites
detected archaeologically. My critique centres on two questions:
1. To what extent does the differential preservation/visibility of
sites today hinder a realistic appreciation of the original patterning
of sites in a landscape; i.e. can we evaluate missing sites'?
2. To what extent does the specifically human attribute of Culture
distort any proposed ideal fit between human settlement units and
the distribution of natural resources across the landscape; i.e. can
the Environmental Archaeologist exist independently of the Social
and even Symbolic Archaeologist?
I would like to illustrate these problems from the data-base of the
Boeotia Project, a regional survey in Central Greece (Figure 6a) of
which I am co-director with Prof. Anthony Snodgrass (Bintliff and
Snodgrass 1985, Bintliff 1985, Bintliff and Snodgrass 1988).
With field-by-field survey in close order, we have uncovered a
remarkable density of rural settlement for the climax era of Classical
Greek civilisation (Fig. 6: note the scale, and a density of some 3-4
sites per km 2 ). Three of these sites are towns, the rest are almost all
interpreted as small farm sites. My initial treatment of this data had
shown two things. Firstly, that the recovered distribution reflected a
preferred soil type in the region. Secondly, that if one assumed a
farm holding module of 5.4 ha, as argued by Jameson (1978) to have
been a common size for the period, a reasonable packing of such
units was revealed over a very limited zone of the landscape (Fig.
7a,b).
However, in Boeotia we could attest to a well-known feature of
field survey, the phenomenon of surface sites that appear one year
only to disappear the next. Clearly the distribution I show here can
only be a fragment of that originally present on the Classical land-
scape but how much is missing? Classical historical sources offer
a possible guide, for they do permit some reasonable estimate of the
size of population in particular city-state territories. A comparison
of recovered density of settlement with that predicted from the ancient
sources for this region suggests that something like 57% of sites have
 V
^^^*sim ix

W UONI
^EQKjr$ "**

THESPIAE

BOEOTIA SURVEY

ARCHAIC/ CLASSICAL

Figure 6. (a) Location of the area subjected to intensive archaeological

survey between 1979 to 1984 within South-West Boeotia, Central Greece.
(Modern villages denoted by filled circles, ancient sites by squares and
triangles), (b) Distribution of Archaic-Classical sites in South-West
Boeotia, derived from intensive archaeological survey. (Definite settlement
sites denoted by filled circles, probableby open circles, doubtful by
question marks).
 I HALIARTOS

THESPIAE

BOEOTIA SURVEY

ARCHAIC/CLASSICAL

I HALIARTOS

THESPIAE

BOEOTIA SURVEY

ARCHAIC/CLASSICAL

Figure 7. (a) Thiessen Polygon analysis of figure 6b. (b) Putative 5.4
hectare farm modules centred on all Archaic-Classical rural sites in South-
West Boeotia.
138 BINTLIFF
been uncovered in the survey area for this period. To the problem
of where to suggest our hypothetical 43% of missing sites should be
placed can be added the obviously related fact that the recovered
distribution and the attached theoretical farm territories occupy only
certain sectors of the land surface taken up by the preferred soil type.
Yet calculations from the sources suggest that population density at
that time was such that all available farming land would then have
been in use.
One explanatory factor must be farming activity carried out by
residents of our three town sites. Yet this still leaves the likely further
participation of our 'missing' farms, assumed to be very numerous
across this same landscape. Total soil mapping and the evaluation
of erosional-depositional conditions over the study area suggest that
no systematic local biasses are present which could be operating to
produce better site survival or site visibility in one sector as opposed
to another. Our working hypothesis is therefore that the site sample
recovered is a reasonable guide to the location of the missing farm
sites.
Now the multiplying-up of almost as many sites again into the
sectors where sites already concentrate will make my earlier farm-
holding territories set around each site quite fallacious and misleading
(Fig. 7). Several considerations support this argument, however.
Firstly, within the sectors of concentrated known sites we find over
the years that some of these sites can disappear and reappear in the
surface record, whilst in areas carefully surveyed previously invisible
sites come strongly into view. Secondly, Classical sources suggest
that farm holdings were often dispersed in several discrete parcels
rather than being concentrated in one location. The known sites could
thus mark a small parcel of land, and hence the multiplication of
sites between known examples becomes permissible. Thirdly, our
techniques of fieldwalking involve a total count of visible surface
pottery densities across every field, allowing us to relate known dense
nucleations of pottery (sites) to non-site or 'off-site' densities. We
believe at present that the abundant off-site pottery reflects primarily
deliberate manuring activities contemporary with our two main peri-
ods of intensive rural settlement, Classical and Late Roman. Sig-
nificantly, there are large sectors of the landscape, virtually all the
less favourable soil, with low off-site pottery. The intense activity is
highly correlated with the known site clusters.
To achieve this working interpretative model we have had to make
many assumptions which are open to question; we are also unusually
favoured by historical sources, although these too can easily be
criticised. But the implications of this case-study (if our preliminary
interpretation is correct), for regional survey strategies elsewhere,
are fairly disturbing. This proves to be even more the case when we
 Site Patterning 139
turn to prehistoric site distributions, for our provisional calculations
suggest that later prehistoric sites are mostly invisible or 'missing'
on the modern land surface even to the most intensive modern field
survey techniques.
However, let us proceed with our working model, whereby the
recovered site distribution for the Classical period is spatially rep-
resentative of the total distribution. As mentioned earlier, with the
exception of a handful of probable shrine sites, our distribution (Fig.
6) consists of small farms and three large urban centres. The farms
are notably non-random in pattern, even within the preferred soil
type. We have just argued that this phenomenon would be merely
reinforced by the discovery of all the 'missing' farms. An obvious
next step for investigation is the interaction with the three towns or
'central places' (cf Fig. 8). Study of site distribution in relation to
putative state boundaries reveals that rural settlement is heavily
concentrated in the immediate neighbourhood of Thespiae and
Askra, but thins out towards the Haliartos borderlands. There is
a minor possible boundary effect between Thespiae and Askra,

I HALIARTOS

THESPIAE

BOEOTIA SURVEY

ARCHAIC/CLASSICAL

Figure 8. Distribution of Archaic-Classical sites in South-West Boeotia

with Thiessen 'catchment' boundaries for the three major town sites of this
diagram.
140 BINTLIFF
however, which is understandable since we know from history that
whereas Thespiae and Haliartos were independent states, Askra was
merely a satellite community of Thespiae.
Beyond the Askra-Thespiae heartlands, borderland settlement
takes two apparent forms: a) In the north-east we see a hamlet-cluster;
a remote area of good land at a distance from any of the three towns
and cut off in most directions by poorer resources; b) In the north-west
and north we see a thin scatter of farms over extensive good soils.
In the hamlet case a small satellite community can fully exploit a
remote territory, but with the scattered farms we could suggest either
a less intensive land use or an exploitation of these sectors from the
relevant town (Askra or Thespiae), or from the dense farmsteads to
be found much closer to those towns. In fact study of off-site surface
pottery densities confirms that our two types of borderland settlement
correlate with indicators of intensive (hamlet) and non-intensive
(scattered farm) land use.
Clearly a complex attitude to land use and settlement is operating.
'Least Effort' principles would probably be best satisfied by an even
spread of rural farms across the best soils, but what we seem to be
finding are too many farms jostling one another round the towns,
and too few in the rural hinterland, except for an unusual pocket of
farms in a remote hamlet. Most plausibly social factors are crucially
operative here; primarily the desirability of living in close proximity
to the civic centre of each polity with its amenities so cherished by
the citizens of the ancient Greek city-states. To this may reasonably
be added the likelihood that partible inheritance gave a farmer several
scattered parcels of land, encouraging him to opt for residence near
the town or at least in a rural hamlet or 'kome' if his lands were
particularly remote.
But a final issue needs to be raised. Having mapped the potential
resources of the survey area, and with some idea of agricultural
technology and crop yields in antiquity, an estimate of the relation
between potential carrying capacity and actual settlement data is in
order. If we take the archaeologically-recovered Classical-(Early)
Hellenistic site density, settlement overall is dense, which befits a
period traditionally viewed as the Classical Greek florescence in both
cultural and economic terms. But if we adopt the higher population
density that the historical sources (epigraphic and literary) suggest,
and see our archaeologically-recovered sample as little more than
half of the original pattern of sites, we find that settlement in Boeotia
was packed at some 90% of maximum food production if not more.
Such a situation is very abnormal for recorded agricultural societies,
and is a dangerously high exploitation.
On the other hand, this picture of a very populous, to alarmingly
over-populated Classical landscape is increasingly confirmed from
intensive archaeological survey in other parts of Greece, and seems
 Site Patterning 141
to arise very dramatically (Snodgrass 1980) and die dramatically. We
believe that this is an example of a long-term 'agrarian cycle' (Bintliff
1982). Initially, in the post-Mycenaean (Bronze Age) Dark Ages,
population seems to have been minimal for some centuries; it then
rose in a few generations to an impressive density, subsequently
overshooting into over-exploitation of fragile resources; the stress
resulting is surely reflected in both Greek colonisation abroad and
in the internecine wars of the Greek city-states and their constant
attempts to appropriate each other's land. Not surprisingly perhaps,
a collapse of settlement in the Greek countryside and a severe urban
contraction can be shown archaeologically in Boeotia from c200 BC,
and lasts for several centuries, before a new cycle begins with a
dramatic revival in town and country in Late Roman times, followed
by a further Dark Age collapse in the Early Byzantine era.
So at no time do we seem to have a predictable match or semi-
equilibrium of settlement and population paired to potential resour-
ces; the picture is either of striking overuse or underuse of the
landscape, or of a rapid upward or downward trend with long-term
social and political implications. Secondly, although available crops
and technology significantly affect the intensity to which exploitation
levels can go, (so that, for example, there is little doubt that Classical
populations peaked far above Bronze Age climax populations in this
area), this markedly unstable picture, with long periods of abnormally
low population alternating via short, dramatic leaps and falls with
climax population inflation near crisis point, cannot be understood
in purely environmental terms, but has to have a socio-political
dimension. Chronic insecurity may be a major feature in the minimal
populations of the pre-Classical and post-Roman Dark Ages, reckless
competitive expansion may underlie the putative overpopulation of
Classical times, elite and state-centred mismanagement of the land
underlie the Late Roman expansion. And increasingly, field evidence
suggests, such overuse of the land may have led inevitably to soil
deterioration and erosion, possibly more than any other factor usher-
ing in the population minima of the post-Classical and post-Roman
areas (cf. Bintliff and Snodgrass 1985, Pope and Van Andel 1984).
Similar conclusions were reached by Prof. R. McC. Adams in his
(1978) analysis of Mesopotamian human ecology over the millennia.
In this case study, the strong distorting effect of community
decisions on the exploitation of the environment, on a social and
political plane, falls into place in terms of what Orlove (1980) recently
characterised as a new era in Economic Anthropology, in which
economic determinism has had to make significant concessions to
cultural factors.
In the era of Classical Greek florescence, the city-state communities
show sufficient security to adopt what might be seen as a partial
Least Effort approach to their land exploitation dispersed rural
 142 BINTLIFF
farms; yet social factors prevent an even spread over the area of
cultivation. By the late Roman era, although our city-states are now
merely town councils within the all-encompassing Roman Empire,
attachment to the citizens' town still conditions settlement patterning,
whilst the relative prosperity and importance of each town is reflected
in absolute rural densities. After the post-Roman Dark Ages a new
settlement pattern comes into being. By at least the llth century AD
and probably from much earlier, the population is increasingly con-
centrated into a network of nucleated hamlets and villages, on or
not far from the modern village sites. These locations are of two
kinds: one is clearly defensive and peripheral to the land exploited,
the other more exposed and on or near ancient town sites. Over time
the trend is clearly towards the former locations. Even where these
villages grow in the recent era to 2,000 or more inhabitants, the
practice of, and even word for, a rural farm has disappeared from
the landscape. The reasons for this post-antique trend are probably
rooted in the need for village solidarity, which became essential
during the Middle Ages and under the Turkish Empire against
insecurity and the aggressive exactions of feudal and imperial powers.
The mega-village is comparable to the slightly less numerous petty
city-states of Classical Greece, but the exploitation of its hinterland
continues to be distorted from ideal access principles, a situation
that has become less detrimental with the advent of mechanised farm
transport.

Archaeological settlement distributions deserve radical reassessment

for a proper appreciation of their original patterning, number and
exploitation of the environment. Even when progress has been made
to reconstruct a plausible original patterning, no one-to-one ideal
spread of settlements can be expected, fitting the resource contours;
rather, aspects of community life, which may be quite period-specific,
or political and military pressures, will systematically distort an ideal
human ecology of the landscape.

BIBLIOGRAPHY
Adams, R. Mc C. (1978) Strategies of maximisation, stability, and
resilience in Mesopotamian society, settlement, and
agriculture, Procs. of the'American Philosophical Soc. 122, 329-
335.
Bailey, G. N. & I. Davidson (1983) Site exploitation territories
and topography: Two case studies from Palaeolithic Spain', J.
of Archaeol. Science 10, 87-115.
Bintliff, J. L. (1977) Natural Environment and Human Settlement
in Prehistoric Greece. British Archaeological Reports, Oxford.
Bintliff, J. L. (1978) The number of burials in the Messara tholoi,
etc., pp. 24-30, 80-83, 83-4, in An archaeological survey of the

L
 Site Patterning 143
Ayiofarango valley, edited by D. Blackman and K. Branigan
Annual of the British School at Athens 72, 13-84.
Bintliff, J. L. (1981) Theory and reality in Palaeoeconomy: some
words of encouragement to the archaeologist, pp. 35-50 in A.
Sheridan and G. Bailey (Eds.), Economic Archaeology. British
Archaeological Reports, Oxford.
Bintliff, J. L. (1982) Settlement patterns, land tenure and social
structure; a diachronic model, pp. 106-111 in C. Renfrew and
S. Shennan (Eds.) Ranking, Resource and Exchange.
Cambridge University Press, Cambridge.
Bintliff, J. L. (1985) The Boeotia Survey, Central Greece, pp. 196-
216 in S. Macready and F. H. Thompson (Eds.)
Archaeological Field Survey in Britain and Abroad. Society of
Antiquaries, London.
Bintliff, J. L. & A. M. Snodgrass (1985) The Boeotia Survey, a
preliminary report: The first four years, J. of Field
Archaeology 12, 123-161.
Bintliff, J. L. & A. M. Snodgrass (1988) The end of the Roman
countryside: A view from the East, pp. 175-217 in R. F. Jones
(Ed.) Europe in the First Millennium AD. British Archaeologi-
cal Reports, Oxford.
Cunliffe, B. (1983) Danebury. Anatomy of an Iron Age Hillfort.
Batsford, London.
Dennell, R. W. & D. Webley (1975) Prehistoric economy and land
use in Southern Bulgaria, pp. 97-109 in E. S. Higgs (Ed.)
Palaeoeconomy. Cambridge University Press, Cambridge.
Finley, M. I. (1973) The Ancient Economy. Chatto and Windus,
London.
Garnsey, P., K. Hopkins & C. R. Whittaker (1983) (Eds.) Trade in
the Ancient Economy. Chatto and Windus, London.
Hoare, C. (1984) Cup and Ring Marked Rocks and the Prehistoric
Environment on Ilkley Moor. Unpublished Undergraduate
Dissertation, Bradford University, School of Archaeological
Sciences.
Jacobi, R. M., J. H. Tallis & P. A. Mellars (1976) The Southern
Pennine Mesolithic and the Ecological Record, J. of
Archaeolog. Science 3, 307-320.
Jameson, M. H. (1978) Agriculture and slavery in Classical
Athens, Classical Journal 73, 122-145.
Jochim, M. (1976) Hunter-Gatherer Subsistence-Settlement
Systems: A Predictive Model. Academic Press, New York.
Orlove, B. S. (1980) Ecological Anthropology, Ann. Rev.
Anthropology 9, 235-273.
Pope, K. O. & T. H. Van Andel (1984) Late Quaternary
Alluviation and Soil Formation in the Southern Argolid: its
History, Causes and Archaeological Implications, J. of
Archaeol. Science 11, 281-306.
Randsborg, K. (1975) Social dimensions of Early Neolithic
Denmark, Procs. Preh. Soc. 41, 105-118.
Renfrew, C. (1973) Before Civilisation. Jonathan Cape, London.
Renfrew, C. and Wagstaff, M. (Eds.) (1982) An Island Polity. The
Archaeology of Exploitation in Melos. Cambridge University
Press, Cambridge.
Sielmann, B. (1972) Die frhneolithische Besiedlung
144 BINTLIFF
Mitteleuropas, pp. 1-65 in J. Lning (Ed.) Die Anfnge des
Neolithikums vom Orient bis Nordeuropa: Westliches
Mitteleuropa. Fundamenta Va. Buhlau Verlag, Kln.
Sielmann, B. (1976) Der Einfluss der geographischen Umwelt auf
die linien- und stichbandkeramische Besiedlung des
Mittelelbe-Saale-Gebietes, Jschr. Mitteidt. Vorgesch. 60,
305-329.
Vita-Finzi, C. & E. S. Higgs (1970) Prehistoric economy in the Mt.
Carmel area of Palestine: site catchment analysis, Procs.
Prehistoric Soc. 36, 1-37.
Snodgrass, A. (1980) Archaic Greece. Dent, London.
Whitehouse D. (1975) Medieval Europe from the Fifth to the
Twelfth Century, Part IV, pp. 281-333 in D. Collins (Ed.) The
Origins of Europe. Allen and Unwin, London.
 An integrative model for the use of different landscapes
in Linearbandkeramik times

A. J. Kalis and A. Zimmermann,

Goethe University, Frankfurt, West Germany

In prehistoric times there were periods for which we have only special
aspects of the archaeological sources (settlements, graves, hoards
and so on). For example many settlements are known from Linear-
bandkeramik time, but only a few graves or graveyards; whereas the
Tumulus Culture in Germany is almost entirely represented by graves.
There are even periods for which we only can assume that there
existed as many people before as after; we cannot therefore be sure
that regions without finds were really unused in prehistoric times.
For instance it is particularly difficult to find evidence for activities
which produced only a little waste. In Bandkeramik time many
permanent settlements are known, which are typically situated on a
very good soil, the loess (Sielmann 1971, 1972; Dohrn-Ihmig 1979;
Linke 1976). However, hunting and fishing camps are rare exceptions
(Taute 1967). It will be argued that landscapes without permanent
settlements in Bandkeramik time were also used for activities which
are related to the food producing type of economy.

ARCHAEOLOGICAL BACKGROUND
The Bandkeramik way of life during the period between 5300 and
4900 B.C. (approximate calibrated C14 dates) was distributed over
large parts of Europe. These first farmers built the first large houses,
made ceramics and used ground stone adzes. The production of
goods in excess of those needed by individual communities is another
aspect of the economy which started at that time.
Because of an open-cast browncoal mine in the Aldenhovener
Platte region in the lower Rhine basin, nearly all features in an area
of 1.3 by 1.0 km were excavated. Along the valley of the Merzbach
Bandkeramik settlements covered the slopes. The analysis of ceramic
decoration has become the basis for a detailed chronology (Stehli,
in preparation). It showed that in a long period of middle Band-
keramik time there existed houses every few hundred meters; and
intensive and systematic survey over a period of eight years in a
region of 85 km has produced a pattern not very different for a larger
146 KALIS AND ZIMMERMANN
area. Regions with Bandkeramik settlements existed up to distances
of 3 km perpendicular to the valleys (Aid. PI. V; 496; Luning 1982).
PALYNOLOGICAL INVESTIGATIONS
Palynological investigations were made in order to reconstruct the
Bandkeramik vegetation cover. We were lucky to find a suitable
deposit near a settlement in the form of a colluvial deposit at the
bottom of a small valley, and a 2.5 cm interval pollen diagram has
been made from this deposit. Two C-14 dates indicate that the pollen
diagram covers the required part of the early Neolithic period.
We will not show the complete pollen diagram of Broekfeld because
of its complexity and the high number of its pollen types. Here only
the most important features are summarized (Fig. 1). At the left, the
three species Tilia and Quercus and the shrub species Corylus is
shown, and at the right, the sum of the NAP, of the grasses and of
the cereals. The diagram demonstrates four important aspects of the
vegetation: a) The tree layer was dominated by lime (Tilia cf. cordata).
b) There was a gradual change in the forest composition. At the
bottom of the diagram we see an increase of hazel (Corylus) con-
nected with a decrease of lime. Throughout the following period,
however, oak (Quercus) and hazel decrease and lime gradually
increases, c) The percentages of the total NAP as well as the percen-
tages of the grasses are low. d) The NAP show a high species diversity,
especially from synanthropic plant species. We propose the following
interpretation:
The synanthropic plants show farming activities, but the arable
land is assumed to be restricted. Considering the bad pollen dispersal
the high lime pollen values suggest a dense lime forest situated on
the loess plateau near the deposit; a forest which according to Iversen
(1973) should be dark and without undergrowth. These palynological
data corroborate the reconstruction of Bakels (1978) and Greig
(1982). The rather low percentages of oak pollen demonstrate that
the clearance should have been quite restricted. Light and open areas
would favour the sprouting of oak and hazel, whereas the regener-
ation of the lime as a shade-sprouting plant species is hindered.
Our picture of the Bandkeramik environment, therefore, is that
the loess plateaux were covered by a dark and dense lime forest with
only small fields, which contradicts the reconstruction of field sizes
by Luning (1979/80). Oak, elm and ash grew in the flood plains of
broad valleys (Auen) and bordered the small brooks.

EXPLOITATION OF THE ENVIRONMENT

To what extent was feeding of cattle possible in the loess regions?
No evidence exists for pastoral activities in the pollen diagrams,
which is to be expected, because lime forest seems not to be suitable
 Integrative Model for the use of different landscapes 147

SUMMARY POLLEN DIAGRAM OF BROEKVELD

-I ^-i-^ .J 1 1 1 I I 1 J

250-

300-

350-

r
40% 50% 40%
AP pollen types NAP pollen types

Quercus Poaceae
- Tilia sum NAP
-g Corylus ~^ sum Cerealia

Figure 1. Summary Pollen Diagram of Broekveld.

148 KALIS AND ZIMMERMANN
for this purpose. Moreover, it is doubtful if the requirements for
building materials could have been satisfied by the small number of
oak trees in the neighbourhood of the settlements (Fig. 1 and Bakels
1982). Space in the loess regions was sufficient only for the arable
land (Liining 1979/80). Other landscapes were needed to satisfy the
requirements for cattle fodder and building materials. Possibilities
include the regions north and south of the Bandkeramik settled
landscape, and the intermediate lower mountain ranges.
In order to get an idea of the potential use of other landscapes we
will try to reconstruct the forest composition of the middle Atlantic
period. This reconstruction is based on palynological data from Dutch
Limburg (Janssen 1960), Aldenhovener Platte (Kalis in preparation),
Hohes Venn (Persch 1950; Schutrumpf and Schmidt 1977) and Eifel
(Straka 1975). It is assumed that this reconstruction be generalized
for the mountains with a Subatlantic climate and therefore with
vegetation belts of the 'Vosges-type' (Haeupler 1970). The plains
were covered with a dense lime forest on both poor and rich soils,
except where there was a high groundwater level. Those soils were
covered with a mixed oak-elm forest on rich soils, and an oak-alder
forest on poor soils. The gently sloping landscape with hills up to
300 m was also covered with lime forest, but the natural morphology
gave the vegetation pattern more diversity (e.g. more oak on sun
exposed slopes, elm and ash bordering the brooklets and small rivers).
This was the landscape where the people of Bandkeramik time settled.
Above 300 m oak became a more important part of the vegetation
cover, because lime with its preference for a more temperate climate
and deep fertile soils was limited in its distribution. Above 500 m
oak became the major component of the vegetation. At that time the
soils were still base-rich and this, combined with the high precipita-
tion at that altitude, encouraged a forest vegetation with a dense
shrub layer of hazel and and a luxuriant undergrowth. According to
Firbas (1949), these forests should be compared with the recent
'Auen- und Schluchtwlder'. In the valleys and on the northern slopes
extensive elm forests were present.
Areas suitable as pasture for cattle would be confined to the forest
types found in the river valleys of the lowlands and at middle
altitudes, and all the forest above 500 m. The area covered by the
river valleys is restricted but the mountain forest types cover large
areas. This could have been a reason for the behaviour described by
the term 'transhumance'. A transhumance system could explain the
lack of evidence for pastoral activities in the loess areas. An economy
of this type in Neolithic times is considered for Poland (Bogucki
1982, 1984) and for the eastern Pyrenees (Jalut 1974; Geddes
1983).
 Integrative Model for the use of different landscapes 149
EVIDENCE FOR HUMAN ACTIVITIES AT MIDDLE AND
MOUNTAINOUS ALTITUDES
Even if there were no permanent settlements, there are some archaeo-
logical indicators that the lower mountains really were used in Band-
keramik time. One indicator is the presence of Neolithic shoe-
last adzes in those regions without permanent settlements. From our
point of view it is not important that in most cases the hoes cannot
be dated with certainty to Bandkeramik times. There are examples
from Bohemia (Zapotocka 1982), the Pfalz (Kriesel 1978), the region
around Trier (Lhr, in preparation) and in the northern Sauerland
(R. Blank 1984, personal information). It is most understandable
that the adzes, which are woodworking tools, are found in the forests
which were visited during Bandkeramik time.
Another important item of evidence for the use of the lower
mountain regions is the distribution of artefacts made from stone
from different sources. If the hypothesis were true that only the
loess-regions were used in Bandkeramik times, then the main direc-
tion of the exchange system should be oriented along these regions.
If, on the other hand, the hypothesis were true that all regions were
used, then the smallest distance between sources of raw material and
the settlements where the finds were last deposited would be of
greater importance than the distribution of the loess-areas. As a test,
the relative frequency of western flint in Westphalia (Gabriel 1974)
is compared with that in the Trier region (Schmidgen, in preparation)
and that in Hesse (Hampel 1984; Zimmermann in preparation). Most
of the western flint comes from the Rijckholt region (Southern
Netherlands). The shortest route from Rijckholt to the settlements
in Westphalia led mainly through the loess regions. The shortest
route to the Trier region crosses the Eifel Mountains. The way from
Rijckholt to the Trier region along the Loess-region along the loess
zones is increased from 130 km to ca. 210 km. The way to Hesse
should have led at least partly through the lower mountains. The
correlation between the shortest distance and the frequency of
western flint is much greater than the correlation between distance
along the loess zones and the frequency of western flint. If these
observations can be confirmed by the raw material spectra in other
regions, then our assumption is justified that the lower mountain
ranges were regularly visited in Bandkeramik times.
Our observation of Rijckholt-flint in a Bandkeramik settlement of
the oldest phase in Hesse (Friedberg-Bruchenbrucken) has confirmed
the impression that the regions which were not occupied by per-
manent settlements were regularly visited, because in the direction
of Rijckholt, Bandkeramik settlements did not exist in this period
(Dorn-Ihmig 1979).
We assume that possible Neolithic land use in the mountain ranges
can explain features in the pollen diagrams. The regional pollen
150 KALIS AND ZIMMERMANN
deposition in the loess landscape during Bandkeramik times is
characterized by slight decrease in the frequency of elm pollen and
a marked increase of ash pollen, features which are most probably
connected with Neolithic forest exploitation (Kalis, in preparation).
The features are visible in the regional pollen diagrams of the neigh-
bouring mountain ranges (for the Venn mountains see Persch 1950,
Schutrumpf and Schmidt 1977; for the Eifel, Straka 1975; for the
Vosges, Kalis 1984). The decrease of elm and the increase of ash
pollen characterize the boundary between the pollen zones Firbas
VI and VII (Firbas 1949). A satisfying explanation for this feature
has not been given until now, because only climatic reasons have
been discussed, and because anthropogenic explanations did not
seem reasonable, assuming that the mountain ranges were not used
at that time. The correlation between the first elm in the pollen
diagrams (for the data compare Huntley and Birks 1983), and the
expansion of the Neolithic economy in Central Europe including the
mountain ranges, suggests such a causal relation.
Archaeological and botanical evidence from the region just north
of the Alps has been discussed as a 'Pre-Pottery Neolithic' before
Bandkeramik time (Kossack and Schmeidl 1974/75), or perhaps as
contemporary with the Bandkeramik (Menke 1978). It is possible
that these phenomena correspond with our observation in the lower
mountains. Late Mesolithic sites contemporaneous and not too far
away from Bandkeramik settlements are so far unknown either north
or south of the Bandkeramik settled land, therefore it is possible that
even these regions belonged to the sphere of economic interests for
the Bandkeramik people.

CONCLUSIONS
What has been presented is a hypothesis of a more extended and
more specialised land use in early Neolithic time than that held until
now. If all types of landscape in between the loess regions were used
in Bandkeramik time, they should be included in future research
strategies to verify our hypothesis. What are needed now are dated
pollen diagrams from the mountain ranges; investigations of
geomorphological features such as early colluvial deposits in the
valleys (an example for later periods is by Poeteray et al. 1984);
investigation of archaeological sites with a special environment as a
priority and macro-regional archaeological investigations even if they
include regions which seem to be without sites.
Regional archaeological work, at least for Bandkeramik materials,
has been undertaken in areas of about 150 km diameter (Dohrn-Ihmig
1979, Gabriel 1979, Meier-Arendt 1966, Linke 1976, Sielmann 1972).
Possibilities for comparison of larger areas exist only on a very
general level (e.g., Pavlu 1977). What we need are studies at a
 Integrative Model for the use of different landscapes 151
macro-regional scale, which focus on landscapes wider than the
regions with concentrations of sites.

BIBLIOGRAPHY
Aid. PI. V (1975) R. Kuper, H. Lohr, J. Luning, W. Schwellnus, P.
Stehli, A. Zimmermann, Untersuchungen zur neolithischen
Besiedlung der Aldenhovener Platte V, Bonner Jahrb. 175, 191
ff.
Barker, G. W. W. (1975) Early Neolithic land use in Yugoslavia,
Proc. Prehist Soc. 41, 85-104.
Bakels, C. C. (1978) Four Linearbandkeramik settlements and
their environment: A Paleoecological study of Sittard, Stein,
Elsloo and Hienheim, Analecta Praehist. Leidensia XL
Bakels, C. C. (1982) Zum wirtschaftlichen Nutzungsraum einer
bandkeramischen Siedlung, Kolloquium Nove Vozokany 1981,
Nitra 1982.
Bogucki, P. I. (1982) Early Neolithic Subsistence and Settlement in
the Polish Lowland. BAR 150, Oxford.
Bogucki, P. I. (1984) Ceramic sieves of the linear pottery culture
and their economic implications, Oxford Journal of Arch. 3,
15-31.
Dohrn-Ihmig, M. (1979) Bandkeramik an Mittel-und Niederrhein,
Rheinische Ausgrabungen 19, 191-362.
Firbas, F. (1949) Spat-und nacheiszeitliche Waldgeschichte nordlich
der Alpen. Jena.
Geddes, D. S. (1983) Neolithic transhumance in the Mediter-
ranean Pyrenees, World Archaeology 15.1, 51-66.
Greig, J. (1982) Past and present lime woods of Europe, pp 23-25
in : M. Bell and S. Limbrey, (Eds) Archaeological Aspects of
Woodland Archaeology. BAR 146, Oxford.
Grooth, M. E. Th. & G. J. Verwers (1984) Op goede gronden.
Leiden 1984.
Haeupler, H. (1970) Vorschlge zur Abgrenzung der Hohenstufen
der Vegetation im Rahmen der Mitteleuropakartierung,
Gttinger Floristische Rundbriefe 70.1, 3-15.
Hampel, A. (1984) Die Linienbandkeramische Siedlung im
Frankfurter Osthafen. Schriften des Frankfurter Museums fr-
Vor-und Frhgeschichte VII. Bonn.
Huntley, B. & H. J. B. Birks (1983) An atlas of past and present
pollen maps for 0-13.000 years ago. Cambridge.
Iversen, J. (1973) The development of Denmark's nature since the
last glacial. Danmarks Geologiske Undersogelse 5.7-C.
Kobenhavn.
Jalut, G. (1974) Evolution de la vgtation et variations climatiques
durant les quinze millnaires dans Vextremit oriental des
Pyrnes. Thse d'Etat, Toulouse, No. CNRS A08190, 1815.
Janssen, C. R. (1960) On the late-glacial and post-glacial vegeta-
tion of South Limburg (Netherlands), Wentia 4, 1-112.
Kossack, G & H. Schmeidle (1974/75) Vorneolithischer
Getreidebau im bayerischen Alpenvorland, Jahresbericht der
Bayerischen Bodendenkmalpflege 15/16, 7-23.
Kriesel, O. (1978) Zur vorgeschichtlichen Besiedlung des Kreises
Kirchheimbolanden (Pfalz). Speyer.
152 KALIS AND ZIMMERMANN
Linke, W. (1976) Frhestes Bauerntum und geographische Umwelt.
Bochumer Geographische Arbeiten 28.
Luning, J. (1979/80) Bandkeramische Pflge?, Fundberichte aus
Hessen 19/20, Festschrift U. Fischer (1980) pp 55-68.
Luning, J. (1982) Siedlung und Siedlungslandschaft in
bandkeramischer und Rossener Zeit, Offa 39, 9-33.
Menke, M. (1978) Zum Frhneolithikum zwischen Jura und
Alpenrand, Germania 56.1, 24-52.
Pavlu, I. (1977) Das linienbandkeramische Ornament und seine
Westausbreitung, Arbeits-und Forschungsber. zur schsischen
Bodendenkmalpflege 22 205-218.
Persch, F. (1950) Zur postglazialen Wald-und Moorentwicklung
im Hohen Venn, Decheniana 104, Bonn, 81-93.
Poeteray, F. A., P. A. Riezebos, & R. T. Slotboom, Rates of
subatlantic surface lowering calculated from mardel-trapped
material (Gutland, Luxembourg), Zeitschrift fr
Geomorphologie N.F. 28, 467-481.
Schutrumpf, R. & B. Smidt (1977) Zusammenarbeit zwischen
Vegetationsgeschichte und Dendrochronologie, aufgezeigt
an zwei ausgewhlten Beispielen, Erdwissenschaftliche
Forschung 13, 28-42.
Sielmann, B. (1971) Der Einfluss der Umwelt auf die neolithische
Besiedlung Sdwestdeutschlands unter besondere
Bercksichtigung der Verhltnisse am nrdlichen Oberrhein.
Acta Arch, et Praehist. 2.
Sielmann, B. (1972) Die Frhneolithische Besiedelung
Mitteleuropas. Fundamenta Va, Ed. H. Schwabedissen,
Reihe A, Bd. 3.
Stehli, P. (in prep) Die Chronologie der bandkeramischen
Siedlungen im Merzbachtal.
Sterud, E. (1978) Prehistoric populations in the Dinaric Alps: An
investigation of interregional interaction, pp 381-408 in: C.
Redman et.al. (Ed.), Social Archaeology: Beyond Subsistence
and Dating. New York, Academic Press.
Straka, H. (1975) Die sptquartre Vegetationsgeschichte der
Vulkaneifel, Beitrge zur Landespflege in Rheinland-Pf ah 3,
1-163.
Taute, W. (1967) Das Felsdach Lautereck, eine mesolithisch
neolithisch bronzezeitliche Stratigraphie an der oberen
Donau, Palaeohistoria 12, 483-504.
 Prehistorie Agriculture in the New Guinea Highlands:
Problems in Defining the Altitudinal Limits
to Growth

Tim Bayliss-Smith,
University of Cambridge
Attempts to 'predict' the size or density of prehistoric populations
based on the calculation of 'carrying capacities' have proved remark-
ably unsuccessful in cases where it has been possible to test the
accuracy of the predictions made. For example, on atolls in Polynesia
or on small islands in Fiji it has been shown that populations at
European contact were less than 70 per cent of calculated carrying
capacity, and in almost all non-insular situations the shortfall is much
greater (Bayliss-Smith 1974, 1980). A multitude of cultural factors
intervene to prevent the calculated limits to population growth having
any connection with reality, in cases where that reality has been
ethnographically observed or has been painstakingly reconstructed
from archaeological sources.
For archaeologists there is a more useful concept than carrying
capacity. Instead of calculating the maximum population that a
prehistoric technology could have supported, we should be consider-
ing what range of populations could be supported at specified levels
of welfare (Bayliss-Smith 1978, 1980). These so-called 'standard
populations' are constituted with reference to specified levels of
surplus production from agriculture, and/or specified levels of labour
input into cultivation, so that the demographic implications of various
standards of wealth and/or leisure can be assessed. This approach
is less deterministic: it enables the archaeologist to modify his popula-
tion predictions according to what welfare standards his supplemen-
tary information about a society suggests are reasonable. Conversely,
if estimates of prehistoric population levels are available from some
independent source (for example, number of house sites or settle-
ments), then a reverse procedure can be adopted. The population
data can be used to predict the levels of production surplus that
could have been achieved by, and/or the level of labour intensity
necessary to sustain, such a population.
The aim of this paper is to illustrate a further way in which the
'standard population' approach can be used. The general question
to be posed is this: what were the altitudinal limits to prehistoric
agricultural settlement in the New Guinea Highlands? In New
154 BAYLISS-SMITH
Guinea (Fig. 1) an agricultural prehistory stretching back at least
6,000 years has been demonstrated for the mid-altitude zone, par-
ticularly at the archaeological site at Kuk in the upper Wahgi valley,
excavated by Jack Golson (1977, 1981, 1982). Kuk is at 1,600 m
above sea level, an altitude at which many tropical crops can still be
cultivated. For the high-altitude zone, however, above 2,000 m, the
archaeological evidence for agriculture is much more ambiguous.
The evidence from high-altitude comes from a variety of sites and
can be used to support a variety of interpretations concerning resource
use. On the Sirunki plateau vegetation changes consistent with inter-
mittent forest clearance, possibly for swiddens, have been inferred
for one site at 2,500 m starting about 4,300 years ago (Walker and
Flenley 1979). At Tambul (2,340 m), in the upper Kaugel valley, a
wooden spade has been found identical to types used elsewhere for
digging ditches associated with intensive swamp cultivation. This
artifact is dated to 4,000 BP (J. Golson, personal communication).
The earliest dates for the occupation of rock shelters at 2,050 m in
the Manim valley are quite similar (4,340 BP), but here hunting and
gathering in the montane forest is regarded as the most likely reason
for occupation of these sites (Christensen 1975).
These three examples indicate the ambiguity concerning high alti-
tude resource use prior to the introduction of the sweet potato
(Ipomoea batatas) in the last two or three centuries. Of the pre-
Ipomoean staples, taro (Colocasia esculenta) appears to be the only
crop capable of being successfully cultivated at these altitudes but
its economic role is difficult to assess because of its almost total
replacement by sweet potato, white potato and other temperate crops
in present-day agricultural systems. It is in this situation of ambiguity
and ignorance concerning the nature of prehistoric resource use
throughout an ecological zone that the standard populations

-4000m

Maximum altitude
of cultivation
-3000
TARO
Pandanus
Puera ri a
BANANAS
-2000
YD- alata
-D bulbitera
M~D. pentaphylla
S D. esculenta
-SAGO. BREADFRUIT
-1000

-0

Figure 1. Altitudinal range of the major Highlands crops in New Guinea.

 Prehistoric Agriculture in the New Guinea Highlands 155
approach becomes valuable. What kind of pre-Ipomoean economy
could have been maintained at such altitudes? What would have
been the 'costs' of high altitude settlement if based on taro as the
staple crop, and what effect do assumptions about the level of
agricultural intensity have on predicted labour requirements?
The archaeological and palaeoenvironmental data suggest three
possible alternatives for land use in the high altitude zone:
1. Ipomoean Revolution hypothesis. The notion that prior to the
sweet potato the highlands economy was based on small patrilocal
bands of hunters and intermittent gardeners (Watson 1965) has been
refuted for the main valleys such as the upper Wahgi, but could still
be justified for the high-altitude zone. It may be that even taro
performs so poorly above 2,000 m that hunting and gathering would
have been the only viable basis for subsistence.
2. Taro swiddening hypothesis. On a small scale, taro is cultivated
in swiddens today as high as 2,700 m in places, even though its
present-day subsistence role is very restricted. Bowers (1968) and
Clarke (1977) both consider that a low intensity system of land use
based on taro would be a viable high-altitude economy, although
greatly inferior to swidden cultivation using the recent crop introduc-
tions.
3. Intensive taro cultivation hypothesis.The spade discovered at
Tambul suggests the possibility that as well as swidden cultivation,
an intensive agricultural system based on swamp drainage for taro
might also have existed, at least in the upper Kaugel valley. The date
for the Tambul artifact (4,000 BP) is synchronous with the beginning
of Phase 3 at Kuk, where the evidence suggests that as well as drainage
for dryland cultivation there were 'chains of small interconnected
basins probably for the water-tolerant crops', and that 'it is highly
likely that taro was prominent among the water tolerant crops'
(Golson, 1981: 58). The digging of networks of ditches to drain
wetland for taro cultivation is an occasional practice today in the
upper Kaugel valley (Bayliss-Smith 1985). This hypothesis would
demand that such practices, instead of being small-scale adjuncts to
agricultural systems based on the sweet potato, were instead the
primary focus of the economy.
Given that environments like the upper Kaugel were, in pre-sweet
potato times, distinctly marginal for agriculture, can data be gener-
ated on the economics of taro subsistence which can help us to accept
as plausible, or reject as unlikely, any of these three hypotheses?
In any ecologically marginal area settlers have to accept three
kinds of 'cost'. One cost is the reduced range of crops that can be
grown at all (e.g Fig. 1). Second, there is a reduction in yield if crops
are being grown outside their optimal range. In a pre-industrial
society this decline in yield is felt either through the increased labour
156 BAYLISS-SMITH
input needed to achieve the same amount of production, or through
the need to accept a smaller surplus. Thirdly, and accompanying this
decline in diversity and productivity there is normally an increase in
risk. For example, there is a positive correlation between the length
of the dry season in the tropics and variability in total rainfall: the
more seasonal environments are also the more risky ones (Walsh
1981). On tropical mountains the fall in average temperature with
altitude is accompanied by an increased risk of occasional frost
damage (Brookfield 1964). Very marginal, and therefore very hazar-
dous environments may be unoccupied altogether because crop losses
occur so often as to outweigh the benefits to be gained from land
use during intervals of more 'average' conditions. Much depends on
the accuracy of hazard perception and the availability of alternative
strategies such as pastoralism, exchange or migration (Watts 1983).
For New Guinea the evidence suggests that in the high-altitude
zone, between 2,000 m and 2,700 m, frost damage is an intermittent
hazard, but that severe crop damage only occurs every 2-3 decades.
Moreover there existed in some areas long distance exchange relation-
ships between high and intermediate altitude groups which enabled
populations to escape from famine by migration (Waddell 1975). In
this and other ways mechanisms existed to enable high-altitude
groups to cope with the severe effects of occasional disasters.
Nevertheless, it is clear from the ethnographic literature that in
normal times high-altitude groups had to be effectively self-sufficient
in their staple food supply.
The question of the productivity of taro at high altitude thus
becomes crucial for an understanding of population-environment
relationships in the New Guinea Highlands in prehistory. Little
information is available for taro in this ecological zone, so in 1980-81
a number of surveys were carried out to provide data on this crop.
Surveys were done at Baisu (1,600 m), which is close to the Kuk
archaeological site, at Bimin (1,400-1,900 m), and in the upper
Kaugel valley at Kiripia (2,180 m) and Tambul (2,340 m). These
studies covered swamp cultivation technology, labour inputs and
taro yields, and were carried out partly through observations on
indigenous agricultural systems where taro is still grown, and partly
through the monitoring of experimental plots (Bayliss-Smith 1985).
The data enable us to reconstruct the subsistence economics of taro.
Fig. 2 shows a comparison between yields and labour intensity for
three swidden systems and four swamp cultivation systems. Three of
the wetland systems are my experimental plots, one at Baisu and two
at the high-altitude site of Tambul.
It is clear that the swamp cultivation systems at Tambul are con-
sistently less productive and more labour intensive than anything at
intermediate altitude, either the three swidden systems or the swamp
 Prehistorie Agriculture in the New Guinea Highlands 157

Swidden c u l t i v a t i o n
A Swamp c u l t i v a t i o n fBAISU
20- + E x p e r i m e n t a l plot

.16-

KAPAUKU
12-
BIMIN - W O K A I M I N

o TSEMBAGA W A R I N G
CC
BOMOGAI-ANGOIANG

TAMBUL VALLEY +

TAMBUL TERRACE+

0 2000 4000 6000
WORK INPUT hours/ha-yr

Figure 2. Yields and labour requirements of taro (Calocasia esculenta) in

the New Guinea Highlands (for data sources, see Bayliss-Smith 1985).

drainage systems at Kapauku and Baisu. We can calculate the impli-

cations of these data for the amount of work needed to maintain a
population in the high-altitude zone. To this end we use the higher
Tambul yield figure (5.2 tonnes per ha) and its labour requirement
of 4,642 person-hours per ha. How dependent would a prehistoric
population have been upon this yield of taro? Today most New
Guinea highlanders depend upon the sweet potato for over 90% of
dietary energy, but if we look at fringe highlanders still cultivating
mainly taro there is a lesser degree of rootcrop dependence. A
Mountain Ok group studied by Hyndman (1982) depended upon
taro for 66% of their energy. This group had access to numerous
supplementary foods not available at high altitude. Nevertheless if
we take this consumption level as a conservative estimate of taro
requirements for a pre-Ipomoean population at Tambul, then the
labour inputs necessary for 66% dependence on intensive taro culti-
vation at Tambul can be calculated.
For calculation purposes (fuller details are given in Bayliss-Smith
1985) we assume a household of six persons plus one domestic pig,
158 BAYLI SS-SMITH
requiring in total 20,664 MJ of food energy per year, 13,638 MJ of
which must come from taro. The 5.2 tonnes of gross yield would
produce 21,260 MJ per ha (Wills et al. 1983), implying that 0.64 ha
would be required to sustain the household, which in turn would
require 2,971 hours of work. This work input amounts to 57 hours
per week for one producer per household, or 28.5 hours per week if
there are two producers; in either case a very heavy burden, par-
ticularly considering that through this work only two thirds of food
needs are being met and that other nonagricultural tasks still have
to be done. Today, even in areas of high human and pig population
density in Enga province, men only spend 15 hours per week and
women 25 hours in all gardening activities (Waddell 1972: 101). In
Kainantu district, Eastern Highlands, married men spend 12.5 hours
per week on gardening and women 21.4 hours (Grossman 1984:
270-275).
These data suggest that intensive taro cultivation would not be
regarded today as a viable basis for high-altitude subsistence.
Archaeological evidence for wetland drainage systems, as at Tambul,
is best interpreted as representing an intensive but minor component
in subsistence economies which are much more broadly based.
Hypothesis 3 concerning pre-Ipomoean subsistence can therefore be
rejected.
A similar procedure can be carried out to test the plausibility of
high-altitude swidden cultivation (hypotheses 1 and 2). Here the data
indicate that a similar yield can be gained from swidden as from
wetland cultivation but suggest that clearance of forest plots would
require much less work (see Fig. 2). If we use the same yield figure
but assume for swiddens 1,000 person-hours of work per hectare,
approximately the median level of input in the observed systems,
then we can calculate that this alternative technology would provide
two-thirds of subsistence needs in return for 640 hours of annual
work per household. This represents 12 hours per week per producer
if there is one productive adult in the household, or 6 hours per week
if there are two. Such an activity level seems a good deal more
reasonable, and indeed conforms quite closely to work levels among
swidden cultivators studied soon after colonial contact (e.g. Rap-
paport 1967). Shifting cultivation may also reduce the risk of taro
being lost because of frost damage. Forest clearings on east-facing
slopes have a protected microclimate, whereas crops grown on valley
floors may be vulnerable to cold air drainage from surrounding slopes,
and the absence of a tree cover increases the likelihood of radiation
frosts.

We can conclude that in the high-altitude zone of New Guinea

shifting cultivation based on taro would have constituted a viable
 Prehistoric Agriculture in the New Guinea Highlands 159
economy, at least for a sparse population with access to supplemen-
tary resources. Agricultural colonisation of this ecological zone was
not necessarily dependent on an Ipomoean Revolution. Instead we
can hypothesise that it was the much earlier diffusion of taro-based
agriculture that provided the stimulus for agricultural settlement at
high altitude. The revolution was Colocasian not Ipomoean.
From the methodological point of view, this paper has shown that
alternative models of prehistoric land use can be tested by means of
a reversal of the usual standard populations procedure. A specified
population (i.e. one household) and specified levels of consumption
have been used to predict the labour demands of alternative forms
of agriculture. The predictions enable us to suggest that swidden
(hypothesis 2) represents an attractive and reasonable economy,
supplying basic energy needs and a small surplus with a tolerable
work load, whereas swamp cultivation alternatives require an excess-
ively high level of labour intensity without any increase in yield. If
ethnoarchaeological data are available for the relevant crops and
cultivation techniques, then there seems to be scope for this approach
to be more widely adopted. It constitutes an objective technique for
assessing the welfare implications of alternative hypotheses concern-
ing prehistoric land use, and as such could be applied to a wide
range of archaeological problems.
BIBLIOGRAPHY
Bayliss-Smith, T. P. (1974) Constraints on population growth: the
case of the Polynesian Outlier atolls in the pre-contact period.
Human Ecology 2, 259-296.
Bayliss-Smith, T. P. (1978) Maximum populations and standard
populations: the carrying capacity question, pp. 129-151 in
Social organisation and settlement: contributions from
anthropology, archaeology and geography edited by D. Green,
C. Haselgrove and M. Spriggs. British Archaeological Reports
International Series (Supplementary), No. 47, Oxford.
Bayliss-Smith, T. P. (1980) Population pressure, resources and
welfare: towards a more realistic measure of carrying
capacity, pp. 61-94 in Population-environment relations in
tropical islands: the case of Eastern Fiji, edited by H.C.
Brookfield. MAB Technical Papers No. 13, UNESCO, Paris.
Bayliss-Smith, T. P. (1985) Pre-Ipomoean agriculture in the New
Guinea Highlands above 2000 metres: some experimental
data on taro cultivation, pp. 285-320 in Prehistoric intensive
agriculture in the tropics, edited by I.S. Carrington. British
Archaeological Reports International Series, No. 232, Oxford.
Bowers, N. (1968) The ascending grasslands: an anthropological
study of ecological succession. Ph.D dissertation, Department
of Anthropology, Columbia University, and University
Microfilms, Ann Arbor, No. 69-394.
Brookfield, H. C. (1964) The ecology of highland settlement: some
suggestions. American Anthropologist 66(4), part 2, spec.
pubn., 2-39.
160 BAYLISS-SMITH
Christensen, O. A. (1975) Hunters and horticulturalists: a
preliminary report of the 1972-4 excavations in the Manim
Valley, Papua, New Guinea. Mankind 10, 24-36.
Clarke, W. C. (1977) A Change of staple in prehistoric New
Guinea, pp. 159-163 in Proceedings of the Third Symposium of
the International Society for Tropical Root Crops, edited by
C. L. A. Leakey. International Institute of Tropical Agricul-
ture, Ibadan.
Golson, J. (1977) No room at the top: agricultural intensification
in the New Guinea Highlands, pp. 601-638 in Sunda and
Sahul, edited by J. Allen, J. Golson and R. Jones. Academic
Press, London.
Golson, J. (1981) New Guinea agricultural history: a case study
pp. 55-64 in A history of agriculture in Papua New Guinea: a
time to plant and a time to uproot, edited by D. Denoon and
C. Snowden, Institute of Papua New Guinea Studies, Port
Moresby.
Golson, J. (1982) The Ipomoean Revolution revisited: society and
the sweet potato in the upper Wahgi valley, pp. 109-136 in
Inequality in New Guinea Highland societies, edited by A.
Strathern, Cambridge University Press.
Grossman, L. S. (1984) Peasants, subsistence ecology and
development in the Highlands of Papua New Guinea. Princeton
University Press, New Jersey.
Hyndman, D. C. (1982) Biotype gradient in a diversified New
Guinea subsistence system, Human Ecology 10, 219-260.
Rappaport, C. A. (1967) Pigs for the ancestors. Ritual in the
ecology of a New Guinea people. Yale University Press, New
Haven.
Waddell, E. W. (1972) The Mound Builders. Agricultural practices,
environment and society in the central Highlands of New
Guinea. American Ethnological Society, Monograph No 53,
University of Washington Press, Seattle.
Waddell, E. W. (1975) How the Enga cope with frost: responses
to climatic perturbations in the central Highlands of New
Guinea, Human Ecology 3, 249-273.
Walker, D. & J. R. Flenley (1979) Late Quaternary vegetation
history of the Enga province of upland New Guinea,
Philosophical Transactions of the Royal Society of London, ser.
B, 286, pp. 265-344.
Walsh, R. P. D. (1981) The nature of climatic seasonality, pp.
11-21 in Seasonal dimensions to rural poverty, edited by R.
Chambers, R. Longhurst and A. Pacey, Pinter, London.
Watson, J. B. (1965) From hunting to horticulture in the New
Guinea Highlands, Ethnology 4, 295-309.
Watts, M. (1983) On the poverty of theory: natural hazards
research in context, pp. 231-262 in Interpretation of calamity,
edited by K. Hewitt, Allen and Unwin, London.
Wills, R. B. H., J. S. K. Lim, H. Greenfield & T. P. Bayliss-Smith
(1983) Nutrient composition of taro (Colocasia esculenta)
cultivars from the New Guinea Highlands, Journal of the
Science of Food and Agriculture 34, 1137-1142.
 Trial By Dure: A Review of Historical-Geographical
Concepts Relevant To The Archaeology Of
Settlement On Corsica And Sardinia

James Lewthwaite, Lancaster University

The aim of this essay is to interest Mediterranean archaeologists in

certain concepts of a geographical and historical nature: more
specifically, those associated with the French traditions of human
geography (founded by P. Vidal de la Blache) and of a genre of
historical scholarship which encompasses much that would pass for
social science elsewhere; that associated with the journal Annales in
its various guises and with its successive editors at least up to 1968
(L. Febvre, M. Bloch and F. Braudel). By way of introduction, lest
a tonic so alien to the archaeological palate be served a little too
cold from the cellar, some objections a priori may be refuted.
Firstly, why should archaeologists look back for conceptual inspir-
ation? Because there are all too few genuinely new ideas under the
sun, only a lot of old ones going round in the wash (environmental
and evolutionary perspectives on culture, the supposed significance
of the 'long term', positivism etc.) which periodically resurface to be
discussed as if for the first time, in ignorance of all previous debate.
Knowledge of itself not being cumulative, however many books and
journals now jostle for space on library shelves, it is incumbent on
each generation to familiarise itself with the history of ideas and the
controversies of their forerunners, understood as far as possible in
their contemporary context rather than in the light of their utility to
some present-day quarrel (Berdoulay 1981b, Buttimer 1978, 1981):
otherwise we merely tread water.
Secondly, why should archaeologists interest themselves in history
and geography? Because, particularly during the great late nine-
teenth-century evolutionary radiation and niche differentiation of the
numerous modern species of academic, questions which are now
regarded as properly falling within the province of one or other
established discipline were far more widely discussed; problems now
considered to be purely the preserve of the prehistorian, and to be
soluble primarily in terms of archaeological data, such as 'The Human
Occupation of the Earth' or 'The Genesis of Pastoralism' would
formerly have been discussed by ethnologists and historical geogra-
phers making inferences from distributional data. True synthesis must
162 JAMES LEWTHWAITE
pay attention to all such sources of insight (e.g. in the case of
pastoralism, Sherratt 1981). Secondly, to borrow a trope from Clarke
(1972), because archaeologists burden themselves with a rummaged
lumber-room full of historical and geographical clichs when dealing
with a region as travelled and studied as the Mediterranean; by
identifying and confronting the sources of such unconsciously
imperative modes of thought we may hope to save ourselves from
perpetuating such errors to the detriment of another generation. In
effect, by pausing to empty our mental rucksacks of such intellectual
paraphernalia, we may then make faster progress.
Thirdly, why look to French scholarship in particular? Because of
the manifestly enduring and ubiquitous acceptance of the high quality
of Vidalian and Annaliste scholarship, evident in the sheer quantity
of the literature devoted in recent years to chronicling their respective
life histories (cf. bibliography). This seems to be a consequence of
the institutional centralisation of French academic life (Clark 1973;
Coutau-Bgarie 1983, 1-28, 248-82; Hexter 1972, 450-98) which, in
the first place, concentrated the brightest minds of all France into
the intellectual hot-house of fin de sicle Paris, with its formal and
informal opportunities for inter-disciplinary discussion (the salons,
the ambience of the Quartier Latin, cf. Buttimer 1981); secondly,
through the system of institutionalised patronage and authoritarian
control ('le patron et son cercle', Clark 1971) which enabled the
greatest, most ambitious and most enterprising to propagate their
doctrines throughout France, and indeed over several generations,
through the judicious management of the careers of their protgs
and the other resources at their disposal, such as junior posts, journal
space, access to publishers, ministerial contacts. (Clark 1973; Claval
1969, 21; Coutau-Bgarie 1983, 16-28, 245-316; Rhein 1982,234-40.)
As a result, particular points of view have come to be disseminated,
distilled and elaborated to a greater degree than elsewhere, so that
a critical mass of erudition sharing a common orientation has
accumulated, whereas the fragmentation of research has otherwise
caused the artisanal efforts of individual scholars to fade into
oblivion.
Fourthly, why bother to examine the context of concepts? Because
'good' ideas have not simply ousted 'bad' ones (Berdoulay 1981b),
however much successful traditions seek to secrete a retrospective
legend in which their triumph is depicted as inevitable, despite their
opponents' intrigues, with alternative viewpoints travestied as intel-
lectual culs-de-sac and ideological opponents condemned to infamy
or non-person status (the Sorbonnistes, P. Aris, G. Dumzil until
recently...: Berdoulay 1981a; Braudel 1972a, 456-62; Burguire
1979; Claval 1968a; Coutau-Bgarie 1982, 295-316; Hexter 1972,
507-8; Revel 1979). The success of ideas is dependent on that of
their proponents, in turn determined by external constraints, (e.g.
 Trial By Dure 163
ideological compatibility with the Zeitgeist; institutionalisation prior
to or during a phase of growth), and individual destinies (the fates
of some - e.g. C. Vallaux, J. Brunhes - to teach in marginal institutions
allowing little opportunity to foster an intellectual progeny; of others
to be strategically well placed to step into a central chair through a
fortuitous demise or a well planned coup; of Durkheim's brightest
young students to die or be maimed in World War I). Moreover, it
required an effort of intellectual dpaysement to understand the
factors polarising a debate after the passage of so many years; indeed
Andrews (1984) following Rhein (1982) and Berdoulay (1978) has
argued that the Vidalian-Durkheimian polemics turned largely on
mutual misunderstandings between close contemporaries. It is only
too easy to attribute a specious consistency or contradiction to a
scholar through the bricolage of selective quotations taken from the
voluminous literature heaped up by lively, original and eclectic minds
without due regard to the context of some episodic utterance or
ephemeral jotting.
Fifthly, can an old concept be revived in a different context?
Logically, one supposes that a concept divorced from its original
context and stripped of its connotational nuances would, even more
than Heraclitus' river, not be the same concept (but cf. Buttimer
1978, 67). However, there is a clear difference in purpose between
striving to understand as accurately as possible the subtleties of
meaning conveyed to a first-generation Vidalian by terms such as
milieu, genre de vie, civilisation etc. ('the archaeology of knowledge'),
and evaluating their continuing utility to archaeologists in the present
day (cf. Buttimer 1978, 59, 67-72). The strongest argument in favour
of this is surely the fact that such concepts changed their significance
for successive generations of scholars within the Vidalian tradition
(Buttimer 1971, 111-20, 156-95).

THE VIDALIAN TRADITION OF HUMAN GEOGRAPHY

Without in any way belittling Vidal's personal achievements, par-
ticularly on acceding to a Sorbonne chair in 1898, which permitted
him to supervise the doctoral theses of a generation of future eminent
geographers (Rhein 1982, 240), it is legitimate to emphasise the
intellectual debt which he owed to the various schools of thought
that contributed to his personal synthesis of la gographie humaine
(Vidal de la Blache 1922) and the external circumstances which
favoured geographers in France with a place in the sun in terms of
material resources (Berdoulay 1981a, Broc 1974,1976, Buttimer 1971,
Claval 1968b, 1984, Lukermann 1965, Rhein 1982).
With respect to the latter:
(i) the trauma of the defeat of 1870 at the hands of Germany, a
country long admired by French academics for its freedom and
164 JAMES LEWTHWAITE
intellectual achievements, occasioned mixed feelings of inferiority
and revanchism (Berdoulay 198la, 17-43); Goethe's jibe that the
French were a people with moustaches who were ignorant of geogra-
phy touched a raw nerve in the acknowledgement that French army
officers were lacking in the topographic skills appropriate to modern
warfare. In the long run, the open sore of the question of Alsace-
Lorraine raised questions of the criteria on which nationality should
be denned, opening up a clear polarisation between the views of
Ratzel on the one hand and Vidal (in La France de l'Est, 1917) on
the other (Allegra and Torre 1977, 139-43);
(ii) the perception of geography as a useful fount of practical
knowledge, by the various political pressure groups which were
promoting the creation of a colonial empire, inspired the formation
of many regional geographic societies and certain national associ-
ations and journals. By 1970, this movement had largely achieved
its ends and won its cause, although Vidal's principal collaborator
in the editorship of the Annales de Gographie set up in 1891, M.
Dubois, continued to teach and promote a specifically colonial con-
ception of geography (Berdoulay 1981a, 45-75; Broc 1974; Rhein
1982, 235);
(iii) the institutionalisation of geography through the establish-
ment of several university chairs during the last decade of the nine-
teenth century (in 1896, fifteen new regional universities were created)
owed much to the role assigned by the ideologues of the Third
Republic, and in particular successive reformist ministers of educa-
tion such as V. Duruy, J. Simon and J. Ferry, to academics in general:
in brief, the staff of the new universities were expected, as the 'Jesuits'
of the lay democratic republican state, to mould a generation of
informed, responsible citizens. The Dreyfus affair, in the course of
which the term 'intellectual' was first coined, and in which many of
Vidal's pupils were actively engaged, provided a litmus test. One
consequence of this attention on the part of the state was the transfor-
mation of the provincial Faculties from mere seats of the administra-
tion of examinations into independent centres of research activity
(Berdoulay 1981a, 77-108).
In purely formal terms, therefore, the groundwork for the organisa-
tion of geographic teaching was in place when Vidal came to occupy
the Sorbonne chair of the conservative A. Himly in 1898, after a
distinguished teaching career at Nancy and, since 1877, at the Ecole
Normale Suprieure. Thanks to the labours of pioneers such as L.
Drapeyron and E. Levasseur a network of academic posts, amateur
and professional groupings, publishing outlets (Armand Colin,
Hachette), cartographic bureaux and other material assets, had been
established, all focused on the Parisian institutions which Vidal was
in a position to influence (Berdoulay 1981a, 7-16, 107-39, 175-81,
230).
 Trial By Dure 165
In the same fashion the purely intellectual derivation of the
Vidalian inheritance has been progressively identified and
categorised. V. Berdoulay claims to distinguish no less than eight
cercles d'affinit which contributed to Vidalian geography its
humanistic and social-scientific orientation (Berdoulay 1981a, 141-
81), this intermingling being set against the backdrop of a less tangible
waxing of neo-Kantian idealism as a reaction against the excesses
of positivism (Berdoulay 1976, 1981a, 109-39) while A. Buttimer
traces the legacy still further back to Montesquieu and the historians
of the French people, J. Michelet, E. Renan and H. Taine (Buttimer
1971, 9-26). Certain key names figure in every list: the conservative
Catholic circle of social reform and research founded by F. Le Play,
whose concern with social solidarity and the conditions of the work-
ing class struck a responsive chord; the regionalism in politics and
culture impelled by Vidal's fellow southerner, F. Mistral, and the
popular regional novels dealing with the themes of enracinement and
dracinement such as the Lorraine cycle of M. Barrs; the polymath
E. Levasseur, who anticipated the Vidalian esprit gographique but
lacked the personality and the position to disseminate a doctrine
(Broc 1976; : Buttimer 1971, 25; Claval 1968b, Nardy 1968; Rhein
1982, 234-5); finally, Vidal's near-contemporaries, F. Ratzel and E.
Durkheim, polarised protagonists of the geographical and sociologi-
cal perspectives on cultural phenomena, who defined the bounds of
the contextual debate within which Vidal's latest writings must be
understood (Andrews 1984; Berdoulay 1981a, 109-39; Buttimer 1971,
27-40; Rhein 1982, 238-51). Curiously, very few authors have drawn
attention to the rather obvious fact that Vidal was a professional
historian drawn into geography relatively late (Claval 1969, 19, 47).
What then was the nature of the Vidalian synthesis? What key
concepts can be identified and what reasons offered for their applica-
bility to archaeology towards the end of the twentieth century?
The 'possibilist' gloss of L. Febvre (1922) may be rejected at the
outset as an ideological distillation much more extreme than that of
Vidal himself (Berdoulay 1981a, 214-5; Claval 1969, 49-53; Luker-
mann 1965, 129). Even scholars impressed by the coherence of the
Vidalian epistemology recoil at its implicit and ambiguous formula-
tion and expression, which seems like mere woolly-mindedness: a
trait no doubt due to his classical, literary upbringing, but one which
particularly irritated the tidy-minded positivists of the Durkheimian
school (Allegra and Torre 1977, 221-2; Andrews 1984, Berdoulay
1978, 1981a, 212; Claval 1984, Ribeiro 1965); it is symptomatic that
the researches of scholars such as F. Lukermann (1965) and A.
Buttimer (1971, 1978) revolve around the discussion of potent mots-
clefs in the absence of authoritative definitions.
Secondly, as F. Lukermann (1965) and V. Berdoulay (1978; 1981a,
183-277, esp. 201-14) have established, the Vidalian opacity and
166 JAMES LEWTHWAITE
apparent failure to come to grips with causality are due to a phil-
osophical orientation opposed to positivism and determinism; a
philosophy of'conventionalism' and 'contingency' derived ultimately
from A. Cournot through the mathematicians E. Boutroux, H. Poin-
car and P. Duhem, and the philosopher-historians G. Tarde and H.
Berr. Briefly, the Vidalian concept of a holistic esprit gographique
unconfined to particular bodies of data, operating freely within a
three-dimensional universe of time, space and social organisation
(Claval 1984), so far from denying the existence of causal chains,
emphasised the multiple enchanement of causes impinging on any
given situational ensemble. Precisely because of the interdependence
of all global phenomena (principe de connexit, de l'unit terrestre)
an attempt to specify positivistically a hierarchy of co-determinant
causes for an individual event would require a recapitulation of the
history of the universe. Therefore, for F. Lukermann (1965), Vidalian
epistemology, based on the concept of the calcul des probabilits,
would restrict the application of even probabilistic laws to whole
classes of phenomena only, explanations of individual events owing
more to chance. V. Berdoulay (1978, 86, 1981a, 2-14) claims that H.
Berr's arguments in favour of the stability of certain individual
phenomena, in apparent violation of the probabilistic philosophy,
conditioned the formulation of the concept of a regional national
'personality'.
The utility of the concept of contingency is obvious, particularly
when combined with that of histoire-problme (cf. infra). Archaeology
oscillates between a self-categorisation as a historical (idiographic)
and as a 'scientific' (nomothetic) form of knowledge, but an esprit
archologique in principle open to the study of any phenomenon
defined by the dimensions of space, time and man cannot afford to
be either exhaustively descriptive, in view of the enormous complexity
of causal chains intersecting in the process of formation of but a
single stratum on a single site, or superficially 'scientific' in reducing
such complexity to law-like statements of inevitably tedious triviality
(cf. K. Flannery's dismissal of such 'Mickey Mouse laws' in prefer-
ence to 'Natural regularities').
Thirdly, every bio-social discipline early in this century formulated
a fundamental conceptual entity and/or hierarchy of entities for
research during the difficult transition from pure inductive
classification to more problem-orientated viewpoints: for archaeol-
ogists the 'culture', borrowed by Childe from the Vienna school of
cultural anthropology and defined as the material expression of a
'people' (in the Danube in Prehistory, 1929); for the phytosociologists
of the Sigmatist school, a whole hierarchy of latinisms; for the
Vidalian geographer, this was the genre de vie (Berdoulay 1978, 87-8;
Buttimer 1971, 43-58; 1978, 60-2; Vidal de la Blache 1911). But
 Trial By Dure 167
whereas the continued utility of the culture concept has come into
question within archaeology on account of the very difficulty of
applying the culture-people equation to prehistory without generating
an absurd pseudo-history, detailing the kaleidoscopic interminglings
of peripatetic peoples and indeed individuals, the quandary facing
the mid-century Vidalian geographers was quite simply that the world
itself had changed: a concept ideal for the study of the isolated pays,
the rural way of life handed down from generation to generation,
had simply run out of contemporary examples (Buttimer 1971, 178-
90; Claval 1969, 118).
In several respects, the genre de vie ('way of life' or 'lifestyle') is
superior to the 'culture'. The former is bounded neither in time nor
space in the manner of the culture, to be equated, that is to say, with
a 'people' occupying a specific territory and producing a material
expression which is necessarily finite, i.e. ending by definition with
the production of the last Bell-beaker or Hallstatt D metalwork. Thus
the Vidalian pays is given its specific personality by the interaction
of several genres de vie (Allegra and Torre 1977, 149-50), one of
which may be dominant (cf. 'modes of production' within the social
formation in Marxist doctrine) while a genre de vie defined in terms
of its milieu or means of livelihood may date from the point at which
human occupation of that niche, or the introduction of a new crop,
farming system or economic specialisation began, for as long as it
may persist. The concept of genre de vie classically conveys the notion
of an adaptation of the physical milieu which is neither totally
determined nor totally determinant, but gradually elaborated over
the centuries in the light of the body of beliefs and values peculiar
to its adherents' civilisation (Buttimer 1978, 60-2; Claval 1969, 51);
recognising, that is to say, both the materialistic and the subjective,
cognitive dimensions of human culture (Buttimer 1981, Ley 1980).
Although Vidal has been charged with sociological navet (But-
timer 1971, 52) in not recognising the existence of class interests
within or across genres de vie, this seems to be a question of actual
application rather than inherent incompatibility. Indeed, the historian
Henri Pirenne expressed the rise of the urban mercantile bourgeoisie
of the Middle Ages precisely through the progressive differentiation
of stratified genres de vie (Allegra and Torre 1977, 67). In any case,
the demonstration that some genres de vie are impracticable without
some form of role, if not status, differentiation (one thinks of certain
forms of pastoralism and cereal cultivation) may act as a long-overdue
response to the crasser intrusion into archaeology of a species of
vulgar Marxism owing more to subcultural fashion than theory.
The variable antithetically opposed in Vidal's eyes to the tranquil
persistence of the traditional genres de vie - circulation - is a term
which can refer to all forms of intensified spatial interaction, ranging
168 JAMES LEWTHWAITE
from commodity flow to the conquest of space and of the constraints
of the milieu through the revolution in transport and communication,
which led in the long run to the differentiation and specialisation of
the earth's surface (Buttimer 1971, 190-3).
Such a concept appears to be particularly applicable to the periods
in prehistory when the evidence for inter-regional contacts appears
temporarily to blur the anticipated pattern of neatly bounded cultures
without recourse to the currently unfashionable (if misrepresented)
concept of diffusion.
The accusation that Vidalian geography never rose beyond the
patient accumulation of regional monographs (cf. Claval 1969,57-60,
67-71) is patently at odds with the grand themes considered in
Principes de Gographie Humaine and Vidal's known passion for
reflection (Claval 1969, 47-9; Ribeiro 1965; Buttimer 1981). Rather,
it seems as if Vidal conformed only too closely to the Clark model
of the patron, in only revealing such aspects of the greater masterplan
(to approach probabilistic regularities through a comparison of
descriptive syntheses of manageable dimensions) to his pupils; to
each of whom, 'like God the Father', he allotted a region of the
projected Gographie Universelle, as he considered strictly necessary.
The inevitable result was that the grander design was lost to his
epigonoi, who tended in the long run rather towards sectorial studies
(Buttimer 1971, 58-120; cf. Rhein 1982, 242, 248).

THE ANNALIST TRADITION OF HISTORICAL SCHOLARSHIP

To define as a 'school' the three generations of historians associated
with the journal Annales: conomies, Socits, Civilisations (founded
in 1929 as the Annales d'Histoire conomique et Sociale) and to
discuss their shared 'paradigms', are generally agreed to be much
more difficult problems for the history and sociology of knowledge
than is the case with the Vidalian geographers; particularly in France
itself, where the well ensconced establishment of the New Historians
has been endowed with a status, if not beyond comprehension, at
least criticism, according to H. Coutau-Bgarie. The historiography
of the Annales tends to fall, in consequence, into three distinct
categories:
(i) More or less laudatory articles by insiders, such as official
histories and biographies of such key personalities as L. Febvre or
F. Braudel; personal reminiscences; programmatic statements by
leaders and overviews of the current state of the movement by anxious
disciples, which can hardly fail, even when striving to be modest,
frank and self-critical, to portray the rise of the Annales as a chanson
de geste of the new scientific history, as the victory of light and
learning over the uncomprehending hostility and overt malice of
obscurantist pedants. The virtues of this new learning are said to be
 Trial By Dure 169
its openness to innovation, especially from the quarter of the social
sciences (n.b. sciences de l'homme); its quest for total history, 'the
sum of all possible histories' and its readiness, in furtherance of this
aim, to pose problems (histoire-problme) and to quantify and seriate
data. Its hallmark is the massive enqute, the research quipe backed
by the computational and cartographic resources capable of tackling
the most ambitious projects. Its antithesis is the narrow learning
advocated by the turn-of-the-century positivists C. Langlois and C.
Seignobos, with their cult of objectivity, their fetish for textual criti-
cism, their myopic compartmentalisation and concern only for the
short-term, their erudition wasted on the recitation of politico-diplo-
matic events (histoire-rcit, histoire vnementielle) (Aymard 1972;
Braudel 1958, 1972a, 1976; Burguire 1979, Chattier and Revel 1979,
Furet 1975, Stoianovich 1976).
(ii) For obvious reasons, the contributions of more or less admiring
outsiders are much less homogenous. Since the appearance in 1972
of a second, much revised edition in English of Braudel's magisterial
survey of the sixteenth-century Mediterranean world (Braudel
1972b), much more attention has been devoted to the post-war
Annaliste scholarship, to Braudel's leadership and the 'paradigmatic'
(in another sense) role and status of his first major work. Although
the tone of such writings is frequently fulsome and rarely less than
objectively positive in respect of the movement's place in twentieth-
century historiography, much more attention is paid than in the case
of the former category to the identification of the overlooked contri-
bution of forerunners and contemporaries to the Annaliste synthesis,
to the identification of contradictions within the tradition between
Bloch, Febvre and Braudel and to the apparent loss of direction and
coherence since 1968. Particular emphasis has been devoted to
Braudel's concept of history as a plurality and of dialectic between
time-scales, each roughly equated with a sub-discipline of history
proper, to his fondness for the longue dure and for such literary
traits as verbosity, fondness for lists and digressions, personalisation
of inanimates, cavalier attitude to statistics and superb evocation of
place and period. In general, critics of this category have come to
regard Braudel's Mediterranean study as a magnificent but funda-
mentally flawed undertaking, which a successor generation has
admired but hardly emulated, evoking the image of a three-storied
palace, each floor of which consists of a maze of endless book-lined
corridors leading hither and thither, of a sprawl of rooms sumptuously
if fussily furnished, but which the architect has neglected to provide
with an adequate stairway (Allegra and Torre 1977, Cedronio 1977,
Hexter 1972, Kinser 1981, Mann 1971, Prince 1975, Santamaria and
Bailey 1984, Trevor-Roper 1972).
(iii) In sharp contrast, H. Coutau-Bgarie (1983) has etched an
170 JAMES LEWTHWAITE
acidly cynical sketch of the Annalistes, outlining the alleged
manoeuvre whereby the radical freemasonry of the '20s and '30s
came to be transformed into a pagan travesty of the Catholic church
establishment by the '60s and '70s (Coutau-Bgarie 1983, 284). The
Annalistes are prosecuted for intellectual complacency and corrup-
tion on two major charges: firstly, on the grounds that their success
owes more to the skill with which their successive leaders have
planned and practised a strategy of the seizure of power than to pure
academic superiority over rival schools (Coutau-Bgarie 1983, 16-33,
247, 250-82, esp. 254, 262; cf. Braudel 1976, 15-16; Prince 1975;
Revel 1979,1370); secondly, on the count that this sense of intellectual
superiority is in fact singularly ill-founded.
Coutau-Bgarie's iconoclastic 'conspiracy model' may be sum-
marised, appropriately enough, as a strategy operating on three levels
simultaneously. Firstly, an internal strategy, designed to eliminate
rival schools and paradigms through the adoption of a pseudo-
scientific gloss (history as puzzle-posing and puzzle-solving rather
than narration, appeals to innovations from the social sciences) held
to be neither original (anticipated in Germany by K. Lamprecht, W.
Dilthey and the Methodenstreit and in France itself by G. Monod)
nor fair (to Langlois and Seignobos, themselves proponents of a
history scientific in comparison with the foregoing romanticists) nor
unique (paralleled in Britain by the founding of the Economic History
Review by Tawney and Power), still less sincere; and, above all, a
pedagogic disaster (Coutau-Bgarie 1983,43-90). Secondly, an exter-
nal strategy, aimed at disarming and dominating the juvenile social
sciences through propaganda on behalf of a total history based on
quantified, serial data distinguished in quality by its unique property
of dure (Coutau-Bgarie 1983, 91-126) of which this critic notes
that Annaliste history is more totalitarian thati total, excluding not
only personalities and paradigms but much non-French scholarship,
most ancient and contemporary history and many historical genres
(e.g. political, diplomatic, military) out of obstinately-conditioned
reflexes (Coutau-Bgarie 1983, 169-98). Thirdly, a deeper, structural
strategy in tune with the prevalent ideology of the twentieth century,
permeated by 'economism', attentive to the history of the anonymous
masses, 'Collective Man', akin to Marxist or Toynbeean historicism
but with the Ideological core removed and the causality left pragmati-
cally open (Coutau-Bgarie 1983, 198-243). Coutau-Bgarie pursues
his thesis of the determination of ideology by strategy in arguing that
the abandonment of narrative and the event in favour of repetitive,
quantifiable data is to be explained principally through the need to
adapt the practice of historiography to that of the social sciences in
order the better to dominate them (Coutau-Bgarie 1983, 61) and
that L. Febvre's (1922) defence of Vidalian geography against the
 Trial By Dure 171
attacks of Durkheimian positivists such as F. Simiand was due less
to reverence for the teaching of the matre than to a cynical tactic
of killing geography with kindness, by assigning it a modest and
introverted niche which posed no threat to the imperialist pretensions
of history (Coutau-Bgarie 1983, 71-2; cf. Claval 1969, 124, 148).
What defence can be offered against this diatribe? Perhaps a
threefold response is adequate:
(i) Some of the minor cavils are shared by other scholars and must
be admitted, such as the injustice of the castigation of Langlois' and
Seignobos' histoire historisante out of its context (Coutau-Bgarie
1983, 296-7; cf. Allegra and Torre 1977, 123-8; Chattier and Revel
1979; Cedronio 1977, 3-18); the failure to acknowledge Monod's
priority as a critic of 'eventful' narration and advocate of social and
economic history (Allegra and Torre 1977, 50-3) influencing Febvre
in particular (Mann 1971, 25-7); the reluctance to respect the
influence of the German 'new history', even the evident modelling
of the first series of the Annales on the Vierteljahreshefte fr Sozial-
und Wirtschaftsgeschichte (Coutau-Bgarie 1983, 18-19; cf. Chartier
and Revel 1979, 434-7; Mann 1971, 30-1, 62-3; Stoianovich 1976,
33, 41). But are these not errors of detail rather than substance?
(ii) Much is mere value-judgement, particularly such assertions
as that the focus on problems rather than narrative has led to a loss
of content or that an over-emphasis on mass action and impersonal
forces has dehumanized history, while only time will verify the
allegation that the contemporary clatement of Annales scholarship
will engender a dissolution into incoherence (cf. Cedronio 1977, 48,
66-72; Stoianovich 1976, 204-31). Archaeology, too, has had its
outbursts of gobbledygook-bashing in the name of humanism!
(iii) In the final analysis, the precise motivations and level of
political adroitness of Febvre, Bloch and Braudel are not at all
relevant to the question posed here, that of the utility of their ideas
as ideas to archaeology; neither are their literary qualities, nor even
the degrees to which their practice (se. La Mditerrane) lives up to
the promise of their paradigm.
What this controversy does highlight is the extent to which it is
impossible to understand the ides fixes of the Annaliste school
without reflecting upon the milieu in which they coagulated: i.e. the
period between 1896/7 or 1898 (appearances of the Anne
Sociologique and the Langlois-Seignobos manual of historical method
respectively) and 1929, or even 1947, date of the re-establishment of
the Annales and the foundation of that citadel of their post-war
fortunes, the Vie section of the EPHE: a Kuhnian pre-paradigmatic
chaos during which history in France was beset on several fronts.
Firstly, the violent onslaught of the Durkheimians (particularly the
economist F. Simiand) on its unscientific (sc. non-positivist) epis-
temology; secondly, the no less wounding sarcasms of P. Valry
172 JAMES LEWTHWAITE
(Stoianovich 1976, 65); thirdly, the countervailing current against
positivism inspired not only by the Cournot-Poincare conventionalist
philosophy of science but the spiritualism of the charismatic Bergson
(Allegra and Torre 1977, 182, 226-9; Cedronio 1977, 8).
Of the various streams which flowed into the Annaliste synthesis
these are invariably mentioned: the human geography of Vidal, the
social morphology of the Durkheimian school and the much less
coherent inspiration of the circle which formed round Henri Berr,
finding its outlet from 1900 in his Revue de Synthse Historique and
later in his Semaines and publishing ventures (Braudel 1972a, Char-
tier and Revel 1979; Coutau-Bgarie 1983, 68-90; Mann 1971, 25),
out of the last of which the Annales school emerged from its bastion
in the lite (showcase) university of Strasbourg (Allegra and Torre
1977, 190-1; Chattier and Revel 1979, 434-7; Stoianovich 1976, 50).
More minor rivulets were the contributions of F. de Saussure and
especially of his pupil A. Meillet in linguistics (Allegra and Torre
1977,111-12, 238-41; Cedronio 1977, 31; Coutau-Bgarie 1983, 84-5;
Mann 1971, 76, 131-7), of ethnologists such as L. Lvy-Bruhl (Allegra
and Torre 1977, 229-38; Chartier and Revel 1979, 437-41; Coutau-
Bgarie 1983, 79-84) and of psychologists such as H. Wallon (Allegra
and Torre 1977, 229-38; Coutau-Bgarie 1983, 85-7); literary
allusions shared with the 'generation of'98' (Unamuno e.g.) are most
probably coincidental (Chartier and Revel 1979; Mann 1971, 57,
157-64; Stoianovich 1976, 64). Insofar as the founding fathers looked
to history for inspiration it was to 'marginals' (Aymard 1972, 492-7);
to the older generation of populist historians such as J. Michelet
(Cedronio 1977, 8-9; Mann 1971, 19-20, 25, 73), later to contem-
poraries such as G. Roupnel (Kinser 1981, 71-2); to progressive
foreign scholars such as H. Pirenne and J. Huizinga (Allegra and
Torre 1977, 65-70, 243-74; Cedronio 1977, 29) or to historicising
economists and economic historians such as F. Simiand (Allegra and
Torre 1977, 65-70, 243-74; Chartier and Revel 1979, 432), H. Hauser
(Allegra and Torre 1977,65-70,168,282-4; Cedronio 1977,29; Mann
1971, 19, 26), and, a major influence on the post-war orientation of
the Annales under Braudel, F. Labrousse (Allegra and Torre 1977,
328-31; Coutau-Bgarie 1983, 132-4; Stoianovich 1976, 194-203);
on the other hand, M. Weber had little impact (Allegra and Torre
1977, 113-8, 128; Cedronio 1977, 56-8; Coutau-Bgarie 1983, 96;
Stoianovich 1976, 200-1) while Marx and Marxism were regarded
with a cordial and judicious diffidence by these adherents of the
socialism of J. Jaurs (Aymard 1972, 496; Coutau-Bgarie 1983,
225-43); Stoianovich 1976,68-9,111-13,134-53; Revel 1979,1366-7.
In effect, so far from being heretical outcasts (Braudel 1976, 14)
Febvre and Bloch pursued a strategy of a voluntary intellectual
isolation or exile, whence they could all the more easily indulge in
 Trial By Dure 173
polemical combats and deal out savagely critical reviews (Burguire
1979, 1350, 1352; Coutau-Bgarie 1983, 251-3 esp. 251 n. 13).
This is surely the essential paradox which explains the genesis of
the Annales: that while their 'internal' strategy, in the sense of
Coutau-Bgarie (i.e. to achieve a hegemony within French his-
toriography) took the form of a deliberate 'externalisation', their
'external' strategy in his sense (viz. not only to resist the appeal of
Vidalian geography and the imperialism of the Durkheimian sociol-
ogists, but in the long run to devour both) must be understood in
terms of the essentially 'internal' nature of the polemic - for these
were not three warring tribes but three tendencies among contem-
poraries drawn from similar backgrounds, attending the same classes,
engaging in the same street-fights as Dreyfusards, drawn into the
socialism of Jaurs by the same Lucien Herr, above all permeated
by the same intensely critical esprit normalien of the rue d'Ulm
(Andrews 1984; Chartier and Revel 1979, 430-2; Clark 1973 passim).
(i) The Vidalian 'tendency' instilled a spiritual emphasis on human
creativity and intelligence; an evocation of the individuality and
organic unity of the myriad pays of rural France and of the repetitive
rhythms of the peasants' way of life which seemed impervious to the
superficial drama of battles and treaties in which the historians of
the Sorbonne chiefly gloried (Claval 1969, 57-64, esp. 63), but above
all in the identification of the 'substance' of history as neither the
actions of great men nor the collisions of impersonal masses and
forces but the intimate experiences of cohesive rural societies within
a regional framework, even as their ingrained solidarity was being
undermined by the industrial revolution, by urbanism and by the
abrasion of circulation (Allegra and Torre 1977, 118-19, 136-50,
215-19; Burguire 1979, 1351; Cedronio 1977, 6-7; Chartier and
Revel 1979,429-30; Claval 1969,49-50; Coutau-Bgarie 1983, 68-75;
Mann 1971, 29-30, 73-4; Stoianovich 1976, 76-8, 85). In terms of
personalities, it was not so much Vidal himself as J. Sion and A.
Demangeon who had the greatest influence on the future Annales
circle through their mediating role as members of both the Vidalian
and the Durkheimian circles and, in the case of Demangeon, as an
actual collaborator who introduced the circle to the publisher A.
Colin (Chartier and Revel 1979, 434-7).
(ii) The Durkheimian circle seems to have provided the founders
of the Annales with a model (Allegra and Torre 1977, 79-88, 219-23;
Braudel 1972,456; Burguire 1979,1352; Cedronio 1977,12; Chartier
and Revel 1979, 430-2; Coutau-Bgarie 1983, 25-7, 75-7, 251 n. 13;
Mann 1971, 18, 26, 82-92): that a group of scholars linked, no longer
by mere mutual self-advancement, but by a common ideology and
a collective methodology, whose goal was nothing less than a major
institutional coup (the establishment of a discipline dedicated to the
174 JAMES LEWTHWAITE
scientific study of society within the university system) whose articula-
tion was achieved through collaboration in a journal (the Anne
Sociologique) and whose patron was not only ambitious but articulate
and, to a degree, charismatic. As previously stated, it was rather
disciples such as M. Mauss, M. Halbwachs and F. Simiand who
interacted with the inchoate new history movement, since the
authoritarianism and ambition of the austere and dogmatic Durkheim
alienated the latter as much as their traditionalist opponents in the
Sorbonne (Allegra and Torre 1977, 111; Mann 1971, 57); after all
the very existence of history as an academic discipline was being
threatened. Since, however, the Durkheimians were largely excluded
from the system by the structural inertia of the lyce-agrgation nexus,
their critique was confined to the ideological sphere, so that it was
not too difficult for the future Annalists to deflect their blows and
indeed turn them to advantage by undertaking the necessary ideologi-
cal aggiornamento while maintaining their grip on the institutional
base provided by history; in effect, the sociologists merely provided
one historical school with a stick with which to beat the rest, while
themselves gaining only belated admittance to the universities and
only as individuals. By the 1920's circumstances favoured the his-
torians: the Vidalian geographers were past their peak and losing
coherence, the Durkheimians had suffered the losses of their leader
and key personnel while the economists were confined to another
faculty altogether (Chartier and Revel 1979, 434-7; Coutau-Bgarie
1983, 67, 123; Revel 1979, 1367-8).
(iii) Neither Febvre nor Bloch was the first to form a circle intent
on the rejuvenation of history through criticism and the appropriation
of what was most useful from the sociological armoury; rather, they
were successively recruited to the circle founded by Henri Berr which
found its outlet in the Revue de Synthse Historique (Allegra and
Torre 1977,133-5,191-242; Aymard 1972,495; Braudel 1972a, 458-9;
Burguire 1979, 1351; Cedronio 1977, 16; Chartier and Revel 1979,
434-43; Coutau-Bgarie 1983, 7; Kinser 1981, 73 n. 18; Mann 1971,
74-82). Berr's philosophical syncretism, inspired by the humanism
and vitalism of Bergson, provided an inadequate paradigm for a new
history, however, while Berr proved to be a less than effective patron,
and his group began to lose coherence long before the founding of
the Annales; nevertheless it provided Febvre and Bloch with an
entre to a stimulating circle at a critical moment in their careers
(the very term histoire vnementielle being coined by another mem-
ber, P. Lacombe). The foundation of the Annales in 1929 represented
the fulfilment of a latent potential within a multi-generational circle
which Berr himself did not realise.
To treat the Annales phenomenon as a single, coherent entity is,
however, to misunderstand one of the most significant aspects of the
 Trial By Dure 175
creativity of the approach which is only belatedly receiving attention:
that of the dynamic tension between the founders (what might be
called the 'Charles Atlas principle') which can be identified in a
number of dimensions (Aymard 1972, 497-502; Cedronio 1977, 18-
30; Coutau-Bgarie 1983, 258 n. 25; Mann 1971, 41-56) and which
can be seen as underlying changes of orientation and policy in the
second and third generations. Thus Bloch appears to have been the
more thoroughly imbued with the spirit of the social sciences - a
sociologist of the past - advocating and practising a history of
institutions, economies, societies and landscape over the long-term,
drawing on masses of social data, abstracting ideal-types in order to
make comparisons, achieving universal recognition through the
publication in 1931 of Les Caractres Originaux de Histoire Rurale
Franaise, yet exposing himself to the very charges of dehumanising
history and depriving it of dynamism once laid at the door of the
Sorbonnistes (Allegra and Torre 1977, 198-300; Cedronio 1977, 8,
11, 33). Febvre, more under the influence of Bergson and Berr, was
more inclined towards Pascal's esprit de finesse, more drawn to the
historiography of personalities such as Luther, Rabelais - and the
Franche-Comt - of mentalities, sensibilities and the irrational in
general; of the Zusammenhang of a civilisation; in short, towards a
humanistic rather than a social-scientific conception of history
(Allegra and Torre 1977, 223-9; Chattier and Revel 1979, 437-41;
Mann 1971, 57-122 esp. 29-39). Febvre was, however, burdened with
involvements in Berr's collective publication projects, and by doubts
and dualism throughout the '30s; and his distillation of the concept
of outillage mental lagged behind Bloch's rapid mastery of agrarian
history (Allegra and Torre 1977, 312-3). The oscillation between the
Bloch and Febvre poles can be detected in the orientation and even
the title of the Annales; with Bloch outshining Febvre throughout
the '30s, his influence waning with the defeat of 1940 and afortiore
his death in 1944; Febvre's triumph, however, was ephemeral, as
Braudel, apparently impressed by the solid progress of the Labrousse
tendency towards economism and the identification of cycles in serial
data, reinstated the Bloch pole (Cedronio 1977, 57-9) until the final
florescence of the histoire des mentalits in the '60s and '70s in which
the ghost of Febvre was reincarnated in Aris, Dumzil and Foucault.
For all their enduring fame, Febvre and Bloch have had far less
impact on recent global historiography than their successor as editor
of the Annales and godfather of the movement, Braudel. More to
the point, given the supposed problematic of this essay, is the hal-
lowed status which prehistorians now accord Braudel's Mditerrane,
a reference to this being now regarded, it seems, as de rigueur in any
study aspiring to intellectual seriousness and credibility. The reason
for this unanimity is the obvious relevance to the archaeologist of
176 JAMES LEWTHWAITE
the problem confronted by Braudel - how to describe historical
processes occurring at very different 'wavelengths' - and of his
solution, the tripartite division into mere surface froth and foam
(courte dure - vnements = politico-diplomatic history), the more
regular beat of the waves (moyenne dure = conjonctures = the cycles
of economic history) and, dearest to Braudel, the deep rhythms so
slow that to the observer they may appear motionless (longue dure =
structure = geohistory) (Braudel 1972b). It is not necessary to repeat
at length the criticisms of historians, even those sympathetic to the
Annaliste paradigm. They include the artificiality of the tripartition,
the arbitrariness of the attribution of whole categories of the historical
process to specific time-scales, the profound underlying pessimism,
the 'prison' image of the structure of history, plausibly explained as
the projection of the circumstances (a POW camp) in which it was
written (Allegra and Torre 1977, 333-45; Cedronio 1977, 40-59;
Hexter 1972, 509-10, 513-39; Kinser 1981; Prince 1975; Santamaria
and Bailey 1984). But two other points must be made.
(i) Braudel's perspective is in many ways the antithesis of that of
the founders of the Annales. The fetish of circulation, inherited from
Simiand through Labrousse, has elbowed out of the middle ground
the rural rhythms of the way of life of small peasant communities,
imprisoning some historically contingent practices in the structure
of the longue dure, while the humanism and possibilism of Febvre,
which can be traced back to Vidal, appears to have been lost without
trace. Febvre, after all, believed that 'l'homme est prisonnier non pas
de la nature, mais de l'humanit' (Mann 1971, 35).
(ii) It is the duty of archaeologists to test and if necessary to refute
such an immobile and over-generalized account of the longue dure
of the human occupancy of the Mediterranean, since La Mditerrane
is no mere reference work but a 'controlling model' in its own right,
the conditioning influence of which surpasses mere disciplinary
frontiers.

APPLICATION TO THE ARCHAEOLOGY OF SETTLEMENT

ON CORSICA AND S A R D I N I A
It is clearly impossible, in an essay of this length, to apply the
historico-geographical concepts outlined above systematically and
with primary data, with the aim of demonstrating utility in suggesting
both solutions to existing problems and points of departure for a
further generation. The object of this section is much more modest:
to consider ways in which the best features of the Vidalian and
Annaliste perspectives may be reconciled, or, better, reintegrated.
The apparent need to choose between a totality- and a problem-
orientation may be denied at the outset as a luxury which is inap-
propriate to the qualitatively inferior database at the disposal of the
 Trial By Dure 177
archaeologist. As long as whole periods, provinces and classes of
data are, and are likely to remain, black holes as far as our knowledge
of prehistory is concerned, the familier dichotomies of syn-
thesis/analysis and nomothesis/idiography must be rejected as
irrelevant. Progress is most likely to be achieved in the long run
through a continuing dialectic between the complementary poles of
synthesis, theory-building and the various forms of problem-oriented
research, such as field and literature survey projects. For purely
practical reasons, including the need to consolidate recurrent costs,
acquire linguistic skills and contextual knowledge, negotiate with
bureaucracies and develop bilateral collaboration with local scholars,
the regional rather than the pan-Mediterranean perspective is likely
to continue as an important medium-term strategy. Therefore the
applications sketched here will involve a region, one with which the
writer is most familiar, that of Corsica and Sardinia.
The crux of the problem must be that of the relationship to be
established between the Vidalian genre de vie and the Annaliste (more
specifically neo-Annaliste or Braudelian) dure(s). The theoretical
flexibility of the genre de vie has already been noted: adaptation to
the purposes and perspective of the pre-historian, as for the historian
of longue dure, revolves around a few key problems: definition,
dynamism, differentiation, demonstration.
A definition of genre de vie which is either extremely broad or
excessively narrow contrives a (pre)history on the one hand practi-
cally static, on the other fragmented and lacking in comparability.
Thus, to choose from the genres de vie of Corsica distinguished by
Vidalian geographer R. Blanchard (1914), the transhumant pastoral
genre of the mid-montane zone appears reasonable enough within a
time-depth of ca. 50 years, but if extrapolated, as some prehistorians
and others would allow, to the terminus a quo of evidence for the
husbandry of domestic animals in Corsica (Lewthwaite 1983a,
1985a,b, 1986; Ravis-Giordani 1975; Vigne 1983, 1984; Weiss
and Lanfranchi 1976) would imply a continuity over nearly eight
millennia (perhaps nine with calibration). In effect, the lack of
expression of such undifferentiated labels needlessly prolongs the
mutually conditioning circular reasoning which has obtained among
archaeologists, historians and geographers in Corsica, as more gen-
erally in the Mediterranean; appearing to confirm a geographical
determinism all the more ironic since Blanchard emphasises the
historical contingency of this way of life and the very rapidity of its
decline (or transformation) under the impact of circulation (Blan-
chard 1914; Lewthwaite 1981, 1985b). As an example of the opposite
tendency, one might point to the specificity of the way of life of the
Castagniccia, synonymous with the culture of the chestnut, which
reached its developmental zenith at the end of the nineteenth century
178 JAMES LEWTHWAITE
from an origin of unknown antiquity, but one at any rate unlikely
to predate the Early Middle Ages (Lcke 1976). How readily can
this Corsican regional form be subsumed within a wider civilt
del castagno (Chestnut Culture) common to Apennine Italy, the
Cvennes, Prigord, the Limousin or the Tras-os-montes (Bruneton-
Governatori 1984; Cherubini 1981)?
Little, moreover, can be admitted by way of analogy with another
form of man-tree relationship with roots deep in the prehistory of
the Mediterranean, that consisting of the management and exploita-
tion of evergreen oak forest, insofar as this characteristic element of
the Mesogean landscape, so far from being the postglacial lowland
climax as formerly believed, is now known to have replaced mar-
cescent and deciduous oaks with increasing effect from the third and
second millennia BC onwards, pari passu with the expansion of
human settlement and of extensive pastoralism in particular (Lew-
thwaite 1982). This symbiosis has proceeded so far in Iberia that a
holm oak variety has evolved which produces a 'sweet' (tannin-
depleted) acorn more suitable as a crop (R. J. Harrison, personal
communication); a striking example of the Vidalian simile of the
mutual adaptation of genre de vie and milieu as that of a snail and
its shell. It seems obvious, therefore, that the definition of genres de
vie for synthesis must be either hierarchical (lacking flexibility?) or
polythetic (with an attendant loss of clarity and precision?).
The integration into the concept of genre de vie of much more
dynamism, duration and differentiation must be approached as a
single complex problem rather than severally and successively; that
is to say, the differentiation must be capable of expressing variation
in time, space and socio-economic complexity in a way which the
classifications of Blanchard (1914) and Le Lannou (1941), organised
as they were on the bases of region and altitudinal (bioclimatic) life
zones, do not. The expression of synchronie socio-economic
differentiation alone, even within a single grossly defined genre such
as pastoralism, requires discrimination of widely different scales of
enterprise, of mobility and of integration into wider tax-exacting or
market economies, quite apart from the problem of linking the various
forms so defined with other genres (agriculture, arboriculture etc.).
For the moment, the simple typology offered by Zbl (1982) offers
a point of departure.
Next, it must be acknowledged that differentiation took place over
time, a point to which early Vidalians paid much attention. C. Vallaux
argued that the very progress of a culture might be measured in terms
of the differentiation of genres (Buttimer 1971, 79) while P.
Deffontaines, adopting the more concrete language of J. Brunhes,
has written of the increase of horizons de travail (Dickinson 1973,
 Trial By Dure 179
242; cf. Deffontaines 1932). To represent the differentiation of prehis-
toric and historic genres purely as a cumulative unilinear progression,
a ramification of the more complex from the more simple, culminating
in the overlapping, interpenetrating pastoral lifestyles of the last
century, would however be as erroneous as that of transposing the
most centrifugal forms of recent transhumant pastoralism (those of
the Niulinchi of Corsica and Barbaricini of Sardinia) from the
moyenne dure to which they properly belong to the longue dure to
which most historians, geographers and archaeologists assign them
(Lewthwaite 1981, 1985b).
A further problem which the static classification by regions and
life zones obscures is that of representation of settlement as a variable
affecting the definition of genre de vie, particularly that of long-term
patterns in the occupation of space; or, to put it differently, of the
patterning of settlement mobility in dimensions less familiar than
those of altitude or centrality (i.e. upland/lowland, centre/periph-
ery). An analogy can obviously be made between the discovery by
the historian J. Day (1975, 1976a, b) that the medieval Sardinian
village was a peculiarly labile phenomenon, extraordinarily liable to
depopulation and dissolution even in regions of relative fertility, and
what appears to be evidence of a similar discontinuity of occupation
of supposedly long-lived and populous middle-late Neolithic sites
in the same regions (Michels et al. 1984). Are these patterns to be
seen as recurrent conjonctures or as manifestation of a chronic 'struc-
tural' maladaptation?
Such a diachronic differentiation of individual genres raises a
further problem. The terminology of Vidalian geography lacks a
high-order category capable of expressing the multisecular coherent
equilibrium of several co-existing genres de vie, since the concept
civilisation properly applies only to the cognitive, non-material
sphere. The geographer P. Claval has himself pointed out the utility
of the Annaliste structure (Claval 1969, 60-4). As a partial model of
such an application one may turn to Casanova's (1980) study of the
relationship between the social history and the collective representa-
tions (legends) of feudal Corsica as related by the chronicler Giovanni
della Grossa. It is partial because Casanova's ideologically deter-
mined interest in class conflict effectively marginalises the highly
relevant variables of subsistence (genres de vie!) settlement (par-
ticularly the phenomenon of incastellamento; cf. Cancellieri 1975)
and, indeed, of the structure and functioning of the village com-
munities from whose better-off elements (the coqs de village) the
feudal class was derived. At a final level of integration, the Vidalian
coverage of the personnalit of a region, and the Braudelian longue
dure, converge in directing the attention of archaeologists towards
the testing of certain familiar historico-geographic generalisations
180 JAMES LEWTHWAITE
against the fullest body of data in existence, that of prehistory. Thus,
the persistent myth that Corsica and Sardinia are isolated by virtue
of being islands may be challenged not only on the historical record
(Tangheroni 1977) but through the archaeological facts of the great
antiquity of their settlement (Cherry 1981, 1984) and their participa-
tion in such wide-ranging cultural phenomena as that of the
Impressed Ware sphere of the Early Neolithic (Lewthwaite 1985a,
1986) and of the sphere of a precocious Copper-Bronze Age 'incastel-
lamento' (Lewthwaite 1983b, 1985c, 1987). The impression that these
Tyrrhenian islands repeatedly functioned as nodes in the mediation
of cultural innovations between the eastern and western Mediter-
ranean is supported by preliminary calculations of their relative
accessibility and centrality within the western basin (R. Chapman,
personal communication; cf. Chapman, 1987).
The weakest plank in this platform may seem to be that of the
inability to provide a prompt and precise response to the question:
how do you identify a genre de vie using archaeological data? The
answer is, through prolonged reflection and analysis. But in this
respect does the historico-geographical hierarchy of entities (genre
de vie, structure, personnalit) differ in any way from that once
advocated by the late D. L. Clarke (culture, culture group, techno-
complex) since both are in practice deductive concepts inseparable
from a particular theoretical perspective? They are projected through
a more or less lengthy process of approximation to real-world data
prone to all the biases, errors and interpretative stereotypes which
came to preoccupy Clarke at a later stage of his development as a
scholar (Clarke 1968, 1978). It is at least clear that artefactual
evidence is inadequate for the task. A warrior pastoral genre de vie
is not to be equated with assemblages of supposed 'milkboilers'
and/or weaponry, but with a comprehensive analysis of settlement
patterns, palaeoecological and bioarchaeological data (Lewthwaite
1984; cf. Castaldi 1982) and above all, through the application of an
as yet inchoate 'middle range theory' relating pastoral activity to
such archaeologically visible manifestations, a theory which must be
created through the systematic liaison of archaeology proper with
ethnoarchaeology and historical rural ethnography within regions
such as the Mediterranean (cf. Moreno 1982).

From the viewpoint of the 1980's archaeology appears to be recover-

ing from a mid-life crisis of introspection inseparable from the
profound transformation of contemporary Western society as a
whole, during which archaeologists have repeatedly been tempted
to emulate the apparent vigour, elegance and objectivity of the natural
 Trial By Dure 181
and the 'relevance' of the social sciences. For this reason alone, a
comparison with the crisis undergone earlier this century by the
histori co-geographical sciences is nothing if not pertinent. The assimi-
lation of archaeology to a historico-geographic form of knowledge
is advocated on these grounds:
(i) Archaeology may thus be enabled to develop from being a
science of things (lake villages, shellmiddens, tells, barrows, and their
contents and contexts) into a science of the co-evolution of man and
milieu, which while accomplished through the patient reflection on
regionally differentiated lifestyles, structures and 'personalities', may
aspire to the identification of regularities in man's occupancy of the
earth (Vidal de la Blache 1922);
(ii) These regularities must be appropriate to the historical rather
than experimental-observational quality of the data, that is to say
not merely subject to contingency but to a context of asymmetrical
temporality ('time's arrow') which precludes the formulation of either
the 'timeless' theorems of the mathematician or the statistical norms
of the 'eternal present' of the social scientist (Braudel 1958).

Acknowledgements. The writing of this article would have been

impossible without the speed and efficiency with which the inter-
library loans staff of Bradford and Lancaster University Libraries
and of the British Lending Library responded to requests for relatively
obscure (one might say, recherch) sources. My sincerest thanks are
due to Mrs. Janet Atkins for her skill and care in deciphering and
typing up a manuscript of inordinate complexity, and to the editors,
particularly Dr. J. L. Bintliff, for their patience.
The cost of attending the conference in Oxford was in part
defrayed, thanks to the generosity of Mr. A. Aspinall, through the
Undergraduate School of Studies in Archaeological Sciences of the
University of Bradford.

BIBLIOGRAPHY
Allegra, L. & A. Torre (1977) La Nascita dlia Storia Sociale in
Francia dalla Comune all "Annales". Turin, L. Einaudi.
Andrews, H. F. (1984) The Durkheimians and human geography;
come contextual problems in the sociology of knowledge.
Transactions of the Institute of British Geographers n.s. 9, 315-
36.
Aymard, M. (1972) The Annales and French historiography
(1929-72). Journal of European Economic History 1, 491-511.
Berdoulay, V. (1976) French possibilism as a form of neo-Kantian
philosophy. Proceedings of the Association of American
Geographers 8, 176-9.
182 JAMES LEWTHWAITE
Berdoulay, V. (1978) The Vidal-Durkheim debate, 71-91 in
Humanistic Geography. Prospects and Problems, (Eds.)
D. Ley and M. Samuels. Chicago, Maaroufa and London,
Croom Helm.
Berdoulay, V. (1981a) La Formation de l'Ecole Franaise de
Gographie (1870-1914) ( = Ministre des Universits Comit
des Travaux Historiques et Scientifiques, Mmoire de la
Section de Gographie II). Paris, Bibliothque Nationale.
Berdoulay, V. (1981b) The contextual approach, 8-16 in
Geography, Ideology and Social Change, (Ed.) D. R. Stoddart,
Oxford, Basil Blackwell.
Blanchard, R. (1914) Les genres de vie en Corse et leur volution,
Recueil des Travaux de Institut de Gographie Alpine de
Grenoble 2(2), 187-238.
Braudel, F. (1958) La longue dure. Annales: Economies, Socits,
Civilisations 13, 725-53.
Braudel, F. (1972a) Personal testimony. Journal of Modem History
44, 448-67.
Braudel, F. (1972b) The Mediterranean and the Mediterranean
World in th Age of Philip II. (Engl. tr. Fr. orig. of 1949),
2 vols. London, Collins.
Braudel, F. (1976) Foreword, 9-17, to French Historical Method.
The Annales Paradigm, (T. Stoianovich). Ithaca and London,
Cornell University Press.
Broc, . (1974) L'tablissement de la gographie en France:
diffusion, institutions, projects (1870-90), Annales de
Gographie 83, 545-68.
Broc, N. (1976) La pense gographique en France au XIXe
sicle: continuit ou rupture? Revue de Gographie des
Pyrnes et du Sud-Ouest 47(3), 225-47.
Bruneton-Governatori, A. (1984) Le Pain de Bois. Ethnohistoire de
la Chtaigne et du Chtaignier. Toulouse, Ech.
Burguire, A. (1979) Histoire d'une histoire: la naissance des
Annales. Annales: Economies, Socits, Civilisations 34(6),
1347-59.
Buttimer, A. (1971) Society and Milieu in the French Geographic
Tradition ( = Monograph of the Association of American
Geographers 6). Chicago, Rand McNally.
Buttimer, A. (1978) Charism and Context: the challenge of la
gographie humaine, 58-76 in Humanistic Geography. Prospects
and Problems, (Eds.) D. Ley and M. Samuels, Chicago,
Maaroufa and London, Croom Helm.
Buttimer, A. (1981) On people, paradigms and 'progress' in
geography, 81-98 in Geography, Ideology and Social Concern,
(Ed.) D. R. Stoddart. Oxford, Basil Blackwell.
Cancellieri, J. A. (1975) Formes rurales de la colonisation Gnoise
en Corse au X l I I e Sicle: un essai de typologie. Mlanges de
cole Franaise de Rome. Moyen ge-Temps Modernes
93(1), 89-146.
Casanova, A. (1980) Rvolution fodale, pense paysanne et
caractres originaux de l'histoire sociale de la Corse. tudes
Corses 15,71-91.
Castaldi, E. (1982) La civilt Appenninica 1959-1976. Per un
bilancio critico, 57-79 in Studi in Onore di Ferrante Rittatore
 Trial By Dure 183
Vonwiller Parte Prima, vol. 1, Como, Lito-tipografia G.
Malinverno.
Cedronio, M. (1977) Profilo dlie Annales attraverso le pagine
dlie Annales, 1-72 in Storiogrqfia Francese di leri e di Oggi,
(Ed.) M. Cedronio, F. Diaz and C. Russo. Naples, Guida.
Chapman, R. W. (1987) Beaker Studies and Beaker Conferences:
an Introduction 1-4 in Bell Beakers of the Western
Mediterranean (Eds.) W. H. Waldren, R. C. Kennard,
(BAR S331 ( )). Oxford, Archaeological Reports.
Chartier, R. and J. Revel (1979) Lucian Febvre et les sciences
sociales. Historiens et Gographes (Feb. 79, no vol. no. given),
427-42.
Cherry, J. F. (1981) Pattern and process in the earliest
colonization of the Mediterranean Islands. Proceedings of the
Prehistoric Society 47, 41-68.
Cherry, J. F. (1984) 'The initial colonisation of the West
Mediterranean islands in the light of island biogeography and
palaeogeography', 7-27 in The Deya Conference of Prehistory,
(Ed.) W. H. Waldren, R. Chapman, J. Lewthwaite and R. C.
Kennard, (BAR S229 (i)). Oxford, British Archaeological
Reports.
Cherubini, G. (1981) La civilt del castagno in Italia alla fine del
medioevo. Archeologia Mdivale 8, 247-80.
Clark, T. N. (1971) Le patron et son cercle: clef de l'universit
franaise. Revue Franaise de Sociologie 12, 19-39.
Clark, T. N. (1973) Prophets and Patrons: the French University
and the Emergence of the Social Sciences. Cambridge Mass.,
Harvard University Press.
Clarke, D. L. (1968) Analytical Archaeology, 1st edn. London,
Methuen.
Clarke, D. L. (1972) Models in Archaeology. London, Methuen.
Clarke, D. L. (1978) Mesolithic Europe: the Economic Basis.
London, Duckworth.
Claval, P. (1968a) L'histoire de la gographie comme discipline
fondamentale, 9-33 in Pour la Cinquantenaire de la Mort de
Paul Vidal de la Blache (P. Claval and J. P. Nardy) ( = Cahiers
de Gographie de Besanon 16, Annales Littraires de
l'Universit de Besanon 93). Paris, Les Belles Lettres.
Claval, P. (1968b) Vidal de la Blache et la Gographie franaise,
91-125 in Pour la Cinquantenaire etc., op. cit.
Claval, P. (1969) Essai sur l'Evolution de la Gographie Humaine,
2nd edn. ( = Cahiers de Gographie de Besanon 12, Annales
Littraires de l'Universit de Besanon 67). Paris, Les Belles
Lettres.
Claval, P. (1984) The historical dimension of French geography.
Journal of Historical Geography 10(3), 229-45.
Coutau-Bgarie, H. (1983) Le Phnomne 'Nouvelle Histoire'.
Stratgie et Idologie des Nouveaux Historiens. Paris,
Economica.
Day, J. (1975) Malthus dmenti? Sous-peuplement chronique et
calamits dmographiques en Sardaigne au Bas Moyen Age.
Annales: Economies, Socits, Civilisations 30(4), 684-702.
Day, J. (1976a) L'insediamento precario in Sardegna nei secoli
XII-XVII, 228-42 in Atti del Colloquia Internazionale di
184 JAMES LEWTHWAITE
Archeologia Medievah (Palermo 1974). Palermo, Istituto di
Storia Mdivale dell'Universit di Palermo, vol. I.
Day, J. (1976b) Villaggi abbandonati e tradizione orale: il caso
Sardo. Archeologia Mdivale 3, 203-39.
Deffontaines, P. (1932) Les Hommes et Leurs Travaux dans
les Pays de la Moyenne Garonne (Agenais, Bas-Quercy).
Lille, Facults Catholiques. Reprinted 1978: Agen, Lib.
Quesseveur.
Dickinson, R. E. (1973) The Makers of Modern Geography, 2nd
edn. London, Routledge and Kegan Paul.
Febvre, L. (1922) (With L. Bataillon). La Terre et L'Evolution
Humaine, Introduction Gographique l'Histoire (l'Evolution
de l'Humanit). Paris, La Renaissance du Livre.
Furet, F. (1975) Frorn narrative history to history as a problem.
Diogenes (Engl. edn.) 89, 106-23.
Hexter, J. H. (1972) Fernand Braudel and th monde Braudellien.
Journal of Modem History 44, 480-539.
Kinser, S. (1981) Annaliste paradigm? The geohistorical
structuralism of Fernand Braudel. American Historical Review
86, 63-105.
Le Lannou, M. (1941) Ptres et Paysans de la Sardaigne. Tours,
Arrault.
Lewthwaite, J. G. (1981) Plains tails from the hills: transhumance
in Mediterranean archaeology, 57-66 in Economic
Archaeology, (Eds.) A. Sheridan and G. Bailey (BAR S96).
Oxford, British Archaeological Reports.
Lewthwaite, J. G. (1982) Acorns for the ancestors: the prehistoric
exploitation of woodland in the West Mediterranean, 217-30
in Archaeological Aspects of Woodland Ecology, (Ed.) M. Bell
and S. Limbrey. (BAR S146). Oxford, British Archaeological
Reports.
Lewthwaite, J. G. (1983a) The Neolithic of Corsica, 146-83 in
Ancient France, (Ed.) C. J. Scarre. Edinburgh, Edinburgh
University Press.
Lewthwaite, J. G. (1983b) La culture des castelli de la Corse:
dernier tmoin d'un genre de vie ouest-mditerranen
d'autrefois. Archeologia Corsa 6-7 (1981-82), 19-33.
Lewthwaite, J. G. (1984) Pastore, padrone: the social dimensions
of pastoralism in prenuragic Sardinia, 251-68 in The Deya
Conference of Prehistory, (Eds.) W. H. Waldren, R. Chapman,
J. Lewthwaite and R. C. Kennard. (BAR S229 (i)). Oxford,
British Archaeological Reports.
Lewthwaite, J. G. (1985a) From precocity to involution; the
Neolithic of Corsica in its West Mediterranean and French
contexts. Oxford Journal of Archaeology 4(1), 47-68.
Lewthwaite, J. G. (1985b) Transhumant and Sedentary Pastoralism
in Earlier Corsican Prehistory. Unpublished Ph.D. thesis,
Cambridge University.
Lewthwaite, J. G. (1985c) Colonialism and nuraghismus, 220-51
in Papers in Italian Archaeology IV. The Cambridge
Conference, (Eds.) C. Malone, S. Stoddart (BAR S244).
Oxford, British Archaeological Reports.
Lewthwaite, J. G. (1986) The Transition to food production: a
Mediterranean perspective in 53-66 Hunters in Transition.
(Ed.) M. Zvelebil. Cambridge University Press.
 Trial By Dure 185
Lewthwaite, J. G. (1987) The Braudelian Beaker: a Chalcolithic
conjoncture in the Western Mediterranean? 31-60 in Bell
Beakers of the Western Mediterranean, (Eds.) W. H.
Waldren, R. C. Kennard (BAR S331 ( )). Oxford, British
Archaeological Reports.
Ley, D. (1980) Geography without Man: a Humanistic Critique
( = Research Paper 24). Oxford, School of Geography.
Lukermann, F. (1965) The 'calcul des probabilits' and the Ecole
Franaise de Gographie. Canadian Geographer 9(3), 128-37.
Lcke, H. (1976) Ostkorsika. Untersuchungen zum
Kulturlandschaftswandel einer insularen Region (Mainzer
Geographischer Studien 5) 2 vols. Mainz, Geographischer
Institut der Johannes-Gutenberg Universitt.
Mann, H. D. (1971) Luden Febvre. La Pense Vivante d'un
Historien ( = Cahiers des Annales 16). Paris, SEVPEN.
Michels, J. W., E. Atzeni, I. S. T. Tsong & G. A. Smith (1984)
Obsidian hydration dating in Sardinia, 83-113 in Studies in
Sardinian Archaeology, (Eds.) M. S. Balmuth and R. J.
Rowland, Jr., Ann Arbor, University of Michigan Press.
Moreno, D. (1982) Storia e archeologia forestale. Una premessa.
Quaderni Storici 49, 7-15.
Nardy, J. P. (1968) Levasseur, gographe, 35-89 in Pour la
Cinquantenaire etc., op. cit.
Prince. H. (1975) Fernand Braudel and total history ( = review of
Braudel 1972b). Journal of Historical Geography 1, 103-6.
Ravis-Giordani, G. (1975) Le troupeau errant. L'utilisation de
l'espace par les bergers du Niolu (Corse). Fourrages 63, 3-21.
Revel, J. (1979) Histoire et sciences sociales: les paradigmes des
Annales. Annales: Economies, Socits, Civilisations 34(6),
1360-76.
Rhein, C. (1982) La gographie, discipline scolaire et/ou science
sociale? Revue Franaise de Sociologie 23, 223-51.
Ribeiro, O. (1965) En relisant Vidal de la Blache. Annales de
Gographie 77(424), 641-62.
Santamaria, U. & A. M. Bailey (1984) A note on Braudel's
structure as duration. History and Theory 23(1), 78-83.
Sherratt, A. G. (1981) Plough and pastoralism; aspects of the
secondary products revolution, 261-305 in Pattern of the Past.
Studies in Honour of David Clarke, (Ed.) I. Hodder, G. Isaac,
and N. Hammond. Cambridge, Cambridge University Press.
Stoianovich, T. (1976) French Historical Method - The Annales
Paradigm. Ithaca and London, Cornell University Press.
Tangheroni, M. (1977) Lunghi secoli di isolamento. Nuova Rivista
Storica 61, 149-81.
Trevor-Roper, H. R. (1972) Fernand Braudel, the Annales and the
Mediterranean, Journal of Modern History 44, 468-79.
Vigne J. D. (1983) Les Mammifres Terrestres Non Volants du
Post-Glaciaire de Corse and Leurs Rapports avec l'Homme:
Etude Palo-Ethno-Zoologique Fonde sur les Ossements (3
vols.). Thse de 3me cycle, Universit Paris VI.
Vigne, J. D. (1984) Premires donnes sur le dbut de l'levage du
mouton, de la chvre et du porc dans le sud de la Corse
(France), 47-65 in Animals and Archaeology 3. Early Herders
and their Flocks, (Eds.) J. Clutton-Brock and C. Grigson
(BAR S202). Oxford, British Archaeological Reports.
186 JAMES LEWTHWAITE
Vidal de la Blache, P. (1911) Les genres de vie dans la gographie
humaine. Annales de Gographie 20(111), 193-212 and (112),
289-304.
Vidal de la Blache, P. (1922) Principes de Gographie Humaine
(posthumously edited by E. de Martonne). Paris, A. Colin.
Weiss, M. C. & F. de Lanfranchi (1976) Les civilisations
nolithiques en Corse, 432-442 in La Prhistoire Franaise II:
Les Civilisations Nolithiques et Protohistoriques, (Ed.) J.
Guilaine. Paris, CNRS.
Zbl, D. (1982) Die Transhumanz (Wanderschafhaltung) der
Europischen Mittelmeerlnder im Mittelalter in Historischer,
Geographischer und Volkskundlicher Sicht ( = Berliner
Geographischer Studien 10). Berliner Institut fr Geographie,
Technische Universitt Berlin.
 Section 3
Conceptual Issues Associated with Environmental
and Population Change
 Introduction

Eric Grant,
Middlesex Polytechnic

The papers in this section are all concerned with change. Some look
mainly at environmental change and some mainly at changes within
human populations, but a strong theme throughout this section is
the relationship between cultural and environmental change.
As human culture developed and man's actions became more
complex we perceive his growing impact on the natural environment
through forest clearance, grazing and cultivation. Professor Thornes
investigates how man's actions affect soil erosion in a Mediterranean
environment. Thornes questions the traditional idea that man-
controlled grazing accelerates soil erosion in the Mediterranean,
mainly because the basis of our assertions concerning the role of
man in changing the erosional environment in prehistoric times is
very slender. As we are better able to reconstruct the initial vegetation
and 'natural' erosion, Thornes concentrates on their impact on graz-
ing systems through a series of differential equations that describe
the interaction between soil erosion and vegetation cover on the one
hand and vegetation cover and grazing on the other.
The key issue highlighted in this paper is that we should not take
accelerated erosion as an immediate consequence of poorer vegeta-
tion cover. Thornes refers to his work in southeast Spain where,
despite poor vegetation cover, erosion rates were spatially very vari-
able and in many areas the rates were unexpectedly low. The role
of grazing in degrading the vegetation cover and its subsequent role
in the erosion process is seen to be a much more complex
phenomenon than hitherto supposed. On being questioned about
how the time factor operates in his model, Thornes emphasised that
most of his work has been on how the variables interact on each
other in a spatial sense as there is a great deal to be learned about
phase space without getting involved in time space.
Burrin and Scaife develop a theme similar to that of Thornes in
their examination of the impact of man on valley alluviations. Though
one might assume that the larger the valley fill the greater the
anthropogenic impact, they stress the need for great care in interpret-
ing the valley response, as the alluvial record is variable through
190 GRANT
both space and time. Catastrophe theory is now recognised as a
useful framework for investigating major environmental changes,
and Burrin and Scaife see the cusp catastrophe model as providing
a conceptual framework within which possible responses of fluvial
environments to the impact of man on valley sides could be indicated.
They also apply systems concepts, perceiving the critical issue to be
system sensitivity, as it is very dangerous to make direct suggestions
about population density etc. based on the magnitude of valley-fills.
For example, a significant alluvial accumulation in a sensitive
environment might have developed a response to valley-side activity
by a less technologically advanced culture with considerably smaller
population numbers.
The advantage of combining archaeological and geomorphological
survey work is stressed in the paper by Hours and Copeland. Where
there is excellent geological stratigraphy with prehistoric material
embedded in it, a great deal of useful information can be gathered.
However, reality is usually less good than this sequences are rarely
complete, and mixing of artefacts and levels has always to be con-
sidered. A close appreciation of the geomorphic processes is thus
vital in coming to some understanding of the archaeology embedded
within.
The impact of prehistoric agriculture on the environment, one of
the great issues of prehistory, is taken up by Dr Green who sees the
clearing and planting associated with the spread of farming into new
habitats as setting off a complex series of ecological responses and
cultural adjustments. The palaeoecological record, in particular pol-
len data, has been used for many years now to chart the rise and fall
of particular species. Green argues that we need more systems-based
understanding of ecological communities with species as their inter-
active components. His study of the impact of agriculture in Denmark
during the Atlantic period uses principal components analysis to
derive associated groups of species which are interpreted in terms
of the ecological processes. The disruption caused by forest clearance
and farming operates through community processes, and in this way
the ecological responses to agriculture can be clearly shown in the
pollen record. However, in the discussion that followed this paper
other workers on Danish pollen data reported that by using a combi-
nation of cluster analysis and discriminant analysis they found that
a fairly homogenous forest zone persisted long after the initial agricul-
tural indicators arrived, suggesting that the impact of agriculture on
the natural vegetation was very slow. Also, the assumption that the
pre-neolithic woodland was a climax assembly may not hold so well
if the effect of mesolithic communities on the forest cover could be
established. The ability to isolate ecological communities clearly was
also questioned, as attempts to do this in a study of mammalian
communities found that what appeared to be associated species was
in fact separate ecological groups that varied seasonally.
 Introduction 191
In concentrating their enquiries on the production and consump-
tion of food, palaeoecologists have perhaps ignored the role of disease
in environmental archaeology. Dr Brothwell argues that studies of
disease in the past should not be divorced from the main themes of
archaeology as, far from being uncommon in the past, disease is an
all-pervading phenomenon. The study of palaeopathology is develop-
ing fast, and Brothwell makes a plea that morbidity as well as
mortality should be recorded at the site level, even if specialist
evaluation of the pathology is not possible.
The influence of pathogens (such as viral diseases) and pest
restraints on crop productivity must now be considered as an integral
part of studies of plant domestication. Plant disease can have a major
impact on human settlement, being responsible for movement and
resettlement, and may even be a possible explanation of the collapse
of the classic Maya civilisation.
The papers by Groenman-van Waateringe, Boyle and Wright, and
Chapman are all concerned with change within human populations,
and more specifically look at pressure on human populations in
different ways. The need to see man as part of a wider ecosystem
embedded within, and interacting with, the physical environment is
stressed by Professor Groenman-van Waateringe. The question of
why and how humankind changed from hunter/gatherer to food
producer is approached through research on non-human animal
populations to examine the relationship between stress and the ability
to change birth rates and population dynamics, for example by
limiting the number of female births. In the early post-glacial period,
the food available to humans expanded dramatically as the popula-
tion of large herbivores increased. Once a ceiling density had been
reached the herbivore population began to decline, but the human
population increase was still in full swing due to long oscillating
movements. This in turn gave rise to food shortages, social stress,
etc. which can only be checked by the acceptance of a new ecotech-
nology such as food production. The impetus for change is not the
original lack of food, but rather its abundance leading to population
growth and then food imbalances. On being questioned about other
theories concerning the mesolithic/neolithic interface, Professor
Groenman-van Waateringe disagreed with the idea that mesolithic
communities were adaptively successful and that some form of stress
precipitated a change towards food production.
Boyle and Wright examine the structure of hunter/gatherer hunting
strategies by explicitly modelling appropriate features of the environ-
ment. Using as a simple example the nutrients that can be obtained
from reindeer and eels, they show how linear programming can be
used to predict the optimal allocation of scarce or limited resources
among competing needs. They argue that hunter/gatherers did have
192 GRANT
a foraging strategy which was partly determined by culture and partly
by physiological needs. The group that satisfied human needs more
easily was more likely to survive when resources were scarce and
competition high. Discussion of this paper raised the general question
of nutritional needs and the authors, on being asked why they had
concentrated on protein rather than energy, answered that most
studies have not distinguished between the needs for fat and carbohy-
drate, which are in fact important but different energy requirements.
The whole gamut of nutritional needs should be considered; but this
gives rise to a very complex situation that has not yet been worked out.
Dr Chapman examines in some detail the concepts of population
pressure, carrying capacity and population growth in prehistoric
societies. However, traditional formulations of population theory,
from Malthus to Boserup, are largely eschewed for more modern
elaborations of the core concepts of population pressure and carrying
capacity. Recent ideas see cultural relativism as being of greater
importance in controlling population growth, though concepts of
population pressure and carrying capacity can be accommodated
within cultural relativism, albeit in diluted form.
This leads to the idea that individuals should not be seen simply
as breeding units but as contributors of labour to a given mode of
production, as the number of people available for work is one of the
main factors in the scale of the economy. In the domestic mode of
production, for example, population growth is usually high because
children are needed as part of the domestic labour force, a need that
outweighs other limiting factors. Chapman strongly believes that the
case for continuous population pressure can no longer be maintained,
and alternative explanations require urgent development for many
major aspects of cultural evolution.
The broad theme of human and/or environmental change that
dominates the papers in this section has raised many important
conceptual issues. Most authors stressed the complexity of their
subject and the difficulty of dealing with data sets that are incomplete
in space and time. Professor Adams, however, observed that almost
all the case studies in this section are spatially tied to temperate
Europe or the Mediterranean. The past technological and biological
systems in this area, though complex enough, pale beside the prob-
lems that are raised when examining environmental change in the
tropics. There one is dealing with often unknown ecotechnologies
and biological systems that include hundreds of tree and animal
species, thousands of insect species, unknown numbers of viruses
and pests. What the papers in this section show is that though we
have started conceptualising the problems of studying past environ-
ments, the field is hardly touched and we already have more problems
than answers.
 Erosional Equilibria Under Grazing

J. B. Thornes,
University of Bristol

Empirical relationships between erosion and human activities have

been established to varying extents for a variety of processes. Unfortu-
nately it is difficult to extrapolate these results to different places and
different times. As a result, the basis for assertions concerning the
role of man in changing his erosional environment in prehistoric
times is very slender. There are two steps involved, the reconstruction
of the initial vegetation and the 'natural' erosion, and then the
investigation of the effects of human activities on the reconstructed
palaeo-environment. Considerable progress has been made in estab-
lishing the former in recent years, provided that some guess can be
made about the pre-existing climate (e.g. Kirkby and Neale 1987),
and these techniques have been utilised in environmental archaeology
(Gilman and Thornes 1985). In this paper an attempt is made to
provide a basis for the second. The problem may be stated as follows:
under a given climate and initial soil type and thickness, how will
erosion perform in the presence of vegetation of varying cover
amounts which is being degraded by grazing? The general conceptual
base is outlined in Thornes (1985). The essential approach is to write
down a set of differential equations describing the interaction between
soil erosion and vegetation cover on the one hand, and vegetation
cover and grazing on the other. The general strategy is derived from
theoretical ecology and exemplified by Noy Meir's (1979) investiga-
tion of stability in grazing-vegetation relationships.
We shall assume first that the rate of erosion is a function of excess
overland flow and slope (generally called Hortonian overland flow),
though the model can easily accommodate saturated overland flow.
This erosion rate increases as a result of erosion through the decreas-
ing permeability of the freshly exposed layers and decreases to zero
as the available soil is used up. The rate of erosion decreases as the
amount of vegetation increases. A second case considers gully erosion
by running water, where gully growth is regarded as a function of
gully area and upstream drainage area. In both cases vegetation
grows logistically, but its rate of increase is restricted both by soil
194 THORNES

erosion, which transports away litter and nutrients, and by grazing,

which reduces the overall biomass (Thornes 1986).

SOIL EROSION BY WASH

We start with an analysis of a simple wash model that is shown
schematically in Figure 1. In this model we shall assume that erosion
has the effect of increasing the rate at which it takes place up to
some maximum value which is determined by the available transport
capacity. This condition, represented by line (a) is the transport
limited case. If there is a fixed depth of soil, then once that soil has
been used up the rate of removal falls to a very low rate determined
by the rate at which it is supplied by weathering, and this is called
the weathering limited case. In Fig. 1 this is shown by line (b) where
the rate of weathering is assumed to be zero. Even when there is
material available an upper bound to its removal may be set by the
vegetation cover, as shown by line (c).
The rate of soil erosion (in kg/m 2 /yr) is given by:
= f(q,s) (1)
in which q is the overland flow and s the surface gradient. Several
variations of this formula are available (e.g. Band 1985). In this case
we shall use
E = k q2 s 1.66 (2)
which is generally called the Musgreave equation. If excess precipita-
tion is given as (I-f) where I is the rainfall intensity, k a constant
and f is the mean infiltration rate in a storm of fixed length T, and
we measure ground loss instead of mass, expressing the loss as normal
to the ground surface then we can write
dz/dt = k ( T - ( I - f ) ) 2 - s 1.66 (3)

WASH CASE
(a)

Actual erosion amount

(a) transport limited (b] supply limited (cj vegetation limited
C= capacity set by runoff
Z'= Value at which all material has disappeared

Figure 1. Basic terminology of wash model.

 Erosional Equilibria Under Grazing 195
where T is a constant, providing either a fixed area, or a fixed time,
or both. In what follows we use it as a time constant and assume
area to be constant but not specified, f could also be considered as
total intercepted or stored water to allow for the saturated overland
flow model which may be appropriate in semi-arid and arid environ-
ments when the soils are very thin. In this case T represents the
average storm length, the integral over dz/dt representing the cumula-
tive erosion over the time for many storms, i.e. a time scale appropriate
to vegetational effects rather than individual storms. In the calcula-
tions which follow it is assumed that the value of T is one hour.
The value of f at the surface is taken to be fo. As soil is removed
horizons with slower infiltration rates further down through the soil
are exposed and it is assumed that infiltrability decreases with depth
such that
fz = fo exp(-b z) (4)
where fz is the infiltrability at depth z cm below the surface and b
is a coefficient. In addition we need to accommodate the effects of
the lower boundary of the soil. As this is approached, there is a rapid
fall off of the erosion rate to zero, as no more soil is available. This
is called the weathering limited case. We wish the erosion rate to fall
only as z approaches close to z', the initial depth of the soil, z'-z is
the residual depth after erosion. Such a rapid drop is given by
Dl = (l-exp(z-z')) (5)
and a plot of equations (4) and (5) is given in Fig. 2 for some typical
values.
The erosion rate is also diminished by the presence of vegetation.
To model this effect the results of Elwell and Stocking (1976) have
been used. They showed that the rate of erosion caused by particular
0 0.2 0.4 0.6 0.8 1.0

91-
92-
93-
94
95

Figure 2. Typical infiltration rate decay function (left) and proportionate

reduction in erosion near soil boundary (right).
196 THORNES
storms was damped exponentially as a function of the vegetation
cover in the growing season (Fig. 3). This effect is represented by
the equation:
D2 = (l-exp(-k 5 -v)) (6)
where D2 is the fraction of the bare soil value which is removed with
different amounts of seasonal vegetation cover. K5 is a constant
which is taken here to have a value of 7.0 based on the empirical
data of Elwell and Stocking. The actual lowering of the erosion rate
due to the vegetation cover is then given by (T I)2 D2. Setting f = o
assumes that the soils are impermeable or that there is no loss to
storage. This is equivalent to saying that in the absence of vegetation,
the erosion is assumed to be operating at the maximum possible rate
for that environment. With increasing cover this falls exponentially
to about 3% of the maximum value by the time cover has reached
75%.
The three terms, equations (4), (5) and (6) are brought together as:
dz/dt = (T (I-fo exp(-b z)) 2 (l-exp(z-z'))
-(I-T) 2 -(l-exp(-k 5 -V))
which is equation (7).
The behaviour of this equation, which assumes slope to be constant,
is shown in Fig. 4. At first the rate of erosion increases as the lowering
of the surface induces lower infiltration rates and consequently more
run off and higher erosion rates. However, as the rate of decrease of

\
1
1 X
1
1 o
1
Dss as percenta je of bare plot r

>
as percentage Df mean annual r

t Soil loss
\
1
1
\
>

c3

\

<

\
\
\

\ \
r
U

gj
'0
(
\ \ Si
Runoft^^^
-V^_ |

Percent mean seasonal vegetal cover

Figure 3. Elwell and Stocking's (1976) curve showing reduction in soil loss
and sediment yield with increasing seasonal vegetation cover.
 Erosional Equilibria Under Grazing 197
(a) constrained by permeability
(b] constrained by depletion and permeability
CcJ constrained by vegetation, depletion and permeability
(a)

10 20 30 40 50 60 70 80 90 100 110 120 Ccm)

Depth of soil removed -

< Soil thickness

Figure 4. Change in soil loss as soil thickness decreases due to changes in

permeability, boundary conditions and vegetation cover.

infiltration falls off the rate of erosion asymptotically approaches the

limit provided by q = I2. As the amount of lowering reaches the
original thickness of the soil, the rate of erosion falls dramatically
to zero. If z' is very large relative to z then the erosion rate continues
to increase up to a limit set by the b value in the infiltrability decay
curve and determined by the vegetation cover.
The question arises as to the effect of the interaction between
vegetation and erosion, i.e. by setting dz/dt = o and solving for V in
terms of z. The resulting curve (shown in Fig. 5) is called the erosion
isocline. It indicates the values of vegetation required to maintain
equilibrium, i.e. such that there is no further erosion. The diagram
shows that as the amount of soil-loss increases there has to be an
increase in the vegetation cover to offset it. As the depletion effect
sets in erosion falls to zero, i.e. for z>z' there is no erosion so that
no vegetation is required to keep the level of erosion at zero. This is
consistent with observations that in many arid areas, although there
is little or no vegetation, there is also little or no erosion.
There is one other interesting feature to be observed in this diagram.
By calculating the relative magnitudes of the changes in dz/dt at any
point in the phase space, we can decide in which direction the system
will move if starting at that point. For example, with a residual soil
of only 30 cm a 42% erosion cover will increase the erosion, moving
the system to the higher limb of the isocline. At 43% the erosion rate
198 THORNES
Unstable Stable

Depth of soil removed

Soil thickness

Figure 5. Isocline for erosion by wash.

is negative and the system is carried (by deposition) to z = o, i.e. a

full soil cover. Since the isocline is a set of equilibria, it can be seen
that to the left of the peak all these equilibria are unstable, whereas
to the right, they are all stable. To the left, a slight shift of the erosion
rate will carry the system either to a condition of complete soil cover
(on the z = o axis) or one of nearly complete soil loss on the right
hand limb. Above the isocline, all trajectories lead to complete soil
cover. It is important to recall that in this example slope has been
held constant, and that there is in reality a family of such curves
representing different slope values.

SOIL EROSION U N D E R GULLYING

Consider now the case in which the erosion is taking place by
concentrated flow, and that we consider the area occupied by gullies
as a measure of erosion. The case is rather more complicated than
that discussed above and a simplified model is outlined here. For
example the network comprises a single channel which is not bifurcat-
ing (c.f. Begin 1983, Thornes 1984). The general framework is shown
in Fig. 6. We suppose a gully whose area is growing headwards into
a uniformly inclined slope and whose rate of increase in area is
proportional to the area of the gully itself, i.e.
da/dt = k, a (8)
This is argued from the effect of the increased runoff resulting
from within the gully itself, in which convergence of flow is occurring
and slopes are propagating by mass failure and wash at a higher rate
than the hillslope into which the gully is cut, because of basal incision
 Erosional Equilibria Under Grazing 199

area of cell=A Period | Period j Period

t 2 3

Figure 6. Basic notions of gully erosion model.

and because vegetation is assumed to be zero in the gully. Such

growth has been observed empirically (e.g. by Bradford, Piest and
Spoomer 1977) and inferred inductively for systems of prairie gullies
by Faulkner (1976). In the early stages, however, growth is determined
dominantly by the contributing area of the regional slope into which
the gully is growing. We shall consider the simple case in which this
growth is a linear function of the available area (-a). In Fig. 6b
this is the first period of growth, before the self-sustaining growth
described by equation 8 has become significant in period 2. The
equation accommodating both of these terms is now:
da/dt = k, - a + k2 (A-a) (9)
As gullies grow they eventually come into competition with each
other, both across the headward divides and laterally. Since there
is a minimum discharge required to sustain a channel, we follow
Schumm (1956) in arguing that there is also a minimum area. In
practise what is really required is a minimum stream-power, and here
we are assuming that this is provided by a constant slope and
minimum area. This area ac is inversely related to drainage density
and can therefore be related back to the rainfall and infiltration
controls. We shall assume it characteristic of a given region, and call
it drainage capacity. In the evolving situation the limiting area avail-
able for growth is the (A-ac). As the gully area approaches this limit,
the rate of growth progressively declines. This is ultimately reflected
in the lower sediment yields of highly dissected areas and the
'erosional paradox' noted by Thornes and Gilman (1984). Setting
amax = (A-ac) then:
da/dt = (k, a + k 2 (A-a)) (l-(a/amax)) (10)
gives us the three effects together. The first term inside the left-hand
bracket is the exponential growth term, the second term the contribut-
ing upslope area term, and the term in the right-hand bracket is the
limiting area (or spatial competition) term. Graphs of this equation
200 THORNES
for constant values of k, and different values of k2 are shown in Fig.
7. k, is the growth per unit area of gully, k2 is the control on the
discharge production per unit area of the contributing upslope area.
When k2 = o it is assumed that there is no contribution from the
upslope area to growth. This is typical of the situation in which
gullies initiate on arcuate slips and are sustained by erosion and
propagation from the unvegetated failure surface. The condition in
which amax is very large occurs when there is no constraint on growth
from spatial competition (the 'open prairie' case of Faulkner (1976)).
Now consider again the effects of vegetation on this growth model.
It may be introduced as a divisor of the term providing for growth
as a function of the catchment area, on the assumption that it is in
this area that the effects of vegetation are most likely to be felt.
A plot of the growth rate against area for the equation:
da/dt = (k, a + (k2/v) (A-a)) (1-a/amax) (11)
is shown in Fig. 8. It reveals that under this model, with sparse
vegetation cover, the fastest growth rates occur when the gully is
small relative to the contributing upslope area. If the vegetation is
greater, then the required gully growth area for maximum growth
rate is greater, though above about 50% these vegetation cover effects
are less important. The erosion isocline for this case (when growth
rate is zero) is obtained as:
V=((A-a/a)-(k 2 /k 1 )) (12)
In this model, as a increases the equilibrium vegetation cover
required to keep it in check declines exponentially at a rate which

I 5 9 13 17 21 25 29 33 37 41 45 49 COO's)

Figure 7. Gully growth rate as a function of maximum area and unslope

control.
 Erosional Equilibria Under Grazing 201
1000-
900-
800-

O 700-
cc
^ 600-
| 500-

r
o "-
< 300-
200-
100-
0
1 4 9 13 17 21 25 29 33 37 41 45 49 COO's)
Area

Figure 8. Gully growth rate in relation to vegetation cover on hillslope

upstream of gully head.

depends on the ratio of the contributing area effect to the positive

feedback effect.
If we simply assume that vegetation reduces the overall growth
rate as a linear proportion of its cover fraction (i.e. add a term -k3 V
to equation 10) i.e.,
da/dt = <ki a + k2 (A-a)) (l-(a/amax))-k 3 V (13)
then the growth curves have the form shown in Fig. 9 and the erosional
isocline shown in Fig. 10. This reflects the form shown in Fig. 7 by
requiring the greatest vegetation cover to keep medium-sized gullies
in check. For this isocline, gully growth is checked completely when
V = A-(k 2 /k 3 ). As with the previous model, the left limb of the
isocline is unstable and the right limb is stable.
VEGETATION COVER
Although there is a -continuing debate about the nature of succession,
especially in relation to the importance of a species diversity approach

9 13 17 21 25 29 33 37 41 44 48 COO's)
Area

Figure 9. Growth rate in relation to vegetation cover with linear vegetation

effect, k, is the coefficient of vegetation reduction of erosion rate.
202 THORNES
500-
450
400-

350-

300-

250-

200-

150-

100-

50-

1 4 9 13 17 21 25 29 33 37 41 44 49 (OO'sJ
Area

Figure 10. Gully erosion isocline for linear vegetation effect.

versus a functional approach, we adopt the latter in this paper for

three reasons. Firstly, our ability to deal with vegetation composition
in terms of productivity and water stress is still very limited despite
over forty years of agronomic research. Secondly, we are even more
ignorant of the relations between soil erosion and individual species,
though work is proceeding along these lines (e.g. Francis et al. 1986).
Thirdly, our knowledge of selective foraging by sheep and goats, at
least in the Mediterranean region, does not permit an elaborate model
of that process, even though it may be reasonably established else-
where. At best we can provide rough estimates of the annual biomass
removed both by erosion (as litter and soil organic matter) and by
sheep and goats.
If we assume that vegetation cover growth is logistic and is inhibited
by overall erosion loss, then a suitable model for the process in a
moisture stressed situation might be:
dv/dt = g V (l-V/Vmax)-f(z,V) (14)
where V is again vegetation cover, z the depth of erosion, Vmax is
the probable maximum vegetation biomass and g is the early growth
rate for vegetation.
The first term on the right-hand side is the logistic growth equation
for vegetation which, for moisture stressed vegetation, is reasonably
well based in theory (see Noy Meir 1978) and is to some extent
confirmed by post-fire growth studies in Mediterranean areas. If V
and Vmax are expressed in terms of biomass, then dv/dt is the
biomass productivity (gm/m 2 /yr). Vmax is the expected equilibrium
standing crop given temperature, light and moisture constraints. We
could assume for dry environments that it is a linear function of
precipitation and that the appropriate coefficients are provided, for
example from the work of Wagner (1975) as cited in Lundholm (1983).
 Erosional Equilibria Under Grazing 203
For Mediterranean data the literature (e.g. Mooney 1981) suggests
2
for evergreen forests a value of about 33 kg/m 2
of total biomass,
with a shoot production of about 650 gm/m /yr; for chapparral 5
kg/m 2 and 400 gr/m 2 /yr and for degraded matorral (phrygana) about
2.7 kg/m 2 and 400 gr/m 2 /yr. These figures suggest a value of g of
about 0.1, which would require about seventeen years recovery time
to reach a low matorral after clearance, and a maximum productivity
of about 125 gm/m 2 /yr.
The second term to the right of equation (14), f(z,V), expresses
productivity loss as a function of erosional loss and vegetation cover.
This is a complex relationship which is rather poorly understood.
Nevertheless, there is abundant evidence from field experiments to
show that the productivity of eroded land relates directly to the depth
of top soil which has been lost (Anon 1981) and reclamation tech-
nology specifically attempts to incorporate these effects. From the
other side, however, both agronomists and geomorphologists have
paid little attention to the organic matter actually being removed
from slopes. Even EPIC (Erosion Productivity Impact Calculator,
Williams 1985) considers only the top 10 mm, though this is undoub-
tedly of great importance. Enrichment ratios, the ratio of organic
matter in eroded soils to those in parent materials, have been studied
for a long time (e.g. Massey and Jackson 1952) and the evidence
indicates organic matter ratios of 2, total nitrogen of up to about 3
and available phosphorus up to 4, reflecting the selective transport
of finer particles to which the nutrients and organic matter are
selectively bound.
A simple model would express these losses in terms of a linear
product of vegetation biomass and the amount of soil lost, leading
to an expression for vegetation production of the form:
dV/dt = gV(l-V/Vmax)-k 4 V - z (14)
and the behaviour of these curves, with different values of Vmax is
shown in Fig. 11. The inset to this diagram shows the vegetation
isocline:
z = (g/k 4 )-(l-V/Vmax) (15)
which is a linear negative function intercepting the V = o axis at g/k 4
and with V transformed from biomass to cover using:
Vc = k 5 -(V/0.0625) 0 5 (16)
following Whittaker and Marks (1975) and assuming for the Mediter-
ranean that k s = 16.6. On this inset the erosion required to keep
vegetation cover constant decreases linearly as the vegetation cover
increases. This is because vegetation is included in the f(z,V) con-
straint. With the constraint set simply at -k4 z then the isocline
reflects the growth curves for different vegetation covers.
204 THORNES

20 25 30 50 COO's)
Vegetation cover g/m2

Figure 11. Vegetation growth curves for different upper limits (expressed
in gm/m 2 ) as a function of vegetation cover. The inset shows the
vegetation isocline, with the dashed arrows giving the vectors of change in
different part of the phase portrait.

These models are still relatively crude. A simple improvement

would be to make the effects of erosion a non-linear function of
depth, thus recognising that there are at least two important effects
in what Ellison (1940) called fertility erosion. These are:
(1) reduction in available moisture. There is less available moisture
as more soil is removed if porosity, permeability and infiltrability
decrease with depth. Plant productivity is directly related to soil
moisture in the root zone.
(2) loss of soil nutrients. The soil organic matter tends to decrease
away from the surface as the materials supplied by leaf fall and other
surface processes are mixed into the surface. Although inorganic
minerals tend to increase through the profile towards the unweathered
boundary (except where there are precipitated layers) with active
vertical mixing and surface decomposition, the appropriate form for
productivity loss with erosion is assumed to take the exponential
form, providing:
dV/dt = g V (1-V/Vmax) - k6 exp(b(z-z')) (17)
which isocline is again a symmetrical function controlled by Vmax.
Since we are a very long way from being able to parameterise this
model or even to understand its more general properties we shall
restrict our discussion in the next section to the case defined by
equation (13).
In the case of gully growth (as opposed to wash) it may be assumed
that the vegetation of the upslope area is unaffected by the expansion
 Erosional Equilibria Under Grazing 205
of the gully itself since in equation (11) and (13) vegetation control
is implemented over the residual area. In practice there is some
evidence to show that the gully head actually drains out the slope
causing some vegetation death (Bull 1979); but this effect is generally
restricted to a very small area immediately adjacent to the gully head,
and so is not related to the overall gully size. Under these conditions,
the isocline for vegetation is a vertical line at V = Vmax. This
apparently uninteresting situation becomes important when grazing
effects are taken into account.
EROSION-VEGETATION INTERACTION
Before considering the impact of grazing, let us consider the effects
of interactions between erosion and vegetation. There are two ele-
ments to be considered. The first is the joint equilibrium between the
two constrained processes, erosional soil loss or gully growth on the
one hand, and vegetation growth oft the other. This is obtained by
solving the two isoclines simultaneously, as illustrated graphically
in Fig. 12, which shows the intersection of the wash and vegetation
isoclines. In this diagram the heavy dots represent the equilibrium
solutions, i.e. the value at which there is zero change in erosion depth
or vegetation cover. A is the only joint equilibrium because for B,
C, D and O, one or other or both of the variables have zero values.
The second operation consists in solving the resultant vectors for
the growth given by both equations to obtain the direction of change
of the system, assuming it to be at a particular value of V and z to
start with. A family of all solution curves as subsets of this two
dimensional system of equations is called the 'phase portrait' and
recalls the ideas on page 198. On Fig. 12 the arrows and the shading

ABCB
COAA
OADD

Vegetation cover
Figure 12. Combined erosion (for wash case) and vegetation isoclines.
Areas are shaded according to direction of change of the system.
206 THORNES
indicate the general character of the phase portrait. Whatever the
starting value, in sector CAB the system is driven by the modelled
behaviour to a condition of maximum erosion at point B. Likewise
from sector OAD the system must end up with the maximum possible
vegetation cover. From sector CAO, and from the area outside the
two curves, the trajectories carry the system towards the equilibrium
point at A which is a saddle. A slight shift from this point in vegetation
cover, caused by effects external to the system (artificial or natural),
will in turn drive the system to one of the two extreme conditions,
so that point A is seen to be an unstable equilibrium by contrast to
points B and D which are stable equilibria. Like point A, point C is
a saddle point whereas the system is always moving away from O
which is therefore a repellor. Mathematical analysis can be carried
out to confirm what has been demonstrated here graphically (e.g.
Waltman 1983). This situation is directly comparable to the competi-
tive exclusion principle in theoretical biology. The probability that
the system will occupy some intermediate position exists, but is very
small.

GRAZING
There is no doubt that grazing results in loss of biomass and there
is some evidence that it enhances surface runoff and erosion. Lusby
(1970) for example recorded the effects of grazing by comparing
grazed and ungrazed catchments near Grand Junction, Colorado.
Measurements made ten years apart showed that cover on ungrazed
watersheds had remained virtually unchanged. Runoff on the
ungrazed catchments was about 30% less than the grazed watersheds,
and sediment yield about 45% less. Gullies on grazed watersheds
had about twice as much erosion as those of ungrazed watersheds.
The differences in runoff and sediment-yield took place soon after
livestock were excluded, and were attributed to trampling effects on
soil that was loosened by frost-heave. There is of course a great
danger in transferring the results drawn from one ecological environ-
ment to another, but it is clear that the effects of both physical impact
of cattle and their cropping is to increase runoff and soil erosion.
Detailed studies of both mechanisms are still awaited.
There are at this stage two possible strategies. The first, adopted
in this paper and by far the simpler, is to consider only the effects
of grazing on biomass productivity. The second is to assume that
soil erosion and biomass reduction themselves have feedbacks on
grazing, so that the populations of grazing animals are limited in
their growth by shortage of adequate forage. The latter case is clearly
of great relevance in the contemporary drought problems of Africa.
In present-day Mediterranean environments, with the reduction in
herding and the control exercised by herders to achieve maximum
 Erosional Equilibria Under Grazing 207
productivity in terms of body weight and product yield, it is probably
a good deal less important. Certainly it can have had little importance
in prehistoric times, for although there is evidence of overgrazing it
is likely that stocks simply moved on to greener pastures rather than
experience decline in numbers. Although there is some excellent
work on vegetation-grazing prdation (e.g. Noy Meir 1978, 1979,
Caughley 1976) the competition between erosion and vegetation,
with grazing prdation on the vegetation, appears not have been
tackled before. We consider here density independence among the
grazers, and leave the more complicated case for a later paper.
Consider the effects of grazing on wash process given by equation
(7), the erosional effect on vegetation as -k4 V z and allow the
grazing to reduce the biomass growth rate by -k9 G, where G is the
consumption rate obtained as the average number of sheep grazing
events per unit of time multiplied by the average consumption per
event. In an interaction model we should again want the consumption
rate to be a function of the availability of forage, there usually being
an ungrazable residual. In this simpler case we have:
dV/dt = gV(l-V/Vmax)-k 4 V z-k9 G (18)
whose isocline is:
dV/dt = (g/k 4 ) (1-V/Vmax) (19)
The plot of this case is shown in Fig. 13a. In this diagram the
heavy lines are the isoclines given in Fig. 12, though in this case they
have been simulated for different parameters for illustrative purposes.
The dashed lines are the vegetation isoclines for different grazing
densities. The diagram illustrates that the peak equilibrium vegetation
is lower as the grazing densities increase, that the peak equilibrium
erosional loss shifts to higher vegetation cover values and that the
'safe' segment, for which complete vegetation cover is the end result
(shaded in Fig. 13) is reduced in size from both sides.
Similar analyses of the canonical forms can be provided for the
other cases. Fig. 13b shows the wash case with the vegetation isocline
conditioned by a linear effect of erosional depth (i.e. -k4 z). In the
gully case the isocline for V is independent of the area of the gully,
and hence is a vertical line. Fig. 13c is, in effect, Fig. 10, the isocline
for gully area (and hence erosion) with the axes transformed and
the vegetation isocline added. With no grazing the vegetation in the
hillslope area upstream of the gully head is unaffected. With grazing,
however, the lower vegetation cover interacts with gully growth
(through equation 13) providing first a single stable equilibrium at
intermediate gully sizes. A further reduction in vegetation due to
grazing will result in two possible equilibria, a stable one with large
208 THORNES

1 2 3 4 5 6 7 9 10 1 2 3 4 5 6 7
Vegetation cover

Gully Gully

Figure 13. Phase portraits for simulations of basic governing equations

with competition between erosion and vegetation and grazing prdation on
vegetation cover. The grazers do not respond to low forage availability.

gullies or an unstable saddle point at small gully areas. Fig. 13d

shows the same relationships for equation 12. Each intersection of
the two isoclines is here a stable equilibrium point, the equilibrium
occurring at higher and higher gully areas as grazing becomes more
intense.

In this work we have tried to show how the competition between

erosion and vegetation cover may be modelled, and how the models
may be roughly parameterized for Mediterranean environments. In
addition we have attempted to add the effects of density-predation
by grazing animals. The simple competition models result in typical
exclusion effects whereby either vegetation cover or soil erosion
dominates in the struggle. The models provide an approach for
developing predictive and management strategies based on the iden-
tification of the phase portrait zones leading to vegetation domination.
By adding the grazing component stable equilibria appear in some
cases, accommodating the coexistence of erosion and a less than
 Erosional Equilibria Under Grazing 209
complete vegetation cover and suggesting the conditions under which
this might occur. These results go some way to explaining the
erosional paradox noted by Thornes and Gilman (1983) whereby
despite the poor vegetation cover in south-east Spain, erosion rates
are spatially very variable and in many areas unexpectedly low.
Current work is being directed to the improvement of both the
theoretical and empirical basis of these models. In particular, a model
for grazing population growth appropriate to nomadic herding is
being developed and data on the amount of organic matter eroded
and the amount of biomass removed is being collected.
BIBLIOGRAPHY
Anon (1981) Soil erosion effects on soil productivity: A research
perspective. /. of Soil and Water Conservation 36(2), 82-90.
Band, L. (1985) Field parametrization of an empirical sheetwash
equation. Catena, 12, 281-290.
Begin, Z. (1983) Applications of 'diffusion' degradation to some
aspects of drainage net development, pp. 169-180, in Badland
Geomorphology and Piping, (Eds.) R. Bryan, and A. Yair, Geo
Books, Norwich.
Bull, W. F. (1979) Threshold of critical power in streams.
Geological Society of America Bulletin 90, 453-64.
Caughley, G. (1976) Plant-herbivore systems, pp. 94-113 in
Theoretical Ecology (Ed.) R. M. May. Blackwell Scientific
Publications.
Ellison, W. D. (1950) Fertility erosion. The Land, 9:7.
Elwell, H.A. & M. A. Stocking (1976) 'Vegetal cover to estimate
soil erosion hazard in Rhodesia'. Geoderma 15, 61-70.
Faulkner, H., An allometric growth model for competitive gullies.
Zeitschrift fr Geomorphologie, Supplement 20, 36-63.
Francis, C. F., Thornes, J. B., Romero Diaz, A., Lopez-Bermudez,
F. and Fisher, G. C. (1986) Topographic controls on soil
moisture, vegetation cover and land degradation in a
moisture-stressed Mediterranean environment. Catena 13(2),
211-225.
Gilman, A. & J. B. Thornes (1985) Land Use and Prehistory in
South East Spain. London, George Allen and Unwin.
Kirkby, M. J. & R. H. Neale (1987) A soil erosion model
incorporating seasonal factors'. International Geomorphology
(Ed.) V. Gardiner, v. 2, 189-210, Wiley.
Lundholm, B. (1976) Domestic animals in arid ecosystems.
Ecological Bulletin, (Stockholm), 24, 29-42.
Lusby, G. C. (1970) Hydrologie and biotic effects of grazing
versus non-grazing near Grand Junction, Colorado. United
States Geological Survey, Professional Paper 700-B, B232-B236.
Massey, H. F. & M. L. Jackson (1952) Selective erosion of soil
fertility constituents. Soil Science Society America, Proceedings,
16, 353-56.
Mooney, H. A. (1981) Primary production in Mediterranean-
climate regions, pp. 249-256 in F. Castri, D. W. Goodall and
R. L. Specht (Eds), Mediterranean Type Shrublands. Elsevier.
Noy Meir, I. (1978) Stability in simple grazing models: effects of
explicit functions. J. of Theoretical Biology 71, 347-380.
210 THORNES
Noy Meir, I. (1979) Graphical methods in ecology, pp. 453-472 in
G. P. Patil and M. Rosenzweig, (Eds.), Contemporary
Quantitative Ecology and Related Ecometrics. Cooperative
Publishing House, Maryland.
Piest, R. F., J. M. Bradford & R. G. Spoomer (1975) Mechanics of
erosion and sediment movement from gullies, pp. 162-176 in
Present and Prospective Technology for Predicting Sediment
Yields and Sources. United States Department of Agriculture,
Agriculture Research Service, ARS-S40.
Rogowski, A. S. (1985) Evaluation of potential top soil
productivity,. Environmental Geochemistry and Health 7(3),
87-97.
Schumm, S. A. (1985) Evolution of drainage systems and slopes
in badlands at Perth Amboy, New Jersey. Bulletin Geological
Society of America 67, 597-646.
Thornes, J. B. (1984) Gully growth and bifurcation, pp. 140-156
in Man and Erosion. Proceedings of XI Symposium of
International Soil Erosion Conference, Denver, Colorado,
1984.
Thornes, J. B. (1985) The ecology of erosion. Geography 222-34.
Thornes, J. B. (1987) The paleoecology of erosion. In Wagstaff,
J. M. (Ed.) Landscape and Culture. Blackwells, Oxford, 37-55.
Thornes, J. B. & A. Gilman (1983) Potential and actual soil
erosion around archaeological sites in south-east Spain.
Catena Supplement 4, 91-113.
Wagner, F. H. (1976) Integrating and control mechanisms in arid
and semi-arid ecosystems. Considerations for impact
assessment. Symposium on Biological Evaluation of
Environmental Impact, New Orleans, 1976, 40pp.
Waltman, P. (1983) Competition Models in Population Biology.
Regional Conference Series in Applied Mathematics, no. 45,
Society for Industrial and Applied Mathematics, 77 pp.
Whittaker, R. H. & P. L. Marks (1975) Methods of assessing
terrestrial productivity, pp. 55-118 in Leith, H.L. and
Wittaker, R.H., (Eds.) Primary Productivity of the Biosphere.
New York, Springer Verlag.
Williams, J. (1985) The physical components of the EPIC model,
pp. 272-284 in El-Swaify, S. A., W. C. Moldenhauer and A.
Lo, Soil Erosion and Conservation, Soil Conservation Society
of America.
 Environmental Thresholds, Catastrophe Theory and
Landscape Sensitivity: Their Relevance to
the Impact of Man on Valley Alluviations

P. J. Burrin, Goldsmiths' College, University of London

and
R. G. Scaife, Institute of Archaeology, University of London
There is widespread and increasing evidence of the significant role
played by prehistoric cultures in causing or exacerbating inorganic
inwash from valley sides to valley floors (Iversen 1941, Godwin 1944,
Knox 1972, Becker and Shirmer 1977, Limbrey 1978, Bell 1981,1984,
Burrin and Scaife 1984). This is often recorded as episodes of acceler-
ated sedimentation within bog or floodplain lithostratigraphy. Studies
of this alluvial architecture have led to the suggestion (Limbrey 1978,
Bell 1982) that many British river valleys exhibit a basically compar-
able sequence of Holocene sediments. Early Flandrian depositionary
episodes are apparently characterised by low rates of sedimentation,
waterlogging and the development of grey-black deposits with high
organic matter contents. Conversely, later Flandrian deposits found
above these basal organic-rich materials are essentially oxidised
sediments with a much higher mineral content. There is evidence at
several sites (Bell 1982) that the organic-rich basal sediments were
laid down under woodland conditions, whereas the inorganic inwash
was deposited following subsequent periods of forest clearance by
man. This finding raises the possibility of making some tentative
archaeological inferences regarding population numbers within an
area, size of settlement and technological ability from the consider-
ation of rates and amounts of valley accumulation.
All else remaining equal, it might appear reasonable to assume
that the larger the valley fill, induced by increased soil erosion and
sediment yield as a result of anthropogenic valley side disturbance
of the vegetation cover, the greater would be the probable impact.
With the advance of technological and agricultural skills and increas-
ing population numbers, an apparent increased environmental impact
in later prehistory might also be expected, partly reflecting an increase
in arable land use. Such a relationship is evidenced in the alluvial
record of some British river valleys, such as the Severn-Avon (Shotton
1978) and the Thames (Hazleden and Jarvis 1979, Limbrey 1983,
Robinson and Lambrick 1984) where there is good evidence of Bronze
Age and more recent activity exacerbating valley-side inwash as a
result of changing land use. A similar response can also be detected
212 BURRIN AND SCAIFE
in the colluvial record of a number of British dry valley networks
(Bell 1981). However, consideration of valley bog and floodplain
lithostratigraphies elsewhere demonstrates that a more variable and
complex situation exists than has previously been recognised. It can
be suggested from this diverse pattern and magnitude of valley
response through space and time that caution needs to be exerted in
the archaeological assessment, interpretation and significance of such
environmental change. The concepts of environmental threshold,
complex response and landscape sensitivity are of specific relevance
here and together fit neatly into a conceptual framework offered by
catastrophe theory. The aim of this paper is to discuss these concepts
and to indicate the value of their application in environmental
archaeology with reference to valley sediment accumulation in low-
land Britain.

BRITISH HOLOCENE VALLEY ACCUMULATION

Valley sediments can generally be sub-divided into a number of larger
sedimentary associations. British river valleys are usually comprised
of two major sedimentary associations, the floodplain coastal zone
association and the inland floodplain association. The floodplain
coastal zone consists of a thick sequence (up to 30 m or more) of
interbedded freshwater, estuarine, littoral and shallow marine
deposits and is found within the lower valleys within the sedimentary
coastal hinterland. These sediments are replaced inland by a thinner
sequence of interdigitated and interbedded freshwater alluvial and
colluvial materials (valley-side derived material not significantly
reworked by riverine processes). A generalised correlation of
Holocene floodplain deposits for a selected sample of British valleys
is presented here (Fig. 1) in order to demonstrate a number of points.
It should be appreciated that this is not an exhaustive collection of
data, for in many valleys the sedimentary sequence is known although
its chronology remains less clear. Nevertheless, the selection is
sufficient to indicate the following aspects of alluviation. First, many
valleys exhibit a considerably more complex sequence of interbedded
Flandrian deposits than recognised by Limbrey (1978) or Bell (1982).
Materials differ from one valley to the next, reflecting geographical
location and geological variation, and distance from river mouth (see
Fig. 1). Typically sediments range from inorganic gravels, sands, silts
and clays, to calcareous marls and tufas and organic accumulations
including peats. Second, phases of erosion and deposition are
frequently asynchronous both at different sites within one valley and
between different valleys. This helps to demonstrate the third point,
namely the problems and complexities of identifying the impact of
climatic change on colluvial and fluvial environments from the
sedimentary record, an increasingly well-known problem (e.g. Bell
 Environmental Thresholds 213

'000 YEARS B. P

Figure 1. A generalized correlation of Holocene alluviation for selected

British river valleys.
214 BURRIN AND SCAIFE
1982, Burrin 1983, 1985). Finally, the onset of alluviation is also
seemingly asynchronous, varying between sites from one valley to
the next. Hence, the alluvial record for the most recent period of
earth history is variable in both space and time.
If the alluvial sequences influenced by possible coastal factors
are removed from this generalised correlation, the sedimentary
record exhibited indicates the inland floodplain alluvial fill
sequences (Fig. 2). It should be appreciated that knowledge
concerning many British inland floodplain fills is limited, and that
in many instances only part of the alluvial stratigraphy has been
exposed in river bank sections by fluvial erosion or excavated for
archaeological purposes. Hence, relatively little is known about
the deeper and presumably older Holocene deposits found
beneath contemporary inland floodplains. A notable exception is
research currently being undertaken by deep augering projects in
the Weald (Burrin and Scaife 1984, Scaife and Burrin 1983, 1985).
Nevertheless, evidence from many inland locations suggests that
the bulk of Holocene sedimentation had been accomplished by the
end of the Bronze Age, with relatively small and minor subsequent
phases of alluviation (Burrin and Scaife 1984). Much late

VALLEY
RECORD OF ANTHROPOGENICALLY INDUCED VALLEY ALLUVIATION
LOCATION

MIDDLE NENE

LOWER GIPPING
i i ? P?
-? P?
WrNDRUSH -* ,,;,,...,..,.:-,.-

I 1= ?

UPPER THAMES No evidence of earlier alluviation - -- ?

UPPER OUSE No evidence of earlier alluviation ---. e
::
-!
ES silty clay
;: e --- ;
1
UPPER CUCKMERE No evidence of earlier aliuviation <! : :::::|V

MEDINA
' l l l l l l l l l l H"! ^' * * *' '*' !"
jh^s -I ,\lt Alfr ,*, ,
E

significant eroiion

minor erosion
SEVERN -AVON JJllil -- ?
p pedogenesis
IDLE 7 ,\, organice
? :=::

COUNTY RECORD OF ANTHROPOGENICALLY INDUCED VALLEY COLLUVIATION

period of hillwa"
aarlieat recorded
'initiation of
colluviation

MESOLITHIC | NEOL. l BRONZE j IRON|RO.|SAXON|MECH [PM

AL l BofeV | ATLANTIC SUB-BOREAL SUB-ATLANTIC

0 '000 YEARS B.P.

Figure 2. A comparison of anthrogenically induced phases of valley

alluviation and colluviation (highly generalized).
 Environmental Thresholds 215
Flandrian sedimentation frequently appears to be largely inorganic
(as noted by Limbrey 1978) and to have developed in response to
anthropogenic valley-side activity, although the patterns of erosion
and deposition found within valleys is frequently asynchronous,
demonstrating a variable response in alluviation. Consideration of
the colluvial record (Bell 1981, 1982) indicates a similarly diverse
response which is apparent even when simplified diagrammatically
(Fig. 2).
Such variation is perhaps not surprising and might be explained by
consideration of a range of archaeological factors. These could
include changing population numbers and their distribution through
space and time; migration around the country; changes in technology
and agricultural practices; cultural displacement; locational preferen-
ces, etc. However, such interpretation is masked by the variable and
frequently complex response of natural environmental systems to
changing phenomena. Previously such dynamic interaction has been
modelled by means of a square wave lag function (Knox 1972; Fig.
3). The inter-relationships were considered between climatic change
and three other variables: relative vegetation cover, the potential for
hillslope erosion and potential sediment supply. Consideration of
the model suggests that rapid climatic change from arid to more
humid conditions results in an accompanying but slower increase

CLIMATE
HUMO WET

POTENTIAL FOR HILLSLOPE

POTENTIAL S DIMENT YIELD

Figure 3. The square wave lag function illustrating climatic change and
biogeomorphic response with respect to sediment yield (from Knox 1972).
 216 BURRIN AND SCAIFE
in vegetation density, such that the potential for hillslope erosion
decreases in a reciprocal manner (Fig. 3). Maximum or minimum
sediment yield correlates strongly with the maximum or minimum
potential for hillslope erosion. This diagram can be similarly
employed in considering the possible impact of man on the vegetation
cover. As vegetation cover decreases because of forest clearance or
other agricultural practices, so the potential for hillslope erosion and
sediment yield increases accordingly. The nature of the changes when
man is involved is probably faster (i.e. a squarer wave function) than
under natural climatic changes when there is more of a lagged
response.
While this is a very useful model it is limited in that it does not
provide an indication of the possible range of responses by the valley
floor to these pulsed inputs of colluvium a factor of interest to
environmental archaeologists. The variation in the sediment record
demonstrated by British valley fills alone (Fig. 2) suggests that the
alluvial responses can be very different from one site to another.
While it is recognised that a variety of endogenic geomorphological
factors can contribute towards the production of complex floodplain
sequences in the short term (e.g. the interaction of discharge, slope,
stream power, etc.), inputs of materials from interfluves to valley
bottoms are frequently controlled in the longer term largely by the
extent of vegetation cover and hence soil stability within a catchment
(Schumm and Lichty 1965). This in turn is dependent upon climate
and, over more recent time, the impact of man.

CATASTROPHE THEORY
Catastrophe theory was developed by the French mathematician
Ren Thorn (1975) and is derived from differential topology (the
study of those properties of geometric figures that are unaffected by
deformation). It has been described and discussed theoretically by
many authors including Woodcock and Poston (1974), Zeeman (1976,
1977) Croll (1976), Wagstaff (1976), Zahler and Sussman (1977),
Woodcock and Davis (1978), Poston (1979), Renfrew and Poston
(1979) and Renfrew (1974). Geological and geomorphological appli-
cations have been noted by Cubitt and Shaw (1976), Henley (1976),
Graf (1979) and Thornes (1979), while Wagstaff (1976, 1978) and
Wilson and Kirkby (1980) have outlined its suitability for geo-
graphical studies. The versatility and wider application of catastrophe
theory has been demonstrated by Poston and Stewart (1978) and
also by Zeeman (1979, 1980, 1982), while its potential role in
archaeology has been considered by Renfrew (1979, 1984), Renfrew
and Poston (1979) and Renfrew et al. (1982). Given this range of
more detailed discussion, a full analysis of the theory is not provided
here, but a summary of the main ideas follows.

_
 Environmental Thresholds 217
Thorn was interested in describing the origin of forms (mor-
phogenesis) and in particular in changes in form. His theory focusses
on spatio-temporal structures undergoing change, and is concerned
with discontinuous changes brought about by the operation of con-
tinuous change in the variables controlling a given system. Usually,
differential calculus is used to describe the transition from one state
to another where a smooth change occurs. However, the flexibility
of calculus is limited when discontinuous operations or abrupt transi-
tions occur, as in step functions. Catastrophe theory is designed as
a language to replace calculus in such situations and indicates that
all changes in the four dimensional universe (length, width, height
and time) seemingly proceed in only a limited number of ways
depending upon the number of control factors and response variables
considered. Thorn has demonstrated that the space-time continuum
can be characterised by seven elementary catastrophies or types of
change. Each catastrophe represents a pattern of behaviour deter-
mined only by the number of control factors, not by their nature, or
by the interior mechanisms that connect them to the system behaviour,
or by their rate of change.
Systems have a tendency to behave in similar ways in that they
seek a minimum potential (i.e. tend towards stability or maximise
entropy) often by homeostasis. The concept of potential is closely
linked to that of equilibrium (Fig. 4a) of which there are many kinds.
The graphical representation of potential and equilibrium exemp-
lified in Fig. 4a shows there are four equilibrium points. Only the
two minima are points of stable equilibrium; the point of inflection
is semi-stable and the local maximum is unstable. This picture can
be extended to three dimensions (Fig. 4b). Instead of a point of
inflection, the surface has a saddle point: a place which is at the
crest of a ridge when approached along the z axis but at the bottom
of a trough when approached along the axis. These examples are
fixed graphs of potentials, but in reality the factors that determine
potentials are constantly changing. In many instances continuous
change in these factors can bring about discontinuous change in the
behaviour of systems.
Catastrophe theory grows out from these generalisations. In any
system governed by a potential and in which the systems behaviour
is determined by up to four different factors, only seven qualitatively
different types of discontinuity are possible (Table 1). System
behaviour or response is measured by a variable (x) that is controlled
by two variables (a and b). These three variables are related to each
other by the relationship E(a,b,x,), where = an energy function.
For each combination of a and b there is a corresponding value of
x that minimises E, representing an equilibrium state, such that
E/x = 0 (i.e. the first derivation of the energy function equals zero).
218 BURRIN AND SCAIFE
(a) a curve
local maximum
jump or catastrophe (unstable equilibrium)

point of inflection
(semi-stable equilibrium)
V /.
local minimum
(stable equilibrium)

(b) a surface

local minimum

local maximum

Figure 4. Singularities of a curve and a surface (modified from Woodcock

and Davis 1978).

Table 1. Elementary Catastrophes (based on Thorn 1975, Wagstaff 1976 and Graf 1979)

Derivative
(singularity
Control Behaviour Energy denned when
Singularity Factors Factors Function equal to zero)
Fold 1 1 x 3 /3-ax x 23 -q
Cusp 2 1 x4/4axbx 2 /2 x 4 -a-bx
Swallow's Tail 3 1 x 5 /5-ax-bx 2 /2-cx 3 /3 x -a-bx-cx 22
Butterfly 4 1 x 6 /6-ax-bx 2 /2-cx 3 /3 x 5 -a-bx-cx
-dx"/4 -dx3
Hyperbolic 3 2 x3 + y3 + ax + by + cxy 3x2 + a + cy
3y2 + b + ex
Elliptic 3 2 x3 xy2 + ax + by + ex2 + cy2 3x2 y 2 + a + 2cx
-2xy + b + 2cy
Parabolic 4 2 X 2 y + y4 + ax + by + ex2 + dy2 2xy + a + 2cx
 Environmental Thresholds 219
Usually co-variations of a, b and that satisfy the equation form a
plane such that each combination of a and b corresponds to one
value of x. In some instances, however, one combination of a and b
produce two or more corresponding values of x. The result is a surface
characterised by convolutions or folds. Each surface is defined by
an equation and characterises how system change proceeds. They
are called catastrophies because a small change in one variable may
result in a catastrophic change in the behaviour or response variable
as its value changes radically across a fold. Thus, a catastrophe is a
surface that is a graph of all (a,b,x) points where the first derivative
of the energy function is zero and each system state plots as a point
on that surface. System change is defined by movement from one
equilibrium condition to another. If the system is not in equilibrium,
the point does not plot on the surface.
The simplest system has one control factor and one behaviour or
response variable and is represented by a two dimensional graph
(Fig. 5a) that is a fold catastrophe. For one range of values of the
control factor there are both maxima and minima and the system
will usually tend to move spontaneously to the minimum (point of
stable equilibrium). Beyond a certain range of values for the control
variable the system breaks down and is undefined. The fold catas-
trophe has only limited application because few systems are control-
led by one variable. Furthermore, there are only a few things than
can happen in such a system, all of them rather obvious. The system
can: (i) move to a state of minimum potential (if the system condition
allows one to exist); (ii) it can be balanced (semi-stable or quasi-
equilibrium) at the point of inflection; or (iii) it can be essentially
unstable, having no available minimum.
A system with two control factors and one response variable
produces a complex three dimensional surface that is characterised
by a tuck or cusp, i.e a surface with smooth sections separated in
places by a reverse slope (Fig. 5b). As the control variables (a and
b) change, so the response variable (x) responds and a point rep-
resenting an equilibrium state moves about the surface. Some move-
ments may be smooth (z-z' in Fig. 5b) whilst others consist of
dramatic leaps or falls across the cusp in order to maintain equili-
brium. All the points on the underside of the cusp are unstable
maxima. All the points along the fold line or cusp are semi-stable
points of inflection. The remaining points are stable minima.
The cusp is the highest dimensional catastrophe than can be drawn
in its entirety. Systems of more complexity than this catastrophe are
definable in a topologie and algebraic sense (Table 1 ), but they cannot
be drawn in three dimensions unless one or more dimensions are
held constant. This perhaps limits their suitability as it is difficult to
conceptualise in more than three dimensions. A second important
220 BURRIN AND SCAIFE
(a) the fold catastrophe

minima - response
(system stability) (behaviour)
variable
threshold
point of inflection
(semi-stable state)

control factor

system moves
spontaneously to
minima (stability)
maxima -"
(system completely unstable)

(b) geometry of the cusp catastrophe

response (behaviour) surface

threshold

7 bifurcation set
control surface (behaviour in the cusp)

Figure 5. The fold and cusp catastrophes.

consideration is that it becomes increasingly difficult to define the

control factors for more complex systems and consequently, higher
order surfaces are not considered here. A very good account of these
more complex catastrophies is provided by Woodcock and Davis
(1978). Despite these limitations, employment of catastrophe theory
is useful in that it enables the prediction of types of system response
by considering the interaction of a stated number and type of control
variables. For instance, the variety of responses by the fluvial system
to man's activity on the interfluves can be demonstrated. Evidently,
the more control variables included, the more complex and sophisti-
cated the final range of responses recognised by the application of
the theory. This paper will focus on the application of a relatively
less sophisticated, low order catastrophe model, the cusp catastrophe,
as a first approximation to the understanding of a very complex
phenomenon. The cusp model indicates that changes during a sys-
tem's operation can be determined providing the system can be
 Environmental Thresholds 221
adequately described by two control factors and one response variable
(Table 1). The surface denned by the appropriate relationships in
Table 1 has six basic characteristics (Fig. 6) which can be utilised
as a conceptual framework within which a range of system responses
can be identified. These characteristics are: (i) abrupt and smooth
changes; (ii) divergence and convergence; (iii) hysteresis; (iv)
bimodal behaviour; (v) an unstable or inaccessible region; and
(vi) structural stability.

Smooth (AA') and abrupt (BB') b. Divergence and convergence

change (CC'and DD') (C'C and D'D)
Cusp boundaries are thresholds

c. Hysteresis d. Jump conventions:

simple hysteresis (EE') random (RR 1 ); perfect delay (PP1);
complex hysteresis (FF1) Maxwell convention (MM')

e. Bimodal behaviour f. Stable structure

Figure 6. Characteristics of the cusp catastrophe.

222 BURRIN AND SCAIFE
Each of these characteristics is now briefly considered in turn
prior to applying these facets of the model to the environmental
impact of prehistoric man on fluvial systems.

(i) The cusp catastrophe indicates that a system can change

smoothly or abruptly (Fig. 6a) from one equilibrium condition to
another. The actual response or relaxation path taken will depend
upon the antecedent state or condition of the system. If a point
representing the system is located away from the fold, then a small
change in a and b will result in a relatively small and smooth change
in x. This is the typical response of a relatively insensitive system.
Conversely, if a system is located near the fold, a small increase in
the control factors causes the system to cross the cusp whereby a
dramatic change in the response variable results as the system crosses
a threshold. The system can then be regarded as being sensitive.
Renfrew (1979) has suggested that the term anastrophe be applied
when the cusp is traversed by a system undergoing change from the
lower to the upper cusp surface, with the term catastrophe being
reserved for movement from the upper to the lower cusp surface.
Whilst it can be valuable to have a clear indication to the direction
across the response surface that a system takes, the differentiation
made by Renfrew may lead to possible confusion. This is because
the term 'catastrophe' takes on two slightly different meanings. Thorn
(1975) originally used the term catastrophe to define a surface describ-
ing system response, rather than a specific direction of system
response within a given catastrophe model. Consequently, it is recom-
mended that if the term anastrophe is found to be valuable in
catastrophe theory, an alternative term be introduced to cover system
response from an upper to a lower surface via a cusp.

(ii) The second characteristic is divergence (Fig. 6b). Given a

similar starting point, the system response to a particular stimulus
can move along very different relaxation paths to a new equilibrium
state. Antecedent conditions play a very important role in that they
indirectly determine system behaviour. One response is on the upper
surface of the pleat or cusp, whilst its neighbour travels beneath the
cusp. If both of them pass on the same side of the cusp, the behaviour
of the two systems remains similar. It is also possible that two systems
can start at two very different equilibrium conditions (above and
below the cusp respectively) but their relaxation paths move them
to a very similar end point. This is system convergence.

(iii) This third characteristic refers to a combination of cyclical

activity involving lags in the system. Hysteresis occurs when there
 Environmental Thresholds 223
are two responses for a given increase or decrease in the control
variables (Fig. 6c). System behaviour may respond with smooth
adjustments or by smooth transitions interrupted by abrupt changes
as the fold is crossed. The jumping or crossing from one state to
another is governed by one of these rules (Fig. 6d): perfect delay,
the Maxwell convention or the random approach. In perfect delay
the point moves into the fold or inaccessible region (which represents
change through time) and jumps to the other surface at the farthest
edge of the fold. Under the Maxwell convention the system jumps
as soon as possible and the systems adjust in a similar manner. Should
a jump occur at any time in the fold, then a random approach is
signified.

(iv) For each combination of the control factors (a and b), two
stable values of the behaviour variable (x) are possible (Fig. 6e). The
same conditions permit either of the two stable states. There is, in
fact, a third possibility where the system moves to the unstable
maxima on the underside of the cusp but this is generally inaccessible:
if the system occupies this state any disturbance will force it to the
stable point above or below. The exact system response is again
determined by antecedent conditions as well as the relaxation path.

(v and vi) These two characteristics are included together for they
are an inherent part of the cusp. The cusp catastrophe is a stable
structure (Fig. 6f) in that it is unchanging from one situation to the
next, unlike other models. If system behaviour is controlled by two
variables, then its change will always be governed by a cusp catas-
trophe. The inaccessible area or unstable region is a local maximum
of the energy function and has no significance in interpreting the
change process.

THE CUSP CATASTROPHE AND ANTHROPOGENIC ALLUVIATION

The characteristics of the cusp catastrophe described above indicate
that it should be possible to apply catastrophe theory to the problem
of both natural and anthropogenically-induced alluviation. It pro-
vides a conceptual framework within which possible responses of
fluvial environments to the impact of man on valley sides can be
indicated. If two control factors and one response variable are
defined, then it should be possible to predict the range of responses
the fluvial system can make. In applying the theory it should be
appreciated that the graphs depicted have no scale and only show
the canonical shape of the catastrophe surface. Furthermore, the
models are highly idealized by the assumption that only a single
Potential is involved. Theoretically, it is possible for a process which
224 BURRIN AND SCAIFE
shows cusp behaviour to evolve into a more complex butterfly or
umbilic catastrophe (see Woodcock and Davis 1978).
It is argued here that the amount of alluviation or colluviation in
valley bottoms, which determines the growth of floodplain and
alluvial fill deposits, is controlled by the interaction of two controlling
variables: the rate or amount of sediment inwash supplied to the
valley floor, and the transport capacity of the fluvial system to remove
that inwash. Given these controlling factors and the response variable,
it should be possible to describe the range of theoretical system
responses with reference to the characteristics of the cusp catastrophe.
There is evidence in some valleys that both alluvial (Shotton 1978)
and colluvial (Bell 1981) sediments have probably aggraded over
long periods of time during the Holocene. Colluvial sequences appear
to have accumulated gradually by the accretion of increments of
sediment rather than by the occurrence of a few major, high magni-
tude low frequency episodes (Bell 1982). These are examples of
smooth changes in sedimentation, where increased sediment inwash
occurs with little or no increase in transport capacity. There are many
examples of this type of accumulation and it appears to be most
frequently documented in many peatland or bog environments (e.g.
Simmons 1964, Scaife 1980, 1982). Perhaps a useful example is the
Gatcombe Withy Bed in the Medina Valley, Isle of Wight, where the
full Flandrian sequence is found in the peat biostratigraphy. Evidence
of prehistoric occupation of the valley sides (including the classic
Landnam clearance) is evidenced by palynological and archaeologi-
cal artefacts and by thin layers of inorganic inwash interspersed
within the organic rich accumulation (Scaife 1980). It is also possible
for significant thicknesses of inorganic alluvium to accrete gradually
in valley bottoms as in parts of the Severn-Avon valley (Shotton
1978, Brown 1983) or the Weald as exemplified by the upper Cuck-
mere at Stream Farm (Scaife and Burrin 1985). However, there is
also evidence that not all valley accumulation has occurred in this
fashion. Consideration of some Wealden valley fills (e.g. at
Sharpsbridge: Burrin and Scaife 1984) indicates that abrupt changes
accompanied by apparently rapid alluviation have resulted at inter-
vals during the Flandrian. This situation, marked in the biostratigra-
phy by very low organic contents and minimal pollen counts at some
levels, is modelled by the abrupt change characteristic of the cusp
catastrophe. Such sedimentation might be more frequently identified
in transitions from infrequent low phases of inwash to rapid pulsed
accumulations as a result of primary forest clearance. Where such
abrupt changes are involved the importance of environmental thresh-
olds cannot be overlooked, for it is the crossing of the threshold
(identified as the boundaries of the cusp in the response surface)
that is responsible for such a dramatic response. The closeness of a
 Environmental Thresholds 225
system to the cusp on the response surface also determines a further
significant factor, i.e. system sensitivity. If relatively insensitive, any
ensuing change will take place smoothly i.e. the relaxation path
of the system avoids the cusp (Fig. 7a); conversely, if the initial state
of the valley system is such that it is located close to the cusp (Fig.
7a), then it is liable to be relatively sensitive. Any disturbance of
the valley side vegetation cover could cause the relaxation path to
cross the fold and abrupt alluviation may prove to be the result.
The rate of response (relaxation time) will depend upon the
convention in which the cusp is crossed (random, Maxwell
convention or perfect delay, see Fig. 6). Thus, it cannot be
assumed the all valley fills have developed in a similar manner
even in areas adjacent to valley sides cleared repeatedly or
(a) smooth and abrupt alluviation

(b) bimodality

SS'-smooth change (insensitive system)

AA 1 - abrupt change (sensitive system)

aggradation

(c) hysteresis

complex hysteresis (d) divergence and convergence

(involving thresholds)
cut-and-fill sequences

- valley fill thickness

U- b -sediment inwash
* divergence
a-transport capacity
- convergence

Figure 7. Application of the cusp catastrophe to anthropologicaly induced

valley alluviation.
226 BURRIN AND SCAIFE
seasonally (e.g. winter wheat) for arable purposes. Much can
depend on system sensitivity.
It follows from this discussion that consideration of the
magnitude of valley fills or their accumulation rate, without
independent corroborative evidence, may provide only limited
assistance in attempts to assess more detailed archaeological para-
meters, such as consideration of population numbers, technological
impact on the environment, or geographical distribution. For inst-
ance, a seemingly small amount of sediment inwash in an insensitive
environment could have been produced by a technologically
advanced culture with either large total population numbers or high
population densities. Similarly, a significant alluvial accumulation
in a sensitive environment might have developed in response to valley
side activity by a less technologically advanced culture with consider-
ably smaller population numbers. Caution is required in assessing
archaeological information of this type from the amounts or rates of
valley sedimentation. Hence, environmental sensitivity is also a very
important conceptual consideration, and has been discussed in more
detail from a geomorphological viewpoint by Brunsden and Thornes
(1979).
These types of change have been most frequently documented in
the literature, but consideration of the cusp catastrophe also indicates
that other types of response can also occur as a result of natural or
man-induced sediment inwash. For example, the bimodality charac-
teristic is of interest in that it indicates that a system experiencing
an increase in sediment inwash can respond in one of two ways (Fig.
7b). First, it can aggrade or create a fill sequence, for the rate or
amount of sediment inwash increases faster than the transporting
capacity of the fluvial system is able to remove it. Conversely, an
increase in sediment inwash can result not in alluviation but in erosion
and excavation of the fill of previous accumulations. This occurs
because of a possible increase in the transport capacity of the system
which might develop as runoff increases due to decreased interception
and evapotranspiration as a result of the removed vegetation, or as
a consequence of other endogenic factors influencing the behaviour
of the fluvial system. This possible response is well-known to geomor-
phologists (see Schumm 1979) but may be less familiar to archaeol-
ogists. Hence, an increase in vegetation disturbance by man on the
valley side need not necessarily give rise to a phase of valley accumu-
lation. An example of this phenomenon has been described for the
historic development of some Wealden floodplains (Burrin 1983,
1985). It may prove difficult to identify such a response within
floodplain lithostratigraphies as being purely attributable to man,
but a possible example is offered by the small phases of late Flandrian
incision noted in some Wealden fill sequences (Burrin 1983) which
 Environmental Thresholds 227
may be the result of such activity. Such a possibility also demonstrates
the problem of recognising the impact of climatic change and
palaeohydrologic fluctuations within the sedimentary record (Bell
1982, Burrin 1985).
A further response of fluvial systems to man's activities on valley
sides is the well documented phenomenon known as cut-and-fill.
This can be exemplified by the hysteresis property of the cusp catas-
trophe and at least two examples can be recognised which occur at
the opposite ends of the hysteresis spectrum (Fig. 7c). In simple
hysteresis, as applied to valley alluviation, an initial increase in
sediment yield leads to the creation of an alluvial sequence (fill).
This results in a (local) increase in stream gradient, thereby increasing
transport capacity to the extent that excavation of the new deposit
takes place and the system returns to its original position. A similar
sequence of events might be envisaged for complex hysteresis except
that thresholds are involved because of antecedent conditions and
the relaxation paths the system is forced to take. The jump convention
across the fold will also help to determine the relaxation time of the
system.
Although cut-and-fill sequences are well known, it is comparatively
rare for the fluvial system to excavate completely a fill deposit. More
often, the system attempts to return to its original position but actually
reaches a state very similar to its original.
The divergence characteristic is an important one for it demon-
strates that one valley site or section alone can rarely be used to
typify the whole valley sequence. For example, two valley sites (A
and B) may have an initially similar starting position but, on
experiencing an increase in sediment from valley side soil erosion,
can 'arrive' at very different end positions (fig. 7d). For instance, the
relaxation path A-A' suggests that although an increase in both
sediment yield and transport capacity occurs, control factor b >
control factor a, such that alluviation occurs and a fill sequence
results. Conversely, where the position a > b occurs, the relaxation
path B - B' is followed and incision is the result. Hence, one clearance
event in a catchment can result in very different responses from one
location to the next.
It also appears logical that if divergence can occur, then there is
no reason why convergence or equifinality cannot also take place,
although this has rarely been considered by previous authorities in
applying catastrophe theory (e.g. Cubbitt and Shaw 1976, Henley
1976, Graf 1979). The principle of convergence is a common difficulty
encountered by earth scientists in that different processes and relaxa-
tion paths can produce the same, or very similar, end forms. With
respect to the cusp catastrophe, convergence occurs when two systems
start at positions above and below the tuck respectively, but their
228 BURRIN AND SCAIFE
relaxation paths move them towards a very similar state of equili-
brium (Fig. 7d). An example of this condition is given by considering
two valley floor sites one where the channel is flowing on the
valley rockhead with no alluvial fill, the other further downstream
where previous alluviation has resulted in the development of a
shallow accumulation. Should widespread forest clearance take place
throughout the catchment, then blanket alluviation might result giving
rise to a similar end form: i.e. an alluvial carpet laid down throughout
the inland floodplain sequence. This appears to have occurred within
the Sussex Ouse valley (Burrin 1983) at one stage during the later
Flandrian (probably c. Bronze Age). Where similar phases of wide-
spread anthropogenic activity on valley sides can be recognised, as
in the Thames or Severn-Avon valleys (Fig. 1 and 2) convergence
may also have occurred as these systems attempted to re-equilibriate.

A range of theoretical alluvial responses can therefore be recognised

for given anthropogenic valley side activities resulting in sediment
inwash to valley floors. The utilization of two control variables within
a cusp catastrophe has illustrated that a series of alluvial responses
can be identified which are based on the characteristics of the cusp
model. These are summarised in Table 2. Given these possible
responses it seems clear that the impact of man on valley alluviation
is far from simple and that a range of complex responses can be
envisaged. The types of response that will take place depend not
only upon the control variables but also upon the concepts of environ-
mental thresholds, the jump convention followed when the cusp is
crossed as a system follows its relaxation path and the idea of
landscape sensitivity. It is argued that these concepts are of relevance
to archaeologists concerned with evaluating the significance of valley
fills in terms of the impact of prehistoric man on the natural environ-
ment. While catastrophe theory appears to offer a suitable framework
within which these concepts can sit, there are some limitations that
should be appreciated. For instance, the descriptions and predictions
offered are purely qualitative rather than quantitive. The theory
merely describes process responses; it does not explain them or even
help to understand them further. No precise prediction can be made
as to when a discontinuity will take place and some attempts to apply
the theory have generated more questions than answers. Nevertheless,
it is a language which is well suited to the description of the behaviour
of environmental systems in that it allows the conceptualization of
a wide variety of problems and enables complex phenomena to be
modelled. In particular, the theory focusses on spatio-temporal struc-
tures undergoing change and is concerned with discontinous change
brought about by the operation of continous change in the controlling
variables. As such it provides a useful tool which can be applied to
 Environmental Thresholds 229
Table 2. Theoretical alluvial responses in a fluvial system for given
anthropogenic^?] stimuli utilising a cusp catastrophe framework.

Cusp
characteristic Theoretical alluvial/fluvial rponse

Smooth change Sediment inwash: small pulsed increments in sediment yield allow
progressive aggradation for given (limited) transport capacity.
Abrupt change more rapid rates of sediment inwash thresholds crossed
more rapid phases of pulsed accretion for a given time span.
Bimodality increased sediment yields result in either:
(a) creation of alluvial fill/floodplain by aggradation;
(b) removal of fill by erosion or creation of terrace by incision
(initial aggradation quickly followed by increase in transport-
ing capacity).
Hysteresis cut and fill sequences maintaining a similar equilibrium condition
to one prior to sediment inwash increases. Many involve single
or complex (thresholds involved) hysteresis relaxation pattern.
Divergence sediment inwash at two sites in a similar state of equilibrium gives
rise to two very different relaxation paths such that the systems
end up in different parts of the cusp: aggradation at one site;
erosion at another. Important implications for climatic and
palaeohydrologic inferences.
Convergence systems at very different equilibrium starting points in different
parts of the cusp converge as a result of their relaxation paths to
a similar end point of new equilibrium e.g. blanket alluviation
throughout an inland floodplain association.

N.B. Where relaxation paths involve crossing the cusp (thresholds) relaxation time
may be complicated by the jump convention followed.

the problems of eludicating man's impact on alluviation in fluvial

systems. Future research might attempt to apply higher order catas-
trophes to the problem of environmental system responses to a wider
range of control factors involving prehistoric man, thereby further
encouraging the interaction of geomorphological, ecological and
archaeological spheres of interest.
Acknowledgements. The authors would like to thank Helen Scoging
for helpful discussion, David Lawes for his cartographic assistance
and Nicky Pope for typing the manuscript.

BIBLIOGRAPHY
Becker, B. & W. Schirmer (1977) Palaeoecological study of the
Holocene valley development of the River Main, Southern
Germany. Boreas 6, 303-321.
230 BURRIN AND SCAIFE
Bell, M. (1981) Valley sediments and environmental change, pp.
75-91 in ed. M. Jones and G.W. Dimbleby. The environment
of man; the Iron-Age to Anglo-Saxon period, Oxford, BAR 87.
Bell, M. (1982) The effects of land use and climate on valley
sedimentation, pp. 127-142 in ed. A. Harding. Climatic
change in later prehistory. Edinburgh, Edinburgh Univ. Press.
Bell, M. (1984) Valley sediments as evidence of prehistoric land-
use on the South Downs. Proc. Preh. Soc. 49, 119-150.
Brown, A. G. (1983) Floodplain deposits and accelerated
sedimentation in the lower Severn basin, pp. 375-397 in ed.
K. J. Gregory. Background to Palaeohydrology. New York,
Wiley and Sons.
Brunsden, D. & J. B. Thornes (1979) Landscape sensitivity and
change. Trans. Inst. Brit. Geogr. New Series 4, 463-84.
Burrin, P.J. (1983) The character and evolution offloodplains with
specific reference to the Ouse and Cuckmere, Sussex. Unpubl.
Ph.D.thesis, Univ. of London.
Burrin, P.J. (1985) Alluviation in S.E. England and some implica-
tions for palaeohydrological studies. Earth Surf. Processes and
Landforms 10, 257-271.
Burrin, P. J. & R. C. Scaife (1984) Aspects of Holocene valley
sedimentation and floodplain development in southern
England. Proc. Geol. Ass. 95, 81-96.
Croll, J. (1976) Is catastrophe theory dangerous? New Scientist,
77,630-632.
Cubitt, J. M. & B. Shaw (1976) The geological implications of
steady-state mechanism in catastrophe theory. Math. Geol. 8,
657-662.
Godwin, H. (1944) Age and origin of the Breckland heaths of
East Anglia. Nature 154, 6-10.
Graf, W. L. (1979) Catastrophe theory as a model for change in
fluvial systems, pp. 13-32 in ed. D. A. Rhodes and G. P. Wil-
liams. Adjustments of the fluvial system. George Allen and
Unwin, London.
Hazleden, J. & M. G. Jarvis (1979) Age and significance of
alluvium in the Windrush valley, Oxfordshire, Nature 282,
291-292.
Henley, S. (1976) Catastrophe theory models in geology, Math.
Geol. 8, 649-655.
Iversen, J. (1941) Landnam i Danmarks Stenalder. Danm. Geol.
Unders. 66, 1-67.
Knox, J. C. (1972) Valley alluviation in southwestern Wisconsin.
Ann. Ass. Am. Geog. 62, 401-410.
Limbrey, S. (1978) Changes in quality and the distribution of the
soils in lowland Britain, pp. 21-26 in ed. S. Limbrey and J.G.
Evans. The effect of man on the landscape: the lowland zone.
London, C.B.A. Rep. No. 21.
Limbrey, S. (1983) Archaeology and palaeohydrology, pp. 189-
212 in ed. K.J. Gregory. Background to Palaeohydrology. New
York, Wiley and Sons.
Poston, T. (1979) The elements of catastrophe theory or the hon-
ing of Occam's Razor, pp. 425-436 in d. C. Renfrew and
K.L. Cooke. Transformations: Mathematical approaches to cul-
ture change. London and New York, Academic Press.
 Environmental Thresholds 231
Poston, T. & I. Stewart (1978) Catastrophe theory and its applica-
tions. London, Pitman.
Renfrew, C. (1979) Systems collapse as social transformation:
catastrophe and anastrophe in early state societies, pp. 481-
506 in ed. C. Renfrew and K..L. Cooke. Transformations:
Mathematical approaches to culture change. London and New
York, Academic Press.
Renfrew, C. (1984) Approaches to social archaeology. Edinburgh,
Edinburgh University Press.
Renfrew, C. & T. Poston (1979) Discontinuities in the endogenous
change of settlement pattern, pp. 437-461 in ed. C. Renfrew
and K. L. Cooke. Transformations: Mathematical approaches
to culture change. London and New York, Academic Press.
Robinson, M. A. & G. H. Lambrick (1984) Holocene alluviation
and palaeohydrology in the upper Thames valley. Nature 308,
809-814.
Scaife, R. G. (1980) Late Devensian and Flandrian palaeoecological
studies in the Isle of Wight. Unpubl. Ph.D. thesis, Univ. of
London.
Scaife, R. G. (1982) Late Devensian and early Flandrian vegeta-
tion changes in southern England, pp. 57-74 in ed. S. Lim-
brey and M. Bell. Archaeological aspects of woodland and
ecology. Oxford, BAR 146.
Scaife, R. G. & P. J. Burrin (1983) Floodplain development in and
the vegetational history of the Sussex High Weald and some
archaeological implications. Sussex Arch. Colls. 121, 1-10,
Scaife, R. G. & P. J. Burrin (1985) The environmental impact of
prehistoric man in the upper Cuckmere valley as recorded in
the valley fill deposits at Stream Farm, Chiddingly. Sussex
Arch. Colls.
Schumm, S.A. (1979) Geomorphic thresholds; the concept and its
applications. Trans. Inst. Brit. Geogr. New Series 4, 485-515.
Schumm, S. A. & R. W. Lichty (1965) Time, space and causality
in geomorphology. Am. J. Sei. 263, 110-119.
Shotton, F. W. (1978) Archaeological inferences from the study of
alluvium in the lower Severn-Avon valleys, pp. 27-32 in ed.
S. Limbrey and J. G. Evans. The effect of man on the
landscape: the lowland zone. CBA Res. Rep. 21, London.
Simmons, I. G. (1964) Pollen diagrams from Dartmoor. New
Phytol. 63, 165-80.
Smith, A. G. (1970) The influence of Mesolithic and Neolithic
man on British vegetation: A discussion, pp. 81-96 in ed. D.
Walker and West, R. G. Studies in the vegetational history of
the British Isles. Cambridge, C.U.P.
Thorn, R. (1975) Structural stability and morphogenesis: an outline
of a general theory of models. Reading, Benjamin.
Thornes, J. B. (1979) Structural instability and ephemeral channel
behaviour. L.S.E. Discussion Paper No. 72.
Turner, J. (1964) The anthropogenic factor in vegetational history
1. Tregaron and Whixall mosses. New Phytol. 63, 73-90.
Turner, J. (1965) A contribution to the history of forest clearance.
Proc. Roy. Soc. B161, 343-354.
Wagstaff, J. M. (1976) Some thoughts about geography and
catastrophe theory. Area 8, 319-310.
232 BURRIN AND SCAIFE
Wagstaff, J. M. (1978) A possible interpretation of seulement
pattern evolution in terms of catastrophe theory. Trans. Inst.
Brit. Geogr. New Series 3, 165-178.
Wilson, A. G. & M. J. Kirkby (1980) Mathematics for geographers
and planners. Oxford, O.U.P.
Woodcock, A. E. R. & T. Poston (1974) A geometrical study of the
elementary catastrophes. New York, Springer-Verlag.
Woodcock, A. E. R. & M. Davis (1978) Catastrophe theory.
London, Penguin.
Zahler, R. S. & H. J. Sussman (1977) Claims and accomplish-
ments of applied catastrophe theory. Nature 269, 759-763.
Zeeman, E. C. (1976) Catastrophe theory. Scientific American 234,
65-83.
Zeeman, E. C. (1977) Catastrophe theory: selected papers 1972-
1977. Addison-Wesley.
Zeeman, E. C. (1979) A geometrical model of ideologies, pp. 463-
497 in ed. C. Renfrew and K. L. Cooke Transformations:
Mathematical approaches to culture change. London and New
York, Academic Press.
Zeeman, E. C. (1980) Catastrophe models in administration, Ann.
Forum Proc. Ass. Inst. Res. 3, 9-24.
Zeeman, E. C. (1982) Decision making and evolution, pp. 325-436
in ed. A. C. Renfrew, M. J. Rowlands and B. A. Segraves,
Theory and explanation in archaeology. London and New
York, Academic Press.
 Cultural and Environmental Change During the
Palaeolithic: Research Methods Combining
Geomorphology With Prehistory

F. Hours and L. Copeland,

Maison De L'Orient, Lyons and Institute of Archaeology, London
One of the earliest concerns of prehistorians has been to find ways
of locating and dating human presence in a landscape by using
geological formations as a chronostratigraphic framework. It was in
this way that the successive stages of the Lower Palaeolithic were
gradually defined. The process started with the series collected in
Europe in the Somme, Seine and Thames river-terraces, and con-
tinued more recently in Africa where, for example it was again the
geomorphology which provided the stratigraphie framework for the
well-known site of Olduvai.
It was, however, soon realised that this approach to prehistory
lacked precision. To begin with, it concerned archaeological material
that had been re-worked in geological deposits by water. This meant
that the find-spot was not the original location of an artefact, i.e. the
place where it had been dropped by its owner. In turn, this made
impossible all attempts at palaeoethnological interpretation. On the
other hand, a geological context which reveals certain chronological
indications may have serious drawbacks. One example is when forma-
tions occur in close proximity to each other in river valleys, in the
form of river terraces, which are very difficult to seriate and correlate.
The discussion concerning the Chellean, and its final suppression as
an autonomous industry, is a good example of such a problem.
Moreover, it can happen that some deposits, especially fluviatile
formations, have been eroded from earlier formations which also
contained artefacts; this sometimes results in an inextricable mixture.
Finally, artefacts collected in this way are sometimes far from being
characteristic.
Links between prehistory and geology are not always regarded as
significant. A growing awareness of the chances of error has cast a
certain amount of discredit on anything that is not archaeologically
in situ, and it is obvious that living-floors, such as the one at Latamne
in Syria, or those at Melka Konture in Ethiopia, which have come
to us intact, sealed by deposits that guarantee the absence of all
contamination, provide more information than scattered pieces
embedded in the thickness of a terrace. It is also clear, however, that
234 HOURS AND COPELAND
living-floors are rare and that, to limit oneself to the study of these
alone means setting aside a priori the greater part of the prehistoric
material available, especially that which concerns the Lower and
Middle Palaeolithic. It is therefore important to find ways of com-
pensating for the risks indicated above, and to see under what
conditions it may be possible to retrieve data which, although difficult
to make use of, we have no right to ignore in order to reach a better
understanding of an age or a region. Reflections on the method to
be described here are to be found in Besanon (1979), Butzer (1980),
Wymer (1976).
Safeguards are necessary if one wants to identify and define the
geological context of the assemblages recovered, but conditions will
vary according to the nature of the formations. First observations
concerned the continental deposits, as it was soon realised that, in
certain favoured places, there was an abundance of artefacts in
geological positions. This, in fact, formed the basis of the thinking
of Boucher de Perthes (1847) at the dawn of Prehistory. Since then
the study of these deposits has made much progress. Distinctions
are now made between facies which are glacial, periglacial, fluviog-
lacial, fluviatile, pediments or aggraded terraces. These formations
may be climatic in origin, due to phases of rhexistasie (breakdown
of climatic balance) which has been the case more or less everywhere
in Europe, or of tectonic origin in unstable regions, such as along
the length of the Rift in Africa and the Near East. Although one can
recognise that the valleys have been filled, eroded and refilled, the
picture often becomes extremely complex and difficult to sort out
and amateurs would be well advised not to venture alone into this
imbroglio. Geomorphologists and geologists usually avoid it by divid-
ing terraces into 'low', 'middle' and 'high'. Sometimes they even add
a 'very low' or a 'very high' terrace to the series. Beyond the simplicity
of the description it serves little useful purpose, as the correlations
are very hazardous and there is no guarantee that the terraces named
'middle' or 'high' along the river are, in fact, covering the same entity.
Geomorphologists, by identifying formations through longitudinal
profiles rather than by cross-section, provide more useful indications
and their identification of the upstream and downstream succession
makes it possible to follow more closely the evolution of the topogra-
phy during the Pleistocene.
At one time, about forty years ago, it was thought possible to work
out a chronostratigraphic scheme covering an extensive region (such
as Europe, for example) by establishing the correlation between river
terraces and glacial moraines, following a four-cycle rhythm defined
by Penck and Bruckner for the northern Alps (Penck and Bruckner
1905). The project was taken up later under different names, in
northern Europe, in England and as far away as America.
 Cultural and Environmental Change 235
As early as the 1950's, temperature readings were being taken by
the 0 16/0 18 method on core samples obtained from deep-sea
borings. From these, and from numerous pollen analyses (even those
from ancient periods), as well as improved cl4 dating methods, from
thermoluminescence and other radiometric dating systems, we know
that the geomorphological chronometer does not record everything
and that there are likely to be gaps even in the most complete
succession of continental deposits (see Labeyrie 1984). For the pres-
ent one can only try, more modestly, to establish local sequences
and to endeavour to connect them with each other step by step,
which the evolution of the typology of the assemblages collected will
come to support.
One rather special aspect of these continental deposits concerns
lake beds which have lacustrine formations suitable for study. Sec-
tions, produced there by erosion, may be visible, there may be the
remains of shorelines along the lake fringes, or there may be the
possibility of obtaining sediment samples by coring the floor of the
lake. An advantage in studying lacustrine formations is that the
rhythm of sedimentation is slower than that along rivers and the
assemblages are likely to be less reworked, more complete and more
homogeneous. In temperate regions, however, lakes are often of
glacial origin and were formed at periods when their shores were
scarcely habitable.
On the other hand, the sites at Olduvai (Tanzania), Ubeidiyah
(Israel), Venosa (Italy), Ambrona and Torralba (Spain) owe the
quality of their assemblages in part to the lacustrine origin of the
deposits that contain them. Indeed the speed of burial in a lacrustrine
environment often enables fauna and pollen to be well preserved,
thus giving valuable indications about the environment, and elements
for dating which are sometimes corroborated by paleomagnetic
measurements recorded in clay as at Ubeidiyah and Olduvai.
Marine deposits set other problems. Shorelines or raised beaches
are not generally very thick and form fairly accurate pointers.
However, the difficulty of distinguishing between transgressive and
regressive ocean movements, the complex links between various sea
levels and the movement of continents tend sometimes to make
interpretations unreliable. The succession of Mediterranean sea-
levels: Calabrian, Sicilian, Milazzian, Tyrrhenian, though classic and
well known in its day, has been shown to be of no real significance.
As regards the Tyrrhenian, the Strombus beaches are not even con-
sistent from one end of the Mediterranean to the other: there are
several in the western Mediterranean, but there seems to be only one
on the Levantine coast. No explanation for these discrepancies has
been put forward.
236 HOURS AND COPELAND
However, despite all the difficulties that have just been pointed
out, the use of prehistoric material in a sound geological position,
even though it may not be in situ from an archaeological point of
view, does give irreplaceable information and provides a framework
both from the chronological viewpoint and from that of ecology and
environment.
As regards prehistory the contents of the formations are of two
orders. There are the remains of human activity that have been able
to withstand corrosion and abrasion stone tools and potsherds:
these are the artefacts. Then there are the ecofacts: fragments of
bone, pollen, and possible vegetable debris and seeds. The chrono-
logical value and the climatic indications of these ecofacts is undeni-
able but less accurate than one might expect. Finds are rare, and as
the dfinition of fauna and flora is made primarily by the grouping
of species the scarcity of the discoveries makes it difficult to establish
them. When there are pollen or rodent remains it is possible to be
precise, and the environmental variations are more easily discernible,
but this requires that suitable methods of collection are foreseen and
that the conditions of conservation, which taphonomy is beginning
to discover, should be taken into account. To envisage the study of
a formation which would include all the research suggested by
modern technology would require the presence of personnel trained
in detailed work methods: water sieving, flotation, and sorting under
binocular magnifying lenses. There is not always the time or the
means to undertake work programmed in such a way.
The artefacts themselves do, theoretically, provide some cultural
indications and should be able to tell us to which culture a population,
living at such and such a time in the Quaternary and in such and
such a region, would belong. For the periods of interest to us, this
concerns the durable remains of equipment; basically, the stone
tool-kits and, for later years, potsherds. Given the risks of contamina-
tion described above, all this material has to be assessed critically
to ascertain its homogeneity. The state of conservation can give some
indications: the degree of abrasion, mechanical wear, chemical
change, the flint patina, with all the uncertainties that accompany
such conclusions based on phenomena as yet imperfectly known and
difficult to evaluate without being too subjective (Shackley 1974,
Besanon 1982). One can, however, manage to sort out some series
whose significance becomes clearer by comparison with other criteria
derived from the technological and typological study of the artefacts.
Occasionally the term 'technology' is regarded as the taking of
simple measurements which have the merit that they can be carried
out by people who are not fully trained students for example
and which can be easily processed by a computer. In this way it is
possible to work out curves, statistical tables, dispersal diagrams and
 Cultural and Environmental Change 237
histograms which can help trace anomalies in an assemblage. This,
however, forms a very rudimentary conception of technology. Studies
concerning debitage and the technique of shaping by re-touching
are, without doubt, a more efficient means of checking the
homogeneity of an assemblage and of revealing extraneous
intrusions, Here again the results can be presented by various graphic
processes and provide a basis for eventual calculations. However,
these observations take a long time and, in practice, only a few criteria
are used to describe the forms of for example, the flake butts or
core-platforms.
There remains, finally, typology which classifies assemblages by
noting the proportions of various types of one particular form of
artefact. It is clear that we now increasingly expect to find Oldowayan
pebble-tools in Early Pleistocene formations, and Neolithic polished
axes in the Holocene terraces, and Oldowayan is usually quite easily
distinguishable from Neolithic. The use of typology as a classifying
criterion remains, however, subject to many conditions if the
classification is to be reliable. One must first remember that 'type
fossils' are rare, and some people maintain that there are none.
Franois Bordes, who fought so hard to promote typology, strongly
stressed the importance of the proportions between the various types
in order to get an overall view of the assemblage and, so that not
too much importance be attached to isolated pieces, he insisted that
only a consistent whole could give valid information. This leads to
the fundamental idea of the statistically representative sample. Taking
these precautions into account, it is possible to identify an assemblage
with a fair degree of assurance and to place it within the prehistoric
sequence of stages: Archaic Palaeolithic, Acheulean, Middle and
Upper Palaeolithic, Epi-Palaeolithic, Neolithic. In the same way it
is also possible to place a succession of geomorphological episodes
in the course of the Pleistocene or even, at a stretch, to date a simple
residual, isolated formation. There are those, however, who refuse
to make use of typological criteria on the grounds that the definition
of types is based on an priori which is too empirical. Discussions
about this have been very lively but, as with all extreme views, this
one is losing its credibility. It is, in the end, reasonable to take the
typological argument into consideration.
It should be added that the need to work on a sufficiently large
collection has value beyond its validity for typological identification.
If one wants to check the homogeneity of an assemblage (by seeing
whether series separated by differences in the degree of abrasion,
patina, or deterioration are consistent), one must have collected a
'statistical sample', i.e. a large enough number of items in each
category for the conclusions drawn to be plausible. This entails
spending sufficient time prospecting and inspecting sections, and
238 HOURS AND COPELAND
here the work of the prehistorian is slower than that of the geomor-
phologist. The longer the prehistorian can linger in front of a cut,
the greater his chance of finding material on which to base his
interpretations, whereas the more the geomorphologist moves around
the greater his chance of finding evidence of different formations
which will enable him the better to understand the landscape. This
mutual adaptation to dissimilar rhythms of work is not always easy.
There are, therefore, ways not only of compensating for the incon-
veniences that have been mentioned, but of checking up on the
information by combining the data concerning the position of the
formations and their prehistoric contents. It is also necessary that it
should be possible to apply this combined method of work.
The terrain must fulfill certain requirements. An isolated, discon-
tinuous formation does not readily lend itself to identifications and
comparisons. It is the ensembles, or suites, which tend to produce
results. In the continental domain it has been the complex series of
terraces that have made famous the Somme and the Thames in
Europe, and the Orontes and Euphrates in the Middle East. The
successive levels of the fossil beaches at Lake Tiberias made it
possible to date the sites at Ein Gev. All this arose because the
geological formations contained archaeological material. It can hap-
pen, however, that fine terraces like those in the Rhone valley, well
identified and well studied, turn out to be sterile. On the other hand,
when formations have been completely wiped out by erosion, the
chronological assessment of the industries which have survived some-
times becomes problematical, as has been the case for some time
with the Acheulean at Joubb Jannine in the Lebanese Bekaa.
In the maritime domain favourable conditions are also necessary
to achieve satisfactory results. Under good conditions, in places
where the fossil raised beaches have no deformities, one can study
an impressive series such as that at Ras Beirut or on the coast of
Morocco. Where deformities do occur interpretation becomes more
difficult as in Tunisia, where the significance of various episodes has
altered several times (Paskoff and Sanlaville 1983). It sometimes
happens, also, that a good sequence may be sterile, as in Greece.
As regards geomorphology, the best possible situation is, finally,
the one where intercalations of various formations are found: i.e.
marine and fluviatile, fluviatile and lacustrine. It is there that stratigra-
phy is most complete and most reliable, and when prehistoric material
is found there in association with geomorphology, one comes across
fabulous sites like the one at Ubeidayah in Israel (Goren 1981), or
the mouth of the Nahr el Kebir in northern Syria (Sanlaville 1983).
From the point of view of the stone tools, also, conditions must
lend themselves to the combined method. The first condition, as
emphasised already, is that there should be sufficient material. There
 Cultural and Environmental Change 239
is in general nothing to be learned from an isolated piece. Unfortu-
nately, geomorphologists have a tendency, with the idea perhaps of
throwing a bone to the prehistorians who may not be able to follow
them over the whole terrain, to come back triumphantly bearing two
or three pebble tools picked up in the course of their investigations,
hoping that they can be dated. This is unrealistic. If research is to
be multi-disciplinary, the rules should also apply to fieldwork and
can only be carried out on numerically viable samples.
When conditions exist for the combined method to be used, the
results are interesting though different in each case, and one should
not expect more than the site situation allows. The most promising
example is, of course, that of assemblages found in a controlled
geological position and that have not been disturbed. These are sites
which may have been used as a setting for various occupations: i.e.
kill-sites, temporary hunting camps, or permanent habitation of fairly
lengthy duration. It can then be said that the objects are archaeologi-
cally and geologically in situ. The homogeneity of the material is
guaranteed as one can be certain that there has been no mixing.
These are the standards to be applied when it comes to making a
typological assessment of other assemblages needing greater caution.
Moreover, thanks to geomorphology the security of their position
makes it possible to place them in a relative chronology. The only
drawback is that they are so rare. In fact, apart from the African
sites such as Olduvai or Melka Konture, the known inhabited sites
in the rest of the world and dating from the lower Palaeolithic can
be counted on the fingers of both hands. This is where the comment
made earlier can be appreciated. A study of prehistory limited to
this type of site would mean the rejection of the greater part of the
material available.
Sites in a marine or, more often, in a continental formation where
assemblages are compacted and reworked, are much more common.
One can say they are in situ geologically but not archaeologically.
The stratigraphie position of the artefacts does provide some basis
for dating; i.e. they cannot be more recent than the formation in
which they are contained. But this is where the control factor should
come into play to verify the homogeneity of the assemblage and to
discount the possibilities of pollution. Even the position of the
formation should be critically assessed because, whether the terrace
is fluviatile or a beach, it consists of material deposited there by
water; humans would not live among pebbles churned up or flooded
by a river. We are therefore faced with information whose constituent
parts are all 'floating'. However, by making use of convergent criteria
it is possible to reduce the margin of inaccuracy provided that here,
as in the case of typology, there are a sufficient number of sites and
samples at our disposal. It would certainly not be possible to know
240 HOURS AND COPELAND
how people lived when their tools have been recovered mixed with
the pebbles of a terrace, and paleoethnology cannot operate; but it
would be possible (by exploring the whole region) to know when it
was occupied and what had been the evolutionary stages of the
industries.
It should be added as a reminder that information provided by
surface sites, which have no link with stratigraphy, is even weaker
although in certain cases it is possible to assess the assemblages and
to establish a certain degree of homogeneity. But that is very much
a question of luck.
By applying the criteria outlined above, i.e. criteria for assessing
the physical condition of the artefacts based on wear, abrasion,
deterioration, patina, and criteria to evaluate the technological and
typological characteristics of the assemblages, it is possible eventually
to re-group ensembles that are consistent, and that are numerically
valid. By working together in the field, geomorphologists and prehis-
torians should make it possible to establish links between geological
deposits, topographic formations and prehistoric assemblages and,
using as much of this data as possible, to construct a relative
chronology. Given favourable conditions, by making comparisons
with well dated industries, meaningful results can be achieved.
Results can be considered on several levels, and the examples we
shall give are obviously influenced by the location of our own work,
i.e. in East Africa and the Middle East. For the Middle East, informa-
tion concerning the results of the combined work of geomorphology
and prehistory is to be found in Bar Yosef 1975, Kirkby and Kirkby
1976, Sanlaville 1984, Vita Finzi 1968.
In the case of an isolated site the research method combining
prehistory and geomorphology does not give particularly accurate
information. In the case of a chronologically well-placed formation,
the industry it contains is also dated grosso modo. In Syria, the
well-known site of Latamne on the Orontes can thus be attributed
to the Middle Pleistocene; and the assemblage of Middle Palaeolo-
lithic aspect at Biache Saint Vaast in northern France is contained
in a formation thought to be 'Rissian'. Conversely an assemblage
that is well characterised from a typological point of view can help
to place in time a geological formation whose position may be difficult
to interpret.
In the Nahr el Kebir region of northern Syria a formation that
had at first been thought to date from the Lower Pleistocene turned
out to contain Middle Palaeolithic: more detailed research has led
to the discovery of a valley-fill that had not previously been noticed.
Similarly, on the banks of the Zarqa in northern Jordan, a formation
thought to be Pliocene contained flint flakes which appeared to have
been knapped: this necessitated a difficult reappraisal. All this
 Cultural and Environmental Change 241
assumes that the assemblages have been critically assessed and their
homogeneity established. It is above all necessary to have some idea
as to the contemporaneity of the formation with the lithic assemblages
it contains. In any case an isolated site is only a dot on the map and
does not teach us much.
It is quite different when one studies a more extensive territory.
The formations and assemblages contained there, distributed in a
defined area for example along a 50 km stretch of river allow
for a detailed local chronostratigraphy to be drawn up, and for the
definition of local typological facies. In Syria a fairly complete
succession of Lower Palaeolithic stages of development were located
on the Nahr el Kebir and the Orontes. Similarly, the Middle
Acheulean of the Orontes, already known through the site at Latamne,
showed itself to be different from the one on the Levantine coast,
identified at Ras Beirut, at the Wadi Aabet, and at the mouth of the
Nahr el Kebir. To the east of Lake Turkana in Kenya, J.W.R. Harris
has been able to establish a sequence which gives a connection
between the Pebble Culture and the Acheulean, and at the same time
shows up the existence of an original facies of the Lower Palaeolithic.
All this provides the elements for creating a synthesis on a much
wider scale. By considering data which allow block by block com-
parison to be made, the relationshps of the assemblages and sequen-
ces are less uncertain than would be the case where it was desired
to 'slot in' a single, isolated, site.
This method does in fact reveal its full potential when one moves
from a micro-region to a macro-region, as for example from the
valleys of the Orontes, the Euphrates, the Nahr el Kebir, the Litani,
the Zarqa, to the whole of the Levant. Local sequences of geological
formations and prehistoric assemblages can be completed and corre-
lated to form a fairly comprehensive picture of the populating of a
whole country. It is here that distribution maps are helpful because
dots on a map, which have no meaning when they are isolated,
become significant when they are assembled, and show the great axes
of human penetration in a region. The example of the Levant,
described earlier, demonstrates this particularly well (Copeland 1981,
Copeland and Hours 1981, Hours 1981). Work carried out there
during the past ten years by geomorphologists has shown that there
is a group of four fairly complex terraces to be found along the main
rivers (Besanon et al. 1980). At some points, as at Beyrouth or at
the mouth of the Nahr el Kebir, they overlap with the ancient
shorelines. We thus possess a sequence of marine and continental
formations covering the whole Quaternary, from Early Pleistocene
down to Holocene from the north of Jordan to the south of Turkey.
All these deposits contain industries. The oldest have yielded fairly
uncharacteristic flint flakes on the Orontes and on the coast, as well
242 HOURS AND COPELAND
as an Acheulean of primitive aspect near Lattakiah. The most recent
include potsherds attributable to the Bronze Age, mingled with Neo-
lithic. Thus one can follow the whole evolution of the Acheulean:
Early Acheulean was succeeded by Middle Acheulean, known by
two facies, one with spheroids and trihedral picks, found along the
rift valley course of the Orontes and the Litani; the other, more
classical with wide, thick hand-axes found on the Mediterrean coast.
After this, one can distinguish a Late Acheulean, an Evolved Late
Acheulean rich in amygdaloids and cordiforms, and a Final
Acheulean. The fine distinctions of this evolution were only dis-
covered through collections made in well-established geological posi-
tions. Ten years ago one could only distinguish Early-Middle
Acheulean and Late Acheulean (Hours, Copeland and Aurenche
1973; Tomsky 1982).
Beyond this improved understanding of the evolution of tools it
is interesting to note that Early and Middle Acheulean is only found
on the coast and in the central rift as far as southern Turkey. The
interior of the land, the Syro-Jordanian steppe, seems to have been
occupied only from the Late Acheulean. The same Late Acheulean
is found from southern Turkey and northern Iraq as far as Arabia.
If one takes into account the latest datings which place the Final
Acheulean and Yabrudian at around 150,000, the complete spatio-
temporal framework of Levantine Lower Palaeolithic is thus put in
place thanks to the combining of prehistory and geomorphology.
There are further examples which could be given of the success of
this spatial archaeology on a geological foundation which has made
it possible for some of the problems, real or imagined, to be solved,
even in classic regions like the Perigord (Laville, Rigaud and Sacket
1980, White 1985).

One should not exaggerate the scope of this method. Geomorphologi-

cal prospecting and a verified stratigraphie collection will never give
the same results as those obtained from a dig on a well-preserved
living-floor. We have pointed out the disadvantages and the lacunae
in our method of working. Sequences are rarely complete, mixing is
always a possibility and difficult to sort out. The thorny question of
contemporaneity of assemblages and formations does not always find
satisfactory arswers. The interpretation of the placing of deposits is
not without ambiguity because the movement of sand, clay and gravel
can arise equally well from deforestation (dryness) or sheetwash
(humidity). The result of all these difficulties is a fairly loose
chronology whose two constituents, geomorphology and typology,
cannot be correlated with absolute accuracy. Moreover, paleoethno-
logical observations are impossible.
 Cultural and Environmental Change 243
Having said all this, the fact remains that the combined method
of geomorphological and prehistoric surveying is the only one which
allows difficult information to be recovered. To set aside everything
that is not archaeologically in situ would give an extremely false
picture of prehistory and especially of the Lower Palaeolithic. On
the other hand, this way of working does open up a spatial
archaeology that provides an overall view and which gives some
perception of the stages and scale of the populating of a whole region.
There is no other method, it seems, that offers the same advantages.
From the conceptual point of view it is, however, necessary to take
some precautions to ensure the validity of the procedure.
Geomorphological analysis is not feasible everywhere, the terrain
must be of a suitable kind. Regions that are too dry, where desert
conditions have prevailed for long periods, especially inland in places
where there is a tendency to develop landlocked basins, do not
present clearly denned geomorphological features such as buried fills
or raised terraces. Information therefore is scarce and of little relev-
ance. Regions too thickly covered with vegetation (Equatorial Africa)
cannot easily be explored, and when human occupation is too dense
(Europe) unheavals of anthropological origin distort the landscape
to the point of making it unrecognisable. On the prehistoric level it
is necessary that the assemblages themselves should be reliable. A
few isolated pebbles mean nothing. To date a formation there must
be a statistically representative sample that is as large as possible in
order to eliminate errors of interpretation due to aberrent or intrusive
pieces or to individual variations.
There remains a final obstacle which is of a practical rather than
a conceptual nature, and which has already been indicated. The
different methods used by geomorphologists and prehistorians, which
makes their work complementary, also makes their co-operation
somewhat difficult. This, no doubt, is why these inter-disciplinary
enterprises are rare, and all the more valuable. One,, can only wish
most devoutly that they would increase because, for a certain category
of material, there is no other way.
BIBLIOGRAPHY
Bar Yosef, O. (1975) Archaeological occurrences in the Middle
Pleistocene of Isreal, in Butzer K. and Isaac G. (eds) After
the Australopithecines: Stratigraphy, Ecology and Culture
Change in the Middle Pleistocene. New York, Aldine.
Besanon, J. (1979) Prhistoire et Gomorphologie. Quelques,
rflexions d'ordre mthodologique, in Sanlaville P. (ed)
Quaternaire et Prhistoire du Nahr el Kabir septentrional, pp.
121-135. Lyon, Maison de l'Orient.
Besanon, J. (1982) Les figures d'altration des silex au Liban et
leurs implications, in Archologie au Levant; Recueil Roger
Saidah, pp. 21-31, Lyon, Maison de L'Orient.
244 HOURS AND COPELAND
Besanon, J., L. Copeland, F. Hours, S. Muhesen, & P. Sanlaville,
(1980) Gomorphologie et Prhistoire de la valle moyenne
de l'Euphrate. Essai de chronologie du Pleistocene et du
Palolithique de Syrie. Comptes rendus de l'Acadmie des
Sciences de Paris 290, serie D, 167-170.
Boucher de Perthes, J. (1847) Antiquits celtiques et diluviennes.
Mmoire sur l'industrie primitive et les arts leur origine. T.I.
Butzer, K. (1980) Context in Archaeology: an alternative Perspec-
tive. Journal of Field Archaeology 7, 417-422.
Copeland, L. (1981) Chronology and distribution of the Middle
Paleolithic as known in 1980, in Lebanon and Syria, in
Cauvin J. and Sanlaville, P. (eds) Prhistoire du Levant, pp.
239-263, Paris, C.N.R.S.
Copeland, L. & F. Hours, (1981) La fin de l'Acheulen et
l'avnement du Palolithique moyen en Syrie, in Cauvin J.
and Sanlaville, P. (eds) Prhistoire du Levant, pp. 225-238,
Paris, C.N.R.S.
Goren, N. (1981) The lithic assemblages of the site of Ubeidiyah,
Jordan valley. Ph.D. Thesis. Jerusalem, Hebrew University.
Hours, F. (1981) Le Palolithique infrieur de la Syrie et du
Liban. Le point de la question en 1980, in Cauvin, J., and
Sanlaville, P. (eds) Prhistoire du Levant, pp. 165-183. Paris,
C.N.R.S.
Hours, F., L. Copeland, & O. Aurenche, (1973) Les industries
palolithiques du Proche-Orient, essai de corrlation.
L'Anthropologie 77, 229-280 and 437-496.
Kirkby, A. & M. J. Kirkby (1976) Geomorphic processes and th
Surface Survey of Archaeological Sites in semi arid areas, in
Davidson, D. A. and M. L. Shackley (eds) Geoarchaeology:
Earth Science and the Past, pp. 229-254, London, Duckworth.
Labeyrie, J. (1984) Le cadre paloclimatique depuis 140,000 ans,
L'Anthropologie 88, 19-48.
Laville, H., J. P. Rigaud & J. Sacket (1980) Rockshelters of the
Perigord. Geological stratigraphy and archaeological suc-
cession. New York, Academic Press.
Paskoff, R. & P. Sanlaville (1983) Les ctes de la Tunisie. Vari-
ations du niveau de la mer depuis le Tyrrhenien. Lyon, Maison
de l'Orient.
Penck, A. & E. Bruckner (1905) Die Alpen in Eiszeitalter.
Sanlaville, P. (ed) (1979) Quaternaire et Prhistoire du Nahr el
Kebir septentrional. Paris, C.N.R.S. et Maison de l'Orient.
Sanlaville, O. (1984) Gographie et Archologie. Le Courrier du
C.N.R.S. Supplment no. 57, pp. 52-56.
Shackley, M. L. (1974) Stream abrasion of flint implements.
Nature 248, 501-502.
Tomsky, J. (1982) Das Altpalolithikum im Vorderen Orient.
Beihefte zum Tbinger Atlas des vorderen Orients. Reihe B,
no. 18. Wiesbaden, Ludwig Reichert.
Vita Finzi, C. (1968) The Mediterranean Valleys. Cambridge, Cam-
bridge University Press.
White, R. (1985) Upper Palaeolithic Land Use in Perigord. A Topo-
graphie Approach to Subsistence and Settlement. Oxford, BAR
S253.
Wymer, J.J. (1976) The interpretation of Palaeolithic cultural and
 Cultural and Environmental Change 245
faunal material found in Pleistocene sediments, in Davidson,
D. A. and M. L. Shackley, (eds) Geoarchaeology: Earth
Science and the Past, pp. 327-334. London, Duckworth.
 Palynology and Human Ecology: Theory and Method for
the Study of Prehistoric Agriculture

Stanton W. Green,
University of South Carolina, U.S.A.

The spread of farming into new habitats raises a classic problem in

human ecology: the apparently simple processes of clearing and
planting fields set off a complex series of ecological responses and
cultural adjustments. The anthropological study of farming, however,
has been heavily weighted to the cultural side. Much attention has
been given to the making of agricultural decisions (Bartlett 1981,
Earle and Christensen 1980, Boserup 1965) without systematic study
of the dynamics of agro-ecosystems. Agricultural models treat farmers
as optimal decision-makers who attempt to minimise their costs and
risks, but the ecological responses to these decisions are rarely sorted
out. To anthropological archaeologists concerned with the origins
and development of farming, such a short term and limited perspec-
tive provides no more than a starting .point, and calls for theory and
method designed to broaden and lengthen our interpretative models
(Green 1980a,b).
A research strategy to investigate the longterm dynamics of agro-
ecosystems requires a linkage of ecological theory (Margalef 1968)
and palynological method. First, ecological theory must guide us to
a working knowledge of how and why ecosystems change in their
natural states, and then in culturally-induced states. The working
models must then be tested against palaeoecological data in ways
capable of reflecting ecological relationships and change.
Such a perspective requires a problem-directed approach to
palaeoecology which draws upon the study of vegetational history
in ways that treat the natural environment as a set of ecological
communities. Palaeoecological interpretations need to do more than
describe the rise and fall of particular species, moving toward a
systemic understanding of ecological communities with species as
their interactive components. Community process as well as environ-
mental factors affect forest composition. Because of this, vegetational
change as expressed in palaeoecological data can only be understood
through the application of ecological theory and modelling. Patterns
of pollen cannot simply be described; they must be interpreted within
a problem-directed linkage of theory, method and data.
 Palynology and Human Ecology 247
This paper addresses these conceptual issues with specific regard
to the agricultural prehistory of temperate Europe. It begins with an
examination of the models and assumptions of palaeoecology to
make clear the necessity for problem-directed research. This is fol-
lowed by a general discussion of agricultural ecology with specific
development of an operational model of ecological succession. Next,
a general statistical methodology is outlined to put into effect this
ecosystemic model for testing against palynological data. Finally, the
effectiveness of this approach is demonstrated through the interpreta-
tion of vegetational change during the spread of farming into Den-
mark during the early Neolithic.

Palynology, of course, has been a long-term partner of archaeological

and historical research (Godwin 1956, Iversen 1937,1941, 1949,1960,
Troels-Smith 1960, 1966). Its objective has been largely descriptive,
focussing on reconstructing palaeo-environments rather than inter-
preting ecological change. For description, the recognition of patterns
is often enough. For a given time-period one can, for example,
enumerate the relative proportions of plant species. Studies of change,
however, call for models of process. A description of the elm decline
in Europe during the Atlantic-Subboreal climatic epoch, for example,
is not completely satisfying to European prehistorians because it
leaves unaddressed the questions of why and how this ecological
change occurred. Ecological theory and modelling, when coupled
with appropriate statistical methodologies aimed at distilling patterns
from pollen data, can provide the ingredients for addressing such
processual questions. To begin with, the assumptions and limitations
of palynology must be made explicit and their implications for
archaeology made clear.
Normal palynological science aims to reconstruct vegetational
variability through time and over space with the ultimate objective
of interpreting both synecological (community) and autecological
(species-specific) relationships. Archaeologists use these reconstruc-
tions as aids for interpreting the prehistoric and historic record. Site
catchment analysis, for example, utilises palynological studies to
reconstruct the floral environs of a site as part of an evaluation of
resource availability. Human impact studies take on a diachronic
perspective by attempting to reconstruct the environmental effects of
human subsistence activities. These types of research are based on
the fundamental assumptions that the pollen record provides a large,
representative sample of past vegetation and that the employment
of appropriate field and laboratory methods can recover representa-
tive samples of the pollen deposits (Faegri and Iversen 1975).
The links between pollen and vegetation, however, are not so
simple. The pollen record is the result of a complex linkage of
 248 GREEN
pollination ecology and site formation processes. As in archaeology,
therefore, unless the processes that generate these patterns are
explicitly modelled, no one account of the pollen record can be
granted more credence than another. With a problem-directed
approach well founded hypotheses can be formulated, tested and
finally evaluated within a logic that incorporates behavioural models
of the social and ecological systems in question. Hypotheses concern-
ing the behaviour of ecosystems can be tried out through models of
community succession and plant demography, while pollination
ecology and site formation models can be used to control depositional
processes. Appropriate statistical and other methods of analysis can
then be designed and applied to test rigorously these hypotheses,
using the rich data of the pollen record. When coupled with
behavioural models of agricultural decision-making (Barlett 1982,
Maclachlan 1984) we can begin to sort out natural and culturally-
induced vegetation change as expressed in the archaeological and
palaeoecological record. The research methodology I am suggesting
links the study of agricultural ecology with pollination ecology in
ways that allow for testing against the palynological record. Let us
begin with a discussion of agricultural ecology.

AGRICULTURAL ECOLOGY
The dynamic of agro-ecosystems derives from a fundamental conflict
between agricultural and ecological systems: agricultural systems aim
to increase productivity, while ecological systems mature in the
direction of stability and low productivity (Odum 1969, Watt 1973).
The principles of ecosystem development have an important bearing
on the relationship between man and nature because the strategy of
'maximum protection' (that is, trying to achieve maximum support
of complex biomass structure), which characterise ecological
development, often conflicts with man's goal of 'maximum produc-
tion' (trying to obtain the highest yield) (Odum 1971, 267).
Societies regulate ecological systems, through such actions as clear-
ing and cultivation, in order to increase a habitat's productivity. This
maintains the ecosystem in immature, fast-growing states. The food
energy derived from these practices comes at the cost of maintaining
the ecosystem at these simpler states of succession. In a conceptual
sense the ecosystem is pushed back in 'successional time', resulting
in decreases in biomass and its associated respiration costs, and
increases in net productivity.
The relatively simple structure of agro-ecosystems makes them
more vulnerable to environmental change, and farmers incur both
the costs of maintaining ecosystems at simpler states (through weed-
ing, for example), and buffering these systems against environmental

_
 Palynology and Human Ecology 249
exigencies, e.g. through irrigation and drainage. As the intensity of
farming practices increases, these tasks are added to by long-term
needs to maintain the nutrient balance and structure of the soils.
Agricultural systems vary widely in their diversity and therefore
their vulnerability. The extremes of this variation are probably rep-
resented by slash-and-burn tropical fields with their tens of species
(Conklin 1961), and the monoculture fields of modern temperate
agriculture. In all cases, decreased diversity in numbers of species
and the horizontal and vertical patterning of a habitat provides a
more homogeneous structure for pests and foraging animals, thereby
increasing the possibility of outbreaks. The farmer, therefore, must
bear the cost of preventing such potential disasters.
Two interrelated questions become important as a result of thinking
about agriculture in this way: first, how much energy does it take to
manage an ecosystem and second, how will an ecosystem respond
to such management in the long run? One operational way to address
these questions is to measure the effects of agriculture on ecological
succession.
During the course of ecological succession the processes of produc-
tivity and biomass accumulation follow logistic curves; that is, they
increase at increasingly smaller intervals until their rate of change
approaches zero. The difference between gross productivity and
respiration approaches zero as the ecosystem matures, so that a
relatively constant state of limited growth is eventually reached. By
'pushing succession back', agricultural practices attempt to recreate
the earlier immature stages of higher net productivity. A comparison
of a variety of natural ecosystems with an alfalfa field illustrates
these relationships (Table 1). The most mature system in this table,

Table 1. Annual Production and Respiration (Kcal/m 2 /yr) in Three Ecosystems

Oak/ Pine
Alfalfa Medium Mature
Field Age Rain Forest

Gross Primary Productivity (GPP) 24,000 11,500 45,000

Autotrophic Respiration 9,200 6,400 32,000
Net Primary Productivity (NPP) 15,200 5,000 13,000
Heterotrophic Respiration 800 4,600 13,000
Net Community Productivity (NCP) 14,400 2,000 little or
none
Ratio NPP/GPP(%) 62.3 43.5 28.9
Ratio N C P / G P P ( % ) 59.0 17.4 0.0

Note the decreased net to gross productivity ratios as one shifts from simple (alfalfa)
to complex (mature rain forest) ecosystems (after Odum 1971, p. 46).
250 GREEN
the tropical rainforest, has both the highest gross primary production
and the lowest net-to-gross production ratio.
This is the result of the high respiration cost of maintaining the
tropical forest's complex biomass structure. The alfalfa field reaches
only approximately half of the primary productivity of the tropical
forest, but produces seven times the net community productivity,
and over twice the net-to-gross productivity ratio of the tropical forest
(Odum 1971, 46-8).
Given the agricultural goal of creating more productive habitats,
the important long-term factor becomes the ability of an ecosystem
to absorb or recover from the disruptions necessary to maintain
young, simple ecological states. This seems to be determined by an
ecosystem's stability and resilience. Stability refers to the ability to
absorb an impact without structural change; while resilience signifies
the ability to recover from impacts that cause structural change.
The more complex an ecosystem is the more difficult it is to change
its basic structure. Tropical forests, for example, are composed of a
long-lived and complex biomass that is not easily modified, and small
clearances do not threaten the integrity of their complex structure.
However, if the structure of the forest is significantly modified it is
not likely to recover. Long-lived species naturally have low turnover-
rates, and complex structures are very difficult to rebuild. As a result,
complex habitats are often unable to recover from significant disrup-
tion. The most effective long-term agricultural strategy for a stable
human ecosystem within such a situation would therefore be exten-
sive, low-impact cultural management. In this way, small increases
in the ecosystem's productivity can be gained without disrupting the
system. The pattern of tropical swidden agriculture clearly fits this
pattern.
Very simple ecosystems, such as tundra habitats, are both unstable
and non-resilient. Relatively minor cultural modification or environ-
mental fluctuation can restructure the habitat to the point of a system's
collapse. Only very extensive exploitation strategies, such as foraging,
can be long-lasting in these types of habitat.
Habitats of intermediate complexity, such as temperate forests,
tend to be more difficult to model and interpret. Since they are less
complex than tropical forests they are easier to disturb. At the same
time, the faster growing tree species and more complex soil structure
of the temperate forest increase its ability to recover from significant
modifications, such as clear-cutting. Temperate forests, in other
words, tend to be quite resilient, and this is a significant point with
regard to the history of agriculture in Europe.
Succession theory helps us to sort out the types of variables and
processes that underlie the dynamic ecology between farming and
natural environmental processes. The next task is to outline a model
 Palynology and Human Ecology 251
of how forest communities actually operate, in order to model the
effects of farming on temperate forests.

MODELS OF SUCCESSION
Ultimately we would like to be able to predict and observe natural
and culturally-induced change in temperate forests as reflected in
the palaeoecological record. Operational models that apply general
succession theory to the measurement of habitat change range from
formal simulation models to specific descriptions of the mechanisms
of succession. Some of these models illustrate the operation of suc-
cession theory through the explication of community and plant
behaviour. The adaptive strategies of plants are described in terms
of their reproductive processes, shade tolerance and soil require-
ments. The successional pattern of a particular forest can be predicted
or modelled through a study of these variables for the species ofthat
forest community. The community response to farming can then be
studied by simulating the effects of particular farming practices, such
as fallowing, on the factors that effect succession. As we shall see,
these predictions can then be tested through the statistical analysis
of pollen data.
The first step is to develop a general, but operational, model of
forest succession. One such model which has wide application is
based on species replacement. In the simplest model each plant has
a certain probability of being replaced by another of its kind, or by
another species. A successional sequence is modelled by repeated
multiplications of an initial distribution by a matrix of these prob-
abilities (Horn 1974, 27; 1975).
The sequential states of a community can be simulated, using this
approach, by predicting the likelihood of the replacement of one
species of tree by another; for example an oak tree succeeding a
birch. The probability of replacement varies according to both com-
munity and environmental factors. For example, the relative inability
of birch to live under the canopy of oak, along with the birch's
relatively short life span, results in its inability to compete with oak
for a place in a given habitat. All other things being equal, the
likelihood of an oak tree replacing birch would be high. Whether or
not oak succeeds birch within a given habitat, however, depends
partly on the environmental conditions at a site, such as drainage and
soil acidity. In certain situations it is conceivable that short lived
colonizers such as birch could out-compete climax species such as oak.
Several computer simulation models use species replacement
models to predict successional patterns (Horn 1975a, b; Leak 1970;
Likens et . 1972; Shugart et . 1973; Siccama 1969). Although a
variety of factors are used, all of these simulation models funda-
mentally characterise species on the basis of their colonizing ability.
252 GREEN
Several of them take environmental requirements into account, to a
greater or lesser degree.
For example, the model of Botkin and others (Likens et al. 1972;
Siccama et al. 1969) uses a large variety of factors to characterise
their tree species, including both behavioural characteristics and
environmental tolerences. These characteristics are matched against
environmental variables, and succession is simulated through appli-
cations of the species attributes to the environmental constraints.
Leak (1970) creates a simulation on the basis of birth and death
rates and forest density as it effects these rates. He derives them from
empirical observation. The difference between this and Botkin's
model is that Leak uses death and birth rates as indicators of species
survival and growth, whereas Botkin and his colleagues are attempt-
ing to examine the factors underlying these differential fertility and
mortality rates. Depending upon the problem at hand, both can be
effective approaches. Both attempt to simulate succession by classify-
ing species in terms of attributes congruent with our general model
of succession; that is, properties relating to productivity (birth rate,
growth), biomass (size) and lifespan.
Plant demography offers an additional angle for understanding
community succession at the species level (Harper and White 1974).
Demography also views plants as adaptive to both natural constraints
of the environment and the changing ecological states that occur
during community succession. Plants adapt to these conditions
through several mechanisms, including seed dispersal, growth and
maturation rates, longevity, size, form and tolerance of environmental
conditions (Harper and White 1974). Two basic strategies emerge
from these sets of characteristics: colonizers tend to disperse pro-
lifically and indiscriminately and to grow quickly and be short-lived;
climax species disperse carefully, grow slowly and live long. Describ-
ing species in these demographic terms also helps to develop models
of community change.
In sum, succession theory and plant demography clearly set out
the variables that underlie community change. Our next step is to
design a methodology to apply these concepts to the palaeoecological
record. For this let us turn to a study of the spread of farming into
Denmark during the Neolithic period.

A MODEL OF THE DANISH FOREST DURING THE ATLANTIC

PERIOD
Farming spread to Denmark during the Atlantic period (Iversen
1975), dated to the fifth millenium BC. The forest at this time was
typical temperate climax forest that included such species as linden,
oak, elm, hazel, birch, willow, alder and ash (Iversen 1975). Our first
methodological task is to develop a model of this forest as an
ecological community. Guided by successional theory we can begin
 Palynology and Human Ecology 253
to do this by rank-ordering each arboreal species with regard to its
essential environmental requirements and its position on a colonizing-
climax continuum of community change. As we saw above, the
adaptive characteristics of plant species can be described in a variety
of ways. For the purposes of this application I have chosen to focus
on three basic variables: lifespan, maximum height and relative shade
tolrance. Lifespan and size indirectly measure the birth and death
rates (that is, the turnover rate) of a species. Long-lived, larger species,
for example, generally have lower dispersal and germination rates.
Shade tolerance measures the competitive ability of a species; shade
tolerant species generally crowd out light-requiring species. When
we combine these three variables we can begin to rank-order the
ability of a species to colonize a habitat and to maintain itself in
face of competition from other arboreal species. This allows us to
predict the relative likelihood of one species replacing another under
a given set of environmental conditions.
The most important species of the Danish Atlantic forest are listed
and ranked according to this continuum as the first step in modelling
the forest community (Tables 2, 3 and 4). This relative ranking
schedule, along with information on the specific environmental
requirements of each species, is used as a basis for predicting forest
histories, including those that occur as the result of farming dis-
turbance.

Table 2. Ranking of Arboreal Species for Size and Light

Tolerance

Species Size' Tolerance"

Alder 2 2
Ash 2 2
Beech 1 1
Birch 2 3
Elm 1 1
Hazel 3 2
Hornbeam 2 1
Linden (Basswood) 1 1
Oak 1 2-3
Pine 2 3
Willow 3 3
1
Size Classes
1 greater than 25 meters (full grown)
2 between 10 and 25 meters
3 less than 10 meters
Tolerance Classes
1 High tolerance of shade
2 Medium tolerance of shade
3 Low tolerance of shade
254 GREEN
Table 3. Rank Classes of Colonizing Ability
Class Ranking Species

I 1 Elm, Linden
2 Oak
II 3 Alder1, Ash 2
III 4 Hazel3, Pine, Birch
1
Alder is very sensitive to water level and requires
moist soil; and even tolerates flooded land.
2
Ash is considered part of the Mixed Hardwood
Forest by several botanists (Vuorela 1972, p. 14); but
here is considered secondary to Oak, Elm and Linden.
3
Hazel, though a pioneer species, is slower in
colonizing than Birch and will shade out Birch seed-
lings. It might be expected to compete with Birch as
a pioneer and follow it in a successional sequence.

Table 4. Percentage of Variance Accounted for by First 3 Principal Components

Mean % Range %
Component variance variance Mean combined % variance
number accounted for (8 sites) 1 &2 1, 2, & 3

1 39.1 27.3-55.4 60.3

2 21.2 18.2-25.5
3 12.8 8.4-13.5 73.1

Since our primary interest concerns the effect of farming on temper-

ate forests, we can supplement this model of an arboreal community
with information on the types of non-arboreal species that are most
sensitive to farming practices such as clearance. These so-called
indicator herbaceous species (Iversen 1967) respond in very specific
ways to clearance and agriculture, and can be divided into three
basic groups (Vuorelli 1970, 1972, 1973, 1975).
A. Species which increase proportionately with agricultural
intensity: sheep sorrel (Rumex acetosella), twitch couch grass
(Asropyron repens).
B. Species which increase or appear with the presence of agriculture
but do not change in response to increases in agricultural intensity:
ribwort plantain (Plantago lanceolata), green plantain (P. major),
common sorrell (Rumex acetosa).
C. Species which increase or appear with early interference but
whose proportion of the herbaceous vegetation decreases with
increases in agricultural intensity: Mugwort (Artemesia vulgaris).
In general, the presence of dramatic increases of pollen of groups
A or B indicates agricultural activity in the area of the site, while
 Palynology and Human Ecology 255
decreases in their representation reflect a pause or termination of
agriculture. Increases in type C vegetation indicate the initiation of
agricultural activity, while subsequent decreases indicate continued
farming in the area. Patterns of these indicator species can be used
to mark the introduction of farming into the area, and to indicate
whether farming was continued beyond initial clearance and planting.
We can now move towards predicting forest succession based on our
ecological modelling.

The initial stages of agriculture within a temperate forest must, of

course, involve clearance. If these clearances are planted with crops
for a short period only, the predicted ecological response to this
cultural pattern is a secondary succession with an initial decrease in
all trees and increases in herbaceous species. This is followed by a
succession of pioneer trees, followed by the re-establishment of the
dominant (climax) tree species, and a concomitant decrease in her-
baceous species. Assuming no degradation of the soil has occurred,
the vegetation may be expected to recover toward its original 'climax'
state. If the clearing was maintained, of course, secondary succession
would be held off.
Secondary succession would be expressed in a rise in agricultural
indicators and a decline in all tree species during the agricultural
phase of the process; with a subsequent succession from colonizer
to climax species (that is, a shift from group III to group I type
arboreal species; see Table 2). Intensification of this pattern would
be indicated by an increase in species A (Rumex acetosella, Asropyron
repens) and/or decreases in group C species (A. vulgaris). The precise
nature of this successional pattern would presumably vary with local
climatic and soil variables.
In order to match these expected community changes with those
expressed in the pollen data, a technique is needed that can distil
species associations from the individual pollen curves for each site.
For this purpose, the multivariate technique of principal components
analysis is proposed. As pioneered by John Birks and T. Webb, this
approach allows for the simultaneous consideration of all species
represented in the pollen record, thereby providing a means to
examine the relationships between all species and how these relation-
ships change through time. As we will see below, this overcomes
many of the problems incurred when we attempt to interpret the
pollen record one species at a time.

THE STATISTICAL ANALYSIS OF FOSSIL POLLEN

The palynological record is suitable for statistical analysis because
pollen grains offer a large spatial sample of comparable data. The
256 GREEN
transfer of pollen grains (often numbering millions) generates the
so-called 'pollen rain' (Faegri and Iversen 1975, Tauber 1965) which
eventually settles over a region. The pollen site then becomes a sample
of this pollen rain. As noted by Faegri and Iversen:
'Although there is... no difference between identifying pollen
grains and... seeds, cone scales etc., there is yet one important
distinction: the fossils dealt with in pollen analysis are all com-
parable, consequently their relative frequency can be expressed
as percentage of the total' (Faegri and Iversen 1975, 123).
These characteristics of fossil pollen form the basis for the major
underlying assumption of palynology, namely that pollen is produced
in great quantity and uniformly dispersed through space (Faegri and
Iversen 1975, 123).
Although the general tenor of this assumption is valid (especially
relative to other fossil data classes), a complex set of sampling errors
needs to be accounted for before proceeding with a statistical recon-
struction of vegetation communities from pollen data. These errors
are the result of pollination and depositional processes and are
summarised below (see Faegri and Iversen 1975 for an extensive
discussion):
(1) The wide variability in pollen production between species
generates a non-representative sample of an area's vegetation in
the pollen rain.
(2) The variety of ways in which species disperse their pollen
results in a very complex and uneven distribution of pollen over
space (Tauber 1965, 1967).
(3) The size of the catchment area of a pollen site and the
redeposition of pollen, for example by moving water, bias the
representivity of pollen at a given site (Birks and West 1973).
Palynologists have developed two general approaches for dealing
with these problems. The first has been to measure differential pollen
production and dispersal in order to understand the sampling biases
they introduce into the pollen record. Differential pollen dispersal
rates are then compensated for in the calculation of relative propor-
tions of tree species in a region. This is complemented by information
on local depositional processes in order to reconstruct the vegeta-
tional history at particular sites in a specific region. Relative pollen
production, for example, has been investigated by comparing the
vegetational census of an area with a sample of its pollen rain, and
developing a rank order of representativity. The results of Andersen's
study of a Danish forest (1970), for example, is a series of coefficients
which serve to decrease the pollen percentages of over-represented
 Palynology and Human Ecology 257
species. These studies have been complemented by research on pollen
dispersal, in which attempts are made to model and measure the
differential pollen displacement distances of species (Tauber 1965).
A second more general approach to the reconstruction of plant
communities from pollen data involves the application of multivariate
statistical methods (Adams 1974, Birks et 1975, Birks and Saarnisto
1975, Gordon and Birks 1972, Webb 1974). This approach follows
research in contemporary ecology, where the object is to isolate the
ecological factors that effect vegetational associations (Gittins 1969,
Goodall 1954). This approach is particularly appealing because it
focusses on the factors that effect community patterns, rather than
attempting to treat mechanically the relative representativity of each
species in the pollen rain.
Principal components analysis has been utilised to derive associ-
ated groups of species which are then interpreted in terms of ecologi-
cal processes. By focussing on the pattern of plant associations, such
multivariate methods have proved powerful in deriving correspond-
ing patterns from both model pollen and contemporary vegetational
samples. This has lent strong support to the use of pollen data for
vegetational reconstruction. For example, in comparing the pollen
found in cores from 64 lakes with that captured in 1,500 pollen traps
in the adjacent forests in Michigan, Webb concludes:
the major patterns apparent in the vegetation are apparent in
the pollen data.... The result clearly supports the use of fossil
pollen to reconstruct the relative composition of past forest types
and changes in plant associations (Webb 1974, 26).

The present analysis draws upon both of the above approaches for
its treatment of pollen data. The primary methodological objective
is to be able to derive patterns from the palynological record in a
way that they can measure the effect of farming on community
succession. For this a multivariate statistical approach is suitable and
powerful. Interpretation of the derived patterns, however, must also
take into account systematic biases in the pollen record.
To account for sample bias, research on pollen production and
dispersal is utilised in two respects. First, patterning derived from
the multivariate analysis is complemented by close examination of
the indicator species. Because of their low production and limited
dispersal, these species are often not present in large enough quan-
tities to effect statistically multivariate analysis. Their presence alone,
however, yields essential information on the presence of farming,
and therefore a presence-absence measure is an important aspect of
the interpretation of the pollen record. Second, the appearance
of pollen types is considered in terms of the relative dispersal
258 GREEN
distances of the particular species. The presence of wind-blown pine
pollen at a site, for example, cannot be interpreted in the same way
as the presence of insect-transported linden pollen. This combined
methodology provides for the interpretation of forest communities
and ecological change.
Eight sites from eastern Denmark (the Danish islands and eastern
Jutland) and western Jutland constitute our sample for this analysis
(see Fig. 1). The distribution of sites was selected to encompass the
major environmental variability of the region. The reliability of the
data was judged according to the sample size for each stratum of
the pollen core. Pollen diagrams from these sites represent the relative
frequency of a set of pollen taxa sampled along a given pollen core,
and can be mathematically conceptualised as a data table with
stratigraphie levels (rows) and taxa (columns). Since the core is
sampled stratigraphically, the series of levels represents a temporal
sequence of deposition with the upper levels being most recent.
The technique devised to convert pollen diagrams into data tables
suitable for principal components analysis involves 'reading' the
pollen diagrams (curves) into a computer with a digitizer (Green and
Moore 1975). A short Fortran program (Readpol) transforms the
digitized data into percentages and prepares it for input as a data

Figure t. Location of pollen sites used in analysis.

 Palynology and Human Ecology 259
matrix to SPSS program Factor which performs principal components
analysis (Nie 1975).
Principal components analysis was applied to this data to derive
the patterns of associations between the pollen taxa, thereby making
it possible to interpret underlying factors causing the variation in the
data set. These factors are mathematically described as principal
components. It could be that these underlying factors are based upon
the environmental requirements and competitive interrelationships
of the plant species (cf. Gittens 1969). Principal components analysis
allows us to examine formally the interrelationships between all the
species at a given site simultaneously (mathematically speaking)
rather than having to compare visually the frequency of graphs of
groups of pollen taxa. For the purpose of our argument this simply
allows us formally to describe which plant species associate with one
another, and which do not. It does not explain the reasons for these
associations. Principal components analysis thereby organises the
data into a form for evaluating our model of community change as
affected by agriculture.
A general discussion of how principal components analysis works
in mathematical terms is not a prerequisite for understanding the
substance of the analysis. Several aspects, however, are important to
understanding its application and interpretation.
If the original data matrix is conceptualized as geometric sample
space with stratigraphie levels as dimensions and taxa as sample
points, the interrelationships between the plant taxa can be visualised
as the geometric scatter of the sample points. For example, in Fig. 2
below it is clear from inspection that the sample points have a certain
pattern. Species which are frequent in sample A are also frequent in
sample B.
Principal components analysis allows us to examine interrelation-
ships between many species by reducing the dimensionality of com-
plex data matrices. Geometrically this is done by placing reference
axes along the major dimensions of the scatter of the data by calculat-
ing a least-squares line. This new reference axis, or principal com-
ponent, accounts for the greatest possible share of the data variance.
Successive principal components are calculated in the same manner,
with the basic requirement that they are orthogonal (perpendicular)
to all preceding components and, therefore, not correlated with each
other. Each of these succeeding axes accounts for the greatest possible
share of the residual variance. The result is a series of principal
components which account for successively smaller parts of the data's
original variance.
Variables, in this case pollen taxa, which correlate in the same
way on a principal component, can be interpreted as behaving
similarly with respect to the process underlying that component; in
260 GREEN

Sample

(Level)

*. pollen t a x a

Sample (Level) A
Figure 2. Conceptual model of principal components analysis in two
dimensions. The points represent pollen taxa (vegetational species), and
the axes pollen samples. In this case those taxa found in sample A are
frequently found in sample B.

this case the ecological processes affecting the forest community.

How strongly a variable is correlated with a principal component is
measured as its component loading. The significance of a component
loading is tested by treating the loading as a product-moment correla-
tion coefficient, and utilising the appropriate test (Child 1970).
The amount of a particular variable's variance explained by a
component is measured as the square of the component loading
(equivalent to the r2 measure in correlation analysis). The total
amount of a variable's (that is species') variance accounted for by a
set of principal components is the sum of these squared components,
and termed the communality (Gittins 1969, 51).
Finally, the relationship between the samples (in this case, strati-
graphic levels) in terms of the derived principal components is
described by the calculation of component scores for each level. This
is done by combining the standardised value of the variables of a
given sample with their respective component loadings on a particular
component (Nie 1975). Component scores measure the strength of
each component for each level; in other words the association
between the group of species represented by the component and the
stratigraphie level.
 Palynology and Human Ecology 261
In sum, the essential elements of principal components analysis
are the following:
(1) A list of principal components for each site with associated
lists of component loadings for each species, represents the
factors affecting each group of associated species and the relative
importance of these factors for the particular species.
(2) A set of component scores for each stratigraphie level,
that represents the relative strength of a component at a par-
ticular level. The component scores are graphed stratigraphically
like the original pollen curves. The scores represent combinations
of species, that is components of the ecological community.
The present analysis uses the program SPSS Factor (Nie 1975),
according to the following procedures. After the principal com-
ponents were delimited, component scores were calculated for the
major components. These were plotted against the stratigraphie order
of the levels. These component score graphs were then graphed
against the appearance and increase of agricultural indicators to
mark periods of significant agricultural activity in the area of the
site. Decreases in their frequency were interpreted as periods of
agricultural decline (except in the case of Group C indicators). The
combination of the indicator graph and the component score graphs
thereby provide a diachronic account of ecological change for each
site.
RESULTS OF ANALYSIS: ECOLOGICAL GROUPS
Our analysis begins with two methodological questions prerequisite
to any palaeoecological interpretation of the data. First, we need to
know whether there were significant patterns in the pollen data which
could be indicative of the Danish forest community. Second, assum-
ing patterning was inherent in the data, we need to discern whether
these patterns matched our colonizing-climax model of the forest.
Initial analysis of the data clearly affirmed both aspects of patterning.
With respect to the first question of patterning, the mean total variance
for eight sites explained by the first three components for the eight
sites, was 73.1 per cent. (The first component contributing a mean
of 39.1 per cent, the second component 21.2 per cent, and the third
12.8 per cent.) This means that the raw data made up of eight arboreal
species can be reduced to two to three community components, while
retaining a high proportion (60 to 75 per cent) of the information of
the original data set (Table 4).
The precise nature of this patterning was then evaluated with
respect to the proposed ecological groups. This was done by tabulat-
ing the significant component loading of all the taxa with the derived
principal components (Table 5). This was then compared to the
expected associations of these taxa with the colonizing ability ranking
262 GREEN
Table 5. Association of Component Loadings of Proposed Ecological Groups
Highest Component Loading Secondary Loadings

Same PC' Same PC/ Same PC/Same PC/

Same Opposite Different PC Same Opposite Different PC
Plant groups Loading Loading (Independent) Loading Loading (Independent)

Elm/Linden 6 0 1 2 0 2
Elm/Oak 0 0 7 0 0 5
Oak/ Linden 2 0 6 1 0 6
Oak/ Elm/ Linden 1 0 0 0 0 0
Oak/ Ash 4 0 4 1 0 5
Ash/ Alder 1 0 5 0 1 2
Hazel/ Pine 0 0 6 1 1 1
Hazel/ Birch 0 1 5 0 3 1
Birch/ Pine 1 0 5 0 0 2
1
PC stands for principal component.

(Table 3). Positive associations between taxa occur when the species
are associated in the same way on a principal component, whereas
negative associations are indicated when species associate in an
opposite way (one positive, and one negative). Independence occurs
when species load on different components. The tabulation was done
in two steps. First, only the highest loadings of each species were
considered, then all secondary loadings were examined.
The results of this tabulation support the proposed ecological
groupings (Table 5). In terms of the dominant species, elm and linden
are consistently associated with the same principal component. In
only one instance are their highest loadings not positively associated
with the same principal component, and in this case two secondary
loadings are associated. In no case are they negatively associated.
The intermediate species (group II), oak and ash, also fit the
expected pattern. Oak is associated positively with elm at several
sites, while at no time is it associated negatively with either linden
or elm. Ash, on the other hand, is shown to be more competitive
with linden or elm, since it loads opposite from them a combined
total of four times. In relation to each other, one half of oak and
ash primary loadings are positively associated, while most of their
secondary loadings are independent. This pattern, along with their
interaction with the dominant species, reflects their intermediate
pioneering or competitive ability, and their different environmental
requirements. Oak is the less exacting of the two species in terms of
soil fertility and water; it can, therefore, co-exist with elm and linden
in the poorer areas of the habitat. Ash, on the other hand, requires
rich and moist soils with good drainage. It must therefore compete
with elm and linden for the better area of the habitats.
The pioneer species, pine, birch and hazel, reflect an interesting
and complex pattern and generally support the proposed ecological
 Pa/ynology and Human Ecology 263
grouping. Hazel and pine appear primarily independent of each
other. All their primary loadings are on different principal com-
ponents, while their secondary loadings show one positive and one
negative association. These patterns reflect the relatively low presence
of pine during Atlantic and Subboreal times in Denmark. Its pattern
probably reflects its presence in early Atlantic times, and its sub-
sequent quick decline. This results in pine representing basically a
temporal factor, independent of the short-term patterns of Atlantic-
Subboreal successions.
Hazel and birch show a pattern of competitive interaction. Their
primary loadings show the species to be basically independent of
each other, though one negative association does occur. Three out
of four of the secondary loadings, however, are negatively correlated,
resulting in a total of four out of ten loadings being negatively
associated. This pattern seems to indicate that these species are in
competition with one another within the early phases of succession.
This concurs with the fact that birch does grow more quickly than
hazel; however, after hazel colonizes an area, birch seedlings cannot
grow under its shade (Iversen 1967). This consideration, then,
differentiates the pioneering abilities of birch and hazel, so that birch
would be expected to precede hazel in a successional pattern.
The final two species, alder and willow, show an independence
from the ecological groups with their respective pioneering ability.
Alder, an intermediate species according to our ranking, is basically
independent from ash. This can be explained in terms of the differing
environmental requirements of the two species. Alder is particularly
sensitive to water levels of a site, requiring very moist conditions and
tolerating even flood. Ash requires only moderately moist soils and
does not tolerate flooding.
Willow, the final pioneer species, consistently loads independently
from all other pioneer species (it is present in great enough quantity
at only three of the eight sites). Like alder, willow is common in very
moist conditions, and can be considered the pioneer version of alder.
Its independence from the other pioneer species is again explicable
in terms of environmental requirements.
In sum, the ecological groups outlined in the pioneering ability
ranking (Table 2) are reflected in the principal components analysis.
Variations from the expected patterning is explicable by the particular
environmental requirements of the species. Elm and linden form a
consistent group; ash and oak form an intermediate group; and birch,
hazel, pine and willow form a final colonizing group. This allows us
to proceed with our analysis of ecological change.
We can now begin to interpret ecological change from the com-
ponent scores. The response of the ecosystem to agricultural activity
can then be inferred by coinciding these scores with the temporal
264 GREEN
pattern of the agricultural indicators. Site specific analyses are sum-
marised (Tables 6-13). A few general comments on the overall pattern
and its fit to the model will now be discussed, followed by two
illustrative examples.
The analytic results meet our general expectations. Changes in
ecological groups reflect the patterns expected from the prescribed
synecological and autecological processes, and these patterns in turn
seem to be attributable to agricultural activity. Rises in agricultural
activity as indicated by weeds and cereals correspond with declines

Table 6. Response of Ecosystem Components to Agricultural Activity.

Site: Tinglev S0

1. Alder-Oak-
Components Ash' 2. Elm-Linden 3. Birch/Hazel'

Levels of Agricultural Activity

(1) None Generally Generally Fluctuating

negative2 positive
(2) Traces of Indicators Smooth increase Smooth decrease Decrease then
fluctuating

Agricultural Indicators: (A) Rumex acet., grasses; (B) Plantago lanceolata;

(C) Artemesia
References: Andersen 1954; Iversen 1949
1
A dash (-) indicates that the species are loading in the same way on the
component; a slash (/) that they are loading oppositely.
2
These descriptions refer to changes in the component scores with respect to the
levels of agricultural activity. (Numbers in left hand margin refer to Figure 6)

Table 1. Response of Ecosystem Components to Agricultural Activity.

Site: Korup S0

1. Elm-
Components Linden/Hazel 2. Oak-Ash/Pine 3. Alder

Levels of Agricultural Activity

(1) First trace Decrease No effect Decrease

Plantago 1.'
(2) Plantago 1. peak 1 Decrease No effect Sharp decrease
(3) Plantago 1. peak 2 Small decrease Decrease Fluctuating
(4) Plantago 1. peak 3 No change Increase Fluctuating
(5) Plantago 1. peak 4 Beginning of Beginning of Beginning of
new decrease decrease decrease

Agricultural Indicators: (B) Plantago lanceolata

References: Iversen 1937; 1941; 1949
1
Numbers in left hand margin refer to figure 7.
 Palyno/ogy and Human Ecology 265
Table 8. Response of Ecosystem Components to Agricultural Activity.
Site: Dyrholmen
2. Elm-Hazel-
Components 1. Linden-Oak Linden 3. Ash

Levels of Agricultural Activity

None Positive Decreasing trend Fluctuating

Indicator Peak Negative Sharp decrease Fluctuating

Agricultural Indicators: (B) Grasses, Rumex acet.; (C) Artemesia

References: Troels-Smith 1942, 1960

Table 9. Response of Ecosystem Components to Agricultural Activity.

Site: Aamosen

1. Elm-Linden-Oak/
Hazel, Rumex,
Components Artemesia, Plantago 2. Ash 3. Grasses'

Levels of Agricultural Activity

Traces of cereals Beginning of Beginning of Fluctuating

decrease decrease negative
Cereals, significant Sharp decrease Sharp Fluctuating
increase in indicators decrease
Peak of indicators Sharp decrease Beginning of Sharp increase
(except Artemesia) increase
Decline of all Small increase Sharp Sharp decrease
indicators increase

Agricultural Indicators: (A) Rumex acet., Grasses; (B) Plantago L; (C) Artemesia;
and cereals
References: Troels-Smith 1960, 1966
1
Because of the limited number of species recorded on the pollen diagram, the
nonarboreal species were included in the principal component analysis. Cereals are
therefore the only independent indicator of agriculture in this analysis.

in most arboreal species. The high forest takes longer to recover than
the other species; while the colonizing species (primarily hazel and
birch) compete for the newly opened habitat. Rises in agricultural
indicators are consistently followed by their decline, with this pattern
being mirrored in the forest ecosystem (declines then rises).
Agricultural intensification does not occur until the end of the
Subboreal or beginning of the Subatlantic periods. Since this period
is only marginally considered in the statistical analyses, the reflection
of intensification is slight. However, the varying behaviours of the
indicator species does reflect intensification in the expected fashion.
Increases in indicator groups A and B and decreases in group C
species do coincide, and are interpreted as reflection of
266 GREEN
Table 10. Response of Ecosystem Components to Agricultural Activity.
Site: Svanemose
Components 1. Elm-Linden 2. Oak-Hazel/Alder 3. Pine-Birch

Levels of Agricultural Activity

Traces of cereals Decrease Strong fluctuations Sharp increase

Plantago 1., Fluctuating Strong fluctuations Fluctuating
Artemesia Grasses positive negative
Same as above Decrease Increase Small increase
Rise in indicators Fluctuating Starts decreasing Increasing
negative

Agricultural Indicators: (A) Rumex acet., Grasses; (B) Plantago 1.; (C) Artemesia;
and cereals
References: Jessen 1939b, Jonnassen 1950, Nillson 1948

Table 11. Response of Ecosystem Components to Agricultural Activity.

Site: Bunds0

Components 1. Elm-Linden 2. Oak-Ash 3. Birch-Hazel

Levels of Agricultural Activity

First trace of Plantago 1. Beginning of Sharp decrease Beginning of

sharp increase
decline
Indicator Peak Fluctuating Decrease Small increase
negative
Indicator Peak Increase Small decrease Small increase
Indicator Peak Decrease Negative Fluctuating
negative

Agricultural Indicators: (A) Rumex Acetosella, Grasses; (B) Plantago 1.; (C)
Artemesia
References: Andersen 1954; Jessen 1938, 1939

Table 12. Response of Ecosystem Components to Agricultural Activity.

Site: Skalles0gaard

Components 1. Oak-Ash-Linden 2. Birch-Linden-Elm 3. Pine-Birch

Levels of Agricultural Activity

Plantago peak Decrease Decrease Small decrease

Plantago peak Negative Slight decrease None
Plantago peak Decrease Sharp decrease None
Plantago peak Decrease Negative Skyrockets

Agricultural Indicators: (B) Plantago 1.

References: Jonnassen 1950
 Palynology and Human Ecology 267
Table 13. Response of Ecosystem Components to Agricultural Activity.
Site: Stevningen

1. Elm-Linden/
Components Ash-Birch-Beech 2. Ash-Oak/Hazel 3. Pine/Ash

Levels of Agricultural Activity

First trace Plantago Rise Decrease Sharp rise

First trace cereals Rise Decrease Sharp decrease
and grasses,
Plantago 1.
Significant rise Rise Rise then decline Beginning of
in cereals significant rise

Agricultural Indicators: (A) Grasses, (B) Plantago 1. and cereals

References: Nillson 1948

intensification. This pattern is clear in the diagram from Dyrholmen

(Table 8) where Artemesia drops significantly at the point where
Plantago lanceolata is increasing sharply; and at Svanemose (Table
10) where Artemesia mirrors the indicator grasses and Plantago
lanceolata at the beginning of the Subatlantic period.
Two sites are now specifically discussed to illustrate the analysis,
and to exemplify the derived patterns. Tinglev S0 (Andersen 1954)
exemplifies a slightly interrupted ecological succession, Korup S0
(Iversen 1949) the effects of long fallow agriculture on the surround-
ing temperate forest.
The first two component score curves for Tinglev S0 (Fig. 3; Table
6) are smooth and complementary. Component 1 reflects the alder,
oak and ash aspect and shows a succession from negative to positive
values, indicating its increasing importance through time. Component
2, elm and linden, reflects exactly the opposite pattern, a succession
from positive to negative values. The third component, representing
mostly hazel but including birch, fluctuates strongly, and does not
seem to respond to the slight agricultural activity indicated.
The general pattern at this site, then, is a shift from elm and linden
to oak, ash and alder, with fluctuations in hazel and birch. Although
no major agricultural activity is indicated through the pollen, this
change in the forest could reflect human exploitation at the very low
levels of intensity such as tree coppicing for fodder and the exploita-
tion of hazel nuts. Linden and elm have historically been the most
desirable trees for cropping and at the same time they are most
affected by these practices because of their long reflowering period
(Green 1973, Ovington 1962, Troels-Smith 1960). The changes in
birch may reflect regional patterns in clearing or succession, since
birch disperses its pollen widely.
268 GREEN

COMPONENT
2

Traces of
indicators
(Atlantic
Subboreal
transition)

Alder-Oak-Ash Elm-Linden Birch

neg:- Hazel

Figure 3. Component score plot for Tinglev S0. The relative strength of the
three principal component scores are stratigraphically plotted alongside the
agricultural indicators. Refer to Table 6, references Andersen 1954, Iversen
1949.

At Korup S0 (Fig. 4; Table 7), there are obvious signs of cultivation

and the expected ecological responses are reflected in the analysis.
The major impact of the agriculture at this site is seen in component
1, consisting of elm and linden with hazel loading negatively.
Responses to agricultural activity coincide with sharp decreases in
the scores of this component, while the periods between indicator
peaks show recovery of the forest with increases in the component
scores. The oak-ash component (2) reflects this same pattern to a
lesser degree. Finally, the alder-linden component (3) reflects impacts
of agriculture strongly, probably a reflection of linden's response to
clearing.
These analyses systematically support our model by showing the
initial indication of agriculture in the vegetational record to be
reflective of an extensive strategy, and by reflecting the expected
ecological response to this impact. The ecological responses to
agriculture are reflected through the association of species on the
derived principal components (Table 3), and the interaction between
components through time (Tables 6-13; Figs. 3,4). The methodology,
 Palynology and Human Ecology 269
COMPONENT
PEAK 5
1 2 3

PEAK 4

PEAK 3

**t

t PEAK 2
JC

PEAK 1

Elm-Linden Oak-Ash Alder

neg:- Hazel neg:-Pine

Figure 4. The component scores at Korup S0 plotted stratigraphically

against the 4 peaks of Plantago lanceolata, an indicator of agricultural
activity (indicator group B). Refer to Table 7, references Iversen 1937,
1941, 1949.

therefore, proves to be appropriate to the prescribed goals of the

research, as it provides for the analysis of ecological interaction and
diachronic patterning both of which contribute to our understand-
ing of the prehistoric agricultural ecosystem of Atlantic-Subboreal
Denmark.

Forest clearance and farming practices disrupt ecological processes

in short and long term ways. This disruption operates primarily
through community processes, and must therefore be observed using
methods which can detect such community relationships. The key to
understanding the historical effects of farming on forest habitats,
therefore, is the development of ecological theory, particularly suc-
cession theory and plant demography, and its application through
operational models of forest communities. When combined with an
understanding of the human side of farming, these models can be
made to work for us in the interpretation of the palaeoecological
and archaeological record. This linking of theory, method and data
270 GREEN
is essential to the process of historical interpretation in which
archaeologists are interested. Pattern recognition alone is insufficient
for interpretation, since infinite numbers of pattern-process combina-
tions are possible for any given case. Nowhere is this more obvious
than in the interpretation of changing prehistoric environments.
However, pattern recognition within a problem-directed approach
can yield powerful interpretations of things past. Taking the process
to the pattern is the work of human palaeoecology.
BIBLIOGRAPHY
Adam, D. P. (1974) Palynological applications of principal
components analysis and cluster analysis. J. Research, U.S.
Geological Survey 2(6), 727-41.
Andersen, A. (1954) Two standard pollen diagrams from South
Jutland. Danmarks Geologiske Undersogelse, II R, No. 80.
Andersen, S. T. (1970) The relative pollen production and pollen
representation of Northern Europe trees and correction
factors for tree pollen spectra determined by surface pollen
analysis from forests. Danmarks Geologiske Undersogelse. II
R, No. 96.
Bartlett, P., (ed) (1982) Agricultural decision making. New York,
Academic Press.
Birks, H. J. B. & R. G. West (1974) Quaternary plant ecology.
Oxford, Blackwell Scientific Publications.
Birks, H. J. B., J. Deacon, & S. Pelgar (1975) Pollen maps for the
British Isles 5000 years ago. Proc. of Royal Soc. London 189,
87-105.
Birks, H. J. B. & M. Saarnisto (1975) Isopollen maps and
principal components analysis of Finnish pollen data for
4000, 6000 and 8000 years ago. Boreas 4, 77-96.
Bornebusch, C. H. & K.. Milthers (1935) Jorbundskort over
Danmark. Danmarks Geologiske Undersogelse, III R, No. 24.
Boserup, E. (1965) The conditions of agricultural growth: the
economics of agrarian change under population pressure.
Chicago, Aldine.
Child, D. (1970) The essentials of factor analysis. New York, Holt,
Rinehart and Winston.
Clapham, W. B. (1976) An approach to quantifying the
exploitability of human ecosystems. Human Ecology 4, 1-31.
Clark, C. (1967) Population growth and land use. New York, St.
Martin's Press.
Clark, J. G. D. 1975 The earlier stone age settlement of
Scandinavia. Cambridge, Cambridge University Press.
Clark, R. M. (1975) A calibrated curve for radiocarbon dates,
Antiquity 49, 251-66.
Faegri, K. & Iversen, J. (1975) Textbook of pollen analysis. (Third
edition, revised). Copenhagen, Munksgaard.
Gittins, R. (1969) The application of ordination techniques pp.
37-66 in I. H. Rorrison (ed.) Ecological aspects of the mineral
nutrition of plants. Oxford, Blackwell Scientific Publications.
Godwin, H. (1956) The History of British flora. Cambridge,
Cambridge University Press.
 Palynology and Human Ecology 271
Goodall, D. W. (1954) Objective methods for the classification of
vegetation: An essay in factor analysis. Australian J. of
Botany 2(1), 304-24.
Gordon, A. D. & H. J. B. Birks (1972) Numerical methods in
Quaternary palaeocology 1 : zonation of pollen diagrams. New
Phytologist 71, 961-79.
Green S. W. (1973) Population growth and birth spacing: a
simulation using model life tables, Manuscript on file,
Department of Anthropology, University of South Carolina.
Green S. W. (1980a) Broadening least cost models for expanding
agricultural systems, in T. Earle and A. Christiansen (eds.).
Modelling subsistence change in prehistoric economies. New
York, Academic Press.
Green, S. W. (1980b) Toward a general model of agricultural
systems, in M. Schiffer (ed.). Advances in Archaeological
Method and Theory Vol. 3. New York, Academic Press.
Green S. & J. Moore (1975) The computerized analysis of pollen
data, Manuscript available, Department of Anthropology,
University of South Carolina.
Harper, J. L. & White, J. (1974) The demography of plants, Ann.
Rev. of Ecology and Systematics 5, 419-63.
Harris, D. (1969) Agricultural systems, ecosystems and the origins
of agriculture, pp. 3-16 in P.J. Ucko and G.W. Dimbleby
(eds.). The domestication and exploitation of plants and
animals. Chicago, Aldine.
Harris, D. (1972) Swidden systems and settlement, pp. 245-262 in
Man, Settlement and Urbanism, edited by P.J. Ucko et al.
London, Duckworth.
Horn, H. (1974) The Ecology of Secondary Succession, Ann. Rev.
of Ecology and Systematics 5, 25-37.
Horn, H. (1975) Markovian properties of forest succession, pp.
196-211 in M. L. Cody and J. M. Diamond (eds) Ecology and
Evolution of Communities. Harvard University Press.
Iversen, J. (1937) Pollenanalyser tidsbestmesler af midjyske
mesolithiske bopladser. Aarboger 25, 182-6.
Iversen, J. (1941) Landnam i Danmarks stenalder. Danmarks
Geologiske Undersogelse II R, No. 66.
Iversen, J. (1949) The influence of prehistoric man on vegetation.
Danmarks Ceologiske Undersogelse IV R, Bd. 3, No. 6.
Iversen, J. (1960) Problems of early postglacial forest development,
in Danmarks Geologiske Undersogelse IV R, Bd. 4, No. 3.
Iversen, J. (1967) Naturens udvikling siden sidste istid, pp. 345-
455 in Danmarks Natur I.
Iversen, J. (1973) The development of Denmark's nature since the
last glacial. Danmarks Geologiske Undersogelse IV R, No. 7-C.
Jessen, K. (1938) Some west Baltic pollen diagrams. Quartr Bd.l,
124-39.
Jessen, K. (1939a) Naturfoldene og mennesket i Danmarks oltid.
Fortis og Nutid 13. Mimeograph.
Jessen, K. (1939b) Bundso II: kornfund. Aarboger 29, 169-98.
Jonnassen, H. (1950) Recent pollen sedimentation and Jutland
heath diagrams, Dansk Botanisk Arkiver 13 (7), 1-168.
Likens, G. E. & F. Boorman et al. (1978) Recovery of a deforested
ecosystem, Science 199, 492-95.
272 GREEN
MacLachlan, M. (1984) Why they did not starve. ISHI press.
Margalef, R. (1968) Perspectives in ecological theory. Chicago,
University of Chicago Press.
Moore, J. A. & S. W. Green (1979) 'Drive he said: evaluating
models of agricultural expansion'. Paper presented in
workshop on Simulation in archaeology. Lipe and
Hackenberger organizers. 1979 Annual meetings of the society
for American Archaeology, Vancouver, British Columbia.
Nie, N. et al. (1975) SPSS systems manual. Second edition, New
York, McGraw-Hill.
Nillson T. (1948) On the application of the Scanian post-glacial
zone system to Danish pollen diagrams. Kgl. Danske Vidensk.
Selskab Biol. Skrifter 5. Copenhagen.
Odum, E P. (1971) Fundamentals of ecology. Philadelphia, W.B.
Saunders.
Ovington, J. D. (1962) Woodlands. London: London, English
Universities Press.
Schou, A. (1949) Atlas of Denmark II: the landscapes.
Copenhagen, Royal Danish Geographic Society.
Skaarup, J. (1973) Hesselo-solager: Jagdstationen der
Sudskandinavischen Trichterbeckerkulturen. Akademisk Forlag.
Copenhagen.
Skaarup, J. (1975) Stengade: Ein Langlandischer Wohnplatz mit
Haysresten aus der fruhneolithischen Zeit. Medd. fra
Langelands Museum Rudkobing, Denmark.
Tauber, H. (1965) Differential pollen dispersal and the
interpretation of pollen diagrams. Danmarks Geologiske
Undersogelse II R, No. 89.
Tauber, H. (1967b) Investigations of the mode of transfer of
pollen in forested areas, Rev. Paleobot., Palynology 3, 277-86.
Tauber, H. (1972) Radiocarbon chronology of the Danish
mesolithic and neolithic. Antiquity 46, 106-10.
Troels-Smith, J. (1942) Geologisk datering af Dyrholm fundet.
Kgl. Danske Vidensk; Arkeol. Kunst Hist. Skrifter. Bd. 1, No. 1.
Troels-Smith, J. (1960) Ivy, mistletoe and elm: climatic indicators:
fodder plants. Danmarks Geologiske Undersogelse IV R, Bd.
4, No. 4.
Troels-Smith, J. (1966) The Ertebolle culture and its background.
Paleohistoria 12, 505-28.
Vuorela, I. (1970) The indication of farming in pollen diagrams
from south Finland. Acta Bot. Fennica 87, 1-40.
Vuorela, I. (1972) Human influence on the vegetation of
Katinhauta Bog. Acta Bot. Fennica 98, 1-21.
Vuorela, I. (1973) Relative pollen rain around cultivated fields.
Acta Bot. Fennica 102, 1-27.
Vuorela, I. (1975) Pollen analysis as a means of tracing settlement
history in southwest Finland. Acta Bot. Fennica 104, 1-48.
Walker, D. & R. G. West (eds.) (1970) The vegetational history of
the British Isles. Cambridge, Cambridge University Press.
Watt, K. E. F. (1973) Principles of environmental science. New
York, McGraw-Hill.
Webb, T., Ill (1974) Corresponding patterns of pollen and
vegetation in lower Michigan: A comparison of quantitative
data. Ecology 55, 17-28.
 Smut, Scab and Pox:
Disease and Environmental Aarchaeology

Don Brothwell,
Institute of Archaeology, London

In a paper entitled 'Problems in the identification of environmental

problems', Vayda and McCay (1977) are clearly aware of the difficul-
ties of establishing a proper balance in environmental studies. 'Under-
lying our discussion', they say, 'is a belief in the desirability of
focussing on environmental problems and how people respond to
them. The kind ... with which we are especially concerned are those
constituting hazards to the lives of the organisms experiencing them'.
In other words, we are particularly concerned with problems that
carry the risk of morbidity or mortality, the risk of losing an 'existen-
tial game' in which success consists simply in staying in the game'.
They recognise that this is not the way the subject has altogether
shaped up; and claim, 'A main criticism of the work of human
cologiste is that they have tended to concentrate their enquiries
upon the production and consumption of food to a degree that
amounts to the sin of "nutritional reductionism".'
In the case of human palaeoecology there is also evidence of bias,
although one can understand why the food quest and agriculture
have received somewhat distorted emphasis, for the evidence of food
resource exploitation, as well as actual food debris, occurs commonly
enough in the archaeological record.
Clearly there is a need to keep an eye on bias in the subject, and
in particular areas neglected or with potential for more detailed study.
If we are ultimately considering ecological, or more correctly bio-
social, existentialism, then clearly morbidity and mortality, not only
of human populations but of species of economic relevance to Man
need to be considered. Or is disease so divorced from the major
themes of archaeology that it can be safely excluded? Previously,
the health status of human populations has been seen as being
contained within the medical sciences, but anthropologists concerned
with living groups now recognise its considerable relevance to them,
and indeed medical anthropology in the United States is a fast
growing subject both in terms of research and the growing number
of courses in it. It certainly seems logical to include evidence of
human health and disease in human palaeoecology, and the range
274 BROTHWELL
of evidence is considerable, from congenital abnormality, to trauma,
oral pathology, arthropathies, neoplasms (tumours) and infectious
diseases. Such information can be related to a host of factors, from
the occurrence of detrimental mutants to forms of environmental
stress and disequilibrium between parasites and people. Archaeology
may in fact eventually contribute significantly to the history and
evolutionary biology of some disease groups. In the case of
pathogenic trponmes, which include the Great Pox (venereal
syphilis), it is begining to be possible to substantiate archaeologically
a theoretical framework for these Man/parasite relationships, where
there is initially adaptation of the trponmes to human groups in
tropical environments (producing yaws), further changes associated
with rural groups in arid situations and poor hygiene (producing
endemic syphilis) and finally, in the colder but socially more
advanced societies of Europe, a transformation into the venereal
form and a transmission usually restricted to adults.
Evidence of disease in other animal species has appeared in
material as diverse as medieval East Anglian fish, early Amerindian
macaws and Pleistocene Olduvai mammals. The identification and
interpretation of the considerable range of evidence is very much in
its infancy, and is not helped by the fact that veterinary departments
do not usually keep reference collections of dry bone pathology. So
again, it is a matter of developing this subject ourselves, asking the
appropriate questions, examining large samples, and attempting to
obtain meaningful answers from morbidity and mortality evidence.
As in our own species, there are a variety of diseases which can
leave their mark on skeletal remains, if only as a distorted age-group
composition. In the case of domesticated species, the potential value
of the disease evidence is considerable, the conclusions to be drawn
being as varied as the disease processes. For instance, at the Peruvian
site of Telarmachay (Wheeler, pers. comm.), 72% of the pre-
dominantly camelid remains were from newborn animals, a fact
which strongly suggests the ocurrence of a high mortality from
enterotoxaemia, a bacterial condition associated with domestication
and unsanitary corralling. A roughly equivalent condition in the Old
World, in terms of heavy lamb mortality, would be sheep pox,
recorded in Asia as early as the 2nd century A.D. (but certainly with
a prehistory). Caused by a highly contagious virus, there is evidence
today that it acts differentially against certain sheep varieties (Singh
et /., 1979), and presumably did so in the past also.
For contrast, one might consider stress related arthropathies. These
have been employed to indicate traction use of mammals by human
groups, and in the case of the water buffalo from Ban Chiang,
Thailand, (Higham et al. 1981) the changes were considered in
relation to plough use and the development of early rice cultivation
in the area. Yet again, the variable occurence of enamel hypoplasia
 Smut, Scab and Pox 275
in archaeological pigs is probably linked to other environmental
factors, especially levels of nutrition.
These new developments can be disconcerting to those caught up
in the basic work of site sample reporting. The important thing is
that aspects of morbidity as well as mortality are fully recorded even
when specialised evaluations of pathology are not possible. Indeed,
at this stage in the study of zooarchaeological diseases, precise
description and the collection of a sound body of data must be a
prerequisite for later theorising.
Although my own research interests have been especially with the
poxy side of this subject, I must confess to a growing interest also
in smut. A decade ago it would have seemed unlikely that there was
any potential in considering plant disease in environmental
archaeology, but it is a reflection of the extending horizons of the
subject that one can see relevance in it today. The significance and
economic seriousness of plant diseases was of course not lost on
ancient societies many centuries ago, the most directly devastating
of all these to human groups being St Anthony's Fire, caused by
ergot. Even today in advanced societies, crop losses from disease
can be enormous, and clearly in any evaluation of the success of
past crop plants to human communities, the health and productivity
of the plants must be assessed as thoroughly as possible.
At a theoretical level one can postulate, as Brewbaker (1979) has
done, that plant disease led to human movement and resettlement.
Indeed, he has presented an interesting case that the Classic Maya
civilisation collapsed in the ninth century A.D. as a result of a major
plant epidemic, probably of maize mosaic virus. As yet there is no
material evidence of plant tissue changes indicative of the virus, or
of the corn plant hopper Peregrinus maidis which probably transmit-
ted the condition. And alas, while the Aztecs to the north observantly
recorded (Ainsworth, 1981) in their art maize cobs damaged by a
form of smut ( Ustilago maydis), the Mayans appear not to have done
so. Nevertheless, the hypothesis is attractive and certainly biologically
reasonable.
In the Old World there is of course ancient literary evidence for
cereal rust from the Bible, as well as Greek and other sources, and
these early peoples clearly took such things seriously. Direct evidence
of wheat rust, Puccinia graminis, has been demonstrated as early as
Bronze Age times in Israel (Kislev 1982). Here, spore germination
and wheat lemma penetration could be clearly seen, and while a few
cases of plant disease don't make an epidemic, it is relevant that it
could have had a devastating effect on crop yields. In the case of
plant disease caused by nematode parasites of the genus Helerodera,
Webley and Dennell (1978) were certainly able to demonstrate the
common occurrence of this genus at the Bulgarian neolithic site at
 276 BROTHWELL
Chevdar (in contrast to the neolithic site of Kazanluk 100 km away)
and postulated that crop losses through serious infestation of cereals
by either the hordecalis or avenue species might have been one of
the reasons why for some time the settlement did not survive more
than brief occupation.
This question of pathogen and pest restraints on crop productivity
is thus a potentially very interesting one, and it seems difficult to
consider plant domestication fully without keeping in mind the
variable impacts of disease. As Jennings and Cock (1977) have put
it, the evolution of parasitic insects and diseases followed a course
parallel to crop evolution. Just as centres of crop origin are distin-
guished by their wealth of varietal diversity, they are also centres of
variability of pathogens and pests'. They also agree with previous
botanists who have considered that the major reason why the prin-
cipal areas of production of many important crops are outside their
centres of origin is that losses in crop yields due to insect and disease
damage are greatest within centres of crop origin'.
We have tried to argue, as briefly as possible, that environmental
archaeology must consider disease as a major ecological topic, and
that there is potentially a wide range of significant information
possible. Studies on ancient diseases, especially in plants, are still in
their early stages, and we are at an exciting level of exploration. With
new techniques, even the remains of the micro-organisms may yield
far more than we thought possible a few years ago.
Disease is not an uncommon phenomenon, rather, it is all pervad-
ing. It has affected our own evolution and that of our domesticates.
It has also been formative in cultural terms, whether in determining
rituals, impeding armies, or influencing the elaboration of therapeutic
practices. To view disease as an insignificant topic in archaeology
was sad twenty years ago, but today would be a very serious error.

BIBLIOGRAPHY
Ainsworth, G. C. (1981) Introduction to the History of Plant
Pathology. Cambridge University Press.
Brewbaker, J. L. (1979) Diseases of Maize in the west lowland
tropics and the collapse of the Classic Maya civilisation'.
Econ. Bot. 33, 101-118.
Higham, C. F. W., A. ICijngam, B. F. J. Manly, & S. J. E. Moore,
(1981) The bovid third phalanx and prehistoric ploughing, J.
Arch. Sei. 8, 353-365.
Jennings, P R. & J. H. Cock, (1977) Centres of origin of crops
and their productivity Econ. Bot. 31, 51-54.
Kislev, M. E. (1982) Stem rust of wheat 3300 years old found in
; Israel. Science 216, 993-994.
Singh, I. P., R. Pandey & R. N. Srivastava (1979) Sheep pox: a
review. The Veterinary Bulletin, 49, 145-154.

-
 Smut, Scab and Pox 277
Vayda, A. P., B. J. McCay (1977) Problems in the identification of
environmental problems, pp. 411-418 in T. P. Bayliss-Smith
and R.G Feacham (eds) Subsistence and Survival. Rural
Ecology in the Pacific. London, Academic Press.
Webley, D. & R. W. Dennell (1978) Paleonematology: Some
recent evidence from Neolithic Bulgaria. Antiquity 52, 136-37.
 Interaction Between The Environment and
Social Subsystems

W. Groenman-van Waateringe,
University of Amsterdam

Mankind forms part of an ecosystem, consisting of a biological

community bound together by an intricate web of relationships and
embedded in the physical environment with which it interacts and
which is modified by it. Man, therefore, cannot be isolated from his
environment, although he, more than any other branch of the biologi-
cal community, is capable of modifying both its biological and
physical components. Mankind, however, is also subject to the law
of limiting factors, although by technological advances it is possible
to expand the possibilities, and as human ecotechnology develops a
wider field of opportunity opens up. Thus, given an ecotechnological
structure, the prehistorian (who is, after all, studying human activity)
should be interpreting the choices made by human groups in terms
of social behaviour. In other words, the environmental subsystem
and the social subsystem are in a condition of continual interaction
on account of the technostructure created by man (Sachs 1974), and
this is most sharply defined in relation to the whole problem of
population growth and human evolution.
The origin of agriculture has been and still is an intriguing problem
and many scholars have tried to answer it (e.g. Binford, 1968). The
Oasis' theory of Childe (1928) assumed a serious desiccation at the
end of the Ice Age, aggregation of plants, animals and man in the
river valleys and the start of a new type of symbiosis. As it gradually
became clear that no such desiccation had taken place, the Oasis'
theory was replaced by Braidwood's idea of favourable circumstances
in the Fertile Crescent, the presence of suitable plant and animal
domesticates and increasingly receptive human communities, leading
to more settled environmental responses (Braidwood 1963). The
school of Higgs (Higgs and Jarman 1969, Higgs 1972) followed more
or less the same line. Domestication is the result of a long-term
process of learning about the possibilities the environment has to
Offer in the way of plants and animals, assuming the ever existing
symbiosis of man with plants and animals to have intensified. Binford,
in his 1968 paper, favours the idea of population growth as the main
impetus for the start of domestication, although a direct link between
 Interaction Between Environment and Social Subsystems 279
population growth and available food is nowadays denied. Ethno-
graphic research has taught us that hunter-gatherer population groups
are homeostatically regulated below the carrying capacity of their
habitat. Population size could consciously be kept under control by
different means, for example marriage/mating rules, lactation taboos,
abortion, infanticide. Food supply was no problem for hunter-
gatherers in such a stable equilibrium. According to Binford (1968)
only a reduction in the biomass of a region, upsetting the balance
between population and available food, or a change in the demo-
graphic structure of the region, could trigger off a disequilibrium. In
this respect he brings forward the concept of closed versus open
systems. Immigration into, or emigration towards, a region could
evoke stress situations, and when these occurred in areas with suitable
plants and animals, domestication could be a potential solution,
although increased birth control would be an alternative possibility.
Population growth could be the result of increased sedentism, while
reduced mobility might cut out the necessity of spacing births. Flan-
nery (1965) saw the change to a broad spectrum of food resources
as a means to allow for sedentism, and as a precursor of domestica-
tion. Binford (1983) has revised his ideas in so far as he still thinks
that population growth triggered off domestication, but that the main
impetus was demographic packing and the inevitable loss of mobility;
whereas this would be absolutely necessary as a security option for
a hunter-gatherer group. With the sound biologial principle in mind
that every species has its preferred ecological zone, Binford analysed
population densities for different environmental zones. The result
was that Homo sapiens seems to be most successful in the temperate
zone, because of its higher reproduction potential there in comparison
with other zones.
But the question still remains, if man was consciously able to
control his population size, could it grow out of control, leading to
manipulation of plants and animals in order to regulate food resour-
ces more effectively? After colonizing a region with sufficient food
resources, an animal population will, in principle, increase until one
of the vital resources becomes a limiting factor. However, this is an
over-simplified representation of a growth curve, because the popula-
tion will tend to expand over the carrying capacity of the region as
the starting population (which includes many immature individuals
with immature appetites) grows up into an equivalent adult popula-
tion eating adult-sized portions. The limiting factors then come into
play. Part of the population dies off or emigrates, bringing the
population once more below the carrying capacity. The effects of
external limiting factors will be more pronounced in the case of
small, short-lived and rapidly-reproducing animals and will give rise
to more marked oscillations in the growth curve than would be the
case with large, long-lived and slowly-reproducing creatures, where
280 GROENMAN-VAN WAATERINGE
strong adult animals can survive a temporary food shortage more
easily. With animals we may therefore speak of an oscillatory move-
ment around a general equilibrium position. Not so, however, with
human populations. Because man possesses the ability to improve
his ecotechnology, he is capable of transcending the limiting factors
by artificially increasing the carrying capacity of a particular region.
Furthermore, the negative effects will only slowly become apparent
because the human being is a long-living, relatively slow-reproducing
creature and the oscillation will cover at least a generation, if not
more. The social subsystem might be able to cope if the limits of the
environmental subsystem can no longer be stretched with the ecotech-
nology available; but only so long as the social subsystem realizes
in time that the limits have been reached. The mely signal of danger
is, however, considerably impeded by the length of the oscillation.
Two Dutch zoologists, Van Schaik and Van Noordwijk (1983) have
analysed the sex ratio of non-human primates as given in the scientific
literature. This has led them to the conclusion that in stress situations
(i.e. situations which include food shortages and/or high population
densities in captivity) fewer females are born and those females which
are born have a lower survival rate. This phenomenon, which they
observed only with polygamous animals (e.g. apes and goats), is
termed by them 'enforced density-dependent sex ratio change'. Situ-
ations entirely free of stress were found exclusively in groups living
under uncrowded conditions and receiving enough food. The data
from the literature were divided into three categories on the basis of
food availability, social tension and the type of social system, i.e. 1)
wild groups, having a natural composition and not fed by man, 2)
captive groups with an artificial composition and living under crow-
ded indoor conditions, and 3) provisioned groups, having a natural
composition, living in large compounds and fed by man. High sex
ratios at birth, i.e. a higher percentage of males than females, occurred
in groups 1 and 2. Only in group 3 with enough food and low stress,
were more females than males born.
The fact that fewer females are born in times of stress may be
explained by the observation that females carrying female foetuses
are more likely to be attacked by other conspecifics (i.e. other adult
females) than those carrying male foetuses. The effect only becomes
apparent after the sex hormones of the foetus start to circulate through
the mother's blood. Moreover, female young, especially those of
low-ranking mothers, are more heavily attacked in stress situations
than young males of low-ranking mothers, and thus have less chance
for survival. Man, as noted above, has other means to check his sex,
ratios and from ethnographic sources it is known that he did and
still does so deliberately by female infanticide (Hausfater 1984, Penn
Handwerker 1983).
 Interaction Between Environment and Social Subsystems 281
This suggests that wild populations live in a constant state of
shortage and thus stress, except after calamities such as epidemics,
drought, severe cold etc. (Hillenius 1985). Only in such periods of
stress for the species - but which for the individual means sufficient
food and absence of aggression - are more females born, resulting
in an excess of females per potent male. Bird species (see Hillenius
1985) showing both polygamous and monogamous behaviour, prac-
tise the first only in times of abundance, when a female can look
after her offspring on her own; the second type of behaviour is shown
in lean periods.
This research has important consequences for what could have
happened to human populations in the period preceding the transi-
tion from hunter-gatherer to food producer. The retreat of the ice
cover from extensive areas of the globe and the consequent expansion
in biomass in the form of primary producers, must have created an
almost ideal food situation for the primary consumer, such as the
large herbivores. Since all large herbivores are polygamous, the
research outlined above leads us to expect a spectacular increase in
population. If the pattern for the primary consumers is also valid
for polygamous hunter-gatherers, the next in the food chain, we may
likewise expect a population explosion due to an increase in female
individuals. But this whole development follows that of the her-
bivores, so that when the population of the large herbivores has
reached its ceiling density and starts declining, human population
increase would still be in full swing due to the long oscillatory
movements mentioned previously. The result is food shortages, pack-
ing, social stress, and the downward movement which can only be
checked by the acceptance of a new ecotechnology, for example food
production. This leaves, of course, Binford's ideas on loss of mobility,
packing etc. intact. Successful hunter-gatherers could be successful
because they did not suffer from unexpected food abundance, dis-
rupting their normal procedures to keep their population size in
check. Thus groups living in unchanged conditions, either favourable
or unfavourable, for a long period, as the Northwest Coast Indians
or the Eskimos before civilization struck, learned how to cope suc-
cessfully with their own environment.
We may conclude that the impetus for change in ecotechnology
is not lack of food, but just the opposite; the danger lies in changes
m the environment leading to an increase in food, when the environ-
mental pacifies the social subsystem.

BIBLIOGRAPHY
Binford, L. R. (1968) Post-Pleistocene adaptations, pp. 313-341 in
S. R. Binford and L. R. Binford (Eds.). New Perspectives in
Archaeology. Chicago, Aldine.
282 GROENMAN-VAN WAATERINGE
Binford, L. R. (1983) On the origins of agriculture, pp. 195-213 in
L. R. Binford (Ed.). In Pursuit of the Past: Decoding the
Archaeological Record. London.
Braidwood, R. J. (1963) Prehistorie men. 6th ed. Chicago, Aldine.
Childe, V. G. (1928) The Most Ancient East: The Oriental Prelude
to European Prehistory. London.
Flannery, K. V. (1965) The ecology of early food production in
Mesopotamia. Science 147, 1247-1256.
Hausfater, G. (1984) Infanticide: comparative and evolutionary
perspectives. Current Anthropology 25, 500-502.
Higgs, E. S. (Ed.) (1972) Papers in economic prehistory.
Cambridge, Cambridge University Press.
Higgs, E. S. & M. R. Jarman, (1969) The origins of agriculture: a
reconsideration. Antiquity 43, 31-41.
Hillenius, D. (1985) De aanvallige vrouw. NRC/ Algemeen
Handelsblad 26.1.1985, Zaterdags bijvoegsel, 2.
Penn Handwerker, W. (1983) The first demographic transition:
an analysis of subsistence choices and reproductive
consequences. American Anthropologist 85, 5-27.
Sachs, I. (1974) Environment et styles de dveloppement. Annales
29(1), 553-570.
Schaik, C. P. van & M. A. van Noordwijk (1983) Social stress and
the sex ratio of neonates and infants, among non-human
primates. Netherlands Journal of Zoology 33, 249-265.
 Foraging Theory: Mathematical Modelling of
Socioecological Change

K. V. Boyle and C. D. Wright,

Cambridge University

Since environmental archaeology is the study of man's total environ-

ment, it is of interest to consider his use of available resources; and
then to examine prehistoric subsistence ecology. The aim of this
paper is to present a technique which, it is believed, can be used to
analyse relevant data and learn more about the structure of hunter-
gatherer foraging strategies by explicitly modelling appropriate
features of the environment.
When adopting a new technique and applying it to archaeological
data we are forced to realise that the attempt may not succeed but,
just as statistical analysis can prove to be useful even when correla-
tions and significances are not found, so it is that we can learn much
from the apparent failure of other approaches. The technique
employed here is that of Linear Programming, one well known to
economists and commonly used in the study of optimal allocation
of resources.
Throughout this discussion three important assumptions are made.
The first is that there is indeed a strategy used by the hunter-gatherers
in question. This may be determined partly by culture and partly by
physiological needs. The second is that nutritional and non-food
requirements are satisfied. Thirdly, it is assumed that the strategy is
such that it requires less effort than alternative feasible strategies.
Justification for these lies in the following argument: any group of
humans that satisfies legitimate needs with less effort than another
group is more likely to survive when resources become scarce and
competition fierce.
To demonstrate this technique we shall consider an example chosen
solely for its illustrative value. The resources considered number only
two, and the nutrients three. This is to prevent the example from
becoming not only unwieldy, but totally unmanageable. The selection
of resources and nutrients used may be thought of as surprising, but
are chosen for the same reasons.
If we let be the amount of reindeer captured, in units of 100g,
and y be the amount of eel caught, also in 100g units, then the
284 BOYLE AND WRIGHT
following can be written :-
3x + 4y > 14 for vitamin B,
27x + 14.4y 5 80 for protein and
7x + 2y s= 14 for niacin
These can be represented graphically as in Fig. 1.
It can be seen that there is a large, indeed infinite, area correspond-
ing to possible combinations of quantities of resources satisfying

Table 1. Table 1 shows some of the nutritional content

of eel and reindeer, the two resources chosen for this
example.

Vit. B Protein Niacin

Reindeer 3 Mg 27 g 7 mg
Eel 4 Mg 14.4 g 2mg
Needs 14 Mg 80g 14 mg
(Daily)
Contents /lOO g

Feasible Combinations

1 2 3 4 s *
(Reindeer)

Figure 1. Nutritional Constraints.

 Foraging Theory 285
these needs. What should now be done is to analyse these data to
determine which combination should be used, and why.
In order to do so, relative costs of procurement must be decided
upon, taking into account such factors as risk, search time, pursuit
time, level of technology, seasonality etc. These will be mentioned
again later, but for the sake of simplicity in this example it will be
assumed that for the effort involved in acquiring one reindeer one
could, by use of nets etc., catch around 100 eels. Using the fact that
the average reinder will yield around 55 kg of useable meat, or 550
units of 100g, and that the average eel will yield in the region of
l.lkg, or 11 units of 100g, we can deduce that for the effort involved
in acquiring 550 units of reindeeer, one could obtain 1100 units of
eel. This shows that reindeer is, kilo for kilo, twice as hard to procure
as eel. Using our symbols as above we can therefore state:-
Total cost = C = 2x + y
(It should be noted that while the data in this example are zoologi-
cally correct, the values have been chosen within the available ranges
so as to simplify the numbers that emerge. Hence the beautifully
simple 2.)
Given this formulation of the cost (effort), close examination of
Fig. 2 shows that all possible combinations of resources that satisfy
the requirements at given costs fall on straight lines.
It can be seen that lower costs correspond to lines closer to the
lower left. What must be done is to find the line of lowest cost
intersecting the area corresponding to the feasible combinations. This
is shown in Fig. 3.
The point of intersection is then the combination of resources that
satisfies the requirements (since it is in the appropriate area) and
has lowest cost. This shows that 89g of venison are needed per person
per day, or one reindeer per group of thirty individuals every three
weeks (assuming that the entire mass of available meat was used).
In comparison, 390g of eel are required per person per day, or ten
eels per day for a group of thirty.
When one expands the model and takes into consideration a larger
number of resources and requirements, the model becomes more
representative of reality, but only at the cost of complexity. If this
example contained just one more resource it would require a three-
dimensional diagram, and if even more are added it becomes imposs-
ible to solve by the above method.
Linear Programming, however, is a widely used technique and
methods have been developed to allow solution of these more com-
plex systems. In particular, the Simplex Method can be used to solve
such systems and, while this requires the use of a computer, most
packages available are easy to use, reliable and provide a great deal
286 BOYLE AND WRIGHT

1 2 3 4 S *

Figure 2. Lines of Given Costs.

of information about the model. Specifically, the Simplex Method

allows the determination of which factors, in the original problem,
were important or critical. This information comprises what is
known as post-optimality analysis.
If the Simplex Method is used on our simple example the following
post-optimality information is obtained. We are told that the limiting
factors are protein and niacin. The requirements for these are only
just satisfied by the solution, whereas the vitamin needs are met
with an excess of 4.22 mg over the 14 mg. The most important
information that can be gleaned from this analysis however, is that
concerning how critical the initial data were.
Again, for the sake of simplicity, let us centre our attention only
on the question of the relative costs of reindeer and eel. Returning
to Fig. 2 it can be seen that the ratio of costs manifests itself in the
slope of the lines of equal cost. Changing the ratio will change the
slope of the lines and may thus affect the solution. Close examination
of Fig. 3, however, reveals that the slope of the line can be increased
to as much as 3.5:1 or decreased to 1.875:1. In other words, the
relative cost of 2 that was used could be as high as 3.5 or as low as
1.9 without changing the essential results of the analysis (see Fig. 4).
 Foraging Theory 287

1 2 3 4

Figure 3. Line of Minimum Cost.

Similar analyses can be performed concerning the requirements

and indeed even the nutritional contents of the resources, providing
still further information which is not immediately apparent from the
initial model.
The application of this technique to a real archaeological example
presents several immediate difficulties. Apart from the large range
of resources available (both food and non-food) and the equally
large number of nutritional and other requirements, one must deter-
mine which resources were in fact available and what their nutritional
contents actually were. For example, how does one discover the
amount of riboflavin in a freshly killed woolly rhinoceros? Such
difficulties compel us to turn to modern equivalents to provide
estimates, and the importance of the accuracy of these can then be
analysed by post-optimality work. An adequate diet today
incorporates some forty to fifty different nutrients and, while not all
of these would turn out to be important in the determination of
Palaeolithic man's hunting strategy, it is better to start with a wide
range and allow the technique to eliminate those which are relatively
unimportant.
288 BOYLE AND WRIGHT

Relative Cost of 3.5:1

;
1 2 3 4 S

Figure 4. Lines of Minimum Cost for Different Relative Costs.

A further problem arises in the seasonal availability of resources

and the spatial variability in resource distribution, but both of these
can easily be included by the placing of upper bounds on the amount
of any particular resource in the solution. Consider, for example, the
possibility that the group of 30 could only catch six eels per day (or
2.2 units per person per day). This means that y must be no greater
than 2.2. The inclusion of the equation
y 2.2

therefore takes that restriction into account (see Fig. 5).

In order to obtain a suitable cost function a formula similar to
that advocated by Keene (198la) could be used. Again, all estimates
and approximations can be examined in the post-optimality analysis
so as to prevent a wild guess from critically affecting the solution
without knowing that this was happening. Keene employs the follow-
ing equation:
C = (Ts + Tp + Tc)r
 Foraging Theory 289

x
1 2 3 4 5

Figure 5. Feasible Combinations with Availability Constraints.

where
Ts = Search time,
Tp = Pursuit time,
Tc = Consumption and processing time
and
r = Risk (of loss or unfortunate outcome).
For a full discussion of these and their derivation the reader is
referred to Keene (1981a, 24-36) and Jochim (1976, 27-28). Their
discussions include such topics as the relevance of the technology
available at the time, as well as prey density etc.
No model of prehistoric subsistence economies can completely
ignore social and ideological factors. The superstitions held by
hunter-gatherers will determine, to some extent, which species are
killed and when. The cropping of bovids may require co-operative
hunting techniques: the species' exploitation may have to be severely
limited, for example, to periods of population aggregation (see
Conkey 1980). It may well be that these superstitions etc., were the
means by which optimal foraging strategies were maintained in spite
of the existence of seemingly easier but not nutritionally complete
290 BOYLE AND WRIGHT
alternatives. If so, then they have already been considered, but this
may not be the entire case.
It must also be taken into account that some prey may be hunted
partially or solely for their prestige value. Such considerations
however cannot be included until the later stages of the development
of the complete model for to quote from Kenne (1981b, 193): 'we
have argued that... we must expand the scope of archaeological
investigations, and that this can be done only by augmenting the
archaeological data base with behavioural models. The Linear
Programming models offer a number of insights which we might
otherwise lack were we to rely solely on the archaeological or
ethnographic record'.
Given the various constraints of nutritional composition and
human requirement, it is Linear Programming which enables us to
predict the optimal allocation of scarce/limited resources among
competing needs, using the nutritional environment as an organisa-
tional framework in which interdisciplinary studies of hunter-
gatherer subsistence decisions can be made.

Acknowledgement: The authors would like to thank Dr. P. A.

Mellars for reading this paper and providing many valuable com-
ments.

BIBLIOGRAPHY
Conkey, M. (1980) The identification of Prehistoric
Hunter/Gatherer Aggregation Sites: the Case of Altamira.
Current Anthropology 21, 609-30.
Jochim, M. A. (1967) Hunter/Gatherer Subsistence and Settlement:
A Predictive Model. Academic Press. London and New York.
Keene, A. S. (1981a) Prehistoric Foraging in a Temperate Forest: A
Linear Programming Model. Academic Press. London and
New York.
Keene, A. S. (1981b) Optimal Foraging in a Non-Marginal
Environment: a Model of Prehistoric Subsistence Strategies in
Michigan, pp. 171-193 in Winterhalder, B. and E. A. Smith
(eds.). Hunter /Gatherer Foraging Strategies: Ethnographic and
Archaeological Analyses. Prehistoric Archaeology and Ecology
Series. Chicago, Chicago University Press.
 Putting Pressures on Population: Social Alternatives
to Malthus and Boserup

John Chapman,
University of Newcastle upon Tyne

The most fertile individual whose procreative achievements are recor-

ded for posterity (Freedman 1979) is held to be the last Sharifan
emperor of Morocco (1672-1727), who, with the fulsome aid of
polygyny, produced 888 offspring at an average rate of over 20
children per annum. Fortunately, not all males are blessed with such
meritorious achievements; were this the case, today's world popula-
tion would have occurred in the Early Pleistocene. The moral of the
story is clear: there is a large gap between man's inherent breeding
potential and what actually happens. This apparently inherent
restraint on 'natural' breeding marks the logical starting point for a
discussion of the concepts of population pressure, carrying capacity
and population growth in prehistoric societies.

BIOLOGICAL AND CULTURAL BACKGROUND

The notions of 'carrying capacity', 'population pressure' and 'popula-
tion growth' with respect to the resources to population balance
derive from formulations as early as those of Malthus on population
(Malthus 1st ed. 1798; 2nd ed. 1807). In the first edition of Essay on
the Principle of Population, Malthus stressed the imbalance created
by the potentially geometric increase in population size compared
with the potentially arithmetic rate of increase in food supply. He
concluded that population growth was limited by available food
supplies, in turn constrained by technological capability. The theo-
retical assumptions underlying this conclusion are twofold:-
(a) there is a maximum level at which the environment can gener-
ate food at the technological level of the inhabitants.
(b) there is a more or less constant demand to achieve the
maximum level of food extraction from a given environment
in any given cultural context.
The first definition corresponds to at least one sense of the term
'carrying capacity', while the second is equivalent to 'population
pressure'. The twin aspects of stability and technological advance
have led Boulding (1946, 97-98) to characterize Malthus' work in
292 CHAPMAN
terms of two theories: the 'Dismal Theory', by which equilibrium is
maintained by increasing mortality leading to increasing starvation;
and the 'Utterly Dismal Theory', by which innovations enable larger
numbers of people to live in misery than before. Whilst this simplistic
depiction of a population resources interaction untrammelled by
intervening social variables was modified in Malthus' second edition,
the less refined concepts of his earlier work were incorporated into
Charles Darwin's studies of the evolution of species (Darwin 1859).
Darwin's edifice was constructed on the proposition that differential
reproduction was a direct reflection of the natural selection of
superior traits which arose by chance mutation within the variability
open to individuals in the same or different competing species. Hence,
relative superiority in what was later defined as the genetic inheritance
of competing individuals led to the spread of more successful popula-
tions at the expense of less successful groups. In this way, the
Darwinian theory of evolution was interpreted as lending support to
the notion that higher population densities connoted evolutionary
successes over lesser competing forms.
It was a short but highly significant step in the debate to suggest,
as have neo-Darwinians and sociobiologists alike, that the evolution-
ary success of species or individuals was reflected not only through
their relative reproductive success but in their capacity to maximize
it (Trivers 1971, Hamilton 1964, Alexander 1974, Wilson 1975). In
a series of highly innovative papers, biologists and ecologists defined
mathematical models for the calculation of the genetic advantages
accruing to competing forms of animal behaviour (Clutton-Brock
1974; Clutton-Brock and Harvey 1976, 1977). Since the outcome of
rival strategies was calculated in terms of maximising their reproduc-
tive success, there have been strong pressures on students of human
population to accept the principle of maximization in relation to the
resource quest. Despite the persuasive arguments of sociobiologists,
it should be stressed that models of maximization provide no more
than standards against which to test field data, not an a priori answer
to a set of questions still under investigation.
At least one major assault has been mounted on population
maximization models: the studies of self-regulating animal popula-
tions of Carr-Saunders (1922) and Wynne-Edwards (1962). In both
studies, the absence of population pressure on species-specific carry-
ing capacities was explained with reference to socially-inspired,
density-dependent feedback mechanisms which dampened popula-
tion growth below levels of intense pressure. But the citadel of
maximization was defended by Williams (1966), who pointed out
that species of socially regulating individuals would soon be geneti-
cally swamped by individuals behaving as if to maximize their own
inclusive fitness. Indeed, in the only anthropological field test of the
 Putting Pressures on Population 293
Wynne-Edwards hypothesis of self-regulating mechanisms, Irons
(1979) demonstrated that, far from adjusting their rates of population
growth to maintenance of their resource base, the Yomut Turkmen
were primarily concerned with turning resources into people. The
apparent falsification of Wynne-Edwards's kin selection model by
evidence from both animal and human behaviour has reinforced
maximization theorists' confidence in their model-building.
In the anthropological sciences, the year 1965' marked the birth
of what has come to be known as the 'demographic paradigm'
(Hassan 1975). This rite of passage was symbolized in two stimulating
studies, by Boserup (1965) and Dumond (1965). Both authors pro-
posed the inversion of Malthus's notion of the independent role of
technology in the population resources interaction, in favour of the
specification of population increase as a prime mover stimulating
economic and social evolution. Whereas Dumond (1965) was content
to explore various examples of population resources interactions,
Boserup (1965) expanded the hypothesis of population growth into
a unilinear, evolutionary model of population, settlement and tech-
nology, using intensification and de-intensification of agricultural

TECHNOLOGICAL
CHANGE PERMITS 1
POPULATION GROWTH j PHASE OF
POPULATION
PRESSURE
(a)

POPULATION
SIZE

POPULATION GROWTH
STIMULATES
TECHNOLOGICAL (b)
CHANGE /

CARRYING
CAPACITY"
POPULATION
SIZE

Figure 1. Two step models of the population-technology interaction (a)

after Malthus (1798) (b) after Boserup (1965).
294 CHAPMAN
technology as an indicator of population growth and decline (for
discussions, see Spooner 1972, Sherratt 1980).
One important point relating to the views of Malthus and Boserup
has received little attention, namely the isomorphic role of population
pressure and carrying capacity in the structure of both models (Fig.
1). The constant pressure of population against a technology-specific
ceiling of resource procurement derives from the analogy from phys-
ics of a gas stream under pressure. The first opportunity for expansion
leads to a release of pressure, an expansion of gas and a new
equilibrium. While Malthus and Boserup differ on the instrument
which releases the pressure in their mechanistic model, the constituent
parts of the model are identical.

THE DEIFICATION OF DEMOGRAPHIC GROWTH

Esther Boserup's hypothesis of the independence of population
growth vis--vis technological resources found ready disciples in the
early days of the processualist breakthrough, or 'New Archaeology'.
In two key early analyses of the transition from hunting and gathering
to food production, both Binford (1968) and Flannery (1973) viewed
population expansion from optimal to sub-optimal zones of the Near
East as the key factor leading to economic change. The latter study
supplements a much-quoted examination of the transformation from
band society into tribes, chiefdoms, and ultimately states (Flannery
1972), in which population growth was highlighted as the cause rather
than the outcome of social evolution (for similar views, see Harner
1970, 1975; Kunstadter 1972, Smith 1972). Indeed, by the time of
the Santa Fe meeting on the dynamics of culture change in 1972,
population growth had emerged as the prime mover of world-wide
culture change, a lynch-pin of the general systems theory approach
of the early 1970's (Hill 1977). This reliance on 'population' as a
stimulus to change underlined the very real weakness in systems
theory for explaining internal change rather than equilibrium in
culture systems.
To analyse the more complex if more chronologically limited world
of contemporary subsistence farmers, researchers such as Basehart
(1973), Brookfield (1972), Clarke (1966) and Vermeer (1970) built
models of agricultural systems in which the Boserup model was
examined. The most valuable insight was perhaps that of Brookfield
(1972), who introduced environmental variables into the over-
simplified relationship between population density and agricultural
intensification. The latter indicated that, if population density were
held constant, the expectation would be that intensification would
be greater on optimal and marginal farming land than on moderate
land. These synchronie relations were not, however, tested statisti-
cally until the late 1970's (Brown and Podolefsky 1976, Turner et.
 Putting Pressures on Population 295
. 1977). Here the relationship between population density, agricul-
tural intensification and other intervening variables was examined,
using rank order analysis in the former, regression analysis in the
latter. In both cases, strong positive correlations were identified for
the relationship between population density and intensification. The
limitation of these approaches for the prehistorian is the lack of time
depth, which vitiates any attempt to relate measures of correlation
to arguments of causality.
Fortified perhaps by these conclusions, and despite the mounting
critique of population as a prime mover in socio-cultural evolution
(see below), the claims of the supporters of population pressure have
become more strident over the last decade. Two particular examples
are noteworthy: Cohen's attempt to explain the origins of food
production as a response to global population pressure (Cohen 1977)
and the efforts of the Cambridge palaeoeconomists to integrate a
demographic perspective into their neo-evolutionary studies ( Jarman,
Bailey and Jarman 1983).
Cohen is satisfied with no less than a global explanation of the
adoption of agriculture in all its diversity, and to that end, he builds
a challenging new theory of population pressure and carrying capac-
ity (1977). Population pressure is defined as an imbalance between
a population, its choice of foods and work standards which forces
a change in eating habits or increased work load (1977, 50). Cohen
dispenses with the idea of carrying capacity (1977, 48-50), claiming
that population pressure need not approach or even threaten this
threshold. He goes on to list a total of fourteen indicators of pressure
on a population (1977, 78-83).
Apart from his disregard of natural fluctuations in available resour-
ces and population sizes, Cohen adopts such a loose definition of
'population pressure', with its all-embracing criteria, that the concept
is in danger of being relegated to the role of epiphenomenon. His
version of 'population pressure' is indeed so diffuse that it can
plausibly be related to otherwise scattered and divergent trajectories
of cultural evolution in the varied regions in which food production
was taken up; but only as the lowest common denominator found
in many different processes of change. Cohen's theory is a catch-all,
and lies at the opposite end of the spectrum from the deterministic,
unitary optimization theories of the Cambridge palaeoeconomists.
For Jarman, Bailey and Jarman (1983), population pressure is
inherent in human and animal existence; economic systems are forced
by population pressure to adopt increasingly productive patterns of
exploitation. Birdsell (1953) is quoted approvingly when he suggests
that the population density of most Pleistocene situations approxi-
mated to operational saturation. On a theoretical plane, the main
evidence quoted for population pressure acting as a constant factor
296 CHAPMAN
is the selective advantage accruing to more efficient economies over
less efficient groups.
Group size is adduced in favour of population pressure, since why
regulate numbers if there is no pressure to stop greater population
size? Indeed, the palaeoeconomists attempt a pre-emptive strike in
their statement that 'we must be suspicious of general assumptions
of the absence [my italics] of population pressure or warfare at certain
levels of economic development' (1983,9). The methodological impli-
cation of the assumption of constant population pressure is the use
of optimization models in such techniques as Site Catchment Analy-
sis. As with the sociobiologists, the palaeoeconomists are distorting
Darwinian theory by smuggling in an assumption of optimal strategies
rather than the relative advantages prescribed in natural selection of
competing groups2.
These relatively pure formulations of the hypothesis of population
pressure and growth as a prime mover in cultural evolution stand in
marked contrast to some recent elaborations and refinements of the
core concepts of pressure and carrying capacity. All these contribu-
tions reflect to a greater or lesser extent current theoretical interests
in disequilibrium rather than equilibrium models.
Glassow's main contribution lies in formulating a formal relation-
ship between the concepts of population pressure and carrying capac-
ity, as well as other economic concepts (Glassow 1978). Fig. 2
envisages a hierarchical relationship in which population pressure
theory encompasses three lower-order theories: carrying-capacity
models, Hayden's (1975) resource over-exploitation models (in which
'carrying capacity' is modified to allow for disequilibria in resources
availability), and cost-comparison models relating changes in sub-
sistence strategy to the unit costs of production.

POPULATION PRESSURE
THEORY

CARRYING CAPACITY RESOURCE COST-COMPARISON

MODELS OVER-EXPLOITATION MODELS
MODELS

Figure 2. A hierarchy of population concepts (based on data in Glassow

1978).

Glassow (1978) is less helpful in distinguishing two different uses

of the term 'carrying
carrying capacity
capacity':: (1) an abstract concept min theoretical
instructs, and (2}
constructs, (2) a measure of snecific asnects of a relationship
specific aspects relationshio
 Putting Pressures on Population 297
between a population and its environment. Such a distinction
corresponds to no more than the difference between a concept and
its operationalisation. More useful is the classification of the divergent
uses to which the term 'carrying capacity' is put by economic
anthropologists (as the maximum ability of an environment to provide
a culturally-determined subsistence level) and ecologists (as the point
at which population growth experiences limits (Bates and Lees 1979).
The most significant difference in these two uses is that, while ecol-
ogists envisage more or less direct constraints on growth, an element
of cultural relativity is introduced into the equation by the economic
anthropologists.
A significant degree of cultural relativism is introduced by both
Anderson (1978) and O'Shea (1981) in their discussion of population
pressure and carrying capacity. For Anderson, there is a strong
contrast between the economic systems of hunter-gatherers and
agriculturalists: the former under a stable equilibrium model, where
population pressure is circumvented by negative feedback mechan-
isms, and the latter with a dynamic equilibrium model where popula-
tion pressure is lowered by density-dependent mechanisms increasing
the incidence of mortality (Anderson 1978, 34). The significance of
the expandibility of agricultural food supplies is held to explain this
contrast. More significantly, Anderson avers that hunter-gatherers
maintain their populations at between 20% and 50% of the actual
carrying capacity (1978, 32), thereby challenging the Cambridge
palaeoeconomists's claims of constant population pressure.
O'Shea (1981) directly addresses the notion of environmental
variations and the effect of resource unpredictability on prehistoric
economies. He argues that there are regular, predictable periods of
shortage for agriculturalists, for which there are coping strategies
(food storage, 'social storage', etc.), and to these regular shortages
Liebeg's law of the minimum does not apply. However, carrrying
capacity applies directly to population size and density in times of
unpredictable scarcity or severe food shortage. The ability of a region
to cope with such scarcity, O'Shea maintains, is closely related to
overall subsistence diversity and the degree of available social storage.
O'Shea's contribution is complementary to Anderson's in questioning
the validity of constant population pressure in societies which nor-
mally cope with routine shortages.
One of the question-marks against the O'Shea model is the ease
with which bulky foodstuffs can be moved within, let alone between,
regions under conditions of simple transport technology. Another
concerns the idealist assumptions that prehistoric groups will usually
be considerate enough to help their less fortunate relations.
In many of the recent refinements of the demographic paradigm
as simulated by Boserup and expanded by Binford, Flannery and
298 CHAPMAN
others, there has been a marked tendency away from the purist notion
of population growth as a prime mover, and a movement towards a
position of comfortable cultural relativism in which the concepts of
population pressure and carrying capacity are retained in a diluted,
sometimes distorted form. Two questions arise: do such concepts
retain validity in their original, undiluted forms? are such concepts
defensible in their modified forms?
It would appear that the refinement of these concepts has been a
dialectical process in which the original notions were modified
because of genuine theoretical difficulties (e.g. Glassow 1978), and
the resulting ideas were so far from the original population concepts
that they required to be embedded in a rather different theoretical
background (e.g. Cohen 1977). Or have the notions of population
pressure and carrying capacity outlived their usefulness in archaeo-
logical theory and practice?

A THEORETICAL CRITIQUE OF 'POPULATION PRESSURE'

AND 'CARRYING CAPACITY'
Over the last decade, archaeologists and anthropologists have made
determined attempts to undermine the theoretical validity and
methodological utility of the demographic paradigm. On a theoretical
front, the attacks have embodied Renfrew's comment that social
attitudes and relations should govern or influence attitudes to popula-
tion growth, which can no longer be considered as a prime mover
(1982). We shall turn to this point shortly.
In terms of methodology, the main criticisms have focussed on
problems of operationalising these concepts. Plog (1975) notes that
it is a circular argument to use estimated biomass to calculate prehis-
toric populations and then use both data sets to discuss population-
environment relationships (cf criticisms by Bates and Lees 1979).
Equally, Hayden (1975) and Ellen (1978) have outlined the difficulties
of accurate quantifiable measurements of carrying capacity, in view
of the variable, complex and cyclical nature of most resource environ-
ments and the effects of agriculture on soil fertility (cf. Street 1969).
Hayden (1975) has extended his critique to Liebeg's law of the
minimum and its relationship to cyclic resource disequilibria of
varying magnitudes. Lastly, Bayliss-Smith (1978) has denied the
possibility of any quantitative definition of population pressure infer-
ring that it is dangerous to use the term. And, in any case, Moore
(1981) reminds us that energy and calories may be convenient units
of measurement but survival involves additional dimensions of
behaviour.
Supporters of population pressure theories have only one defence
to offer to such charges, namely that merely because the archaeologi-
cal or environmental data base is biased, unrepresentative or partial,
 Putting Pressures on Population 299
this is no reason to reject a theoretically valid concept. So it is to the
realm of theory that we turn.
Perhaps the most controversial area of the work of Malthus,
Boserup, and the Cambridge palaeoeconomists is their assumption
that subsistence, resources, population and scarcity are all absolute
concepts whose very meaning is taken as given by the empirical form
of enquiry in which the discussion is framed, and whose interrelation-
ships can be studied to provide 'natural laws' of population (e.g.
Malthus' law through which population inevitably places pressure
on the means of subsistence). Contrary to such empiricist conclusions,
Marxist authors stress the relational rather than the absolute view of
the world, in which there is no such thing as 'a resource' in abstract
but rather 'resources' which can be defined only by reference to the
mode of production which seeks to use them (Oilman 1971). Just as
each type of social relations his and each structural level in society
are subject to the functioning of specific demographic conditions, so
each demographic structure has a causal effect on the functioning
of different societies (Godelier 1975).
Several inferences can be drawn from a dialectical approach to
population. First, it is clear that there can be no eternal law of
population, since each population-to-environment relationship is
related to a specific mode of production. Secondly, subsistence cannot
be regarded as absolute, as Malthus apparently believed, but is
relative to socially and culturally determined needs as well as physio-
logical requirements (Orans 1966). Thirdly, resources can only be
defined with respect to a particular technical, cultural and historical
stage of development (Harvey 1974). Fourthly, scarcity cannot, pace
O'Shea (1981), be assumed to be inherent in nature but rather is
inextricably social and cultural in origin. It is social aims and objec-
tives which define scarcity.
The anthropological evidence of culturally relative strategies of
production, distribution and consumption is voluminous. On a gen-
eral level, the principle has been established that most aspects of
economic behaviour are related to a culture-specific set of core beliefs,
classifications and strategies whose purpose is to make the world
explicable as much as to provide satisfaction for physiological needs.4
Rappaport (1967) has provided the most general statement so far
on this position, in the form of his distinction between the cognized
(emic) and operational (etic) environment. For Rappaport, E0 is the
operational model of the environment, as measured by archaeologists
and anthropologists, whilst Ec is the cognized model of reality as
defined by the individual or group under study. Ec overlaps with E0,
is often simpler than E0 yet may contain elements not found in E0.
It is important to remember that behaviour is in more direct response
to Ec than to E0; it is assumed that E0 is never identical to Ec, since
300 CHAPMAN
perfect tracking systems of the environment have not yet been
developed. Conversely, wide discrepancies between Ec and E0 are
unlikely to persist, being potentially maladaptive (Fig. 3).
On this basis it would be difficult to talk about the degree to which
hunters or collectors were specialized or generalized, unless there
was clear evidence of the original biomass, part of E0. Some creative
research is attempting to establish E0 by a variety of means.5 By
contrast, the domesticated portion of food consumption is by
definition a set of deliberate cultural choices which are more likely
to be reflected in total consumption patterns. If anything, the concepts
of Ec and E0 introduce an extra level of complexity into the population
environment interaction, but one which is an essential component
of the relativisation of these two concepts:
The sceptical population-purist's rejoinder to these objections may
well be to accept a relativistic definition of population pressure and
carrying capacity as quite acceptable, either in itself or as a welcome
tool for the application of natural selection processes through the
medium of reproductive success. But there are both cultural and
biological objections to the single-minded optimization of reproduc-
tive success, whether in human groups of over 100 individuals, or in
rich environments or in the creation of new selective pressures during
cultural change. Durham (1979) has argued that relatively beneficial
compromises between individual and group selection pressures are
more likely than attempts at universal adaptive optimization, which
are constrained by the pull of social against individual behaviours.
Similarly, Caplan (1980) asserts that, because of the multiplicity of

1 2 3 4 5 6 7 8 9 10 11 12 TOTAL AVAILABLE
EDIBLE BIOMASS

4 5 6 7 8 9 BIOMASS RECOGNIZED
AS EDIBLE

CULTURAL PREFERRED STAPLES

CHOICE

FOOD CONSUMPTION
REFUSE AND DISCARD

MATERIAL ARCHAEOLOGICAL
REMAINS EXCAVATION

Figure 3. The relationship between the operational environment (E 0 ), the

recognised environment (E c ) and archaeological discovery. [Each number
in the boxes represents a different species].
 Putting Pressures on Population 301
selectional demands (environmental, biotic, inter-specific and con-
specific), a great variety of alternative behaviours can viably co-exist
in resource-rich environments without ever being selected on prin-
ciples of optimization.
Fuller (1978) has speculated that, in groups of over 100 humans,
factors such as status, behavioural traits, genotypes and reproductive
success may vary independent of each member's behaviour; and to
the extent that they so vary, optimization theory will prove to be
inadequate on an explanatory level.
Finally, it is increasingly difficult to maintain the proposition that,
for a given environment and level of technology, there is but one
single optimally adaptive solution to a subsistence problem.6 Hence,
the greater the density of cultural factors inter-relating with physio-
logical mechanisms of survival and reproduction, the greater the
frequency of viable alternatives to optimization strategies of human
behaviour. In short, Optimization' is as subject to cultural relativity
as other component parts of the demographic paradigm.
Accepting this relativistic framework at face value, what difference
will it make to prehistoric economic strategies in relation to the
environment? An initial answer is to examine flexible social responses
to resource scarcity focussing on tribal societies. Laughlin and Brady
(1978) devote a book to the proposition that, when human groups
are faced with disaster, unremitting deprivation, or even scarcity, the
key response lies in social co-operation. Laughlin's (1974) refinement
of the concept of reciprocity in terms of an 'accordian effect' is
illuminating. With abundant resources, reciprocity tends towards the
balanced state in which exploitative exchanges tend to be avoided,
the closer the kin. But in bad times, there is a tendency towards
negative reciprocity, a decline in food-sharing between kin, higher
dispute frequencies and consequent increase in ritual activity and
the need for elders to adjudicate. A similar picture is found in Sierra
Leone, where, in tribal societies, a substantial part of basic annual
production is devoted to ceremonial and feasting in good years
(MacCormack 1978). However, when crop yields are low, such cer-
emonies are deferred and the foodstuffs used as a communal store.9
These instances reflect the tendency for tribal economies to produce
not merely for survival, but also for seed corn, food reserves and
ceremonial/ritual consumption as a norm. The significance of
buffering strategies in tribal economies has been clearly identified
(Hassan 1978; Anderson 1978; cf. recent TAG 84 symposium on
buffering strategies). Hassan (1978) has formalised the buffering
alternatives to constant population pressure in his Economic-Demo-
graphic Equilibrium (or EDE) model (Fig. 4).
The important question remains: what is it that stimulates economic
demand for optimization ? The evidence for population growth world-
wide is incontrovertible, so what are the social or institutional factors
302 CHAPMAN
100%
CARRYING BUFFER ZONE
CAPACITY SHORT TIME TO RESPOND HIGH-RISK OPTIMIZING
IF RESOURCES FAIL STRATEGY

50%
SECURE ZONE LOW-RISK BUFFERING
LONG TIME TO RESPOND STRATEGY
IF RESOURCES FAIL

Figure 4. An economic-demographic equilibrium model (based on Hassan

1978).

which stimulate it, damp down population increases or lead to

agricultural intensification in egalitarian societies?
THE SOCIAL ALTERNATIVE
An alternative viewpoint begins with a consideration of individuals
not merely as breeding units but also as contributors of labour to a
given mode of production. As an important part of the means of
production, the number of individuals available for work is one of
the key determinants of the scale of the economy and the scope of
investments in time and energy, if any.8 In this respect Harner (1975)
has proposed two alternative models for man-environment relation-
ships:- (1) a situation where land is plentiful and labour is scarce,
(2) a situation where land is scarce and labour is plentiful. Harner
argues that land and labour are rarely in balance and that whichever
of the two is in short supply becomes the object of competition and
the keystone of the search for power and prestige. In diachronic
terms, in the earlier stages of food-production in Europe, fertile land
was still abundant, but was becoming increasingly scarce by the time
of early state-formation (Bintliff 1984, Chapman in press, a, Milisaus-
kas and Kruk 1984, Sherratt 1980, Starling 1985).
In the first case, where land exceeds labour, there is a positive
advantage for group leaders to add to the labour pool, either through
breeding or alliance,9 so as to increase the number of producers of
resources available to the group for conversion into prestige goods,
etc. (cf. Chagnon 1979). Clearly, population growth will be
encouraged by cultural specifics, such as the allocation of rewards
and other incentives by group leaders (Cowgill 1975), or the social
demands on production (Bender 1985).
An elegant example of this argument is the hypothesis that popula-
tion growth in evolving societies is most pronounced in groups where
the Domestic Mode of Production10 is in operation and remains
partially reliant on children's labour (Blanton 1975). The key idea
is, that since the family is the main source of domestic labour for
 Putting Pressures on Population 303
production, there is a tendency to increase family size, held back
only by factors such as the economic cost of dependent children and
higher dispute rates in larger families. It should be noted that
throughout his discussion Blanton considers such population growth
as remaining well below any notional carrying capacity.
An alternative context, in which population growth to provide
additional labour is critical, is the agricultural colonization of temper-
ate landscapes hitherto supporting groups of foragers (Green 1980,
Green this volume; Chapman in press, b). Green's model of coloniz-
ation posits a series of unbalanced family groups, whose demographic
characteristics of rapid growth and predominance of young males
lead to the development of exchange relations involving gifts and
ultimately mates with local foragers. The crucial problem for coloniz-
ing populations is the lack of labour at important times of the
agricultural cycle, a shortage which prompts alliance formation and
the gradual restoration of more balanced demographic characteris-
tics. In this example, the social conditions of labour-demand in
colonizing populations can be responsible for stimulating a logistical
type of population growth, whose pace slows down not because of
any notional carrying capacity but because suitable levels of agricul-
tural labour have been reached. An archaeological example of the
lack of population increase consequent upon the stabilization of
colonizing agricultural groups is the Middle and Late Neolithic of
Central Germany (Starling 1985). Similarly, analysis of Neolithic
population sizes and landuse potential in Little Poland led Milisaus-
kas and Kruk to conclude absence of population pressure throughout
the Neolithic (1984, 15-18). A third example of the desirability of
increased population in the context of excess land over labour is the
case of the expanding segmentary lineage formation (Sahlins 1961,
Vayda 1961). Using swidden agriculture in a forested environment,
primary economic segments of up to 250 people were integrated by
means of 'mechanical solidarity' into tribal units exhibiting internal
ranking defined in relationship to ancestor figures. One of the main
strategies for external expansion and increases in internal densities
in such groups is population growth through rapid fissioning of
structurally similar units from established farming core areas. An
example postulated for European prehistory is the Battle Axe culture
of southern Jutland, c 2800-2300 BC (Kristiansen 1982). Here a
rapid expansion of small pastoral families from the now less fertile
core arable lands of the earlier farmers took place onto the lighter
interfluvial soils. After the initial expansion in which rapid population
growth was advantageous, a period of flux ensued, defined by cyclical
expansion and contraction in the same areas.
This example is valuable in introducing the notion of cyclic change
into population considerations, as an alternative to the assumption
304 CHAPMAN
of steady but continuous unidirectional population increase since
the Pleistocene. In addition, it specifies a type of social formation
in which the imbalance between land and labour fluctuates through
time, providing an additional social dynamic.
The second case, where there is an imbalance of labour over land,
provides a precise definition of population pressure. But how soon
did this occur in European prehistory? The first clear settlement
evidence for colonization and intensive utilization of large areas of
marginal land, as well as all the major areas of core agricultural soils,
is the later Bronze Age, c2500-1000 BC (Bintliff 1982, 1984, Coles
and Harding 1979, Harding 1984); a peak of settlement density
possibly not repeated in temperate Europe until the high Medieval
period (c. AD 1300-1500). The enclosure of large areas of land and
the construction of defended centres at nodal points in the landscape
gives clear evidence for competition over, and control of, the most
crucial but scarce resource; land. At this juncture in the trajectory
of European socio-economic development, Boserup's hypothesis of
agricultural intensification finally becomes relevant, since land
efficiency the extent to which the subsistence technology reduces
the quantity of land needed for each unit of production is directly
related to production for conversion to prestige goods; the ultimate
aim of Bronze Age elites (Bradley 1984).

The foregoing analysis has demonstrated that, although methodologi-

cally vacuous and empirically almost impossible to quantify, the
concepts of population pressure and carrying capacity cannot be
readily despatched to the refuse-tray of redundant theory. Both terms
appear to take on new life once it is accepted that their only place
lies within an explanatory framework of cultural relativity, defined
in terms of specific social structures. The case for continuous popula-
tion pressure as a norm of human behaviour can no longer be
maintained; alternative explanations of this primus inter pares of
prime movers require urgent development for many major aspects
of cultural evolution. It is surely sensible, pace Bettinger (1980), to
view population as a dependent variable in periods of long-term
stability, such as much of the Pleistocene, and to explore the possible
effects of population growth in contexts of more rapid change. With
the introduction of farming, there arose social formations, such as
the lineage-based tribe, some with household units of production,
and the segmentary lineage-based conical clan, in which the positive
incentives for increasing family size and overall population densities
outweighed the resultant economic and social disadvantages. The
concept of population pressure may well play a more significant
theoretical role in discussion of periods where there is a demonstrable
imbalance of labour over land. An example in some areas of European
 Putting Pressures on Population 305
prehistory is the Late Bronze Age, where a shift in status competition
from prestige goods to land is suggested, leading to population
increase, agricultural intensification and major political alliance
building.
The dialectical relationship between population and social struc-
ture is one of the key stimulants to cultural change. In an archaeologi-
cal, and therefore diachronic, context, this relationship is but one
facet of the paradox of human agency; the fact that history and
society are made by constant, more or less purposive human action
and that, simultaneously, individual action, however purposive, is
made by history and society. The most powerful tool for illuminating
this paradox is the study of long-term culture process.

Acknowledgements: My grateful thanks to John Bintliff, Jim Moore,

John Lazarus, Jeremy Johns and Peter Fowler for commenting on
previous drafts of the paper and helpful discussions. Thanks too to
Cathy Rutherford for work on the word-processor and to the Univer-
sity of Newcastle upon Tyne Audio Visual Centre for drawing the
illustrations.

NOTES
1. For a discussion of the literature on the effects of population pressure on
agriculture prior to 1965, see Turner et al (1977).
2. An identical procedure is followed by those who have applied the theory of
Optimal Foraging Strategy to prehistoric economies (e.g. Christenson (1980), who
argues that population pressure is directly related to variations in niche-width). For
a recent auto-critique of a hitherto Optimal Foraging Strategist see Keene (1983;
cf. Keene 1981). See also Bellinger (1980, pp. 208-210).
3. Il should be noted lhal the Cambridge palaeoeconomists altack social
archaeology for its paucity of evidence on art, religion and social struclure (Jarman
et al. 1983), whilsl placing their trust in such empirically indefinable concepts as
population pressure and carrying capacity.
4. Three examples of relativislic behaviour conslilule a sample of many:-(l)
Hassan's (1975) assertion lhat man's recognition of whal is 'food' is far from
cross-culturally uniform; (2) Bronson's (1972) discussion of how 'effort', 'labour-
inlensificalion' and indeed Ihe 'law of Minimum Effort' itself are all socially relative;
(3) Douglas and Isherwood's (1978) picture of consumption goods as the hardware
and software of an informalion system whose principal aim is to monitor its own
performance, and their definition of consumption as rilual - an aclive process in
which all social calegories are being redefined.
5. See Deilh and Shackleton, this volume; J. Rackham's unpublished research
on Pleistocene faunas as represented in natural deposits such as river gravels.
6. In primatology, Tallersall (1977) has recenlly concluded that any primate
species is capable of more than one adaptive solution to any given environmental
situalion and lhal Ihe choice amongs them may be due as much to chance as to
anything else.
7. Cf. Gledhill's (1978) account of the Trobrianders, who used yams as prestige
goods in good years, as a food reserve in poorer years.
8. See Blanton (1975) for discussions of family size and economic stralegies.
306 CHAPMAN
9. Cf. Levi-Strass' (1960) comment that, in certain New Guinea tribes, the purpose
of marriage is to obtain not so much a wife as extra brothers-in-law.
10. The Domestic Mode of Production defines an economic system whereby the
domestic household acts as the basic unit of production, distribution and consump-
tion (Chayanov 1966, Sahlins 1974, Tringham n.d.).

BIBLIOGRAPHY
Alexander, R. D. (1974) The evolution of social behaviour,
Annual Review of Ecology and Systematics 5, 325-383.
Anderson, A. (1978) The role of a competition model in the
archaeological explanation of economic change, pp. 31-45, in
Green, D., Haselgrove, C. & Spriggs, M. (eds). Social
Organization and Settlement. Oxford, BAR S-47.
Basehart, H. W. (1973) Cultivation intensity, settlement patterns
and homestead forms among the Matengo of Tanzania,
Ethnology 12, 57-73.
Bates, D. G. & S. H. Lees (1979) The myth of population
regulation, pp. 273-889, in Chagnon N. A. and W. Irons,
(eds). Evolutionary biology and human social behaviour.
Sciutate, Duxbury.
Bayliss-Smith, T. (1978) Maximum populations and standard
populations: the carrying capacity question, pp. 129-151, in
Green et al. (eds). Social organisation and settlement. Oxford,
BAR S-47.
Bender, B. (1985) Emergent tribal formations in the American
Midcontinent. Am. Antiquity 50, 52-62.
Bellinger, R. L. (1980) Explanalory/Prediclive models of hunler-
galherer adaplalion. Advances in Arch. Method & Theory 3,
189-255.
Binford, L. R. (1968) Posl-Pleistocene adaptalions, pp. 313-341, in
Binford L. R. and S. R. (eds). New perspective in archaeology.
Chicago, Aldine.
Binlliff, J. L. (1982) Seulement patlerns, land tenure and social
structure: a diachronic model, pp. 106-111, in Renfrew, C.
and Shennan, S., (eds). Ranking, resource and exchange.
Cambridge, CUP.
Bintliff, J. L. (ed). (1984) European social evolution: Archaeological
perspectives. Bradford, Bradford UP.
Birdsell, J. B. (1953) Some environmenlal and cultural faclors
influencing Ihe structure of Australian aboriginal populations.
American Naturalist 87, 171-207.
Blanton, R. E. (1975) The cybernetic analysis of human
populalion growlh', pp. 116-126, in Swedlund A. C. (ed).
Population studies in archaeology and biological anthropology.
Memoirs of SAA No. 30:
Boserup, E. (1965) The conditions for agricultural growth. Chicago,
Aldine.
Boulding, K. (1946) The economics of peace. London, M. Joseph.
Bradley, R. (1984) The social foundations of prehistoric Britain.
London, Longmans.
Bronson, B. (1972) Farm labor and the evolution of land
production, pp. 190-218, in Spooner, B. (ed). Population
Growth. MIT Press.
 Putting Pressures on Population 307
Brookfield, H. C. (1972) Intensification and disintensification in
Pacific agriculture: a theoretical approach. Pacific Viewpoint
13, 30-48.
Brown, P. & A. Podolefsky (1976) Population density, agricultural
intensity, land tenure and group size in the New Guinea
Highlands. Ethnology 15, 211-238.
Caplan, A. (1980) A critical examination of current sociobiologi-
cal theory: adequacy and implications, pp. 97-121, in Barlow
G.W. and Silverberg, J., (eds). Sociobiology: beyond nature.
AAAS Selected Symposium 35. Washington DC, Westview
Press.
Carr-Saunders, A. M. (1922) The population problem, a study in
human evolution. Oxford, Clarendon.
Chagnon, N. A. (1979) Is reproductive success equal in egalitarian
societies?, pp. 374-401, in Chagnon, N. A. and W. Irons,
(eds). Evolutionary biology and human social behaviour.
Sciutate, Duxbury.
Chapman, J.C., in press a, The development of Balkan agriculture
and pastoralism.
Chapman, J.C., in press b, Demographic trends in Neothermal
South-east Europe. Paper presented to Edinburgh Mesolithic
Symposium.
Chayanov, A. V. (1966) (1923) The Theory of Peasant Economy.
Irwin.
Christenson, A. L. (1980) Change in human food niche in
response to population growth, pp. 31-72, in Earle T.K. and
Christenson, A.L. (eds). Modelling change in prehistoric sub-
sistence economies. London, Academic Press.
Clarke, W. C. (1966) From extensive to intensive shifting cultiva-
tion: a succession from New Guinea. Ethnology 5, 347-359.
Clutton-Brock, T. H. (1974) Primate social organization and
ecology. Nature (London) 250, 539-542.
Clutton-Brock, T. H. & P. H. Harvey (1976) Evolutionary rules
and primate societies, pp. 195-237, in Bateson, P.P.G. and
Hinde, R.A., (eds). Growing points in ethology. Cambridge,
CUP.
Clutton-Brock, T. H. & O. H. Harvey (1977) Primate ecology and
social organisation, J. of Zoology 185, 1-39.
Cohen, M. N. (1977) The food crisis in prehistory. Yale Univ.
Press.
Coles, J. and A. F. Harding (1979) The Bronze Age in Europe.
London, Methuen.
Cowgill, G. (1975) Population pressure as a non-explanation, pp.
127-31, in Swedlund, A., (ed.) Population studies in
archaeology and biological anthropology. Memoirs of SAA No.
30.
Darwin, C. (1859) On the origin of species by means of natural
selection. London, J. Murray.
Douglas, M. & Baron Isherwood (1978) The world of goods. Lon-
don, Penguin.
Dumond, D. E. (1965) Population growth and cultural change, S.
W. Journal of Anthropology 21, 302-325.
Durham, W. H. (1979) Towards a co-evolutionary theory of
human biology and culture, pp. 39-59, in Chagnon, N.A.,
308 CHAPMAN
and Irons, W. (eds). Evolutionary biology and human social
behaviour. Sciutate, Duxbury.
Ellen, R. F. (1978) Ecological perspectives on social behaviour,
pp. 121-127, in Green, D. Haselgrove, C. and Spriggs, M.,
(eds). Social organisation and settlement. Oxford, BAR S-47.
Flannery, K. V. (1972) The cultural evolution of civilisation. Ann.
Review of Ecology and Systemalics 3, 399-426.
Flannery, K. V. (1973) The origins of agriculture. Annual Review
of Anthropology 2, 271-310.
Freedman, D. G. (1979) Human sociobiology. A holistic approach.
New York, Free Press.
Fuller, J. L. (1978) Genes, brains and behaviour, pp. 98-115, in
Gregory, M. S. A. Silvers and D. Sutch, (eds). Sociobiology
and human nature. San Francisco, Jossey-Bass.
Glassow, M. A. (1978) The concept of carrying capacity in the
study of culture process. Advances in Arch. Method & Theory
1 31-48.
Gledhill, J. (1978) Formative developments in the North
American Southwest, pp. 241-290, in Green, D. C.
Haselgrove and M. Spriggs, (eds). Social organisation and
settlement. Oxford, BAR S-47.
Godelier, M. (1975) Modes of production, kinship and
demographic structure pp. 3-27, in Bloch, M. (ed). Marxist
Analyses and Social Anthropology. ASA Studies 2. London.
Green, S. W. (1980) Broadening least-cost models for expanding
agricultural systems, pp. 209-241 in T. Earle and A.
Christenson (eds). Modelling Change in Prehistoric Subsistence
Economies. London, Academic Press.
Hamilton, W. D. (1964) The genetical evolution of social
behaviour. J. of Theoretical Biology 7, 1-16.
Harding, A. (1984) Aspects of social evolution in the Bronze Age,
pp. 135-145 in Bintliff, J., (ed). European Social Evolution.
Bradford, Univ. of Bradford Press.
Harner, M. J. (1970) Population pressure and the social evolution
of agriculturalists. S.W. Journal of Anthropology 26, 67-86.
Harner, M. J. (1975) Scarcity, the factors of production and social
evolution, pp. 124-138, in Polgar, S., (ed). Population, ecology
and social evolution. The Hague, Mouton.
Harvey, D. (1974) Population, resources, and the ideology of
science. Economic Geography 50, 256-277.
Hassan, F. (1975) Determination of the size, density and growth
rate of hunting-gathering populations, pp. 27-52, in Polgar,
S., (ed). Population, ecology and social evolution. The Hague,
Mouton.
Hassan, F. (1978) Demographic archaeology. London, Academic
Press.
Hayden, B. (1975) The carrying capacity dilemma: an alternative
approach, pp 11-21, in Swedlund, A.C., (ed). Population
studies in archaeology and biological anthropology: a
symposium. SAA Memoirs 30.
Hill, J. (ed). (1977) The explanation of prehistoric change.
Albuquerque, Univ. of New Mexico Press.
Irons, H. (1979) Investment and primary social dyads, pp. 181-
213, in Chagnon, N.A. and Irons, W., (eds). Evolutionary
 Putting Pressures on Population 309
biology and human social behaviour. An anthropological
perspective. Sciutate, Duxbury.
Jarman, M. R., G. N. Bailey & H. N. Jarman (1983) Early
European Agriculture. Cambridge, Cambridge Univ. Press.
Keene, A. (1981) Prehistoric foraging in a temperate forest: a linear
program model. New York, Academic Press.
Keene, A., 1983 'Biology, behaviour and borrowing, a critical
examination of Optimal Foraging Theory in archaeology, pp.
137-155, in Moore, J. and A. Keene (eds). Archaeological
h >mmers and theories. London, Academic Press.
Kristiansen, K. (1982) The formation of tribal systems in later
European prehistory: Northern Europe, 4000-500 BC., pp.
241-280 in Renfrew, C. Rowlands, M.J. and Segraves, B.A.,
(eds). Theory and explanation in archaeology. London,
Academic Press.
Kunstadter, P. (1972) Demography, ecology, social structure and
settlement patterns, pp. 313-351, in Harrison, G.A. and A. J.
Boyce (eds). The structure of human populations. Oxford,
Clarendon.
Laughlin, C. D. Jnr. (1974) Deprivation and reciprocity. Man 9,
380-396.
Laughlin, C. D. Jnr. & I. A. Brady (eds). (1978) Extinction and
survival in human populations. New York, Columbia, U.P.
Levi-Strauss, C. (1960) The family, pp. 261-285, in Shapiro, H. L.
(ed). Man, culture and society. Oxford, Oxford Univ. Press.
MacCormack, C. P. (1978) The cultural ecology of production,
Sherbro coast and hinterland, Sierra Leone, pp. 197-210, in
Green, D., C. Haselgrove and M. Spriggs (eds). Social organi-
sation and settlement. Oxford, BAR S-47.
Malthus, T. (1798) An essay on the principle of population 1st edi-
tion. London, J. Murray.
Malthus, T. (1807) An essay on the principle of population 2nd edi-
tion. London, J. Murray.
Milisauskas, S & J. Kruk (1984) Settlement organisation and the
appearance of low level hierarchical societies during the Neo-
lithic in the Bronocice Microregion, Southeastern Poland.
Germania 62, 1-30.
Moore, J. A. (1981) The effects of information networks in hunter-
gatherer societies, pp. 194-217, in Winterhalder, B. and A. E.
Smith (eds). Hunter-gatherer foraging strategies. Chicago,
Univ. of Chicago Press.
Oilman, B. (1971) Alienation: Marx's conception of man in capital-
ist society. London, Cambridge Univ. Press.
Orans, M. (1966) Surplus. Human Organisation 25, 24-32.
O'Shea, J. (1981) Coping with scarcity, pp. 167-183, in Sheridan,
A. and G. N. Bailey (eds). Economic archaeology. Towards an
integration of ecological and social approaches. Oxford, BAR
1-96.
Plog, F. (1975) Demographic studies in Southwestern prehistory,
pp. 94-103, in Swedlund, A. (ed). Population studies in
archaeology and biological anthropology. SAA Memoirs 30.
Rappaport, R. (1967) Pigs for the ancestors. New Haven, Yale
Univ. Press.
Renfrew, C. (1982) Explanation revisited, pp. 1-23, in Renfrew,
310 CHAPMAN
C., M. R. Rowlands and B. A. Segraves (eds). Theory and
explanation in archaeology. London, Academic Press.
Sahlins, M. D. (1961) The segmentary lineage: an organisation of
predatory expansion. Amer. Anthrop. 63, 322-345.
Sahlins, M. D. (1974) Stone Age economics. Chicago, Aldine.
Sherratt, A. (1980) Water, soil and seasonality in early cereal cul-
tivation. World Arch. 11, 313-330.
Smith, P. E. L. (1972) Changes in population pressure in archaeo-
logical explanation. World Arch. 4, 5-18.
Spooner, B. (ed) (1972) Population growth. Anthropological impli-
cations. Cambridge, Mass., MIT Press.
Starling, N. J. (1985) Social change in the Later neolithic of
Central Europe. Antiquity lix, 30-38.
Street, J. M. (1969) An evaluation of the concept of carrying
capacity. Professional Geographer xxi, 104-107.
Tattersall, I. (1977) Ecology and behaviour of Lemur fulvus
mayottensis (Primates, Lemuriformes). Anthrop. Papers of the
Amer. Museum of Nat. Hist., 54/4.
Tringham, R., n.d. The development of the household as the
primary unit of production in Neolithic and Eneolithic South
East Europe. Paper delivered to 48th. SAA meetings,
Pittsburgh (1983)!
Trivers, R. L. (1971) The evolution of reciprocal altruism. Qtly.
Review of Biology 46, 35-39 and 45-47.
Turner, B. L. II, R. Q. Hanham & A. V. Portararo (1977)
Population pressure and agricultural intensity. Annals of the
Assoc. of Amer. Geographers 67/3, 384-396.
Vayda, A. (1961) Expansion and warfare amongst swidden
agriculturalists. Amer. Anthrop. 63 346-358.
Vermeer, D. E. (1970) Population pressure and crop rotational
changes among the Tiv of Nigeria. Annals of the Assoc. of
Amer. Geographers 60, 299-314.
Williams, G. C. (1966) Adaptation and natural selection: a critique
of some current evolutionary thought. Princeton, Princeton
Univ. Press.
Wilson, E. O. (1975) Sociobiology. The New Synthesis. Cambridge,
Mass., Bellknap P.
Wynne-Edwards, V. C. (1962) Animal dispersal in relation to social
behaviour. Edinburgh, Oliver and Boyd.
 INDEX

Aamosen, Denmark, 258, 265 red deer, 52, 53, 68

Achnacree, Black moss of, Argyll, Scotland, sheep, 52, 53, 55, 274
40 animal diseases and arthropathies, 274, 275
Adams, R McC, on spatial and temporal con- animal husbandry, 45,52, 55,56,177,274,275
texts, 1-15 animal populations, growth rates and sex
agriculture, see also under animal husbandry, ratios in, 280, 281, 292
cultivation.grazing etc, 41, 45, 46, 74, 76, Annales:
138, 140 de Geographie, 162
altitudinal limits to, New Guinea High- de Economies, Societies, Civilisations, see
lands, 153-159 also Annalist, 161, 168-176
chestnut culturing, 177, 178 Annalist tradition of Historical Geography,
cultural variation in, 66, 67, 68, 69 168-176, 177
Denmark, Neolithic introduction to, 252, anthropogenic alluviation, 211, 212, 214, 216,
253, 254, 255, 258-269 223-229
diseases in, 274, 275, 276 anthropogenic erosion, 193,194,202,205-209
ecosystems in, 248-255 anthropogenic modification of soils, 25, 27,
effects on artefact distribution, 85-89 29-35
impact of on environment, 190 artefact distribution analysis, see also under
intensification of re population pressures, Spatial analytical techniques, 82-90, 138
294, 295, 303 artefacts:
New Guinea, problems in defining limits adzes, 145, 148
to growth, 153-160 ards, 76
origins of, 278, 281 axes, stone, 131, 132
pastoral activities, 146, 148 biface fragments, 29
prehistoric re palynology and human flakes, 29
ecology, 190, 246-272 palaeolithic assemblages, 236, 237, 238,
slash and burn, 249 239, 240
spatial limits of exploitation, 66, 67 querns, 75, 76
swiddens, 154, 155, 156, 158, 159, 250, 303 spade, 154, 155
terracing, 69 Assur, Iraq, 8, 9
transition to from foraging, S Argolid, Assyrian Period, Tell Seh Hamad area, Syria,
Greece, 49, 52, 53-55 111, 113, 114-118, 120-123, 126
zoning in, rainfall as indicator of, 110, 111 astronomy, observation in, relevance to
alluviation, anthropogenic and cusp catas- environmental archaeology 2-3
trophe theory, 223-229 Aztecs, maize disease, 275
valley, impact of man on, see also valley
alluviation, 211-232
Ambrona, Spain, 235
Anatolia, 8, 9 Bandkeramik:
Anglo-Saxon cemeteries, distribution of, 129 integrative model in landscape reconstruc-
animal bone: 73, 74, 79, 80 tion, 145-150
cattle, 53, 55, 56 landscapes reconstruction of, 145-150
fallow deer, 68 palynologial investigations, 146, 147
goat, 52, 53, 55 pottery 145
Pig, 52, 53, 55, 56, 275 settlement sites, 145, 148, 149, 150
312 INDEX
Bayliss-Smith, T: 92, 93, 298, 299 catchment analysis, see under Spatial Analyti-
on prehistorie agriculture in New Guinea cal techniques
Highlands, 153-160 cereal processing, 74-75
bedouins, 111, 116, 117 cereal remains: see also under pollen, 53, 64,
Berlin, Free University of, 108 68
bibliographies: Barley, 53
Section 1 14-15, 23-24, 35-37, 47-48, 56- Emmer, 53
58, 70-71, 80-81, 89-90 Oats, 53
Section 2 106-107, 126-128, 142-144, 159- Chapman, J: 192
160, 181-186 on putting pressures on population:
Section 3 209-210, 229-232, 243-245, 270- alternatives to Malthusu and Boserup, 292-
272, 276-277, 281-282, 290, 306-310 310
Bintliff, J: 6, 10 Chayanov, A V, 67
introduction to conceptual issues on micro- chemical analysis of soils, 20, 21, 25-37
to macro- region, 92-93 chestnut culture, 177, 178
on site patterning: Chevdar, Bulgaria, 276
separating environmental cultural and pre- Childe, V G, 278
servation factors, 129-144 city-states, Classical Greek, 135, 140, 141, 142
Boetia, Greece, regional survey, 135-142 Classical period, Greece, rural settlement,
Boserup, E, 192, 293, 298, 299 135-142
Boucher de Perthes, J, 234 Cleddau Valley, Wales, 72, 74, 75, 76
boundary definition, 87, 88, 89, 96, 99, 101 climate:
Boyle, K V: 191 effects of on alluviation rates, 215-216
on foraging theory, mathematical evidence re Habur Valley, Syria, 110, 111,
modelling in, 283-290 114, 116, 117
Broekfeld,Germany, Bandkeramik site, pol- cluster analysis, 21
len analysis, 146, 147 Cohen, M N, 295, 298
Bronze Age: concepts, Historical-Geographical relevant to
colonization and settlement density in, 304, Archaeology, review of,161-186
305 contexts, spatial and temporal, 1-13
erosion in British river valleys, 211,212,213 Copeland, L: 190
Habur Valley, Syria, 111, 116,117,118-122, on cultural and environmental change,
126 palaeolithic, geomorphological research
rock carvings, Yorkshire, 129, 130 re, 233-245
Brothwell, D: 191 core sampling, O16/O18 method in, 235
on disease and environmental archaeology, Corsica, Historical-Geographic concepts rel-
273-277 evant to, 176-181
Bulgaria, Chevdar, 276 Coygan Camp, Wales, 73, 74, 75, 76, 79
Bunds So, Denmark, 258, 266 Cuckoo Grove Rath, Wales, 77, 78
Burrin, P J: 9, 189, 190 Cul a' Bhaile, Jura, environmental reconstruc-
on environmental thresholds, catastrophe tion at, 21, 38-47
theory and landscape, re valley alluvi- cultivation, see also under agriculture, 41, 46,
ation, 211-232 53, 74, 85, 88
Butzer, K W, 20 New Guinea Highlands, problems of in
high altitudes, 153-159
Cunliffe, B, 133
C14, see under dating techniques and radio- cusp catastrophe theory, 218, 219, 220-229
carbon dating Cyprus, Tenta catchment area, resource
calcium, 20, 25, 26, 27, 28, 29, 31, 32, 34 exploitation in, 22, 64-71
Canada:
Caniapiscau river basin, 25-35
La Grande river basin, 25 Darwin, C, 292
Quebec, soil analysis of sites, 20, 25-37 Dating techniques:
canals, irrigation, Habur valley, Syria, 108, dendochronology, 11, 12
113, 115, 117-126 geological and geomorphological, 234, 235
Caniapiscau river basin, Canada, 25-35 palaeomagnetism, 235
Cantabria, Italy, 132 radiocarbon, 12, 34, 235
carrying capacity, 153, 291, 297, 298-302 Davidson, D A on conceptual issues at site
Castell Henlys, Wales, 72, 73, 74, 75, 76 and microregion scale, 19-24
catastrophe theory, 190, 212, 216-229 defensive siting of enclosures in Iron Age
catastrophe theory, cusp 218, 219, 220-229 Wales, 72, 73, 77
 INDEX 313
Deith, M R: 21 eels, 283, 284, 285
on shell material from Franchiti Cave, Ein Gev, Israel, 238
Greece, 49-58 elm decline, see also ulmus under pollen and
demographic growth re Population Theory, spores, 247
283-298, 301 England, Ilkley Moor, Yorkshire, 129, 130
demography, see also poulation, 61, 93 environment:
dendochronology, 11, 12 E models in, 300
Denmark: interactions with social subsystems 278-282
Aamosen, 258, 265 environmental archaeology, and disease, 273-
Bunds So 258, 266 277
Dyrholmen, 258, 265, 267 archaeology, conceptual issues,introduc-
introduction of agriculture in, 252,253,254, tion to, 19-23
255, 258-269 change, Palaeolithic period, research
Korup So, 258, 264, 268, 269 methods, 233-245
pollen from, statistical analysis of, 258-269 issues associated with population change,
Skallesogaard, 258, 266 189-192
Stevingen, 258, 267 reconstruction, Habur River, Syria, 108-
Svanemose, 258, 266, 267 128
Tinglev So, 258, 264, 267, 268 thresholds, re impact of man on valley
Denton, D: 9, 20 alluviation, 211-232
on soils in archaeological sites in Northern Ergenzinger, PJ: 92
Quebec, 25-37 on environment, irrigation and settlement,
diet, 19 Habur.N E Syria, 108-128
disease: erosion, see also under alluviation,
and environmental archaeology, 273-277 anthropogenic disturbance in, 211, 214, 216
animal, 274, 275 Bronze Age, British, 211, 212
human, 273, 274 erosion, soil: 189, 193-210
plant, 275, 276 erosion-vegetation interaction in, 205-206
role of in environmental archaeology, 191, grazing in, 206-208
273-276 gullying, 198-201
documentary evidence: Mediterranean environments, 202, 206,
Assyrian re irrigation canals, 126 208, 209
Classical sources, 135, 138, 140 vegetation cover in, 201-205
Roman, inscription, Materia, Yugoslavia, wash, 194-198
104, 105, 106 erosion-vegetation interaction re soil erosion,
settlements and socio-economic variability 205-206
8, 9 Ethiopia, Melka Konture, 233
Dumond, D E, 293 Euphrates, 238, 241
Dur-katlimmu, Syria, see also Tell Hamad,
113, 114
Durkheimian school of Historical Geography farm sites, classical, survey of, Boetia.Greece,
163, 165, 173, 174 135-142
Dyrholmen, Denmark, 258, 265, 267 faunal remains,see animal bones
flint artefacts, see under artefacts
earthworm activity, 40, 46 flint distribution, Bandkeramik culture, 149
East Hook Rath, Wales, 77, 78 food resource exploitation study, bias
ecofacts, problems with re Palaeolithic, 236 towards, 273
ecological approach to archaeology, 20 foraging theory:
ecological units, concept of as spatial units, mathematical modelling of re socioecologi-
102, 103 cal change, 283-290
ecology: foraging-farming transition, Franchiti Cave,
agricultural, 248-255 S Argolid, Greece, 49, 52, 53-55
human re palynology and prehistoric forest clearance, New Guinea Highlands, 154
agriculture, 190, 246-272 forest cover:
theory in, 246 Mediterranean, 177
economic space, definition of re Rodik, reconstruction, Bandkeramik culture 146-
Yugoslavia, 95-107 149
ecosystems: forests: 249, 250, 251, 252, 253, 254
agricultural, 248-255 arboreal species, pollen from, 252, 253, 254,
interaction between environment and social 256, 262, 263-269
subsystems, 278-282 species replacement in, 251, 252, 253, 254
314 INDEX
forests: (cont.) Green, S W: 9, 190
temperate, succession models in, 251, 252, on palynology and human ecology, re pre-
253 historic agriculture, 246-272
France: Groenman-van Waateringe, W: 191
Annalist tradition of Historical Geography, on interaction between environment and
168-176, 177 social sub systems, 278-282
Historical-Geographical concepts relevant gullying, re soil erosion, 198-201
to Archaeology, 161-186
Paris, as centre of 19th C French academic Habur River, Syria, reconstruction of
life, 162 environment, irrigation and settlement,
Vidalian tradition of Historical Geography, 108-128
161, 163-168, 172, 176-180 Hassan, F, 300, 301
Franchiti Cave, Greece, shell material from, health, human, study of, 273, 274
21, 49-56 hearths, chemical analysis of soils from, 25-35
Frey, W, on environment, irrigation and Higgs, E S, 99, 132
settlement, Habur,N E Syria, 108-128 Hillforts:
Friedberg-Bruchenbrucken, Germany, Band- Kras region Yugoslavia, 95-106
keramik site, 149 S English, as community focii, 133, 134
Friedel, D, 10, 11 typological variations in, 97
Historical Geography.
Gaffney, C F: 4, 5, 23 Annalist tradition of, 168-176, 177
on Quantitative approaches to site territory Vidalian tradition of, 161, 163-168, 172,
and land use from surface record, 82-90 176-180
Gaffney, V L; 4, 5, 23 Historical-Geographical concepts :
on Quantitative approaches to site territory a review of re Corsica and Sardinia, 161-
and land use from surface record, 82-90 186
geographical-historical concepts, a review of culture, concept of, 166-168
re Corsica and Sardinia, 161-186 duree(s), longue, concept of, 169, 176, 177
geography: 93, 96 Historical-Geographical concepts, develop-
historical, Vidalian tradition of, 161, 163- ment of.France, academics relevant to:
168, 172, 176-180 Aries, P, 162
links with Archaeology, 21, 59-63 Berdoulay, V, 161, 162, 163, 164, 165, 166
geological formations re Palaeolthic, 233, 234 Berr, H, 166, 172, 174
geomorphological processes in Palaeolithic Blanchard, R, 177
research, 233,234,235,238,239,240,241 Bloch, M, 161, 169, 171, 174, 175
geomorphology, research methods in re Boutroux, E, 166
Palaeolithic, 233-245 Braudel, F, 161,168,169, 171,172,175,176
Germany: Brunhes, J, 163
Broekfeld, Bandkeramik site, pollen analy- Buttimer,A, 161, 163, 165, 167, 168
sis, 146, 147 Childe, V G, 166
Friedberg-Bruchenbrucken, Bandkeramik Clark, T N, 162, 168
site, 149 Clarke, D L, 162, 180
Linearbandkeramik landscapes, 145-150 Claval, P, 179
Merzbach, Aldenhovener Platte, Band- Colin, A, 173
keramik sites, 145 Cournot, A, 166
Glassow, M A, 296, 297, 298, 302 Coutau-Begarie, H, 168, 169, 170, 171,172,
grain, see under plant remains 173
granaries, Iron Age, 74, 75 Demangeon, A, 173
Grant, E, on conceptual issues associated with Draperon, L, 164
environmental and population Dubois, M, 164
change, 189-192 Duhem, P, 166
grazing, re soil erosion, 206-208 Dumziel, G, 162
Greece: Durkheim, E, 163, 165, 172
Boetia, regional survey, 135-142 Duruy, V, 164
Franchiti Cave, shell material from, 21, 49- Febvre, L, 161, 165, 168, 169, 171, 174
56 Ferry, J, 164
Melos, agricultural land use survey, 67 Flannery, K, 166
Greek Classical period, rural settlement in, Hauser, H, 172
135-142 Himly, A, 164
 INDEX 315
Historical-Geographical concepts, develop- isochronic distance.see under Spatial analyti-
ment of.France, academics relevant to: cal techniques
(cont.) Israel:
Huizinger, J, 172 Ein Gev, 238
Jaurs, J, 172 Ubeidiyah, 235, 238
Labrousse, F, 172 Italy:
Langlois, C, 169, 171 Cantabria, 132
Le Play, F, 165 Venosa, 235
Levasseur, E, 164
Levy-Bruhl, L, 172 Jochim, M A, 19
Luckerman, F, 165, 166 Jura, Isle of, Cul a'Bhaile, pollen from, 21
Meillet, A, 172
Michelet, J, 165, 172 K see under potassium
Mistral, F, 165 Kalis, A J: 92
Pirenne, H, 167, 172 on Linearbandkeramic Times, 145-152
Poincare, H, 166 Karouska-Soper, V, 98
Ratzel, F, 165 Kenya, Lake Turkana, 241
Renan, E, 165 Knock Rath, Wales, 72
Roupnel, G, 172 Korup So, Denmark, 258, 264, 268, 269
Sauserre, F de, 172 Khne, H, on environment, irrigation and
Seignbois, C, 169, 171 settlement, Habur.N E Syria, 108-128
Simiand, F, 172, 174 Kuk, New Guinea Highlands, 154
Simon, J, 164 Krschner, H, on environment, irrigation and
Sion, J, 173 settlement, Habur.N E Syria, 108-128
Taine, H, 165
Tarde, G, 166 La Grande river basin, Canada, 25
Vallaux, C, 163 labour/land models in population theory,
Vidal de Blanche, P, see also Vidalian tradi- 302, 303
tion, 161, 163-168,171 lake deposits, 235, 238, 241
Wallon, H, 172 Lake Turkana, Kenya, 241
Weber, M, 172 land use:
Historical-Geographical concepts : Bandkeramik culture, 145, 146, 148-150
genres de vie, 166-168, 177, 178, 180 Boetia, intensity of, 135, 138, 140
Holgan Camp, Wales, 73 evaluation, 92, 93
Holland, Rijkholt, Hint from, 149 high altitude zone, New Guinea, 155-159
Holling, C S, 12, 13 patterns, 25, 26
Hours, F: 190 reconstruction of re artefact distribution
on cultural and environmental change: analysis, 82-89
palaeolithic, geomorphological research re, landscapes, integrative model in reconstruc-
233-245 tion of, Bandkeramic times, 145-150
hunter-gatherer strategies, 191, 192 land/labour models in population theory,
hunter-gatherers, nutritional requirements, 302, 303
281, 283-285, 287, 289, 290 Latamne, Syria, 240
hydroxides, Fe-Al, 26, 32, 34 'least effort' principles, 140, 141
Leiden University, linking of Archaeology
Ilkley Moor, Yorkshire, England, 129, 130 with Geography, 59
inscription, Materia, Yugoslavia, 104, 105, Lethwaite, J: 93
106 on Historical-Geographical concepts, a
Integrative Model, Bandkeramik landscapes, review, 161-186
145-152 lichens, 27
Iron Age: Linearbandkeramik, see also under Band-
changes in subsistence economy in West keramik
Wales, 70-81 Linearbandkeramik times, integrative model
Habur Valley, Syria, 111, 113,115,122,123, in landscape reconstruction, 145-150
126 lithic assemblages:
Hillforts.S English, 133, 134 Canadian Indian, 26, 29
Rodik Hillfort, Yugoslavia, 95-106 Palaeolithic, 236-240
"rigation canals, Habur valley, Syria, 108, Ljubljana, University of, 95
113, 115, 117-126 Llawhaden, Wales, 72, 74
316 INDEX
Mg see under magnesium economy in Iron Age and R-B Wales,
Maddle Farm, Berkshire, England, Romano- 72-81
British villa, 83, 84-89
magnesium, 20, 25, 26, 27, 28, 29, 31
Malthus, T, 192, 291, 292, 293, 299 Nahr el Kebir, Syria, 238, 240, 241
manuring, 21, 23, 45, 84, 85, 89, 138 nematodes, plant diseases caused by, 275, 276
mapping, importance of in archaeology, 61, Neolithic:
62,63 agriculture and population, 303
marine deposits, 235, 238, 241 axes 131, 132
marine mollusca and palaeoenvironmental Bandkeramik culture, see also under Band-
reconstruction, 49-56 keramik, 145-150
marine mollusca: Denmark,introduction of agriculture, 252,
Cerastodermum, 50, 51 253, 254, 255, 258-269
Cerithium, 50, 51, 52 resource exploitation in Tenta, Cyprus, 64-
Columbella, 50, 51 71
Cyclope, 50, 51 transition to in S Argolid, Greece, 49-56
Donacilla, 50, 51 plant disease in, 275, 276
Donax, 50, 51 network reconstruction, re settlements, 132,
Gibbula, 50, 51 133, 134, 135
Monodonta, 50, 51 New Guinea:
Murex, 50, 51 Kuk, 154
Patella, 50, 51 Tambul, 154, 155
Tapes, 50, 51 New Guinea Highlands, agriculture in, 153-
Materia, Yugoslavia, inscription, 104, 105, 160
106 New Guinea Highlands:
Mathematical modelling, see also under altitudinal limits to agriculture, 153-159
Spatial analytical techniques forest clearance, 154
Mathematical modelling: land use, high altitude zone, 155-159
animal behaviour, 292 resource use, prehistoric, 154-159
catastrophe theory re landscape, 212, 216- settlement, agricultural, in 153-159
229 swamp cultivation, 154, 156
erosion equilibria under grazing, 193-209 swiddens, cultivation in, 154, 155, 156,158,
Linear Programming, 283, 285, 290 159
Musgreave equation re soil erosion, 194 nutrition re hunter-gatherers, 283-285, 287,
re foraging theory and socioecological 289, 290
change, 283-290 nutrition, needs in foraging-farming com-
Simplex Method, 285, 286 munities, 53, 54
Maya: 10, 11
Classic, collapse of re maize mosaic virus,
275 O'Shea, J, 297, 299
Mediterranean environments, re soil erosion, Olduvai, Tanzania, 233, 235, 237, 239
202, 206, 208, 209 Orontes Valley, Syria 240, 241
Melka Konture, 238
Melos, Greece, agricultural land use survey, P see under phosphorus
67 palaeoenvironmental reconstruction:
Merryborough Camp, Wales, 72 marine mollusca in, 49-56
Merzbach, Aldenhovener Platte, Germany, palynological techniques in, 38-47
Bandkeramik sites, 145 Palaeolithic:
Mesolithic period, transition from re S Achaulean, differentiation of phases in,
Argolid, Greece, 49-52, 54, 55 241, 242
mobility in pre-industrial environments, 61, artefact assemblages, 236,237,238,239,240
62 Chellean, 233
Modderman, P J: 12, 21 cultural and environmental change during,
on Archaeology, Geography and Scales, 233-245
59-63 ecofacts, problems with, 236
Moore, T R: 9, 20 geomorphological processes in, 233, 234,
on soils from archaeological sites, Northern 235, 238, 239, 240, 241
Quebec, 25-37 living floors, 233, 234
Mytum, H C: 22, 23 sequences in, 237
on evidence for changes in subsistence undisturbed assemblages, 239, 240
 INDEX 317
palaeopathology, 273, 274 palaeopodsols, Cul a'Bhaile, Jura, 39-47
palynological techniques in palaeoenviron- pollen and spores.see also under forests and
mental reconstruction, 38-47 plant remains
palynology: see also pollen pollen and spores:
history and aims of, 247, 248 Alnus, 253, 254, 262, 263-269
re human ecology and prehistoric agricul- Artemesia vulgaris, 254, 255, 265
ture, 190, 246-272 Asropyron repens, 254, 255
Paris, France as centre of 19th C French Betula, 253, 254, 262, 263-269
academic life, 162 Calluna, 43, 44
pathogens, influence of on crop productivity, Cerealia, 43, 44, 45, 46, 146, 147
191, 275, 276 Corylus, 146, 147, 148, 253, 254, 262, 263-
peat exploitation, 39 269
peat, see also under soils, 39, 46 Fraxinus 253, 254, 262, 263-269
Pelcombe Rath, Wales, 77, 78 Graminae, 42, 43, 44,45,146,147, 264, 265,
Pembrey Mountain, Wales, 74, 75 266, 267
pests, influence of on crop productivity, 191, Lycopodium, 42, 43, 44, 45, 46
275, 276 Pinus, 253, 254, 262, 263-269
pH values 20, 23, 26, 27, 28, 31, 34 Plantago lanceolata, 254, 264, 265, 266
phosphorus fractionation analysis, 20, 21, 25, Polypodium, 43, 44, 45
32-35 Pteridum, 42, 43, 44, 45
phosphorus, 25-35 Quercus, 146, 147, 148, 253, 254, 262, 263-
pigs: see also under animal bones, 269
domestic in New Guinea Highlands, 157, Rumex acetosa, 254, 264, 265, 266
158 Rumex acetosella, 254, 255, 264, 266
plaggen soils, 25 Tila, 146, 147, 148, 253, 254, 262, 263-269
plant demography, 252 Ulmus, 148,149,150,253,254,262,263-269
plant diseases: Vaccinium myrtillus 46
ergot 275 pollen, fossil, statistical analysis of, 255-261
mosaic virus, 275 population, animal, growth rates and sex
parasites, 275, 276 ratios in, 280, 281, 292
rust, 275 population approach, standard, re prehistoric
colocasia esculenta, 154,155,156,157,158, resource use, 153, 154, 155, 159
159 population densities, Classical Greece, 140,
plant remains, see also under pollen and 141
spores population, human:
plant remains: control of, 279-281
Ipomoea batatas, 154, 155, 157, 158, 159 growth rates in, 278-281, 292, 293, 303, 304
Almond, 53 interactions between environment and
Artemisia herba-alba, 114, 116 social subsystems, 278-282
barley, 68 sex ratios in, 280, 281
Chenopodiaceae, 114 population pressure, 291, 293, 295-302, 303,
emmer, 80 304
graminaea 74, 75, 76, 79, 80 population size, prehistoric, re carrying
lentil 68 capacity, 153
Pear, 53 Population Theory: 291-305
Pistacia, 53, 114 biological and cultural background to, 291-
Poa sinaica, 114 293
Populus euphratica, 114 demographic growth in, 293-298, 301
Prosopis farcta, 114 social alternative in, 302-304
Salix sassaf, 114 theory of population pressure and carrying
Tamarix jordanis, 114 capacity in, 298-302
wheat, 68 populations, change within, 191, 192
Pl
gg, F, 11, 12 populations, conceptual issues, 189-192
Pollen, 19, 20, 21 potassium, 27, 28, 31
Pollen analysis: pottery:
Banderamik Culture, 146, 147, 149, 150 Bandkeramik,( Linearbandkeramik), 145,
Danish Neolithic, 252-270 150
Habur valley area, Syria 114, 116 Classical Greek, 138
statistical analysis of, 255-269 Iron age, 79, 80
theory and method 276-272 Roman, 79, 80, 83, 84, 85-89
318 INDEX
pottery: (con/.) reconstruction of development of, Habur
statistical approach to sherd distribution, Valley, Syria, 110, 111-126
84, 85-89 rural, Boetia, Greece, 135-142
pox, evolution of 274 settlement sites:
distribution, 130-142
quantitative approaches to site territory and Bandkeramik, 145, 148, 149, 150
land use, 82-89 Corsica, classification of, re cultural con-
quantitive measurement, 2 cepts, 179, 180
Quebec, Canada, soil analysis of sites, 25-37 disappearance of, Boetia, 135, 138
querns, 75, 76 network reconstruction, 132, 133, 134, 135
reconstruction, surrogates in, 130, 131, 132
radiocarbon dating, 12, 34, 145, 154 Sardinia, classification of, re cultural con-
rainfall, as agricultural zone indicator, Syria, cepts, 179, 180
110, 111 Shackleton, J C: 21
raised beaches, Mediterranean, 235, 238 on shell material from Franchiti Cave,
Rappaport, R, 299, 300 Greece, 49-58
Ras Beyrouth, 238 shells: 20, 21
redeposition, problems of re Palaeolithic, 233, marine,see under marine mollusca for
234 details, 49-56
refuse/rubbish disposal, see also manuring, shifts, environmental, problems of interpreta-
76, 80 tion, 9
region, as unit of observation, 96 shoreline reconstruction, 52
reindeer, 283, 284, 285 site patterning; environmental, cultural and
Renfrew, C, 216 preservation factors 129-142
resource use, prehistoric, New Guinea High- site territorial determination, problems in, 82,
lands, 154-159 83, 85, 86-89
Rijkholt, Holland, flint from, 149 sites:concept of, see also under settlement, 4,
river terraces, 233, 234, 238 5,22
rock art, Bronze Age, Ilkey Moor, Yorkshire, environmental research in, 19
129 siting, defensive re Iron Age, Wales 72, 73, 77
Rodik, Yugoslavia, Iron Age hillfort, 95-106 Skallesogaard, Denmark, 258, 266
Roman Period, Yugoslavia, 99 Slade Farm, Wales, 77, 78
Romano-British period: Slapsack, B: 93
changes in subsistence eeconomy in West on defining the economic space of an Iron
Wales, 70-81 Age Hillfort:
Maddle Farm Project, artefact distribution Rodik, Yugoslavia, 95-107
analysis, 82-89 social alternative in Population Theory, 302-
Rudictes, 105, 106 304
rural settlement, Boetia, Greece, 135-142 social subsystems, interaction with environ-
ment, 278-282
sampling strategies, 82, 83 Socioecological change, mathematical
Sardinia, Historical-Geographic concepts rel- modelling,in re foraging theory, 283-290
evant to Archaeology of, 176-181 soil maps, 62, 63
Scaife, R G: 9, 189, 190 soil properties, spatial variation in, 26-32
on environmental thresholds, catastrophe soil type, effects on small farm distribution,
theory and landscape, re valley alluvi- Boetia, 135, 138, 140
ation, 211-232 soils:
Scotland: alluviation processes, 211-232
Black Moss of Achnacree, 40 anthropogenic modification, see also culti-
Scotland, Cul a'Bhaile, Jura, environmental vation etc, 25, 27, 29-35
reconstruction at, 21, 38-47 brown earths 40, 76, 78
sedimentation, rates of re river valleys, 211, buried, 21
212-216 capability, modern, 76, 78
seeds, 19 chemical analysis, see also under specific
settlement: techniques, 20, 21, 25-37
agricultural, in New Guinea Highlands, colluvial deposits, Holocene, 224
153-159 cultivation, see also agriculture, 41, 46
definition of units of, 97, 98, 99, 100-6 earthworm activity, 40, 46
patterns, dynamics of in West Wales.Iron erosion, see also under erosion, 189, 193-
age to R-B period, 70-81 210
 INDEX 319
soils: (con/.) ment, irrigation and settlement, 108-128
erosion processes, 193-210 Latamne, 233, 240
floodplain deposits, Holocene 212-214 Nahr el Kebir, 238, 240, 241
226 Orontes Valley, 240, 241
gypsiferous 111 Syria, Tell Ailun, 108, 109
histosols, 39, 46 Tell Aqab, 108, 109
Loess 145, 146, 148, 149 Tell Barri, 108, 109
.palaesols, 39-47 Tell Brak, 108, 109
plaggen, 25 Tell Chagar Bazar, 108, 109
podzols and podzolisation, 26, 38, 39, 40 Tell Fecheriya, 108, 109
42, 46 Tell Halaf, 108, 109
pollen analysis in, 39-47 Tell Hamidi, 108, 109
subarctic, 25-35 Tell Leilan, 108, 109
Spain, Ambrona, 235 Tell Mozan, 108, 109
Spain, soil erosion in, 189, 209 Tell Seh Hamad, 108, 109, 111-126
Spain, Torralba, 235
Spatial Analytical techniques: Tambul, New Guinea, 154, 155
activity analysis, 92 tannur, oven, 116
artefact distribution analysis, 23, 82-90,138 Tanzania, Olduvai, 233, 235, 237, 239
catchment analysis, see also under site Taro, (colocasia esculenta), 154,155,156,157,
catchment analysis, 132-134 158, 159
cluster diagrams, 97, 98 Tell Agaga/Sadikanni, Syria, 108, 109, 117
isochronic distance, 66, 77, 78, 87, 99, 100, Tell Ailun, Syria, 108, 109
101, 103 Tell Aqab, Syria, 108, 109
land use evaluation, 92, 93 Tell Barri, Syria, 108, 109
site catchment analysis, 22, 82, 83, 87, 88, Tell Brak, Syria, 108, 109
89, 92, 99 Tell Chagar Bazar, Syria, 108, 109
site territorial analysis, 22, 23, 64-71 Tell Fecheriya, Syria, 108, 109
spatial analysis, see also under various tech- Tell Halaf, Syria, 108, 109
niques, 59-63, 96 Tell Hamidi, Syria, 108, 109
Theissen catchment boundaries, 139 Tell Leilan, Syria, 108, 109
Thiessen polygons, 23, 77, 79, 88, 89, 92, Tell Mozan, Syria, 108, 109
99, 100, 101, 132, 133, 134, 135 Tell Seh Hamad, Syria, 108, 109, 111-126
spatial units, concept of, 4, 5 Tenta, Cyprus:
spatial variation, in soil properties, 26-32 resource exploitation in catchment area, 22,
standard population approach, re prehistoric 64-71
resource use, 153, 154, 155, 159 problems in catchment area definition, 66-
statistical analysis, pollen, 255-261 68
Stevingen, Denmark, 258, 267 terracing, 69
subarctic soils, chemical analysis in, 25-35 terrain, as problem in Site territorial analysis,
subsistence economy, changes re West 69,70
Wales, Iron age to R-B period, 70-81 territorial definition, re Rodik, Yugoslavia 96,
survey strategies, 82, 83, 84 99, 102, 104, 105
survey, geomorphological and archaeologi- Theissen polygons, 23, 77, 79, 88, 89, 92, 99,
cal, 190 100, 101, 132, 133, 134, 135
Svanemose, Denmark, 258, 266, 267 Thorn, R, 216, 217
swamp cultivation, New Guinea Highlands, Thornes, J B: 189
154, 156 on erosional equilibria under grazing, 193-
Sweet Potato (Ipomoea batatas), 154, 155, 210
157, 158, 159 thresholds, environmental, re impact of man
swiddens, cultivation in : on valley alluviation, 211-232
New Guinea Highlands, 154, 155, 156, 158, Tinglev So, Denmark, 258, 264, 267, 268
159 Torralba, Spain, 235
Tropics, 303, 250 Trade route models, 129
syphilis: trade routes, Bandkeramik times, 149
endemic, 274 transhumance, 148, 177, 179
venereal, 274 tree cover reconstruction, see also under for-
Syria: ests, 146-149
Dur-katlimmu, 113, 114, 122, 126 tribal societies, buffering strategies in, 301,
Habur River, reconstruction of environ- 302
320 INDEX
Tubingen University, 108, 110 Knock Rath, 72
Llawhaden, 72, 74
Ubeidiyah, Israel, 235, 238 Merryborough Camp, 72
Upham, S, 12 Pelcombe Rath, 77, 78
Pembrey Mountain, 74, 75
valley alluviation: 189, 190 Slade Farm, 77, 78
accumulation, British Holocene, 212-216 subsistence economy, changes in Iron Age
anthropogenic impact on, 211-232 and Roman period, 70-81
cusp catastrophe theory in, 223-229 Walesland Rath, 73, 74, 75, 76, 77, 78
valley formation and deposition, re Palaeo- Woodbarn Rath, 72
lithic, 233, 234 Walesland Rath, Wales, 73, 74, 75, 76, 77, 78
variability, socio-economic, 6, 7,8 wash, re soil erosion, 194-198
Vasilikos river, Cyprus, 64, 65 Whittington, G: 21
vegetation cover, re soil erosion, 201-205 on environmental reconstruction at Cul
Venosa, Italy, 235 a'Bhaile, Jura, 38-48
Vidalian tradition of Historical Geography, Woodbarn Rath, Wales, 72
161, 163-168, 172, 176-180 Wright, C D : 191
Vita-Finzi, C, 99, 132 on foraging theory, mathematical
modelling in, 283-290
Wagstaff, M: 6, 22
on Tenta catchment, resource exploitation
in Neolithic Cyprus, 64-71 yaws, 274
Wales: Yorkshire, neolithic axes from, 131, 132
Castell Henlys, 72, 73, 74, 75, 76 Yugoslavia:
Cleddau Valley, 72, 74, 75, 76 Materia, inscription, 104, 105, 106
Coygan Camp, 73, 74, 75, 76, 79 Rodik, Iron Age hillfort, 95-106
Cuckoo Grove Rath, 77, 78
East Hook Rath, 77, 78 Zimmermann, A, on Linearbandkeramic
Holgan Camp, 73 Times, 145-152