Regional Approaches to Adaptation in Late Pleistocene Western Europe 9781841711553, 9781407352381


182 11 194MB

English Pages [238] Year 2000

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Front Cover
Copyright
Table of Contents
List of Contributors
Editors' Preface
1. Regional Approaches to Adaptation in Late Pleistocene Western Europe
2. The Magdalenian of Grotte XVI (Dordogne, France) and Regional Approaches to Magdalenian Settlement and Economy
3. The Relevance of Regional Analysis for Upper Paleolithic Archaeology: A Case Study from Portugal
4. Intra-Regional Similarities in Resource Exploitation Strategies: The Late Magdalenian in the Vezere Valley
5. Late Upper Paleolithic Environments, Subsistence, and Zoogeography in Cantabrian Spain
6. Variability and Context of Magdalenian Visual Imagery, Western Languedoc-Roussillon, France
7. Les groupes humains au Tardiglaciaire dans le Bassin parisien: diverses voies pour une approche regionale
8. Settling Down or Moving Around? The Development of Regional Traditions during the European Magdalenian
9. Readaptation: Changes in Magdalenian Subsistence and Social Organization
10. Specialized Final Magdalenian Hunting Technology in Southwest France
11. Regional-Scale Variation and the Magdalenian Record of Northwestern Europe
12. Patterns of Ethnogeographic Variability in Late Pleistocene Northwestern Europe
13. Final Palaeolithic in the Northwest: Migrations and Seasons
14. The Eastern Magdalenian: Hunters, Landscapes, and Caves
15. Coming Out from the Cold: Western Europe in Dryas I and Beyond
16. Intraregional Variability in the Mesolithic of Atlantic Coastal Iberia
17. Regions and Late Pleistocene Hunter—Gatherers
Recommend Papers

Regional Approaches to Adaptation in Late Pleistocene Western Europe
 9781841711553, 9781407352381

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Regional Approaches to Adaptation in Late Pleistocene Westem Europe

Edited by Gail Larsen Peterkin and Heather A. Price

BAR International Series 896 2000

Published in 2019 by BAR Publishing, Oxford BAR International Series 896 Regional Approaches to Adaptation in Late Pleistocene Western Europe © The editors and contributors severally and the Publisher 2000 The authors’ moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. ISBN 9781841711553 paperback ISBN 9781407352381 e-book DOI https://doi.org/10.30861/9781841711553 A catalogue record for this book is available from the British Library This book is available at www.barpublishing.com BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 1974 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by John and Erica Hedges in conjunction with British Archaeological Reports (Oxford) Ltd / Hadrian Books Ltd, the Series principal publisher, in 2000. This present volume is published by BAR Publishing, 2019.

BAR

PUBLISHING BAR titles are available from:

E MAIL P HONE F AX

BAR Publishing 122 Banbury Rd, Oxford, OX2 7BP, UK [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

Table of Contents List of Contributors . . .. . ........... Editors' Preface . .. .. ..... l

.....

. . .. . ... . .. ......... . . . .. . .. . .......

. .......

. ......

. .........

.. .. ....

....

.. . . .. .. . . . . . .. ... . ....

... .. . ... .. . . . . ... . . . ....

Regional Approaches to Adaptation in La.te Pleistocene Western Europe Heather A. Price and Gail Larsen Peterkin . . .. ..... .... ...... .. . . .. .. . . ...........

. .....

. .. v

. .. . ... vii

. .....

....

..1

2

The Magdalenian of Grotte VXJ (Dordogne, France) and Regional Approaches to Magdalenian Settlement and Economy Jean-Philippe Rigaud , Jan F. Simek, and Maureen A. Hays .. . .. .... . .. . . . .. . .... . .. . ..... . ........ . ..... 9

3

The Relevance of Regional Analysis for Upper Paleolithic Archaeolog y : A Case Study from Portugal Paul T. Thacker ..... . .. . ... . .. . .. . ... . . .. .. . ... .... .. . . . .. . .. . ........ .. ... . .. . .........

. ... 25

Intra-Regional Similarities in Resource Exploitation Strategies : The Late Magdalenian in the Vezere Valley Katherine V. Boyle ... . .. .. .. . ... .. . . .... . ... .. .... .. .... . ... ... ... . . .. ........ .. .. . ......

.... 47

4

5

6

late Upper Paleolithic Environments , Subsistenc e, and Zoogeograph y in Cantabrian Spain James T. Pokines ... . . .. . . .. . . .. .... . .. . ..... . ............ . . . ... .. . .. .. ....

... . .. .....

Variability and Context of Magdalenian Visual Imagery, Western Languedo c- Roussillon, France Heather A. Price ....... . . .. ....... . ......... ... . . .... . .. .... ... ... . . ... . ..... . ....

. . ... . .61

....

....

. .. 75

7

Les groupes humaines au Tardiglaciaire dans le Bassin parisien: diverses voies pour une approche regionale Boris Valentin . . ..... . . .. . . ......... .... .. .. .. . ..... . .. . ......... . .. .. ..... . . . . .. .. . .. ... ... .93

8

Settling Down or Moving Around? The Development of Regional Traditions during the European Magdalenian Todd A. Koetje ................ . .. ...... . .. ...... ... .... . . . ......... .... . . .. . . .. . . .. . . . . . ... 109

9

Readaptation: Changes in Magdalenian Subsistence and Social Organization James G. Enloe .. ..... . ........ . .... . .... . . .. .... . . . ......... .. .. . . . ........

10

11

12

13

14

Speciali zed Final Magdalenian Hunting Technology in Southwest France Gail Larsen Peterkin . . ........ . ....... ... ......... . ........ ... .....

Regional-Scale Variation and the Magdalenian Record of Northwestern Europe Eelco Rensink ... ... ......... . . . .... ... . . . ... .. .... ... . .... . . . .. .. .... Patterns of Ethnogeographic Variability in La.te Pleistocen e Northwestern Europe Berit Valentin Eriksen . ... .. ... .. . . ...... ... .. . .. . .. . ....... . . . ... . .... Final Palaeolithic in the Northwest: Migrations and Seasons Marcel Otte . . . . . . . . . . . . . ..... . . . ..... . . . .. . .. . .. . ......... The Eastern Magdalenian: Hunters, La.ndscapes, and Caves . ...... . .. .... .... . .. . ....... . .. .. ....

Jirf Svoboda

15

l6

17

. . . .....

Coming Out from the Cold: Western Europe in Dryas I and Beyond Lawrence Guy Straus .. . .. ... . . . . . ........... . .... . . . .........

Regions and La.te Pleistocene Hunter-Gatherers H. Martin Wobst ... . ..... . . . ...... .. . ....

. . . .......

. . . .. . .. . .. . .. .......

. .....

. . . ......

. . ... ... . .............

. .. . . . ......

. ... . . 115

121

. . . . . . . .. . . .. . .. .. .. . 133

... 147

. ....

.. . 169

. ....

. .. 179

. .. . . . ... 191

. ... . .. . ... .. .. . ....

.. . .. . . . .. ... . . . . .. . . .. . . . .. ......

Ill

.. ......

. .. .. ... .. .. . ... . . . .....

.. ................

Intraregional Variability in the Mesolithic of Atlantic Coastal Iberia Geoffrey A. Clark ...... . ...... .. .. . ... .. ..... ... ... . ....

. . .. ......

.. .. 205

. ... . .. . ... . . . .. .221

List of Contributors Katherine V. Boyle McDonald Institute for Archaeological Research University of Cambridge Cambridge CB2 3ER U.K. E-mail [email protected]

James T. Pokines U.S. Army Central Identification Laboratory 310 Worchester Avenue Hickam AFB HI 96818 U.S.A. E-mail j-pokines-3@alumni .uchicago.edu

Geoffrey A. Clark Department of Anthropology Arizona State University Tempe AZ 85287 U.S.A. E-mail [email protected]

Heather A. Price 1056 Overlook Road Berkeley CA 94708 U.S.A. E-mail [email protected] Eelco Rensink Godsweerdersingel 54 6041 GM Roemond The Netherlands E-mail [email protected]

James G. Enloe Department of Anthropology University of Iowa Iowa City IA 52242 U.S.A. E-mail [email protected]

Lawrence Guy Straus Department of Anthropology University of New Mexico Albuquerque NM 87131 U.S.A . [email protected]

Berit Valentin Eriksen Department of Prehistoric Archaeology University of Aarhus Moesgard DK-8270 H0jbjerg Denmark E-mail [email protected]

Jirf Svoboda Akademie Ved Cr archeologicky Ustav Brno Oddelenf paleolitu a paleoethnologie 691 29 Dolni Vestonice 25 Czech Republic E-mail [email protected]

Maureen A. Hays Department of Sociology and Anthropology College of Charleston 66 George Street Charleston SC 29424 U.S.A. E-mail [email protected]

Paul T. Thacker Department of Sociology and Criminal Justice Texas A & M University-Commerce P.O. Box 3011 Commerce TX 75429 U.S.A. E-mail [email protected]

Todd A. Koetje Department of Anthropology, MS-9083 Western Washington University 516 High Street Bellingham WA 98225 U.S.A. E-mail [email protected]

Boris Valentin UMR 7041 "Archeologies et Sciences de I' Antiquite" Centre de Recherche Prehistorique UFR03 Universite de Paris I (Pantheon-Sorbonne) 3, rue Michelet F-75006 Paris France E-mail Boris [email protected]

Marcel Otte Service de Prehistoire Universite de Liege 7, place du XX AoGt, Batiment Al Liege B-4000 Belgium E-mail [email protected]

H. Martin Wobst Department of Anthropology University of Massachusetts, Amherst P.O. Box 34805 Amherst MA O1003 U.S.A. E-mail [email protected]

Gail Larsen Peterkin 370 I Clifford Drive Metairie LA 70002 U.S.A. E-mail [email protected] V

vi

Editors' Preface The present volume, Regional Approaches to Adaptation in Late Pleistocene Western Europe, was inspired by a pair of symposia held at two successive annual meetings of the Society for American Archaeology. The first, "Situational Adaptation in the Late Pleistocene of Western Europe," was held at the 60 th Annual Meeting in Minneapolis , Minnesota , May 3-7, 1995. The success of and the enthusiastic response to the first symposium resulted in another symposium, "Regional -Scale Variation in Late Pleistocene Western Europe," at the 61 st Annual Meeting in New Orleans, Louisiana, April 10- 14, 1996. From the inception of the first symposium, we intended that this collection of insightful and informative symposium presentations would find their home , and a wider audience, in an edited volume such as this one . As is inevitable , the ensuing volume has been trans formed from a mere reproduction of the original symposium papers. In an effort to give the volume a wider geographic and theoretical scope, we thoughtfully solicited contributions from other scholars who were unable to participate in either symposium but whose regionally based research complemented and supplemented the work originally reported by the symposiasts. We are delighted to include in this volume new papers by Katherine V. Boyle, Boris Valentin, and Jirf Svoboda We would also like to thank those who participated in the original symposia , but who , for a variety of reasons , chose not to include their papers in this edited volume: Ariane Burke; Margaret W. Conkey; Marcy Krupa; Federico Bernaldo de Quiros and Anne Pike-Tay; Blythe E. Roveland; and Silvia Tomaskova. We are especially grateful to the participants in the two original symposia who stuck with the project during all the ensuing ups and downs, and whose papers are included in this volume: Geoffrey A. Clark; James G. Enloe; Berit Valentin Eriksen; Todd A. Koetje; Marcel Otte; James T. Pokines; Eelco Rensink; Jean -Philippe Rigaud, Jan Simek , and Maureen A. Hays; Lawrence Guy Straus ; and Paul T. Thacker. Thank you for your patience and forbearance , and

for your timely submissions . We assume full responsibility for the inevitable publication delays! The role of thoughtful discussant was assumed first by John Parkington of the University of Cape Town, then by H. Martin Wobst of the University of Massachusetts, Amherst. Their cogent comments and discussion strengthened the original symposium papers and provided much food for thought, and Martin Wobst graciously agreed to reprise his role as discussant by writing the concluding chapter for this volume. We thank them both for their frankness and enthusiasm. We would both like to thank our former dissertation advisors, Harvey M. Bricker of Tulane University and Meg Conkey of the University of California at Berkeley, for their moral support and encouragement throughout the publication process. You can rest assured that they both made constructive and substantive comments and criticisms that were incorporated into the final publication. The appearance of the volume benefitted greatly from the professional graphic design services of Sharon Freeman of Freeman Graphics, Inc. Sharon tweaked edited text and graphics and ably melded them into a polished and professional format. She, too, is to be commended for her patience and diligence. Thanks are also due to former archaeologist Vickie Carpenter, who assisted with typing and other essential editorial tasks, and, of course , to John and Erica Hedges for their faith in and commitment to the publication of this and other volumes of interest to Palaeolithic archaeologists around the world. Finally, we both wish to acknowledge our families (especially Lewis, Gavin, and Haley) and our various employers for their understanding and equanimity during the planning and organization of the original symposia and during the long production process that accompanied the ensuing edited volume. What a long, strange trip it's been! Gail Larsen Peterkin Heather A. Price

vii

1 Regional Approaches to Adaptation in Late Pleistocene Western Europe Heather A. Price (University of California, Berkeley) Gail Larsen Peterkin (Tulane University) INTRODUCTION

THE REGIONAL FRAMEWORK

The Late Pleistocene of Western Europe stretched from the last glacial maximum, at roughly 18,000 years ago, to the onset of the Holocene, at roughly 11,000 years ago. Accounts of this time period in Western Europe have moved from describing a rather homogeneous phase in human prehistory, in which hunter-gatherers maintained a sort of stasis until climate change stimulated the development of agriculture and all that came with it, to more diverse accounts portraying a fascinating and variable period during which hunters and gatherers diversified, with some developing a significant degree of cultural complexity.

The term region has multiple meanings in Late Pleistocene studies. When North American archaeologists use the term region in reference to hunter-gatherer studies, they evoke an anthropologically based framework of subsistence-settlement systems, a landscape upon which societies interact. Many Europeans, on the other hand, have used region in a culturehistorical sense, delineating a spatial zone within which the material culture demonstrates a certain homogeneity. This distinction can be traced to the differential development of archaeology as a scholarly discipline in North America and on the continent. In the Americas , there was, at least in some cases, a direct historical link between ethnographically known cultures and the prehistoric archaeological record. This fact, along with the perception of a relatively shallow chronological depth for the peopling of the Americas, has resulted in a strong intellectual link between anthropology and prehistoric archaeology . Further developments in hunter-gatherer studies by scholars teeped in this tradition, and who specifically modeled differential organization of activities across the landscape, formed the basis for using the idea of the regional framework in the evaluation of archaeological data.

Over the past 15 years, the issue of variability in the Late Pleistocene has achieved a high profile in the archaeological literature, with several edited volumes focusing specifically on the archaeological record of various geographic domains during this general time period. A close reading of the individual contributions within these volumes reveals a tremendous range in terms of spatial and temporal scales of analysis. In addition, the socio-politics underlying regional construction within one or more specific contexts has also been addressed. Other areas that would benefit from formal examination include the scope and nature of the research question, the methodologies and variables addressed, and the scientific paradigm favored by the researcher. The current volume brings together examples of research carried out at a variety of spatial and temporal scales of analysis, although all are focused explicitly on Late Pleistocene hunter-gatherers in Europe. More than one author has suggested that the flexibility and resulting variability of Ice Age hunter-gatherers has been underestimated (Otte 1989; Wobst 1990). In order to identify and, ultimately, to explain this variability, it is necessary to define the different scales at which we work and to move between these scales of analysis. Even within the seemingly delimited arena of Late Pleistocene Western Europe, the task remains a complicated one. In the past, prehistoric hunter-gatherers experienced varying degrees of mobility and spheres of interaction, while, in the present, researchers define and conceptualize their regional subject in different ways. How, then, do we compare results derived from different geographical domains or from different theoretical research traditions? How do we choose and link appropriate theoretical models to archaeological data (e.g., Binford 1965; Thomas 1985)? And, finally, how do we move between different scales of analysis and interpretation, from the micro- to the macro-scale (cf., DeWalt and Pelto 1985)?

Such a direct historical link was not present in Western Europe. The substantially greater time depth for the occupation of Europe, along with a long tradition of explicit typological and chrono-stratigraphic studies, meant that Pleistocene archaeology became more closely linked to disciplines such as history and geology than to anthropology. Thus, the concept of region developed in order to distinguish between archaeological cultures as they were originally manifested and as they changed over time. A major distinction can be drawn between these two basic concepts of region. The American concept of region is strongly linked to questions of process and to the goal of understanding how different sites may have been formed as a result of interaction and differential use by diverse groups of hunter-gatherers, while the European concept of region tends to be more closely linked to real-world geographical or administrative units, as well as to questions of ethnic identity and to interactions between ethnic units through time and across space. Unfortunately, these concepts are sometimes portrayed as fundamentally incompatible (e.g., Clark 1994; Otte and Keeley 1990), when, in fact, they can be complementary. They simply operate at different scales of analysis. Despite their underlying settlement system framework, North Americans who work in Western Europe must acknowledge

PRICE AND PETERKIN

CURRENT RESEARCH SETTING

that most of their data has been collected and organized within the culture-historical framework, as well as within the socio-political framework of their chosen area of research. This discrepancy obviously increases the potential for theoretical disjunction (e.g., Roveland n.d.; Tomaskova n.d.). Within each of these major models of region, there is additional variability in the ways regions are defined. For example, Willey and Phillips (1958:18-21) have discussed the archaeological subdivision of geographic space. They move from the site as the minimal unit to, in order of increasing size, the locality, the region, the subarea, and, finally, to the culture area (discussed in Mueller 1974, 1975). This terminology has not been widely applied to the archaeology of Late Pleistocene Europe, where region is and has been used indiscriminately as a convenient way to refer to the particular constellation of sites within the cultural framework one wishes to investigate. Archaeologists from both traditions who work in Late Pleistocene Western Europe must become more precise in their use of the word region, formally defining and discussing the term and its specific application to the chrono-geographic landscape at the center of their investigation (cf., Rossignol and Wandsnider 1992). This formality will necessarily entail more explicit discussion of the scale of analysis.

Most articles of relevance to the regional study of Late Pleistocene hunter-gatherers have appeared in edited volumes similar to this one (cf., Dibble and Montet-White 1988; Gamble and Soffer 1990; Soffer 1987; Straus et al. 1996), as well as in the various publications of the British Archaeological Reports International Series (cf., Boyle 1990; Burke 1995; Koetje 1987; Otte 1988; Pike-Tay 1991; White 1985). Only a few scattered articles dealing with regions or with regional studies have appeared in the major archaeological journals during the past 15 years. The most recent discussion of region as a unit of analysis has been almost exclusively limited to articles reviewing the socio-politics of their construction (cf., Arnold 1990; Gonzalez Morales 1992; Kristiansen 1990; Milisauska s 1990; Otte and Keeley 1990; Van Reybrouk 1994).Complementing the basic research components of most publication s are a few theoretically oriented articles that have helped to identify some of the problems and limitation of, as well as the potentials for, the application of a regional approach to Ice Age variability (Brown 1987; Conkey 1987; Schortman 1989; Wobst 1990). Current research presents many different scales of analysis, including: 1) site-specific analyses formulated within a regional interpretive framework; 2) regional survey on the scale of hundreds of square kilometers; 3) the description of environmental and cultural variability on the scale of thousands of square kilometers; 4) the comparison between and among such regions; and 5) critical assessments of the construction and epistemology of regions and of regional research. It is interesting to note that scholars working with hunter-gatherers in a similar chronological framework in places such as Japan, the Northwest Coast of the United States, or California work at a considerably smaller scale of analysis and with much smaller regions; perhaps this is because they are working primarily with the archaeological record of sedentary hunter-gatherer . Western Europe in the Ice Age is, by contrast, the archaeology of more mobile peoples . Hence, regions are not always explicitly defined or discussed, and methodological and/or paradigmatic approaches differ to some extent (cf., Clark 1994) . Yet the overwhelming direction of this research is towards detailing and evaluating behavioral and cultural differences.

As mentioned above, the region has long been established as the appropriate analytical framework for examining questions of hunter - gatherer adaptations, at least among North American researchers (Binford 1964; Brown 1987; Judge et al. 1975; Read 1975; Struever 1968, 1971; Thomas 1985), yet an appropriate scale of analysis for region has never been, and probably cannot be, codified. Prehistoric hunter-gatherers, like their ethnographic counterparts, interacted within a variety of different social contexts, including the nuclear family, the extended family, specialized work parties, the minimum band, the maximum band, and probably in even more extensive networks involving long-distance movement of raw materials and artifacts (e.g., Binford 1980; Steward 1938; Wobst 1974 ). Thus, scales of interaction can overlap . Furthermore, hunter-g atherers in different environments, and who thus have different ecological adaptations, must thereby occupy regions of different spatial extent. Linking hypothesized social manifestations of hunter gatherer lifeways with spatial domains on the ground is, by nature, a very tenuous undertaking. The spatial extent of these organizational poses and their archaeological correlates varies in relationship to variables such as the local carrying capacity of the physical environment and the historically unique demographic and political setting of the region. As part of the analytical process, regions have been variously defined on the basis of a loose combination of similarity in material culture, natural geographical divisions and subdivisions of the landscape, theoretical perspectives and expectations for hunter-gatherer spatial organization, and, last but by no means least, modern administrative units . Rarely are each of these variables explicitly recognized or addressed by archaeologists.

SITE-SPECIFIC RESEARCH Although there has been considerable criticism of the traditional Western European practice of basing large-scale general interpretations on excavations from a single site, these criticisms are usually based on studies that fail to place single-site findings within a regional framework. All regionally framed work must ultimately draw upon both areal survey and the intensive excavation of individual sites. If models of hunter-gatherer behavior on the scale of the settlement system or the regional scale are to serve as a baseline, then sitebased interpretations must place the individual site within this larger behavioral dynamic, as several chapters in the present edited volume illustrate (e.g., Rigaud et al.). As long as the 2

REGIONAL APPROACHESTO ADAPTATION

researcher treats the analysis as an investigative technique rather than as a way to create models for regional patterns of the past (Conkey 1987; Geertz 1966), this site-specific scale of inquiry remains integral to the overall effort of understanding the many and varied adaptations of hunter- gatherers in the prehistoric past.

1968; Mithen 1990; Nitecki and Nitecki 1987; Peterkin, Bricker, and Mellars 1993; Price and Brown 1985; Schrire 1984; Winterhalder and Smith 1981). As a result , many scholars advocate a combination of extensive survey and the intensive excavation of selected sites as the most appropriate methodology for investigating the archaeological traces of prehistoric hunter-gatherer groups (e.g., Thomas 1985; Wobst 1990). Although some Palaeolithic archaeologists recognize the need for systematic survey (e.g., Straus and Clark 1986; Rigaud and Simek 1987), there are only a few examples of this kind of survey chronicled in the literature (e.g ., Conkey n.d.; Rensink 1995; Thacker 1996 and this volume).

One set of authors (Riguad et al.) argue that archaeologists must do a better job of more fully examining a single site before moving beyond it to consider the region. While the editors agree with this general premise, we are concerned that such thorough and exhaustive investigation of a single site could become an end in and of itself, thus discouraging hypothesis building at the larger, regional scale of analysis. Instead , we advocate moving back and forth between the different scales of analysis, acknowledging the inevitable mistakes along the way. It is unrealistic to insist that research must always begin at the most minuscule of scales , without progressing to further interpretive scales of analysis until everything has been done completely and correctly at the level of the individual site excavation. Merely using the phrase hunter-gatherer invokes obvious , implicit assumptions that have been made regarding mobility, interaction, group size and composition, and different behavioral poses for this way of life. These assumptions come into play at all analytical scales of analysis, from the single assemblage to larger analytical units, such as the site, a cluster of sites, or even larger spatial expanses. As evidence accumulates from many different individual excavations, the data will inevitably be placed within a larger regional context. Thus, it is perhaps most useful to work simultaneously on dual fronts: the presentation of important, fundamental data obtained from small-scale, individual excavations, along with preliminary attempts at placing this site-specific information within a more comprehensive regional framework.

Thus, the archaeological record of Western Europe is primarily known from the intensive excavation of isolated sites. The lack of systematic survey leaves us on shaky ground when it comes to making behavioral inferences on the basis of our known sample (Brown 1987; Conkey 1980, 1987; Rigaud and Simek 1987; White 1985; Wobst 1974). The record, as it stands, is simply not repre entative. Results from systematic surveys could serve as the basis for a more accurate representation of the past by using a more extensive landscape framework in the reconstruction of Late Pleistocene hunter- gatherer lifeways. Although theoretically sound, there are several potentially serious methodological limitations for these kinds of regional surveys. For example, the geological changes of the past 15,000 years have undoubt edly obscured many surface or non-cave manifestations of the Upper Palaeolithic; it is also certainly true that the intensive occupation and exploitation of much of Western Europe over the last two millennia has disturbed, if not destroyed, considerable evidence dating from the Upper Palaeolithic (cf., Rensink this volume). A few standard exceptions are recognized as rare and unusual instances of preservation, including, for example, the open-air reindeer butchery pavement of Lassac, preserved on a thalweg in southern France (Sacchi 1987). Nevertheless, the list of exceptions is becoming intriguingly, and encouragingly, long (e.g., Audouze 1987; Basinski 1979, 1982; Franken and Veil 1983; Gaussen 1980; Koetje 1987; Leroi-Gourhan and Brezillon 1972; Poplin 1976; Rust 1958; Veil 1982).

In addition, such a stringent approach is able only to utilize data acquired from the most modern of multidisciplinary excavations. While these data are vastly superior to those collected by earlier excavators at many of the classic sites of Western Europe, museum collections from these older excavations may still retain the potential to contribute information about prehistoric hunter-gatherers at a grosser scale of analysis-in other words , larger, more comprehensive geographic regions with less exact temporal resolution, the broad brush rather than the fine stroke. Nevertheless, in this era of emphasizing the conservation and stewardship of archaeological resources around the world, it is indispensable to be able to utilize this older data in some meaningful way and to incorporate it into contemporary heuristic models of huntergatherer behavior in Late Pleistocene Western Europe.

The judicious use of a preliminary geomorphological study gives the archaeologist some ability to predict where sediments dating from the Late Pleistocene are likely to remain, either on the surface or buried beneath it (e.g., Conkey 1995; Rigaud and Simek 1987; Simmonet 1985; Stafford 1995; Thacker 1996). These sediments can then be sampled using appropriate surface or subsurface techniques. Even when remains have been located and identified, it is still notoriously difficult to date surface features such as lithic scatters, especia11y in the absence of relevant diagnostic artifacts. Furthermore, many surveying decisions remain open debates. It is difficult to delimit both the spatial extent of a prehistoric hunter-gatherer group and the scale at which to conduct the survey. Given the presumed variability and diversity of hunter-gatherer responses to different local conditions, the problem of determining the scale at which to work will

REGIONAL SURVEY

Ethnographic research has revealed the seasonal and situational fluidity of the composition of hunter-gatherer groups and their movement across the landscape, and these observations have, in turn, been applied to the prehistoric archaeological record (Bailey 1983; Bettinger 1991; Lee and De Vore 3

PRICE AND PETERKIN

always be a controversial aspect of the research process. Survey , especially in combination with more intensive site excavation, is a potentially promising avenue for research , although it must be pursued with the understanding that there are inevitable limitations. For the moment , most of us are working with a biased sample of known sites, accompanied by theoretically derived expectations for how these sites might have been interrelated in the prehistoric past.

although many archaeologists assume that the environment played a causal role in the haping of Ice Age adaptations (e.g., Binford 1978; Foley 1984; Gamble 1984, 1986; Jochim 1983, 1987; Straus and Clark 1986), others question the extent to which or the scale at which the environment wa deterministic (e.g., Bender 1978; Conkey 1987; Schortman 1989; Wobst 1990). This issue cannot be properly addressed until detailed studies are carried out at multiple scales of analysis in order to see how multiple lines of material culture and spatial evidence vary in relationship to the complex set of condition that make up the environment. Global and continental syntheses of environmental trends throughout the Pleistocene (e.g., CLIMAP 1976; COHMAP 1987) have proven useful for grand-scale evolutionary syntheses, while small- cale paleoenvironmental studies establish the nature of micro -environmental variability within the broader, more generalized trends (e.g., Bahn 1984; Jalut 1976, 1977; Peyron et al. 1998; Raike s 1967). The baseline, smaller scale environmental information, when combined with archaeological evidence from different scales of analysis, will help establish an understanding of the role of the environment in hunter-gatherer adaptations in Late Pleistocene Western Europe.

LOCAL-SCALE DETAIL Research projects in Western Europe that have been designed to address questions at the local scale, often cited as the scale at which minimum band interaction occurred , generally encompass hundred to even thousands of square kilometers. Several factors complicate the linking of behavioral models to the archaeological record. As Palaeolithic archaeologi ts, we rarely operate at the nece ary scale of temporal resolution to be able to identify definitively social entities, such as the nuclear family and work party, in the same way modern social scientists identify such entities within extant human groups. Archaeological cultures differ from ethnographic cultures, and it is difficult, if not impossible, to reconstruct meaningful Palaeolithic social entities and personal identities from evidence provided by material culture and landscape use (for an alternate view, cf. Clark 1980; Hodder 1986).

INTER-REGIONAL COMPARISON Several of the chapters in the present volume compare archaeological patterning between local regions (e.g., Enloe, Eriksen, Koetje, Otte, Peterkin, Rensink). This scale of analysis is believed to access the broader scale of inter-group interactions between Ice Age hunter-gatherers and to characterize differences in adaptive solutions to variable local settings. It is only at this scale, or at an even broader scale, that variations in local adaptations and cultural expressions can be evaluated within chronological and spatial perspectives. For example, the traditional tool typology for the Magdalenian period was developed for as emblage from the Perigord region of southwest France. Yet, through time, this same typology has been used to identify and categorize materials from many other parts of Europe. These artifacts, and, by implication, their associated occupations, date from roughly the same period of time and suggest some degree of shared cultural tradition . Although the apparent commonality of such materials has been useful at the level of developing general cultural-historical models, it also raises important questions. For example, what variables are masked by this perceived homogeneity? And what variables might more appropriately reflect behavior at the local scale?

As a result of problems of temporal resolution, we are often doomed to the practice of time averaging (e.g., Stiner 1994). Most archaeologists would be horrified by the enormous spans of time commonly averaged by paleontologists concerned with the behavior of animals in the past, just as contemporary social scientists might, in turn, be horrified by the enormous spans of time routinely lumped by archaeologists . Nevertheless, when dealing with humans, we commonly assume a certain contemporaneity for the con truction of behavioral models when working with assemblages that may encompass, and may be separated from one another, by hundreds or even by thousands of years. The heuristic status of resultant models is often forgotten in our desire to tell the story of the past on a more intimate level. Although we appear to have a sophisticated understanding of the relationship between the availability of resources and the extent of group mobility, debates still rage regarding the extent of reindeer migrations and the extent to which hunter-gatherer groups followed such migrating animals (e.g., Bahn 1984; Boyle 1990; Burch 1972; Burke 1995; Delpech 1988; Gordon 1988; Pike-Tay 1991; Spiess 1979; Sturdy 1975; White 1985). These ongoing discussions demonstrate the problems inherent to the identification of behavioral constructs, such as the minimum band, and their spatial correlates within the archaeological record .

THE SOCIO-HISTORICAL CONSTRUCTION OF REGIONS Archaeological regions as they have been presented heretofore lie in the past, albeit with some more or less tenuous connections to the present in the form of geological landforms and climatological considerations. The very concept of region, however, is often firmly rooted in the present and in a particular sequence of historical development. No one would deny the importance of understanding the historical

The so-called local region is commonly chosen on the basis of environmental parameters, under the assumption that Late Pleistocene hunter-gatherers were strongly influenced by the nature of their physical world. The following dilemmas face researchers at this level of analysis and interpretation. First, 4

REGIONAL APPROACHESTO ADAPTATION

REFERENCES

development of regional concepts , as well as the complex motivations that underlie our reconstructions of the past. It is for this reason that we must strive to examine explicitly some of the historical, social , and political issues that form the basis of our regional constructions. We have already noted the articles in Antiquity (cf., Arnold 1990; Kristiansen 1990; Milisauskas 1990) on this subject, no doubt inspired by the recent political restructuring of substantial parts of Europe. These often unseen or unacknowledged aspects of the archaeological enterprise inevitably affect the regional approach in two important ways. In the first place , there are the practical structuring effects that contemporary administrative divisions impose upon our regions ; it is sometimes difficult to distinguish these practical structuring devices from the reality of regions in the past. Second, as we move from smaller scale studies to those that are comparative and hence broader in scope, we must be aware of all factors that structure our recovery and perception of the archaeological evidence.

Arnold, B., 1990. The past as propaganda: Totalitarian archaeology in Nazi Germany . Antiquity 64 :464-478. Audouze , F., 1987. The Paris Basin in Magdalenian times. In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer , pp. 183-200 . New York: Plenum Press. Bahn, P., 1984. Pyrenean Prehistory: A Palaeoeconomic Survey of the French Sites. Warminster: Aris and Phillips. Bailey , G. (editor), 1983 . Hunter-Gatherer Econom y in Prehistory . Cambridge: Cambridge University Press . Bender, B ., 1978. Gatherer-hunter to farmer: A social perspective. World Archaeolog y 10:204-222. Bettinger , R.L. , 1991. Hunter - Gatherers: Archaeological and Evolutionary Theory . New York : Plenum Press .

CONCLUSION

Binford, L.R. , 1964. A consideration of archaeological research design . American Antiquity 29:425-441.

At the present time, archaeologists and anthropologists have not yet reached a sophisticated understanding of how to move between and among the various scales of analyses, from the micro- to the macro-scale. The process of linking appropriate anthropological, or indeed other social , theory to the archaeological record remains a complex and sometimes controversial endeavor , as we attempt to identify exactly which variables vary in behaviorally meaningful ways and why they do so . Another problem involves moving from the theoretical framework of mobile hunter-gatherers that dictates consideration of broader, multi-site spatial realms to the diverse bodies of evidence that have been collected primarily within a cultural-historical framework. We are asking questions that beg thorough , rigorous , and scientific collection and analysis of every imaginable variable from an enormous number of individual sites. Yet, as Rigaud, Simek, and Hays so lucidly argue in the present volume, we are still hampered by the very lack of this kind of evidence.

Binford, L.R., 1965. Archaeological systematics and the study of culture process. American Antiqui ty 31 :203-2 10. Binford, L.R. , 1978. Nunamiut Ethnoarchaeolog y . New York: Academic Press. Binford, L.R ., 1980. Willow smoke and dogs ' tails: Hunter-gatherer settlement systems and archaeological site formation. American Antiquity 44:4-20. Boyle, K., 1990. Upper Paleolithic Faunas from South-West France. British Archaeological Reports International Series 557. Oxford : British Archaeological Reports. Brown, J.A., 1987. The case for the regional perspective: A new world view. In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer , pp. 365- 376. New York: Plenum Press.

This volume brings together a wide variety of researchers from different theoretical and research perspectives, all of whom have confronted similar epistemological and methodological challenges. The various theoretical and methodological attempts recounted in this volume represent different approaches to the understanding of a single , well studied macro-region during a crucial transitional period , the Late Pleistocene of Western Europe. The economic intensification exhibited by some of the hunter-gatherers of Late Pleistocene Europe laid the groundwork for the subsequent development of agriculture and the increasing cultural complexity that accompanied this revolution in human prehistory . Thus, the papers contained in this volume have broader implications and applications for a more accurate understanding of the role played by the concept of the region in anthropological and archaeological theory, as well as for a more accurate understanding of the people who inhabited them in the prehistoric past.

Burch, E., 1972. The caribou/wild reindeer as a human resource . American Antiquity 37:339-368. Burke, A., 1995. Prey Movements and Settlement Patterns during the Upper Paleolithic in Southwestern France. British Archaeological Reports International Series 619. Oxford: British Archaeological Reports. Clark, G., 1980. Comment on "The identification of prehistoric hunter-gatherer aggregation sites: The case of Altamira" (Conkey). CurrentAnthropology 21:621-622. Clark, G., 1994. Migration as an explanatory concept in Paleolithic archaeology. Journal of Archaeological Method and Theory 1:305-343 .

5

PRICE AND PETERKIN

Gaussen, J., I980. Le Paleolithique Superieur de Plein Air en Perigord. XIveme Supplement a Gallia Prehistoire. Paris : Centre National de la Recherche Scientifique.

Clark, G., and L. Straus, 1983. Late Pleistocene huntergatherer adaptations in Cantabrian Spain. In Hunter-Gatherer Econom y in Prehistory: A European Perspective, edited by G. Bailey, pp.131-167. Cambridge: Cambridge University Press .

Geertz, C., 1966. Religion as a cultural system. In Anthropological Approaches to the Study of Religion, edited by M. Banton, pp. 1-46. Association of Social Anthropologists Monograph 3. London: Tavistock.

CLIMAP, 1976 . The surface of the Ice Age earth. Science 191:1131-1137.

COHMAP, 1988. Climatic changes of the last 18,000 years. Science 241: 1043-1052 .

Gonzalez Morales, M.R. , 1992. Racines: La Justification Archeologique des Origines Regionales dans l'Espagne des Communautes Autonomes. In The Limitations of Archaeological Knowledge, edited by J. Shay and J. Clottes, pp. 15-27 . Etudes et Recherche Archeologiques de l'Universite de Liege 49. Liege: Universite de Liege .

Conkey, M.W., 1980. The identification of prehi toric hunter-gatherer aggregation sites : The case of Altamira. Current Anthropology 21 :609-630. Conkey, M.W., 1987 . Interpretive problems in hunter gatherer regional studies. In The Pleistocene Old World , edited by 0. Soffer, pp. 63-77. New York: Plenum Press.

Gordon , B., 1988. Of Men and Reindeer Herds in French Magdalenian Prehistory. British Archaeological Reports International Series 390. Oxford: British Archaeologi cal Reports.

Conkey, M.W., n.d. Between the caves: Open-air survey and Magdalenian social geography in the French MidiPyrenees. Paper presented at the 60 th Annual Meeting of the Society for American Archaeology, Minneapolis, MN, May 3-7, 1995.

Hodder, I., 1986. Reading the Past: Current Approaches to Interpretation in Archaeology. New York: Cambridge University Press. Jalut, G., 1976. La vegetation au Pleistocene superieur au debut de )' Holocene dans les Pyrenees. In La Prehistoire Franraise, Tome I: Les Civilisations Paleolithiques et Mesolithiques de la France, edited by H. de Lumley, pp. 512-516. Paris: Centre National del la Recherche Scientifique.

Delpech, F., 1988. Review of "Men and reindeer herds in French Magdalenian prehistory" (Gordon). Bulletin de la Societe Prehistorique Franraise 85:280-282. De Walt, B.R., and P.J. Pelto (editors), 1985. Micro and Macro Levels of Analysis in Anthropology: Issues in Theory and Research . Boulder: Westview Press.

Jalut, G., 1977. Donnees chronologiques, paleosylvatiques, et paleoclimatiques sur la tardiglaciaire et le postglaciaire de l'extremite orientate des Pyrenees. In Approche Ecologique de !'Homme Fossile, edired by H. Laville and J. Renault-Miskovsky, pp. 343- 345. Supplement au Bulletin AFEQ 47. Paris: Universite Pierre et Marie Curie.

Dibble, H., and A. Montet-White (editors), 1988. Upper Pleistocene Prehistory of Western Eurasia. University Museum Monograph 54 . Philadelphia: The University Museum, University of Pennsylvania. Foley, R., 1984. Putting people into perspective: An introduc tion to community evolution and ecology . In Hominid Evolution and Community Ecology, edited by R. Foley, pp. 1-24. London: Academic Press. Franken, E., and S. Veil, 1983 . Die Steinartefakte Gonnersdorf Wiesbaden: Frantz Steiner .

Jochim, M ., 1983. Paleolithic cave art in ecological perspec tive. In Hunter-Gatherer Economy in Prehistory, edited by G. Bailey , pp . 212-219. Cambridge: Cambridge University Press .

von

Jochim, M., 1987. Late Pleistocene refugia in Europe. In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer, pp. 317-332. New York: Plenum Press.

Gamble, C., 1984. Regional variation in hunter-gatherer strategy in the Upper Paleistocene of Europe. In Hominid Evolution and Community Ecology, edited by R. Foley, pp. 237-260. London: Academic Press.

of

Judge, W.J., J.I. Ebert, and R.K. Hitchcock, 1975. Sampling in regional archaeological survey. In Sampling in Archaeology, edited by J. Mueller, pp . 82-123 . Tucson: University of Arizona Press.

Gamble, C., and 0. Soffer, 1990 . Introduction. In The World at 18,000 B.P., edited by 0. Soffer and C. Gamble , pp. 1-23. London : Unwin and Hyman.

Koetje, T.A., 1987. Spatial Patterns in Magdalenian OpenAir Sites from the Isle Valley, Southwestern France. British Archaeological Reports International Series 346. Oxford: British Arch -aeological Reports.

Gamble, C ., 1986. The Paleolithic Settlement Europe. Cambridge: Cambridge University Press.

6

REGIONAL APPROACHES TO ADAPTATION

Raikes, R., 1967. Water,Weatherand Prehistory. London: Baker.

Kristiansen, K., 1990. National archaeology in the age of European integration. Antiquity 64:825-828.

Read, D ., 1975. Regional sampling. In Sampling in Archaeology , edited by J.W. Mueller, pp. 45-60. Tucson: University of Arizona Press.

Lee, R.B. , and I. DeVore (editors), 1968. Man the Hunter. Chicago : Aldine. Leroi-Gourhan, A., and M. Brezillon, 1972. Fouilles de Pincevent: Essai d'Analyse Ethnographique d'un Habitat Magdalenian. vneme Supplement a Gallia Prehistoire. Paris : Centre National de la Recherche Scientifique

Rensink, E., 1995. On Magdalenian mobility and land use in North-West Europe: Some methodological considerations. Archaeological Dialogues 2:85-119. Rigaud, J.-Ph., and J. Simek, 1987. "Arms too short to box with god:" Problems and prospects for Paleolithic pre history in Dordogne, France. In The Pleistocene Old World: Regional Perspectives , edited by 0. Soffer, pp.47-62 . New York: Plenum Press.

Milisauskas, S., 1990. People's revolutions of 1989 and archaeology in Eastern Europe. Antiquity 64:283-285. Mithen, S.J., 1990. Thoughtful Foragers: A Stud y of Prehistoric Decision Making. Cambridge: Cambridge University Press .

Rossignol , J., and L. Wandsnider (editors) , 1992. Space, Time, and Archaeological Landscapes. New York: Plenum Press .

Mueller, J.W., 1974. The Use of Sampling in Archaeological Survey . Memoir 28. Washington , D.C.: Society for American Archaeology . Mueller , J.W. (editor), 1975. Sampling in Archaeology . Tucson: University of Arizona Press .

Roveland, B.E ., n.d .. In the eye of the beholder: Changing views of Hamburgian variability on the Northern German Plain. Paper presented at the 60 th Annual Meeting of the Society for American Archaeology, Minneapolis, MN, May 3- 7, 1995.

Nitecki, M., and D. Nitecki (editors), 1987. The Evolution of Human Hunting. New York: Plenum Press.

Rust, A., 1958. Die Jungpalaolilthischen Zeltanlagen van Ahrensburg. Neumunster: K. Wachholtz .

Otte, M. (editor), 1988. De la Loire a I 'Oder: Les Civilisations du Paleolithique Final dans le Nord-Ouest Europeen . British Archaeological Reports International Series 444 and Etudes et Recherche Archeologiques de l'Universite de Liege 25. Oxford: British Archaeological Reports and Liege: Universite de Liege.

Sacchi, D., 1986. Le Paleolithique Superieur du Languedoc Occidental et du Roussillon. XXIeme Suplement a Gallia Prehistoire. Paris: Editions du Centre National de la Recherche Scientifique. Schortman, E.M., 1989. Interregional interaction in prehistory: The need for a new perspective. American Antiquity 54:52-65.

Peterkin, G.L., H.M. Bricker, and P. Mellars (editors), 1993. Hunting and Animal Exploitation in the Later Palaeolithic and Mesolithic of Eurasia. Archaeological Papers of the American Anthropological Association 4. Washington, D.C.: American Anthropological Association.

Schrire, C. (editor), 1984. The Past and Present in Hunter-Gatherer Studies . New York: Academic Press. Simmonet, R., 1985. Le silex du Magdalenien final de la Grotte des Eglises dans le Bassin de Tarasconsur-Ariege. Bulletin de la Societe Prehistorique de l 'Ariege 40:71-97.

Peyrone, 0., J. Gulot, R. Cheddadi, P. Tarasov, M. Reille, J.-L. de Beaulieu, S. Bottema, and V. Andrieu, 1998. Climatic reconstruction in Europe for 18,000 B .P. from pollen data. Quaternary Research 49: 183-196.

Soffer , 0., 1987. Upper Paleolithic connubia, refugia, and the archaeological record from Eastern Europe. In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer, pp. 333-348. New York: Plenum Press.

Pike-Tay, A ., 1991. Red Deer Hunting in the Upper Paleolithic of South- West France: A Study in Seasonality. British Archaeological Reports International Series 569. Oxford : British Archaeological Reports.

Spiess, A., 1979. Reindeer and Caribou Hunters: An Archaeological Study. New York: Academic Press.

Poplin, F., 1976. Les Grands Vertebres de Gonnersdorf: Fouilles 1968. Wiesbaden: Frantz Steiner.

Stafford, C.R., 1995. Geoarchaeological Perspectives on paleolandscapes and regional subsurface archaeology. Journal of Archaeological Method and Theory 2:69-104.

Price , T.D., and J.A. Brown (editors), 1985. Complexity Among Prehistoric Hunter-Gatherers. Orlando: Academic Press.

7

PRICE AND PETERKIN

Steward, J.H., 1938. Basin-Plateau Aboriginial Sociopolitical Groups. Bureau of American Ethnography Bulletin 120. Washington , D.C.: Bureau of American Ethnography.

American Museum of Natural History 59. New York: American Museum of Natural History. Thomas, D.H., 1985 . The Archaeology of Hidden Cave, Nevada. Anthropological Papers of the American Museum of Natural History 61, Part 1. New York: American Museum of Natural History.

Stiner, M., 1994. Honor Among Thieves: A Zooarchaeological Study of Neandertal Ecology. Princeton: Princeton University Press.

Tomaskova, S., n.d. Europe viewed from the East. Paper presented at the 61 st Annual Meeting of the Society for American Archaeology , New Orleans, LA , April 10-14, 1996.

Straus, L.G. (editor), 1986. The End of the Paleolithic Old World. British Archaeological Reports International Series 284. Oxford: British Archaeological Reports. Straus, L.G., and G. Clark (editors), 1986. La Riera Cave: Stone Age Hunter- Gatherer Adaptations in Northern Spain. Anthropological Re earch Paper s 36. Tempe: Arizona State University.

Van Reybrouk, D., 1994. Changing perspectives on huntergatherers in Continental and in Anglo-American archaeology. Antiquity 68:831 -837.

Straus, L.G., B.V. Eriksen, J.M. Erlandson, and D.R. Yesner (editors), 1996. Humans at the End of the Ice Age: The Archaeology of the Pleistocene-Holocene Transition. New York: Plenum Press.

Veil, S., 1984. Siedlungsbefunde von Magdalenien -Fundplatz Andernach. Zwi chenbericht Uber die Grabungen 1979-1983. In Upper Palaeolithic Settlement Patterns in Europe, edited by H. Berke, J. Hahn, and C.J. Kind, pp. 181- 193. Tlibingen: Archaeologica Venatoria.

Struever, S., 1968. Problems, methods, and organization: A disparity in the growth of archaeology. In Anthropological Archaeology in the Americas, edited by BJ . Meggars, pp. 131-151. Washington, D.C.: Anthropological Society of Washington.

White, R., 1985. Upper Paleolithic Land-Use in the Perigord: A Topographic Approach to Subsistence and Settlement. British Archaeological Reports International Series 253. Oxford: British Archaeological Reports .

Struever, S., 1971. Comments on Archaeological data requirements and research strategy. American Antiquity 36:9-19.

Winterhalder, B., and E. Smith (editors), 1981. HunterGatherer Foraging Strategies: Ethnographic and Archaeological Analyses. Chicago: University of Chicago Press.

Sturdy, E., 1975. Some reindeer economies in prehistoric Europe. In Paleoeconomy, edited by E. Higgs, pp. 55-95. Cambridge: Cambridge University Press .

Willey, G.R., and P. Phillips, 1958. Method and Theory in American Archaeology. Chicago: University of Chicago Press.

Thacker, P., 1996. Hunter-gatherer lithic economy and settlement systems: Understanding regional assemblage variability in the Upper Paleolithic of Portuguese Estremadura. In Stone Tools: Theoretical Insights into Human Prehistory, edited by G.H. Odell, pp. 101-124 . New York: Plenum Press .

Wobst, M.H., 1974. Boundary conditions for Paleolithic social systems: A simulation approach. American Antiquity 39: 147-178. Wobst, M.H., 1990. Afterword: Minitime and megaspace in the Paleolithic at 18K and otherwise. In The World at 18,000 B.P., edited by 0. Soffer and C. Gamble, pp. 331-343. London: Unwin Hyman.

Thomas, D.H., 1983. The Archaeology of Monitor Valley 2: Gatecliff Shelter. Anthropological Papers of the

8

2 The Magdalenian of Grotte XVI (Dordogne, France) and Regional Approaches to Magdalenian Settlement and Economy Jean-Philippe Rigaud (Institut du Quaternaire) Jan F. Simek (University of Tennessee) Maureen A. Hays (College of Charleston)* ABSTRACT Regional analyses of Magdalenian sites in southwestern France suffer from the use of old collections unsuitable to their goals and from a lack of new sites that can help to advance regional models. A recently excavated Magdalenian deposit from the Grotte XVI (Dordogne, France) has yielded a rich archaeological record consistent with the goals of regional reconstruction. This paper discusses analyses of faunal assemblages, stone tool typology and technology, artifact spatial distributions, and use wear on stone tools; all suggest a short-term, specialized use of the cave by mobile task groups, probably hunters . The Grotte XVI results are discussed in relation to regional understandings of land use and site variability.

INTRODUCTION Regional-scale analysis of Magdalenian settlement has been a major goal for Paleolithic archaeologists in southwestern Europe for a number of years. Such analyses have emphasized the economics of Magdalenian land use (e.g., Straus 1986, 1991), social organization (Conkey 1978, 1980; White 1980, 1987), or they have concentrated on specific aspects of Magdalenian site content over space, such as faunal remains (Delpech 1987) or artwork (Conkey 1978, 1989). In many

cases, regional analyses have as their primary goal the reconstruction of ethnographic-like accounts of the Magdalenian past. This goal obviously requires a large sample of sites , and regional analysis of the Magdalenian in southwestern Europe has traditionally employed all available site assemblages, including a number that were excavated many years ago. As we have argued previously , however, the vast majority of available data were collected long before the application of the precise and detailed excavation strategies we now know

a

b

w

a

Cb tl

Figure 2.1- Compari son between old and recent Magdalenian site excavations . 2.1a) Peyrony 's type-profile 2 of Laugerie-Haute Est (after La ville et al. 1980: 302) . 2. lb ) Lavill e's interpretation of the same profile (after Laville et al. 1980: 303).

*Author to whom correspondence should be addressed.

RIGAUD, SIMEK, AND HAYS

interest to behavioral archaeologists. Thus, we remain firm in our view that only modern data generated with strategies explicitly designed to examine economic, social , and cultural process can be used to such ends. We do not see why our original assessments should be different today, since there have been no new techniques for examining old collections that reduce or eliminate the problems we outlined in 1987. The arguments for that position are presented in the 1987 pub lication. However, we will highlight a few of them here as they pertain to the Magdalenian. Two major areas of concern impact the use of old data : biase s in site content s and biase s in site locations (Rigaud and Simek 1987). Both of these bia es, in fact, reflect aspects of site formation processes and how they have been understood in different ways by archaeologists in the past and today . Castelnaud

N

---~-----~--"".'.'.'.'.'=::::=~:::-.. 16 I5

14 , 2 13

S

F~~ ,o 19 1-ln-r-TT l '~~ I 1i I 11 18 17 8 6 4 2 20 22

Figure 2.2- Map showing the Le Conte Cliffs (Cenac-et -St.-Julien) and their location in southwest France. Grotte XVI is numbered " 16" at the left of the cliff drawing .

are required for accurate reconstructions of past behaviors (Rigaud and Simek 1987). For the most part, early excavations had interests other than ethnography and were technically designed to elucidate aspects of chronology and culture history. In particular, site formation processes were all but ignored, resulting in assemblages of unknown structure. Yet, today, some Paleolithic archaeologists continue to use these data toward ends for which they have little or no warrant. In this paper, we reiterate our previous considerations of regional-scale analysis using old data, with specific examples from the Magdalenian. We then illustrate the kind of data we think appropriate to building regional models for Magdalenian land use by presenting results of our recent studies of the Magdalenian from the Grotte XVI (Dordogne). We hope to show that, until enough sites have been excavated using techniques that generate data relevant to reconstructions of site use, regional syntheses are premature at best.

PROBLEMS WITH REGIONAL-SCALE DATA IN SOUTHWEST EUROPE In 1987, we tried to show why we believe that most early archaeological collections from southern France are not useful for modern regional analysis (Rigaud and Simek 1987). We have never advocated "throwing the baby out with the bathwater" (cf. Kuhn 1995), but because of the way they were collected, early assemblages have no "baby" in them, at least as far as paleoethnography is concerned. The basic problem with old excavations, of course, is that they were designed to answer rather general questions concerning culture sequences and climatic history, not those of interest to paleoethnographers ; excavations normally emphasized vertical relations among artifacts rather than the horizontal relationship s of

Biases in site contents come about because the stratigraphic unit defined during excavation must form the countin g unit for inter-site comparisons . If, as is the case in southern France, those stratigraphic units were based on different criteria during the late 19th and early 20 th centuries than they are todaY., then the assemblages resulting from different times simply cannot be treated as equivalent units. Furthermore, because archiving of collections is determined permanently by those unit definitions, we can no longer "deconstruct" the early assemblages into units that are equivalent to those we collect today. An example of this problem for the Magdalenian comes from the important site of Laugerie-Haute Est in Dordogne, France . Extensively excavated by D. and E. Peyrony ( 1938) from 1921-1935, the site yielded a long sequence of late Paleolithic deposits comprising 12 strata (Figure 2.1 a); four strata (Peyrony's I', I", I"', and J) contained rich Magdalenian assemblages. When Bordes re-excavated the site with Laville as stratigrapher (Laville et al. 1980) 42 layers were defined within the same deposit (Figure 2.1 b). Thus, the assemblage units defined at Laugerie-Haute by Bordes cannot be compared in a meaningful way to those derived by Perony since they are based on two different scales of temporal resolution for the same site. Early excavation strata represent palimp-sests of the strata defined later by more refined methods. Micromorphology (e.g., Courty et al. 1989) has taught us that the more-recently defined strata, too, represent palimpsests of individual occupations (Rigaud et al. 1995). Therefore, the prognosis for using such assemblages to recon struct past regional land-use patterns is not good. Biases in site locations result from geological processes (slope processes, mass wasting, asymmetrical valley formation) that affect the nature and location of sites (Rigaud and Simek 1987). Obviously, even at the end of the Pleistocene, regional-scale geological changes obliterated sites, altered stratigraphies, and modified artifact positions and forms (e.g., Laville et al. 1980; Bertran and Texier 1995). These same processes can also produce concentrations of artifacts easily confused with sites (Rigaud and Simek 1987). Thus , even site distribution s must be considered in term s of human land -use 10

THE MAGDALENIAN OF GROITE

XVI

REAR GALLERY

0

23 22

4m

21

20 19

lower passageway "COULOIR" upper passageway ~

14

13 12 11

10

MAIN CHAMBER

9

8 7 6

5 4

3 grid origin

¥

2

grid anchor drip line

entry gate ABCDEFGHI

K

L M N 0

Figure 2.3-Map of Grotte XVI showing grid system and areas defined within the cave. The Magdalenianis concentrated in the Rear Gallery at the back of the cave.

patterning only after the geomorphological context of the region is understood in historical detail.

We must turn to sites excavated with modern techniques if we hope to undertake regional analysis. But there are very few sites, Magdalenian or earlier, that have been excavated _in southwest France using methods and approaches that warrant inclusion in a regional sample designed to elucidate behavior. One such site is the Grotte XVI, currently under excavation, has yielded a late Magdalenian level, Couche 0. In the remainder of this paper, we will try to show how an individual site can be analyzed in such a way as to warrant its integration into regional models for site use. We believe that building upon analyses of sites like Grotte XVI is the best way to eventually understand regional land-use patterns in Paleolithic Europe.

Local and regional historical processes have acted to produce biased assemblages from even late Paleolithic sites in southwestern France. Biases in site contents, resulting from changing paradigms among prehistorians, preclude comparison among assemblages collected at different times, because we simply do not know how comparable the assemblage units are in chronological and geological terms. Biases in site locations are introduced by geomorphological processes that, for many old collections, cannot be reconstructed today. Overall, site formation processes, which we now know must be understood if accurate reconstructions of site use are to be attempted, simply cannot be assessed for the majority of sites typically used in regional-scale analysis of the Magdalenian from southern France. For this reason, accurate reconstructions of individual site use are indispensable if regional models are to be constructed and tested.

THE CASE OF THE GROTTE XVI In 1983, test excavations were undertaken in the Grotte XVI, a large karstic cavity in the Le Conte Cliffs in the Dordogne region of southwest France (Figure 2.2). Grotte XVI is one of 11

RIGAUD, SIMEK, AND HAYS

at least 23 caves and shelters in the massif that include Grotte Vaufrey (Rigaud 1988) and Grotte Noir, among others (Rigaud 1982). Further excavations in 1984 revealed that the Grotte XVI contains a stratigraphic profile representing the entire Upper Paleolithic cultural sequence with underlying Mousterian levels. A long-term project was launched in 1985 and continues to this day as a cooperative venture of the Institut du Quaternaire at Bordeaux and the University of Tennessee. Though excavations at the Grotte XVI have produced a rich archaeological record (Rigaud and Simek 1991, 1993, 1994, 1995; Rigaud et al. 1995, 1996), this paper concentrates on describing the uppermost, intact deposits from Couche O dating to the Magdalenian occupation of the cave. We hope to show that analy es of the associated fauna! as emblage, stone tool typology , technology , raw material analyses , and artifact spatial distribution , as well as functional analyses, all suggest a short-term , specialized use of the site by small, mobile task groups, probably hunter . This interpretation implie that use of Grotte XVI during the Magdalenian was part of a larger, logistically organized collector land-use strategy. The nature and structure of Magdalenian site use at this location is, at least for now, somewhat different than for the earlier Upper Paleolithic occupants of the Grotte XVI.

Excavation Strategy Excavations in the Grotte XVI were designed to obtain a large sample of archaeological and geological information from all

areas of the cave. Over 90 meter-square excavation units have been opened (Figure 2.3). Excavation techniques are designed o that all sedimentary materials over one centimeter in size are recovered with three-dimensional spatial provenience . All ediment is water washed through 2 mm wire mesh to recover small faunal materials and artifacts missed during excavations. These materials are collected by both level and unit provenience. This tight three-dimensional control allows for detailed spatial analysis of the artifacts. Extensive geological and paleoenvironmental sampling, including micromorphology, is undertaken in all areas as a matter of course.

The Magdalenian of Grotte XVI The Magdalenian depo its found in Couche O are isolated in the upper levels of the back portion of the cave . This gallery comprises approximately 30 square meters, nearly one-fifth the area of the entire cave. Micromorphology shows that the sediments have been di turbed by periglacial soil movement in only a few small areas; otherwise, Couche O artifacts are undisturbed and in situ. Moreover, the layer is shallow and limited in its spatial distribution. Therefore, we have been able to excavate it in its entirety . Because of its exceptional preservation, the Magdalenian deposit is the best documented level we have so far excavated (Rigaud and Simek 1995). Analyses of the fauna by Fran~oise Delpech, of the lithic typology and technology by Jean-Pierre Chadelle and Helene Marino, of microwear evidence, and of the assemblages and their spatial distributions have been accomplished. We will discuss these studies briefly in turn.

100 90 80

■ ■

70 60

~ ~

~

C/.l

z

50

'?ft

40

~

10

o-

Birds SmallMammals Bovids

m Equids rmFish

30 20

Reindeer

-



, ,

Carnivores

,/ /

,,,

ITT

Reindeer WithoutReindeer

Figure 2.4-Frequencie s (NISP) of different animal taxa in the Couche O fauna! assemblage. As reindeer dominate the assemblage (left), frequencies are also shown without using reindeer bone s in frequency calculation s (right) .

12

THEMAGDALENIAN OF GROTTEXVI FemP



100

FemD

90

80 70

-

TibP



60 Cote

~

c.:, ~



50 40

• vi • HumD • rars • calc • mnP • RadP e MtD • RadD Pha2 • carp ee Phal e Pha3 e Ax e Max • McD

30 20 10 0

e Pelv • scap

e vt e ribD • HumP

0

10

20

30

e MtP

40

50

60

70

80

• Mand

e McP

90

100

%MAU Figure 2.5-Relationship

between %MAU and MGUl for Couche O reindeer. The curve resembles that for meat transport from a hunting camp (Binford 1978) .

The Faunal Assemblage

frequently at the site, making up 90% of the assemblage (Figure 2.7). Still, there is some use of materials from distant sources, with I 0% of both tools and debitage made of nonlocal material. Flint from sources arnund Bergerac (some 60 km away) is the most common non-local material.

Delpech's faunal analysis indicates that reindeer (Rangifer tarandus) is by far the dominant species present in Couche 0 (Figure 2.4 ), comprising 91 % of the assemblage. The remaining 9% is divided among birds, bovids, carnivores, equids, and small mammals.

Using Chadelle's chafne operatoire (Chadelle 1983), stone tool technology in Couche Ois characterized primarily by three stages of production. Stage 2 products (blades, crested blades, and blade cores) constitute nearly 80% of the lithic material (Figure 2.8). Stage 1 (unworked cores and decortification flakes) is present but quite under-represented if initial core reduction were practiced at Grotte XVI. Thus, technological analyses suggest that fully formed blade cores, procured and shaped elsewhere , were brought to the site for transformation into tools.

Delpech 's comparison of the relation between Minimal Animal Units (MAU) and Modified General Utility Index (MGUI) for reindeer (Figure 2.5) shows a curve similar but not identical to Binford's "Reverse Utility Curve" (Binford 1978). This kind of curve is seen ethnographically as reflecting meat preparation at or near a kill prior to transport, but Delpech views this interpretation with caution . Still, the Couche O fauna} assemblage, dominated by a single species butchered in the Grotte XVI, is one piece of evidence suggesting a short-term hunting camp.

Couche OSpatial Analysis Even though the artifact distribution in Couche O is very dense, it has a rather simple spatial order. Two main areas of accumulation are apparent to the naked eye (Figure 2.9). This structure is confirmed by quantitative assessment using k-means cluster analysis (Kintigh 1990; Kintigh and Ammerman 1982). A two-cluster optimal solution is indicated by a change in the log (%SSE) (Figure 2.10), and two discrete clusters, Cluster 1 near the gallery mouth and Cluster 2 at the back of the gallery, can be mapped onto the Couche 0 area (see Figure 2.9). Cluster 1, it should be noted, is located near the only evidence for a hearth feature recovered from Couche 0. The dichotomous pattern in artifact distributions can be explained, at least in part, by the fact that artifact

The Lithic Assemblage There are 3,788 pieces of lithic debris from Couche O at the Grotte XVI. Debitage comprises nearly 85% of the lithic material, while the other 15% of the assemblage is formally typed tools. Microliths by far dominate the assemblage at 73% (Figure 2.6). Scrapers, burins, and pieces with lateral retouch are roughly equal in abundance and make up the basis of the rest of the lithic collection . Chadelle and Marino's analysis of the lithic raw materials from Couche O indicates that local materials are used most 13

RIGAUD, SIMEK, AND HAYS

100 90

80

Microlith

70

Burin Scraper Retouch

~ 60 C'd

-E ..0

Cl)

50

r./J r./J

< 40

Notch Composite Piercer

'$-

30

20 10 0

With Microliths Without Microliths Figure 2.6-Frequencie s of different stone tool type groups in the Couche O lithic assemblage . As microliths dominate the assemblage (left), frequencies are also shown without using microliths in frequency calculations (right).

■ □

100

Cl)

OJ)

-8

~

Local Exotic

100

90

90

80

80

70

70

Phase 1 Phase 2



Phase 3

~

Cl)

60

-E: 60 0()

i:d

..0

Cl)



.0

50

(lJ

w w

v:i v:i

< 40

50

< 40

?F-

~

30

30

20

20

10

10

0 Total

Debitage

0

Tools

Figure 2.7-F requncie s of lithic raw material types in the Couche 0 as emblage .

Figure 2.8-F requencies of lithic technology stage classes in the Couche 0 assemblage.

14

THE MAGDALENIAN OF GROTIE

hunters dug several holes in the middle of the gallery, dividing the artifact-bearing layers with their efforts. However, there are real content differences between the two cluster areas, which will be the focus of our discussions below. These differences reflect activity variation and organization of tasks in space. Thus, the two spatial clusters visible both intuitively and quantitatively must have some behavioral relevance beyond simple definition by artifact-hunter intrusions. We turn now to examining content differences between Cluster 1 and Cluster 2.

XVI

10 9 Cluster 2

8 7

Fauna

6

Faunal variation between clusters is examined on two dimensions: species representation and anatomy. Figure 2.11 shows the proportion of each cluster comprised by reindeer compared to all other species, and Figure 2.12 illustrates the nonRangifer taxonomic content of each spatial cluster. Reindeer occur in equal proportions in both areas. Equid and bird bones are relatively more frequent in the back of the gallery; bovid and fish bones are relatively more frequent toward the gallery opening near the hearth. Since all of these species are rare when compared to reindeer, it may be that individual animals or parts of individuals were processed in only one or the other of the two cluster areas.

5 4 3

Cluster 1

2 1

Figure 2.13 shows the distribution of reindeer body parts by cluster. There are few or no evident differences between the two clusters; anatomical parts are dispersed in similar proportions. Thus, the two zones do not seem to represent different activity areas with respect to processing reindeer.

T

u

w

V

Figure 2.9-Plan map of artifact distributions for Couche 0. The two cluster k-means cluster analysis solution is overlaid on the artifact point pattern .

2.5

2 1.5

1

1

2

3

4

5

6

7

8

9

1

Figure 2.10-Log (%SSE) plot for Couche O artifacts (heavy line) compared with a series of runs on random data (light lines). A marked inflection at two cluster level indicates an optimal olution at that scale (Kintigh and Ammerman 1982).

15

RI GAUD, SIM EK, AND HAY S

■ □

Reindeer Exotic

In sum , no dramatic distinctions can be defined between the front and back area of the gallery based on the animal species they contain ; activity differentiation in reindeer processin g does not seem to have a spatial dimension here either .

100

Lithics

90 Chipped stone artifact s are examined for three dimensions : raw material , technology , and tool type . Below , we also con sider tool function , as de cribed through high -magnification microwear analysis.

80 70

Figure 2.14 shows that stone raw material s are distributed equally in both area of the Couche O occupation . Both clu ters are dominated by local material s, with small quantities of both Berg eraco is and other exo tic materials present in each cluster. Thi s uniformity is encoura ging, becau e it support s an inference of a single occ upation episode.

60 ~ ~

z

1---4

~

50 40 30

Figure 2.15 illustrate s the distribution of technological groups between Cluster 1 and Clu ster 2. Again, the two areas are indistinguishable from thi s point of view; as for the Magdalenian assemblage as a whole , reduction Stage 2 is the most frequent, indicating that tool manufacture occurred across the occupied area. A few cores are present in each cluster, along with a similar frequency of tools; the debitage to tool ratio is about 5: I in both clusters. Thus, technological characteristics of both clusters are very similar, further warranting the identification of a single occupation episode as sugge sted by raw material frequencies .

20 10 0

Cluster 1

Cluster2

Figure 2.11- Frequencies (NISP) of animal taxa compared between two k-mean s clusters. Here, reindeer are compared to all other taxa combined .

40

Birds

30



~ 20

II

Equids Carnivores

10



~ SmallMammals ~ Bovids

~ C/.) 1---4

z

~ Fish

Cluster 1 Figure 2.12-Frequencies

Cluster 2

(NISP) of non-reindeer animal taxa compared between two k-means clusters.

16

TH E MAGDA LENIAN OF GROTT E

XVI

100 90

80 70





60 50 40

Antlers

~ ~

Cranium Vertebrae Ribs

~

Feet

m Legs

30 20

10

Cluster 1

Cluster 2

Figure 2.1 3-- Frequencies (NISP) of reindeer keletal elements compared between two k-means clusters.

■ □

100

...-

c::

r.n

..a

u 20 ~

10

0

Cluster 1

Cluster 2

Figure 2.19- Frequencies of stone tool use action compared between two k-means clusters.

19

RIGAUD, SIMEK, AND HAYS

50

■ ~

Antler/Bone Hide/Meat

D

Wood

Keeley ( 1980), were employed. In this analysis 100% of the formally typed tools were analyzed, and their use wear interpreted.

In microwear analysis it is possible to assess both the material that the tool was used on and the motion or past action of the tool. The assessment of the kinematics, in combination with the material that the edge came in contact with, provides for a reasonable interpretation of prehistoric activities.

40

~ 0

A wide range of motions have been interpreted from the use wear ob erved on the tools from the Magdalenian occupation of Grotte XVI. These are, however, within the range of typical prehistoric activities and comprise cutting, scraping, graving, and boring. Figure 2.18 shows that, of the implements for which motion could be assessed, the most frequently interpreted motion was cutting. Scraping ranks as the second most prevalent activity, while graving was represented by only a few tools less than scraping. Tools used for boring are very carce.

30

0

E-0

cu C/.l

:=>20 '#-

When viewed independently from spatial patterning, these observations are interesting but predictable at an Upper Paleolithic site. However, activities do have a spatial dimension in Couche 0. Fig ure 2.19 shows the distribution of motions in Cluster 1 and Cluster 2. Although each cluster contains the full range of motions represented at the site, Cluster 1 has a higher percentage of wear resulting from a graving motion whi le Cluster 2 is dominated by evidence of wear from scraping; this is in association with a high proportion

10

0 Figure 2.20-Frequencies in microwear.

of materials worked by stone tools as seen

100 90 80 ....,_)

c:::

70

(l)

....,_)

c::: 0

u '(l) ....,_)

00

60 50

■ ~

Antler/Bone Hide/Meat

40

D

Wood

...2

u

~

30 20 10 0 Cluster 1 Figure 2.21-Frequencies

Cluster 2

of materials worked by stone tools compared between two k-means clusters.

20

THE MAGDALENIAN OF GROTTE

XVI

50

40 ....,_J

= cu

■ ■

....,_J

§ 30

u

"cu

~ ~

....,_J

rn

..2

u 20 '#-



Cut Antler/Bone Grave Antler/Bone Scrape Antler/Bone Cut Hide/Meat Scrape Hide/Meat

10

0

Cluster 1 Figure 2.22-Functional

Cluster 2

variability compared between two k-means clu ster .

Grotte XVI. There appears to be two functionally distinct activity areas (Figure 2.22). Cluster 1, in the front portion of the gallery and associated with a hearth, appears to be dominated by the graving and scraping of both antler and bone. The front of the gallery is also the area that has the highest proportion of microliths . This combination is indicative of the manufacture and repair of bone and antler items and composite projectile points . Cluster 2, in the back portion of the gallery, seems to be dominated by cutting and scraping hide and meat. These activities are associated with the early stages in the processing of hides . Thus , functional assessments in combination with tool distributions in the two spatial clusters at Grotte XVI support the interpretation that the Magdalenian occupation of the site was both short and activity specific.

of typological scrapers in the back section of the gallery. Thus, activity area differentiation is indicated by microwear evidence. Still, to identify the actual activities carried out in specific areas , we must combine all of the above with an assessment of the materials worked during the Magdalenian occupation. There seems to be a fairly limited range of materials worked at Grotte XVI, and they are fairly evenly distributed across the classes defined for motion. These materials, as represented by the presence of distinct polishes, are: bone, antler, hide, meat, and wood. Figure 2.20 shows that, where material could be identified, bone and antler working were the dominant activities. Tool use on hide and meat ranks as the second most prevalent activity, while only a few tools were used to work wood . When these functional data are examined in space (Figure 2.21 ), fairly strong distributional patterns emerge . Cluster 1 is dominated by the processing of bone and antler, while Cluster 2 is dominated by hide and meat processing.

DISCUSSION Reindeer carcass processing , external origins and core production of lithic resources, simple spatial structure , and limited activity performance all contribute to our interpretation of Couche Oas a short-term hunting camp occupation. Meat and hides were processed, and weapons were retooled . No evidence for long-term occupation or more diverse activities is present. Thus, even in a stratified cave context, it is clear that Magdalenian peoples occupied some places for highly specific tasks. In this case, a hunting camp is indicated, reflecting one aspect of a mobile collector land-use pattern. Other functionally specific "structural poses" will certainly be recognized elsewhere.

In a site dominated by microliths, it is hard to ignore their technological and functional roles. Use wear analysis of the 458 microliths showed that only a small percentage (20%) were actually used. The other 80% of the microliths were unused. We believe this reflects production and retooling at the site rather than the extensive use of composite tools . In sum, functional analyses of the Magdalenian assemblage indicates that a limited number of activities took place at the 21

RIGAUD, SIMEK, AND HAYS

CONCLUSION

Conkey , M. , 1989. Structural

In Archaeological We have argued that regional-scale analysis of Magdalenian sites and assemblages, if they hope to be accurate reconstructions of the past, cannot rely on data obtained by early excavations. These data are simply too variable in resolution and scale, and they do not entail information that would allow critical evaluation of the site formation context of the assemblage. We simply do not know how comparable these assemblages are, and they cannot serve to construct models for regional land use if they are not comparable .

analysis

of Paleolithic art. edited by pp. 135-154. Cambridge:

Thought in America,

C.C. Lamberg-Karlovsky, Cambridge University Press.

Courty, M.A., P. Goldberg, and R. Macphail, 1989. Soils and Micromorphology in Archaeology. Cambridge: Cambridge University Press. Delpech, F., 1987. L'Environnement animal des Magdaleniens. In

Le Magdalenien en Europe La Structuration du Magdalenien, edited by J .-Ph. Rigaud, pp. 5-30. Etudes et Recherches Archeologiques de l'Universite de Liege 38. Liege: Universite de Liege.

We are left with looking to new data, and while there are admittedly few site that can serve at this point in time, recent excavations and site survey programs currently underway promise to provide the kinds of information necessary to pur sue regional goals . We have tried to present an example here of how we think individual sites can contribute to regional analysis : by providing accurate and warranted reconstruction of how Magdalenian hunters used individual places. Obviously, Magdalenian subsistence and ettlement can only be understood at the regional scale. But such a regional understanding will be contingent on fine-scale, site-specific analyses like those described here. Without such "building blocks," the nature and complexity of Magdalenian settlement systems will continue to be underestimated.

Keeley, L., 1980 . Experimental Determination of Stone Tool Use: A Microwear Analysis. Chicago: University of Chicago Press. Kintigh , K., 1990. Intrasite spatial analysis: A commentary on major methods. In Mathematics and Information Science in Archaeology: A Flexible Framework , edited by A. Voorrips, pp. 165-200. Bonn: Holos. Kintigh, K., and A. Ammerman, 1982 . Heuristic approaches to spatial analysis in archaeology. American Antiquity 47:31-63.

ACKNOWLEDGMENTS Kuhn, S., 1995 . Mousterian Lithic Technology: An Ecological Perspective. Princeton: Princeton University Press.

The authors wish to thank the Ministry of Culture of France, the U.S. National Science Foundation, the Department of the Dordogne, the National Geographic Society, and the University of Tennessee Professional Development Program for contributing funding to this project. We also thank the hundreds of volunteer excavators who recovered the data we analyze here.

Laville, H., J.-Ph. Rigaud, and J.R. Sackett, 1980. Rockshelters of the Perigo rd. New York: Academic Press. 1938. Laugerie-Haute, pres Archives de l'lnstitut de Paleontologie Humaine Memoire 19. Paris: lnstitut de Paleontologie Humaine.

Peyrony, D., and E. Peyrony,

des Eyzies (Dordogne). REFERENCES Bertran, P., and J.-P. Texier, 1995. Fabric analysis: Application to Paleolithic sites. Journal of Archaeological Science 22:521-535.

Rigaud, J.-Ph., 1982. Le Paleolithique en Perigord Les donnees du sud-ouest sarladais et leurs implications. These de Doctorat d'Etat e Sciences. Universite de Bordeaux I, Bordeaux.

Binford, L., 1978. Nunamiut Ethnoarchaeology. Orlando, FL: Academic Press. Chadelle,

Rigaud, J.-Ph., 1988. La Grotte Vaufrey: Paleoenvironnement, Chronologie, Activites Humaines . Paris: Societe Prehistorique Franc;aise.

J.-P., 1983. Technologie et Utilisation du Silex

au Perigordien Superieur: L'Exemple de la Couche VJJ du Flageolet I. L'Ecole de la Hautes Etudes en Sciences

Rigaud, J.-Ph., and J. Simek, 1987. Arms too short to box with God: Problems and prospects for Paleolithic prehistory in Dordogne, southwestern France . In The Pleistocene Old World: Regional Perspectives, edited by 0 . Soffer, pp. 47-61. New York: Plenum Press.

Sociales. Bordeaux: Universite de Bordeaux I. Conkey, M., 1978. Style and information in cultural evolution: Toward a predictive model for the Paleolithic. In

Social Archaeology Beyond Subsistence and Dating, edited by C.L. Redman and M.J. Berman, pp. 61-85. New York: Academic Press.

Rigaud, J.-Ph., and J. Simek, 1991. Cenac-et-St. Julien: Grotte XVI. In Gallia Informations: Aquitaine , edited by J.-Ph. Rigaud, P. Garmy, B. Bizot, and A. Collier, pp. 13-14. Paris: Centre National de la Recherche Scientifique.

Conkey, M., 1980. The identification of prehistoric huntergatherer aggregation sites : The case of Altamira. Current Anthropolog y 21 :609-630. 22

THE MAGDAL ENIANOFGR0TIE XVI Rigaud , J.-Ph., and J. Simek, 1993. Cenac-et -St. Julien: Grotte XVI. In Bilan Scientifique de la Region Aquitaine 1992 , edited by D. Barraud , p. 24. Bordeaux: Ministre de la Culture.

R. March , and M. Pipemo , pp. 77- 80. Proceedings of the XIIIth International Congress of Prehistoric and Protohistoric Sciences 5. Forli , Italy: ABACO Edizioni . Straus, L.G. , 1986. Late Wi.irm adaptive systems in Cantabrian Spain : The case of eastern Asturias. Journal of Anthropologi cal Archaeolog y 5:330-368 .

Rigaud, J.-Ph., and J. Simek, 1994. Cenac-et -St. Julien : Grotte XVI. In Bilan Scientifique de la Region Aquitaine 1993, edited by D. Barraud , pp. 25- 26. Bordeaux: Ministre de la Culture.

Straus, L.G ., 1991. Human geography of the Late Upper Paleolithic in Western Europe: Present state of the question. Journal of Anthropological Research 47:259 - 278 .

Rigaud, J.-Ph., and J. Simek, 1995. Cenac-et-St. Julien: Grotte XVI. In Bilan Scientifique de la Region Aqui taine 1994, edited by D. Barraud , pp. 24-25. Bordeaux: Ministre de la Culture . Rigaud, J.-Ph., J. Simek, and T. Ge, 1995. Mousterian fires from Grotte XVI (Dordogne , France) . Antiquity 69:902 - 912.

White , R., 1980. The Upper Paleolithic Occupation of the Perigord: A Topographic Approach to Subsistenc e and Settlement. Ph.D. dissertation . Univer sity of Toronto, Toronto .

Rigaud, J.-Ph. , J. Simek , and T. Ge, 1996. Structures de combustion du Mousterian de la Grotte XVI a Cenac -et-St. Julien (Dordogne , France). In The Lower and Middle Paleolithic, edited by 0. Bar-Yosef, L. Cavalli-Sforza,

White , R., 1987. Glimpses of long-term shifts in Late Paleolithic land use in the Perigord . In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer, pp. 263-277. New York: Plenum Press.

23

3

The Relevance of Regional Analysis for Upper Paleolithic Archaeology: A Case Study from Portugal Paul T. Thacker (Texas A & M University-Commerce)

ABSTRACT Upper Paleolithic archaeologists often describe settlement locations or intersite relationships within a geographical area. Such studies frequently assume representativeness of site locations and attempt to reconstruct an articulated settlement system from archaeological patterning. In contrast, regional analysis, by researching structural organization rather than hi toric events, provides a rigorous theoretical framework for explanation complementing site archaeology. Equifinal assumptions and tautologies concerning past group boundaries, survey area size, and hunter-gatherer ranging behavior are resolved within a regional approach . Methodologically, landscape archaeology facilitates inclusion of small sites, plowzone assembl ace , and urface collections through techniques such as adopting the artifact as unit of measure. Data from the Upper Paleolithic of Portuguese Estremadura demonstrate significant advantages to contextualizing site information through regional analysis. In addition to detailing site formation processes and developing models of Late Pleistocene hunter-gatherer land use, a landscape perspective clarifies prehistoric lithic organization and risk management behaviors. Direct relationships between linear distance to a raw material source and mobility are falsified, as the organization of archaeological deposits and assemblages, even in the more problematic Late Pleistocene record, proves critical for explaining the Upper Paleolithic of Portugal. INTRODUCTION Despite the publication of several significant symposia on regional perspectives of the Paleolithic during the last decade (Soffer 1987; Soffer and Gamble 1990), little has changed in the way many researchers are conceptualizing space and approaching archaeology in the Old World. With a few exceptions, the primary result of this intense discussion on research design has been an increase in the number of articles describ ing the difficulties of operationalizing a regional approach and the inclusion of a token paragraph or two on regional patterning or land use in most site reports/monographs. Even studies that explicitly promote regionalism in their research design rarely include a strong theoretical framework for interpretation, instead choosing to describe settlement patterns within a geographical region. This chapter reports on research into Upper Paleolithic land use and settlement of Portuguese Estremadura using a regional approach. While the survey and excavation project developed models of hunter-gatherer cultural 'adaptations in Late Pleistocene Portugal, this paper focuses on specific results that would have been impossibly mired in equifinality or stochastic variation without a regional perspective . The Rio Maior survey and excavation results detail the theoretical and methodological advantages of a regional approach for explaining past human systems. Perhaps most significantly, the Portuguese database is useful for explaining the differences between settlement pattern studies and regional archaeology. REGIONAL APPROACHES IN ARCHAEOLOGY: RESOLVING OBJECTIVES AND OVERCOMING OBSTACLES U ing settlement patterns to infer past human behaviors

began quite early in American anthropological archaeology, perhaps stemming from Morgan 's (188 I) studie of Native American households . While Mindeleff (1900) and Steward's (1938) research in the American southwest are examples of seminal settlement pattern research, it was Willey 's (1953) Viru Valley project that established widely adopted methodologies for inferring past behavior and process from regional patterning. By 1955, archaeologists were discussing problems of vague definitions and theoretical foundations of regional settlement studies at seminars hosted by the Society for American Archaeology (Parsons 1972). Research designs promoted by Binford ( 1964, 1965) in the mid-1960s challenged existing settlement pattern methodologies in archaeology (Parsons 1972). Regional approaches were essential to Binford (1964:44 ) because " .. .it is the structure of archaeological remains that informs about the cultural system ..." Invigorated by this debate, many archaeologists pursued settlement archaeology and settlement system research in the 1970s, but, even in hindsight, it is difficult to untangle which projects were embracing a theoretical framework beyond that established decades earlier. In general , theoretical foundations and methodological choices were often inadequate for yielding insight beyond what became termed "site archaeology." Within hunter-gatherer archaeology, regional approaches to fieldwork suffered (perhaps suffer) from a perception that despite increased labor and monetary expense (Read 1975; Schiffer 1987; Ebert 1992), rarely was understanding of prehistoric cultural systems significantly enhanced. With a few important exceptions, archaeologists were failing to find the neatly articulated prehistoric settlement systems they sought in the archaeological record . The confusion surrounding interpretation and reconstruction of settlement systems in archaeology, coupled with the

THACKE R

0

KEY: UpperPaleolithic Undetermined SeasonalWater



PermanentWater ~

N

Scale:

1 kilometer

Figure 3.1-Map of Rio Major vicinity indicating find location of Upper Paleolithic artifacts or assemblages which were unattributable to specific chronological periods .

26

THE RELEVANCE OF REGIONAL ANALYSIS FOR UPPER PALEOLITHIC ARCHAEOLOGY

0

t N

KEY: Gravettian Workshop Gravettian Residence Seasonal Water

o •

Permanent Water ~ Scale:

1 kilometer

Figure 3.2-Map detailing Gravettian (24- 19 kya) site locations.

27

THACKER

discipline's increased emphasis on non-cultural site formation processes, led to prolific critical commentary on regional archaeology in the 1980 . This critical environment was evident in European Upper Paleolithic archaeology, in the spirit of which Rigaud and Simek ( 1987) commented that "our arms are too short to box with God." White's (1985) innovative study of land use in the Perigord is important in this context, as he demonstrated significant behavioral patterning on a regional scale that withstood the scrutiny of visibility or preservation-bias critiques. The less-than-receptive atmosphere facing regional Upper Paleolithic studies in the 1980s was not without benefit ; today, most archaeologi ts assess and control for formation processes prior to behavioral interpretation. More recently , several studies have assaulted the site concept in archaeology (e.g. , Foley 1981; Schofield 1991; Dunnell 1992; Ebert 1992). This challenge takes many forms, as evidenced in the host of names for alternative approaches to spatial variability in the archaeological record : non-site archaeology, off-site archaeology, landscape archaeology, distributional archaeology, site-formation-process archaeology, etc. Despite incorporating claims of a new theoretical position, these "new" archaeologies are a logical continuation of the formation process preoccupation of the 1980s. Proponents of these methodologies are correct in that an archaeologist must control for how the static pattern of artifact distributions was created by the interaction of numerous dynamic systems. Methodologically, artifact occurrence provides greater spatial resolution and less initial interpretive bias than site location during survey, and the Rio Maior survey incorporated these techniques. Humans do organize space, however, and through exploration of variability, sites were identified in the Rio Maior, Ribeira da Pa, and Penegral river valleys in central Portugal. While these caveats and methodologies have influenced how some archaeologists record spatial variability, the "site-less" challenge's overall theoretical contribution to regional archaeology has been less significant and largely falls within a critical framework that previously existed. Regional Approaches versus Settlement Archaeology: The Importance of Organization

Once formation processes are understood , regional archaeology explicitly assumes "the cumulative character of an archaeological landscape carries diagnostic information about the organizational properties of the systems of origin" (Binford 1987:19, italics added for emphasis). That is, understanding specific historical events within a settlement system is not a goal of analysis, nor is it very likely possible. This emphasis on organization rather than events is a fundamental, although often overlooked, difference between a regional approach and many traditional site and settlement analyses in Europe. Many Upper Paleolithic archaeologists emphasize reconstruction of events from large in situ activity areas. The goal of this event-based research was recently described as an "ethnography of the extinct" (Conard 1994:282). This objective, whether implicit or explicit, is a primary reason so few Upper Paleolithic regional projects have succeeded in building new models or theory. Regional analysis should not search for, or develop, prehistoric ethnology. Rather than focusing on specific individual behavior, or events, a regional analysis seeks organizational structure and strategie of systems. For example, the refitting of a core is only a "just-so story"-trivial, unless contextualized within the variability present in a hunter-g atherer system. More significantly, as will be discussed below, reconstruction of past ranges and territories, while occassionally possible, is not a goal of a regional approach to past organization. The three objectives of the Portuguese Upper Paleolithic regional approach were: 1) to identify and assess patterning invisible from the site level using a framework that allows shifting of spatial scales; 2) to contextualize and explain site data in ways impossible without information on spatial variability; and 3) to build models and hypotheses of past cultural systems that are useful not only for inter-regional comparison, but also for increasing anthropological understanding of hunter-gatherers . The Rio Maior survey project demonstrates that a regional approach yields results that site archaeology cannot and produces models with anthropological relevance beyond reconstructing culture history. SHIFTING SCALES OF SPACE: PHYSICAL, BIOLOGICAL, AND HUMAN LANDSCAPES

It is appropriate to review what regional approaches intend to accomplish. As with many theoretical positions , it is easiest to begin with what regional archaeology is not. Regional archaeology is not simply a collection of observations about patterning of/at more than one site. Further, regional archaeology is much more than analysis of settlement patterns . Details about site location, distances to resources, or catchment analysis, even when applied to a collection of sites, are only descriptive exercises. Such description is no different than a comprehensive site analysis from independent sites and cannot be considered regional archaeology (Isaac 1981). Minimal additional information is gained through this technique, and failure to distinguish descriptive techniques from theory building has contributed to archaeologists' frustration with doing regional archaeology.

Some observations concerning the nature of open-air Upper Paleolithic sites in the Rio Maior vicinity are appropriate prior to discussing patterns recognized from a regional perspective. This paper will only include the Gravettian, Early Magdalenian, and Late Magdalenian periods, as only a few Aurignacian sites and no Solutrean sites were located in the survey area. Within Portugal, the Rio Maior area contains the largest known density of sites and the best documented Upper Paleolithic sequence, in part because of a long history of research (Heleno 1956), but also because of Marks and Zilhao's (Marks et al. 1994) recent extensive excavation and dating project. The Portuguese Gravettian dates to between

28

THE R ELEVANCE OF R EGIONAL ANA LYSIS FOR UPPER PALEOLITHIC AR CHAEOLOGY

0



""' ~,

V'-..

i ~/ / ~

KEY: Early Magdalenian (Small) a Early Magdalenian (Large) El Seasonal Water Permanent Water ~

N

Scale:

Figure 3.3-Map

1 kilometer

detailing Early Magdalenian (16.5-14 kya) site locations .

29

~

THACKER

0

/=

7 ,,>

~ ~_)

t N

Magdalenian (Small) 11 Magdalenian (Large) ■ Seasonal Water Permanent Water ~ Scale:

Figure 3.4-Map

detailing Magdalenian ( 12.5- 1O kya) site locations .

30

~-~ ~

THE RELEVANCE OF REGIONAL ANALYSIS FOR UPPER PALEOLITHIC ARCHAEOLOGY

26,000 and 19,000 B.P.; the Early Magdalenian from 16,500 to 14,000 B.P.; while the Late Magdalenian dates from 12,500 to just after 10,000 B.P. The apparent hiatus between Early and Late Magdalenian is not a function of geological destruction or visibility of sites and appears to correspond to an actual population absence. While there is ome evidence for technological continuity between Early and Late Magdalenian industries, there is no evidence for links between Gravettian organization and later Magdalenian techniques or strategies. Additionally, the setting of Gravettian adaptations in a different climatic trend, deterioration to the glacial, and the Magdalenian, leading out of the Pleistocene, makes direct comparison misleading (Wobst 1990). Artifact occurrences in the valleys were limited to stone tools. Organic preservation is very poor in the well drained sands, and faunal assemblages are unknown from Late Pleistocene openair sites in Portugal. Several cave excavations by Zilhao have yielded faunal assemblages from some valleys within Estremadura, indicating, in those instances, a very general human diet (species choice) compared to the classically studied Late Pleistocene of Cantabrian Spain and southern France (Zilhao 1995). No large migratory herd species are common, while rabbit and deer increase in prevalence throughout the Late Upper Paleolithic. Horse and ibex are present at the glacial maximum but decline rapidly in paleontological assemblages dating to the Magdalenian. No mobilary art has been recovered from these cave sites, and Upper Paleolithic art is very scarce in central Portugal. Overall, the cultural remains of Late Pleistocene hunter-gatherers in Estremadura lack much of what Wobst labeled the "periglacial behavioral package " (Wobst 1990).

Chronology, Surface Collections, and Small Assemblages Technologically, Portuguese Upper Paleolithic stone tool industries are very different from those recovered in France and most of Spain; the closest p_arallels are the Mediterranean industries at such sites as Parpall6 in Valencia (Aura Tortosa 1995). For instance, one distinctive feature of these southern Iberian industries is the absence of a large blade industry during the Magdalenian. Significant technological changes coupled with typological trends identified by Marks and Zilhao (using assemblages with absolute dates) facilitated a decision model of assemblage chronological indicators (Thacker 1996a). The decision model was designed to assign chronological periods to small or surface assemblages, as well as to assess degree of mixing and other contextual observations. Indicators of later prehistoric periods, of course, had to be included, along with diagnostic indicators of gunflint production sites near Azinheira.

best, not to mention the hundreds of isolated finds, such as large flakes or retouched blades, that could have been produced anytime from I 00,000 to l 00 years ago. Surprisingly few cases of badly mixed assemblages from more than one period were found, but these assemblages were excluded from this paper's analysis . This type of decision model would be much more difficult to construct for areas with more similar technological industries, such as Cantabria or the Paris basin . The patterning of finds in Figure 3.1 strengthens land-use patterns and strategies modeled using site assemblage data (Figures 3.2-3.4). All Upper Paleolithic undetermined finds are subsumed within site-based models of Late Pleistocene behaviors. These artifacts/assemblages do not occur in unusual locations and are not outliers in terms of technological or raw material attributes . A parsimonious interpretation of evidence from Figure 3.1 reinforces assemblage-based models, but in no case can the finds be used to falsify these models without chronological control. For example, there are six undetermined assemblages in the lower valleys that could be either Gravettian or Magdalenian. The site locational characteristics fit the Gravettian pattern, but not the Magdalenian one. The technological attributes and types found at these locations all occur in other Gravettian sites in the lower section of the valley (Figure 3.2). Thus, the parsimonious explanation is that these undetermined sites are probably Gravettian, although this evidence is not secure enough to include these sites and assemblages in the Gravettian group. Furthermore, if these few sites are actually Magdalenian, little new is learned. The lower valleys were certainly used by Magdalenian groups , but the land-use model contends that significant artifact assemblages were not produced in those areas. Magdalenian hunter - gatherers procured cobbles of flint in the lower valleys and transported them to campsites in the middle valley sections without much decertification (Thacker 1996b). Thus, the scenario of a knapper producing a few artifacts in the lower valley during the Magdalenian is expected, and the undetermined sites again do not contradict or falsify the model.

Regional Analysis and the Assumption of Articulated Sites A complaint commonly lodged against settlement archaeology, particularly of the Pleistocene record, is that chronological resolution is too coarse-grained for discerning settlement systems. For example, Conkey and Rigaud and Simek (see chapters in Soffer 1987) both caution against interpreting settlement patterns as if they represent the cultural system. Implicit in this critique is the assumption that settlement archaeology seeks articulated sets of sites (Isaac 1981). Reconstructions of regionally defined ethnic-group boundaries and ranging behaviors discussed by many regional archaeologists are deserving of this censure. Regional analysis seeking organizational properties, however, does not treat spatial patterning as an articulated system.

From a practical standpoint, the model was both a success and a failure. No assemblage with under 250 pieces was able to be attributed to any period, and, in most cases, well over 500 pieces and 40 tools were required. Figure 3.1 is a map of some small assemblages that were clearly Upper Paleolithic but could date to any subperiod. Integrating this data without chronological control into a regional framework is difficult at

Cases where the archaeological record preserves strong evidence for large-scale cultural divisions of space, such as 31

THACKER

historical group ranges, ethnicity patterns, or group boundaries/ territories are, and should be expected to be, unusual rather than commonplace. The remnant settlement patterns that comprise the archaeological record may only rarely reflect the dynamic and con tantly renegotiated nature of hunter-gatherer territoriality or ranging (Dewar and McBride 1992; Gamble and Boismier 1991). Also, linking arguments connecting static archaeological remains to meaningful cultural units are often weak. For example, as mentioned above, the assemblages recovered from the Rio Maior vicinity are comprised of stone tools. Few convincing arguments clearly linking technological or typological attributes of stone tools to single bands, ethnicity, or culture exist. Lithic variability may have structure and meaning beyond functional-use properties, but the level of this explanation may not be ethnographic. In contrast to chipped stone, artifacts or behavioral evidence more heavily loaded with cultural meaning may provide increa ed resolution of historical events and intergroup dynamics.

space containing sites. These sites may have been produced by separate groups that had no interaction with each other, or it is possible that not one of these sites, despite functional differences, was occupied within 100 years of one another. The lack of artifact variability supporting a strong argument for past cultural divisions of space is not an obstacle hindering model development if research focuses on organizational structure rather than questions of prehistoric territoriality.

Representative Landscape Patterning: An Essential Starting Point Understanding prehistoric land use requires an awareness of the repre entativenes of archaeological patterning. Three main dimen ions of archaeological variability must be examined: the physical landscape, including geology, climate, and topography; the biological landscape, including paleoenvironmental reconstructions of floral and fauna) communities; and the human landscape, including ite and artifact occurrences in social contexts. Variability of human land use in the Rio Maior, Pa, and Penegral valley does not reveal ethnographic ranges; rather, it distinguishes changing organization of settlement and artifact assemblage .

Granting the improbability of recovering archaeological signature to directly discern past ranging behavior, regional analysis attempts to recover organizational structure across a known landscape. The explanatory value of these structural models contrasts with the limited and equifinal explanations of past human behavior developed through a more traditional settlement approach.

The physical landscape of the surveyed valleys has changed little since the Middle Pleistocene. No major valley-wide erosional or depositional events have severely influenced site preservation. Lowest geological visibility of Late Pleistocene landform surfaces occurs in the upper sections of the Ribeira da Pa, where slopewash from the apron slopes of the limestone highlands of the Serra Dos Candeeirros buried or destroyed sites. Little can be argued concerning land use of these areas, as there is an absence of evidence.

Hunter-Gatherer Ranges and Survey Areas In order to identify group-specific ranging behavior and boundaries using traditional settlement approaches, archaeologists must recover articulated sets of sites within a spatial area encompassing an entire hunter-gatherer round. The total coverage survey area of the Rio Maior project is currently only about 110 square kilometers, between fifty and a thousand times smaller than some ethnographically known hunter-gatherer ranges (Wobst 1976; Kelly 1995). According to some ethnographic examples, all of Iberia may not be large enough to be considered a meaningful ethnographic region (Brown 1987; Jochim 1987). Even if marginality arguments are accepted and ranges in temperate environments were smaller than historically known behaviors (Butzer 1988), the Portuguese survey area would still be smaller than any reasonable range. It is extraordinarily unlikely that the survey area fortuitously corresponded to past cultural divisions of space. Further, such an approach assumes that Upper Paleolithic hunter-gatherer ranges were rather static over thousands of years. A more reasonable assumption, until proven incorrect, is that the Late Pleistocene archaeological record in central Portugal represents a complex palimpsest of shifting ranges and boundaries.

The middle- and lower-valley sections, roughly below 125 meters elevation (asl), have experienced only local slopewash and stream cut and fill/overbanking processes. Most landforms in the lower section of the valleys have been continuously exposed since the Upper Paleolithic, remaining unchanged except for rather minor slopewash events. These local and gradual effects result in an Early Upper Paleolithic visibility that is greatest in the lower valley sections but is present to some degree in the middle valley. This vi ibility bias is evident in the patterning of Gravettian sites shown in Figure 3.2, as sites cluster in the lower valleys, with a few sites located in the middle sections. In contrast, the Late Upper Paleolithic is well represented in the middle-valley sections, with six Early Magdalenian (Figure 3.3), and 20 Late Magdalenian sites (Figure 3.4). Significantly, no Late Upper Paleolithic sites are located in the lower-valley sections. As no major erosional events occurred, and as Epipaleolithic sites have been found in the lower sections, sometimes on the same landform as Gravettian sites, this pattern is evidence of absence (Thacker 1996a). Magdalenian groups did not produce recognizable sites or concentrations of artifacts in the lower valleys.

From a regional or landscape perspective, the size of the survey area and the scale of past ranging behaviors are minimally important. Arguments linking various site types into a short, time-exposed snapshot of a subsistence round are avoided, as prehistoric ethnography is not a goal. Analytically, the Portuguese survey area only represents a window of

Lithic raw material in the form of quartz and quartzite cobbles occurs throughout the valleys in Middle Pleistocene gravels, 32

THE RELEVANCE OF REGIONAL ANALYSIS FOR UPPER PALEOLITHIC ARCHAEOLOGY

and high quality flint cobbles are located in sands on the ridge separating the Rio Maior and Penegral drainages. The distribution of the flint deposits was mapped during survey and is displayed in Figures 3.1-3.4 . Paleoenvironmental data and reconstructions place Estremadura in a transition zone between cold northern Iberia and nearly Mediterranean climates of southern Portugal, even during the glacial maximum (Straus 1992). The climatic impact of the glacial peak was probably minimal in the Rio Maior region, as the lower-valley sections sheltered arboreal and temperate species throughout the Late Pleistocene. Warm events result in an expansion of these lower-valley biotic communities into the middle and upper valleys, as confirmed by charcoal analysis from several Magdalenian hearths at CPM (Figueira} 1991). The limestone uplands were probably open throughout the Late Pleistocene, with grasses and low cover vegetation prevalent (Zilhao 1995). These reconstructions indicate that central Estremadura was extremely patchy in bioresource terms (Johnson 1977).

Figure 3.5 illustrates that sites in both the Early and Late Upper Paleolithic occur in a range of elevations but cluster between 75 and 99 meters above sea level. The absence of sites above 150 meters is a combination of the poor visibility of Late Pleistocene landforms in the upper valley sections and the lack of occupation on the limestone uplands. Another interesting conclusion from this examination of spatial variables is that in situ assemblages are found more frequently on certain landforms. For example, while a significant number of Upper Paleolithic sites occur on low terraces, modern land use in the middle valley, principally plowing for vine-yards or planting olive trees, has disturbed most sites at these locations. If plowed sites and assemblages are not included in a regional approach to the Upper Paleolithic of Estremadura, human activity taking place on entire landforms and within corresponding biological communities would be omitted . Regional analysis, thus, does not simply examine the location of sites. The physical, biological, and human background is analyzed within a window of space in order to solve problems of equifinality in ite patterning. In many instances, only this detailed and time-consuming research design can segregate patterns created by human behavior and those patterns resulting from non-cultural formation processes.

Sampling of space is critical to regional analysis (Read 1975; Thomas 1975), and site patterning across space (both intersite and intrasite) must be proven as representative of huntergatherer behavior and not purely a result of geological or other visibility biase . First, the possibility that formation processes other than past human spatial organization created site assemblages was falsified using Portuguese artifact concentrations. Scatter areas, and tool and core frequencies across space, consistently vary from the random structure expected if artifacts were lost or discarded through either undifferentiated or even consistent use of geomorphologic landforms (Thacker 1996b ). Consequently, site assemblages in the study do not match the configurations expected if artifact concentrations represent stochastic variation and clustering of artifacts displaying a background noise pattern such as those discussed in Schofield ( 1991), Ebert ( 1992), or Rossignol and Wandsnider ( 1992).

1.0r----------------------

D

Background

0 EUPAII

0.8



LUP All

0 .6 0.4

0.2 0.0 ., u >

:J

c.>

i f

A

'§ >,

u

0..

~

~

• • • • • • • f!

~

·5 !-

"'

...J

.c u

Cl)

·5

j,!

~

~

v;

"' E

.s:i 0

u

>


II)

u

...J

0

0::

Figure 4.7-Figure

01)

showing size ranking and the multipurpose nature of sites, summarised in Table 4.3.

55

u

.c

e u

:::,

t)I) C:

0

...J

0C

t)I)

C:

·~

-~ -~ u u

u

e

!

• •

A

i::

:,

.:

u

~

>

9

.:,/.

]

~

~

"' ~0:, j,! "' ·o. -e

"E

:i

%



uC: 2

"'

0..

Crl

cu

...J

u

i:

a::i -0

u 0

c::

0

LI..

.A.

&

= u=

.!:!

~ :,

"'

LI.

"E ~

:i 0

:i

.::.: 9 .:,/.

]

..

ij

f! 0

"'

...J

0

E ~

.:!

~

BOYLE

as Peyrat, 7800 m away from the Vezere and possible herd migration routes (Gordon 1988; White 1985: 189), represent neither large migration hunting sites nor small look-out posts. Perhaps they served a totally different purpose. Abundant art found at these sites (e.g., La Mouthe, Les Combarelles) points at least partially to a non-subsistence function.

CONCLUSIONS During the Final Magdalenian, reindeer hunting is often assumed to have dominated subsistence. Instead, multiple, rather than single, resource exploitation provided the subsistence diversity necessary for survival. By systematically exploiting alternative, additional resources, an element of variety, as well as ecurity, is maintained in the diet.

Small site size ranges from 45 m2 (Jolivet) to 120 m2 (the latter Villepin and Roe de Barbeau) , while only Faurelie II and Villepin, the latter perhaps part La Madeleine, lie at low altitudes. Font Brunel, Le Pouzet, and Jolivet occur at wide points in the river valley ( ee Table 4.2), the regional/period mean being just over 668 m, while only La Greze and Faurelie IT lie at a greater than average distance from the nearest ford . Variable site-orientation data lends support to the suggestion that these ites were used only briefly and for specific purpo es; natural retention of warmth appears not to have been a major concern. Sites facing WNW and NNW ( uch as Jardel II and Jolivet) do not provide the degree of protection from winds and low temperatures that those facing outh or east are known to do. A single hearth provided sufficient heat for a short stay, and the view from the site was perhaps of greater importance.

While the major species were taken in bulk at river crossings and other such locations, secondary specie were taken on a short-term basis similar to "e ncounter hunting," a strategy which Binford ( 1978:265) describes as used when prey is not abundant. Longer term, planned hunting of the large secondary taxa (bovids, hor e, etc.) may also have occurred, especially when reindeer was either scarce or undesirable . The carcity of relevant seasonality data is, at this point, a problem . It becomes clear that, wherever reindeer dominates, its over-abundance was, in all probability, balanced by exploitation of something else. The range of possible alternative resources is often large and is not confined to large herbivores, although these did

Table 4.2 Final Magdalenian Site Data Including Topography and Location

Area

Rank

Altitude

River

Valley Width

RSR Size

Water

Dry Valley

Lithics

No. Taxa

Major Taxon

Laugerie-Basse

6000

I

15

50

425

6000

0

50

7716

5

reindeer

La Madeleine

3000

2

5

0

500

3000

0

100

5600

5

reindeer

Les Eyzies

2000

3

35

125

500

2000

100

700

Bout du Monde

1500

4

5

150

400

1500

150

750

Chez -Galou

750

5

10

50

550

1200

50

0 0

% Major Taxon

2nd Taxon

% 2nd Taxon

97 .7

bovids

1.84

reindeer reindeer

* * * *

>200

Abzac

680

6

40

250

125

1000

40

La Mouthe

550

7

95

875

50

875

500

0

6

reindeer

Limeuil

460

8

10

25

2250

667

25

25

4774

7

reindeer

91.37

hor e

7.64

Cap Blanc

450

9

20

6500

250

600

150

75

405

2

reindeer

93.85

horse

4.2

Rocher de la Peine

400

10

10

200

600

522

100

0

Liveyre

400

II

20

75

500

522

50

25

124

3

Peyrat

350

12

30

7800

300

462

0

0

La Tuiliere

300

13

35

250

500

429

250

150

Cazelle

270

14

2

5000

200

400

50

0

reindeer

93.83

bovids

4.35

Les Combarelles

251

15

20

2500

50

375

25

0

Valojouix

250

16

5

2250

50

353

100

0

Peyrille

200

17

30

600

300

333

100

100

Longueroche

200

18

2

1500

300

316

10

25

Richard

185

19

30

600

300

300

100

100

Yillepin

120

20

5

10

500

279

10

50

Roe de Barbeau

120

21

35

800

50

279

50

0

Le Pouzet

100

22

75

4001

3000

255

100

0

Font-Brunel

100

23

70

75

2250

255

25

0

Fauretie II

96

24

5

6000

200

240

25

0

Jardel II

60

25

35

150

650

231

150

250

La Greze

50

26

10

6250

250

222

125

100

Jolivet

45

27

45

500

3000

214

20

0

56

Art

* * *

* 3

* 117 295 278

4

reindeer

6

reindeer

3

horse

*

reindeer

red deer

*

* *

3 110 2 453

3

reindeer

red deer

626

2

reindeer

hor e

THE LATE MAGDALENIAN

probably form the major portion of the resource spectrum. Their potential is significant given that some species provide quantities of fat, protein, and, in the case of horse, sugars at a time when reindeer meat is poor (Ducommun 1982:14). Thus, intentional reindeer culling formed just one part of the Upper Magdalenian subsistence economy, an economy practised at sites of variable size and importance and which continued until temperate species such as red deer, roe deer, and boar became important elements in both environment and subsistence (Boyle 1990, 1993, 1996). Reindeer was usually abundant, and its appearance was probably sufficiently predictable to preclude the need to process all carcases intensively, especially if the potential yield from each carcase was not great because of seasonal variation in physiological condition. Best parts could be selected. Reindeer is a low-fat resource at the best of times (Stern et al. 1980: 127), and, during the late winter and spring when the site was occupied, many were likely to be in poor condition. Late winter and early spring are the most stressful time of year for all species, including man, and, in the case of the male reindeer, condition may be very poor indeed. Any animal in poor condition (i.e., at its leanest) is likely to be rejected by hunters (Speth 1982:146-148; Jochim 1981:81-90), who may therefore: a) select only those carcase parts in better condition; (b) practise grease extraction; or c) concentrate on females in calf, which are generally in better condition. The low percentage of reindeer bones which bear processing marks suggests that many reindeer were either abandoned or only partly processed. Species abundance probably warranted at least partial discard when animals were in poor condition. The presence of stiff carcase processing marks, within the context of the relatively low percentage of other marks on reindeer bones, lends support to the argument that more reindeer were killed than could be immediately used. Although the frequency of filleting and skinning marks implies that sufficient reindeer was available to warrant selective processing, it is unlikely to have provided everything that was required by the hunting community. Thus, while selective reindeer processing appears to be important, other resources were processed more intensively in order to augment the resource base. The secondary species provided variety in the diet and a way in which to supplement the existing subsistence base. Just as the Rank Size Rule predicts, we see more smaller Magdalenian VI sites than larger ones in the Vezere valley. In general, these sites yield fewer large herbivore species, and, where quantitative data are available, the assemblages are specialised, i.e., heavily dominated by a single species, usually reindeer. Neither a strict primate pattern, in which the largest site is too large or others are too small (Johnson 1981: 150), nor convex patterning, in which some of the smaller sites are too large for their rank values, may be proposed. Instead, a concave pattern is seen, with site size divided into three categories, implying a degree of organisation not apparent from a simple distribution map. Each of the medium and small sites may have been relatively activity specific, irrespective of whether that activity was 57

IN THE VEZERE VALLEY

social or economic. Medium-sized sites such as Les Combarelles yield parietal art, a possible indication of nonhunting behaviour (see Table 4.2), whether or not that art is of Magdalenian VI date. This is a characteristic that distinguishes the Vezere valley from the Altamira region of northern Spain considered by Conkey (1980). The large quantities of both portable and parietal art seen at the aggregation site of Altamira are not seen in the V ezere Valley. Instead, parietal art is usually found away from the aggregation sites, often in or in close proximity to the smaller tributary valleys. Such sites, removed from the main thoroughfare, may have been deemed to have been special by the hunters and may have been used for symbolic, ideology-related purposes, such as huntinginitiation rites. This might explain the small cultural and faunal assemblages often recovered; long-term subsistence was not of primary concern here. At other non-art sites, there is evidence of more intensive hunting and carcase processing (Boyle 1990), while, at the smallest sites in the region where short-term occupation is suggested, faunal assemblages are often small. Despite their size, the function of these small sites in the annual subsistence system was critical. They perhaps represent the primary economic unit or territoirefamilier recently proposed by David et al. (1994) for the final Magdalenian in eastern France. The largest sites provide greater evidence of a mixture of the subsistence activities and social dimensions of life. Reindeer Table 4.3 Site Type and Characteristics during the Final Magdalenian SITE CATEGORY Small

CHARACTERISTICS Single purpose Location governed by view and/or ease of access Small lithic/faunal assemblages Presence of art ("to pas the time")

Look out posts Hunting stands

Medium

A small number of activities, one major A few associated minor activities? Social activity reflected by art

Carcaseprocessing sites Ritual sites

Large

Multiple activities Hunting - Sites close to kill Processing - Disarticulation, etc. to remove meat Marrow and grease rendering Consumption Some (limited) art Non-subsistence activities : -exchange -social network maintenance, etc .

Aggregationsites Large-scale, multi-purposebase camps

BOYLE

abundance and the associated butchery data point to largescale killing of herds at or close to the site (Boyle 1994). Presence of mobiliary art points to some non-subsistence functions, indicating perhaps that additional non-economic activities occurred there as well. The existence of large tool assemblages of stone and bone, however, confirms the economic role of the site in que tion . The multipurpose nature of these sites is indicated in Figure 4.7 and summarised in Table 4.3. Here potential function i indicated, "attached" to the site. Whether Laugerie-Basse, La Madeleine, Chateau des Eyzies, and Bout-du-Monde were occupied a few times by a relatively large group of hunters and their dependents or by fewer people many more times, the importance of these sites cannot be denied. Whether they congregated only for sub istence-related purposes in order to gain maximum shared yield from migrating herds of reindeer or for social reasons as well, including , for example , information exchange and network extension/maintenance , is only speculation at this time , but the presence of works of art implie s more than a single purpose. The large, mixed assemblage give the impression of a general purpose nature for the sites . This i more a reflection of combined functions than singlepurpose activity at a site which may have been of importance to all Magdalenian VI hunters in the region.

Boyle, K.V., 1990. Upper Palaeolithic Faunas from Southwest France: A Zoogeographic Perspective. British Archaeological Reports International Series 557. Oxford: British Archaeological Reports . Boyle , K.V., 1993. Upper Palaeolithic procurement and processing strategies in Southwest France. In Hunting and Animal Exploitation in the Later Palaeolithic and Mesolithic of Eurasia, edited by G.L. Peterkin, H.M. Bricker, and P. Mellars, pp. 151-162. Archaeological Papers of the American Anthropological Association 4. Washington, D.C.: American Anthropological Association . Boyle, K.V., 1994a. La Madeleine (Tursac, Dordogne): Une etude paleoeconomique du Paleolithique superieur. Paleo 6:55- 77 . Boyle, K.V., 1994b. Late Magdalenian carcase management trategies. The Perigord data. Paper presented to the 7'h Congress of the International Council for Archaeozoology. September, 1994. Konstanz, Germany. Boyle, K.V., 1996. From Laugerie-Basse to Jolivet: The organization of Final Magdalenian settlement in the Vezere valley. World Archaeology 27:477-491. Brain, C.K., 1981. The Hunters or the Hunted: An Introduction to African Cave Taphonomy. Chicago: University of Chicago Press.

ACKNOWLEDGMENTS Some of the data referred to here was collected while working at the Musee National de Prehistoire in Les Eyzies. I should like to thank M. J.-J. Cleyet-Merle for allowing me to examine material from La Madeleine and Limeuil. The French Embassy (London) generously provided financial backing for my second, and longer, visit to Les Eyzies.

Capitan, L., and J. Bouyssonie, 1924. Limeuil: Son Gisement a Gravures sur Pierre de l 'Age de Pierre. Paris: Nourry. Conkey, M.W. , 1980. The identification of prehistoric hunter-gatherer aggregation sites: The case of Altamira. Current Anthropology 21 :609-630.

REFERENCES

David, S., F. Seara, and A. Thevenin, 1994. Territoires magdaleniens: Occupation et exploitation de l' espace a la fin du paleolithique superieur dans l'Est de la France . L 'Anthropologie 98: 666-673 .

Binford, L.R., 1977. Forty-seven trips: A case study in the character of archaeological formation processes. In Stone Tools as Cultural Markers, edited by R.V.S. Wright , pp. 24-36 . Canberra: Australian Institute of Aboriginal Studies.

Delibrias, G., M.-T. Guillier, J. Evin, and J. Chevallier, 1986. Sommaire des datations 14C concernant la prehistoire en France III: Dates effectuees de 1979 a fin 1984. Bulletin de la Societe Prehistorique Franr;aise 84: 208-223.

Binford, L.R., 1978. Nunamiut Ethnoarchaeology. New York: Academic Press. Binford, L.R., 1980. Willow smoke and dogs' tails: Hunter-gatherer settlement systems and archaeological site formation. American Antiquity. 45:4-20.

Delpech, F., 1983. Les Faunes du Paleolithique Superieur dans le Sud-Ouest de la France. Cahiers du Quaternaire 6. Paris: Centre National de la Recherche Scientifique.

Binford, L.R., 1981. Bones: Ancient Men and Modem Myths. New York: Academic Press.

Delpech, F., and P. Villa, 1993. Activites de chasse et de boucherie dans la grotte des Eglises. In Exploitation des Animaux Sauvages a travers le Temps, edited by J. Desse and F. Audoin-Rouzeau, pp. 79-102. Juanles-Pins: Associations pour la Promotion et la Diffusion des Connaissances Archeologiques .

Binford, L.R., and J.B. Bertram, 1977. Bones frequencies and attritional processed. In For Theory Building in Archaeology: Essays on Fauna[ Remains, Aquatic Resources, Spatial Analysis and Systematic Modeling, edited by L.R . Binford , pp. 77- 153. New York: Academic Press . 58

THE LATE MAGDALENIAN IN THE VEZERE VALLEY

Ducommun, J.-F., 1982. Gastronomie de la Viande de Cheval. Paris: Federation de la Boucherie Hippophagique de France. Gamble, C.S., 1986. The Palaeolithic Settlement of Europe. Cambridge: Cambridge University Press.

Wilrm recent en Dordogne. Paleo. 1:36-46. Mellars, P.A., 1973. The character of the Middle-Upper Palaeolithic transition in south-west France. In The Explanation of Culture Change: Models in Prehistory, edited by C. Renfrew, pp. 2552-2576. London: Duckworth.

Gordon, B.C., 1988. Of Man and Reindeer Herds in French Magdalenian Prehistory. British Archaeological Reports International Series 390. Oxford: British Archaeological Reports.

Mellars, P.A., 1985. The ecological basis of social complexity in the Upper Palaeolithic of southwestern France. In Prehistoric Hunter-Gatherers: The Emergence of Cultural Complexity, edited by T.D. Price and J.A. Brown, pp. 271-297. London: Academic Press.

Haggett, P., 1972. Geography: A Modern Synthesis. New York: Harper and Row.

Murphey, R., 1982. The Scope of Geography. 3rd edition. New Haggett, P.,A.D. Cliff, and A. Frey, 1977. locational Analysis in Human Geography. Volume/: Locational Models. 2nd edition. London: Edward Arnold.

York:Methuen. Olsen, S.R., 1987. Magdalenian reindeer exploitation at the grotte des Eyzies, SW France. Archaeozoologia l: 171-182.

Hassan, F.A., 1981. Demographic Archaeology. New York: Academic Press.

Regnault, F., 1894. Les haltes de chasse et de peche de l' epoque du renne. Association Franr;aise pour l'Avancement des Sciences, Congres de Caen, pp. 747-752.

Hayden, B., B. Chisholm, and H.P Schwarcz, 1987. Fishing and foraging: Marine resources in the Upper Palaeolithic of France. In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer, pp. 279-291. New York: Plenum Press.

Roussot, A., 1976. Abri du Cap Blanc, commune de Marquay. In Livret-Guide de I 'Excursion A4 Sud-Ouest (Aquitaine et Charente), edited by J.-Ph. Rigaud and B. Vandermeersch, pp. 85-88. Nice: ixemeCongres de l'Union Internationales des Sciences Prehistoriques et Protohistorique.

Hodder, I., and Orton, C., 1976. Spatial Analysis in Archaeology. Cambridge: Cambridge University Press.

Small, J., and M. Witherick, 1989. A Modem Dictionary of Geography. 2nd Edition. London: Edward Arnold.

Jochim, M.A., 1981 Strategies for Survival: Cultural Behavior in an Ecological Context. London: Academic Press.

Speth, J.D., l 983. Bison Kills and Bone Counts: Decision Making by Ancient Hunters. Chicago: Chicago University Press.

Johnson, G.A., 1977. Aspects of regional analysis in archaeology. Annual Review of Anthropology 6: 479-508. Johnson, G .A., 1981. Monitoring complex system integration and boundary phenomena with settlement size data. In Archaeological Approaches to the Study of Complexity, edited by S.E. van der Leeuw, pp. 144-188. CINGVLA VI. Amsterdam: Albert Egges van Giffen Institut voor Prae- en Protohistorie,

Stern, R.O., E.L. Arobio, L.L. Naylor, and W.C. Thomas, 1980. Eskimos, Reindeer and land. Fairbanks: Agricultural Experiment Station, University of Alaska.

Kelly, R.L. 1995. The Foraging Spectrum: Diversity in Hunter-Gatherer Lifeways. Washington, D.C.: Smithsonian Institution Press.

White, R., 1985. Upper Paleolithic Land Use in the Perigord: A Topographic Approach to Subsistence and Settlement. British Archaeological Reports International Series 253. Oxford: British Archaeological Reports.

Vita-Finzi, C., 1978. Archaeological Sites in their Setting. London: Thames and Hudson.

Lyman, R.L., 1985. Bone frequencies: Differential transport, in situ destruction and the MGUI. Journal of Archaeological Science 12:221-236.

Whitelaw, T.M., 1990. The Social Organisation of Space in Hunter-Gatherer Communities: Some Implications for Social Inference in Archaeology. Ph.D. Dissertation. University of Cambridge.

Lyman, R.L., 1994. Vertebrate Taphonomy. Cambridge: Cambridge University Press.

Zipf, G.K., 1949. Human Behavior and the Principle of Least Effort: An Introduction to Human Ecology. Cambridge: Cambridge University Press.

a

Madelaine, S., 1989. Contribution des anciennes fouilles la connaissance des ongules et de leur milieu durant le

59

5 Late Upper Paleolithic Environments, Subsistence, and Zoogeography in Cantabrian Spain James T. Po kines (Field Museum Department of Mammals and U. S. Army Central Identification Laboratory) ABSTRACT The question of what constitutes a distinct region has become increasingly important to archaeologists as analytical frameworks have expanded beyond site -specific analyses. Useful parameters include physical geography and areal distributions of artifact types, artwork styles , and settlement hierarchies. Since the ecological systems of hunter-gatherers are based upon the harvest of wild species, defining regions biogeographically places a logical emphasis upon food and raw material (furs , fibers , etc.) availability as a basis for regional definition. This analytical framework necessarily emphasizes faunal over floral remains because the latter are preserved and recovered with much lower frequency. The hunter-gatherer groups of Late Upper Paleolithic Cantabrian Spain developed within and as one part of a distinct biogeographic region in Europe, where the cultural systems reflect the unique array of exploitable food species present and juxtaposition of ecological zones. This biographic region is explored primarily through its terrestrial mammalian microfauna, which allow the greatest resolution of ecological parameters , even though these species themselves were rarely human prey during the Upper Paleolithic.

INTRODUCTION Cantabrian Spain is defined politically by the adjoining provinces of Asturias , Cantabria proper (or Santander), Vizcaya , and Guipuzcoa (Figure 5.1). These four provinces encompass a unique geographic area in which to examine Late Pleisto cene regional adaptations. This coastal strip along the Cantabrian Sea is bordered abruptly by the Cantabrian Cordillera, which , within 25 km of the present-day shore, rises as high as 2,600 m, although peaks of 1,000 to 1,500 m and a slightly greater width to the coastal lowlands/foothills are typical. Even during glacial maxima, a lowering of ocean level of 120 m (Bailey 1983; Butzer 1983) would have exposed only

an additional 4 to 12 km wide strip of land (Straus 1992). Glaciation in the Cordillera also would have reduced the habitable upland area, with terminal moraines from the last glacial maximum reaching as low as 550 m (Clark 1981). Hence, in this relatively small and circumscribed region there are the closely proximate environmental zones of littoral, coastal lowland , foothill, and upland, with each of these zones bisected by river valleys. Cantabrian Spain is also defined by its unique biogeographic status. While the remainder of the Iberian Peninsula, excluding the alpine areas, is characterized today by a Mediterranean climate and its associated floral and faunal composition, the

CantabrianSea

France

CantabrianCordillera

50 km 1 2 4 5

Las Caldas Tito Bustillo,La Lloseta La Riera, Coberizas, Balrnori, Cueto de la Mina Altamira El Castillo

6 7

El Juvo, El Pendo Cue\~ Morin

8

Rascaflo

16

3

9

10 11

Ot~o . La Chora

El Mir6n Bolinkoba

12

Lwnentxa

13

Lezctxoo

14 15

Ekain , Urtiaga , Ennittia

Amalda.Eralla Aitzbitarte,Torre

Figure 5.1-Cantabrian Spain, with sites discussed in the text.

POKINES

Table 5.1

Cantabrian region is characterized as part of the Atlantic climatic/biogeographic zone. The ranges of many Palearctic mammal species reach their southernmost limits in this area, while the distributions of many Mediterranean species reach their northernmost limits. The Cantabrian fauna! community also includes many species endemic to the Franco-Iberian region, especially those whose distribution center on the Pyrenees Mountains. The fauna] species available for exploitation limit and shape the hunter-gatherer adaptations possible in this region (Coughenour et al. 1985; Little et al. 1984), including the technology and hunting strategies employed to harvest these resources. The availability of these food species and the manners in which they may be acquired most efficiently also shape other aspects of hunter-gatherer adaptations beyond the direct acts of food acquisition, including settlement patterns, seasonal movement and aggregation, group size, food torage , and task specialization. Hunter gatherer regional adaptations are thus usefully studied where that region is defined biogeographically (Boyle 1990). Cantabrian Spain during the Late Pleistocene presented a unique species mix exploited by humans , which in many ways shaped the local regional hunter -g atherer adaptations of the Late Upper Paleolithic.

Distribution Zones for Cantabrian Microfauna Endemic Galemys pyrenaicus Sorex coronatus Pitymys lusitanicus Pitymys pyrenaicus Arvicola sapidus Atlantic Talpa europaea Sorex minutu s* Neomys fodiens* Sciurus vulgaris* Glis glis* Clethrionomys glareolus * Microtus agrestis* Microtus arvalis* Arvicola terrestris* Apodemus flavicollis* Micromys minutus* Mediterranean Talpa occidentalis Crocidura suaveolens Mus spretus Genetta genetta**

General Erinaceus europaeus Neomys anomalus Crocidura russula Oryctolagus cuniculus Lepus capensis Eliomys quercinus Apodemus sylvaticus Vulpes vulpes Mustela erminea Mustela nivalis Mustela lutreola Mustela putorius Martes martes Marte s foina Meles meles Lutra lutra Fe/is lynx Fe/is silvestris Boreal/1\mdra Microtus oeconomus Gulo gulo Alopex lagopus Microtus gregalis

Alpine Marmota marmota Microtus nivalis

During the period from the last Glacial Maximum c. 22,00018,000 B.P. until the start of the Holocene, the human groups inhabiting Cantabria were also faced with a gradually ameliorating but fluctuating climate . Evidence from sea-core foraminifera indicates that the average North Atlantic surface ocean water temperatures at 42° north latitude were around 10° C colder at 18,000 B.P. than at present, indicating polar water conditions (McIntyre et al. 1976). Air temperatures at glacial maximum for the coast had a February average of around -1.6° C and an August average of 10.4° C (Butzer 1986:215). Correlation with global sea-level fluctuations lies in the waning pha e of Isotope Stage 2.

*Atlantic species whose distributions extend into the Iberian Peninsula only as far as Cantabrian Spain. **A primarily African species whose distribution reaches into the Iberian Peninsula and southern France. roughly with that of modern France and extends into Cantabrian Spain. Its primary association is with humid meadows, but it can be found in most habitats with sufficient cover (Corbet and Ovenden 1982). The southwestern water vole (Arvicola sapidus) is found throughout the Iberian Peninsula and into France along riverine habitats supporting dense cover (Corbet and Ovenden 1982: 166). The presence of any of these three species in an archaeological site indicates humid local climate and proximity to open water. Two other vole species are endemic to this region : Pyrenean pine vole (Pitymys pyrenaicus) and Lusitanian pine vole (Pitymys lusitanicus). The former species is found today only in northeastern Spain and southwestern France, while the latter is found only in the northern and western Iberian Peninsula. P. pyrenaicus is fossorial and herbivorous, preferring deep soils and dense vegetation cover (Borghi et al. 1994). P. lusitanicus is associated with open meadows, agricultural land, and shrubland (Niethammer and Krapp 1982:482) and is plentiful in coastal low land Cantabrian Spain (Po kines 1998c ). All of these species, which today inhabit a temperate regime, were present in Pleistocene levels of Cantabrian Spain, even during severe cold phases, although in reduced proportions compared with cold-associated microfauna.

TERRESTRIAL MAMMALIAN MICROFAUNA The unique biogeography of Cantabrian Spain is demarcated most clearly through its terrestrial mammalian microfauna (Table 5.1). This group is defined by having an average adult body mass less than 5 kg and includes in this region rodents, lagomorphs, insectivores, and small carnivores. Only some of these species were preyed upon directly by humans during the Paleolithic, and most of these are generalists found across a range of habitat zones in the Palearctic. This terrestrial mammalian microfaunal community, by association with multiple habitat types , illustrates the unique environmental character of this region. Several microfaunal species are found only in and around the Cantabrian region (Figure 5.2). These taxa include the Pyrenean desman (Galemys pyrenaicus), an alpine species endemic to the Pyrenees Mountains, the Cantabrian Cordillera, and some montane zones further south in the Iberian Peninsula. This semi-aquatic insectivore is found only in riverine habitats from 300 to 2200 m (Palmeirim and Hoffman 1983). The range of Millet's shrew (Sorex coronatus) corresponds

Many Atlantic habitat zone species have modern ranges that terminate in or just southwards of Cantabrian Spain. Most of these species' ranges extend across temperate Europe, and into Asia in many cases. Talpa europaea (mole) is found 62

LATE PALEOLITHIC ENVIRONMENTS

Spain's zoogeographic separation from the remainder of the Iberian Peninsula is also apparent, as is the richness of its microfaunal community.

Figure 5.2-Range urrounding region.

of specie

Mediterranean elements are also a part of the Cantabrian mammalian microfaunal community (Figure 5.4). The genet (Genetta genetta) is a viverrid species whose Holocene range is primarily African but extends into the Iberian Peninsula and southern France (Corbet 1978: 177). It is lacking from Upper Paleolithic levels corresponding with glacial episodes. Its appearance in Azilian Level I at El Pendo (Fuentes Vidarte 1980) corresponds with the warming of the Terminal Pleistocene. The Iberian mole (Talpa occidentalis) is endemic to the Iberian Peninsula and the southern portion of France. This species was formerly designated as T. caeca, which has a discontinuous Mediterranean range and differs karyotypically (Jimenez et al. 1984). The range of T. occidentalis corresponds to a warmer region than the more widely distributed T. europaea. Both the lesser white-toothed shrew ( Crocidura suaveolens) and Algerian mouse (Mus spretus) have primarily Mediterranean ranges which extend into Cantabrian Spain. The presence of this Mediterranean microfaunal component indicates the relative warmth of the local climate, as ameliorated by coa tal proximity and in particular the warm Gulf Stream current. The presence of these species among microfaunal components associated with cooler and wetter climates also indicates the mixed habitat composure of the region, which was also the case during the Pleistocene.

endemic to Cantabrian Spain and the

across the steppes of European Russia and throughout temperate Europe, generally not extending into the Mediterranean and Boreal zones. Its western range terminates across the northern half of the Iberian Peninsula, where it overlaps with the range of Talpa occidentalis. The range of Talpa europaea is primarily dependent upon burrowable topsoil (Corbet and Harris 1991). The range of Sorex minutus (pygmy shrew) extends from the Lake Baikal region of Siberia, through European Russia, Scandinavia, the United Kingdom, and through Cantabrian Spain and part of the Spanish Meseta. Sciurus vulgaris (red squirrel) is similarly distributed, but also is found in discontinuous patches in central Spain. Other Atlantic zone taxa do not extend as far into the Iberian Peninsula. Neomys fodiens (water shrew) extends across an expanse similar to that of Sorex minutus, yet its Iberian range terminates along the north coast. Glis glis (fat dormouse) has an Atlantic/Mediterranean range that terminates westwards in Cantabrian Spain, as a result of this species' association with deciduous forests (Corbet and Harris 1991). Its appearance in Cantabrian Spain during warmer phases indicates the return of forest cover (Altuna 1972). Clethrionomys glareolus (red backed vole) is also a terrestrial forest species and has a similar range. Microtus agrestis (field vole), Microtus arvalis (common vole), Arvicola terrestris (water vole), Apodemus flavicollis (yellow necked mouse), and Micromys minutus (harvest mouse) also follow this distribution pattern.

The many species which are found across a broader range of environmental zones do not serve to differentiate Cantabrian Spain from neighboring regions, except that they contribute to the characteristic richne s of the local faunal community (ranges not illustrated). Most of these species belong to the Order Carnivora and were increasingly preyed upon by Magdalenian hunters (Straus 1982). The pattern of this exploitation indicates that many of these species were hunted or

The confluence of overlapping ranges ending in Cantabrian Spain is illustrated in Figure 5.3. This distribution pattern illustrates the humid, temperate, and partially forested (under climax vegetation conditions) character of this region which is typical of large expanses of Europe. Some species extend past the Cantabrian region into drier habitats of the Meseta. Range extension into peripheral zones around the main range is typical for small mammal species, the juveniles of which normally disperse from their natal territories. Cantabrian

Figure 5.3-Ranges

63

of species inhabiting the Atlantic zone.

POKINES

Figure 5.4- Ranges of species inhabiting the Mediterranean zone.

trapped for their furs (Pokines 1998c). These species belong primarily to the Family Mustelidae: weasel (Mustela nivalis), ermine (M. erminea), mink (M. lutreola), polecat (M. putorius), pine marten (Martes martes), beech marten (M. foina), badger (Meles meles), and otter (Lutra lutra). Also common in Cantabrian Upper Paleolithic sites are red fox (Vulpes vulpes), wildcat (Felis silvestris), and lynx (Felis lynx). Other species with expansive ranges are the insectivores hedgehog (Erinaceus europaeus), Miller's water shrew (Neomys anomalus), greater white-toothed shrew (Crocidura russula); the lagomorphs rabbit ( Oryctolagus cuniculus) and brown hare (Lepus capensis); garden dormouse (Eliomys quercinus); and wood mouse (Apodemus sylvaticus). Some alpine microfaunal species are found in the Cantabrian lowlands during glacial episodes when their habitat zones had been pushed to lower altitudes by climatic deterioration (Figure 5.5). Both of these species are found elsewhere in the alpine zones of Europe: marmot (Marmota marmota) and alpine vole (Microtus nivalis). M. marmota has been reintroduced into portions of its former range in the Pyrenees and Cantabrian Mountains (Herrero et al. 1992). This species inhabits open pastures above the tree line, in areas where soil depth is sufficient to allow construction of extensive burrow systems. M. nivalis inhabits loose rock and grass areas above the tree line and more open montane woodlands (Krystufek and Kovacic 1989). The presence of either of these species in coastal lowland archaeological deposits indicates significant downward altitudinal shift of habitat zones during glacial episodes. The distribution of these species in Holocene Cantabrian Spain overlaps with or is in close proximity to multiple lowland species, illustrating the mixed habitats of this region. Contributing to the uniqueness of the Cantabrian mammal community are those northern species whose ranges were pushed southwards during glacial episodes and inhabited this

64

region alongside more temperate species (Figure 5.5). Nordic voles (Microtus oeconomus) are common in Cantabrian Upper Magdalenian sites during glacial episodes (Altuna 1972; Pokines 1998c). The European portion of this species' modern range extends no farther south than northern Germany. Within this range, M. oeconomus prefers marshy areas and humid grassland, also inhabiting tundra and boreal woodland (Corbet and Ovenden 1982; Ligtvoet and Van Wijngaarden 1994 ). This species is found commonly in archaeological levels alongside temperate zone microtid species like Arvicola terrestris, Microtus arvalis and M. agrestis. Wolverine (Gulo gulo) is found today in Holarctic tundra and boreal zones (Pasitschniak -Arts and Lariviere 1995). This species occurs rarely in Cantabrian Upper Paleolithic levels, as at Lezetxiki Solutrean/Gravettian Level II (Altuna 1972). Arctic fox (Alopex lagopus, modern range beyond map boundaries) is similarly rare in this context, having been found to date only at Amalda Perigordian V Level VI (Al tuna 1990). Tundra vole (Microtus gregalis, modern range beyond map boundaries) has been identified at Eralla Lower Magdalenian Level V (Peman 1985) alongside temperate Sorex , Arvicola, and Pitymys . M. gregalis was an infrequent inhabitant of the alpine tundra zone during glacial episodes and was therefore found in lowland contexts only rarely, and as a result of transportation by predators. The terrestrial mammalian microfauna of modern Cantabrian Spain is very diverse compared to most other temperate-zone regions and includes species from multiple environmental regions and habitat types in close proximity. In this respect it is comparable to Israel, whose terrestrial mammalian microfaunal community includes Palearctic and Ethiopian species found in Mediterranean, semi-arid, and arid habitats (Tchernov 1982). The diverse habitats indicated by the presence of these species were in easy access of one another: heathland, forest, humid meadow, dry meadow, wetland, and alpine zones. The terrestrial mammalian microfauna of Pleistocene Cantabrian

Figure 5.5-Range s of species inhabiting the Alpine and Boreal/ Tundra zones.

LATE PALEOLITHIC ENVIRONMENTS

Spain was even more diverse, since it included arctic species whose environmental zones were pushed southwards with the spread of continental ice sheets.

species. The importance of this resource is often overlooked because of the archaeological undervisibility of fish remains (Lubinski 1996; Pokines and Krupa 1997), which can be partially rectified through the use of fine-screening procedures (Freeman et al. 1998). Drainage from the Cantabrian Cordil lera occurs via multiple abrupt rivers, which run from the peaks north into the Cantabrian Sea. These rivers are typically no more than 50 km long and drain a limited area. This hydrographic situation contrasts with the remainder of the Iberian Peninsula, where much longer rivers such as the Ebro, Guadalquivir, or Douro drain much greater areas. More significantly , salmon rivers play unique hydrographic and zoogeographic roles; all breeding salmon returning from the ocean to their spawning streams must enter the freshwater system via the mouth of a river emptying into the ea. This food resource therefore concentrates seasonally at natural chokepoints along its migratory route. The larger a river's drainage area, the more salmon that must enter it at the ame point. Smaller rivers draining the same size land mass apportion the same biomass of adult salmon traveling to spawning steams at multiple, smaller chokepoints along the coast. Hunter-gatherers utilizing this seasonal resource in Cantabrian Spain had a more dispersed influx of salmon entering multiple, abrupt, parallel river valleys.

TERRESTRIAL MAMMALIAN MACROFAUNA The terrestrial mammalian macrofauna of Cantabrian Spain are more general in their modern distributions. Most of these species are found across a range of Palearctic ecological zones, with their ranges in some cases extending into the Nearctic. They therefore do not demarcate modern Cantabrian Spain from neighboring regions. This biogeographic picture was different during glacial episodes of the Late Pleistocene, when a cliserial shift forced by advancing ice sheets pushed the environmental zones of many species southwards. True steppe/tundra conditions were reached in France during peaks of extreme cold, with the concomitant species mix that included saiga antelope (Saiga tatarica), arctic fox (Alopex lagopus), and reindeer (Rangifer tarandus). Human predation on large mammal specie therefore followed a different pattern on the other side of the Pyrenees (Boyle 1990; Hemingway l 980; Pike -Tay 1991; Straus 1995) . At the French Pyrenean site of Dufaure, reindeer are the most abundant prey species (by NISP) in Levels 6 (Magdalenian), 5 (Middle Magdalenian), and 4 (Final Magdalenian), dropping off by Level 3 (Azilian) to a very small fraction in favor of red deer as the climate ameliorated (Altuna and Marriezkurrena 1995) . Cantabrian Spain remained relatively temperate during glacial episodes and thus sheltered a different macrofaunal species mix. True arctic species are rare in its Upper Paleolithic levels. Rangifer tarandus has been found only in very small amounts across many sites, as at Amalda Solutrean Level VI (Altuna 1990) , El Castillo Level 16 (Cabrera 1984), and Tito Bustillo Upper Magdalenian Level 1b (Altuna 1976). The range of this species may have extended at times from the Pyrenees along the alpine tundra/boreal zone, thus transported to coastal lowland sites by hunters.

Efficient wild harvesting of salmon impacts the form of ethnographically known hunter - gatherer societies, such as the chiefdoms of the Northwest Coast of North America. These richly complex cultures had as their caloric and protein base the rich salmon resources of the North Pacific Ocean (Mills 1989) . This rugged coastal region is marginal for agriculture and parallels Late Pleistocene Cantabrian Spain in its basic food resources . Intense, controlled seasonal harvesting and storage of a wild resource allowed the growth of settled villages, dense population, elaborate art, and intense craft specialization. Territorialism also was a practical option, since this reliable and non-dispersed resource was worth expending large amounts of energy to protect. Elaborate ceremonials (e.g., the potlatch) redistributed temporary surplus resources and were a cultural response to the parameters of this environment.

AQUATIC SPECIES Riverine geography also demarcates Cantabrian Spain from surrounding regions. Cantabrian Spain's coast and multiple river valleys provided access to many exploitable aquatic

Hydrogeography of major salmon rivers affected the growth of the cultural systems of the Pacific Northwest of North America. This region has a richer variety of salmon species: six species in the genus Oncorhynchus, whose runs do not coincide, inhabit the North Pacific, while one salmon species in the genus Salmo inhabits the North Atlantic. The larger drainage systems cause a greater number of salmon to pass through the bottlenecks of river mouths . The salmon resources of the Pacific Northwest therefore allow greater concentration of multiple salmon species at key chokepoints during multiple seasonal runs and hence a more prolonged harvest. Late Upper Paleolithic Cantabrian Spain had a more dispersed and less rich salmon resource. The latter region also had other important resources, so that primary reliance upon salmon was less viable and also unnecessary.

Table 5.2 Terrestrial Mammalian Macrofauna of Late Pleistocene Cantabrian Spain

Cervus elaphus

(boar)

(red deer)

Ursus arctos

Dama dama

(bear)

(fallow deer)

Felis Leo

Capreolus capreolus

(lion)

(roe deer)

Felis lyn.x

Capra pyrenaica

(lynx)

(Pyrenean ibex)

Canis Lupus

Rupicapra rupicapra

(wolf)

(chamois)

Rangifer carandus

Bison bonasus

(reindeer)

(bison)

Bos taurus

Equus caballus

(wild cattle)

(horse)

The coastline of Cantabrian Spain supplied another important gathered resource : marine molluscs. While the two most

Sus scrofa

65

POKINES

commonly taken species (Patella vulgata and Littorina littorea) have extensive modern distributions along North Atlantic coasts, in this region their ranges overlap with other abundant food resources. Mollusc exploitation was sparse during the Early Upper Paleolithic but accelerated during the Late Upper Paleolithic, when true shell midden accumulated in many sites, some many kilometers from the coast. These include levels at El Juyo (Krupa 1994), El Castillo (Cabrera 1984), La Riera (Straus and Clark 1986), Tito Bustillo (Moure Romanillo and Cano Herrera 1976), Altamira (Altuna and Straus 1976), La Lloseta (Clark 1971 ), Cueto de la Mina (Vega del Sella 1916), and Otero (Gonzalez Echegaray et al. 1966 ). Exploitation of this resource culminated in the concheros (shell middens) of the Asturian Period (Clark l 971, 1983). The availability of this easily harvested, year-round ource of protein further demarcates Cantabrian Spain as a distinct region .

warmer episodes (Sese 1995). This refugium status further defines Cantabrian Spain as a distinctive region during the Upper Paleolithic. These species include the now extinct Pliomys lenki, which persisted in this area far later than in other portions of its former European range (Alcalde Gurt 1982; Pokines 1998b) . Pliomys lenki first evolved at the beginning of the Plei tocene and has been found from Spain to European Russia (Bartolomei et al. 1975). It is known from multiple sites in Cantabrian Spain during the Middle/Upper Paleolithic, as at Lezetxiki (Ahuna 1972; Chaline 1970), Ekain (Zabala 1984), and Amalda (Peman 1990). This late urvival in Pleistocene Cantabrian Spain may be linked to this species ' association with forests, which persisted in this region during colder episodes, at least in localized patches. Palynological analysis al o indicate the presence of thermophile refugium plant species during periodic colder episodes , whose ranges re-expanded during warmer epi odes . These most likely existed in more heltered location in the coastal lowlands , particularly in and around sinkholes in the karst topography. An example of this situation i El Juyo; during the Lower Magdalenian, this coastal lowland site yielded pollen from the thermophiles oak and walnut (LeroiGourhan 1994).

AVIAN SPECIES The relatively temperate climate of Plei tocene Cantabrian Spain attracted a rich avian fauna, including year-round and sea anally migratory species. The mobility of birds decreases their use as direct climatic indicators; it is difficult to determine in which portion of their seasonal range they were taken. Their large ranges similarly make them less useful in defining zoogeographic regions. For example, despite accumulation over many fluctuations in climate, the avifauna of Amalda do not show marked increases in cold favoring species during glacial episodes (Eastham 1990). This richness, however, did add another food resource to the already abundant local mix, thus indirectly enhancing the region's uniqueness.

LATE UPPER PALEOLITHIC ENVIRONMENTS Fluctuations in Late Upper Paleolithic environments are also indicated most clearly by terrestrial microfaunal accumulations, as compared with and supplemented by data from other sources (Sese 1995). Reasons for this include the sensitivity of small mammals to changes in local habitats, the large sample sizes obtainable using modern excavation techniques, and the usual lack of a cultural filter affecting their inclusion in an archaeological site (Pokines 1996). Large terrestrial fauna suffer from this cultural bias in their proportions, as well as a broad range of occupied habitats for many frequently taken pecies, especially red deer. Questions of pollen taphonomy also limit its usefulness in reconstructing local environmental conditions (Sanchez Gofii I 993, I 994), in that differences in pollination syndrome (Dunwiddie 1987; Prentice 1986), differential transport (Goodwin 1988; Holmes 1990; Traverse 1990), and differential destruction (Havinga 1967) may all affect the proportions of pollen species present in archaeological sediments. Macrobotanical remains are generally recovered in small amounts and suffer from a cultural bias in their site introduction and differential survivability and identifiability (Stettler 1995).

Waterfowl exploitation intensified during the Late Upper Paleolithic in Cantabrian Spain. The coastal wetlands attracted many of these taxa, such as mallard (Anas platyrhynchos), teal (A. crecca), pintail (A. acuta), and goose (Anser spp.) found in La Riera Late Upper Paleolithic levels (Eastham 1986). Magdalenian levels yielding avian remains include Balmori, Coberizas, Altamira, El Pendo, Otero, La Chora, Lumentxa, Torre, Erralla, Ekain, Urtiaga, and Aitzbitarte (summarized by Straus 1992: 151). Tundra-dwelling species such as snowy owl (Nyctea scandiaca) appear only rarely in Cantabrian Spanish sites. These include the Lower Magdalenian Level V at Erralla (Eastham 1985), which also yielded the remains of ptarmigan (Lagopus mutus), an arc-tic/alpine species. This site is close to France, where true tundra conditions prevailed during cold spells. From that region, tundra-species ranges may have temporarily expanded, as is the case with "irruption" years for snowy owls, when cyclical winter range expansion results from crashes in lemming and hare prey populations in their arctic home ranges (Burton 1973).

The Solutrean Period in Cantabrian Spain (20,500 to 16,500 B.P.) corresponds approximately with the last glacial maximum (Figure 5.6). It is underrepresented by microfaunal data, so paleoenvironmental inferences must be drawn by the presence/absence of indicator species. Cold and treeless conditions are indicated by the sample from the Upper Solutrean levels (!Vb and !Va) from Amalda (Peman 1990). These have been radiocarbon dated to after the glacial maximum; dates of 17,580 ± 440, I 6,200 ± 300 and 16,090 ± 240 B.P. indicate some overlap with the succeeding Lower Magdalenian

REFUGIUM SPECIES During glacial conditions, some species were distributed in Cantabrian Spain whose ranges had contracted from surrounding regions , with range expansion occurring during 66

LATE PALEOLITHIC ENVIRONMENTS

period. The microfaunal samples from these two levels are dominated by open meadow microtines (Microtus arvalis, M. agrestis, Pitymys pyrenaicus, and P lusitanicus), indicative of drier and cooler conditions. M. nivalis is also abundant for a coastal Cantabrian site (total n = 22) and indicates open rocky areas near Amalda. Humid-indicator species are reduced, including Arvicola cf. terrestris and M. oeconomus. Some tree cover is indicated by the presence of Apodemus.

Initial Magdalenian Level 5 has been dated to 16,433 ± 131 B .P., and Lower Magdalenian Levels 4 and 3 have yielded dates respectively of 15,988 ± 193 B.P. and 15,173 ± 160 B.P. The small microfaunal sample from Level 4b includes the anomalous presence of Glis glis, indicating some forest cover. Other species reflect open and humid conditions. The Lower Magdalenian levels of La Riera ( 18, 19 and 20) also indicate open and humid conditions and include one alpine indicator (Microtus nivalis) at a site adjacent to the modern coastline (Altuna 1986). These levels have yielded comparable determinations of 16,420 ± 430, 15,520 ± 350, and 15,230 ± 300 B.P. (Straus and Clark 1986).

The Solutrean level microfaunal sample of Aitzbitarte exhibits a definite cold character (Altuna 1972). While A. terrestris has the highest frequency among the microtines (n = 120), M. oeconomus is similarly abundant (n = 109) and far outweighs both M. agrestislarvalis (n = 36) and Pitymys (n = 13). Alpine-dwelling M. nivalis is also present (n = 3), while no Apodemus has been recovered from this level. Talpa is the most abundant taxon (n = 616) as a result of its greater ease of identification from multiple durable elements. This microfauna assemblage indicates a humid, cold, open environment with perhaps sparse trees and open water nearby.

The Lower Magdalenian Levels Vlla-f at Ekain show a similar pattern of open, cold, and humid conditions (Zabala 1984). Microtines are dominated by Arvicola and Microtus oeconomus with M. nivalis present. Pitymys and M. agrestislarvalis, indicative of dryer grasslands, are present in much smaller numbers. Altamira Lower Magdalenian Level 2+3 reflects a similarly open and humid environment, but with more indication of grassland and a forest species (Apodemus sylvaticus) present (Castro Bernardez 1986). Greater tree cover is indicated by the Magdalenian (Level 6) and Middle Magdalenian (Level 5) from Dufaure in nearby Pyrenean France, where Apodemus is a significant portion of the microfaunal assemblage (Eastham 1995). A cold, open, and humid environment is also indicated by the dominance of Arvicola terrestris, Microtus ratticeps (= oeconomus), and M. agrestis. The microfaunal sample of the undesignated Magdalenian level at Aitzbitarte (Altuna 1972) reflects a harsher environment and is hence most probably Lower Magdalenian in age. Cold, open, and wet conditions are indicated by the dominance of A. terrestris (n = 195) and M. oeconomus (n = 144) and the presence of M. nivalis (n = 5), while no Apodemus have been recovered. Some dryer grassland is present based upon the presence of M. agrestis/arvalis (n =65) and Pitymys (n = 18).

The Middle Solutrean levels ( 15, 16, 17) of Las Caldas have yielded Apodemus cf. flavicollis, M. arvalis, Microtus spp., Arvicola sapidus, and Pitymys cf. lusitanicus (Corch6n 1981 ). These species are also common modern Cantabrian fauna, although the microtines favor open environments and Apodemus favors woodland. Remains from Solutrean Level 3 at Cueva Morfn is less informative, with only a single remain each of Arvicola and Talpa reported (Altuna 1978). The Solutrean level at Ermittia (Altuna 1972) has yielded only Microtus agrestislarvalis, M. oeconomus, Arvicola terrestris, and Talpa europaea. The presence of M. oeconomus agrees with ocean-level data indicating a cold phase during accumulation. La Riera has yielded a small microfaunal sample from its Solutrean levels (4, 5, 7- 11 and 14-17) that consists mostly of Arvicola terrestris. The presence of Galemys pyrenaicus in Level 7 is a strong indicator of downward alpine zone shift, since La Riera is almost upon the modern coast.

The microfaunal sample from Lower Magdalenian Level V at Erralla is more informative (Peman 1985). The Erralla Level V occupation has been radiocarbon dated to 15,740 ± 240

The Magdalenian period has sufficient amounts of recovered microfaunal remains to allow paleoenvironmental interpretation beyond indicator species for portions of this time. This culture stratigraphic phase spans 16,500 to 13,500 B.P. (Lower Magdalenian) and 13,500 to 11,000 B.P. (Upper Magdalenian) in Cantabrian Spain, with some overlap between these two major divisions. This archaeological period spans the late Oxygen Isotope Stage 2 and the onset of Stage 1, representing sea-level rise from the glacial maximum during the Solutrean period towards the high levels of the Holocene. Fluctuations between these two extremes complicate interpretation, since ocean level rise was not steady and went through repeated temporary lowerings before obtaining Holocene levels after 10,000 B.P.

Relative Ocean Level

Upper Paleolithic

Tool laduslry Azilian

Upper

MacdaJenian

AllerOd Dryas lI

12.000

Bolling

Lower

MacdaJeniH

Solutrnn

Lower Magdalenian environments in Cantabrian Spain correspond to generally higher ocean levels and cold conditions ameliorated from glacial-maximum conditions. Lower Magdalenian levels just postdating the glacial maximum have been excavated at the upland site of Rascafio (Altuna 1981 ).

Dryas le Pre-BOIiing Dryas lb Angles Dryasla Lascaux

14.000

16,000

18,000

2

Laugene 20,000

22.000

Figure 5.6---Relative ocean level, adapted from Sanchez Gofii 1993: 144.

67

POKINES

B.P. and 16,200 ± 240 B.P. (Altuna 1985:26). This time corresponds with rising ocean levels just postdating the glacial maximum c. 22,000-18,000 B.P., so climatically rigorous conditions were likely. This level has the largest total microfaunal MNI (n = 7469) for any Cantabrian Upper Paleolithic level to date. The microfaunal profile is dominated by humid meadow/ water edge species. This total includes a very high proportion of Microtus oeconomus, indicating significantly cold as well as humid climate. This open landscape also had significant areas of dry grasses . Forest cover was minimal near the site, amounting to well under 1% of the land area . That some for est was present is also indicated by Glis glis, the presence of which in this level contradicts the general environmental pic ture . Any tree cover was found in sheltered areas on the landscape. Harsh conditions are reinforced by the presence of the alpine species Microtus nivalis and the only Cantabrian Upper Paleolithic occurrence of Microtus gregalis (steppe lemming) . This latter species is found today in the tundra and steppe bio mes of Asia. These two species indicate significant altitude zonal shift downwards to within a few kilometers of this lowland site.

represented by the singular appearance of Pliomys lenki. This occurrence represents the latest survival in Cantabrian Spain of this now extinct species . The abundance of P. lenki in Pleistocene deposits correlates with interstadial conditions and the presence of extant forest species (Bartolomei et al. 1975) in its European distribution; hence, it was likely a for est species. The presence of some forest cover is also supported by the continued presence of Erinaceus europaeus (NIT = 40). Humid, open vegetation species still dominate the assemblage : Talpa, Sorex coronatus , S. minutus , Microtus oeconomus, Pitymys lusitanicus , and Arvicola cf. terrestris. Also present are two water-associated shrews , Neomys Jodi ens and N. anomalus . Species totals associated with dry meadow increased slightly from Level 6. Climate during the time of Level 4/4S show s some deteriora tion . The alpine species Microtus nivalis first appear s at El Juyo during this time . Th e presence of M. nivalis indicates the downward attitudinal shift of the exposed alpine rock zone occupied by this specie s with the concomitant downward spread of alpine glaciation (Krystufek and Kovacic 1989). The alpine zone must have intruded to much lower altitudes during the time of Level 4/4S deposition. Cold conditions are further reinforced by the continued presence of M. oeconomus.

The Lower Magdalenian levels of El Juyo show gradual climatic deterioration during their deposition. Microfauna from Levels 9, 8, 7 (dated to 14,400 ± 180 B.P.), and 6 indicate a cold, very humid, and open environment. Palynological and macrobotanical analysis indicates that heath predominated, along with many wetland taxa (Freeman et al. 1988). Some forest cover and areas of grassland are indicated. These levels correspond with the relatively temperate Pre-Bolling pollen zone. Microfauna from Level 4/4S (dated to 13,920 ± 240 B.P.) indicate environmental deterioration corresponding with temporary lowering of ocean level and the Dryas le pollen zone. An increase in alpine species is noted among the microfaunal, pollen, and macrobotanical remains, which indicates significant lowering of alpine zones.

The pollen profile for El J uyo is in agreement with these findings. In general, non-arboreal pollen dominates the spectrum and includes large proportions of Graminae (grasses) and Ericaceae (heather). Arboreal pollen oscillates around 25% of the total per sample from Levels 12 through 7, then drops off to lower levels in Levels 6 through 4. After a hiatus, arboreal pollen for Levels 2 and 3, Bronze Age and more recent, has returned to higher levels around 35%. Levels of Pinus (pine) pollen show a slight tapering from Level 12 through 7, with values in the neighborhood of 1% to I 0%. Other tree taxa represented include fir (Abies), oak (Quercus), walnut (Jug/ans), elm ( Ulmus), ash (Fraxinus), birch (Betula), alder (A/nus), and hazel (Cory/us) . Forest cover is also indicated by the frequent inclusion in sediments of carbonized wood, primarily from the taxa Pinus, Salix, and Populus (Freeman et al. 1988: 12). Patches of mixed forest are thus indicated, with pine and oak dominating. Levels 12 through 7 are dominated in particular by Ericaceae, which Leroi-Gourhan (1994:123) interprets as occupying most of the terrain, indicating a coastal heath vegetation with interspersed trees in more sheltered locations. Proportion of these species oscillates around 43% of pollen totals. Climatic stability is indicated for an extended period of time encompassing these lower levels. Temperate conditions are indicated by the presence of thermophiles such as walnut and oak. Levels 12 through 7 of El Juyo are interpreted as having been deposited during the relatively temperate Pre-Bolling phase. Climatic deterioration is indicated in Levels 6 through 4 by the decrease in arboreal pollen, the sharp drop in Ericaceae pollen (from a peak of 67% down to 2% ), and the sharp increase in Graminae to levels of over 50% in some samples. In particular, Cicoriaceae (chicory family) and Poaceae (meadow grasses) jump in importance. The interpretation for the Levels (6-4) of the Dryas le phase is for an open steppe environment with much

The climate of Level 6 was cool and humid, with coastal heath interspersed with small forest patches and some grasslands. Forest-associated taxa (Apodemus) appear in small amounts, as does the forest-edge species Erinaceus europaeus. Humid meadow- and/or water-edge species continue to dominate the faunal profile, especially Arvicola cf. terrestris. These include, for the first time in this site, Neomys fodiens and N. anomalus . Each of these shrew species is normally associated with water-edge habitats, but they can also be found away from water in humid vegetation settings. Cold, humid climate is reinforced by the large amounts of Microtus oeconomus. Both Microtus arvalis and M. agrestis were probably present, as supported by biometrical analysis (Pokines 1998c). These species indicate the presence of some dry grassland . Carbon from Level 4 has been dated to 13,920 ± 240 B.P. (Freeman et al. 1988:12). Open, humid heath continued to dominate the landscape with small patches of forest in more sheltered areas . In addition to Apodemus , forest species are 68

LATE PALEOLITHIC ENVIRONMENTS

colder and dryer conditions than previously. True tundra is not indicated, and oaks have nearly disappeared from this open landscape, as have any other thennophilic species.

shows relatively little microfaunal turnover from its Magdalenian level (6) through its Middle (5) and Final Magdalenian levels (4 inf. through 4 sup.) (Eastham 1995).

Cantabrian Upper Magdalenian environments change in a fluctuating manner from relatively harsh interstadial to warmer and more forested conditions. Smaller samples are available from Upper Magdalenian Level 2 at Rascafio (Altuna 1981) and Level I at Lezetxiki (Altuna 1972). The Upper Magdalenian Levels lac of Tito Bustillo have been dated to the very early Upper Magdalenian period and overlap with the end of the Lower Magdalenian. These radiocarbon determinations range from 13,520 ± 300 to (perhaps anomalous) 15,400 ± 300 B.P. (Moure Romanillo 1975). A date just preceding 13,000 B.P. places Tito Bustillo's occupation amid the relatively temperate Bolling pollen zone. The microfaunal sample from this site (Altuna 1976) suggests the earlier dating is correct, placing occupation during the end of the colder pollen phase Dryas le. While small, the sample has only one element of a forest species (Apodemus) and is dominated by open-environment microtines: A terrestris (n = 61), M. oeconomus (n = 7), and M. agrestis/arvalis (n = 6). A coldphase attribution for this occupation is further supported by the presence of a single reindeer (Rangifer tarandus) remain from level lb (Altuna 1976:153). A similar microfaunal sample has been recovered from the Upper Magdalenian Level Vlb of Ekain (Zabala 1984 ).

The Azilian period in Cantabrian Spain (11,000 to 9,000 B .P.) occurred during the Terminal Pleistocene/Initial Holocene and therefore spans a period of abrupt ocean level rise and climatic amelioration. This period begins during the cooler Dryas III pollen zone and continues into the Preboreal. Global sea level is in Oxygen Isotope Stage 1, where levels rapidly rose to the approximate modern level after l 0,000 B .P. The changing Cantabrian Azilian environment reflects warming climate with spreading forests, the cliserial shift northwards of cold-adapted species' ranges, and the retreat upwards of alpine biotic zones (Penalba 1994 ). This transition is recorded at La Riera: the small microfaunal sample from Azilian Levels 27-30 includes the arboreal dormouse species Glis glis for the first time (Altuna 1986) and is strongly indicative of forested conditions. A similar pattern is found at El Pendo (Fuentes Vidarte 1980). The microfaunal samples from the Azilian levels of Rascafio (Altuna 1981), Ermittia and Urtiaga (Altuna 1972), and Ekain (Zabala 1984) are uninformative, as they yield primarily Arvicola and Talpa. The situation is similar for Azilian Level 3 from the nearby French site of Dufaure, although the arctic vole Microtus oeconomus persists locally (Eastham 1995). Aitzbitarte has yielded a more substantial microfaunal sample from its Azilian level (Altuna 1972). Apodemus shows up for the first time in this site. Aitzbitarte has a late persistence of Microtus oeconomus and M. nivalis, although the former species decreases relative to other microtines in the Magdalenian level at this site.

The later occurrences of the Upper Magdalenian show mixed amounts of environmental amelioration and may correspond with the Bolling (warmer), Dryas II (cooler), Allerod (warmer), and Dryas III (cooler) pollen zones. Levels 24 and 21-23 of La Riera have been dated to the end of this phase (Straus and Clark 1986). This site retains alpine vole Microtus nivalis in Level 24 (Altuna 1986). The sample size is too small to conclude more about the environment other than probable assignment to a cooler phase. The Final Magdalenian Level Vla at Ekain has a very similar microfaunal sample (other than the relative reduction in Arvicola), as does the Upper Magdalenian Level Vlb at this site (Zabala 1984).

DISCUSSION The Late Upper Paleolithic Cantabrians harvested key species with increasing intensity and broadened their resource base during this time of fluctuating environment leading to the wanner Holocene (Freeman 1973, 1975, 1981; Straus 1977). Red deer (Cervus elaphus) was generally the most abundant large taxon taken. This situation is illustrated at sites such as Upper Magdalenian Tito Bustillo (Altuna 1976) and at the Solutrean, Magdalenian, and Azilian sequence at La Riera (Altuna 1986). The red deer taken at El Juyo exhibit a catastrophic age profile, a probable indication that drive and surround techniques taking whole herds at once were practiced (Klein and Cruz Uribe 1987). Mass-game harvest at selected times is also reflected in the projectile technology of this period. A gradual shift occurred from lithic and oval-profile antler projectile tips during the Solutrean to quadrangularprofile antler projectiles typical of the Cantabrian Lower Magdalenian (Pokines 1998a). These latter points are both more reliable in an engineering sense (Bleed 1986) and exhibit a pattern of more systematic manufacture and use. This projectile type was replaced by the harpoons typical of the Upper Magdalenian and Azilian periods.

The Final Magdalenian levels at Erralla are more illuminating due to their greater sample size. Level III has been dated to 12,310 ± 190 B.P. (Altuna 1985:26), coinciding with the Dryas II or Allerod pollen phases. Assignment to the temporary warming conditions of the Allerod is more likely, based upon the microfaunal sample (Peman 1985). The arboreal Glis glis shows up for the first time at Erralla in these levels, as does forest-dwelling Clethrionomys glareolus. Apodemus is also abundant, while Microtus nivalis declines from previously higher levels. The other alpine species (M. gregalis) does not recur. While M. oeconomus is still present, it no longer dominates the microfaunal sample. Other indicators of humid, open cover have declined (Arvicola, Talpa europaea and Sorex spp.). The changed microfaunal profile indicates that the cool, humid heath with sparse trees of the Lower Magdalenian has shifted towards dryer grassland with increasing tree cover and a retreat of the alpine zone away from Erralla. In contrast, the nearby French site of Dufaure

Another shift in faunal exploitation is the use of specialized hunting camps, such as the upland sites of Rascafio and El Mir6n, where the Magdalenian and Azilian levels are 69

POKINES

dominated by ibex (Altuna 1981; Straus and Gonzalez Morales 1997). Ibex are also plentiful at other sites, such as Erralla and Bolinkoba . This macrofaunal species was also selectively harvested, most likely using drive and surround techniques (Straus 1987), and came from an attitudinal zone not normally occupied by these hunter-gatherers. These sites may have been occupied during the warmer months, with migration to the coastal lowlands during the colder months. Warm-season aggregation of dispersed bands has also been interpreted at the site of Altamira by Conkey ( 1980), including the purpose of hunting cooperation (Freeman 1994). Marine molluscs also show an increasing pattern of exploita tion throughout this period (Krupa 1994; Straus 1981). While of limited dietary importance during both Mousterian and Early Upper Paleolithic occupations, molluscs assume a ignificant role during the Solutrean and were accumulated in true shell middens beginning with the Lower Magdalenian . The major occupation levels at El Juyo have each yielded midden -like quantities of shells , primarily of limpets (Patella vulgata) and periwinkles (Littorina spp.) . Significantly large abundances have also been found at sites such as Altamira, El Castillo, Otero, and Tito Bustillo during the Upper Magdalenian (Krupa 1994). This trend was to accelerate, culminating in the true concheros of the Mesolithic, regionally described as the Asturian Period (Clark 1971, 1983; Straus 1991).

this same element representation discrepancy , nor do smaller, more fragile mammalian taxa exhibit such a bias. The preponderance of fur-bearing mammal foot elements is therefore attributable to their selective introduction to the site. This pattern is the logical result of foot bones left in a pelt that was skinned from a trapped or hunted animal away from the site, with the remaining carcass discarded in the field. Mustelid meat has a strong, unpleasant flavor and therefore may have been a less desirable food choice. This same pattern of fur exploitation has been posited for other Cantabrian sites (Straus 1977, 1992). Aitzbitarte has yielded several remains from Mustela nivalis , M. erminea , and M. putorius from its Solutrean and Magdalenian levels (Altuna 1972). Bolinkoba has also yielded the remains of mustelids and red fox (Vulpes vulpes). Population increases during thi same period is shown by the increasing number of sites and in the intensity of their occupation (Straus 1977, 1992). The conflation of population increa e and the intensity of resource exploitation at this time of fluctuating climate indicates a growing separation of human behavior from the constraints of natural setting. Microfaunal and other indicators of paleoenvironment portray a humid, cold and open landscape moving through multiple cycles of poorer and more favorable conditions, approaching the temperate and more forested Holocene. Red deer, ibex, molluscs, salmon, and furs were obtained in an increasingly efficient and cooperative manner. These trends are reflected in the faunal remains themselves, as well as in the changing patterns of site use and technology.

Fish also exhibit a pattern of increasing exploitation during this period. El Juyo has currently yielded over 500 intact vertebral centra and many times this amount of vertebral fragments of large fish species, primarily Atlantic salmon (Salmo salar) (Krupa 1996). The majority of these remains await final identification. Far more compelling are the additional remains of fragmentary fish ribs and spine yielded by the fine screening of all sediments from this site. Nearly 45,000 of these fragments have been recovered from all levels and areas of El Juyo. These remains are extremely fragile, far more so than terrestrial microfauna recovered from the same levels, and thus represent only a fraction of all fish remains introduced to the site. A complete lack of far more durable otoliths and only tiny amounts of cranial elements indicate that processing took place away from the site, with heads disposed of elsewhere. Many other Cantabrian sites have yielded fish remains, although in smaller amounts probably due to differences in recovery system. These include most notably La Riera, the remnant stratigraphy of which has yielded 81 identified Salmo remains from its Late Upper Paleolithic sequence (Straus and Clark 1986).

This pattern of fauna} exploitation developed within Cantabrian Spain and was made possible by its unique zoogeographic character. The relatively temperate Late Pleistocene climate supported abundant large game across coastal heathland, grassland, forest patch/edge, and alpine zones. Aquatic habitats supplied stable, abundant, and easily harvested salmon and molluscs. The former could be preserved and stored for later use, and the latter was available year round and is more accurately cla sified as a gathered resource than a hunted one. The diverse array of fur-bearing mammals was obtainable by trapping across many of the local habitat types . These resources remained available and were exploited with increasing intensity throughout the Late Upper Paleolithic, despite major fluctuations in climate. The diverse microfaunal community was associated with equally rich plant communities that included thermophiles in more sheltered locations alongside cold and humid associated species, patches of drier grassland, and alpine zones. This mosaic provided a multitude of plant resources that are only now being recovered in sufficient amounts for meaningful analysis of Paleolithic plant exploitation (Stettler 1995).

Faunal exploitation in this period includes hunting/trapping of small- to medium-sized fur-bearing mammals specifically for their pelts (Straus 1982; Pokines I 998c). El Juyo so far has yielded 72 foot elements, primarily phalanges and metapodials, of carnivores in the size range of polecat (Mustela putorius), European mink (Mustela lutreola), marten (Martes spp.), and red fox (Vulpes vulpes). The presence of these elements is in striking contrast to the virtual lack of any other elements, including teeth, from these taxa. Non-carnivores of this size range, including rabbit and hedgehog , do not show

The Pleistocene environment was very different in nearby France and may have greatly affected the cultural systems that developed in that region. The habitats of France often included areas of true tundra, more expansive grassland areas, and boreal forest. Human groups may have migrated along with reindeer herds in order to harvest more efficiently this primary resource (Boyle 1990; Hemingway 1980) . Salmon 70

LATE PALEOLITHIC ENVIRONMENTS

away from the coast would have been a more dispersed resource as it migrated into multiple small spawning streams. Territorialism by human groups would have been a less likely response to this set of resources, given their mobile states and lower density, respectively. Harsher periglacial environments may have also supported lower density human populations during colder phases and offered fewer plant resources. The Magdalenian settlement systems of France, therefore, were a very different adaptation to a zoogeographic region distinct from Cantabrian Spain, despite their similarities in technology, tool forms, and art styles . The cultural systems, and most especially the differences between them, are best analyzed starting with the zoogeographic parameters defining each region.

Echegaray and L.G. Freeman, pp. 201-209. Santander: Instituci6n Cultural de Cantabria. Altuna , J ., 1981. Restos oseos del yacimiento prehist6rico del Rascafio. In El Paleolitico Superior de la Cueva del Rascafio (Santander), edited by J. Gonzalez Echegaray and I. Barandiaran Maestu, pp. 221-269. Centro de Investigaci6n y Museo de Altamira Monografias 3. Santander: Centro de Investigaci6n y Museo de Altamira. Altuna , J., 1985. Introducci6n. Munibe 37:7-9. Altuna, J ., 1986. The mammalian faunas from the prehistoric site of La Riera. In La Riera Cave, edited by L.G. Straus and G.A. Clark, pp. 237-274 . Arizona State University Anthropological Research Papers 36. Tempe: Arizona State University.

Zoogeography is just one of many ways in which regions may be defined. The rationale behind the assignment of any boundary around an area of archaeological investigation must continually be questioned, given the inherent artificiality of any imposed modern definition . Archaeologists should also examine parameters such as geology when defining a region, since this has de facto influenced the distribution and preservation of archaeological sites. Cantabrian Spain's karst topography also defines it as a region in that it allows for the preservation of habitation and art sites within caves. The availability of these sites for excavation, as in France, in many ways has delineated this region, yet may place a modern boundary where none existed in the minds and actions of its Paleolithic inhabitants. Analyses therefore should be framed by factors that would have shaped the economic and social activities of these groups.

Altuna, J., 1990. Caza y alimentaci6n procedente de macromamiferos durante el Paleolitico de Amalda . In La Cueva de Amalda (Zestoa, Pais Vasco), edited J . Altuna, A. Balde6n , and K. Mariezkurrena, pp. 149-192. San Sebastian: Sociedad de Estudios Vascos . Altuna, J. , and K. Marriezkurrena, 1995. Les restes osseux de macromammiferes. In Les Demiers Chasseurs de Rennes du Monde Pyreneen . l'Abri Dufaure: Un Gisement Tardiglaciare en Gascogne, edited by L.G . Straus, pp. 181-211. Memo ire de la Societe Prehistoriq ue Franr;aise 22. Paris : Societe Prehistorique Fran9aise. Altuna, J., and L.G. Straus, 1976. The Solutrean of Altamira: The artifactual and faunal evidence. Zephy rus

ACKNOWLEDGMENTS

26-27:175-182.

This research would not have been possible without the efforts of and access to materials provided by Dr. Leslie Freeman and Dr. Joaquin Gonzalez Echegaray. Financial support was provided by the Field Museum, Division of Mammals (Chicago, USA), the Instituto para Investigaciones Prehist6ricas (Santander, Spain), and the National Science Foundation.

Bailey , G., 1983. Economic change in Late Pleistocene Cantabria . In Hunter-Gatherer Economy in Prehistory, edited by G. Bailey, pp. 149-165. Cambridge: Cambridge University Press . Bartolomei, G., J. Chaline, 0. Fejfar, D. Janossy , M. Jeannet, W. von Koenigswald, and K. Kowalski, 1975. Pliomys lenki (Heller 1930) (Rodentia, Mammalia) en Europe. Acta Zoologica Cracoviensia 20:393-467.

REFERENCES Alcalde Gurt, G ., 1982. Presencia interessant de Pliomys lenki i de Microtus oeconomus en el reompliment de la cova

1

~ ~

2

16

G)

c:5i5iiiiiiiiiiilv r\ 3 \.Ji)

(v\ \6)

l.fe- 61 c:==I::> ,.~~ ;;,~ ' 47 'MIL j 62

31

46

I

~ 11 32

18

~

~

E··--~~~~

~~~~~ 6



63 ~78

(_~]64~9

20

35

21

36

0

:=~, : ,'

CS···r£::> ~ ~ 5

,

l~~-~~~

CJITD48

33



'---./

50

~ 65

51 ~6

81

c:z:=:::>~~~~ 7

80

E-:::,

22

37

52

~7

----

82

~ ~ ~ ~ ~ 68L/d83 ~~ 9

24

~

~ c~~-~

~

39

54

69

84

~~~~cs 10

25

40

\

55

~ ~ ~ 11

26

85

70

r

: ~ ~

41

56

71

86

~~~~~ ~ 12

27

~ ~

42

c=z::,

57

~

~~~c ::/; 14

29

44

~ 59

~~~~ 15

Figure 6.2-Design

30

45

60

elements identified on portable objects in the We tern Languedoc-Roussillon.

79

87

72

~

C==®



73

~

74

u

75

~88

89

.

PRICE

0

10cm

Figure 6.3--Tracings of a selection of the wall engravings from different surfaces in La Grotte Gazel (after Meroc and Simonnet 1965; Sacchi 1986, 1993b).

80

VARIABILITYAND CONTEXT OF MAGDALENIAN VISUAL IMAGERY

as much as 3,300 years. The following analysis does not assume that these assemblages were created contemporaneously, but simply assumes that they were created by people following a way of life and using a material culture that had not changed significantly.

Canecaude I This medium-sized cave with a very low ceiling is located at 250 meters altitude in a narrow stream valley. The lowceilinged cave offers roughly 75 m2 of potential living space. The Middle Magdalenian period occupation of the site, dating to 14,230 ± 160 years B.P. (Gif-2708) (Sacchi 1986: 143), has been interpreted as a short-term reindeer hunting camp. Over 60 percent of the faunal assemblage is composed of reindeer remains (Fontana 1994). This is one of several cavities in the enclosed, cramped valley of the Rieu Sec, a tiny affluent of the Orbiel River. The small obscured opening of Canecaude I lies just below a steep and loose scree slope at the edge of a limestone plateau. There is a minor porch at the front of the modern entrance, large enough for a small group of up to five people to stand. The entrance itself is narrow and restricted, limiting the amount of light that enters the cavity . The cavity opens to a small room off to the left and a small room to the right. Upon entering, the cavities quickly become dark. The Magdalenian deposits are only a few centimeters below the present surface. It is probably the case that people visiting and occupying this cave 14,000 years ago experienced a similarly restricted space.

20 people could have sheltered in this chamber alone. Gazel has been interpreted as an aggregation site based on the size of the site and the richness and size of the assemblage with relation to others in the region. Seasonality studies suggest it was occupied from late spring through early fall. Reindeer make up 50 percent of the faunal assemblage, and horse make up 30 percent (Fontana 1995a). The two modern entrances to the cave are located at roughly 20 meters above the bed of the intermittent stream known as La Ceize. The view from the porch in front of either entrance includes this small but rather steep stream valley and the opposite hill. The climate is Mediterranean. Today the surrounding land is cultivated in grape vines and small-scale family gardens . The soils are abundantly strewn with limestone debris from the exposed formations that house numerous cavities such as Gaze). From the opposite hill looking south over the extensive plain of the Aude River, the snow capped peaks of the eastern Pyrenees are visible on a clear day. Springs related to the karstic system are scattered across the landscape. Local streams are intermittent, with notable variation in wetness and associated vegetation from year to year. La Caune de Belvis

This cave is of medium size with 65 m 2 of available floor space. It is largely open to daylight due to a roof collapse that occurred just prior to the Late Pleistocene occupation. La Caune de Belvis is located in the eastern Pyrenees to the west southwest of the town of Quillan in the Aude departement. At 960 m above sea level, it is the highest altitude site included in the present study. The cave has an extensive view of the broad Espezel plain below. The faunal and palynological evidence suggests that the occupants were taking advantage of a diverse set of ecological niches, including lake, stream, or river, as well as open and wooded zones, from early spring to late fall. The identifiable specimens of the faunal assemblage are dominated by ibex (60 percent). The remaining 40 percent of identifiable specimens are divided among horse, deer, hare, bison, aurochs, chamois, fox, and trout (Fontana 1995b ). The assemblage has been interpreted as a multipurpose site seasonally occupied by a relatively small group of huntergatherers, possibly a microband or extended family .

No more than five people could have sheltered in the left chamber. The chamber to the right, the location of excavations that yielded the study assemblage, could have accommodated as many as fifteen people in cramped conditions. The interior walls of this cave are not as smooth as those of other limestone caves in the region. They are a reddish limestone with uneven surfaces full of breaks and niches. The climate today is Mediterranean. From the mouth of this cave, even in colder climates with less vegetation, one could look out across the narrow stream valley and see only the steep face of the opposite bank. Rather than a vantage point, this cave provided a protected and sheltering space. However, standing atop the plateau above, the view opens up considerably, encompassing the stream valley and the continuation of the plateau on the other side.

Today there are reservoirs in the area to supplement the meager natural water supply. The nearest source of natural water is at a distance of 700 m from the site. The climate today is sub-Atlantic and sub-Mediterranean and, therefore, somewhat more temperate than the climate of the other sites. The cave receives snow and is subject to frost from the second half of October through the first half of May .

La Grotte Gaze/

Gazel is a large cave with well over 1000 m2 of potential living space and many chambers . The cave is located at 250 meters altitude, in a winding stream valley at the transition between the broad Aude River plain and the Montagne Noir. This is by far the largest site in the region, with the densest occupational debris from the Magdalenian period. Gazel has engraved imagery on the walls of a chamber over 100 meters from the entrance and primary habitation area. The Middle Magdalenian assemblage from Gazel, dated to 15,070 ± 270 years B.P. (Gif-2655) (Sacchi 1986:127), was recovered by Sacchi from the main room at the southern entrance. At least

The cavity holds over three meters of fill. Four thin layers of Late Magdalenian deposits dated to 12,270 ± 280 B.P. (Gif-2950) lie near the present day surface (Sacchi et al. 1993). Sacchi has excavated nearly 50 percent of these deposits . The cavity is limited to a single room that could have accommodated a group of five people comfortably and 81

PRICE

up to ten in cramped conditions. The porch area is rather limited and abutted by a distinct talus slope. The interior wall surfaces are very smooth light grey limestone that could have served as surfaces for painting or engraving; however, no evidence of these activities exists today.

Procurement/Production/Enigmatic At a general level, dividing the bone and antler industry into procurement, production, and enigmatic categories reveals a consistent distinction between the three sites. La Grotte Gazel has the greater number of object classes in all three subdivisions. Canecaude I exhibits a subset of the range of classes extant at Gazel. La Caune de Belvis also exhibits a subset of the Gazel repertoire, one that is even more limited than that of Canecaude I, except in the case of procurement classes. Canecaude I and La Caune de Belvis share the same repertoire in some cases (spear, needle, smoother, mineral bead, tube, and cut-out) , but differ in others .

compared with the seven classes at Gazel and the five at Canecaude I. When the actual frequencies of these objects are taken into account, however, it becomes clear that there is a relatively high concentration of objects involved in production at La Caune de Belvis; these objects are primarily needles. Needles are present in all stages of manufacture, use, and discard, and there appears to have been both the production of needles, as well as activities in which the needles were employed, at La Caune de Belvis. The number of object classes having to do with production is rather limited at La Caune de Belvis, more numerous at Canecaude I, and the most numerous at La Grotte Gazel. Enigmatic object classes are rather limited at La Caune de Belvis, where mineral bead, tube, cut-out , and engraved bone fragment are the four classes present. Canecaude I exhibits three of these, including mineral bead, tube , and cut-out, along with additional classes including pierced bone, teeth, and shell. La Grotte Gazel yields the broadest range of object classes , including engraved cobble, engraved rod, antler bead, and pendant, as well as those present at Canecaude I and La Caune de Belvis.

The greatest difference between Canecaude I and La Caune de Belvis occurs in the classes of procurement. Canecaude I has spear , spear throwers, and half-round rods. La Caune de Belvis has spears, but no spear throwers or half-round rods. The majority of procurement implements at La Caune de Belvis are harpoons, points, and darts. The difference can be explained in large part by referring to the faunal evidence. Reindeer dominates the faunal assemblage at Canecaude I, although fish and bird are present, along with a variety of other fauna. The fauna at La Caune de Bel vis is dominated by ibex, and a variety of fish, birds, and other fauna as well. The difference in tool repertoire is most likely one of difference in game. Reindeer, spear points, and spear throwers are associated. Ibex are associated with barbed points and darts. The small hunting points at La Caune de Belvis would have been appropriate for catching fish and birds. Their absence at Canecaude I may either be explained by the use of different kinds of implements (nets and traps) that may not have been preserved in the archaeological record or, as is always possible in cave sites, by their introduction via non-human agents.

For the following analysis of the assemblages, I modified the assemblage total, removing unelaborated bone and antler fragments (worked debris), diverse, and indeterminate categories. These are three classes for which it was not possible to make the distinction between procurement, production, or enigmatic. The Gazel assemblage total is adjusted to 761 (from 1282) items, Canecaude I to 128 (from 152) items, and La Caune de Belvis to 162 (from 195) items. Table 6.1 illustrates the three subdivisions of procurement, production, and enigmatic for the three assemblages. There is a significant difference between the three assemblages in terms of the three categories of bone and antler use (Chi-square test with a 95 percent confidence level). These results offer a complementary perspective on the assemblage make-up and activities undertaken at each site . Each of the bone and antler assemblages is dominated by procurement implements. Canecaude I is primarily procurement, with only a very mall percentage of the assemblage devoted to either production or enigmatic objects. Gazel and La Caune de Belvis exhibit more evidence that production as well as procurement activities took place on-site, and enigmatic objects, although still in the minority, contribute more to the overall

In the realm of production, Canecaude I and La Caune de Belvis share the classes of needle and smoother. The only other object class made of bone and/or antler and associated with production at La Caune de Belvis is the chisel. These three classes comprise a limited repertoire, especially when

Table 6.1

Bone and Antler Industry Assemblage Division into the Classes of Procurement, Production, and Enigmatic La Grotte Gaze!

Canecaude I

La Caune de Belvis

Procurement

59.3 %

82.8%

55.6%

Production

22.3%

10.2%

32.1%

Enigmatic

18.4%

7.0%

12.4%

761

128

162

Total in Assemblage

82

VARIABILITY AND CONTEXT OF MAGDALENIAN VISUAL IMAGERY

Table 6.2 Bone and Antler Industry Assemblage Elaboration La Grotte Gaze!

Canecaude I

La Caune de Bel vis

Elaborated

37.4 %

15. 1%

14.9%

Unelaborated

41.7 %

61 .2%

60.5 %

Unfinished

20 .9%

23 .7%

24 .6%

1282

152

195

Total number of items

assemblages . These percentage subdivisions of the assemblages, when taken together with what we already know of each site from faunal evidence, site size, and location, can be integrated as follows.

and enigmatic objects are much more prevalent than at either of the other sites. A greater number of different individuals not only deposited beads and pierced objects, but at least one individual was making antler beads on site. What appears to have been a rabbit tooth necklace was left behind. Gaze} may have been an aggregation site where a relatively diverse group or groups gathered, hunted, produced objects, and worked materials, and where a greater proportion of items used in visual imagery and communication was produced, used, and deposited . It has been suggested that the organizational and communication tensions at such aggregation sites might have made elaborated portable objects more useful in the sense of mediating interactions between individuals and maintaining or establishing order, group cohesion, and cooperation.

Canecaude I is a special purpose reindeer hunting camp. The bone and antler assemblage was predominantly dedicated to procurement activities (reindeer hunting), and very little attention was given to production or activities associated with enigmatic objects. Although Canecaude I had a more limited range of different classes of objects used for procurement, the assemblage is heavily dominated by procurement tools, and, in light of the faunal assemblage, work was concentrated on a limited range of procurement tasks associated with reindeer hunting. Some of the procurement tools may have been fabricated in place, as there is some evidence for the working of bone and antler in the assemblage. The lithic assemblage is proportionately quite minimal, with roughly 5,000 pieces compared to over 100,000 at La Caune de Belvis from the same volume of sediment (Price 1997). In the case of pierced objects, it appears as if a very small number of different individuals had passed through the camp and lost or deposited items of personal adornment. It is possible that one individual began crafting a cut-out on a horse hyoid bone. Alternatively, this practically finished "enigmatic" piece could have been part of a personal kit carried with someone and lost at the site. It is certain that an individual lost or left behind a fragment of an elaborately engraved mammoth spear thrower, perhaps broken in use and abandoned.

Elaborated/Unelaborated/Unfinished Table 6.2 shows the degree to which each bone and antler industry is elaborated at the three sites. The proportion of elaboration is not significantly different between the two smaller sites (Chi-square test at the 95 percent confidence level). The proportion of assemblage elaboration between these two sites and La Grotte Gazel is significantly different (Chi-square test at the 95 percent confidence level). The most notable result of this comparison is the degree to which Canecaude I and La Caune de Belvis resemble one another. When the non-lithic assemblage is taken as a whole, 15 out of 100 objects are elaborated, 60 out of 100 are unelaborated, and 24 out of 100 are unfinished . By contrast, La Grotte Gazel exhibits more than twice the degree of overall assemblage elaboration. Nearly as many objects are elaborated as are unelaborated (37 percent to 41.7 percent). This greater frequency of elaboration reflects the higher frequency of social interactions that would have been reinforced by visual display, supporting the model of more intensive social action and display at aggregation sites.

At La Caune de Belvis, there was more of a balance between procurement and production activities for which bone and antler were crafted. Enigmatic objects played a greater role than at Canecaude I. Given the significant difference in size between the bone and antler assemblage of La Caune de Bel vis ( 195 pieces) and Gazel ( 1282), the similarity in lithic assemblage size between the two sites (> 100,000 pieces) is striking. A relatively small group of people lived for the months between late spring and early fall at La Caune de Belvis and carried out, among other activities, ibex hunting and processing, needle production and use, and the exploitation of a wide variety of ecozones. They were not making or depositing pierced teeth or shell, but they were engraving images on bone and antler implements. One such implement is the tube, a section of bird long bone; one of its functions may have been to carry small objects such as needles. At Gaze}, procurement played a greater role than production,

Imagery What is the nature of the images and elaboration on objects at each site? Although grouped together in the broad category of elaboration , pierced objects may have had different social contexts than engraved objects. Within the category of engraved objects, items of procurement may have had different uses and meanings than flat fragments of bone. Some of these differences are explored below. 83

PRICE

Pierced Objects

Engraved Objects

Pierced objects show an uneven distribution among the three assemblages. When the total number of pierced items are standardized by the volume of sediments excavated, Gazel appears to yield more than twice as many as either of the other sites. The volume of sediments at any given site, and even from location to location within a site, can be affected by differential rates of both erosion and accumulation, so it is interesting to evaluate the frequency of piercing or engraving for the total number of objects retrieved during excavation. These measurements, frequency of pierced objects by volume and by total assemblage size, are presented in Table 6.3.

A similar standardization of the frequency of engraved objects, illustrated in Table 6.4, shows somewhat different results. The sites show increasing frequency of engraving, from Canecaude I to La Caune de Belvis and, finally, to La Grotte Gazel. There are significant differences in terms of the frequency of engraved objects (Chi-square test, 95 percent level of confidence). The Brainerd-Robinson test further reveals that the assemblages of Canecaude I and La Caune de Belvis are, in fact, quite similar (BR= 195), that La Grotte Gazel and La Caune de Belvis are fairly similar (BR=l81), and that more difference lies between La Grotte Gaze} and Canecaude I (BR=167).

Neither the Chi-square test (95 percent confidence level) nor the Brainerd-Robinson test show significant difference between the three assemblages in terms of the frequency of pierced objects for the volume of sediment excavated. This class of object is rare in each assemblage, and the mall sample size limits the kinds of que tions that can be posed of this data. The actual make-up of each pierced assemblage yield more information. La Grotte Gazel yielded teeth from an unidentified species of leporid ( 18) and from fox (5), cervid (7), reindeer (3), bear (1), wildcat (1), and horse (1). In addition, two fragments of stalactite were fashioned to resemble small pierced teeth. Several of these pierced teeth had also been engraved at the root with single or multiple engraved lines. Of the 39 pierced shells, 32 of these are from the Atlantic species Littorina obtusata. Many different materials were formed into beads, including antler (22), jet (5), tubular bone (5), and reindeer sesamoid bone (4). 1Series of similar items, such as 18 pierced leporid incisors found in association at La Grotte Gazel, suggests that an entire necklace or other object of adornment was deposited at once. The 22 antler bead blanks and finished antler beads suggest onsite production.

Taking sample size into account, La Grotte Gazel has a significantly higher frequency of engraved objects. Canecaude I and La Caune de Belvis appear similar with comparable rates of both pierced and engraved objects. Canecaude I has slightly more pierced objects than La Caune de Belvis, although this category of object is rare at both sites. La Caune de Bel vis has a greater frequency of engraving than Canecaude I, but this is a relatively low frequency in both cases. In these terms, perhaps the most one can say is that these objects are present at both the hunting camp, a semi-private setting, and at a seasonal base camp, a public setting. The characteristics of elaboration on portable objects are often presented as a general phenomenon occurring throughout the modeled Magdalenian "culture," regardless of sitespecific contexts. When placed within the site-specific contexts of La Grotte Gazel, Canecaude I, and La Caune de Belvis, differences within this body of materials become interesting and informative. A thorough description and analysis of the imagery is beyond the scope of this chapter (cf. Price 1997). Some of the results are presented below.

The Magdalenian deposits of Canecaude I yielded one pierced cervid tooth, one pierced horse tooth, three pierced shells, and a single jet bead. La Caune de Belvis was devoid of pierced teeth and shell, yielding one jet bead and one bead made from hematite. No evidence of bead production is present at either site.

A total of 49 different design elements were found on portable objects at La Grotte Gazel, compared to 14 at Canecaude I and 20 at La Caune de Belvis. 2 At first glance, La Grotte Gazel appears to have the most design elements and, therefore, the greatest diversity. This kind of disparity has been

Table 6.3 Frequency of Pierced Objects per Cubic Meter of Sediment Excavated and by Assemblage Size

Total # Pierced Objects

La Grotte Gaze!

Canecaude I

La Caune de Belvis

132

7

4

(m 3)

22

3

3

Total pierced per m 3

6

2.3

1.3

Assemblage size

1282

152

195

% of Total Assemblage

10.3%

4 .6%

2.1%

Volume

84

VARIABILITY AND CONTEXT OF MAGDALENIAN

VISUAL IMAGERY

Table 6.4 Frequency of Engraved Objects per Cubic Meter of Sediment Excavated and by Assemblage Size La Grotte Gazel

Canecaude I

La Caune de Belvi s

352

17

27

22

3

3

Engraved per m

16

5.7

9

Assemblage size

1282

152

195

27.5%

11.2%

13.8%

Total # Engraved Objects Volume (m3) 3

% of Total Assemblage

interpreted as support for the interpretation that design elements represent group identity, and the occurrence of a larger number of different design elements at a site reflect the gathering together of a greater number of groups, as at an aggregation site. This may, in fact, be true. After all, the site of La Grotte Gazel has yielded a much greater concentration of materials, and this may have been the result of more intensive use of the site by a greater diversity of people.

1979). The horse head engraved and sculpted on the distal end of spear throwers, as found at Gazel, has also been found at sites in the Dordogne (e.g., Laugerie-Basse , La Madeleine), in the Haut Garonne (Gourdan), in the Tarn-et-Garonne (Bruniquel), in the Hautes-Pyrenees (lsturitz), and in Switzerland (Kesslerloch) (Delluc and Delluc 1987). Perhaps more interesting are those that appear unique to particular sites. For example, Canecaude shares most of its 14 design elements with other sites in the region (or outside of the region), except for the sculpted mammoth on the distal end of a spear thrower . La Grotte Gazel exhibits many design elements not found in the region, but found in regions such as the Ariege (e.g., Chollot 1964) and north coastal Spain (Barandiaran 1973; Corch6n Rodrfguez) (for more detail see Price 1997). Elements that appear to be truly unique to La Grotte Gazel include the discontinuous sinuous groove (design element 15) found most often on spears (Bertrand 1995), the sculpted antler bead, and the complex of associated design elements found on the smoothed pendant (design elements 39, 46, and 85). La Caune de Belvis yielded many design elements shared within the Western Languedoc Roussillon and many that are characteristic of the Ariege. There are no uniquely occurring design elements at Belvis when the scope of analysis is expanded to include neighboring regions . However, within the context of the smaller region of this study, Belvis design shows a unique tendency towards stylization in animal imagery, as well as the representation or evocation of a different fauna, including birds and fish. The elongated figure that Sacchi ( 1986: 187) has likened to a waping bird is vaguely reminiscent of other late Magdalenian bird representations (Breuil and Begouen 1937; Glory 1965; Vayson de Pradenne 1934), yet the actual realization of this particular figure is as unique in its abstraction as is the mammoth spear thrower from Canecaude.

When sample size is considered using a simple index, the results are reversed . When the number of design elements are divided by the total assemblage size, the index for La Grotte Gazel is 10, 61 for Canecaude I, and 69 for La Caune de Belvis. This measurement could be interpreted as indicating that, for the amount of activity related to the production, use, and discard of bone and antler industry at each site, a relatively greater diversity of design elements was employed at La Caune de Belvis and at Canecaude I than at La , Grotte Gazel. Had the sample size been greater at either of these smaller sites, they might have yielded a larger number of different design elements. But does this mean that a multipurpose seasonal base camp, had it been more repeatedly used by the hypothetical microband or extended family unit, would have yielded a greater diversity of different designs? This is not clear. With the data at hand, it is not possible to separate out roughly equivalent chunks of time from each assemblage, so we are left with intriguing possibilities.

In number and in frequency, the majority of design elements are geometric. In the region as a whole and including these three sites, the only object class from procurement and production categories that shows animal imagery is the spear thrower. All other objects that are presumed to be implements, such as spears or smoothers, show geometric design. With the exception of the spear thrower, animal imagery is limited to enigmatic items such as cut-outs or pendants. The animals depicted at Gazel are horse (6), ibex (1), bear (1), and aurochs ( 1). Animals depicted at Canecaude are limited to a single mammoth. La Caune de Belvis yielded engravings of one example each of ibex, bird, horse, and fish.

There is only a handful of designs that appear to have been uniquely associated with a single site and assemblage. Canecaude and Belvis are distinguished by their treatment of subjects (a mammoth and a bird, respectively) that are known elsewhere but executed distinctively . If the occurrence of similar objects, similar design elements, and styles across space implies behavioral linkages in the past, whether by migration, by trading, or by imitation, then we are left to wonder at the significance of unique occurrences. The possibility that unique execution or design represents individual

Many of the design elements identified in this region have also been observed in neighboring regions. For example, the horse head cut-outs found at Gazel fit stylistically with a collection of similar horse heads from the Ariege (e.g., Sieveking 85

PRICE

aggregation , if, indeed, La Grotte Gazel was such a site. Furthermore, and somewhat paradoxically, this form of symboling was not part of the public domain . It is interesting to note that the portable and pierced objects at La Grotte Gazel are not only plentiful and diverse, but they are concentrated in the habitation deposits and even appear to have been produced there . The wall imagery was of limited access (private) , while the portable imagery wa generally accessible and visible (in the public domain).

efforts at change or expressing new identities or ideas must be considered. Portable imagery is found among the habitation remains at all three sites. This includes the public contexts of aggregation and multipurpose base camp and the semi-private context of the hunting camp. Production of pierced objects is limited to La Grotte Gazel. Production of engravings on items of procurement, production , and even enigmatic objects could have occurred in all three contexts, as there is evidence of bone and antler working at all three locales. Animal imagery is limited to spear throwers and enigmatic objects . The case is different for stationary imagery .

Stationary Imagery La Grotte Cazel The most widely accepted Magdalenian period wall engravings are located in the cave of La Grotte Gazel. These are dated on the basis of stylistic and contextual criteria . Despite testing in the area nearest the engravings, no material culture has been found in direct association with the elaborated walls. Sacchi (1986: 142) has recovered an engraved pebble in the habitation deposits in the front portion of the cave that he judges to have been made in a stylistically similar period. The pebble was found in association with deposits that have been stylistically dated to the Middle Magdalenian (Style IV of Leroi-Gourhan 1958) and radiocarbon dated to around 15,000 B.P. The horse imagery found on portable objects in the Middle Magdalenian deposits accords well with the horse imagery on the cave walls. The ibex theme and the portrayal in profile-with, for example, selected anatomical details of the horns and facial and body hair-match the material culture and stylistic representation known from this region during this particular time in prehistory.

Fornols-Haut

In the mountainous region between La Grotte Gazel and La Caune de Belvis at 650 m altitude lies the isolated schist outcrop known as Fornols-Haut (Sacchi et al. 1988; Sacchi 1993a, 1993b). Today, the outcrop is less than two meters high and les than three meters wide and sits on the exposed northeast facing side of a mountain with a dramatic panoramic view. The schist is covered with engravings of similar subject and style to those at La Grotte Gazel. Unfortunately , there are no associated deposits to provide clues as to the context of image production and viewing. The ibex images, in particular, are reminiscent of the images at La Grotte Gazel, as is the predominance of incomplete and diminutive animal subjects, most often shown in profile (see Figure 6.4). While horse and ibex are the species represented at La Grotte Gazel, Fornols has several partial ibex, as well as a vulture and a waterfowl. Geometric markings include limited series of cross-hatched lines and other angular lines in various associations. These engravings are so small and subtle that they can only be perceived by one or two people in close proximity to the rock face. Excavations in the immediate vicinity of the rock have failed to reveal Magdalenian period deposits. Interpretation: Ethnography of Communication

If we consider Fornols-Haut to have been produced roughly contemporaneously with La Grotte Gazel, examining the similarities and contrasts offers a way of approaching a plausible context for their creation. Although it is difficult to say how many people might have gathered around the Fornols-Haut rock face, it is obvious that only a few at a time could have actually seen the engravings, and only then if they got within a few feet and looked very carefully. The round room of La Grotte Gazel allows a group of up to ten people to stand in the center, but only one or two at a time could have mounted the slick, sloping walls to view the engravings. The engraved scene just beyond the round room is located in such a narrow passageway that only two or three people crouching or lying down could possibly view the composition . These three locations each create a different effect. Fornols-Haut is exposed to daylight and a wide open expanse in all directions. Within the surrounding terrain, the rock surface appears rather small. By contrast, both La Grotte Gaze} locations are deep within a cave in the dark. As one travels on hands and knees through the narrow passageways of La Grotte Gazel, arrival at the round room brings a sense of opening, although this space is still relatively limited and dark . Crawling up into the very narrow passage of the farthest engraving places the prone viewer

The seven engraved loci of La Grotte Gazel are clustered in and just beyond an open rounded room located over 100 meters into a part of the cave far removed from the habitation areas at the front rooms of the cave. Representations include horse and ibex, alone and in groups (see Figure 6.3). There is one cluster or scene involving two complete ibex in profile, at least two more ibex suggested by partial profiles, and a few associated geometric markings. These include an oval situated on the back of one ibex, a series of parallel nearly vertical lines, and a few isolated lines. A nearly complete ibex, in profile, is located in the round room, where partial, sketchy profiles of the head and neck of at least seven horses are evident with careful and directed inspection. The engravings are diminutive, difficult to gain access to, and, most important, visible to only one or two people at a time, and with difficulty even then. One must either mount a steep and slick calcite flow in the round room or lie prone in the narrow passageway to view the La Grotte Gazel images. These observations suggest that there was some form of visual image making that may have occurred only during times of 86

VARlABILITY AND CONTEXT OF MAGDALENIAN VISUAL IMAGERY

' ..

0

5cm

\ \.

.'' . \

f \

\'

I

Figure 6.4-Tracings

0

10cm

of a selection of the rock-face engravings from different surfaces at Fomols-Haut (after Sacchi 1993b; Sacchi et al. 1988).

87

PRICE

in a very restricted setting; the engraving, although small, occupies a major portion of one wall. An ethnography of communication interpretation suggests that La Grotte Gazel settings are both of relatively limited access in terms of the number of people who could have experienced them, whether in the role of creating or simply of viewing. Of the two settings at La Grotte Gazel, neither the narrow passageway nor the round room, evokes large public group gatherings or showy displays and interactions. Rather , a private experience, such as a spirit quest, or a small group experience, such as an initiation ceremony, are brought to mind . These analogies are apt whether one imagine s a single episode of creation and significance or a series of repeated visitations .

The engravings are diminutive, difficult to gain access to, and, most important, visible to only one or two people at a time. These observations suggest that there was some form of visual symboling that occurred in association with the context of aggregation. Furthermore, and omewhat paradoxically, this form of symboling was not part of the public domain. It is interesting to note that the portable and pierced objects at La Grotte Gazel are not only plentiful and diverse but are concentrated in the habitation deposits, where they appear to have been produced. The wall art was of limited access, while the portable art was fully in the public domain . La Grotte Gazel appears to have been the location of ocial interactions that involved a relatively high degree of all archaeologically known modes of visual elaboration and communication asso ciated with the Magdalenian period . Canecaude I was dedicated to the reindeer hunt , with a relatively low frequency of assemblage elaboration in the form of geometric engravings on procurement and production tool and very few pierced teeth, shell, and minerals . The geomet ric designs are found throughout the region; however, the sculpted mammoth spear thrower is unique.

The apparent similarities in style and subject of the FornolsHaut and the La Grotte Gazel engravings become even more interesting given the contrastive physical settings. Both share the ibex theme . Both have been engraved on a rather diminu tive scale. The subjects are depicted in profile and rarely com pletely rendered. A few simple geometric markings occur in proximity to the ibex . These engravings are presented out in the daylight on an open mountain slope and deep inside a dark passageway within a large cave system.

La Caune de Belvis is a multipurpose base camp with an emphasis not only on ibex hunting, but also on the exploitation of a variety of game, including birds and fish. The production of needles and tubes with which to carry them is notable. No pierced teeth or shell were recovered. Only two examples of mineral beads were found. The range of animal imagery is notable at this site, especially the bird and fish on cut-outs and pendants.

There are two major differences in the engravings. The ibex theme is numerically overshadowed by horse at La Grotte Gazel. This theme is notably prevalent among the portable imagery in the nearby habitation deposits. The ibex theme is complemented by bird subjects at Fornols-Haut. It is likely that ibex were, in fact, present on the landscape in both locations . The prevalence of reindeer and horse in the faunal assemblage of La Grotte Gazel, and the lack of reindeer in visual imagery, obviates the interpretation that the imagery is focused on the most economically important species. The message content, at least at La Grotte Gazel, privileged ibex and horse, possibly exactly as we in Anglo-American culture today tend to imagine them-as economic commodities-or, more likely, as the set of complex metaphorical and oppositional ideological associations that they evoked to the members of this particular culture of the deep past.

The Canecaude I assemblage does not exhibit much evidence for the elaboration of individuals in the form of pierced objects. The rarity of these objects is in accord with the apparently limited occupation of the site in terms of time spent there, the small group size, and the limited kinds of activities carried out. Not enough people from different groups were there long enough to have deposited such items, nor were they there under social or political contexts that called for the making or displaying of such items, aside from the individualized spear thrower and a few beads.

SUMMARY AND CONCLUSIONS

La Caune de Belvis, on the other hand, appears to have been occupied by a group or various groups of people for longer periods of time. They appear to have undertaken a wider range of activities than did the occupants of Canecaude. The inferred social and political context is one of a wider range of activities, but no more cause for the visual signaling of social and personal identity.

This study used the local scale of individual sites within a regional model for hunter-gatherer behavior in order to develop context. Differences in imagery and elaboration at each site are placed within different behavioral and social contexts. Gazel is a large , multipurpose aggregation site. Production of pierced objects, beads, engraved objects, and imagery on walls occurred on-site. Although the imagery in several cases is part of an extra-regional design system, in other cases the imagery diverges to express unique elements. The engraved wall compositions are far removed from the habitation areas in the front sections of the cave. Excavations have not revealed any associated cultural materials. Representations include profiles of horse, ibex, and a few associated signs .

Fornols-Haut, the isolated schist outcrop with no associated deposits, shows similarity to Gazel in the ibex theme and execution; however, it differs with the two birds and with its setting. Although it is difficult to say how many people might have gathered around this rock face, it is obvious that only a few could have actually seen the engravings . As at Gazel, the imagery on stationary surfaces appears to have taken place in a private or semi-private context. 88

VARIABILITY AND CONTEXT OF MAGDALENIAN VISUAL IMAGERY

Where La Grotte Gazel wall engravings are loosely associated with rich habitation deposits and aggregation contexts, the Fornols-Haut engravings are isolated. Where Gazel is enclosed, restricted, and dark , Fornols-Haut is open and light. The similarities lie in the style and scale of the imagery. Animal subjects are presented in partial profile and are often incomplete. The engravings are small, and the viewer must approach the panel closer than an arm's length in order to view them. This in itself suggests that these engravings, in two seemingly different physical contexts, were nevertheless created in private or semi -private settings of limited social access.

Bertrand, A., 1995. Les armatures de sagaies Magda/,eniennesen matiere dure animale dans Les Pyrenees. Nouveau Doctorat, Art et Archeologie These. Universite de Paris I.

The study of local context holds great potential for gaining deeper appreciation for the social settings within which imagery was created. This is limited only by the rarity and slow rate of excavation and analysis of the kind necessary to build context. Materials from these three sites represent a life time of dedicated work. This kind of baseline study produces the kind of information necessary for reflecting upon the nature of elaboration in other regions and the evolution of symbols and imagery .

Binford , L.R., 1980. Willow Smoke and dogs ' tails: Hunter-gatherer settlement systems and archaeological site formation. American Antiquity 45:4-20.

Binford, L.R., 1977. Forty-seven trips: A case study in the character of archaeological formation processes . In Stone Tools as Cultural Markers, edited by R. Wright, pp. 24-36. Canberra: Australian Institute of Aboriginal Studies . Binford, L.R., 1978. Nunamiut Ethnoarchaeology. New York: Academic Press.

Breuil, H., and H. Begouen, 1937. Quelques oiseaux inedits OU meconnus de )'art prehistorique. Congres Prehistorique de France, x11eme Session , ToulouseFoix:475-488. Chollot , M., 1964. Collection Piette : Art Mobilier Prehistorique. Paris : Editions des Musees Nationaux.

ACKNOWLEDGMENTS Conkey, M.W., 1978. An Anal ysis of Design Structure: Variability Among Magdalenian Engraved Bones from Northcoastal Spain. Ph.D. Dissertation. Department of Anthropology, University of Chicago.

In conducting this study, I relied heavily on the accumulated work, experience , and hospitality of Dr. Dominique Sacchi, who excavated the assemblages from La Grotte Gazel, Canecaude I, and La Caune de Belvis. The study was made possible by funding from a General Grant from the L.S.B. Leakey Foundation, the Stahl Endowment Fund, a Humanities Graduate Research Grant and Fellowship from the Regents of the University of California, and a series of Lowie Awards for Anthropological Research administered by the Department of Anthropology at the University of California, Berkeley. My friend and colleague Dr. Gail Larsen Peterkin inspired me with her dedication and determination in the completion of this chapter, as well as of the entire volume.

Conkey, M.W., 1980a. The identification of hunter - gatherer aggregation sites: The case of Altamira. Current Anthropolog y 21 :609-630. Conkey, M.W., 1980b. Context, structure, and efficacy in Paleolithic art and design. In Symbol as Sense , edited by M.L. Foster and S. Brandes, pp . 11- 19. New York: Academic Press. Conkey, M .W., 1985 . Ritual communication , social elaboration, and the variable trajectories of Paleolithic material culture. In Prehistoric Hunter-Gatherers: The Emergence of Cultural Complexity , edited by T.D. Price and J.A. Brown, pp. 299-323. Orlando: Academic Press.

NOTES 1. Excavations and basic description and reporting of objects recovered from La Grotte Gazel were ongoing at the time the present study was prepared; hence subsequent publications may present different assemblage information.

Conkey, M .W., 1987. Interpretive problems in hunter - gatherer regional studies: Some thoughts on the European Upper Paleolithic. In The Pleistocene Old World: Regional Perspectives, edited by 0 . Soffer, pp. 63-78. New York: Plenum Press.

2. Referring to Figure 6.2, design elements found at Gazel include: 1-6, 8,9,12-16, 18-21, 24-30, 32, 39-41, 43, 44, 46, 48, 50, 56-59, 61-63, 66, 67, 70, 81-86; at Canecaude I: 1, 3-5, 8, 9, 14, 22, 25, 56, 58, 60, 80, 81; and at La Caune de Belvis: 1, 3, 8-12, 17, 19, 25 , 32, 33, 44, 56, 61, 63, 65, 69, 71, 72 .

Corch6n Rodrfguez, M.S., 1994. Arte mobiliar e industria osea Solutrense en la cornisa Cantabrica. In Monografico el Solutrense en la Peninsula Iberica: Fervedes 1, pp. 131-148 . Villalba: Museo de Prehistoria e Arqueoloxia de Villalba.

REFERENCES Barandiaran, I., 1973. Arte Mueble del Paleolitico Cantabrico. Monografias Arqueol6gicas 14. Zaragosa: Monografias Arqueol6gicas.

Davidson, I., 1989. Freedom of information: Aspects of art and society in Western Europe during the last Ice Age. 89

PRICE

In Animals Into Art, edited by Howard Morphy, pp. 440-456. London: Allen and Unwin.

Ethnography of Communication, edited by J.J. Gumperz and D. Hymes, pp. 35-71. New York: Holt, Rhinehart, and Winston.

Delluc, B., and G. Delluc, 1987. Le decor des objets utilitaires du Paleolithique superieur. In L'Art des Objets au Paleolithique. Tome 2: Les Voies de la Recherche, edited by Jean Clottes, pp . 39-72. Foix: Ministere de la Culture.

Jochim, M., 1983. Paleolithic Cave art in ecological perspective. In Hunter-Gatherer Economy in Prehistory, edited by G.N. Bailey, pp. 212-219. Cambridge: Cambridge University Press.

Desdemaines-Hugon, C., 1994. Art et Fonction au Paleolithique. Les Decors et Biseaux des Ciseaux du Magdalenien VI de Trois Sites en Aquitaine: la Madeleine (Dordogne), le Morin (Gironde) et Rochereil (Dordogne). D.E.A . d' Anthropologie. Institut du Quatemaire, Universite de Bordeaux I.

Julien, M., 1982. Les Harpons Magdaleniens. xvneme Supplement a Gallia Prehistoire . Paris: Centre National de la Recherche Scientifique. Leone , M., 1982. Some opinions about recovering mind. American Antiquity 47 :742-760 .

Docktor, D., 1983. The significance of rock art setting in the interpretation of form and function: Preliminary investi gation of two Yokuts rock art sites in California. American Indian Rock Art 9:63-71.

Leroi -Gourhan, A., 1958. Repartition et groupement des animaux dan l' art parietal Paleolithique . Bulletin de la Societ e Prehistorique Franfaise 55:515- 528 .

Domenech-Faus, E., 1995. Remarques a propo de matiere premiere de Gazel. Personal communication.

Longacre, W., 1970. Archaeology as Anthropology: A Case Study. Anthropological Papers 17. Tucson: University of Arizona.

Durkheim, E., 1915. The Elementary Forms of Religious life . New York: The Free Press.

Meroc, L., and G. Simonnet, 1965. La grotte omee de Gazel a Salleles-Cabardes. Archeologia 5:60-67.

Firth, R.W., 1973. Symbols Public and Private. Ithaca: Cornell University Press.

Pfeiffer, J.E., 1982. The Creative Explosion: An Inquiry Into the Origins of Art and Religion. Ithaca: Cornell University Press.

Fontana, L., 1994. La faune de Canecaude. In Les Civilisations Mediterraneennes: Les Facies Leptolithiques du Bassin Mediterraneen Nord-Occidental, Milieux Naturels et Culturels, pp. 45-46. Congres Prehistorique de France, :XXIveme Session, Carcassonne.

Price, H.A., 1997. Context and Variation: Upper Paleolithic Visual Imagery in the Western Languedoc-Roussillon of Southern France. Ph.D. Dissertation. Department of Anthropology, University of California, Berkeley.

Fontana, L., 1995a. Les Mammifieres des Couches Magdaleniennes de la Grotte Gazel: Rapport Preliminaire. Manuscript on file with the author.

Price, H.A., 1999. The Kashaya Petroglyph Project: Rock Art in Parklands in Kashaya Territory. Report prepared for California State Parks Silverado District, Sonoma, California.

Fontana, L., 1995b. A Propos de Belvis, de Gazel, et de Canecaude. Manuscript on file with the author.

Sacchi, D ., 1986. le Paleolithique Superieur du Languedoc Occidental et du Roussillon. XXIeme Supplement a Gallia Prehistoire. Paris: Editions du Centre National de la Recherche Scientifique.

Friedrich, M.H., 1970. Design structure and social interaction. American Antiquity 35:332-343. Gamble, C., 1982. Interaction and alliance in Paleolithic society. Man 17:92-107. Gamble, C., 1991. The social context for European Paleolithic art. Proceedings of the Prehistoric Society 57 :3-16.

Sacchi, D., 1988. Remarques generates sur les facies Magdaleniens du Bassin de 1'Aude et du Roussillon. Espacio, Tiempo y Forma. Serie I: Prehistoria. Homenaje al Profesor Eduardo Ripoll Perello 1:177- 195.

Glory, A., 1965. L'oiseau de la Pasiega (Espagne). In Congres Prehistorique de France, xv1e1neSession, Monaco:596-607.

Sacchi, D., 1993. Un rocher grave Paleolithique dans les Pyrenees fran~aises. lntemational Newsletter on Rock Art 4:14-15.

Hodder, I., 1986. Reading the Past. Cambridge: Cambridge University Press.

Sacchi, D., J. Abelanet, and J.L. Brule, 1988. Un temoin de I' art Paleolithique de plein air en Roussillon: Le rocher de Fomols-Haut. In Prehistoria I Arqueologia de la Conca def Segre: Homenaje al Profesor Dr. Joan Maluquer de Motes , pp. 37-42 . Puigcerda: Institut d'Estudis Ceretans .

Hymes, D., 1972. Models of the interaction of language and social life. In Directions in Sociolinguistics : The 90

VARIABILITYAND CONTEXT OF MAGDALENIAN VISUAL IMAGERY

Sauvet , G., 1987. Les signes dans l'art mobilier. In L'Art des Objets au Paleolithique, edited by J. Clottes, pp. 83- 98. Foix: Actes des Colloques de la Direction du Patrimoine.

Rock Art Symposium of Man .

of the San Diego Museum

Weniger, G.C. , 1987. Magdalenian settlement pattern and subsistence in Central Europe. In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer, pp . 201-215. New York: Plenum Press .

Sieveking , A., 1979. Style and regional grouping in Magdalenian cave art. Bulletin of the Institute of Archaeology 16:95-109.

White, R., 1985. Upper Paleolithic Land Use in the Perigord: A Topographic Approach to Subsistence and Settlement. British Archaeological Reports International Series 253. Oxford: British Archaeological Reports.

Soffer, 0., 1985. The Upper Paleolithic of the Central Russian Plain . Orlando: Academic Press. Vayson de Pradenne , A., 1934. Les figurations d' oiseaux dans l'art quatemaire . /pek 9:3- 17.

Whitley, D.S., 1987. Socioreligious context and rock art in east-central California. Journal of Anthropological Archaeolog y 6: 159-188 .

Weinberger , G., 1980. Cupules and Their Context: Some Southern Valley Yokuts Site Examples. Paper presented at the Annual Meeting of the Society for California Archaeology.

Wobst, W.M., I 990. Afterword: Mini time and megaspace in the Paleolithic at 18K and otherwise. In The World at 18,000 B.P. Volume /: High Latitudes , edited by 0. Soffer and C. Gamble, pp. 331- 343. London : Unwin Hyman .

Weinberger , G., 1981. "Public " and "Private" Rock Art Sites in Tulare County. Paper presented at the Annual

91

7 Les groupes humains au Tardiglaciaire dans le Bassin parisien: diverses voies pour one approche regionale Boris Valentin (U niversite de Paris I)

ABSTRACT Cet article montre comment un projet initialement oriente vers la reconnaissance de l'identite culturelle specifique d'un groupe regional du Magdalenien s'est reoriente vers une approche diachronique prenant en compte Jes diverses traditions techniques attestees au cours du Tardiglaciaire dans le Bassin parisien . Les difficultes relatives a la seriation fine des evenements au Paleolithique superieur nous ont conduit a privilegier une approche technologique comparative sur le temps long afin de cerner, par contraste , 1'originalite de chaque tradition. Du modele evolutif ainsi construit pour une region limitee, nous esperons deduire des criteres facilitant les comparaisons avec d'autres territoires . La recherche des decalages chronologiques entre les principaux evenements paleohistoriques, d' une region a l'autre , est une autre voie pour identifier des specificites culturelles locales . This chapter shows how a project initially oriented towards a recognition of the specific cultural identity of a regional group of Magdalenians was reoriented towards a diachronic approach that took into account the diverse technical traditions evidenced during the course of the Tardiglacial in the Paris Basin. Difficulties relating to the detailed, fine-scale seriation of Upper Paleolithic events led the author to privilege a comparative technological approach in his attempt to discern the originality of each tradition . From an evolutionary model constructed for a limited region he deduced criteria to facilitate comparisons with other regions. The research of chronological disparities between the main events of paleohistory , from one region to another , was another method he used to identify specific local cultures. INTRODUCTION Nos recherches sur le Tardiglaciaire portent actuellement sur le Bassin parisien. Dans ce cadre geographique, nous souhaitions, comme beaucoup d'autres prehistoriens, pouvoir identifier des territoires coutumiers et cerner, a cette echelle, l' identite culturelle specifique des groupes prehistoriques. Ces ambitions se heurtent a plusieurs limites methodologiques, en depit du nombre et de la qualite des donnees archeologiques reunies depuis 30 ans, notamment sur le Magdalenien superieur . L'objet de cet article est d' eclairer ces limites avant de montrer quelle piste fut empruntee pour tenter de les surmonter. LA GENESE D'UN PROJET DE RECHERCHE: CERNER L'IDENTITE CULTURELLE DES GROUPES MAGDALENIENS DANS LE BASSIN PARISIEN J usqu' a la decouverte de Pincevent en 1964, on connaissait une vingtaine de gisements magdaleniens dans la moitie nord de la France. La decouverte de Pincevent, ainsi que celles d'Etiolles, de Verberie et de Marsangy marquerent un tournant en raison de 1' exceptionnel degre de preservation des vestiges et surtout de la revolution methodologique que represente le developpement des methodes palethnographiques sur ces gisements (Leroi-Gourhan et Brezillon 1972). La confrontation des resultats obtenus sur ces quatre sites a donne lieu a plusieurs ·syntheses thematiques qui soulignent leur parente dans divers domaines: strategies d'acquisition du silex; pratiques cynegetiques; habitudes

techniques relatives a la confection de 1'outillage lithique; mode d'habitat (Audouze et al. 1988; Julien et al. 1988; Pigeot et al. 1991; Schmider 1984). D'autres travaux qui s'appuient sur ces donnees ont tented' en deduire des modeles pour interpreter l' organisation territoriale des Magdaleniens dans une zone de plus de 10000 km2 correspondant a l'Ile de France (Audouze 1987; Audouze 1992; Audouze et Enloe 1991; Julien 1989; Taborin 1994). Les auteurs ont souligne la difficulte et les limites d' une telle entreprise en rappelant le foible nombre de gisements sur lequel portaient les comparaisons , leur eloignement geographique et l'ecart chronologique qui separe certains d'entre eux. Sans postuler leur complementarite, les auteurs consideraient que ces gisements bien conserves pourraient illustrer, par leurs differences, la complexite d'un unique systeme d'exploitation saisonniere des ressources , qui pourrait relever d' une identite regionale specifique . Au cours des premieres etapes de notre recherche, nous avons voulu trouver des moyens supplementaires pour preciser 1'identite des groupes magdaleniens du nord de la France. Contribuer a definir leur tradition consistait a rechercher tous les traits qui legitiment l'usage du terme Magdalenien, qui sert par ailleurs a individualiser de nombreux autres groupes dans un espace geographique immense. Nous formulions egalement l'espoir d'entrevoir des traits plus specifiques qui donneraient tout son sens a une denomination regionale: Magdaleniens du Bassin parisien voire Magdaleniens d'Ilede-France. 11va de soi que nos ambitions etaient d'emblee restreintes par la nature des sources. Dans le Bassin parisien, les temoins de

VALENTIN

la tradition magdalenienne sont presque exclu ivement reduits aux vestiges de la culture materielle (les temoins "artistiques" etant etonnamment rares). II etait done clair des le depart que la voie d'acces la plus efficace etait une enquete technologique. Nous nous sommes interesse de preference aux artefacts en silex parce qu'ils constituent d'un gisement a l' autre--et quelles que soient les conditions de leur exploitation archeologique-un denominateur commun, en raison des discontinuites taphonomiques. Nos premiers travaux ont consiste a elargir nos connaissances ur le Magdalenien superieur regional en confrontant systematiquement les donnees dej a acquises a celles que nous avon recueillies a propos d ' autres gisements magdaleniens non dates et moins bien preserves, connus eulement par des rama sage de surface, des fouilles anciennes ou des sondages limites . A ce corpus etudie en partie du point de vue de la typologie de I' outillage (Schmider 1984 ; Allain 1989) , nous avons ajoute d'autres sites mieux preserves, decouverts grace a l'acceleration recente des travaux de sauvetage . Un peu plus d'un siecle apre le debut des recherche , le nombre de sites magdaleniens actuellement connu dans la moitie nord de la France s' eleve a une quarantaine. Les sites ont nombreux et plusieurs sont dates ou situes assez precisement dans un cadre chrono-stratigraphique; sur certains, la faune est conservee (et la saison d'occupation peut etre parfois etablie); les activites sont minutieusement analysees grace au developpement des analyses technologiques. Pour le developpement de nos ambitions initiales, la situation pouvait paraitre ideale, au regard de celle qui prevaut dans beaucoup d'autres regions de l'Europe moyenne.

experimentaux. 11ne s'agit bien sfir que d'estimations et ces mesures approximatives ne permettent pas pour !'instant de depasser l'echelle d'un seul moment d'occupation. L'echelle du temps long est definie par les chronologies absolues et relatives qui nous permettent d' es timer la duree de vie des unites culturelles et leurs rapports dans le temps. Le pouvoir de resolution des differentes methodes utilisees pour ordonner les observations sur le temps long est limite. Les methodes de datation physiques ont une marge d'incertitude qui se compte en iecles voire en millenaires. L'imprecision du cadre chronoclimatique est plus grande encore, pour des raisons qui tiennent evidemment a la tongue duree des episodes climatiques mais au si a la difficulte des correlations qui rend la traduction des bio-zones en chrono-zones bien souvent problematique. II exi te done des echelles temporelles intermediaires qui restent hors de notre portee, car elles sont en-de~a de la reso lution des methodes physique et environnementales d'une part , et au-dela de celle des methodes palethnographiques d' autre part . Ces echelles devraient servir a mesurer la duree que couvrent les occupations successives de certains gisements stratifies, ou la duree qui separe 1'occupation de deux sites par un meme groupe. Actuellement, ii est impossible par exemple de savoir i deux occupations geographiquement distinctes et presumees contemporaines-a l'echelle de nos methodes de datation-sont reellement simultanees ou bien si elles correspondent a des haltes saisonnieres successives du meme groupe ou encore si elles sont en fait separees par plusieurs annees voire par plusieurs siecles. Cette estimation serait fondamentale pour valider nos modeles socioeconomiques sur l'exploitation cyclique d'un territoire. De plus, ces echelles de temps peuvent constituer les cadres de certaines variations techniques discretes mais parfois decisives, dont les temoignages, en consequence, peuvent difficilement etre series chronologiquement.

LES LIMITES DE CE PROJET INITIAL: UNE PERCEPTION PARTICULIERE DU TEMPS ET DES ESPACES PREHISTORIQUES Mais cette recherche concernant l'identite du Magdalenien se heurte dans notre region a la rarete des sites stratifies et, plus fondamentalement, a la faible precision des datations. En consequence, ii reste tres difficile d' ordonner la variabilite des industries magdaleniennes, autrement dit de departager ce que nous avons appele des "variantes techniques," c'est-a-dire diverses options techniques pouvant coexister a l'interieur d'un meme groupe et des "variations techniques," c'est-a-dire des versions simultanees ou successives, appartenant soit a plusieurs groupes differents et coexistants soit a une meme societe au cours de son histoire.

Cette dis torsion dans l' appreciation du temps, deja soulignee par de nombreux auteurs (voir notamment Conkey 1987; Rensink 1995), entraine necessairement une perception tres particuliere de I' espace. Au meme titre que certains gisements accueillent des "palimpsestes" d'occupations, lorsque le pouvoir de discrimination stratigraphique est insuffisant, ce que nous percevons des territoires prehistoriques est une juxtaposition de gisements dont les relations chronologiques exactes ne peuvent pas etre restituees. Un exemple d'analyse "micro-regionale" a laquelle nous avons contribue illustre ces difficultes (Julien et Rieu 1999; Valentin et al. sous presse). Cette analyse portait sur la portion de vallee dans laquelle s'inscrit le site de Pincevent, sur lequel nous connaissons actuellement 15 niveaux d' occupation magdaleniens (Figure 7 .1 ). Dans ce secteur de quelques dizaines de kilometres carres, encadre par deux confluences, on connait pour le moment un seul gisement de plateau, intensement frequente: Le Tilloy aVille-Saint-Jacques a livre, dans des sondages limites, une industrie lithique abondante et des restes de faune domines par le renne mais parmi lesquels le cheval est bien represente . Cette situation differe de celle que 1'on conna'it sur la plupart des niveaux de Pincevent, ou le

Nos moyens de seriation chronologique fine restent en effet limites. La recherche sur le Paleolithique superieur utilise actuellement deux grilles chronologiques principales (les reflexions qui suivent a ce propos sont extraites de Pigeot et Valentin sous presse). Une echelle du temps court a pu etre apprehendee depuis peu grace au developpement d'une palethnographie rigoureuse. C' est l' echelle qui correspond sur un site a la duree d'une occupation, dont les limites peuvent etre parfois fixees par la saison de chasse et qui est scandee par !'execution d'un certain nombre d'activites, dont la duree peut etre evaluee d' apres certains modeles 94

LES GROUPES HUMAINS AU TARDIGLACIAIRE

200m

Glacier Sheet 13000 years BP

0 -----

200 km .

...... .....__

-......._,

-.._

1: La Grande-Paroisse, Pincevent 2: Ville-Saint-Jacques, le Tilloy 3: Marolles-sur-Seine, le Tureau des Gardes 7: Varennes-sur-Seine, le () Figure 7.1-Les

sites avec faune conservee

4: Marolles-sur-Seine, le Grand-Canton 5: Barbey, le Chemin de Montereau 6: Marolles-sur-Seine, le Chemin de Sens

Marais du Pont •

gisements magdaleniens de la region de Pincevent.

95

sites sans faune conservee

VALENTIN

renne est presque exclusif. Un seul niveau de Pincevent, le plus recent, associe egalement le renne et le cheval. Cette association a ete de nouveau rencontree sur deux gisements magdaleniens decouverts depuis 1990: Le Tureau des Gardes et Le Grand Canton a Marolles-sur-Seine. Ces deux sites, ou les restes de chevaux sont nettement majoritaires, s'inscrivent dans une zone de confluence ou l'on a recemment mi en evidence une densite exceptionnelle de gisements magdaleniens (trois autres gisements sont apparus, ou la faune n' est malheureusement pas conservee). Les nouveaux modeles proposes s'appuient done sur la confrontation de quatre gisements riches en faune, "contemporains" a l'echelle du 14C. Une premiere constatation s'impose: notre vision des occupations magdaleniennes dans une aire geographique restreinte a ete profondement modifiee, en peu de temps, a la faveur de decouvertes fortuites liees au developpement de l' archeologie de sauvetage en vallee. Notre perception de l'intensite des peuplements prehistoriques est done fortement dependante de }'evolution des strategies scientifiques et elle peut etre severement deformee par 1'exploitation archeologique inegale des differents contextes geographiques et topographiques. On notera a ce titre qu'il existe dans ce secteur geographique restreint comme dans le reste du Bassin parisien un desequilibre dans la representation des differents contextes, au profit des occupations de fond de vallee. Des facteurs scientifiques oeuvrent en faveur de ce desequilibre: "attirance" pour un milieu assurant une bonne preservation; meilleure perception des risques de destruction brutale, generant un contexte d'urgence permanente. Les facteurs taphonomiques jouent egalement car les sites n' ont pas ete soumis aux memes modes d'erosion (naturelle et anthropique).

allochtones provenant des terrain tertiaires du centre de l'Ile-de-France. Le mode d'introduction exact de ces materiaux demeure inconnu (acquisition directe ou echange?) et l'on peut tout aussi bien continuer a defendre l'hypothese de deplacements d'assez grande ampleur (a l'echelle du centre du Bassin parisien) ou privilegier desormais l'hypothese d'une mobilite plus reduite et d'une complementarite tres locale des activites saisonnieres (s' accompagnant d' expeditions ponctuelles vers des sources plus eloignees de silex ou integrant des echanges avec des territoires voisins). Mais la complementarite reelle des gisements reste sujette a caution en rai on du pouvoir de resolution limite du 14C. Si de nombreux arguments nou permettent d'affirmer que les quatre gisements partagent de nombreux traits culturels (dans le domaine technique), rien n'interdit de penser que la diversite des modes d'occupation reflete une evolution des strategies adaptatives au cours de deux millenaires que pourraient couvrir-selon le resultat des calibrations-la duree de la tradition magdalenienne dans le Bassin parisien. A des systemes d'exploitation a sez planifies ont pu succeder des organisations plus souples (au gre des changements climatiques, des fluctuations dans l'effectif des troupeaux, de !'evolution demographique des societes de chasseurs ...). Ce parti pris de reserve, inspire par la prudence, nous a conduit a elargir notre recherche, que nous voulions d'abord faire porter seulement sur la tradition des Magdaleniens du Bassin parisien. LA REORIENTATION DU PROJET INITIAL: UNE APPROCHE COMPARATIVE DES TRADITIONS TECHNIQUES DU TARDIGLACIAIRE DANS LE BASSIN PARISIEN

Depuis cinq ans seulement, de nouveaux sites proches de Pincevent revelent done la frequence d'une situation deja reconnue a Ville-Saint-Jacques: !'association du renne et du cheval dans le tableau de chasse de certains groupes magdaleniens. Les etudes archeozoologiques menees sur ces nouveaux gisements indiquent un mode d'occupation, jusque la inedit, differant assez nettement de celui qui a ete mis en evidence sur la plupart des niveaux de Pincevent ainsi qu' a Verberie: succession de courts episodes d' abattage de chevaux, echelonnes sur toute la duree de l' an nee et reunissant probablement peu de chasseurs a chaque fois (Bridault et Bemilli 1999).

Le projet s' est done oriente vers une approche comparative prenant en compte egalement les industries des groupes tardiglaciaires plus recents que le Magdalenien (Valentin 1995). Dans cette nouvelle optique, la definition du cadre geographique fut seulement guidee par la recherche d'une documentation de qualite et par la necessite de disposer d'une "unite de lieu" ou pourraient se reveler des contrastes sur la longue duree. L' ouest et le sud du Bassin parisien, qui presentent des paysages et des ressources naturelles assez identiques, ont ete choisis comme cadre de cette enquete car nous avons pu y reunir un corpus en partie inedit compose d'une quarantaine de gisements (une vingtaine pour le Magdalenien; une autre vingtaine pour les traditions plus recentes). Les nombreuses donnees recueillies plus au nord, dans la vallee de la Somme, ont contribue, par comparaison, a une meilleure definition du cadre chrono-stratigraphique (Fagnart 1997).

Considerant la proximite des datations 14C obtenus sur les gisements concernes, on est en droit de proposer l'hypothese de leur complementarite. Dans ce cas, plusieurs modeles sont en concurrence: simultaneite d' occupation entre un camp de base et un (ou plusieurs) site(s) specialise; succession de plusieurs etapes au cours d'un cycle saisonnier. Les combinaisons envisageables sont multiples et aucun argument deduit de l'etude des outillages ne vient pour l'instant soutenir l'une ou l'autre hypothese (aucun assemblage d'outil n'indique pour }'instant un degre de specialisation particulier). Plus incertaine encore reste l' articulation de ce cycle eventuel d'activite dans le cadre regional plus vaste que designe la presence, sur chacun de ces gisements, de silex

Le resultat est une enquete paleohistorique s' interessant aux mecanismes qui ont preside a la stabilite des traditions techniques ou a leurs changements. L'etude couvre au moins quatre millenaires scandes par des mutations climatiques, qui ont impose probablement aux communautes de chasseurs l' elaboration de strategies economiques differenciees (Figure 7 .2).

96

LES GROUPES HUMAINS AU TARDIGLACIAIRE

,,. --

rO

o ,,..

,,,,=·:.·•:.

@

Prlndpaux 11tes des groupes l Federm«ISerde la "phae andenne" (en nolr, sites de notre corpus)

Prlndpaux sites du Magdal6nlen11.1p6rieur (en nolr, lites de notre corpus)

(a- ",-

!111111! ::::

• • lleilolslan •

Alnnsbwglen • · Sitesd'Nlbltat d'tttributlonlnoertahl

Prlndpaux sites dee groupes l F•dMrnnHr de la • phne rkente• (on nolr, lites de notro corpus)

AGf.S

de• groupesde rextrtme fin du TardlgladaJre (en nolr, de notre corpus)

PrlndpauxIll"

c,_.,-.----..--- ....-----

sit••

....-----------,

C14

B.P.

Pr6bor6al

Flandrlan

lneer;lac.

10000

,0000

rarmtlon Dryas II (r6cert)

"Bellolalen• YOWlf« Dryn

Stadtal

®

Feder.

11000

11000

·-

AllerOd

ransltlon

----12000

12000 l.atll

Glacial e,stadlal

ea 13000

1300

......

,

..., Late

0eovetlllln /Mein

8ubdMlianl

CUgN

ohronoolmltlqUN

per1aedann..

calagllparlaC1-4

~ at llWIRlnntffltlltllel

._

par~trallur

(cumuldee lntlMllae • 2

•>

Figure 7.'1.-A : Localisation des principaux gisements de notre corpus; B: Chronologie des traditions techniques du Tardiglaciaire dans le Bas in parisien .

97

VALENTIN

Nous avons corn;;uune enquete visant d'abord a cerner, en Ies confrontant, 1'identite des differentes traditions techniques, cette confrontation nous permettant de faire ressortir, par contraste, les traits presume les plus stables de chacune d' entre elles. En outre, pour une tradition comme le Magdalenien les facteurs circonstanciels de variabilite peuvent etre en partie evalues grace a l'avancement des recherches palethno graphiques: dans certains domaines, ii a done ete possible d' observer une diversite des choix qui ne se laisse pas reduire a 1'expression de variantes. Sans pouvoir les serier precisement certains faits sont apparus comme le temoignage de variations significatives . Nous avons done privilegie l'echelle du temps long mais en cherchant a tirer profit de tout ce que l' on sait du quotidien des differentes communautes .

A travers cette etude, on peut acceder a plusieurs ordres de faits techniques organises selon des principes complexes d'arborescence. Une activite impliquant !'usage du silex se situe en effet a }'intersection de plusieurs "cheminements techniques" (c'est a dire plusieurs chaines operatoires articulees): celui qui conduit de !'acquisition de la matiere premiere a sa transformation en outils puis a leur utilisation et tous les autres cheminements qui conditionnent ou motivent ces differentes etapes (la chasse, la transformation de ses produits, la construction des habitats, pour ne citer que ceux qui sont partiellement conserves archeologiquement) . En nous interessant surtout aux modalites d' acquisition du silex, aux modalites de debitage et de transformation des supports en outils ainsi qu ' aux procedes de maintenance de l' outillage, nous avons privilegie I' etude d'un cheminement dont la continuite archeologique est pre que integralement preservee mais nous savons aussi que de multiples articula tions le relient aux autres secteurs du systeme technique . Ces intersections peuvent parfois etre designees archeologiquement par des couples d' objets conserves ou seulement deduits (outils en os et leurs armatures en silex par exemple). Ces intersections meritent toute notre attention car elles peuvent nous informer un peu plus largement sur l 'organisation du systeme technique general.

Nous avons construit a partir de cette recherche un modele evolutif qui cherche a rendre compte de Ia part que tiennent dans Ies changements constates differents facteurs (incitations et limites de l'environnement, mecanismes socio -culturels) . Pour 1'instant , ce modele a e sentiellement une valeur regionale mai nous avons signale a plu ieurs reprises que les evenements per~us dans le Bassin parisien pouvaient constituer les echos de phenomenes paleohistoriques de grande ampleur, qui se sont produits dans un vaste espace, a nouveau decloisonne par la deglaciation du domaine nordique et des regions montagneuses. L'Europe centrale et septentionale forment done la "toile de fond" de cette recherche.

En parlant de technologie globale nous voulons dire egalement qu'au dela d'une description des faits, ce sont des idees que cette methode doit restituer car !'action de l'homme sur la matiere met en oeuvre autant de decisions, de concepts et de representations que de gestes. Ce qui donne a ces idees une coherence-arbitraire-c'est un systeme de valeurs determinant dans chaque societe une appreciation eminemment variable des moyens et des besoins. Pour qu'une technologie comparee puisse contribuer a une approche paleohistorique, son ambition ultime doit etre de chercher a confronter ces systemes de valeurs.

II nous faut revenir maintenant quelques instants sur le choix de la technologie comme methode. Ce choix s'impose, comme nous I' avons dit, en raison de la nature des sources et ii s'accompagne immediatement de deux exigences. Par definition, puisque ce travail repose sur la confrontation de multiples assemblages lithiques, il s'agit necessairement de technologie comparee. L' analyse technologique permet de decrire !'interaction des elements concretement mis en oeuvre dans une chatne operatoire. L'interaction de ces elements definit des "modalites" (Pelegrin 1995), c'est-a-dire "toute forme, mode, version de realisation d'un acte technique" (ibid.). Pour contribuer a une histoire des idees techniques on ne peut se contenter d'une simple controntation des diverses modalites attestees dans differents systemes de production lithique. Confronter ces versions concretes, en vue de faire ressortir des options differenciatrices, suppose que l' on reconnaisse les intentions qu'elle veulent satisfaire (ces intentions etant la concretisation d'images mentales, de concepts, de "manieres de voir" [ibid.]). Les options differenciatrices, les "preferences," peuvent etre decelees des lors que l' on observe "la realisation repetee d'une intention technique selon une certaine modalite alors qu'une autre modalite au moins aurait pu satisfaire la meme intention " (ibid.).

LES GROUPES HUMAINS ET LEURS TRADITIONS TECHNIQUES AU TARDIGLACIAIRE DANS LE BASSIN PARISIEN (Table 7.1)

Les groupes du Magdalenien superieur A travers l'etude d'une vingtaine de gisements repartis sur une assez vaste aire geographique, notre enquete a debute par les groupes du Magdalenien superieur, qui ont occupe la region depuis le debut du Tardiglaciaire jusqu'aux commencements de l'interstade d' Allerod soit durant 2000 a 3000 ans. II faut signaler qu'a l'exception d'Etiolles, dont !'occupation remonterait au debut du Bolling ou juste avant, les sites dates sont attibues pour la plupart a la chrono-zone du Dryas II et sont done assez tardifs, au regard de ce que I' on sait de I' occupation magdalenienne des regions avoisinantes (Belgique et Rhenanie notamment).

L'autre exigence conceme le champ d'application de la methode que nous considerons comme un champ tres large comme nous avons voulu l' exprimer en affirmant que la technologie se doit d'etre "globale." S'interesser a la production de l'outillage en pierre dans une communaute du tardiglaciaire nous plonge au coeur de son systeme technique (on pourrait en dire autant d'autres temoins , s' ils etaient integralement preserves) .

Un forte unite technique peut etre per~ue entre les assemblages magdaleniens , a travers I' expression de preferences 98

LES GROUPES HUMAINS AU TARDIGLACIAIRE

Principaux changement Tradition s Technique s et Facies d 'activite

Degre de specialisation des gisement s

Table 7.1 Trois tradition s techniques et un facies d' activite: relatif aux indu tries lithique s au cours du Tardiglaciaire dans le Bassin pari sien

Approvi sionnement en silex

silex local de bonne qualite Magdalenien

Armatures dominante s

a dos

lamelle s

Objectif s principaux du debitage

variable

percuteur tendre organique quelques support s en silex allochtone

quelque s pointe s a dos

silex local de bonne qualit e pointe s

inconnu

a dos

Degre de predetermination de la methode de debitage

productivit e et normali sation laminaire

superieur

Federmesser

Techniqu e de percussion pour le plein -debitage

eleve

productivit e et normali sation lamellaire productivite et normali sation laminair e

ancien

percuteur de pierr e tendre

eleve

percuteur de pierre tendre

faible

percuteur de pierre tendre

Ires eleve

(bipointe s symetrique s) pas d'obj ectif lamellair e

qu elque s supports en silex allochtone point es

a dos

(monopointe s asy metrique s)

Federmesser

inconnu

recent

silex local de qualite tres variable

utili sation des nombreux eclat s quelques lamelle

a dos

silex local de tres bonne qualite "Belloisien "

tres eleve

produ ctivite et normali sation laminaire reduites

peu nombreuses mais diverses d 'un site a / 'autre quelques support s en silex allochtone

objectif lame Ilaire deriv e

productivite et normalisation laminaire elevees

produ ctivite et normalisation lame Ilaire

dont la recurrence semble transcender la diversite des modes d' occupation. Notre enquete fait ressortir une tres forte stabilite des choix dans un des domaines de la production lithique: la confection des supports de l' outillage de transformation (essentiellement grattoirs, burins et bees) . Sur tous les gisements, des debitages elabores ont ete con~us a cet effet. Leur vocation economique est une production en serie de supports laminaires normalises, longs et de section legere. Ces lames peuvent circuler dans l'es pace des campements (Bodu 1993; Olive et Pigeot 1991) et constituent Jes petites panoplies que les Magdaleniens transportent d'un site a l'autre (attestees par quelques produits en silex non local presents dans chaque

gisement). Aucun concept d'o util particulier ne semble commander particulierement ces exigences. Longueur et normalisation sont plutot en accord avec des exigences generales relatives a la maintenance des outils (facilites de ravivage) et aux possibilites eventuelles de transformation en cours d'usage (polyvalence). II faut souligner que le coGt de cette option fut assez eleve dans l'absolu (quand nous disons dans l'absolu , c'est qu' il n' existe pas, a notre sens, d' unite de mesure universelle pour rendre compte de l' investissement technique d'un groupe humain). Cette option exige notamment une selection attentive 99

VALENTIN

des matieres premieres locales et s'accompagne d'un transport systematique de quelques supports deja debites d'un gisement a 1'autre, y compris quand le contexte geologique local est favorable. De plus ces debitages elabores, realises preferentiellement au percuteur tendre organique, consistent a mettre en oeuvre des preferences techniques difficiles: predetermination elevee, assuree i necessaire par une longue phase de mise en forme; maintien d' un plan de frappe preferentiel fortement incline dont le bord est toujours soigneusement prepare; entretien des surfaces de debitage, si possible, par la progression du debitage ... (Audouze et al. 1988; Pigeot 1987; Pigeot 1991; Pigeot et al. 1991; Ploux et al. 1991; Valentin 1995). La mise en oeuvre reussie de ces preferences suppose une somme de connaissances et un degre de savoir-faire eleves qui n'ont pu etre reunis qu'au prix d'un apprentissage progres if et d' une pratique reguliere (Pigeot 1987; Ploux et al. 1991). On peut reconnattre dan la manifestation recurrente de ces choix, relativement "cofiteux" pour les communautes, l'ex pression d' une planification des besoins en outillage (a l'echelle de chaque campement et a l'echelle territoriale) . Cette option semble en parfaite adequation avec les contraintes definies par certains traits des economies magdaleniennes: mobilite probablement assez elevee, acquisition saisonniere a haut risque de ressources inegalement reparties, expeditions dans des zones marginales pauvres en silex. Cela etant, nous n' avons encore qu' une image tres partielle des economies magdaleniennes et leurs formes ont pu evoluer au cours du temps: ]'exploitation planifiee des ressources animales n'est qu'une tendance et elle ne s'est peut etre pas manifestee a tousles moments. Quoi qu'il en soit, ]'anticipa tion des besoins en supports polyvalents se manifeste dans des contextes divers: a Etiolles (ou la chasse, pour le peu qu'on en sache, fut diversifiee), a Pincevent (ou elle fut dirigee vers une espece au cours d'occupations breves mais rassemblant probablement d'assez nombreux individus), ou encore sur certains sites de la confluence Seine-Yonne (ou semblent s'etre deroules plusieurs courts episodes d'abattage, mobilisant peut-etre a chaque fois un petit nombre de chas eurs). La motivation des options techniques evoquees (anticiper Jes besoins en outils par une production normalisee) pourrait done etre plus complexe que ne le laisserait supposer une equation simple, rapportant la constance de cette anticipation a !'expression d'un systeme economique hautement planifie. Ce choix a pu connattre un long succes parce qu' il garantit au systeme technique dans son ensemble une certaine stabilite, en depit de la fluctuation probable des ressources (selon Jes contextes, selon les moments de l'annee et selon les epoques). Nous ajoutons a ces observations une interrogation complementaire. On peut trouver a cet investissement technique des motivations economiques mais ne s'agit-il pas, pour une part, de justifications que nous Jui reconnaissons a posteriori? On peut se demander s'il n'y a pas dans cet investissement, qui semble egalement atteste dans d'autres territoires magdaleniens , l'indice d'une "valeur" particuliere - peut-etre

"excessive" dans certains contextes-accordee a la normalisation de quelques debitages laminaires (ou a la methode qui permet de l'atteindre). On soulignera a l'appui de cette hypothese qu' une valeur economique "negative" est accordee, par contraste, a la plupart des enlevements predeterminants non laminaires : meme dan des contextes ou l' "efficacite" aurait recommande le contraire, la part des eclats transformes en outils reste excessivement faible. Ce choix, dont Jes motivations ne sont peut etre pas seulement economiques, n' a pas d' equivalent dans les traditions tardiglaciaires plus recentes de la region. Au dela d'une incontestable unite, il existe dans le domaine du debitage des lames une variabilite difficile a reduire seulement a des facteur circonstanciels: Jes debitages de grandes et de tres grandes lames connus dans certains contextes et notamment aEtiolles (Olive et Pigeot 1991; Pigeot 1987) sont absents des sites magdaleniens les plus recents. 11est vrai que les sources de silex neces aires pour ce type de production sont inegalement reparties . Mai il reste assez difficile a comprendre pourquoi, dan l'e ventuel cycle qui rattacherait Etiolles a des gisement comparables a ceux que nous avons etudie , cet investissement technique particulier fut si occasionnel (la production exceptionnelle d'Etiolles ne semble pas particulierement destinee a etre diffusee a l' echelle territoriale ). 11faut ajouter que dans certaines unites d'Etiolles, l'investissement technique ne s' est pas limite a ce secteur des activite (comme en temoigne la nature des structures d'habitat). Nous serions prets a admettre que ce comportement peut constituer un des traits de la societe magdalenienne (surinvestissement ephemere mais repete, a !'occasion d'une opportunite exceptionnelle). Mais ii se trouve que cet investissement a des equivalents dans quelques autres gisements d'Europe septentrionale, d'occupation plutot precoce, notamment a Gonnersdorf dont l'environnement immediat ne comporte pas de sources de silex (Franken et Veil 1983). En conequence, nous nous demandons si la production des grandes lames, avec ce qu'elle suppose comme controle des sources de silex, comme gestion des percuteurs, comme distribution des savoir-faire ne correspond pas a un phenomene culturel limite dans le temps (cette hypothese a deja ete evoquee dans Pigeot et al. 1991). La confection des elements en silex armant les instruments ·de chasse est un autre domaine, ou s'expriment, semble t-il, des variations significatives. Sur la plupart des gisements, et suivant une tendance bien ancree dans la tradition magdalenienne, l' association des lamelles a dos a des tetes de projectile en matiere osseuse constitue la version dominante dans 1'equipement de chasse. Or nous avons reconnu de maniere recurrente un certain nombre de variations concomitantes qui ont trait a la morphometrie des lamelles a dos, aux techniques de retouche qui leur sont appliquees et aux methodes de production de leurs supports (Valentin 1995). Traduisant une evolution possible des systemes de fixation (voire de la nature et du mode de lancer des armes composites) les variations inedites que nous avons reconnues sont encore difficiles a serier mais pourraient 100

LES GROUPES HUMAINS AU TARDIGLACIAIRE

constituer a l' avenir de bons criteres pour une periodisation plus serree du Magdalenien regional. II s'avere en effet que deux assemblages differents de lamelles a dos sont superposes dans la stratigraphie de Pincevent mais cette evidence est encore insuffisante pour batir un modele evolutif solide (Valentin 1995, 1999). Sur deux-tiers des gisements , les lamelles a dos sont associees a des pointes a dos, interpretees comme des armatures axiales (Schmider 1984; Schmider 1992; Valentin 1995). Compte tenu de la presence repetitive de ces outils, attestes dans la plupart des cas en faible nombre, c'est surtout leur absence sur quelques gisements comme Etiolles, Pincevent et Verberie qui pose probleme. 11n'est pas certain que l'anciennete de ces gisements est seule en cause car ii n' est pas encore demontre que I' apparition de ces pointes est veritablement un phenomene tres tardif. La fonction specifique de ces elements pourrait egalement expliquer qu'ils ne soient pas totalement ubiquistes: a titre d'hypothese de travail, dont la validation exige un diagnostic fonctionnel plus approfondi , on rappellera que Jes pointes sont parfois assez regroupees dans les gisements et nous nous demandons en consequence si leur presence ne reflete-t-elle pas des moments de chasse parti-culiers (gibier specifique, technique d' interception differente). 11pourrait y avoir alors une certaine logique ace qu'on retrouve ces outils sur des sites, qui ont peut-etre accueilli plusieurs installations successives, a diverses saisons (c'est probablement le cas sur les deux gisements de Marolles). Inversement, sur des gisements dont les temps d'occupation sont limites par le rythme des crues (Pincevent, Verberie), il est possible que ces moments de chasse particuliers n' aient pas eu lieu. Un fait particulier doit etre signale au sujet de ces pointes. Sur trois gisements de notre corpus, elles sont particulierement abondantes et la production de leurs supports concurrence, a Marsangy, ou supplante, a Cepoy, la production des supports de lamelles a dos. Sur un substrat technique parfaitement con forme a celui des autres gisements, cette inflation des armatures axiales semble s'accompagner d'un trait original: leurs supports, des petites lames robustes, ont ete debitees assez systematiquement au percuteur de pierre tendre. 11 reste a comprendre l' exacte valeur de cette option. A priori, nous n'excluons aucune hypothese et nous signalons qu'a la lumiere de cette nouvelle originalite, la question des relations avec le Hambourgien doit etre reposee. Mais il faut tout de suite rappeler que cette originalite repose sur la recombinaison d' elements preexistants (1'usage du percuteur de pierre tendre est atteste dans les autres assemblages magdaleniens surtout pour des operations demise en forme et d'entretien). Nous voulons souligner par ailleurs la relative simplicite de la methode de debitage qui accompagne a Cepoy et Marsangy l' usage de cette technique et preciser que cette systematisation de l'usage du percuteur de pierre n'a pas necessairement exige un transfert d'idees techniques complexes pour se developper. Dans Jes annees a venir, ii importe egalement de verifier s'il s'agit reellement d' un fait tardif comme le laisserait supposer

l' attribution chronologique de Marsangy (a la transition entre la chrono-zone du Dryas II et celle de l' Allerod) .

Les Groupes

aFedermesser

Cette question parait d'autant plus importante que les derniers moments de l' occupation magdalenienne recouvrent un intervalle de temps ou apparaissent les temoignages d' une tradition technique peut-etre distincte: celle des groupes a Federmesser de la "phase ancienne." Cette tradition, reconnue d' abord par J.-P. Fagnart dans la vallee de la Somme (Fagnart 1997) a ete individualisee, depuis peu, dans la moitie meridionale du Bassin parisien: sur le niveau inferieur du Closeau a Rueil-Malmaison (Hauts-de-Seine) (Bodu 1995) et a La Grottedu Chevala Gouy (Seine-Maritime) (Valentin 1995). Dans !es quelques assemblages connus, on observe une generalisation de l'usage du percuteur de pierre tendre se substituant totalement aux percuteurs organiques preferentiellement utilises au Magdalenien. Au sud du Bassin parisien, on constate ]'affirmation d' un nouveau type d' armature (bi-pointe a dos courbe, etroite et legere) et l'absence de lamelles a dos. Ces transformations ne paraissent pas s' accompagner d' une claire rupture concernant les exigences generales assignees au debitage . On pourrait y voir le temoignage d'un certain conservatisme. Mais est-on vraiment en droit de distinguer nettement les nouvelles options de celles qui fondaient une certaine identite des Magdalenien? L' avenir nous dira si nous ne sommes pas plutot en presence des manifestations ultimes de la tradition magdalenienne. Nous retenons de ces premieres observations que 1' "azilianisation" des industries dans le Bassin Parisien fut un processus progressif, dont la logique trouve, semble-t-il, des echos lointains: notamment dans les niveaux anciens de certaines sequences aziliennes des Alpes ou d' Aquitaine (Celerier 1993; Chollet et al. 1999; Pion 1990). Une rupture beaucoup plus nette est consommee avec le Magdalenien au cours des "phases recentes" de la tradition des groupes a Federmesser, qui s' affirmeraient durant la premiere moitie de l' Allerod, dans un environnement au caractere forestier nettement plus marque (Fagnart 1997). La panoplie d'outils a change depuis le Magdalenien. Les armatures axiales (monopointes symetriques) sont partout nettement predominantes et elles sont constamment etroites et legeres. Les armes de chasse les plus courantes sont probablement d' une autre nature qu'au Magdalenien. II peut s'agir d'une veritable rupture (invention d'un autre mode de lancer) ou seulement d'une restriction des solutions techniques diversifiees qui se sont affirmees a la fin du Magdalenien . Les enquetes fonctionnelles comparatives ne sont pas suffisamment avancees pour conclure definitivement sur ce point mais la generalisation de l'usage de l'arc reste une solide hypothese de travail qui apporte une explication satisfaisante a la multiplication des armatures axiales etroites et legeres (Rozoy 1978). Les outils de transformation ont une diversite plus restreinte qu' au Magdalenien. Certains instruments cornrne les grattoirs 101

VALENTIN

semblent fabriques pour former systematiquement les extremites d'un outillage composite. Les parties actives en silex sont moins soumises a des operations d'affOtage qu'au Magdalenien et leurs transformations en cours d'usage sont plus occasionnelles. Le faible taux de cassure de ces outils s'accorde par ailleurs avec certaines donnees traceologiques acquises dans des contextes voisins, qui suggerent des temps d'utilisation brefs pour ces parties actives (ces resultats attestent egalement la rarete des utilisations multiples) (Plisson 1985, pour la Rhenanie). On pourrait etre tente d'etablir une relation entre la tres foible densite des occupations connues, qui evoque des sejours tres courts et une mobilite elevee, et cette apparente diminution du soin accorde a l' entretien des outils en pierre. Mais l'equation n'est peut-etre pas si simple car c' est en fonction des changements ergonomiques qu' il fout juger la transformation des strategies de maintenance de l'outillage en silex: !'attention est peut-etre desormais accordee de preference a la confection et a I' entretien des manches. En tout etat de cause, le role du silex dans l' equipement materiel semble avoir en partie change , etant donne le faible investissement technique dont il fait desormais 1'objet. Dans l'ouest et le sud du Bassin Parisien, les materiaux traites ont ete exclusivement acquis a proximite immediate des gisements; il ne subsiste plus aucun temoignage d'une anticipation des besoins par le transport de supports allochtones, comme au Magdalenien ou durant la phase ancienne de la tradition a Federmesser. Les modalites de l'approvisionnement local ont egalement change depuis le Magdalenien. Si les sources ont pu rester identiques a certains endroits, on peut affirmer que la collecte ne s'y est pas foite selon les memes criteres: l'heterogeneite dimensionnelle et qualitative des volumes recueillis est toujours plus elevee que dans les occupations magdaleniennes. Doit-on invoquer, pour interpreter ce changement, une restriction des ressources en silex, comme on 1'a parfois propose? Le couvert vegetal est devenu plus dense et son developpement a pu rendre difficile l' acces a certains gites. Mais ces modifications de 1'environnement ne suffisent pas a rendre compte des changements dans ce domaine car il est difficile d'entrevoir dans les strategies d'acquisition et de production les temoignages d'une pression occasionnee par un acces plus difficile aux sources de silex. On ne constate pas en effet d' anticipation des besoins futurs lorsque les gites exploites sont de bonne qualite et l' on n' observe jamais d'intensification particuliere du debitage. A notre avis, ii convient d' ecarter provisoirement les determinations ecologiques comme principaux focteurs de changement dans ce domaine. On remarquera d' ailleurs que pour expliquer en partie la simplification des methodes de production chez les groupes scandinaves de l' Allerod, on invoque parfois une causalite totalement inverse, en soulignant la disponibilite accrue en bons materiaux d'origine morainique (Fischer 1991; Madsen 1993). Pour le moins, ii faut admettre que la relation de causalite a pu etre beaucoup plus complexe. A titre d'hypothese, on pourrait construire un autre modele

fonde sur une hierarchie differente des causes : n'est-ce pas plutot la simplification des methodes de debitage qui permet desormai d'exploiter une gamme tres diversifiee de materiaux? Le nouveau systeme de production fournirait aux societes de 1'Allerod les moyens pour se degager des liens de dependance relatifs entretenus avec l' environnement mineral. On pourrait y trouver une des raisons du succes de ces innovations car cette "liberte" constituerait un gain appreciable dan une economie apparemment fondee sur une plus forte mobilite residentielle . Qu'il y ait eu un lien entre les modifications de l'environnement et les principaux changements techniques est fort probable mais nous voyons plutot s'esquisser une relation de causalite complexe, ou le role des choix economiques parait determinant de prime abord. Une plus forte mobilite et une exploitation plus tactique des res ources animales autant que minerales ont pu favoriser la stabilisation des nouveaux choix techniques. Ce choix consistent en de nouvelles manieres de faire (auxquelles on peut done trouver des motivations economiques, sans oublier qu'il s'agit peut-etre pour une part de justifications a posteriori). Ils expriment egalement des "manieres de voir" differentes. Les armatures d'instruments de chasse et !es "couteaux" soot les seuls outils exclusivement fabriques sur lames. Le statut de "dechet" ne s'applique plus systematiquement aux enlevements non laminaires puisque des outils comme les grattoirs et les burins sont souvent faits sur eclats. La methode de debitage est encore clairement dominee par une intention laminaire mais elle ne s'accompagne pas d'une recherche de productivite en supports normalises. L' objectif laminaire restant preponderant, le debitage exige done un certain degre de predetermination. Mais cette predetermination s' ex prime a court terme, se traduisant par une absence de stricte distinction entre les differentes phases du debitage. A toutes les etapes et tre precocement, enlevements predeterminants et predetermines, detaches exclusivement au percuteur de pierre tendre, s'enchainent selon un ordre fixe par des opportunites circonstancielles: les lames calibrees sont detachees le long des nervures degagees par des eclats allonges et epais; les plans de frappe echangent frequemment leurs roles pour supprimer facilement les nombreux negatifs de reflechissements; la configuration des volumes est corrigee par des sequences d' entretien dispendieuses. On conclura cette evocation en precisant que le savoir-faire investi dans la production lithique n'atteintjamais un tres haut niveau, y compris pour les debitages les plus elabores. Peutetre doit-on rechercher egalement les raisons du succes de cette methode dans sa facilite d' execution et dans le court apprentissage qu'elle exige. Nous ne nous risquons par a en conclure que la taille du silex s'inscrit desormais dans le cadre d'un savoir-faire moins specialise encore qu'au Magdalenien. Nous nous contentons de souligner un assez foible investissement des communautes dans cette activite (qui s'accompagne peut-etre d'une redefinition de la "valeur" qui lui est accordee) sans pouvoir preciser si ce faible investissement est l'indice d'une reorientation des priorites ou s'il temoigne

102

LES GROUPES HUMAJNS AU TARDIGLACIAIRE

d'un moindre degre de division des competences (eventuellement dans des groupes plus petits?). II reste a preciser le rythme de I' "azilianisation," dans sa dimension technique, pour apprecier correctement ses facteurs probablement complexes. En certains endroits du Bassin parisien, ce processus semble done s'etre produit en deux temps au moins. Mais ii se pourrait que la tendance qui s'affirme nettement dans notre region au cours de la premiere partie de l' Allerod ("phase recente") ait pu emerger plus tot dans certains contextes: a ce titre , ii faut signaler que les temoignages comparables a notre "phase ancienne" pourraient trouver des equivalents en Grande-Bretagne a Hengistbury-Head (Barton 1992) mais font defaut pour le moment en Belgique et en Rhenanie (Bosinski et al. 1995; De Bie 1997; Baales et al. comm. pers.). Dans notre region, les donnees disponibles sur l'occupation a Federmesser de Pincevent (niveau III) participent egalement a cette discussion (voit les contributions de Baffier et al. dans Gaucher 1996). Assez precoce d' apres les donnees radiometriques, cette occupation possede une industrie similaire aux assemblages de la "phase recente" mais livre, sur le cortex d' un eclat en silex , une gravure qui presente d'indiscutables convergences avec l' art mobilier magdalenien. Cette association constitue un temoignage precieux sur les ecarts probables entre les rythmes d' evolution des differents faits culturels au cours de cette periode de mutation. Le devenir de la tradition des groupes a Federmesser reste egalement une source d'interrogation . Les donnees chronostratigraphiques acquises recemment sur le niveau superieur du Closeau a Rueil-Malmaison attesteraient que cette tradition a pu perdurer jusqu ' au debut du Dryas recent (Bodu 1995). Un degre supplementaire de simplification des methodes de debitage semble atteint au cours de cette phase tres tardive sans que l' on observe une veritable baisse de qualite des materiaux traites.

Les groupes de l' extreme fin du Tardiglaciaire Une nouvelle rupture majeure dans l'evolution des traditions techniques regionales peut etre datee de la fin du Dryas III et du debut du Preboreal (Barton 1991; Fagnart 1997). Cette nouvelle transformation, dont on ignore la genese , est attestee par l' apparition de facies specialises comparables, dans une aire qui s'etend du sud-est de I' Angleterre et du Bassin de la Somme jusqu' a l' ouest et au sud-est du Bassin parisien, ou une dizaine de nouvelles references ont ete identifiees. La diversite des termes utilises jusqu'ici pour designer ces industries ("Long Blade Assemblages," "facies a elements machures," "Belloisien"), temoigne des hesitations relatives a l'identite culturelle des auteurs de ces facies specialises (voir infra) . Le "Belloisien"-nous preferons ce terme neutre qui ne conrere pas de preeminence a un seul caractere technique, au demeurant relatif-est connu sur des sites devolus en partie a l'acquisition d'une source au moins de tres bon silex et au debitage de lames en vue d'un usage en partie differe. Sur plusieurs gisements, les outils majoritaires sont des dechets de taille endommages par l'usage (elements machures). Les

resultats d'analyses traceologiques et experimentales recentes indiquent que certains de ces outils a posteriori pourraient avoir servi directement dans le cadre des activites de taille (confection et entretien des percuteurs en gres) (Bodu et Valentin 1993; Fagnart et Plisson 1997). Les sites d ' habitat eventuellement lies a ces gisements demeurent pour l'instant inconnus mais l'existence d' etapes , ou }' acquisition et le traitement du silex ont joue un role primordial, evoque une configuration assez particuliere des cycles d' activites. Bien que la destination precise des produits preleves demeure inconnue , on peut affirmer que ces choix traduisent un haut degre d'anticipation des besoins en outillage . Ces options supposaient une bonne connaissance du milieu geologique local. Une incitation assez forte du milieu naturel a pu jouer dans ce domaine: les conditions climatiques de la fin du Dryas recent (reduction du couvert vegetal et capacite erosive accrue des cours d' eau) ont peut-etre favorise la mise a nu de sources tres localisees . En tout etat de cause , meme si l' acces acertaines sources s' est trouve facilite, les criteres qui regissent I' approvisionnement , y compris en milieu alluvial, sont redevenus stricts. II reste assez surprenant-mais seulement au regard d'une logique peut-etre totalement etrangere aux groupes concernes-que les excellentes sources qui ont ete decouvertes acette occasion n' ontjamais fait l'objet d'une exploitation durable. D' autres indices attestent un fort investissement dans l' activite de taille. En temoigne d' abord I' energie mobilisee pour le transport: d' assez nombreuses lames (et peut-etre des preformes de nucleus) ont ete prelevees. L'investissement s'exprime egalement a travers l ' elaboration des methodes de debitage laminaire et le savoir-faire necessairement mobilise pour de executions difficiles (Bodu et Valentin 1993; Valentin 1995). Les contraintes opposees par un debitage au percuteur de pierre tendre ont ete contournees par la mise en oeuvre de principes assez rigides garantissant une forte normalisation de la production: mi e en forme elaboree, longue et dispendieuse ; mise en jeu immediate de deux plans de frappe utilises en alternance tres rapide ; preparation soigneuse des bords de plans de frappe; progression du debitage tres envahissante. II n' est pas certain que ces facies relevent tous de la meme tradition. IIs peuvent temoigner aussi bien de l' expansion geographique d'une seule tradition que de la convergence des choix techniques et economiques partages par plusieurs groupes contemporains. La deuxieme hypothese recueille momentanement notre preference pour des raisons intrinseques d'abord. Les rares armatures associees sont diversifiees: exemplaires evoquant certaines armatures arhensbourgiennes dans le sud-est de 1'Angleterre et dans la Somme (Barton 1991; Fagnart 1997), elements d'affinites laboriennes sur l'occupation "belloisienne " pour l'instant la plus meridionale (Hanta ·i 1997). Les raisons extrinseques sont constituees par quelques indices de choix techniques comparables dans des contextes plus lointains mais a peu pres contemporains: en Aquitaine, dans les 103

VALENTIN

industries laboriennes (Celerier 1993); au Nord de l' Allemagne, dans certains assemblages ahrensbourgiens et en Pologne, dans le Swiderien (Schild 1984). Ajoutons que les choix economiques originaux qui definissent les facies "belloisiens" pourraient egalement trouver des equivalents a longue distance (notamment dans le Swiderien). Les donnees sont encore insuffisantes pour proposer un modele qui rende compte des eventuelles relations intercommunautaires qui ont pu encourage ces emergences. Nous soulignons seulement que les concepts techniques de nouveau rigides qui regissent le debitage laminaire dans certaines de ces traditions sont des "idees fortes" (necessitant un apprentissage) et que leur apparition presque simultanee dans des regions eloignees se con~oit mal sans diffusion. Les changements climatiques du Dryas recent ont peut-etre favorise (par leurs consequences economiques) la concretisation locale de ce nouveaux choix, mais force est de cons tater qu' ils s' expriment encore pleinement lors de }'amelioration climatique brutale qui succede. Cette amelioration n' eut done pas immediatement pour corollaire un desinvestissement dan les activites de taille du silex. A l' inverse, dans certaines regions (Est de la France notamment), la prolongation des traditions techniques relevant de l' Azilien durant les froids rigoureux du Dryas recent designe un autre decalage possible. II importe de verifier si le fait est limite a des regions particulieres (montagneuses notamment), ou s'il ya la l'indice d'une divergence des choix plus generalisee.

CONCLUSION Les formulations interrogatives qui ont ponctue cette presentation designent certaines lirnites de cette recherche. Nous voulons les recapituler pour conclure, car certaines peuvent designer de nouvelles perspectives. Notre recherche a atteint son seuil car le secteur des activites humaines auquel nous nous sommes interesse est lui-meme limite. Mais les intersections que nous avons entrevues avec d' autres spheres de la culture materielle sont nombreu es. Ces intersections designent quelques voies possibles pour des approches technologiques interdisciplinaires. Le champ d'investigation chronologique et geographique que nous nous etions fixe etait limite. Un effort supplementaire de mise en perspective s' impose pour saisir la signification de certains evenements. Pour ce faire, il nous parait souhaitable de poursuivre I' elargissement de cette enquete (a des contextes geographiques voisins ou l' on per~oit des echos des faits que nous avons decrits, a des periodes proches durant lesquelles se sont epanouies certaines tendances que nous n'avons pu percevoir que de fa~on incomplete). Parmi les evenements que nous avons decrits, nombreux sont les faits d' ampleur supra-regionale. Cette ampleur peut designer de veritables phenomenes de diffusion, quand i1 s'agit d'idees techniques difficiles (par exemple, les grandes tendances qui caracterisent les debitages laminaires du Magdalenien ou ceux de l'extreme fin du Tardiglaciaire); cette ampleur peut

temoigner de simples convergences, quand ii s'agit d'idees simples (par exemple, les principaux choix qui marquent l'achevement de l'azilianisation) . Cette dimension supraregionale, qu'atteignent pour diverses raisons certains choix techniques, explique en partie les difficultes que nous avons a cerner des identites regionale specifiques dans ce domaine. II reste que l'adoption de ces choix semble connaitre des decalages chronologiques d'une region a l'autre. Mettre en evidence ces decalages est un moyen qui s' offre parfois pour eclairer des specificites regionales. En outre, des decalages peuvent egalement etre per~us entre les rythmes d' evolution des differents faits culturels (habitudes relatives a l' outillage, choix concernant le debitage, structuration economique, systemes ideologiques ... ). II devient meme pos ible d'entrevoir des decalages entre ces rythmes et ceux des changements climatiques et environnementaux . Mesurer ces decalages constitue a notre avis un des objectifs d'une dernarche paleohistorique, lorsqu'elle pretend rehabiliter la dynamique culturelle qui sous-tend certaine innovations et depa ser Jes interpretations parfois deterministes proposees pour interpreter les mutations socio-economiques qui ont accompagne la transition vers l' Holocene. En fin, l' autre limite de notre etude est inherente au parti pris que nous avons choisi en privilegiant l' approche diachronique. Les approches sur "le temps court" sont plus que jamais indispensables. A l' issue de cette enquete, nous tenons a souligner que le prolongement de cette perspective dialectique nous semble prometteur. De futures approches conjointes, menees a la fois sur le temps court et sur le temps long, ne peuvent que s'enrichir mutuellement et perrnettront probablement d' investir partiellement les echelles chronologiques intermediaires. C' est en ordonnant les faits sur l'echelle du temps long que les choix fondateurs d'une identite prennent une valeur distinctive, par contraste avec d'autres choix possibles attestes archeologiquement; c'est sur le temps court que l' on peut apprecier le plus finement la valeur fonctionnelle, economique, voire sociale de certains choix techniques. Ace titre, l'exceptionnelle preservation de certains gisements plus recents que le Magdalenien, comme Le Closeau (Bodu 1995), laisse augurer de fructueuses comparaisons, prenant en compte plusieurs aspects de la culture materielle et s'inscrivant done dans une perspective palethnographique plus globale.

REFERENCES Audouze, F., 1987. The Paris Basin in Magdalenian times. In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer, pp. 183-200. New York: Academic Press. Audouze, F., 1992. L' occupation magdalenienne du Bassin Parisien. In Le peuplement magdalenien: paleogeographie physique et humaine, pp. 345-355. Actes du Collogue International de Chancelade 10-15 Octobre 1988. Paris: Editions du C.T.H.S. 104

LES GROUPES HUMAINS AU TARDIGLACIAIRE

Audouze, F., et J.G . Enloe, 1991. Subsistence strategies and economy in the Magdalenian of the Paris Basin , France . In The Late Glacial in North-W est Europe: Human Adaptation and Environmental Change at the End of the Pleistocene , edited by R.N .E. Barton , A.J. Robert, et D.A . Roe, pp. 63-71. Research Report 77. London : Council for British Archaeology. Audouze, F.C. Karlin, D. Cahen , E. Croisset , P. Coudret , M. Larriere , P. Masson , M. Mauger, M. Olive, J. Pelegrin , N . Pigeot, H. Plisson, B. Schmider, et Y. Taborin, 1988. Taille du silex et finalite du debitage dans le Magdalenien du Bassin parisien. In De la Loire a l'Oder: Les civilisations du Paleolithique final dans le Nord-Ouest europ een, edited by M. Otte, pp. 55-84. Actes du Colloque International de Liege Decembre 1985 . Etudes et Recherches Archeologiques de l'Universite de Liege 25 et British Archaeolo gical Report s International Series 444 . Liege : Univer site de Liege et Oxford: British Archaeological Reports. Barton , R.N.E. , 1991. Technological innovation and continu ity at the end of the Pleistocene in Britain . In The Late Glacial in North -West Europe: Human Adaptation and Environmental Change at the End of the Pleisto cene, edited by R.N .E. Barton , A.J. Robert, et D.A. Roe, pp. 234-245 . Research Report 77. London: Council for British Archaeology. Barton, R.N.E . (editor), 1992. Hengistbury Head , Dorset. Volume 2: The Late Upper Palaeolithic and Early Mesolithic Sites. Oxford University Committee for Archaeology Monograph 34. Oxford: Oxford University Press. Bodu, P., 1993. Anal yse typo-technologique du materiel lithique de quelques unites du site magdaleni en de Pincevent (Seine-et-Marne): applications spatial es, economiques et sociales . These de Doctorat. Universite de Paris I. Bodu, P., 1995. Le site a Federmesser du "Closeau" a RueilMalmaison (Hauts-de -Seine) . Notae Prehistorica e 15:45-49. Bodu, P., et B. Valentin , 1993 . Nouveaux resultats sur le site tardiglaciaire a pieces machurees de Donnernarie-Dontilly (Seine-et-Marne). Prehistoire Europeenne 4:85-92. Basinski, G., M. Street, et M. Baales (editors), 1995. The Palaeolithic and Mesolithic of the Rhineland. In Quaternary Field Trips in Central Europe : Volume 2, edited by W. Schirmer, pp. 829-999. Mi.inchen: Verlag F. Pfeil. Bridault, A., et C. Bemilli, 1999. Les occupations magdaleni ennes de Marolles-sur-Seine/Le Grand Canton (Seineet-Marne): La chasse et le traitement des animaux. In Occupations du Paleolithique superieur dans le sud-est

du Bassin parisien , edited by M. Julien et J.-L. Rieu, pp. 49-64. Documents d' Archeologie Fran~aise 78. Paris: Editions de la Maison des Sciences de l'Homme. Celerier , G. (editor), 1993. L' abri sous roche de Pont d' Ambon a Bourdeilles (Dordogne). I: Technologie de l' outillage lithique taille ; II: Inventaire et typometrie des pointe s aziliennes; III: Analyse technologique du materiel de la couche 3B; IV: Le materiel lithique non taille . Gallia Prehistoire 35: 1- 136. Chollet , A., P. Fouere, A. Hantai', et G. Lelicon, 1999. L' evolution des choix techniques et economiques entre le Magdalenien superieur et I' Azilien: I' exemple de la grotte du Bois-Ragot (Vienne). In Actes du veme Congres UISPP, Commission XII, Epipal eolithiqu e et Mesolithique en Europe (Grenoble, France), edited by A. Thevenin et P. Bintz , pp. 275-285. Paris : Editions du C.T.H.S. Conkey , M., 1987. Interpretative problems in hunter-gatherer regional studies: Some thoughts on the European Upper Palaeolithic. In The Pleistocene Old World. Regional Perspectives, edited by 0. Soffer, pp. 63- 77. New York: Academic Press . De Bie, M., 1997. L' industrie a Federmesser de Rekem (Belgique ): datation, technologie lithique , et repartition spatiale. In Chronostratigraphie et environnement des occupations humaines du Tardiglaciaire et du debut de /'Holoc ene en Europe du Nord-Ouest, edited by J.-P. Fagnart et A. Thevenin, pp. 381-396. Actes du 119eme Congres National des Societe s Historiques et Scientifiques Amiens Octobre 1994. Paris: Editions du C.T.H.S. Fagnart, J.-P., 1997. La.fin des temps glaciaires dens le Nord de la France. Memoires de la Societe Prehistorique Fran~aise 24. Paris : Societe Prehistorique Fran~aise. Fagnart J.-P., et Plisson, H., 1997. Fonction des pieces machurees du Paleolithique de la Somme: caracteres traceologiques et donnees contextuelles. In Chronostratigraphie et environnement des occupations humaines du Tardiglaciaire et du debut de !'Holocene en Europe du Nord-Ouest , edited by J.-P. Fagnart et A. Thevenin, pp. 95-106. Actes du 119eme Congres National des Societes Historiques et Scientifiques Amiens Octobre 1994. Paris: Editions du C.T.H.S. Fischer , A., 1991. Pioneers in deglaciated landscapes: The expansion and adaptation of Late Palaeolithic societies in Southern Scandinavia . In. The Late Glacial in NorthWest Europe: Human Adaptation and Environmental Change at the End of the Pleistocene , edited by R.N.E. Barton, A.J . Robert, et D.A. Roe, pp. 100-121. Research Report 77. London : Council for British Archaeology . 105

VALENTIN

Franken, E., et S. Veil, 1983. Die Steinartefakte Gonnersdorf. Wiesbaden: Frantz Steiner.

von

Gaucher, G. (editor), 1996. Fouilles de Pincevent II: le site et ses occupations recentes (l 'environnement, l'Epimagdalenien et Les niveaux postglaciaires). Memoires de la Societe Prehistorique Frarn;aise 23. Paris: Societe Prehistorique Fran~aise. Hanta"i, A., 1997. Le "Belloisien" jusque sur Jes bords de la Loire: Les gisements de Muides-sur-Loire (Loir-et- Cher). Revue Archeologique du Centre de la France 36:5 - 22. Julien, M., 1989. Activites saisonnieres et deplacements des Magdaleniens dans le Bassin parisien. In Le Magdalenien en Europe: la structuration du Magdalenien, edited by M . Otte, pp . 177- 191. Acte du Collogue International de Mayence 1987. Etudes et Recherches Archeologiques de l'Universite de Liege 38. Liege: Universite de Liege. Julien, M., F. Audouze, D. Baffler, P. Bodu, P. Coudret, F. David, G. Gaucher, C. Karlin, M. Larriere, P. Masson, M. Olive, M. Orliac, N. Pigeot, J.-L. Rieu, B. Schmider, et Y. Taborin, 1988. Organisation de 1'espace et fonction des habitats magdaleniens du Bassin Parisien. In De la Loire a !'Oder. Les civilisations du Paleolithique final dans le Nord-Ouest europeen, edited by M. Otte, pp. 85-123. Actes du Collogue International de Liege Decembre 1985. Etudes et Recherches Archeologiques de l'Universite de Liege 25 et British Archaeological Reports International Series 444. Liege: Universite de Liege et Oxford: British Archaeological Reports. Julien, M., et J.-L. Rieu (editors), 1999. Occupations du Paleolithique superieur dans le sud-est du Bassin parisien. Documents d' Archeologie Fran~aise 78. Paris: Edition de la Maison des Sciences de l' Homme . Leroi-Gourhan, A., et M. Brezillon, 1972. Fouilles de Pincevent: Essai d'analyse ethnographique d'un habitat magdalenien (la Section 36). vueme supplement a Gallia Prehistoire. Paris: Editions du Centre National de la Recherche Scientifique. Madsen, B., 1993. Hamburgkulturens Flintteknologie I Jels. In lstidsjaegere ved Jelss(/)erne: Hamburgkulturen I Denmark, edited by J. Holm et F. Rieck, pp. 58-130. Skrifter fra AMT 5. Haderslev: Haderslev Museum.

Cahiers du Quaternaire 20. Paris: Editions du Centre National de la Recherche Scientifique. Pigeot, N., 1987. Magdaleniens d'Etiolles: economie de debitage et organisation sociale. xx:veme Supplement a Gallia Prehistoire. Paris: Editions du Centre National de la Recherche Scientifique. Pigeot, N., 1991. Entre Nature et Culture: valeur heuristique de la technologie lithique par des approches systemiques et cognitives. These d'Habilitation. Universite de Paris I. Pigeot, N., M. Philippe, G. Lelicon, et M. Morgenstern, 1991. Systemes techniques et essai de technologie culturelle a Etiolles: nouvelles perspectives. In Vingt-cinq ans d'etudes technologiques, pp. 159-168. Actes des x1 eme Rencontres Internationales d' Archeologie et d'Histoire d' Antibes Octobre 1990. Juan-Jes-Pins: A ociation pour la Promotion et Diffusion des Connaissances Archeologiques. Pigeot, N., et B. Valentin, a paraitre. Les chronologies de la prehistoire dans le Bassin Parisien au Tardiglaciaire: acquis recents, questions et bilan. In Chronologies geophysiques et archeologiques du Paleolithique Superieur. Actes du Collogue International de Ravello, 3-8 Mai 1994. Pion, G. (editor), 1990. L'abri de la Fru a Saint-Christophe (Savoie). Gallia Prehistoire 32:65-123. Plisson, H., 1985. Etudesfonctionnelles des outillages prehistoriques par I 'analyse des micro-usures: recherche methodologique et acheologique. These de 3emecycle. Universite de Paris I. Ploux, S., C. Karlin, et P. Bodu, 1992. D'une chaine l'autre: normes et variations dans le debitage magdalenien. Techniques et Culture 17-18:81-114. Rensink, E., 1995. On Magdalenian mobility and land use in north-west Europe: Some methodological considerations. Archaeological Dialogues 2:85-119. Rozoy, J.-G. (editor), 1978. Les derniers chasseurs: l' Epipaleolithique en France et en Belgique, Essai de synthese. Bulletin de la Societe Archeologique Champennoise 1. Charleville: Societe Archeologique Champennoise.

Olive, M., et N. Pigeot, 1991. Les tailleurs de silex magdaleniens d'Etiolles: vers }'identification d'une organisation sociale complexe. In La pierre prehistorique: Actes du seminaire du laboratoire de recherche des Musees de France I 3-14 Decembre 1990, edited by M. Menu et P. Walter, pp. 173-185. Paris: L.R.M.F.

Schild, R., 1984. Terminal Paleolithic of the North European Plain: A review of lost chances, potential and hopes. Advances in World Archaeology 3:193-274 . Schmider, B., 1984. Les industries lithique du Paleolithique superieur en /le-de-France. v1emeSupplement a Gallia Prehistoire . Paris: Editions du Centre National de la Recherche Scientifique .

Pelegrin, J., 1995. Technologie lithique: le Chatelperronien de Roe-de-Combe (Lot) et de La Cote (Dordogne) , 106

LES GROUPES HUMAINS AU TARDIGLACIAIRE

Schmider, B . (editor), 1992. Marsangy: un campement des derniers chasseurs magdaleniens sur les bards de l'Yonne. Etudes et Recherches Archeologiques de l'Universite de Liege 55. Liege: Universite de Liege. Taborin, Y. (editor), 1994. Environnements et habitats magdaleniens dans le centre du Bassin parisien . Documents d' Archeologie Frarn;aise 43. Paris: Editions de la Maison des Sciences de l'Homme. Valentin, B., 1995. Les groupes humains et leurs traditions au Tardiglaciaire dans le Bassin parisien: apports de la technologie comparee. These de Doctoral. Universite de Paris I.

Valentin, B., 1999. Techniques et cultures : Les chasseurscueilleurs de la fin du Tardiglaciaire au sud du Bassin Parisien . In Actes du veme Congres UISPP, commission Xll, Epipaleolithique et Mesolithique en Europe (Grenoble , France) , edited by A. Thevenin et P. Bintz, pp. 204-212. Paris : Editions de C.T.H.S . Valentin , B., P. Bodu , et M. Julien , sous presse. Lithic raw material economy in the Late Glacial of the Paris Basin: Case studies in the Magdalenian , Federmesser, and Long Blade Technology Hunter - Gatherers. In Lithic Raw Material Econom y in Late Glacial and Early Postglacial Western Europe, edited by L.E . Fischer and B.V. Eriksen. Ann Arbor: International Monographs in Prehistory .

107

8

Settling Down or Moving Around? The Development of Regional Traditions during the European Magdalenian Todd A. Koetje (Western Washington University)

ABSTRACT Intrasite spatial patterning in the southwest French Magdalenian is relatively diverse, but shows some clear intra-regional similarities and contrasts well with spatial patterning from Magdalenian sites in other well defined regions in Europe (the Paris Basin and southwest Germany for example). There is also a substantial amount of evidence supporting the notion of significant regional integrity during the Magdalenian in this area. Artistic styles, raw material distribution, site chronology and chronostratigraphy, and most recently occupational seasonality can all be used to argue that there is substantial within region movement and less extra regional movement during the Magdalenian. When considered together , these factors suggest the development of strong regional traditions within the Magdalenian. INTRODUCTION The Magdalenian time period in European prehistory has always involved an intriguing set of ideas. From the early 20 th century when l'Age du Renne (the Reindeer Age) was recognized as a distinct entity (Breuil 1912), unique types and patterns in stone tools, faunal remains, and sites have differentiated it from other cultures. Modern data and concepts of the Magdalenian have confirmed much of the initially recognized distinctiveness, while adding substantial complexity to our picture of the time period between 15,000 and 9,000 years ago. We can now discern and use patterning in such things ac site location, site seasonality, prey species migration patterns , raw material sourcing and distribution, similarity of artistic renderings, and internal patial organization of sites to address relatively detailed behavioral questions about the people who made up this culture. From the perspective of the other Upper Paleolithic industries, the Magdalenian has always been unusual. This became very clear with the de Sonneville-Bordes ( 1960) refinement of general Upper Paleolithic systematics in the middle of this century. Although eventually divided into six phases, primarily by the presence or absence of various type fossils, the early phases typically contain few of the more chronologically definitive bone tools and, in general, look very different from either a typological or technological point of view, as compared with the later phases. The middle phases are largely indistinct from one another when bone tools are absent, and the well defined lithic and harpoon type fossils from the later phases have occasionally been found on the same surfaces or in contexts dating far earlier than would be expected under the normal model (e.g., Bouvier 1969; Rigaud 1979). The fact that the earliest phases (0 and 1) are now commonly redefined as the Badegoulian industry illustrates these difficulties. There is, for example, still no single site where the early and later phases appear together in a stratified context. Outside of southwest France and the Perigord, where it was originally defined, it is not unusual to find even more striking phase discontinuities with no evidence for explanations such as gaps in dating or general abandonment of the area.

SPATIAL ORGANIZATION OF MAGDALENIAN SITES Recent synthetic studies of the spatial patterning found in Magdalenian sites in the region of southwe t France have suggested some broad similarities in the use of internal site space from this time period (Koetje 1991, 1994a, 1994b ). Although a diverse set of site has been examined, and their patterning does not match directly, when abstracted slightly and when internal features uch as the presence of cave and cliff wall boundaries , rock falls, etc. are controlled, fairly clear behavioral similarities are present. At least in a small sample , sites are very commonly occupied repetitively, and the same or very similar zones are used for at least broadly comparable activities, as inferred from similar lithic tool and bone distributions, as well as from more sparsely occurring feature distributions, refitting , and use-wear information . Less systematic examination of other Magdalenian sites in the region reveals similar pattern (i.e., Dufaure [Petraglia 1992; Straus et al. 1988]). These behavioral choices are clearly recognizable, even, or perhaps especially, in geographic situations like the Dordogne River valley in the Perigord region of southwest France, where there are very large numbers of caves and rock shelter .s . In the Dordogne and similar areas, choice of place should be relatively unconstrained, even with a high population density, because of the large number of possibilities close to any given fixed resource or set of fixed resources . It seems reasonable, then, to expect a wide distribution of sites with little repetition in occupying any specific place and, particularly, little immediate reoccupation. The archaeological signature of these expectations could be seen as a large number of discrete occupation layers in a wide variety of caves and rock sheltersideally, of course, with sterile layers separating them. In addition, some minor spatial overlap of functional zones between occupations within sites might be expected, but not rigid adherence to a single pattern of internal space organization . Several sites from pre-Magdalenian cultures exhibit one or more of these characteri tics; Le Flageolet I, Le Malpas, and Combe-Grenal, among others, all come to mind easily. In

KOETJE

some cases, Magdalenian sites such as Le Magdalene also show these patterns, to the extent that they can be discerned, although a substantial number of sites from the Magdalenian in the Dordogne show very different patterns. Magdalenian sites in general, and particularly late Magdalenian sites from this area, are infamous for relatively thick, often seemingly undifferentiated deposits, suggesting long-term or highly repetitive occupations. They commonly include pavements, remnants of walls, or other substantial architectural improvements to the living areas . Detailed spatial analyses of several of these sites shows striking similarity in overall internal spatial patterning and, in at least one case, clearly repetitive or redundant sets of tools and locational associations over great depth. More ca ual observation of a number of Magdalenian sites suggests very similar patterning. At Le Flageolet II, for example , thi internal consistency la ted for c. 1000 years; cross-cut several of the traditional chronological pha es which subdivide the Magdalenian; and remained unchanged despite substantial climatic change, changes in the type distribution of the tool assemblages beyond the temporal markers , and a shift from single hearth-focused occupations to an occupation containing a large pavement and no clear hearth (Koetje 1991, 1994a). I have used these patterns at Le Flageolet II to argue that Magdalenians were attached to extremely specific places on the landscape and that this narrow focus on place might be best explained ideologically rather than as an ecological or economic tie to some fixed resources in a particular area. This suggestion goes against the grain of typical explanations for site distributions on the landscape, particularly in the Paleolithic, where functional and geological factors prevail as explanations. White ( 1985), for example, has found a marked relationship in the Perigordian Magdalenian between large rock shelter and cave occupations and proximity to river fords and springs and to raw material sources. He explains these patterns as an adaptation to extensive hunting of herd animals, especially reindeer, at fords. In contrast, Rigaud and Simek ( 1987) have suggested that geologic formation processes rather than economic choices may be responsible for some or all of this perceived patterning. While these patterns clearly exist, and irrespective of the economic or geologic arguments that can be applied to explain them in a broad sense, given the number of potentially occupiable sites in the general area and even in the immediate vicinity of large sites, we can also ask why only these particular sites are occupied so intensively? Why not all of them? Why the internal consistency, and why few or no changes despite changes in climate and other factors? To my mind, these are separate issues that are not necessarily best explained by the larger scale general patterning that also exists. I am particularly intrigued by the co-occurrence of these factors. In the absence of any other variables, some of the patterning may be attributed to simple chance. That is, given a small overall number of occupations and a large number of suitable sites, some re-use would be expected and would look like inherently interesting variation against the background , despite being random. However, the

conjunction of re-occupation with striking similarity in internal organization, especially when it cuts across both conventional phase, typological, and climatic periods , suggests a need for explanations that can include both functional and geological conditioning, but also that is not restricted to them. ADDITIONAL FACTORS A number of interesting connections can now be made between factors such as internal spatial consistency in specif ic locations and wider regional and perhaps even transregional processes . For example, recent work with faunal remains and site seasonality by Gordon ( 1988) has suggested that yearly movement patterns by Magdalenians were occurring primarily within the southwest region of France, rather than between the southwest and central or northern regions . In order to intercept seasonal reindeer movements , Gordon (1988) goes so far as to suggest a clear easonal migration pattern between the Pyrenean foothills and the Dordogne. While the details of Gordon's arguments are controversial, regardless of whether Magdalenian s were practicing herd following or interception strategies, the data uggest that their movements were generally restricted to the Perigord and regions south, rather than between the Perigord and northern or central France . Using a very different set of data, some categories of mobil iary art have been interpreted as showing such substantial similarity in form and technique as to suggest that a very limited number of people were responsible for producing it. They seem to have moved exclusively between the Dordogne and the Pyrenees, to the immediate south (Bahn 1982, 1984). Even if this pattern is due to trade contacts rather than actual movement, the exclusivity of the pattern suggests uniquely close ties between the two regions, as well as substantial homogeneity and stability of the groups inhabiting each region. Generally, more limited arguments have been made for similarity in style and content of parietal art in these two regions, implying the direct movement of artisans . Similar patterns are not evident between either the Perigord or the Pyrenees and other regions. Raw material sourcing can also be used to address these issues, although the data is somewhat scanty. Larick (1982) and Geneste (Geneste and Rigaud 1989) provide perhaps the best known examples of raw material sourcing studies in southwest France. Although neither address the Magdalenian specifically, both show a fairly consistent pattern-namely, that the overwhelming majority of stone tools were made on locally available materials. Exotics from extra-regional sources are extremely rare , and even substantial intra-regional movement of specific materials is uncommon . Although there are no similar, large-scale synthetic studies for the Magdalenian in this area, smaller scale studies have been conducted as part of single-site excavations (e.g., Gaussen 1980; Simonnet 1985). Unfortunately, this evidence is hard to interpret. Although extra-regional exotics would be expected if there was substantial interregional movement, their absence 110

THE DEVELOPMENT OF REGIONAL TRADITIONS

or their occurrence in only low numbers might easily be explained because of the general availability of high or very high quality raw materials and simple foreknowledge. Why carry exotics hither and yon when you can be assured of appropriate raw materials at your destination? Especially when the high value placed on extremely well made stone tools, as is likely to have been the case in the preceding Solutrean, has apparently been superseded by an emphasis on bas-relief and other three dimensional sculpture techniques in mobiliary art on bone by the time of the later Magdalenian? While it seems unreasonable to suppose that earlier peoples did not practice seasonal rounds within a limited geographic area, the Magdalenian is the first time period in which this pattern can be clearly seen. It is particularly striking that there are several sources of data that reinforce one another by showing this type of pattern. While the patterns evident in the Perigord are not precisely replicated, and there are some important differences in the details of occupation timing, industry, and industrial succession within other regions, examining the Magdalenian from the greater Paris Basin and southwest Germany suggests at least broadly similar trends.

THE PARIS BASIN MAGDALENIAN Topographically, the Paris Basin is very different from the southwest region . Rather than a series of deeply cut river valleys with many exposed caves and rocks shelters separated by high plateaus, the basin proper is quite flat, bordered by low hills on three sides, and contains a number of lc;trgemeandering rivers. The Magdalenian of the Paris Basin is generally somewhat later than that of the Perigord, and the area is smaller with many fewer sites. Nonetheless, there is quite good evidence for consistent use of the area for very specific purposes. Most of the well known sites (Pincevent, Verberie, Les Etiolles, etc.) are open-air occupations near the major rivers, often spread over a very large area. As a group, they can be characterized as aggregations of small reindeer hunting and/or processing camps. They clearly represent repeated use of the same general location for the same or similar functions and were almost certainly occupied on a seasonal basis, presumably in order to intercept moving herds at strategic hunting points. In general, art is absent or extremely rare from these sites, and lithic raw materials are essentially local, deriving almost exclusively from deposits in the surrounding footslopes. Smaller sites are not common and tend to be found away from the major rivers, although this is not an observation firmly based on extensive survey data. Like most areas of Europe, there has been little emphasis on systematic survey in the Paris Basin. Because of these patterns and the overall size of the general area, Audouze ( 1987), and others conclude that only a fragment of the subsistence system in the Magdalenian is monitored here. Thus , despite the clear evidence for repeated use of specific locations, there is no good indication of where the Magdalenians were coming from, or where they were returning. It is interesting to compare explanations for the strong sitelocation patterning from the Perigord and the Paris Basin .

White (1985) and Audouze (1987) have made essentially identical arguments-i.e., that site location is a function of herd interception strategies focused on using river or stream crossings. I wonder, though, if the same argument cannot also be applied to the Paris Basin cases, as it has been for the Perigord? Why these fords so repetitively? There are many more than four or five pos ible crossing locations for each of the major rivers in the area, and, assuming that the animals were moving roughly north-south through the area, all of the major rivers and streams would have to be crossed. I'm particularly struck by the fact that this pattern appears where there are no major terrain features to guide or channel herds into particular paths. Modern caribou generally travel in small groups along paths that are erratic at best. They do congregate at crossing points along major rivers, although these are not always fords, sometimes returning to the same one over many years. The more extreme cases are generally in topographically bounded circumstances, and, even then, there i al ways variation in time and space. If we can extrapolate this behavior into the past, we would expect a generally slow filtering of small groups through a broad area. Certain crossing points or fords would certainly be commonly used, but which ones and when, especially without clear topographic funneling and with many to choose from, would be difficult to determine. Hunting these small groups of animals by consistently positioning yourself at any particular crossing point, when a number of such points are available, may not yield very predictable results, yet this is apparently what we see. If there are so few crossing points or so many reindeer that this strategy is reliable, we can still ask why this ford was chosen, as the choice becomes even more clearly arbitrary. Habit, aesthetics, or, following Isaac (1981 ), comfortable shade trees (although probably not very applicable to the last glacial period) have all been offered as suggestions to answer this type of question. Perhaps they are all involved in some notion of what we would call ownership or territoriality-that is, fixation on a specific place by one particular group.

THE MAGDALENIAN IN SOUTHWEST GERMANY If we look quickly at this time period in southwest Germany, the topographic situation is more similar to southwest France than the Paris Basin, and the Magdalenian is distinct from that of either the Perigord or the Paris Basin. Especially in the Later Magdalenian, however, some broadly similar patterns begin to emerge. Artistic and lithic traditions indicate substantial internal homogeneity , with no clear indications of extensive trade or migration with the other areas. Site patterning on the landscape is perhaps not as strong as in the Paris Basin or the Dordogne, yet large sites are clearly typical in river valleys and may well be associated with fords and seasonal-intercept hunting strategies. In addition, there are some large sites that demonstrate repetitive occupation over a lengthy time span. To my knowledge , none has demonstrated the kind of internal consistency and redundancy of occupation

111

KOETJE

that sites in the Dordogne have, but I'm not sure that it has been looked for, either. Certainly, the larger patterns are suggestive of the same general processes that can be een in the other regions (see Weniger 1987). There are very few, if any, indications of contact between the Perigord, the Paris Basin, and southwest Germany during the Magdalenian, and certainly no clear indication of regular migration among or between these areas. Thus, we have the first really multifaceted set of evidence for a clear trend towards distinctive regional groups within the same broad culture, apparently pursuing very similar subsistence strategies and, in the process , becoming attached to specific places on the landscape. PARADIGMS AND EXPLANATIONS One of the more striking similarities in the manner in which much of the Magdalenian archaeological record, and . indeed, much of the Paleolithic record as a whole, has been interpret ed is the heavy dependance on functional and economic models. Of course, these approaches are a basic part of processual archaeology per se, the paradigm which dominates analy ses of this time period, and the basis for much of the enthnoarchaeological and ethnographic work that is responsible for our vision of the Paleolithic (e.g., Binford I 980, etc.). Without entering a debate regarding the merits of processual and post-processual approaches, it may be useful to apply models that take into account the so called irrational behavior of social groups ( Cowgill 1993; Mi then 1990, 1991 ), or at least the unintended (thus irrational?) consequences of their behavior (Boudon 1977; Elster 1978) in trying to understand the Magdalenian record. Outlining such a model is beyond the scope of this paper, but perhaps the reasons that some of my questions above are difficult to answer is that we have to look beyond assumptions that detailed positioning on the landscape is a simple functional process based on rational economic decision making or on choices among foraging and collecting feeding strategies. To ground these ideas a little more firmly, perhaps choice of cave and shelter sites in the Dordogne is loosely based on the following criteria, in no particular order: 1) economic utility, i.e., resources must be accessible to some varying degree; 2) history, i.e., where did we stay last time, and what have we brought along this time? 3) social interaction, i.e., are we in competition with other groups or amongst ourselves? 4) group composition, i.e., who is along? 5) encounter success, i.e., where are the animals and plants, and what have we gotten so far?; and 6) intangible elements, i.e., magical, religious, and/or ritual assumptions and patterns. I would like to argue that strong patterns in repetitious positioning, site utilization, and internal organization violates expectations based on a group making a series of independent, rational economic choices alone and suggests very strong influences from the other factors.

repetitive site occupation in earlier time periods? Are these patterns something unique or uniquely dominant in the Magdalenian, or did they exist previously? To what extent do differential preservation and formation processes play roles in these patterns? Particularly since Bender's (1985) work, it is not uncommon to make broad comparisons between the Archaic traditions in the New World and the Mesolithic traditions in Europe. But if the New World Archaic is characterized as the period in which regional traditions in lithic technology and subsistence came to dominate life, po sibly rich sources of comparison are being overlooked by ignoring the Magdalenian. Information from the Azilian and early Mesolithic is fairly canty, but it is not po sible to rule out the possibility that regional traditions that become clear in the later Mesolithic became established prior to the advent of the Holocene.

REFERENCES Audouze, F., 1987. The Paris Basin in Magdalenian times . In The Pleistocene Old World: Regional Perspectives, edited by 0 . Soffer, pp. 183-200. New York: Plenum Press. Bahn, P., 1982. Inter- ite and inter-regional links during the Upper Paleolithic: The Pyrenean evidence. Oxford Journal of Archaeology 1 :247-268. Bahn, P., 1984. Pyrenean Prehistory: A Paleoeconomic Survey of the French Sites. Warminster: Aris and Phillips. Bender, B., 1985. Prehistoric developments in the American Midcontinent and Brittany, Northwest France. In Prehistoric Hunter-Gatherers: The Emergence of Cultural Complexity, edited by T.D. Price and J. Brown, pp. 21-58. New York: Academic Press . Boudon, R., 1977. The Unintended Consequences of Social Action. London: Macmillan. Bouvier, J.-M., 1969. Existence de Magdalenien Superieur sans harpon: Preuves stratigraphiques. Comptes Rendus de l'Academie des Sciences de Paris 268:2865-2866.' Breuil, H., 1912. Les subdivisions du Paleolithic Superieur et leur signification. Congres International d'Anthropologie et Archeologies Prehistorique 7:165-238. Cowgill , G., 1993. Distinguished lecture in archaeology: Beyond criticizing New Archaeology. American Anthropologist 95:551-573. Elster, J., 1978. Logic and Society: Contradictions and Possible Worlds. New York: John Wiley. Gaussen, J., 1980. Le Paleolithique Superieur de Plein Air en Perigord. XIveme Supplement a Gallia Prehistoire . Paris: Centre National de la Recherche Scientifique .

Of course, many important questions remain to be addressed, perhaps most striking among them is this one: What about 112

THE DEVELOPMENT OF REGIONAL TRADITIONS

Geneste, J.-M., and J.-Ph. Rigaud, 1989. Matieres premieres lithiques et occupation de I' espace. In Variations des Paleomilieux et Peuplement Prehistorique , edited by H. Laville, pp. 205-218. Paris: Centre National de la Recherche Scientifique. Gordon, B., 1988. Of Men and Reindeer Herds in French Magdalenian Prehistory. British Archaeological Reports International Series 390 . Oxford: British Archaeological Reports. Julien, M., 1981. Les Harpons Magdaleniens. XVIIeme supplement a Gallia Prehistoire. Paris: Centre National de la Recherche Scientifique. Isaac, G.Ll., 1981. Stone age visiting cards: Approaches to the study of early land-use patterns. In Patterns of the Past: Studies in Honour of David L. Clarke, edited by I. Hodder, G. Isaac, and N. Hammond, pp. 131-155. Cambridge: Cambridge University Press.

Petraglia, M.D., 1992. Stone artifact refitting and formations processes at the Abri Dufaure: An Upper Paleolithic site in SW France. In Piecing Together the Past: Applications of Re.fitting Studies in Archaeology, edited by J.L Hofman and J.G. Enloe, pp. 163-178. British Archaeological Reports International Series 578. Oxford: British Archaeological Reports. Rigaud , J.-Ph., 1977. Apropos des industries Magdaleniennes du Flageolet. In la Fin des Temps Glaciaires en Europe, edited by D. de Sonneville-Bordes, pp. 467-469. Paris: Centre National de Ia Recherche Scientifique. Riguad, J.-Ph., and J. Simek, 1987. Arms too short to box with God: Problems and prospects for Paleolithic prehistory in Dordogne, France. In The Pleistocene Old World: Regional Perspectives , edited by 0. Soffer, pp. 47-62. New York: Plenum Press.

Koetje, T.A., 1991. Dealing with three dimensional site structure: An example using simulated archaeological levels and the Magdalenian site of le Flageolet II (Dordogne, France). Journal of Field Archaeology 18:187-198.

Simonnet, R., 1985. Le silex du Magdalenien final de Ia grotte des Eglises dan le basssin de Tarascon-sur-Ariege. Bulletin de la Societe Prehistorique de l 'Ariege 40:71-97.

Koetje, T.A., 1994a. Human behavior and formation processes: The view from Le Flageolet II. In Formation Processes in the Old World Archaeological Record, edited by M.D. Petraglia, and 0. Bar-Yosef, pp. 113-124. Madison: Prehistory Press.

de Sonneville-Bordes, D., 1960. le Paleolithique Superieur en Perigord. Bordeaux: Delmas.

Koetje, T.A., 1994b. Intrasite spatial structure in the European Upper Paleolithic. Journal of Anthropological Archaeology 13:161-169. Larick, R., 1983. The Circulation of Solutrean Foliate Point Chert: Residential Mobility in the Perigord. Ph.D. Dissertation. SUNY, Binghamton, New York. Mithen, S., 1990. Thoughtful Foragers: A Study Of Prehistoric Decision Making. Canbridge: Cambridge Universtiy Press.

Mithen, S., 1991. A cybernetic wasteland? Rationality, emotion, and Mesolithic foraging. Proceedings of the Prehistoric Society 57 :9-14.

Straus, L.G., K. Akosima, M.D. Petraglia, and M. SeronieVivien, 1988. Terminal Pleistocene adaptations in Pyrenean France: The nature and role of the Abri Dufaure site (Sorde-1' Abbaye, Les Landes). World Archaeology 19:328-348. Weniger, G.-C., 1987. Magdalenian settlement pattern and subsistence in Central Europe: The southwestern and central German cases. In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer, pp. 201-216. New York: Plenum Press. White, R., 1985. Upper Paleolithic land-Use in the Perigord: A Topographic Approach to Subsistence and Settlement. British Archaeological Reports International Series 253. Oxford: British Archaeological Reports.

113

9 Readaptation: Changes in Magdalenian Subsistence and Social Organization James G. Enloe (University of Iowa)

ABSTRACT While lithic technology and typology appear to link Magdalenian culture over a large area of western Europe , there are differences apparent in site size, duration of occupation, production of art, fauna) exploitation , and subsistence among the various regions occupied during the Magdalenian. Many of these differences may be linked to adaptations to slightly or radically different environments . The resource species available in these different environments may actually be the same , but the patterns of exploitation of those resources may constitute significantly different adaptations. This may be manifest archaeologically by more variation in functions between sites. Mellars (1994) suggested that Magdalenians pioneered regions of northern Europe at the end of the Pleistocene in order to maintain exploitation of reindeer and horse in open environments as the glaciers retreated. Magdalenian occupations of these regions, such as the Paris Basin , differ from those of the classic sites in the Perigord . There are no large aggregation sites. Sites are small and highly seasonal, suggesting increased mobility . Elaborate art is much more poorly represented. These suggest decreased social complexity and changes in subsistence. Faunal data from sites in the Paris Basin are contrasted with data from sites in the Perigord to evaluate differential site function and how different site function s relate to situational adaptation in new environments. INTRODUCTION The Magdalenian has often been characterized as the greatest fluorescence of the Upper Paleolithic of western Europe. It is best known for its spectacular displays of parietal art, as at Lascaux and Altamira. Among the earliest scientific excavation in the history of Paleolithic archaeology was the work of Lartet and Christy ( 1877) at La Madeleine, spurred in large part by the search for mobiliary art work. Not all Magdalenian sites offer such spectacular finds , of course. The quotidian tasks of daily life resulted in the accumulation of lithic debitage , animal bones, fire-cracked rock, and other cultural debris that represent the bulk of the archaeological record for this, as for any other, culture or period . Further, regional differentiation may have shaped the degree to which art treasures played a major role in the people's interactions or adaptations. Therefore, it may be more productive to look at subsistence, which probably varied more from region to region than would have cultural practices that served to integrate people. The Magdalenian is most heavily represented in the Perigord. It has been characterized as pioneering, spreading out from its center of origin to occupy previously unoccupied territories, including the Pyrenees to the south and the Paris Basin to the north. We will concern ourselves here with adaptation to new environment to the north, the Paris Basin. Upper Paleolithic occupation in the Paris basis is sparse, compared to that of the Perigord, with very few sites dating to the early Upper Paleolithic (Schmider 1984). The pathway from the Perigord to the Paris Basin can be traced in Badegoulian sites across the upper valley of the Loire, at the Abri Fritsch, La Pluch,

Saint-Fiacre, Grand-Pressigny Silo, and La-Chapelle-SaintMesmin, and into the southern reaches of the tributaries of the Seine, at Beauregards , le Grand Surplomb, Deuxieme Redan, and Ballancourt-sur-Essone (Hemingway 1980; Schmider 1984) . The occupation of the Paris Basin is much better known during the Late Magdalenian, at such sites as Pincevent, Etiolles, Marsangy, and Verberie (Audouze 1987; Enloe and Audouze 1997). While the lithic industries show clear technological (Le Licon 1997) and typological (Otte 1997) continuity and consistency with the Perigord (Laville et al. 1980), the Magdalenian occupation of the Paris Basin differs from that of the classic sites in the Perigord. First of all, the Paris Basin lacks the dense concentration of large rock shelters with preserved stratified deposits that have attracted archaeologists for over a century. It might be suggested that much of the archaeological richness of the Perigord is a function of conservation and investigation , rather than greater abundance in food resources for Paleolithic hunter-gatherers; we cannot address that question here. At any rate , the richness of the Paris Basin lies in well preserved open-air sites. There appear to be no large aggregation sites. Sites are small and highly seasonal, suggesting increased mobility. Elaborate art is very poorly represented. Art is restricted to a very few mobiliary pieces, often not elaborated to any degree . For example, at Pincevent, there are really only two pieces of representational art: a rudimentary sketch of a cervid head on a pierced baton and a horse head scratched on a flint core. At Verberie, one single pierced fox tooth and one pierced shell comprise the totality of adornment or art. These might suggest decreased social complexity and changes or differences in subsistence.

ENLOE

Large site size has been used to indicate higher population during the Upper Paleolithic, with concomitant inferences for social complexity, particularly for the Magdalenian (Mellars 1973; Boyle 1996). White (1982:171) suggested that Mellars (1973) exaggerated the size of Upper Paleolithic sites, especially for the Magdalenian, and he recorded 57 sites less than 500 m2, five sites between 500 and 1000 m2, and eight sites 2 larger than 1000 m • Boyle (1996:483) used a Rank Size Rule analysis to argue that a few very large sites dominated the Magdalenian settlement system in the Vezere valley of the Perigord. Were large sites, generally identified as aggregation sites (Conkey 1980), necessarily occupied by large numbers of people? As White (1982 :171) pointed out in reference to Upper Paleolithic site size, Yellen's (1977:125) ethnoarchaeological research indicates a trong correlation between overall site size and length of occupation. We cannot , of cour e, directly compare the sizes of rock helter and cave site with those of open-air sites . Radically different constraints on the use of space pertain. La Madeleine's size has been estimated 2 2 between 3000 m (White 1985:185) and 10,000 m (Laville et al. 1980 :137; de Lumley 1984:303). Reignac extend over an 2 area of 2000 m (Boyle 1993:151). Limeuil is an open-air site 2 reported to cover 450 m (White 1980:286). Sites in the Paris Basin tend to be either small sites or sites with low popula2 tion. Verberie covers less than 400 m • Although level IV20 of Pincevent has been excavated over 4500 m2, this represents the living space of half a dozen households, probably 25 to 50 people at most. This is surely not an aggregation site in Conkey's (1980) sense. Mellars (1994:76) has suggested that Magdalenians pioneered the Paris Basin at the end of the Pleistocene in order to maintain exploitation of reindeer and horse in open environments as the glaciers retreated. This uggestion implies that the Magdalenians were attempting to continue to exploit those species in the same manner that they had in the Perigord, as a key element in their culture. It would be useful to examine the patterns of subsistence by utilizing recent analyses of fauna) assemblages of several late Magdalenian sites in the Perigord and in the Paris Basin. While Mellars ( 1973, 1989) has contended that specialized hunting of reindeer was an intrinsic part of the Upper Paleolithic from its very beginning, recent work has suggested that apparent specialization may be more related to climatic regimen than to overt strategy (Enloe 1993). The question that must be posed is not "What percentage of a people's food is derived from a given species?" but rather "What is the availability of a species in the region, and how are people exploiting them?" (Enloe 1998). To address this question, this article will draw on selected levels of sites from the late Magdalenian of the Perigord and Paris Basin, in which reindeer was clearly the predominant species exploited. Several Late Magdalenian sites from the Perigord have been recently analyzed. Those considered include La Madeleine (Boyle 1994, 1997), Reignac (Boyle 1997), Limeuil (Boyle 1997), and Flageolet II (Deplano 1994). These sites are compared to Late Magdalenian sites in the Paris Basin which have good preservation of fauna: Pincevent (David and Enloe 1993), Verberie (Audouze and

Enloe 1997; Enloe 1997; Enloe and David 1997), and Tureau des Gardes (Bridault 1996, 1997).

MAGDALENIAN IN THE PERIGORD

In the Magdalenian occupation of couche IX of Flageolet II (Deplano 1994), the fauna is numerically dominated by reindeer, with 17 individuals making up 96% of the specimens; minimal percentages of remains indicate the presence of three chamois, one red deer, one bovid, one horse, and one saiga antelope. Skeletal representation is dominated by the mandible, metatarsal , tibia, femur, humerus, and metacarpal. Accurate counts of elements are difficult to discern. Deplano ( 1994:35) considered that taphonomic and excavation recovery problems probably accounted for most of the differential repre entation and concluded that entire carcasse had been transported to the ite. At least two individuals are repre ented by fetal bones. Crania with and without antlers are present. Males and females appear to be represented. Filleting marks are mo t common , followed by skinning, and, finally, by disarticulation. Long bones were systematically fractured for marrow extraction. Deplano ( 1994:80) concluded that these · data indicate, first, a specialized hunting strategy, primarily for nutrition, but also for raw materials from skin, tendons, and antlers; and, second, year-round occupation by a small group of people. At La Madeleine (Boyle 1994), the Magdalenian VI assem blage is composed of 94% reindeer, representing at least 86 individuals; eight horses are present. Of the reindeer, the metatarsal, astragalus, humerus, metacarpal, mandible, and scapula are best represented. Seasonality for the reindeer, according to Gordon (1988:64), was primarily spring and winter; Burke's (1993: 147) analysis of the horse teeth indicated an overwhelming summer mortality. At Limeuil, 92% of the faunal assemblage consists of reindeer, at least 158 individuals, with astragalus, humerus, tibia, metacarpal, scapula, and radiocubitus best represented (Boyle 1993:142). I have no seasonality information for Limeuil. At Reignac, 92% is also reindeer, best represented by metatarsal, tibia, scapula, mandible, metacarpal, and humerus (Boyle 1993: 152). There were at least 307 reindeer and six horses. Seasonality for the reindeer is primarily winter (Gordon 1988:64); only one horse tooth gave a legible reading, for summer (Burke 1993:147). Boyle (1994, 1997) examined carcass management strategies at the late Magdalenian sites of Reignac, Limeuil, and La Madeleine, considering contrasts between the treatment of reindeer and that of other species. In the Late Magdalenian, reindeer predominate, with up to ten times the number of individuals of any other species; the most prevalent secondary species is often horse. She noted separate strategies for each species, with consistent patterning of reindeer exploitation at the sites she examined . She also noted strong statistical correlations in body part representation with the additional Magdalenian sites of Gare de Couze and Les Eyzies (Boyle 1994:59). She concluded that there was selection primarily of high general utility parts. Gourmet utilization curves (sensu 116

READAPTATJON

Binford 1978) are seen for the major species , where one species clearly dominates the assemblage-reindeer in all cases. Boyle (1994) interpreted this as reflecting the abundance of reindeer, which are relatively predictable on a seasonal basis. Although there are more filleting marks seen on reindeer than on secondary species, she attributed their relative paucity to the abundance of meat; when there were " ... multiple kills, some parts of carcasses may not have been processed at all" (Boyle 1997:290). Secondary species, especially horse, exhibit more complete skeletal representation of fewer individuals, interpreted as bulk curves (sensu Binford 1978), along with more complete processing and fragmentation. Burke (1993) reported a marked seasonality in horse dentition, indicating kills in summer. Boyle ( 1994, 1997) calculated that the total meat weight of all levels at La Madeleine derived from reindeer was 5200 kg and horse was 5000 kg, a result similar to Spiess' (1979) meat weight calculations for relative dependence on reindeer and bovid from earlier levels at the Abri Pataud. In sum, these site suggest year-round occupation, with seasonal shifts in not only the species exploited, but also in the manner in which they were used. MAGDALENIAN IN THE PARIS BASIN At Pincevent, 98% of the fauna} assemblage is reindeer, representing at least 52 individuals, with cranium, mandible, metacarpal, metatarsal, radiocubitus, pelvis, and tibia very well represented. Significantly missing are vertebrae, suggesting that this was a transported assemblage, but the rest of the skeleton is well represented, which suggests that the distance for transport was not very far. Cutmarks are approximately evenly divided between disarticulation and filleting (Enloe 1991). Seasonality is indicated by male and female bois de massacre and is particularly well documented by deciduous dentition of first- and second-year individuals, clearly peaking in the fall (David and Enloe 1993; Enloe and David 1997). This suggests exploitation during the migration. Abundant other data indicate that Pincevent was a residential site, with multiple households (Enloe 1992); it was a consumption location for prey from fall hunting . Verberie is also characterized by over 98% reindeer in the fauna} assemblage. Although analyses are not yet complete, skeletal element representation is similar or slightly higher than at Pincevent. A significant difference is the presence of vertebrae. Numerous articulated vertebral columns are present. These strongly suggest initial butchering. The seasonality is again clearly indicated by antlers and particularly by deciduous dentition, which would have been restricted to a very short period in the fall (Enloe 1997). This site is only shallowly buried; abundant root vermiculation has obscured the surfaces of most of the bones, making cutmark determination difficult or unreliable. A few cutmarks indicating disarticulation and filleting have been discerned. The seasonality and body parts strongly suggest that Verberie wa a hunting camp for a fall migration interception.

2

Tureau des Gardes at Marolles near Pincevent, 75 m over four successive occupations (Bridault 1996), probably dates to a slightly later period than Pincevent or Verberie. This site, like those in the Perigord, is dominated by two species, over 98% of the fauna} assemblage, including 33 horse and 23 reindeer. Skeletal representation of horse includes almost all elements, with mandible, tibia, metacarpal, and metatarsal most frequent. Reindeer element representation also includes most of the skeleton, with humerus, radius, metacarpal, tarsal, and metatarsal most frequent. Scant evidence for seasonality suggests early spring for the reindeer, but direct evidence for horse seasonality is lacking. This leads Bridault (1996: 150) to suggest repeated small-scale hunts for the horse; similar treatment of carcasses may indicate that the reindeer were exploited like the horse, in contrast to the cases at Pincevent and Verberie. In another article, Bridault (1997: 170) cited Poplin's (1994) seasonality determination of beginning of spring for horse mortality at Etiolles to suggest a winterspring horse hunting season in the Paris Basin, which would complement late summer/autumn reindeer hunting. DISCUSSION

If we consider the pattern of carcass exploitation at both of the Paris Basin sites, there is a clear contrast with the pattern we see in the reindeer in the Perigord . In the Perigord, there is evidence for longer term, perhaps year-round, occupation of sites. The large rock shelters, equally attractive for prehistoric hunter-gatherers as for archaeologists, may have been magnets for occupation over the centuries or millennia. We have little fine-grained, large-scale , modern excavations of the classic aggregation sites such as La Madeleine or LaugerieHaute, particularly when compared to the amount of dirt moved and art objects discovered during the 19th and early 20 th centuries. Do aggregation sites exist? I don't know. The question may be more one of numbers and proximity of sites with relatively small populations in each, rather than large numbers of people at any one place. The faunal analyses do not suggest numbers of prey animals of vastly greater orders of magnitude than are present in the very short-term occupations of the open-air sites of the Paris Basin. As Boyle ( 1993, 1994, 1997) has suggested, the contrast may be in the fashion in which the prey were exploited. One glaring difference appears to be in the seasonality of reindeer exploitation; in the Perigord, reindeer was exploited in the winter to spring, while it was exploited in the fall in the Paris Basin. As a function of their different seasonality, the prey offered different kind of resources and were exploited using different strategies. Boyle (1993, 1994, 1997) suggested that the fauna from La Madeline, Reignac, and Limeuil indicate gourmet selection of only a small part of each reindeer carcass. Whether this was due to an overabundance of available or killed prey, or to degradation of the fat content and nutritional utility of certain carcass portions compared to other portions, will require more detailed research to answer. Either possibility would be consistent with the structure and nature of availability at the end of the late winter dispersal and the beginning of the spring migration. Even though the bone 117

ENLOE

surface preservation is very good, filleting on the reindeer is relatively minor, suggesting that the food was immediately or shortly thereafter consumed. There is no indication of mass kills for longer -term storage. This does have implications for the availability or exploitation of other resources. Secondary species, particularly horse , have seasonality that is complementary to that of the reindeer and could provide nutrition when the reindeer were no longer present-if , for example, they had migrated out of the region . Each individual of these species appears to have been much more completely exploited , suggesting they were hunted as individual encounters.

relatively fewer people . This may be the basis for the much less elaborated production of art-the more thinly stretched population at a peripheral location not exactly at the heart of the Magdalenian symbolic interaction sphere. As such, I cannot agree with Mellars 's (1994) contention that the Magdalenians moved into the Pari s Basin to co ntinue the ame kind of exploitation of species that they had practiced and continued to practice in the Perigord. This is adaptation in its most evolutionary , non -teleological sense.

REFERENCES

In co ntra st, the Paris Basin sites of Pincevent and Verberie have very restricted seasonality; mortality occurred in the fall , obviously the best time and place for acquirin g large quantitie of meat to be stored for the winter . Carcasse are more complete and much closer to the bulk curve that Boyl e ( 1993, 1994, 1997) saw for the Perigordian horses. Other reso urces were very limited in their availability , as far as we know from the few sites ava ilable . The sites are differ ent because, as open-air ites, they did not have the geographic constant of the rock shelter to anchor the settlement pattern. The sites occupying them, therefore, imply lower population densities. With lower population , we may have decreased social complexity, smaller groups, and more mobility , perhaps as a result of a less abundant resource base or at least fewer options in non-optimal seasons. The Magdalenians were not moving north to the Paris Basin simply to continue preying on reindeer and horse. Whatever the reason for the pioneering thrusts into previously unoccu pied or underutilized regions, it resulted in adaptational changes in fundamental aspects of Magdalenian culture in the new regions. Such changes can be seen in the Magdalenian occupation of the Allier and Loire valleys in Auvergne, with seasonal and regional alternation of prey species (Fontana 1998). Every adaptation is local. A culture does not merely copy itself onto a new territory; it must meet the problems posed by the physical and social environment, taking advantages of the structure of resource availability as it exists in the new region, and , via the flexibility that culture offer s as an adaptational means , readapt it elf . I am inclined to think that the differences between the Perigord and the Paris Basin are closely related to differences in population density. Although I remain unconvinced that the Perigord was the Garden of Eden or otherwise extraordinarily abundant in food resources , I nonetheless suspect that the Paris Basin was relatively poorer or less abundant. The reindeer were exploited in a less diverse, less vegetatively pro ductive environment than in the south, and there is less evidence for alternative resources to provide substantial subsistence. We have very minor use of other species , including butchered microfauna and birds. I think that the Magdalenians of the Paris Basin, of necessity, exploited their reindeer prey in a more logistical manner than was necessary in the Perigord . While the prey remained largely the same, i.e., reindeer, the structure of the adaptation was changed, with concomitant grea ter mobility over a larger territory and supportin g

Audouze, F., 1987. The Paris Basin in Magdalenian time s. In The Pleistocene Old World: Regional Perspectives, edit ed by 0 . Soffer, pp . 183- 200. New York: Plenum Pre s. Audouze, F., and J.G. Enloe, 1997. High resolution archaeology at Verberie: Limits and interpretations. World Archaeolog y 29: 195-2 07. Binford , L.R. , 1978. Nunamiut Ethnoarchaeology. New York: Academic Press. Boyle , K.V., 1993. Upper Palaeolithic procurement and processing strategies in southwest France. In Hunting and Animal Exploitation in the Later Palaeolithic and Mesolithic of Eurasia , edited by G.L Peterkin, H.M . Bricker, and Paul Mellars, pp 151-162. Archaeological Papers of the American Anthropological Association 4. Washington, D.C.: American Anthropological Association . Boyle, K.V., 1994. La Madeleine (Tursac, Dordogne): Une etude paleoeconomique du Paleolithique superieur. Paleo 6:55- 77. Boyle , K.V., 1996. From Laugerie-Basse to Jolivet: The orga nization of Final Magdalenian settlement in the Vezere valley. World Archaeology 27:477-491. Boyle , K.V., 1997. Late Magdalenian carcase management strategies: The Perigord data. Anthropozaologica 2526:287-294. Bridault, A., 1996. Le probleme de I' exploitation du gibier au Tureau des Gardes (Marolles-sur-Seine, Seine-etMarne ). In Paleolithiquie superieur et Epipaleolithique dans le Nord-Est de la France, edited by Y. Pautrat , pp . 141-1 51. Actes de la Table Ronde de Dijon 1995. Cahiers Archeologiques de Bourgogne 6 . Dijon: Direction Regionale des Affaires Culturelles-Service Regional de I' Archeologie de Bourgogne . Bridault, A., 1997. Chasseurs , ressources animates et milieux dans le nord de la France du Paleolithique a la fin du Mesolithique: Problematique et etat de la recherche . In le Tardiglaciaire en Europe du Nord-Ouest , edited by J.-P. Fagnart and A. Tevenin, pp. 165-176. 118

READAPTATION

Paris: Editions du Comite et Scientifiques.

des Travaux Historiques

Burke, A ., 1993. Applied skeletochronology: The horse as human prey during the Pleniglacial in southwestern France. In Hunting and Animal Exploitation in the la,ter Palaeolithic and Mesolithic of Eurasia, edited by G.L. Peterkin, H.M. Bricker, and P. Me11ars, pp. 145-150. Archaeological Papers of the American Anthropological Association 4. Washington, D.C.: American Anthropological Association. Conkey, M.W., 1980 . The identification of prehistoric hunter-gatherer aggregation sites: The case of Altamira. Current Anthropology 21 :609-630. David, F., and J.G . Enloe, 1993. Chasse saisonniere des Magdaleniens du bassin parisien. Bulletin et Memoire de la Societe d'Anthropologie de Paris 4(3-4) :167-174.

1'Association pour la Promotions la Connaissances Archeologiques.

et la Diffusion

de

Enloe, J.G., and F. Audouze, 1997. Le role de l'environnement dans la vie des chasseurs magdaleniens du Bassin parisien. In Le Tardiglaciaire en Europe du Nord-Ouest, edited by J.-P . Fagnart and A. Thevenin , pp. 177-186. Paris: Editions du Comite des Travaux Historiques et Scientifiques. Enloe, J.G., and F. David, 1997. Rangifer herd behavior: Seasonality of hunting in the Magdalenian of the Paris Basin. In Caribou and Reindeer Hunters of the Northern Hemisphere, edited by L.J. Jackson and P. Thacker, pp. 47-63. Worldwide Archaeology Series 6. Brookfield: Avebury Press. Fontana, L., 1998. Mobilite et subsistance au Magdalenien superieur et final en Auverngne. In Economie

Prehistoric : Les Comportements de Subsistance au Paleolithique, edited by J.-P. Brugal, L. Meignen , and

Deplano, S., 1994. Etude de la Faune des Grands Mammiferes

M. Pathou-Mathis, pp . 373-386. xvmem e Rencontres Internationales d' Archeologie et d'Histoire d' Antibes . Sophia Antipolis: Editions de I' Association pour la Promotions et la Diffusion de la Connaissances Archeologiques.

de la Couche IX de l'Abri du Flageolet II (Dordogne): Approche Taphonomique et Palethnographique. Memoire de Matrise en Prehistoire. U niversite de Paris I. Enloe, J.G., 1991. Subsistence Organization in the Upper

Paleolithic: Carcass Refitting and Food Sharing at Pincevent. Ph.D. Dissertation. University of New

Gordon, B.C., 1988. Of Men and Reindeer Herds in French Magdalenian Prehistory. British Archaeological Reports International Series 390. Oxford: British Archaeological Reports.

Mexico. Ann Arbor: University Microfilms International No . 9208123.

a

Enloe, J.G., 1992. Le partage de la nourriture partir des temoins archeologiques: Une application Ethnoarcheologique. In Ethnoarcheologie: Justification, Problemes, Limites, edited by A. Gallay, F. Audouze, and V. Roux, pp. 307 - 323. Centre de Recherches Archeologiques du CNRS. Juan-les-Pins: Editions de l' Association pour la Promotions et la Diffusion de la Connaissances Archeologiques.

M.F., 1980. The Initial Magdalenian in France. British Archaeological Reports International

Hemmingway,

Series 90 . Oxford: British Archaeological Reports. 1877. Relequiae Aquitanicae, Being Contributions to the Archaeology and Palaeontology of Perigord and the Adjoining Provinces of Southern France. London : H. Bailliere.

Lartet, E., and H . Christy,

Enloe, J.G. , 1993. Subsistence organization in the early Upper Paleolithic: Reindeer hunters of the Abri du Flageolet, couche V. In Before la,scaux: The Complex Record of the Early Upper Paleolithic, edited by H. Knecht, A. Pike-Tay, and R. White, pp. 101-115. Boca Raton : CRC Press.

1980 . Rock Shelters of the Perigord: Geological Stratigraphy and Archaeological Succession. New York: Academic Press .

Laville, H., J.-Ph . Rigaud, and J.R. Sackett,

Le Licon, G., 1997. Magdaleniens du Bassin parisien: Comparaisons entre les groupes du centre et des marges sud-ouest. In Le Tardiglaciaire en Europe du NordOuest, edited by J.-P. Fagnart and A. Thevenin, pp. 187-198. Paris: Editions du Comite des Travaux Historiques et Scientifiques.

Enloe, J.G., 1997. Seasonality and age structure in remains of Rangifer tarandus: Magdalenian hunting strategy at Verberie. Anthropozoologica 25-26:95-102. Enloe, J.G. , 1998. Fonction des sites et chasse specialisee: Variation regionale pendant la periode magdalenienne. In Economie Prehistoric : Les Comportements de Subsistance au Paleolithique, edited by J.-P. Brugal, L. Meignen, and M. Pathou-Mathis, pp. 363-372 . xvmeme Rencontres Internationales d' Archeologie et d'Histoire d' Antibes. Sophia Antipolis: Editions de

H. (editor), 1984. Art et Civilisations des Chasseurs de la Prehistoire, 34000-8000 ans av. J.-C.

de Lumley,

Paris: Laboratoire de Prehistoire du musee de l'Homme et Musee des Antiquites de Saint-Germain-en-Laye. Mellars, P.A., 1973. The character of the Middle-Upper Paleolithic transition in southwest France. In The

119

ENLOE

Explanation of Culture Change, edited by C. Renfrew, pp. 255-276. London: Duckworth. Mellars, P.A., 1989. Major issues in the emergence of modern humans. Current Anthropology 30:349-385. Mellars, P.A., 1994 . The Upper Paleolithic revolution. In The Oxford Illustrated Prehistory of Europe, edited by B . Cunliffe, pp. 42-78. Oxford: Oxford University Press. Otte, M. , 1997 . Paleolithique final du nord-ouest, migrations et saisons. In Le Tardiglaciaire en Europe du Nord Ouest, edited by J.-P. Fagnart and A. Thevenin, pp. 353-366. Paris: Editions du Comite des Travaux Historiques et Scientifique . Poplin, F., 1994. La faune d'Etiolles : Millieu animal , milieu taphonomique , milieu humain. In Environments et Habitats Magdaleniens dans le Centre du Bassin Parisien, edited by Y. Taborin, pp . 94-104. Documents d' Archeologie Fran~aise 43. Paris: Editions de La Maison des Sciences de I'Homme.

Schmider, B., 1984 . Les Industries Lithiques du Paleolithique Superieur en Ile-de-France. VIeme Supplement a Gallia Prehistoire. Paris: Editions du Centre National de la Recherche Scientifique. Spiess, A.E., 1979 . Reindeer and Caribou Hunters: An Archeological Study. New York: Academic Press. White, R.K., 1980. The Upper Paleolithic Occupation of the Perigord: A Topographic Approach to Subsistence and Settlement. Ph.D . Dissertation. University of Toronto. White, R.K., 1982. Rethinking the Upper/Middle Paleolithic transition. Current Anthropology 23:169-176. White, R.K., I 985. Upper Paleolithic Land Use in the Perigord: A Topographic Approach to Subsistence and Settlement. British Archaeological Report International Series 253 . Oxford: British Archaeological Reports . Yellen, J.E., 1977. Archaeological Approaches to the Present: Models for Reconstructing the Past. New York: Academic Press.

120

10 Specialized Final Magdalenian Hunting Technology in Southwest France Gail Larsen Peterkin (Tulane University)

ABSTRACT The statistical analysis of French Upper Palaeolithic weapon armatures indicates that the Magdalenian was characterized by the preferential use or organic weapon armatures (sagaie and harpoons), although, in some cases, backed microliths were inserted into grooved sagaies. During the Final Magdalenian, however, there was a brief return to traditional lithic projectile points (e.g., Azilian points , 12,000

481

-

stone -circle

1986

152 m2

6, 120

67

69

-

1990- 199 1

159 m2

3,414

130

17

hearth ?

Alsdorf

1974

34 m2

15,000

200

-

stone pavement?

Beeck

surface

-

1, 195

51

35

-

Kamph ausen

surface

339

15

33

-

Sweikhuizen-GP * Mesch Eyserheide

-

hearth hearth ?

*Th e numb er of arti facts from Orp , Kanne, and Sweikhuizen-GP includes finds reco vered from sieving the sediments.

structure (Table 11.1 ). In all instances, the lithic scatters have relatively small dimensions; evidence for large settlements, with a high investment in the construction of large habitation and hearth structures, is absent (Arts and Deeben 1987). Tools found in the assemblages include burins, endscrapers, borers and bees, backed blades, and retouched blades (Figure 11.3, nos. 1-13); they are mostly made of good quality flint that occurs in chalk layers, alluvial deposits, and river terraces near the sites (Rensink 1993) . In a number of assemblages, including Eyserheide, Beeck, and Alsdorf , burins and large borers or bees are the main tool category , pointing to a specialization of on-site activities and an emphasis on antler and/or bone -working (Joris et al. 1993). Assemblage diversi ty is also demonstrated by the site of Mesch, where the num ber of retouched tools is remarkably low, and large , roughly shaped precores are numerically well represented among the cores (Rens ink 1993 ). With respect to the range of mobility of Magdalenian groups in the area, one important observation is that some of the used flint materials occur as exotic materials in the assemblages of Gonnersdorf and Andernach in the cen tral German Rhineland, about 120 km to the southeast (Floss 1994; Rensink 1993) . These raw material links suggest that Magdalenian sites in both areas belonged to the same settlement system, although the exact nature of this presumed relationship is still unclear. This is partly due to the position of the concentrations in decalcified loess sediments, which explains why organic material has not been preserved and why radio carbon dates or data on the season(s ) of occupation are absent. As opposed to other regions in northwestern Europe, no artistic manifestations have been found in these northern sites . On the basis of the available evidence , the small lithic scatter s can best be explained as representing short-term camps used by small groups exploiting biotic and abiotic sources on the fringe of the northwest European loess area (Art s and Deeben 1987 ; Rensink 1993 ). Apparently , proximity to perm anent

water cour ses or exploited raw material sources was not the main motive for settling in most of these locations . This premise is best illustrated by the Magdalenian site-record of the Dutch loess area , where minimum distances to streams or exploited flint sources range between 0 .5 and 1 km. This characteristic, together with their high strategic location along plateau-edges above stream valleys, has led to the idea that at least some of the settlement locations were primarily chosen for game monitoring (cf . Mesch [Rensink 1993]) .

Geographical Representativeness The vast majority of Magdalenian settlements in the area (including Orp , Mesch , Sweikhuizen -GP, Eyserheide, Alsdorf , and Beeck ) have been found close to the edge s of loes s-mantled plateaus above stream or dry valleys . In this geomorphic zone, loess layers have been subject to Holocene erosion , as can be inferred from the presence of truncated Holocene soil s observed durin g the excavation of the sites . The process of erosion, together with modern agricultural activity disturbing archaeological layers, led to the exposure of the Magdalenian lithic scatters at the present surface. This particular circumstance , however , is not encountered everywhere along the plateau-edges in the area. For instance, on relatively flat terrain where the top of the loess layer has remained relatively intact, Magdalenian material may still be present , as yet outside the reach of the plough and without archaeological surface visibility . The same applies to plateau edges where erosion did occur , but where concentrations were originally covered by a relatively thick layer of loess , and neither erosion nor ploughing affected the archaeological layers . On the contrary, preservation of sites is not expected to be evident on those plateau-edges that are devoid of loess or where Pleistocene river terraces or older geological formations appear at the surfac e. The se area s most likely included spots that were favored by Magdale nian hunter-gat herer for setting 136

THE MAGDALENIAN RECORD OF NORTHWESTERN EU ROPE

3 0

4



1



2

8

9

11

10 0

I

'

5

·•

i

0

6

7

12

13

15

16

0

Figure 11.3- Retouc hed flint tools (nos. 1-13) and flint core (no. 14) from excavated Magdale nian sites in the Meuse-Rhi ne loess area. Nos. 15 and 16 are stray finds collected from the surface of the Dutch loess area, possibly reflecting Magdalenian off-site activities . All artifacts 55% of actua l size with the exception of no. 14 (centimeters are indicated) . l =borer, 2-4=endscrapers, 5-7=burins, 8-11 =backed blades, l 2- l 3=becs, 14=core, l 5=burin, 16=endscraper.

137

R.ENSINK

up camps, although erosion most probably washed away the loess sediment along with Magdalenian cultural debris. As opposed to the plateau-edges, there is virtually no evidence of Magdalenian occupation of the relativ.ely flat and more central parts of loess-mantled plateaus. Until now, only a single site has been documented from this section; it is located in the loess district of the southern Netherlands (SweikhuizenKool weg). To explain the scarcity of known sites, a crucial question is whether or not loess was deposited in the area after the period of Magdalenian occupation. In the case of loess sedimentation, settlement traces may be obscured by loess sediments. Hence, the virtual absence of Magdalenian surface sites in this zone may not be significant in behavioral terms. If we accept the second option-i.e. no sedimentation of loess that postdates the time of the Magdalenian occupation-the absence of evidence strongly indicates that Magdalenian groups did not use this ection of the landscape or used it only sporadically for the execution of specific activities, leaving few archaeological traces. In view of the stratigraphic position of most of the Magdalenian sites recovered from the area-i.e., about 30-60 cm beneath the present surface and overlain by loess sediments (be it in primary or secondary position)-the option that settlement traces are obscured by loess seems to be the most plausible, although, of course, it is impossible to quantify the number of unknown sites. In view of the observed uniformity in the geomorphological setting of most sites, one could suppose that the plateau-edges represented the main zone of activity of Magdalenian hunter-gatherers exploiting the Meuse-Rhine loess region. Such an interpretation, however, is not deemed very accurate here because it does not take into account the effects of both erosional and depositional processes that operated in the area after the time of the Magdalenian occupation. Both factors are largely responsible for the fact that portions of the Magdalenian settlement record were destroyed or have not yet been identified. Destruction of sites and other traces seems possible, especially for the intermediate zones between plateau areas and valley floors-i.e., valley slopes and other sloping areas. After deforestation and cultivation of these areas, geological processes such as colluviation played a major role, probably resulting in the partial or complete destruction of prehistoric hunter-gatherer sites. In the MeuseRhine loess area, only a single Magdalenian site (SweikhuizenOude Stort) has been reported for this section; it is not clear if the find occurred in primary or secondary position. With relation to the lower parts of the landscape, it is very difficult, if not impossible, to be precise on where and at what depth Magdalenian material might be present. This holds especially true for the two most important river valleys in the area-i.e., those of the rivers Meuse and Rhine. The lack of finds in this region seems to be explained, on the one hand, by the dynamic character of both rivers which, in the Late Glacial and in the Holocene, probably eroded significant parts of their own sediments that might have contained Magdalenian material. On the other hand, non-eroded landforms of early Late Glacial age may be covered by

Holocene alluvium, making them inaccessible or only very incidentally accessible to archaeological investigation. A similar situation holds for the numerous smaller stream and dry valleys in the area, where multiple episodes of erosion and sedimentation changed early Late Glacial landforms. At the bottom of these valleys and at the foot of the valley slopes, dense, concentrated finds and off-site patterns may be found well-preserved in fine-grained sediments; this pattern is undisputedly demonstrated by the Magdalenian site of Kanne in Belgium. As men-tioned before, this site was found by accident, located very close to the alluvial plain of the river Geer and buried primarily by a layer of colluvial sediments (Vermeersch et al. 1985). Magdalenian Off-Site Archaeology The study of archaeological off-site patterns also requires attention within this regional perspective. Long-term excavation projects carried out in the area at, for example, the Middle Paleolithic locality at Maastricht - Belvedere (Roebroeks et al. 1992) have clearly shown the high potential of this field of research; they are one of the key elements of landscape approaches in archaeological research (Rossignol and Wandsnider 1992). As holds true for geologically dynamic areas in general, the interpretation of isolated finds collected from the surface of the Meuse-Rhine loess landscape, outside the context of sites, is beset with serious problems. In the first place, while patinated tools like endscrapers, burins, and backed blades have, on repeated occasions, been found as stray finds (see Figure 11.3, nos. 15-16), it is difficult to decide whether these tools are actuallyMagdalenian artifacts. Tools like the ones mentioned here, although Magdalenian in character, might also represent later Paleolithic material, such as retouched implements left behind by Federmesser or even Ahrensburgian groups. Unfortunately, typologically distinctive projectile points made from stone do not form part of the Late Magdalenian assemblages under discussion here, and, as a result, this tool category cannot be used as a source of information on Magdalenian hunting practices in the area. Secondly, for most finds of isolated tools and even for some typical Magdalenian blade cores, it can not be determined with certainty whether the e reflect off-site activities or, alternatively, are displaced elements originally part of denser scatters. In short, the recovery and documentation of Magdalenian off-site patterns seems to be possible only under specific circumstances-Le., through the exposure of large areas consisting of fine-grained fluvial or aeolian sediments that contain Magdalenian cultural debris in situ. In the Meuse-Rhine loess area, however, such efforts have not yet been carried out in a systematic manner; the excavations undertaken thus far have been directed to the documentation of lithic scatters without giving full consideration to the archaeological record of the surrounding sediments. Conclusion In the Meuse-Rhine loess area, loess sediments relevant to the preservation of Magdalenian material in situ and lying at or very near the present surface are tied to very specific sections 138

THE MAGDALENIAN RECORD OF NORTHWESTERN EUROPE

of the loess landscape: plateaus and plateau edges. Outside these areas, Late Glacial sediments have been partly or completely eroded or are found at a depth of several meters beneath the present surface. The consequence is that, even if we would be able to document all the traces preserved in, for example, the uppermost meter of the soil, we would have only a fraction of the full settlement record in this particular loess landscape. In addition, little information is available about Magdalenian off-site activities in the area. REGIONAL LANDSCAPES AND MAGDALENIAN DATA-SETS

If considered within a larger geographical framework, it appears that the Meuse-Rhine loess area does not represent an exceptional case. In fact, similar observations can be made for other regions in northwestern Europe. In northern France, the Paris Basin consists of an extensive loess landscape characterized by wide river valleys and elevations like cuesta ridges and limestone plateaus, ranging from 40 meters to more than 200 meters in height. Magdalenian open-air occupations have principally been found incorporated into alluvial and aeolian sediments which are found in and close to major river valleys, including that of the Seine, the Yonne, and the Oise, respectively. Until the 1960s, field surveys conducted in the area revealed a number of Magdalenian open-air sites located on loess-mantled plateaus, including Les Bois des Beauregards, dominating the valley of the Loing, a tributary of the Seine (Schmider 1971). As applies to the Meuse-Rhine loess area, the deposition of aeolian sediments during and/or following the period of Magdalenian occupation seems to have been of great importance for the preservation of settlements on and near the edges of plateaus. Most of the sites were exposed at the surface as the result of ploughing, disturbing archaeological layers and providing an opportunity for the collection of lithic artifacts at the surface. These partly excavated plateaus sites were, for several decades, the only source of information about the nature of Magdalenian settlement in the Paris Basin. A rather dramatic change in this situation took place in 1964, when a Magdalenian concentration referred to in French as "habitation no. I" was discovered accidentally in the course of dredging at the bottom of the valley of the Seine at Pincevent (Leroi-Gourhan and Brezillon 1966). This discovery proved, beyond the shadow of a doubt, the great potential of valley floors for the excellent preservation of traces of human activity in fine-grained sediments dating to 13,000-11,500 B.P. At Pincevent, periodic flooding of the Seine, involving the gentle deposition of fine sands and silts between successive phases of occupation, has been responsible for the burial, and thereby preservation, of short-term, seasonal encampments belonging to the Late Magdalenian (Leroi-Gourhan and Brezillon 1966, 1972; David and Orliac 1994). Since virtually no fluvial sedimentation has taken place in the later part of the Late Glacial or in the Holocene, finds occur just below the present surface, making them easily accessible for large-scale excavation. This circumstance is also known from localities such as Verberie, Etiolles, and Marsangy, all Late Magdalenian campsites located at the

bottom of valley floors; all have revealed numerous wellpreserved hearths, concentrations of lithic tools and debitage, and faunal debris (Audouze 1994; Audouze et al. 1981; Coudret et al. 1994; Schmider 1992). All of these sites provide a very detailed picture of activities carried out by Magdalenian groups in specific parts of river valleys (see, for example, Audouze and Enloe 1991), information not available for the Meuse-Rhine loess area. The Pincevent site, in particular, provides us with sound evidence that Magdalenian hunters returned on a regular basis to specific parts of river valleys for intercepting migrating herds of reindeer. At this location, reindeer-dominated habitation units have been found associated with various stratigraphic levels. Moreover, each of these levels contains small quantities of the same type of exotic stone material (silex brun), which was imported over a distance of about 40 km from the northeast (Bodu 1991). Thus, in the Paris Basin, we can see a shift in research priorities beginning in the middle of the 1960s, from mainly surface sites located along plateau-edges to well-preserved sites located at the bottom of valley floors. In the latter, long-term excavation projects at localities such as Pincevent, Etiolles, and Verberie have been, and still are, directed towards the documentation and analysis of small portions of river banks rather than individual locations. This strategy has proved to be important to the discovery of low density distributions, isolated hearths, and small knapping areas, which may be classified as off-site archaeological material. A similar level of detail has not been achieved for the plateau areas, where organic preservation is rare and disturbance of settlement traces much more intensive. Unfortunately, in regional syntheses dealing with the Magdalenian occupation of the Paris Basin, information from these plateau sites is touched upon only briefly. A regional perspective on the Magdalenian settlement record of the uplands of the Ardennes Massif, 2 embracing parts of southern and eastern Belgium and adjacent parts of northeastern France, reveals a very different picture. This area of pronounced relief is characterized by high plateaus, steep valley slopes, and relatively small, narrow valley floors. The first systematic research in the area dates back to 1865-1870, a time period when it was commonly believed that Paleolithic man both used and lived in caves much of the time. As a result, research at that time was principally aimed at the geological and archaeological exploration of caves and rockshelters located in the valley of the river Meuse and some of its tributary valleys (Dupont 1867, 1872). The Late Magdalenian is well-represented in this region, particularly in the valley of the Lesse, where caves like Trou de Chaleux, Trou des Nutons, and Trou du Frontal have yielded rich deposits of Magdalenian lithic and organic tools, hunted fauna, art objects, and ornamental fossil molluscs. In contrast to most of the riverside settlements in the Paris Basin, these assemblages can be viewed as accumulations of cultural debris deposited on relatively stable surfaces, with at least some of the caves used on repeated occasions. Although density of cultural remains differs from one cave to another, one typical characteristic is the use of exotic flint and diversity in animal species (including horse, bovines, reindeer, saiga, chamois, arctic fox, and roe and 139

RENSINK

red deer) . Moreover , engravings of realistic animal figures on slate are an important aspect of the inventories of caves such as Chaleux. A welcome synthesis of these early observations and finds has been written by Dewez (1987) , while part of these old collections is further subject to a systematic reappraisal by the University of Liege (Charles 1994). In recent times, we can see a continuation of the research tradition initiated in the 19th century by Dupont and others- the exploitation and excavation of Magdalenian and other archae ological deposit s in cave and rock shelter s, thi s time address ing new research questions and using modern excavation technique s and detailed documentation (see , for instance , Straus et al. 1994). As in many other regions of Europe , caves and rockshelter s in the Ardenne s Ma ssif are welcome repo itorie s of Paleolithic material , essentially "pre servation halls" of archaeolo gical layer s that , in an open-air situation, would probably have been lost because of pronoun ced relie f and slope erosion . For this particular area, site-o riented research can be put on a par with cave-oriented resea rch, with one notable exception. The open-air site of Roe -la-Tour in north ern France , excavated in 1980 and covering an area of 1 IO m2, is located in a high position in the French part of the Ardennes uplands (Figure 11.4); it has revealed a rich Magdalenian stone industry, together with numerous engravings on slate (Rozoy 1988). It is unique for its preservation ; the material s were found under and between large boulders that fell upon the archaeological layer, hence protecting parts of the settlement area from erosion . No other open-air sites are presently known from the Ardennes Massif . This observation seems to be partly because of the lack of loess or other fine-grained sediments with a high potential for the preservation of openair settlement traces. In addition, much of the surface of the Ardennes is covered by extensive forests, which has been a major factor in the lack of recognition of Magdalenian surface sites and other diagnostic traces . To illustrate further the specific propertie s of regional Magdalenian records, the final region under discussion here is the Central Rhineland, lying to the south of the town of Cologne in the central part of western Germany. Althou gh upland areas like the Eifel and the Westerwald are the domi nant landscape features , the area also includes the Neuwied Basin , one of the most significant regions for Paleolithic research in western Europe (Basinski et al. 1995). In this basin , measuring about 20 x 30 km and drained by the river Rhine, Late Glacial human occupation is evident in the form of both Magdalenian and Federmesser settlements . Unique by northwestern European standards is their stratigraphic position beneath volcanic sediment s from the Laacher -See volcano in the East Eifel , to the west of the Neuwied Basin. The eruption of this volcano towards the end of the Allerod Interstadial (about 11,000 B.P.) covered the land surface of the Neuwied Basin with a layer of pumice and ash, which locally can reach a thickness of more than 10 meters . This layer has served as an excellent shield, protecting much of the early Late Glacial environmental and archaeological record from weathering and erosion .

Figure 11.4- Yiew of the Magdalenian open-air site of Roe-la-Tour on the southern edge of a plateau in the Frenc h Ardennes area. Preserva tion of part of the site is the result of large boulders which have fallen upon the Magdalenian archaeologica l layer . Ot her parts of the site have bee n completely eroded .

In this geolo gically favourable depo sitional environment, the famous and classic Magdalenian sites of Gonnersdorf and Andernach are located in a slope position, on the right and left side of the river Rhine, respectively , approximately two kilometers away from each other (Bosinski 1979; Bosinski and Hahn 1972; Veil 1984). At Gonnersdorf , the find layer occurs under a 2 m thick pumice layer and within loess sediments, indicating that, in the Neuwied Basin , the deposition of loess continued after the time of the occupation of the site. This cir cumstance, in combination with the very quick accumulation of the pumice layer about two thousand years later, has accounted for the fact that much of the Magdalenian archaeological layer has remained intact. It is very likely that the Gonnersdorf and Andernach sites, and perhaps other, yet undiscovered Magdalenian sites, would have been largely or completely eroded without the protection of the overlying pumice layer. Thus , in the Neuwied Basin, we are dealin g with a "Pompei like" situation in term s of excellent condition s for the preser vation of Magdalenian and later Paleolithi c (Federmesser) material, albeit without surface visibility of archaeological traces. It is not surprising , therefore , that both Andernach in 1883 and Gonnersdorf in 1968 were discovered by accident , in the course of pumice quarrying and digging for house con struction, respectively (Bosinski 1979; Basinski and Hahn 1972). Obviously , at both locations , we are dealing with rather substantial part s ( 120 m 2 and 680 m2 ) of larger settle ment area s, the exact size of which , however, is unknown becau se of their location within the modern town area. The assemblages are characterized by huge dwelling structures ; high quantities of stone tools made from several varieties of local and exotic stone material ; a high diversity of preserved faunal remains ; and various "artistic manifestations ," including numerous engravings of animals (horse, mammoth) and female humans on slate plaquette s (Bosinski 1991). Diversity in faunal species, as described for the Belgian cave sites, is in 140

THE MAGDALENIAN RECORD OF NORTHWESTERN EUROPE

agreement with the data obtained from Gonnersdorf and Andernach, pointing to the exploitation of various game resources. Late glacial animals represented in the faunal assemblages include horse, arctic fox, varying hare, birds, and fish. Several authors have stressed the great similarities in the spatial lay-out, raw material utilization, and material culture between Gonnersdorf and Andernach. This comparison has led to the idea that the occupation of both locations may even have been contemporaneous (Basinski 1991). Because of quantitative and qualitative characteristics of the find material, it has also been proposed that the locations had some central function in Magdalenian settlement system; perhaps they were large base camps or even "aggregation camps" (for a discussion on this topic, see Rensink 1993). Elsewhere in the Neuwied Basin, Magdalenian occupational traces have most probably remained unnoticed during the quarrying of pumice , including low density scatters and other traces with little archaeologica l visibility. Such traces of short-term activities, which are very characteristic for the Federmesser record in the area, have not yet been documented for the Magdalenian. This absence might indicate that strategies of land-use of Magdalenian hunter-gatherer were very different from those of Federmesser groups, who inhabited the Central Rhineland during the Allerod Interstadial (Baales and Street 1996). It is clear, however, that further specification of this presumed difference in overall land-use strategies requires insight into the nature of the Magdalenian settlement record outside the site contexts of Gonnersdorf and Andernach. For instance, additional information may still be present in the lower riverine parts of the basin itself. There, islands of pumice may cover fine-grained fluvial sediments with high potential for the preservation of Magdalenian material, including traces of short-term activities very different from the dense, highly visible assemblages of Gonnersdorf and Andernach.

in northwestern Europe have been found in depositional environments where fine-grained fluvial, loess, or other sediments were laid down immediately or shortly after their abandonment. In other words, they represent elements of a buried archaeological record, most of which have come to light accidentally as a result of their disturbance by activities like dredging (Pincevent), ploughing (Bois-des-Beauregards), or commercial digging (Andernach, Gonnersdorf). These factors also explain why our view of Magdalenian settlement and subsistence activities in each region is almost exclusively based on one specific category of sites (see Table 11.2): •

Open-air riverside settlements in the Paris Basin, saved from erosion because of their geomorphic positions at the bottom of valley floors, in the early Late Glacial within zones susceptible to flooding and favourable to the preservation of Magdalenian settlement traces. Information on Magdalenian settlement outside river valleys is less detailed, but available as far as loess-covered plateaus are concerned. Documentation of off-site patterns is limited to some well-excavated portions of valley floors . Cave and rock-shelter sites in the Ardennes Massif, located in steep valley slopes and saved from erosion because of the protective role of the caves themselves. Information from other parts of the landscape is virtually absent; the open-air site of Roe-la-Tour is one notable exception. Traces of off-site activities carried out outside the caves have not yet been identified.



EVALUATION: REGIONAL-SCALE VARIATION

Open-air sites along plateau-edges in the Meuse-Rhine loess area, saved from erosion owing to their positions in loess sediments that can still be found on major parts of plateaus. Information from other parts of the landscape is virtually absent. Although some finds of isolated tools and cores may be a reflection of Magdalenian off-site activities, the exact meaning of these finds remains unclear.

If taken from a site-perspective, the Magdalenian record of each region exhibits some specific characteristics with relation to the topographic location of sites and faunal, lithic, and "artistic" site contents. Together they account for the various types of sites that have so far been recovered from these regions. Taken from a regional perspective, the data suggest that we perhaps must consider different regional settlementsubsistence systems. This would be in agreement with the contrasting natural environments Magdalenian huntergatherer were living in, with the notion that, in the early Late Glacial, regional variations in vegetation and in fauna) communities did exist.

Two large open-air sites in the Central Rhine region, located on the valley slope of the Rhine and saved from erosion as a result of their burial by a thick layer of Table 11.2 Magdalenian Open-Air and Cave Site (Including Geomorphological Setting) Region

Type of Site Cave or Abri

Valley Floor

Valley Slope

Plateau

13

I

7

I

6

I

12

-

12

l

-

l

-

9

-

4

-

Open-Air Paris Basin

At the same time, however, the regions are characterized by different types of landscapes, which, since the period that Magdalenian hunter-gatherers left their "visiting cards," have been subject to various geological processes impacting original site distributions, as well as the surface visibility and/or the chance of discovery of Magdalenian cultural remains. Except for the Ardennes Massif, Magdalenian settlements

Ardennes Massif

141

Geomorphological Position

Meuse -Rhine loess area

IO

Central Rhineland

2

2

RENSINK

pumice from the Laacher-See volcano. Again, information from other parts of the landscape is virtually ab ent. Given the spatially continuous character of the activities of hunter-gatherers, it can be assumed that the identified campsites were associated with other camps erected in both similar and dissimilar geomorphic zones, representing similar or complementary components of former settlement-subsistence systems. Within this context, procurement sites of the flint materials found in abundance in the Magdalenian sites in the area may serve as an example for the Meuse-Rhine loess area. For the Ardennes Massif and the Central Rhineland , we can mention short-term kill sites for animals whose remains have been found in both cave (Chaleux) and open-air localities (Gonnersdorf and Andernach). Thus far, archaeologists have failed to identify these traces either because: 1) conditions for their preservation were unfavourable; 2) their burial by Holocene sediments; 3) small size or low density of finds ; and/or 4) research prioritie s. Together with problems created by site-formation processes (Rensink 1995), the e factors hinder an assessment of particular regional characteristics, not only in material culture, but al o in settlement locations and types and in the spatial organization and subsistence activitie s of Magdalenian hunter-gatherers operating in these areas on a seasonal or yearly round basis . One example is the Paris Basin, where well preserved settlement areas like Pincevent and Verberie have contributed to the common view held by archaeologists that the occupants of these locations, at least during some periods of the year, were specialized hunters of reindeer. However, the notion that specialized reindeer hunting may have been a seasonal practice very specific for the Late Magdalenian of the Paris Basin must be regarded as rather premature. Comparison among regions is complicated by the fact that Magdalenian sites con taining organic material have not been identified from valley floors in the other three regions under discussion. If we take a look at the steep, narrow valleys in the Belgian Ardennes area (Lesse) or in the Central Rhine region (Lahn), for instance, we must consider very different geological processes, leading to the destruction rather than to the preservation of Late Glacial sediments . In large parts of these valleys , therefore, Magdalenian camps probably have been lost due to fluvial activitity. As pointed out earlier for the Meuse-Rhine loess area, lack of surface visibility seems to be a major factor explaining the virtual absence of Magdalenian sites and other traces documented from valley floors. For the time being, it can not be excluded that such traces, if preserved, could include flint or bone scatters very comparable to those that have been found at different spots in valley floors in the Paris Basin (Pincevent) that are indicative of specialized hunting of reindeer.

DISCUSSION The example of the Meuse-Rhine loess area, although not yet worked out in detail, illustrates very well that, if we concentrate on archaeological landscapes rather than on specific sites and if we consider research possibilities concerning off-site

archaeological patterns , more can be said about the potential of regional archaeological records for studying aspects of Magdalenian hunter-gatherer behavior . In this region and the other areas discussed in this paper, early Late Glacial landforms have been subject to postdepositional erosional and sedimentation processes destroying, effecting, and obscuring original Magdalenian site- and non-site distributions. Moreover, as pointed out at the outset of this paper, archaeol ogists have organized their fieldwork in a rather selective way, focusing their attention on localities which are obvious (caves in the Ardennes Massif) , well-preserved (riverside settlements in the Paris Basin) , or dense and highly visible (Central Rhineland). It is evident that both geological and historical factors have limited overall reconstructions of the regional use of the landscape by Magdalenian hunter -g atherers (cf., those formulated for ethnographically observed hunter -ga therers) . It is argued here , therefore , that these factors should be a main focus of any regional research in Paleolithic archaeology . Without giving full consideration to these aspects, interpretation of the available data may lead to co nclu ions about Upper Paleolithic land use that are deceptive or even false . Besides integrating these aspects of source evaluation in regional syntheses, it i necessary to look for alternative approaches that make interregional comparisons more meaningful. For instance, in order to overcome at least some of the problems connected with the interpretation of incomplete and biased regional data-sets , one of the approaches that can be taken is to concentrate on the archaeological record from different regions, from more or less similar geomorphic zones or portions of landscapes. In the case of the Meuse-Rhine loess area, this could mean a comparison with the Magdalenian record of the plateau zones of the Paris Basin. One pertinent area is named Les Bois des Beauregards; it is situated on the right bank of the Loing river about 80 km to the southeast of Paris. On this small plateau, measuring about 1 x 1 km in extent, many years of intensive survey have brought to light a large number of localities associated with the Magdalenian; some of these were investigated by excavation during the 1950s (Schmider 1971 ). Magdalenian tools recovered from this plateau are typologically similar to those from the Pincevent site, providing the site with a provisional date in the Dry as II period. Although stratigraphic or other details are not available, the data point to various artifact concentrations and finds distributed over a relatively large surface, notably along the western , riverside edge of the plateau. Estimates of the number of retouched tools collected from these locations are very high-i.e., more than 15,000. This large number of tools has not been reported for other plateau areas in the Paris Basin. The data from Les Bois-des-Beauregard are consistent with the expectation that, during the Dryas II period, Magdalenian hunter-gatherers favoured specific plateaus above river valleys, returning to these particular locations on a regular basis. This idea is further strengthened by the site of Ville-SaintJ acques, situated on a loess-covered plateau about 60 m above the valley of the river Seine and overlooking the Pincevent area (Fig ures 11.5 and 11.6). From the surface of this

142

THE MA GDALENIAN RECORD OF NORTHWESTERN EUROPE

, ,, ,,,"' ,-'

_,

--...-

,

,

,, ,

_..., ......

......

,--""

,,

.

...... , ''I ,, .:______________ ,' I

,1""

,, ,I

I

I

• loess cover

ancient alluvial sediments

alluvial sediments bottom of the valley

plateau

Magdalenian site

m

Figure 11.5-Location of the Magdalenian sites of Ville-Saint -Jacques and Pincevent to the southea st of Paris in the Paris Basin. The sites are only about one kilometer apart , but situated in two very different geomorphological settings. The Magdal enian concentration s at Pincevent were (and still are) found in fluvial sediments on the bottom of the valley of the river Seine . The site of Ville-Saint-Jacque s is located on the northern edge of a loes -covered plateau overlooking the Pincevent area . Modified after Robin-Jouve (1994 :fig. 2).

plateau, several thousand Magdalenian tools have been recovered over the course of a century , including two small- scale excavations undertaken in 1970 and 1988 (Degros et al. 1994) . Large numbers of burned stone fragments probably relate to the presence of numerous hearth structures during the occupation. Although ploughing has severely affected intra-site patterns and little can be said about the exact timespan of occupation , it seems likely that the finds reflect several or many Magdalenian occupations , which were only one part of a large, intensively used settlement area. One explana tion for this concentration might be that the plateau area of Ville-Saint-Jacques was very attractive for monitoring migra tory herds of reindeer; it was also not far from locations at the bottom of the valley floor of the Seine , where this animal species actually could be captured (cf., Pincevent) . It is also important to note that two AMS dates on bone material, 12,300 ± 160 (OxA-730) and 12,240 ± 160 (OxA - 730), date Ville -Saint-Jacques to the Dryas II period. These dates are comparable to those obtained at Pincevent.

identified seem to reflect small , isolated occupation s rather than components of more or le s continuous distributions . In this context, it is important to note that the surveys of ploughed fields in the immediate surroundings of the

In Binford ' s (1980) terms , both plateau areas discussed here show a great redundancy , which also holds true for extensive sites such as Pincevent and Etiolles at the bottom of valley floors. A rather different archaeological signature has been documented for the loe ss-covered plateau s of the Meu se-Rhine loess area, where the artifact concentrations

Figure 11.6-Section of the valley of the river Seine where the reindeer dominat ed concentration s of Pincevent were found. The photograph is taken from the plateau area of Ville-Saint -Jacques, bordering the valley of the river Seine to the south and overlooking the Pincevent area (cf,. Figur e 11.5).

143

R.ENSINK

documented lithic scatters have remained devoid of Magdalenian artifacts. In other words, find situations are very different from those encountered at Les Bois des Beauregards and Ville-Saint-Jacques, although, of course, it cannot be ruled out that, even in the very direct vicinity of the documented sites, lithic scatters have remained unnoticed because of lack of surface visibility. Observations such as these may be significant to our understanding of how Magdalenian hunter-gatherers periodically settled themselves on the regional landscape. They also support the notion that, in the Meuse-Rhine loess area, occupation of plateau wa less intensive, more sporadic, and/or more constrained in time as compared to specific plateau areas in the Paris Basin. The notion that densities of stone artifacts on plateaus vary within and between regions-the Paris Basin and the MeuseRhine loess area, in the example discussed above- hould be further tested by systematic field research. Given the lack of archaeological surveys of ploughed fields in, for example, the Belgian loess area, it i expected that traditional surface surveys will already yield additional evidence about the nature and intensity of Magdalenian occupation on and/or close to plateau edges. For other plateau areas, it is likely that mo t of the archaeological traces that remain undiscovered are obscured by loess sediments, and, consequently, their discovery will be far from easy. Their recovery will only be possible by using time-consuming methods like drilling, the excavation of small test-pits, or the close inspection of excavation sites for house construction or other building projects. In this context, another research possibility worth considering is looking at the results of large-scale excavations conducted in, for example, the loess district of southern Limburg and neighbouring areas, including the Aldenhovener Platte and the Merzbach Tai (Farruggia et al. 1973). Although these excavations have focused on Neolithic and even later settlement areas, observations regarding the presence and, perhap even more importantly, the absence of Magdalenian campsites and other traces within these large excavated areas may give valuable insight into the use of the regional loess landscape by Magdalenian hunter-gatherers. Such efforts are necessary and are even prerequisite in order to assess, as accurately as possible, the distribution and nature of Magdalenian cultural remains across this strip of loess. Moreover , these and other efforts may allow us to escape from a simple and meaningless translation of actual site patterns into earlier settlement systems.

ACKNOWLEDGMENTS I would like to thank Gail Peterkin and Heather Price for giving me the opportunity to participate in the symposium "Regional-Scale Variation in Late Pleistocene Western Europe," held in the context of the 61 st Annual Meeting of the Society for American Archaeology, 10-14 April 1996, in New Orleans. Financial support was provided by a grant from the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO). Thanks are also due to Henk de Lorm (Leiden) and Peter Vleugels (Roermond) for preparing Figure 11.1 and Figures 11.2 and 11.5, respectively.

NOTES 1. The settlements are linked with the Magdalenian colonization of northern Europe-i.e., the most recent and the best documented example of human expansion following an extremely cold phase-in this case, the Weichselian Glacial Maximum around 20,000-18,000 B.P. Radiocarbon date of Magdalenian sites in the area, for the most part, range between 13,000 B.P. and 11,800 B.P., from the transitional Dryas I-Bolling into the early Allerod. 2. The name Ardennes Massif a used in this paper includes the following three geographical areas (from west to east): the Condroz, the Famenne, and the Ardennes per se. Magdalenian occupations have been found in caves and rockshelters, particularly in the first area; this area consists primarily of Middle and Upper Devonian limestone depo its.

REFERENCES CITED Arts, N., and J. Deeben, 1987. On the northwestern border of late Magdalenian territory: Ecology and archaeology of early Late Glacial band societies in northwestern Europe. In Late Glacial in Central Europe: Culture and Environment, edited by J.-M. Burdukiewicz and M. Kobusiewitz, pp. 25-66. Pace Konisji Archeologicznej 5. Wroclaw: Pace Konisji Archeologicznej. Audouze, F., 1992. L' occupation magdalenienne du Bassin parisien. In Le Peuplement Magdnlenien: Paleogeographie Physique et Humaine, edited by J.-Ph. Rigaud, H. Laville, and B. Vandermeersch, pp. 345-356. Actes du Collogue International de Chancelade, Perigueux, I988. Paris: Editions du C.T.H.S. Audouze, F., 1994. Verberie. In Environnements et Habitats Magdaleniens dans le Centre du Bassin Parisien, edited by Y. Taborin, pp. 167-172. Documents d' Archeologie Fran~aise 43. Paris: Editions de la Maison des Sciences de l'Homme . Audouze, F., D. Cahen, L. Keeley, and B. Schmider, 1981. Le site magdalenien du Buisson Campin aVerberie (Oise). Gallia Prehistoire 24:99-143. Audouze, F., and J. Enloe, 1991. Subsistence strategies and economy in the Magdalenian of the Paris Basin, France. In The Late Glacial in North-West Europe: Human Adaptation and Environmental Change at the End of the Pleistocene, edited by N. Barton, A.J. Roberts, and D.A. Roe, pp. 63-71. Council for British Archaeology Research Report 77 . London: Council for British Archaeology. Baales, M., and M. Street, 1996. Hunter-gatherer behavior in a changing Late Glacial landscape: Allerod archaeology in the Central Rhineland, Germany. Journal of Anthropological Research 52:281-316 . 144

THE MAGDALENIAN RECORD OF NORTHWESTERN EUROPE

Binford, L.R., 1980. Willow smoke and dog 's tails: Hunter-gatherer settlement systems and archaeological site formation. American Antiquity 45:4-20. Binford, L.R., 1982. The archaeology of place . Journal of Anthropological Archaeology I :5-31 . Bodu, P., 1991. Pincevent: Site magdalenien . Les Dossiers d'Archeologie 164:60-67. G., 1979 . Die Ausgrabungen in Gonnersdorf 1968-1976 und die Siedlungsbefunde der Grabung 1968. Der Magdalenien-Fundplatz Gonnersdorf 3.

Basinski,

Wiesbaden: Steiner. Basinski, G., 1988. Upper and Final Paleolithic settlement patterns in the Rhineland , West Germany. In Upper Pleistocene Prehistory of Western Eurasia, edited by H.L. Dibble and A. Montet-White, pp. 375-386. University Museum Monograph 54. Philadelphia: The University Museum , University of Pennsylvania. Basinski, G., 1991. The representation of female figures in the Rhineland Magdalenian . Proceedings of the Prehistoric Society 57 :51-64. Basinski, G., and J. Hahn, 1972. Der Magdalenien-Fundplatz Andernach (Martinsberg). Rheinische Ausgrabungen 11 :81-257. Bosinski, G., M. Street, and M. Baales (editors), 1995. The Palaeolithic and Mesolithic of the Rhineland. In Quaternary Field Trips in Central Europe 15, Vol. 2, pp. 829-999 . 141h INQUA Congress, Berlin. Munich: Verlag F. pfeil.

Dewez, M. , 1987. Le Paleolithique Superieur Recent dans Les Grottes de Belgique. Publications d'Histoire de 1' Art et d' Archeologie de Universite Catholique de Louvain 17. Louvain: Universite Catholique de Louvain. Dunnell, R.C., and W.S. Dancey, 1983. The siteless survey: A regional scale surface collection strategy. In Advances in Archaeological Method and Theory: Volume 6, edited by M. Schiffer, pp. 267-287. New York: Academic Press. Dupont , E., 1867. Etude sur l 'ethnographie de l'Homme de l 'Age du Renne (dans Les Cavernes de la Vallee de la Lesse). Memoires Couronnes 19. Brussels: Memoires Couronnes. Dupont, E., 1872. L'Homme pendant Les Ages de la Pierre dans Les Environs de Dinant-sur-Meuse. Brussels: C. Muquardt and Cie. Enloe , J.G., and F. David , 1992 . Food sharing in the Paleolithic: Carcass refitting at Pincevent. In Piecing Together the Past: Applications of Refitting Studies in Archaeology, edited by J.L. Hofman and J.G. Enloe, pp. 296-315. British Archaeological Reports. International Series 578. Oxford: British Archaeological Reports. Farruggia , J.P., R. Kuper, J. Llining, and P. Stehli (editors), 1973. Der Bandkeramische Siedlungsplat z Langweiler 2: Gemeinde Aldenhoven, Kreis Duren. Rheinische Ausgrabungen 13. Bonn: Rheinland-Verlag. Floss, H., 1994. Rohmaterialversorgung im Paliiolithikum des Mittelrheingebietes . Monographien RomischGermanisches Zentralmuseum 21. Bonn: R. Habelt.

Charles, R., 1994. Towards a new chronology for the Late Glacial archaeology of Belgium. Part II: Recent radiocarbon dates from the Oxford AMS system. Notae Praehistoricae 13:31-39.

Foley, R., 1981. Off-site archaeology: An alternative approach for the short-sited. In Patterns of the Past: Studies in Honour of David Clarke, edited by I. Hodder, G.Ll. Isaac, and N. Hammond, pp. 157-183. Cambridge: Cambridge University Press .

Coudret, P., M. Larriere-Cabiran, M. Olive, N. Pigeot, and Y. Taborin, 1994. Etiolles. In Environnements et Habitats Magdaleniens dans le Centre du Bassin Parisien, edited by Y. Taborin, pp. 132-146. Documents d' Archeologie Fran9aise 43. Paris: Editions de la Maison des Sciences de l'Homme.

Gordon, B .C., 1988. Of Men and Reindeer Herds in French Magdalenian Prehistory. British Archaeological Reports International Series 390. Oxford: British Archaeological Reports.

David, F., and M. Orliac, 1994. Pincevent. In Environnements et Habitats Magdaleniens dans le Centre du Bassin Parisien, edited byY. Taborin, pp. 154-166. Documents d' Archeologie Fran9aise 43. Paris: Editions de la Maison des Sciences de l'Homme.

Joris, 0., R-W. Schmitz, and J. Thissen, 1993. Beeck: Ein "special-task-camp" des Magdalenien. Neue Aspekte zum spaten Jungpalaolithikum im Rheinland, Archaologisches Korrespondenzblatt 23:259-273 . Kelly, R.L., 1983. Hunter-gatherer mobility strategies. Journal of Anthropological Research 39:277-306.

Degros, J., B. Schmider, and B. Valentin, 1994. Ville-SaintJacques: Le Tilloy. In Environnements et Habitats Magda.leniens dans le Centre du Bassin Parisien, edited by Y. Taborin, pp. 176-178. Documents d' Archeologie Fran9aise 43 . Paris: Editions de la Maison des Sciences de l'Homme.

Leroi-Gourhan, A., and M. Brezillon, 1966. L' Habitation magdalenienne no. 1 de Pincevent pres Montereau (Seine-et-Mame). Gallia Prehistoire 9:263-385 . 145

R.ENSINK

Leroi-Gourhan, A., and M. Brezillon, 1972. Fouilles de Pincevent: Essai d'Analyse Ethnographique d'un Habitat Magdalenien. Vlleme Supplement a Gallia Prehistoire. Paris: Editions du Centre National de la Recherche Scientifique .

Lohr, H., 1979. Der Magdalenien-Fundplat z Alsdorf, Kreis Aachen-Land: Ein Beitrag zur Kenntnis der funktionalen Variabilitat jungpaliiolithischer Stationen . Unpublished thesis. Ttibingen . Patou, M. , 1992. Apports de l'archeozoologie a la comprehension du peuplement magdalenien dans le Bassin mosan (Belgique). In Le Peuplement Magdalenien: Paleogeographie Physique et Humaine, edited by J.-Ph. Rigaud, H. Laville , and B. Vandermeersch , pp. 309-316. Actes du Colloque International de Chancelade , Perigueux 1988. Paris: Editions du C.T.H .S. Rensink, E ., 1993. Moving into the North : Magdalenian Occupation and Exploitation of the loess lands capes of Northwestern Europe. Leiden: University of Leiden . Rensink , E., 1995. On Magdalenian mobility and land use in north-west Europe: Some methodological considerations. Archaeological Dialogues 2:85-119. Rensink, E., 1997. Loess sediments and Magdalenian campsites in the southern Netherlands: A pilot study on preservation and site location. In Le Tardiglaciaire en Europe du Nord-Ouest, edited by A. Thevenin and J.-P. Fagnart, pp. 397-408. Paris: Editions du C.T.H.S. Roblin-Jouve, A. 1994. Le milieu physique . In Environnements et Habitats Magdaleniens dans le Centre du Bassin Parisien, edited by Y. Taborin, pp. 12-35 . Documents d' Archeologie Fran~aise 43. Paris: Editions de la Maison des Sciences de l'Homme.

Schmider, B., 1971. Les Industries Lithiques du Paleolithique Superieur en Ile-de-France. VlemeSupplement a Gallia Prehistoire. Paris: Editions du Centre National de la Recherche Scientifique . Schmider , B. (editor ), 1992. Marsang y: Un Campement des Derniers Chasseurs Magdaleniens sur Les Bords de l 'Yonne. Etude et Recherches Archeologiques de l'Universite de Liege 55. Liege: Universite de Liege . Straus, L., P. Lacroix , J.-M . Leotard, M. Otte , A. Martine z, and M. Newman , 1994. Late Pleistocene and Early Holocene research in Wallonia: Abri du Pape, Grotte du Bois Laiterie, and Huccorgne. Old World Archaeology Newsletter 18:11-2 1. Thi sen, J., 1989. Ein Fundplatz des Magdalenien am linken Niederrhein bei Kamphau sen. gem. Jtichen, Kreis Neuss. Archiiologisches Korresponden zblatt 19:315-32 3. Veil, S., 1984. Siedlungsbefunde vom MagdalenienFundplatz Andernach. Zwischenbericht tiber die Grabungen 1979-1983. In Upper Palaeolithic Settlement Patterns in Europe, edited by H. Berke, J. Hahn , and C.J. Kind, pp. 181-193 . Ttibingen: Archaeologica Venatoria. Vermeersch, P., R. Lauwers, and Ph . van Peer, 1985. Un site Magdalenien a Kanne (Limbourg). Archaeologia Belgica 1:17-54 . Vermeersch, P., N. Symens, P. Vynckier, G. Gijselings, and R. Lauwers, 1987. Orp: Site Magdalenien de plein air (comm. de Orp-Jauche). Archaeologia Belgica 3:7-56.

Roebroeks, W., D. De Loecker, P. Hennekens, and M. van Ieperen, 1992. "A veil of stones:" On the interpretation of an early Middle Palaeolithic low density scatter at Maastricht-Belvedere (The Netherlands). Analecta Praehistorica Leidensia 25:1-16.

Waters, M.R., and D.D. Kuehn, 1996. The geoarchaeology of place: The effect of geological processes on the preser vation and interpretation of the archaeological record . American Antiquity 61 :483-497.

Rossignol, J., and L. Wandsnider (editors), 1992. Space , Time, and Archaeological landscapes. New York: Plenum Press.

Zvelebil, M., S.W. Green, and M.G . Macklin, 1992. Archaeological landscapes, lithic scatters, and human behavior. In Space, Time, and Archaeological landscapes, edited by J. Rossignol and L. Wandsnider, pp. 193-226. New York: Plenum Press.

Rozoy, J.-G., 1988. Le Magdalenien superieur de Roe-la-Tour I. Helinium 28: 157-190.

146

12 Patterns of Ethnogeographic Variability in Late Pleistocene Northwestern Europe Berit Valentin Eriksen (University of Aarhus)

ABSTRACT Regional studies have a long tradition in the archaeology of Late Pleistocene northwestern Europe. However , despite an increas ing amount of archaeological and paleoenvironmental data, relatively few comparative studies have attempted a more general interdisciplinary synthesis of cultural and environmental changes during the period in question . Based on an exhaustive review of the archaeological and paleoenvironmental record, the present chapter attempts to make up for this bias. The purpose of the study is to achieve a sound analytical framework for a comparative interdisciplinary discussion of how Late Pleistocene hunter - gatherer groups of northwestern Europe interacted with their environment at a regional scale. In this respect, our ability to determine the role or position of a single site and individual area in the local and regional universes of past hunter-gath erers is emphasized. The spatio-temporal framework of the case study is naturally defined on a macro-regional (i.e., northwestern Europe sensu lato) and long term (13,000 - 10,000 RCY B.P.) scale, and the approach is largely "culture-historical." The issue of ethnogeographic variability is thereby reflected in a source critical discussion of the spatio-temporal and sociocultural relationships of the Late Pleistocene archaeological cultures, groups, or technocomplexes of northwestern EuropeHamburgian, Federmesser, Bromme , Ahrensburgian, Fosna-Hensbacka. INTRODUCTION Regional studies have a long tradition in the archaeology of the northwest European Old Stone Age. The approach taken has often been at a rather local scale, or a landscape approach with a geographical focus on areas varying from a few hundred up to a few thousand square kilometers. The object of these studies was often to establish the socio-economic and natural setting of a specific site or group of sites, such as the site catchment area, or to delineate the settlement history of a particular area and period (see, for example, Andersen 1988; Bokelmann 1979; Cull berg 1996; Fischer 1991; Larsson 1991, 1996; Tromnau 1975, 1976). Many prominent scholars have also attempted a more generalized, macro-regional approach, with a geographical area of interest covering major parts, or even all, of the northwest European lowland. These endeavors have often been devoted to largely descriptive presentations of archaeological cultures or technocomplexes , with a main focus on questions pertaining to chronology and subsistence economy (Burduckiewicz 1986; Schwabedissen 1954; Taute 1968). Relatively few comparative studies have attempted a more generalized interdisciplinary synthesis of cultural and environmental changes and varying land-use patterns during the period in question (Clark 1975; Schild 1976, 1984).

for a comparative interdisciplinary discussion of how Late Pleistocene hunter-gatherer groups of northwestern Europe interacted with their environment at a regional scale. In this case study, the spatio-temporal framework will be naturally defined on a macro-regional (approximately 300,000 square kilometers) and long-term (3,000 RCY) scale, and the approach will be largely culture-historical. However, the purpose of the following discussion is not to present yet another cultural map or another fine-meshed chronological subdivision of Late Pleistocene northwestern Europe. Rather, the purpose is methodological: to discuss our ability to determine with some degree of certainty the role or position of any single site and/or any individual area in the local and regional universes of prehistoric hunter- gatherers. The issue of ethnogeographic variability will accordingly be reflected in a critical discussion of the spatio-temporal and socio-cultural relationships of the various Late Pleistoce~e archaeological cultures, groups, or technocomplexes of northwestern Europe. In this respect , the problem of how to distin guish between geographical and chronological aspects will receive major attention. For methodological reasons, all 14 absolute dates will refer to conventional C radiocarbon years (RCY) before the present. THE METHODOLOGICAL APPROACH

The archaeological record of Late Pleistocene northwestern Europe is well documented , but, despite an increasing amount of archaeological and paleo-environmental data, major gaps remain in our understanding of the complex spatio-temporal and socio-cultural relationships of these prehistoric huntergatherer societies. The purpose of the present study is to explore the possibilities for bridging these gaps and thereby attaining a sound methodological and analytical framework

The origin of the word region is Latin (regio = direction, line, boundary), and, in virtually all Western European languages, it is used synonymously with other geographically descriptive terms such as landscape or district. In common usage, it often refers to an area with a rather indefinite extent-the delimitation of which is frequently perceived rather than defined. In archaeological terminology, a "region" is thus not

ERIKSEN

surprisingly a quite elastic concept that may be used as a more or less distinct geographic frame of reference. A region may cover an area of a few hundred or several thousand square kilometers, and it may be circumscribed by more or less well-defined topographical, political, demographic , economic , or other natural or socio-cultural boundaries . The scale and the geographic delimitation of a region is highly dependent on the nature of its perceived or defined boundaries. Depending on the definition, regions may overlap, include, or eliminate each other and may even build up a hierarchy of behavioral universes. Essentially, the region is an analytical construct associated with sub-division of space (for a more elaborate discussion of these issues see Price and Peterkin in this volume and Wobst 1990). In a "regional -scale approach ," the extent and the delimitation of the study area depends first and foremost on the methodological approach, the questions addressed, and the chronological framework of the study. Together these issues determine the scale of analysis. Obviously, a naturally defined region , such as a topographic region, is a much more stable analytical construct than a socio-culturally defined region. On the other hand, one of the dangers of a well-defined, natural area of investigation is the temptation of seeing it as a prehistoric, behavioral universe in itself or even as a static phenomenon in the consciousness of prehistoric man . In the present case study this is particularly problematic. It is widely acknowledged that Late Pleistocene huntergatherer groups of northwestern Europe were very mobile. The available subarctic resources did not allow any permanent exploitation of the newly deglaciated areas. For a long period, most of the region was quite likely exploited on a more or less (ir)regular seasonal basis. It would thus be a serious error to assume that man-land relationships and land-use patterns remained constant through time. This problem increases with the scale of analysis. In my opinion, a regional analysis of Late Pleistocene land-use patterns may be meaningfully associated with three different hierarchical levels of behavioral universes , depending on the scale of analysis (for a more elaborate discussion see, for instance, Clark 1975: l 2f and Foley 1981): 1. Habitation level. In common usage, this is a "local-scale" approach seeking to establish the local universe-i.e., the socio-economic and natural setting, such as the site catchment area of a specific site or group of sites. 2. Annual territory level. Apart from the chronological studies, this is probably the most widespread "regional-scale" approach in current hunter-gatherer studies. The geographical range is the annual universe-i.e. , the probable territory covered by a group in the course of its seasonal movements on a yearly basis . Analysis typically seeks to integrate different sites in an annual system of subsistence, settlement , and mobility behavior . The extent of this "region" may vary

considerably. In the Magdalenian of southwestern and central Germany, there are indications that annual movements may have been at the scale of 200 or more kilometers (Eriksen in press; Floss 1994:325; Weniger 1991). 3. Social level. At this level the purpose is to explore intergroup contact and communication networks-i.e., the social universe. The approach may be regional or inter-regional depending on the nature of the data . Communication networks of Magdalenian hunter-gatherer groups for example, are known to have covered distances of 600-800 kilometers (Eriksen in press ; Floss l 994 :336f) . Yet we are still dealing with a universe that, at least in theory, may have been perceived by any individual member of the different groups. These behavioral universes or scales of interaction (cf. Price and Peterkin this volume) are obviously interrelated . At the habitation level, it is pos ible to reconstruct the extent and physical properties of the local universe with a high degree of certainty . At the increasing spatio -temporal scales associated with the annual and social levels, the amount of interrelated data and, accordingly , the number of possible sources of error increases markedly . Moreover, due to the prevailing subarctic conditions throughout the period in question, mobility and settlement patterns, as well as contact networks, are likely to have changed significantly over the course of time. Thus, our chances of identifying local and regional behavioral universes among these hunter-gatherers probably decrease exponentially with the increasing scale of analysis. In the present case study, it will be assumed that man-land relationships and land-use patterns, as expressed in the abovementioned behavioral universes, varied across time and space. The purpose is to examine this variation on a macro-regional scale and within a broad chronological framework. This socalled "culture-historical" or "ethnogeographic" level would not have been perceived by prehistoric man as a "behavioral universe." It is an analytical construct allowing us to examine a large number of archaeological sites of varied quality within a consistent spatio-temporal framework, which perhaps allows us to attempt a more generalized synthesis of cultural and environmental changes during the period in question .

THE NATURALSETTING Large parts of northwestern Europe were devoid of human settlement during the last Pleniglacial. Most of the area was covered by extensive inland glaciers, and the remaining part was severely affected by accompanying periglacial phenomena (Figure 12.1). The ice-free areas were thus truly inhospitable, characterized by a markedly arctic-continental climate and covered with treeless tundra. There is no archaeological evidence of human exploitation of the periglacial zone during this period . This observation may, to some extent, result from inadequate conditions for preservation . Nevertheless, the extreme scarcity of contemporary Upper Paleolithic (i.e., Gravettian) finds in the adjacent regions of western Europe makes it seem most likely that population density was close to an absolute minimum (Soffer and Gamble 1990). 148

PATTERNS OF ETHNOGEOGRAPHIC VARIABILITY

The large Fennoscandian ice-sheet retreated rapidly during the early Late Pleistocene and, by approximately 13,000 RCY B.P., the picture had changed significantly. The northwest European lowland, from the British Isles in the west to the Russian plain in the east, had by then become the destination of many hunter-gatherer groups, immigrating from their probable Late Pleistocene refugia in the southwest (Jochim 1987; Straus 1991) or the southeast (Soffer 1987). These immigrations, or re-immigrations, increased markedly over the course of the B0lling chronozone, when the climatic amelioration had finally established itself. Eventually, by the onset of the Holocene at approximately 10,000 RCY B.P., hunter-gatherer groups had expanded throughout virtually all of the ice-free area of northwestern Europe.

lowland that were uninhabitable during the last Pleniglacial because of ice cover or severe climatic (i.e., periglacial) conditions (Figure 12.1). When the ice retreated during the early Late Pleistocene, it left a virgin area open to human exploitation. Thus , we have a unique chance to study settlement history from its very beginning and to examine, on a macroregional basis, the timing and nature of changes in human adaptations to environmental changes during the period of 13,000-10,000 B .P. The region of the present case study is, first and foremost, a naturally defined area; whether it is also a socio-culturally defined area is an entirely different matter and, as previously mentioned, a question of the scale of analysis. For a start, it seems reasonable to assume that the boundaries are completely arbitrary with respect to Late Pleistocene huntergatherer behavior.

In the current paper, discussion will be confined to the parts of southern Scandinavia and the adjacent northwest European

□- 3 ... + 4

+

+

+

+

... ~

.. .. ..

+

+

• ,.

+

+ + +

+

+

..

+

+

+

+

+ ♦

...

+-

.... + 4-

.. ,.

• -+-

.. .. +

,,--

I

..,.,_, ..

+ +-

...

+

+

+ +

--~ ',:~~!

,,.--

~

t-

+

..

~

++

+



+

+

+

+

.,.

+-

+

+ +t-

I"

+

+

"

..... +-

t-

+

t'

....

+

+

+

.. ..

4-

+-

+

t

+-

+

+~

~

t-

♦ ♦

+ t-

.. .... ... .... +

..

+

+



+

+

+-

+

+

... ....

,r~

t-



+

~

+

+

,.

,a..

T

+

...

+

...

+

11

II

.. ... ... ... +-

-

+ ♦

.. ... +-

+-



+ ~

++-

~

...

-



-i-

+

++

+

+

+-

+-

+-

+

-

+ +-

+-

+-

+-

+

+

+

t-

+

+



..

+-

++

+-

++-

.. ... .. .... .. +

+

+

0

100

200

km

Figure 12.1-Map ummarizing the macro-topography (ice-front and vegetation) of northwestern Europe at approximately 18,000 RCY B.P. (I : icefront ; 2: periglacial zone; 3: arctic tundra) .

149

ERIKSEN Table 12.1 Schematic Outline of Geochronology, Climate, Vegetation, Temperature Variations (after Bokelmann 199I :75), and Baltic Ice Lake Phases in Late Pleistocene Northwestern Europe

Years BP ChronozonesClimate

Predominant vegetation

Sea Mean temperatures Baltic January

July

YoldiaSea

Preboreal 10.000 YoungerDryas

subarctic oceanic

Park-tundra(willow,dwarfbirch, juniper,wormwood,crowberry, grasses,sedges)

11.000

BalticIce Lake Aller0d

12.000............. . (OlderDryas) B0lling

cooltemperate increasing oceanic

sensu Lato

cooltemperate

OldestDryas

subarctic continental

Park-tundraand sparseforest (birch,rowan,juniper,aspen, willow,[pine],grasses,herbs) Park-tundra(birch, rowan,willow, wormwood,grasses,herbs)

13.000

14.000.......

Tundraandpark-tundra(dwarf birch,willow,wormwood,rock rose,sea buckthorn,grasses, herbs)

THE GEOCHRONOLOGICAL FRAMEWORK The chronostratigraphy of Late Pleistocene northwestern Europe has been an object of intensive research for more than 80 years and is thus very well documented. In fact, the classical sequence of the Oldest Dryas, B¢11ing, Older Dryas, Aller¢d, and Younger Dryas pollen- or chronozones was first established in southern Scandinavia (Iversen I 942, 1954; Jessen 1934). Despite increasing disagreement concerning the overall applicability and especially the synchronism of this sequence, it is still widely applied. The Older Dryas climatic retreat is now openly disputed even in southern Scandinavia (Berglund 1979: 11O; Kolstrup 1991 :4; Lowe and Gray 1980: 158). Conversely, the Younger Dryas event has been established as an almost global phenomenon between 11,000 and 10,000 RCY B .P. (Gates 1993:57f). Within the present study area, the sequence of environmental changes is known to have been largely synchronous (Lowe and Gray 1980). Moreover, because of the proximity of the Fennoscandian glacier, changes were often distinct and even abrupt, and inhabitants of this region probably had to make frequent adjustments to adapt to changing environmental conditions. Throughout the period in question, we may observe several more or less profound environmental changes with respect to temperature and precipitation (Berglund et al. 1984; Kolstrup 1991; van Geel and Kolstrup 1978); vegetation (Iversen 1954; Usinger 1975); wildlife (Aaris-S¢rensen 1992; Eriksen 1996a, 1996b with references); and changes in available land areas resulting from deglaciation and sea-level rise (Bjorck 1995, 1996; Petersen 1984, 1985) that very likely would have had a major influence on contemporary hunter-gatherer land-use patterns . Figure Table 12.1 synthesize s our present

Firststageof BalticIce Lake I

I I



knowledge with respect to chronostratigraphy, climate, vegetation, temperature variation, and Baltic Ice Lake phases. The figure does not take into consideration local variation between individual landscape situations. The size of the study area and the geomorphological variation within it, however, clearly indicate that paleoenvironmental conditions must have been quite variable on a spatio-temporal scale. Most of the study area was covered by ice during the last Pleniglacial. The rate of the glaciers retreat from the terminal moraines has previously been a matter of some discourse. However, recent research has confirmed that there was a sudden, noticeable warming about 14,000-15,000 B.P., and there is now general agreement that the ice retreated rapidly during the subsequent Oldest Dryas and B¢1ling chronozones (Berglund 1979; Berglund and Morner 1984; Bjorck 1995; Lagerlund and Houmark -Nielsen 1993). A several-hundr~dyear period of stagnation seems to have followed, accompanied by a visible decrease in temperatures and probably corresponding to the now controversial Older Dryas climatic deterioration (Berglund 1979:110). A considerable subsequent retreat of the glaciers in the Aller0d was then followed by a major readvance (in some areas up to 40 kilometers) to the Younger Dryas marginal moraines (Andersen et al. 1995; Bjorck and Digerfeldt 1984). Not surprisingly. the Younger Dryas is also characterized by a drastic temperature decline. In particular, the first part of this chronozone, from about 11,000-10,500 RCY B .P., seems to have been very cold and possibly fairly dry (Andersen et al. 1995; Kolstrup 1991). Obviously , there were significant changes in available landareas durin g the period in question, partly because the retreating glaciers left a large virgin area open for exploitation and 150

PATTERNS OF ETHNOGEOGRAPHIC VARIABILITY

partly because of a very turbulent coastline history. As the ice melted, sea levels rose eustatically, but, at the same time, land that had been previously weighed down by large masses of ice began to rise isostatically. Following the rapid , noticeable warming in the Oldest Dryas and B0lling, sea levels quickly rose. From this period , there are many marine deposits on present-day dry land in northwestern Denmark (Figure 12.2) (Petersen 1984, 1985). During the Aller0d and Younger Dryas, however , the isostatic upheaval of land took the lead in the southern part of the study area. Thus, the North Sea and most of the Baltic coastline of that time are submerged today (Figures 12.4, 12.8). Conversely, in the northern part of the study area (i.e., south central Sweden and southern Norway), these coastlines are often very distinctive (Andersen et al. 1995).

formation of the great Baltic Ice Lake. The initial stage of the lake, dating to the Oldest Dryas, is not well known, but, in the course of the B0lling chronozone, the lake was slowly blocked by surrounding land and ice masses (Bjorck 1995 :21). The Baltic Ice Lake had only a few and insufficient outlets-for example, through the narrow strait of bresund separating Denmark and Sweden. As one of the primary outlets of a vast, dammed lake, the Late Pleistocene bresund presumably was a deep torrent. It is often assumed that this narrow strait represented a major obstacle for animals and humans to enter southern Sweden (Bjorck 1995:23; Larsson 1996:lSlf; Liljegren and Ekstrom 1996). A presumed first drainage of the Baltic Ice Lake at 11,200 B .P.' removed this obstacle and created the first land-bridge between Denmark (i.e., the continent) and Sweden (Bjorck 1995:23, 1996:128). The subsequent readvance of the Younger Dryas glaciers, however, forced the lake to be dammed again, with the narrow bresund once again the primary outlet.

The enormous amount of meltwater from the retreating ice, combined with the isostatic upheaval of land, led to the

NORTH SEA

,

.....,·

-:_;•

'

l ,... .-. ,

,, -

-" -,, - " h

-

-

I\

II

-

"

-

II

-

I•

-

11

-

-

h

-

tr

ti

-

II

-

II

-

-

,,

- "

-

II

11

II

-

-

" -

"

,,

-

-

I•

11

-

II

-

-

1

-

-

-

It

II

11 I/

II -

1

11

,,

-

Ir

1-

-

11

-

1 II

-

-

It

-

h

::j,

II

-

If

-

1 : .)

-

fl

I -

n

-

11

fl

-

-

II

ll

I I -

-

-

II

-

I fl

-

II

-

-

t.~.

,, '".(',..

BALTIC

,, -

/I

-

,, - ,, -

-

•1

-

- ,, - ,, 1,

-

,, - ,, -

11

II

,, -

- f,P:1.5

-

11

{'•·.

...... .,

I

"

,,

)

h ,.,

-

ICE SEA

- ,, -

u

ti





- ,, -

- ,, - "



,,



- ,. - ,, ,, - ,, - ,,

,,

" e " • " •

e

I•



11



,1



,,, __ .. ,11

4-.,,.._, ., •

.. -•

.. •

.

"

• " • i'\~ ,.,-... " • .. • .. • " •

,, .

I/

,.,.

,. e

....

-

t,

,,

-

,, -

,,

-

fl

-

,.

-

,,

- ,, -

,,

"

., -

If

-

-

- ,,

,, - ,,

,,

-

I,

11



,,

e



,,





It



II



II





,.



"



fl



"



'• ,,



• ,,

1,



,,



•I





"



II



"



,,



..



,,



I•



II



•1



'I



I•



h

It



t,





II



..



It



f1



lo

f>

If

II



II



•r



II



It



"

,1

I•

. .. . . . . . . ... ... .. .. . .. . . "

,,

le



-

"

II

II

"

,,

•t



a\-:~-~-. u et'•,'; •



"

••

,,

"

,.

,.



,,



"

II



"



•1

,,

lo





,,

,,

~

,,

"

,,

,,



,.

,,

ti

Figure 12.2-Map showing the distribution of Hamburgian sites (hatched area) in relation to the macro-topography of the B0lling ( 1: ice margin at approximately 12,500 RCY B.P.; 2: tundra ; 3: park-tundra and sparse birch forest) .

151

ERIKSEN

The impending end of the Younger Dryas, and, accordingly, the end of the Pleistocene, was marked by a final dramatic 2 draining of the lake at approximately 10,300 B.P. The surface of the lake was then lowered 25 meters, perhaps in the course of a few years as the result of one more or les instantaneou event (Bjorck 1995:25). The Baltic Ice Lake pha es are central to the environmental history of the area. Even today the Baltic Sea, which is now the world's largest brackish-water body, is of vital importance to the environment, influencing the climate, vegetation, and, to some extent, the fauna of the surrounding land areas. Its influence would have been no less pronounced in the Late Pleistocene. The maps in Figures 12.2, 12.4, and 12.8 illustrate the observation that the recently deglaciated areas of southern Sweden close to the retreating ice generally were characterized by a notably warmer vegetational cover than the more distant old morainic soils of Western Denmark and Germany that border the North Sea (cf. Huntley and Birks 1983:620t) . This is prob ably largely an expression of edaphic difference - i.e., more favorable conditions for growth in the young morainic area in comparison to the washed out, old morainic area (Usinger 1975:154f). On the other hand, the Baltic Ice Lake may also have had a certain cooling effect on the environment. A close examination of pollen diagrams from southern Sweden indicates that the southeastern areas bordering the Ice Lake were characterized by a colder, more subarctic vegetation than the northwestern areas bordering the North Sea (Berglund 1979: 107f). The Late Pleistocene obviously witnessed a number of profound environmental changes. Some, like the final draining of the Baltic Ice Lake, were dramatic and almost catastrophic in nature. Others, such as fluctuations in vegetation and fauna, were more gradual and sometimes even rather localized in

importance . Presumably, these changes must have influenced contemporary hunter-gatherer land-use patterns to varying degrees. The following discussion will review the archaeological evidence from the period in question, with particular attention to the nature and timing of possible relationships between environmental and socio-cultural change on a patiotemporal cale.

THE ARCHAEOLOGICAL RECORD The first hunter-gatherer groups to reach the virgin areas of the Late Pleistocene northwest European lowland were the Hamburgians . Ever since the classic site of Meiendorf was discovered in the Ahrensburg tunnel valley near Hamburg in the early 1930s (Rust 1937) the Hamburgian culture or technocomplex ha been an object of intensive research . For a thorough pre entation of the archaeological remains , relative and absolute chronologies, settlement pattern, sub istence economy , etc., the reader may refer to the exten ive scientific literature on these i sues (e.g., Bokelmann 1979; Bratlund 1994; Burduckiewicz 1986; Gr~nnow 1985; Holm and Rieck 1987, 1992; Rust 1937, 1943; Tromnau 1975, 1976; and variou papers in, for example, Barton et al. 1991 or Larsson 1996). Hamburgian tool assemblages are today known from sites in the Netherlands, northwest Germany, southern Denmark (and Sweden?), and western Poland (Figure 12.2). Contemporary assemblages from the closely related Creswellian are known from southern England (Burduckiewicz 1986; Jacobi 1991). These assemblages will not be included in the present discussion. To summarize briefly, the Hamburgian lithic technocomplex may be characterized as a high quality blade industry. The slender, prismatic blades were usually struck from unifacial, bi-directional cores with oblique, opposed, and faceted

0

A

t

• 0

B Figure 12.3-Hamburgian

lithic tool types (A: shouldered point; B: Havelte point).

152



PATTERNS OF ETHNOGEOGRAPHlC VARlABIUTY



It

11

I

• II

I

to



3

~ ~4

□ .

: ,,...

5



,,

.

"

e

.... e

II

,1



II



11

•t

II



11



/1



h

II



r,

I

.

II •

• II

,,



14

e

• II

II

I

II

II •





II



"

,1

I

II

&

• 11

I•



,,

t

--1r•1"• -·•. -···•' ,,

II



I

,'

'

I/:

,,



/I

/'



1,

II

I•

It

t

ti

I I

• I

t

• I

_: f

I

,,

I

I

•1

:, ~ c:;r·")~ ,.. •

,,,

,,

I

I I

"

0

0

0

0

0

j

0

0

0

(t

0

0

0 0

0

j

0

0 0

0



0

a

0 0

0

0

j

0 0

0

0

0

0

'

0

0

0

Figure 12.4-Map showing the distribution of Federmesser sites (hatched area) in relation to the macro-topography of the Aller~d ( I : ice margin at approximately 11,500 RCY B.P.; 2: tundra; 3: park-tundra; 4: sparse birch forest; 5: birch forest with some pine).

platforms. The predominant reduction strategy seems to have involved a careful preparation and trimming of cores, followed by the application of soft-hammer direct percussion (Hartz 1987:27; Madsen 1992, 1996:65). The Hamburgian lithic tool inventory (Figure 12.3) is characterized by numerous end scrapers, often with lateral retouch; burins, mostly with burin blow on truncation; simple and curved perforators (or Zinken); and shouldered points (Kerbspitzen) or slender tanged points of Havelte type. Multiple tools and combination tools are common. Along with the presence of many exhausted cores (Madsen 1996:62) , this pattern might indicate an inclination to economize scarce, high-quality lithic raw materials.

barbs are prominent, as are simple points, awls, and some unique hafts or handles (Riemenschneider). Regular art objects are rare. Artistic expressions are generally confined to a few naturalistic or geometric engravings on pieces of bone, antler, amber, and stone. The lithic and organic industries of the Hamburgian may now be considered firmly rooted in the western and central European Magdalenian (Burduckiewicz 1986). Unfortunately, most of the Hamburgian sites consist only of lithic scatters, hence our information on subsistence activities and absolute dating is limited to a few classic sites in northwest Germany. Faunal remains from these sites indicate that Hamburgian subsistence economy was, at least to some degree, seasonally specialized on reindeer. The preferred hunting method appears to have been stalking with spears and spear-throwers (Bokelmann 1991; Bratlund 1994). It must be stressed, however , that Hamburgian subsistence was presumably much more varied and opportunistic than indicated by

The organic inventory, which is mostly made on reindeer antler, is known from the classic sites in the Ahrensburg tunnel valley. Uniserially barbed harpoons with slender, curved 153

ERIKS EN

0

0



f



Figure 12.5- Federmesser lithic tool type s.

this handful of sites . Thus, even from a socio-economic point of view, Hamburgian and Magdalenian hunters were closely related (Eriksen 1996a, 1996b). Recent dating efforts place the Hamburgian in the chronozones of B0lling and Older Dryas (Table 12.2) (Fischer and Tauber 1986; Usinger 1975: 123). For methodological reasons, a relative chronological subdivision of the Hamburgian is controversial. However, there is now almost general agreement that the so-called Havelte group-i.e. , sites characterized by the occurrence of tanged points of the Havelte type (Figure 12.3B)-are typologically late. These sites may very likely be dated to the Older Dryas (Holm and Rieck 1992:87; Stapert 1992: l 6f). The classic Hamburgian , characterized by shouldered points (Figure 12.3A), is often referred to the Poggenwisch and Meiendorf groups (Fischer 1991). These sites are convincingly dated to the B0lling (Table 12.2). The almost complete absence of Hamburgian finds in eastern Germany (i.e., the former German Democratic Republic) (Figure 12.2) is conspicuous, and it is tempting to interpret this coincidence of a recent political boundary and an archaeological distribution as a reflection of research strategies. However, it is also certain that the vegetational and faunal colonization of the newly deglaciated areas would have displayed a considerable time lag in response to the climatic amelioration (Aaris -S0rensen 1992: 145; Berglund et al. 1984). Thus, it is quite likely that the young morainic areas (including northeast Germany) were not truly exploitable before the late B0lling-early Aller0d period , as a result of paleoenvironmental conditions. In the same light, it is interesting to note that the rather few Hamburgian sites in southern Denmark all belong to the typologically late Havelte group . Moreover , these sites have all been found within the past 15 years . It is thus plausible that an intensified

search for Hamburgian sites in northeast Germany might produce a few Havelte sites in this area . The Federmesser or Arch-Backed Piece complex, which succeeds the Hamburgian geographically as well as chronologically, is also clearly related to the central European Magdalenian and Late Paleolithic . Thus, the Federmesser cultural groups were first described as an expansion of Magdalenian open-air settlement into the northwest European lowlands (Schwabedissen 1954). In southwest Germany , the genetic relationship between these two archaeological cultures is generally indisputable (Eriksen 1991:92, 112). For a detailed presentation of the archaeological remains, relative and absolute chronologies , etc., the reader may refer to the extensive literature on these issues (Barton and Roberts 1996; Bokelmann 1978; Bokelmann et al. 1983; Caspar and De Bie 1996; De Bie and Vermeersch 1998; Schild 1984; Schwabedissen 1954; Taute 1963; Tromnau 1976). Arch-Backed Piece assemblages are distributed throughout most of northwest Europe; they extend from England, across the Netherlands , Belgium, Germany, and southern Denmark to the Polish Vistula Valley in the east. In the south, they merge with final Magdalenian or Azilian groups (Schild 1996:131) (Figure 12.4). The following summary is especially concerned with the northernmost appearance of this widely distributed , hence rather heterogeneous , technocomplex . This is indicated by the use of the term Federmesser . The Federmesser lithic technocomplex may be described as a typical blade industry characterized by broad bi-directional, sub-cylindrical or conical cores with plain or roughly faceted platforms . The heavy blades and flakes were detached using both soft - and hard -hammer direct percussion (Hartz 1987:27; Madsen 1996 :71). The lithic tool inventory 154

PATTERNS OF ETHNOGEOGRAPHIC

(Figure 12.5) is characterized primarily by backed points (Federmesser, Ruckenspit zen, or pointes a dos) , mostly with a curved back, but more rarely with an angled back; short thumbnail end scrapers; and blunt burins on truncation . In the northernmost inventories, Wehlener scrapers with a retouched tang are especially numerous. As a result of generally poor conditions for preservation, the organic inventory is practically unknown.

(Table 12.2) and that only a few inventories may be attributed to the late Older Dryas (e.g., Klein Nordende CR) or early Younger Dryas, respectively. A suggested Preboreal age of some sites has been rightly dismissed (De Bie and Vermeersch 1998). For methodological reasons, a relative chronological subdivision of the Federmesser complex is still not possible. Obviously , such a widely distributed cultural complex as the Federmesser or Arch-Backed Piece complex must display a range of local and regional differences. However, such differences are reduced and may, to a certain degree , be related to variations in raw material accessibility or to use and procurement strategies (Barton and Roberts 1996; De Bie and Vermeersch 1998), as well as to functional differences among sites or even to idiosyncratic preferences with respect to the morphology of specific tool types. Needless to say,

Within the study area, Federmesser sites often consist of rather small lithic scatters, hence evidence for subsistence strategies is highly inadequate . In fact, only one site within the study area has provided a meager faunal inventory (Bokelmann et al. 1983; Eriksen 1996a, 1996b). The ab olute chronology of the Federmesser groups has been the focus of much debate over the years, but there now seems to be a general consensus that the majority of the sites are of Aller0d age

..



II



II



II

,,

□ □'

VARIABILITY

3

4

0

5

,''~

)

:,' I,',''

I I: ,, , I

, I

/: : I

..:

I

,,

.,, .,, .

.,, . . ,,



II



II



II



II



II



II



I•

•,,

1,



,,

0

II

0

0

0

" 0

0

0

0

0

Q

Q

~

0

0

C)

() Q

0

j 0

0 0

0

0

0

~

0

0

0

0

0 Q

C

0

0

j

0

0

C ()

()

0 0

0

0 0

0

0

'

0

0

0

0

C,

0 0

0 C,

0 0

0 C,

0

j

0

0

0

J 0

0

0 0

0

j

0

j

0

j

j

Q

0

j

0 0 0

0

j

0

0

Q Q

0 0

j

Q

C)

0

0

0

0

0 0

Q

0

0

0 0

0

0

0

0 0

~

0 0

0

Q

j

A

~

Q

Q

0

0 0

Q 0

Q

0

0

0

0

0

0

0

0

0

"

0



0

Q

C)

0

0 0

j

0

j

0

'

0

0 0

()

C,

0

0 0

0

0

0

0

0 0

0

()

0

0

0

0

0

0

0 0

0

~

0

0

0

0

'

Q

0

0

e

,,

0

0 0

"

I I

0

• • •

C,

, t

I

,,

II

0

I

0

• 0

Figure 12.6-Map howing the distribution of Brommian sites (hatched area) in relation to the macro- topography of the Aller~d (1: ice margin at approximately 11,500 RCY B.P.; 2: tundra ; 3: park -tundra; 4: sparse birch forest; 5: birch forest with some pine) .

155

ERIKSEN

the influence of various site formation processes on the preservation and retrieval of inventories in different parts of this large area must also be taken into consideration. For these reasons, the homogeneity of the Federmesser or Arch-Backed Piece complex has been seriously questioned (Madsen 1983:29). On the other hand, the assemblages have also been described as " ... the last technocomplex with great similarities all over western (and partly central) Europe" (De Bie and Vermeersch 1998:37). The northernmost limit of the Federmesser complex is not unambiguous . Genuine, undisturbed Federmesser sites are very rare in southern Scandinavia. The majority of Federmesser artifact types from this area come from potentially mixed assemblages that contain other Late Paleolithic , especially Hamburgian or Brommian, artifacts as well. Moreover, it is quite likely that some Federmesser types constitute an integral aspect of the early Brommian artifact inventory (Fischer 1988:21; Fischer 1991:102). The early phase of the Tanged Point complex- Le., the Bromme culture-is largely contemporary with the Federmesser groups but has a distinctive northerly distribution of sites. Brommian tool assemblages are abundant in Denmark, southern Sweden, and northern Germany (Schleswig-Holstein) (Figure 12.6), but there are no significant Bromme-related finds south of the river Elbe. For the first time in the Late Pleistocene settlement history of southern Scandinavia hunter-gatherer groups were able to exploit the newly deglaciated young morainic areas on a more permanent basis. Not surprisingly, their remains have been

subjected to intensive studies by southern Scandinavian archaeologists (Fischer 1990, 1991; Madsen 1983; Mathiassen 1946; Taute 1968). The Bronune settlement sites are primarily a sociated with the young morainic areas. This locational preference is clearly reflected in the lithic inventories . Brommian flintknappers unquestionably benefitted a great deal from the abundant outcrops and occurrences of high quality Cretaceous flint in these areas (Fischer 1990:41). They had no need to economize lithic resources, and the Brommian lithic industry may be described as an expedient and wasteful technology (Fischer 1991: 113). The broad, mostly uni-polar, semi-conical or sub-cylindrical cores were exploited by hard-hammer direct percussion only (Hartz 1987:27; Madsen 1992, 1996:65). Preparation was generally restricted to a light dorsal trimming. The platform was kept plain and non-facetted. It was a simple and expedi ent technique aimed at the production of heavy and straight blades. The lithic tool inventory comprises only a few retouched tool type . The inventorie s are largely dominated by end scrapers, dihedral burins, and large tanged points (Figure 12.7). Multiple tools and combination tools are infrequent. Unfortunately, the organic inventory is unknown because of generally poor condition for preservation. Again, information on subsistence activities and absolute dating is very scarce . A handful of sites have furnished rather meager faunal inventories, and there is general consensus that the limited species composition at these sites presents an interesting, but not very comprehensive, picture of subsistence strategies (Fischer 1991). With these reservations in mind, the finds nevertheless tend to give an



0



Figure 12.7-Brommian lithic tool types.

156

PATTERNS OF ETHNOG EOGRAPHIC VARIABILITY

Table 12.2 Schematic Outline of Relative and Absolute Chronologie s in Late Pleistocene Northwestern Europe (after Fischer 1991, 1996 ; Fischer and Tauber 1986; Lar sson 1996)

dated finds Years BP ChronozonesRadiocarbon 10.000.......

Archaeological cultures

Stellmoor, upper(10,020BP) YoungerDryas

11.000...... .

Fosna • Hensbacka Ahrensburgian

Arreskov(10,600BP) Fensmark(10,810BP) Mickelsmossen (10,980BP) Trollesgave(11,100BP)

Federm~er

Alleroo .... 12.000.... ......... . (OlderDry~) ....

B01ling sensu lato

13.000.......

KleinNordende CR (12,035BP) Stellnioor, lower (12,200BP) Meiendorf(:::12,300BP) Poggenwisch(:::12,500BP) Poggenwisch(:::12,900BP)

Bromme

Hamburgian

OldestDryas

(South)

14.000.... ...

overall , albeit tentative , impression of a band society charac terized by small social groups with high residential mobility and with subsistence activities focusing on the opportunistic stalking of moose and other large herbivores (Eriksen 1996a, 1996b). The few Bromme sites which may be scientifically dated appear to concentrate in the Aller0d (Figure Table 12.2). However, pollen and sediment analysis, as well as the radio carbon dating from Fensmark , indicate the continued existence of this cultural group into early Younger Drya s (Andersen 1988:525; Fischer 1991:111). For methodological reasons, a relative chronological subdivision of the Bromme is controversial. By comparison with the preceding Hamburgian and the partly contemporary Federmesser lithic technocomplexes, the Bromme complex is technologically and typologically distinct. However, a certain relationship is evident , especially between the two Aller0d complexes. As previously mentioned , some Federmesser types (e.g., Wehlener scrapers ) may constitute an integral aspect of the early Brommian artifact inventory. Reduction strategies also display evident similarities, and some sites and inventories may, with almost equal plausibility, be attributed to a local variant of the Bromme or the Federmesser complexes (e.g., Egtved: see Fischer 1988 and Madsen 1996:71 ). In a major thesis on the Tanged Point cultural groups of northwestern Europe , Wolfgang Taute incorporated a number of such hybrid inventories from northwest Germany into a distinct Tolk-Sprenge group (Taute 1968:213f). However, for source critical reasons, such a subdivision is highly questionable.

(North)

The later phase of the Tanged Point complex , the Ahrensburgian culture , is clearly rooted in the Bromme culture, but it has a distinctive southerly distribution of sites . Ahrensburgian tool assemblage s are widely distributed throughout the northwest European plain - i.e ., in the Netherlands, Belgium, northern Germany, southern Denmark, and Sweden (Figure 12.8). In western Poland, a contemporary and closely related Tanged Point technocomplex is known as the Masovian or Swiderian (Schild 1996:135). There are also a few Ahrensburgian sites in the western foothills of the Mittelgebirge (Baale s 1993; Cziesla 1992). A few of the latter assemblages have been found in caves or rock shelters. For a detailed presentation of the archaeologi cal remains, relative and absolute chronologies, etc ., the reader may refer to the extensive literature on these issues (e.g., Bokelmann 199 I; Bratlund 1991; Fischer 1991; Petersen and Johansen 1991; Rust 1943; Taute 1968; Tromnau 1975, 1976) . To summarize briefly, the Ahrensburgian lithic technocom plex may be characterized as a high quality blade industry . The cores were usually uni-polar , cylindrical or, more rarely , conical , with roughly faceted platforms. Core preparation was sometimes extensive and also included dorsal trimming of the front. Series of slender, prismatic blades were probably produced by soft-hammer direct percussion or a punch technique, while the very large blades (Riesenkling en) were detached using hard-hammer direct percussion (Hartz 1987:27; Schild 1996:133). The lithic tool inventory (Figure 12.9) is characterized by small tanged points, micro-truncations (Zonhoven-points) , end scrapers, and burins . Other retouched tool types , such as perforators , laterally retouched or backed pieces, etc., are rather infrequent. 157

ERIKSEN

,,,, .., I

., t,

. □ □ •

II



II

f

q

.

3

"

4

0

A

,,_... ",

5



I•



II

,,

If

.f

,, • ,ir··;-:,1 -""'- Y

• l ~.: .... ,;,.-~-·ir•';-• • II

,» ,~~~~~~-

II



11

e

1•



11



II

II



If



Yi • • II

. . .,,::-,

,,

It



\ 11

I;:::~ f.1, • ,:;;,~

e

; • I

I

fl





11

• ,, ,;••~ I

II

' ,, ,,,.,,

",'.

• ,, • ,, •II Ir ,,• II II

•,,

II

• • • •,, • ,, • • • •II II• •,,

II

0

II

II

II

0

0 0 0

0

0

0

0

0

b

"

0 0

0 0

" 0

0

--CJ

A

0

C>



0 ()

0_,

_,o,'

0

C

·-..........

0

~

0

0

0

0

'

0

0

0 0

0

0 0

0

0

0

j

0

"

0 0

0

0

C,

0

0

0

0 0

0

0

0

()

0 0

0

0 0

0

" 0

0

0 0

0

0 ,..-~

0 0

0

0 0

~ C,

0

~

0

0

()

"

j

0

0

0

0

0 0

0

0

0

0

0

0 0

0 0

0

0

q

(')

0

"

0

"

0

0

0

0

0

t>

0

0

0

0

0

0

0

0

0

0

"

0 C,

0

0

0 0

0

0 t,

0

0

0

,, •

j

0

0

{>

()

Figure 12.8-Map showing the distribution of Ahrensburgian sites (hatched area) in relation to the macro-topography of the Younger Dryas ( 1: ice margin at approximately 10,300 RCY B.P.; 2 : tundra; 3: park-tundra; 4 : sparse birch forest; 5: birch forest with some pine) .

The organic inventory is known from the classic Stellmoor site in the Ahrensburg tunnel valley. Lyngby antler clubs are prominent, as are uniserially, as well as biserially, barbed harpoons with rhomboid or triangular barbs and various simple bone points. Moreover, Stellmoor has produced the oldest remains of bows and arrows. Arrow shafts, particularly those of pine, were extremely well-preserved (Rust 1943: 188f, pl.91-96), but, unfortunately, they were lost during World War II (Bokelmann 1991:78f). The lithic, and probably also the organic, industries of the Ahrensburgian are firmly rooted in the Bromme culture (Fischer 1991:110) . From a typological and lithic technological point of view, however, there is a truly remarkable similarity between the Ahrensburgia r ,. id the almost 2,000year-older Hamburgian traditions. U··· , ·t nately, most of the Ahrensburgian sites consist only of lithic scatters , but there is

evidence of at least one aggregation camp (Stellmoor) withpossible periodical mass killing of reindeer (Bokelmann 1991; Bratlund 1991; Eriksen 1996a, 1996b, with further references). The geochronological date of the Ahrensburgian is fixed to the Younger Dryas (Table 12.2). The exceptional concordance of nine radiocarbon dates from the Ahrensburgian level of Stellmoor (Fischer and Tauber 1986) may, to some degree , be caused by a plateau in the radiocarbon curve at 10,000 RCY B.P. (Kromer and Becker 1993). The mean value of the Stellmoor dates (at 10,020 B.P.) is, in fact, somewhat younger than expected from the initial palynological dating, which placed the site in early Younger Dryas (Schtitrumf 1943:37) . On the other hand, recent zooarchaeological analysis indicates that the abundant archaeological remains uncovered at this site may actually have been accumulated within a few decades (Bratlund , personal communication) . 158

PATTERNS OF ETHNOGEOGRAPHIC VARIABILITY

The Ahrensburgian is usually referred to as the last of the truly Paleolithic hunter - gatherer groups in northwestern Europe . However, recent research on the geochronology of Late Pleistocene shoreline displacements in western Sweden and southern Norway, combined with radiocarbon dates and artifact analysis, indicate that some sites belonging to the Fosna-Hensbacka culture, previously considered purely Mesolithic, may actually date to the Late Paleolithic-i.e., the late Younger Dryas (Cullberg 1996:185f; Fischer 1996:164f; Kindgren 1995, 1996; Pr0sch-Danielsen and H0gest0l 1995; Schmitt 1995).

some convincing arguments (e.g. , Kindgren 1995, 1996; Pr0sch -Danielsen and H0gest0l 1995) that at least some Fosna-Hensbacka sites are of a Late Paleolithic age and Younger Dryas date , respectively. The lithic tool inventory of these early Fosna -Hensbacka assemblages is characterized by the occurrence of singleedge points and delicate Ahrensburgian-like tanged points, along with burins, end scrapers, and occasional large, Bromme-like tanged points (Figure 12.10). Chronologically later assemblages-i.e. , the classic Fosna-Hensbacka complex-have an increasing number of geometric microliths, especially lanceolate points and slender trapezes, accompanied by micro -burins and flake axes. These assemblages are unquestionably of Mesolithic-i.e. , Preboreal or early Boreal-age.

For methodological reasons, the age, cultural affinity , and the homogeneity of these inventories has been the object of serious debate during the past 40 years (for a summary of this discussion see, for example, Cullberg 1996 or Schmitt 1995). The sites in question often consist of surface scatters or are covered by very thin layers of top soil. For literally thousands of years, they have thus been exposed to various biogenic and anthropogenic erosional forces. The archaeological remains are.often mixed, organic remains are rare, and scientific dating by radiocarbon and/or pollen analysis has been impossible. Estimations of age have usually been based on purely typological observations or on the age determination of presumably contemporary shoreline displacements. The sites are generally situated on raised beach lines , and the question is whether they were, in fact, seashore settlements sensu stricto or perhaps coastal settlements sensu Lato. As a result of the isostatic upheaval of land, the older coastlines are situated higher than the younger ones . Accordingly, an older seashore settlement might well be situated at exactly the same altitude as a younger coastal site, thereby obscuring their chronological significance. Although problems still remain with respect to these issues of relative and absolute chronology, there are

The situation of the early Hensbacka sites in the west Swedish archipelago (Figure 12.11) leaves no doubt that these people must have had boats. It has accordingly been uggested that they represent a coastal aspect of the late Ahrensburgian complex, probably associated with hunting large sea mammals (Schmitt 1995). Some scholars have even argued that the "classic " Ahrensburgian hunter-gatherers might also have had boats (Tromnau 1987), but we still lack unambiguous evidence to upport this supposition. The presumably Late Paleolithic Fosna-Hensbacka phase is probably closely related to the Ahrensburgian technocomplex. The exact nature of this relationship, however , remains a point of future archaeological research (Kindgren 1996:202f). In my opinion, the recently excavated Galta site is thus early Fosna and not Ahrensburgian, as claimed by the excavators (Pr0sch-Danielsen and H!,?Sgest01 1995). The latter designation should be reserved for the classic assemblages described above.

0





• Figure 12.9- Ahrensburgian lithic tool types.

159



ERIKSEN

Figure 12.10 - Fosna-Hensbacka lithic tool types.

SOURCE CRITICISM: HOW TO DISTINGUISH BETWEEN GEOGRAPHICAL AND CHRONOLOGICAL ASPECTS? Relative and absolute chronologies are roughly summarized in Table 12.2. Unfortunately, there are few reliable scientific dates, especially from the Allen~d period. Despite an increasing number of thoroughly excavated and carefully documented Late Pleistocene sites, there is a notable lack of organic remains suitable for radiocarbon dating. In addition, there is also a lack of other biochronological and chronostratigraphical observations, notably pollen and sediment analysis. Thus, most of the archaeological assemblages are datable only within a relative typological framework. Only a few sites have been dated by means of radiocarbon, pollen, sedimentary, or thermoluminescence analyses . Moreover, these so-called "absolute" chronologies, including the Swedish varve chronology, are all floating chronologies and may only be used in the sense of a relative chronostratigraphic framework. Direct comparisons between these chronologies are, therefore, not always valid. The establishment of a reliable radiocarbon chronology for the period in question is 14 severely hampered by the existence of three C plateaus at 12,700 (Ammann and Lotter 1989), 10,000, and 9,600 RCY B.P. (Kromer and Becker 1993), respectively. For method14 ological reasons, a calibration of Late Pleistocene C dates is still controversial. However , when completed , such a calibration is likely to revolutionize our current perception of the geochronological framework of this period (Street et al. 1994). Within the study area, the sequence of Late Pleistocene archaeological cultures or technocomplexes outlined above is often presented as a continuous and largely endogenous cultural development (Eriksen 1996a: 123; Fischer 1991: 111).

The present discussion is in general accordance with this model. For the time being, however, we have significant difficulty with establishing absolute as well as relative chronologies. A major source -critical problem that is particularly relevant to the present case study is the fact that most of the archaeological sites in question consist only of lithic scatters. Some lithic technocomplexes are technologically and typologically distinctive; yet, for various reasons, the degree of variation among contemporary, and even closely related, inventories (i.e., inventories left by the same hunter- gather group at different sites) may be considerable. Sites are traps of pecific behavior, which may vary on a cyclical scale according to season, function, etc., but they are also the product of various site formation and taphonomic processes. The technological and typological differences observed among different lithic inventories are thus not necessarily cultural or chronological; they may also be seasonal, functional, or even accidental. The absolute contemporaneity of and a definite socio-cultural relationship between different sites may be established through refitting significant components of their lithic inventories. The relative contemporaneity of and a plausible sociocultural relationship between sites may also be asserted by means of a thorough analysis integrating all available information on chronostratigraphy , environmental and site-locational context, subsistence economy, lithic technology, etc. However, as in the case of "behavioral universes," our chances of establishing even a relative archaeological contemporaneity between sites decrease on a spatio-temporal scale. The nature of the sample itself makes it very difficult to " ... identify the role or position of a single site and an individual area in the local and regional universes of these prehistoric hunter -g atherers." In other parts of contemporary Late 160

PATTERNS OF ETHNOG EOGRAPHIC VARIABILITY

Pleistocene Europe, studies of lithic raw material use and procurement strategies may help clarify these issues (see, for instance, various papers in Fisher and Eriksen [eds] in press). This is not the case in the morainic areas of southern Scandinavia. The young morainic areas are rich in high-quality Cretaceous flint in secondary glacial deposits , and even the old morainic areas have reliable supplies of redeposited Cretaceous flint suitable for artifact production. Thus , throughout the entire study area, it is generally impossible to distinguish between locally and regionally procured raw materials . Unfortunately, there does not seem to be an abundance of exotic raw materials , such as amber, jet , fossil or subrecent molluscs (used for adornment , etc.; see Eriksen in press), which could supply information on local or regional mobility strategies or contact networks. From the point of view of raw materials , we are left with an almost uniform lithic inventory.

. . . D D• . 11

From a "culture-historical" point of view, it is obvious that the lithic inventories represent a mere fragment of the original material culture . Thus , our possibilities for discussing past socio-cultural relationships are likewise very weakly founded. As previously indicated , different socio cultural (i.e., demographic, subsistence -economic, ethnic , etc.) relationships may be characterized by different regional expression s, depending on the nature of their perceived or defined boundarie s, as well as on their scale of interaction. In the Magdalenian of southwestern Germany, for example, lithic raw material procurement strategies probably reflect seasonal movements throughout the Swabian Alb, while "ornamental " raw materials, especially fossil and subrecent molluscs , more likely express inter-group contact networks. These two networks run transversely to one another (Eriksen in press).

~

D



3

"

4

0

5

o

0

c:, 0

0

0

0

0

0

0

0 0

0

0

0

0

0

0

e

0

0

0 0

0

0

--

0

0

"

0

0 0

0

a 0

0 0 0

0

0

0 0 0

0

0

0

0

0

• 0

0 0

0

C,

0

0 0

0 0

'

0

j

0

0

0 0

~

0

0 0

C!)

~

0

0

' 0

0

0 0

0

0

0

0

0

0

0

0

0

0

0

'o..,-

0

0

0

o 0

0

0

0

°','•

0

0

0

0 0

0

0

0

0

0

j

0

0

0 0

0

0

0

C>

0

j 0

0

0

0

0

j

0 C,

0

6

0

o

0

0

0

0

0

0 0

0

0

0 0

0

C,

0

'

'

0 0

' '

0

0

0

0

0

0

0

0 0

G

a

0

0

0

0

0

0

0 0

0

0 Q

d

0

0

0

0

0

0

0 0

0

a.

0

0

0

0 0

0 0

0

0 0

0

0

0

0

Q

0

0

0 0

0

0

0

0 ()

0

0

0



0

0

0

0

0

0

0

0

0

0

0

0

0

0

6

0

0

0

0

0

_.{

0

0

0

It

0

0

0

(>

0

0

0

0

0

0

,Q,-

o: o, , · o

0

0

C,

0

0

0

0

0

0

0

0

0

0

0

0

0

C)

0

0

0

~,,--6--,,:~p

0

0

0

0

0

0

0

• ' 0

0

0 C

Figure 12.11- Map showin g the distribution (hatched area ) of early Fo na sites (in southern Norway ) and early Hensbacka sites (in western Sweden ) in relation to the macro-topography of the late Younger Dry as ( 1: ice margin at approximately 10,000 to 9,900 RCY B.P.; 2: tundra ; 3: park-tundra ; 4 : sparse birch forest; 5: birch forest with some pine).

161

ERIKSEN

Organic tools, like portable art objects and ornaments, often represent a higher production cost and thus frequently have a longer lifetime than lithic artifacts. As a consequence, an organic inventory might be characterized by a different regional expression than a lithic inventory. Presumably, each of these different "technoterritories" (sensu Clark 1975: 14f) would represent different socio-cultural relationships and/or scales of interaction. This might be the case in Late Pleistocene northwestern Europe. Consider, for example, the so-called Lyngby clubs; these reindeer antler implements occur in great numbers at the classic Ahrensburgian Stellmoor site (Rust 1943:176f) and are often assigned to this cultural group. However, most of the known specimens are stray finds, and their geographical distribution is omewhat different from that of the Ahrensburgian lithic inventories (Figure 12.12). On the other hand, some of the Lyngby clubs without context may, with equal validity , be attributed to the Bromme culture (by way of radiocarbon dating [Table 12.2, Mickelsmossen] and geographical location [Figure 12.12]), especially because the organic inventory for this culture is practically unknown. Therefore, Lyngby clubs should more likely be assigned to the Tanged Point complex sensu Lato, in accordance with

their known geographical distribution and underscoring the probability that lithic and organic technoterritories are likely to be quite different on a spatio-temporal and sociocultural scale. These considerations are in perfect accordance with observations from contemporary hunter-gatherer groups, where territorial boundaries are generally vague (Lee and De Vore 1968: l 56f). Well defined boundaries are rare, and culturally defined boundaries do not necessarily imply sanctions against trespass: "So if we find boundaries in a given case, we should not commit the frequent error of assuming that they enclose a defended and exclusive territory " (Lee and DeVore 1968: 157).

DISCUSSION Spatio-Temporal Relationships A discussion of patterns of ethnogeographic variability in Late Pleistocene northwestern Europe should proceed from a clear understanding of the patio-temporal and socio-cultural relationships of all the cultural groups involved. Obviously, these are rather unrealistic, perhaps even pretentious,

Figure 12.12-Map showing the distribution of Ahrensburgian sites as defined by their lithic inventories (hatched area) and reindeer antler Lyngby clubs (triangle: one specimen; square : many specimens ; 1: Stellmoor; 2: Arreskov ; 3: Mickelsmossen ).

162

PATTERNS OF ETHNOGEOGRAPHIC VARIABILITY

demands . Based on a consideration of the spatio-temporal distribution of sites over the study area, it is obvious that one of the major challenges is to distinguish between geographical and chronological aspects of variation. The lack of absolute dates and accompanying chronostratigraphical observations is a major problem. We are unable to distinguish between absolute and relative contemporaneity for most of these archaeological sites, and, as a consequence, between the degrees of contemporaneity of certain major cultural groups and/or traditions. It is evident that large cultural complexes do not appear suddenly throughout every part of the final distribution area. There will always be a forward movement of people, information, ideas, or devices into new areas; thus, there will also be some degree of contemporaneity or even co-existence of different cultural traditions. It is thus inevitable that we sometimes find ourselves largely unable to differentiate between time trends, spatial trends, and their joint results. In the Late Pleistocene, southern Scandinavia may be regarded as a frontier zone within which various cultural groups were constantly moving back and forth (Petersen 1993). These complex spatio-temporal relationships are best illustrated by fitting the groups or sites into a chronologically significant grid system completely devoid of archaeological interpretation. This is clearly not possible, and the discussion is therefore subject to certain reservations.

course of the Late Pleistocene the recently deglaciated areas became suitable for human exploitation . All of southern Scandinavia may be considered ice free in, or even before, the early B0lling, but there surely would have been a certain time lag before the major plant and animal resources suitable for hunter-gatherer exploitation had recolonized the region, a time lag of perhaps one or two thousand years. During this period, human exploitation is indeed likely to have been both episodic and ephemeral, as well as limited by seasonal or cyclical fluctuations in resource availability. In the Aller0d, the young morainic areas finally flourished with respect to stable plant and animal resources, and this is immediately reflected in a marked increase in population density. On the other hand, the apparent "depopulation" of major parts of this area in the Younger Dryas is probably because of missing evidence, insufficient conditions for preservation , or an absence of chronostratigraphic observations. Thus, the possibility that some of the missing late Tanged Point (i.e., Ahrensburgian) sites have been falsely attributed to the early Tanged Point complex (i.e., the Bromme culture) cannot be dismissed . Moreover, the Hensbacka sites in the west Swedish archipelago indicate an increased reliance on maritime and coastal resources over time, but, because the Younger Dryas coast-line is now mostly submerged, we find only scant evidence of such coastal exploitation.

Socio-Cultural Relationships Based on the two Magdalenian-related groups (Figures 12.2 and 12.4), we find that, despite obvious differences in extent, the distributions of Hamburgian and Federmesser sites are nonetheless very similar. The nature of their boundaries, however, seems to be distinctive. The northern boundary of the Hamburgian sites appears to be based on purely natural, climatic, and environmental barriers, whereas the Federmesser groups probably have a socio-cultural boundary to the north. There is only a minor peripheral overlap, in terms of actual number of settlement sites, between the largely contemporary Federmesserand Bromme groups (Figures l2.4 and 12.6) . Turning to the Tanged Point technocomplex, we find that there are no significant Bromme-related finds south of the river Elbe. Ahrensburgian finds are also confined to the northwest European plain, although they have a more distinctively southerly distribution than the Bromme finds. Finally, it is worth noticing that the Ahrensburgian and the possibly contemporary early Fosna-Hensbacka phase (Figures 12.8 and 12.11) complement each other with respect to the geographical distribution of sites. Unfortunately, this observation is impaired by the very small number of positively identified early Fosna-Hensbacka sites.

In respect to the possible socio-cultural relationships of the Late Pleistocene hunter-gatherer groups, we face a number of methodological questions relating to, among other things, our ability to define convincingly different archaeological cultures, groups, or even technocomplexes in time and space. The "cultural taxonomy" of Late Pleistocene northwestern Europe outlined in this paper is based primarily on the morphological classification of lithic inventories. Nevertheless, there is frequently an underlying assumption of ethnicity or genetic/generic relationships pertaining to each of these cultural groups-Hamburgian, Federmesser, Bromme, Ahrensburgian, and Fosna-Hensbacka. For various reasons, this is rather problematic. In fact, even the asserted socio-cultural boundary between the Bromme and Federmesser groups may very well be completely arbitrary with respect to the actual social and demographic arrangements of these Allen2~d hunter-gatherers. Unfortunately, evidence on subsistence strategies, settlement and mobility patterns, etc., is still far too inconclusive. Therefore, this boundary is, first and foremost, an archaeological construct and must not be confused with a "territorial boundary." In addition, differential access to flint resources over time and across space are likely to have played an important role in respect to differences in lithic technological traditions. These observations may be less "socio-culturally significant" than at first anticipated. For example, the elaborate and very economical application of soft-percussion, blade production techniques in the Hamburgian inventories, in

It is interesting to note that the Hamburgian sites are located on, or very close to, old morainic soils, while the Bromme sites tend to be located on young morainic soils. In my opinion , this distribution of sites is no coincidence; neither is it a methodological question of missing evidence or inadequate preservation conditions. It is rather a question of when in the 163

ERIKSEN Table 12.3 Schematic Outline Showing Pattern s of Spatio-Temporal and Socio -Economic Variability in Late Pleistocene Northwestern Europe (after Eriksen 1996a, 1996b)

Northwest Germany

South Scandinavia

Preboreal

Early Fosna-Hensbacka - increased importance of maritime and coastal resources. Boats! Probable Ahrensburgian influence or origin. Ahrensburgian - possibility for periodical

Younger mass killing of reindeer. Bow and arrow hunting. At least one possible aggregation Dryas camp. Logistic settlement pattern.

Few Ahrensburgian sites in the south. Ephemeral or episodic exploitation of the area.

Allered

Federmesser - opportunistic stalking of e.g. moose. Probably small social groups and high residential mobility. Forest adaptation.

Bromme - opportunistic stalking of e.g. moose. Probably small social groups and high residential mobility. Park-tundra/ open land adaptation.

Belling

Hamburgian - probably seasonal specialization on reindeer. Stalking with spears and spear-throwers?! Magdalenian influence or origin.

Few Hamburgian sites in the south. Ephemeral or episodic exploitation of the area. Small social groups.

comparison to the more wasteful hard-hammer percussion technique in the Bromme inventories, may be a reflection of Hamburgian flint-knappers economizing scarce, highquality raw materials in their old moraine setting versus a more affluent flint-economy at the young moraine Bromme sites. Incidentally, the evidence from the Ahrensburgian inventories on the old moraine would tend to support this picture of a macro-topographical influence on lithic technology. Unfor-tunately , our knowledge of the corresponding young morainic inventories of the Ahrensburgian remain rather incomplete. Finally, some of the observed differences are very likely to be due to the nature of the sample and the "average" site type. Consider, for example, the mutual relationship between the early Fosna-Hensbacka phase and the apparently contemporaneous late Ahrensburgian; we frankly do not know if this is a matter of two distinct hunter-gatherer populations, or if it is the same population group exploiting different resources, perhaps maritime and terrestrial on a seasonal or cyclical round.

The concluding table (Table 12.3) briefly summarizes current viewpoints with respect to the Late Pleistocene cultural groups in northwestern Europe, vis a vis their spatio-temporal habitat and probable socio-economic characteristics. For methodological reasons, the question of specific socio•cultur al relationships can be discussed on only a speculative level. The general opinion seems to be that there is one major, early B0lling wave of immigration into the area, represented by Hamburgian hunter-gatherers and probably originating in the Magdalenian of central and western Europe. The Federmesser groups very likely did maintain regular (in)direct contact with central European late Magdalenian groups, and the early Aller0d may have witnessed another immigration of people. The Tanged Point complex (the Bromme and Ahrensburgian cultures) presumably represent an indigenous development of cultural (i.e., lithic technolog ical) traditions in southern Scandinavia and northwestern Germany. Future research must address the above mentioned issues of absolute and relative chronologies.

ACKNOWLEDGMENTS CONCLUDING REMARKS Obviously , an attempt at a generalized synthesis of cultural and environmental changes during the period in question is severely hampered by a number of limiting factors. One of the main problems is, indeed, our general inability to distin guish between degrees of contemporaneity of major cultural groups or traditions .

I wish to thank Heather A. Price and Gail Larsen Peterkin for their efforts in organizing the symposia in Minneapolis and New Orleans and for persevering in editing the current volume . Critical comments by Martin Street on an earlier draft of this chapter were of great help. Last , but not least, the financial support of the Danish Research Council for the Humanities is gratefully acknowledged. 164

PATTERNS OF ETHNOGEOGRAPHIC

NOTES

VARIABILITY

Millennial Basis, edited by N.-A. Mamer and W. Karlen, pp. 25-36. Dordrecht: D. Reidel Publishing Company. 14

1. It is stressed that this date is based on C dating of lacustrine sediments which may be up to 400 years older than conventional terrestrial datings (Bjarck 1996: 125). 14

2. This date is also based on C dating of lacustrine material, thus " ... possibly corresponding to an age of c. 10,000 BP in terrestrial plant remains " (Bjorck 1996: 131).

REFERENCES Aaris-S0rensen, K., 1992. Deglaciation chronology and reimigration of large mammals: A south Scandinavian example from Late Weichselian-Early Flandrian. In Mammalian Migration and Dispersal Events in the European Quaternary, edited by W. von Koenigswald and L. Werdelin, pp. 143-149. Courier Forschungsinstitut Senckenberg 153. Frankfurt am Main : Forschungsinstitut Senckenberg. Ammann, B., and A.F. Lotter, 1989. Late Glacial radiocarbon and palynostratigraphy on the Swiss Plateau. Boreas 18: 109-126. Andersen , B.G ., J. Lundquist, and M. Saarnisto, 1995. The Younger Dryas margin of the Scandinavian ice sheet: An introduction. Quaternary International 28:145-146. Andersen, S.H., 1988. A Survey of the Late Palaeolithic of Denmark and southern Sweden. In De la Loire a /'Oder: Actes du Colloque de Liege Decembre 1985, edited by M. Otte, pp. 523-566. British Archaeological Reports International Series 444. Oxford: British Archaeologi cal Reports. Baales, M., 1993. "Head ' em-off-at-the -pass:" Okologie und Okonomie der Ahrensburger Rentierjager im Mittelgebirge. Archiiologische lnformationen 16: 116-119. Barton, N., A.J . Roberts , and D.A. Roe (editors), 1991. The Late Glacial in North-West Europe: Human Adaptation at the End of the Pleistocene. CBA Research Report 77. Oxford: Alden Press . Barton, N., and A.J. Roberts, 1996. Reviewing the British Late Upper Palaeolithic: New evidence for chronological patterning in the late glacial record. Oxford Journal of Archaeology 15:245-265.

Berglund, B.E., and N.A. Marner, 1984. Late Weichselian deglaciation and chronostratigraphy of southern Scandinavia: Problems and present "state of the art ." In Climatic Changes on a Yearly to Millennial Basis , edited by N.-A. Marner and W. Karlen, pp. 17-24. Dordrecht: D. Reidel Publishing Company. Bjarck, S., 1995. A review of the history of the Baltic Sea, 13.0-8 .0 ka B.P. Quaternary International 27: 19-40 . Bjorck, S., 1996. Late Weichselian/Early Preboreal development of the Oresund Strait: A Key area for northerly mammal immigration. In The Earliest Settlement of Scandinavia, edited by L. Larsson , pp. 123- 134. Acta Archaeologica Lundensia 24. Stockholm: Almquist & Wiksell International. Bjarck , S., and G. Digerfeldt , 1984. Climatic changes at Pleistocene/Holocene boundary in the middle Swedish endmoraine zone, mainly inferred from stratigraphic indications . In Climatic Changes on a Yearly to Millennial Basis edited by N.-A. Marner and W. Karlen, pp. 37-56. Dordrecht: D. Reidel Publishing Company. Bokelmann , K., 1978. Ein Federmesser -Fundplatz Schalkholz, Kreis Dithmarschen . Offa 35 :36-54.

bei

Bokelmann, K., 1979. Rentierjager am Gletscherrand in Schleswig-Holstein? Ein Diskussionsbeitrag zur Erforschung der Hamburger Kultur. Offa 36: 12- 22. Bokelmann, K., 1991. Some new thoughts on old data on humans and reindeer in the Ahren burgian tunnel valley in Schleswig-Holstein , Germany. In the Late Glacial in North-west Europe: Human Adaptation at the End of the Pleistocene, edited by N. Barton, A.J. Roberts , and D.A. Roe, pp. 72-81. CBA Research Report 77. Oxford: Alden Press . Bokelmann, K., D. Heinrich, and B . Menke, 1983. Fundplatze des Spatglazials am Hainholz-Esinger Moor , Kreis Pinneberg. Offa 40: 199-239.

Berglund, B.E., 1979. The deglaciation of southern Sweden 13,500-10,000 B.P. Boreas 8:89-117.

Bratlund, B., 1991. A study of hunting lesions containing flint fragments on reindeer bones at Stellmoor, SchleswigHolstein, Germany. In The Late Glacial in North-west Europe: Human Adaptation at The End of The Pleistocene, edited by N. Barton, A.J. Roberts, and D.A. Roe, pp. 193-207. CBA Research Report 77. Oxford: Alden Press.

Berglund, B.E., G. Lemdahl, B. Liedberg-Jonsson, and T. Persson , 1984. Biotic response to climatic changes during the time span 13,000-10 ,000 B.P.: A case study from SW Sweden. In Climatic Changes on a Yearly to

Bratlund , B., 1994. A survey of the subsistence and settlement pattern of the Hamburgian culture in Schleswig-Holstein. Jahrbuch des Romisch-Germanischen Zentralmuseums Mainz 41:59-93. 165

ERI KSEN

Burdukiewicz, J.-M., 1986. The Late Pleistoc ene Shouldered Point Assemblages in Western Europe. Leiden: E.J. Brill. Caspar, J.-P., and M. De Bie, 1996 . Preparing for the hunt in the Late Paleolithic camp at Rekem , Belgium. Journal of Field Archaeolog y 23 :437-460. Clark , G ., 1975 . The Earlier Stone Age Settlement of Scandinavia . London: Cambridge University Press. Cullberg, C ., 1996. West Sweden: On the earliest settlements. In The Earliest Settlement of Scandinavia , edited by L. Larsson , pp . 177- 189. Act a Archa eologica Lunden sia 24 . Stockholm : Almqui st & Wiksell International.

Fi scher, A., 1991. Pioneer s in deglaciated landscape s: The expansion and adaptation of Late Palaeolithic societies in southern Scandinavia. In The Late Glacial in north-west Europ e: Human adaptation at the end of the Pleistocene, edited by N. Barton, A.J. Robert s, and D.A . Roe , pp . 100-1 2 1. CBA Research Report 77 . Oxford : Alden Press. Fischer , A., 1996. At the border of human habitat: The Late Palaeolithic and Early Mesolithic in Scandinavia. In The Earliest Settl ement of Scandina via, edited by L. Lars son, pp . 157- 176. Act a Archaeolo gica Lunden sia 24. Stockholm: Almqui st & Wiksell International. 14

Cziesla , E., 1992. Ahrensbur ger Jager in Sildwestdeutschland ? Archdologis ches Korrespond enzblatt 22: 13-2 6. De Bie , M ., and P.M. Vermeersch , 1998 . Pleistoce neHolocen e transition in the Benelu x. In As the World Warmed: Proceedings of an INQ UA Symposium on the Archaeolog y of the Pleistocene-Holocene Transition, edited by B.V. Eriksen and L.G . Straus. Quaternary International 49-50 :29-43. Eriksen, B.V., 1991. Change and Continuity in a Prehistoric Hunter-Gatherer Society . Archaeologica Venatoria 12. Tilbingen: Archaeologica Venatoria. Eriksen, B .V., 1996a. Regional variation in Late Pleistocene subsistence strategies: Southern Scandinavian reindeer hunters in a European context. In The Earliest Settlement of Scandinavia, edited by L. Larsson, pp. 7-21 . Acta Archaeologica Lundensia 24. Stockholm : Almquist & Wiksell International. Eriksen, B.V., 1996b . Resource exploitation, subsistence strategies, and adaptiveness in Late Pleistocene - Early Holocene northwest Europe. In Humans at the End of the Jee Age: The Archeolog y of the Pleistocene Holocene Transition, edited by L.G. Straus, B.V. Eriksen, J.M . Erlandson , and D .R. Yesner, pp. 101-128 . New York : Plenum Pres s. Eriksen, B.V., in press. Fossil molluscs and exotic raw materials in late glacial and early postglacial find contexts : A complement to lithic studies . In Lithic Raw Material Economy in Late Glacial and Early Postglacial Western Europe, edited by L.E. Fisher and B.V. Eriksen. Ann Arbor: International Monographs in Prehistory.

Fischer, A., and H . Tauber, 1986. New C datin gs of Late Palaeolithi c culture s from northwestern Europe. Journal of Danish Archaeology 5:7- 13. Fisher, L.E ., and B.V. Eriksen (editors), in press. Lithic Raw Material Economy in Late Glacial and Ea rly Postglacial Western Europe. Ann Arbor : International Monograph s in Prehi tory. Floss , H., 1994. Rohmaterialversorgung im Paldolithikum des Mittelrheingebietes.Romisch-Germanisches Zentralmuseum Monographien 21. Bonn : Rudolf Habelt. Foley, R., 1981. A model of regional archaeological structure. Proceedings of the Prehistori c Society 47:1- 17. Gates, D.M. , 1993 . Climate Change and Its Biological Consequences. Sunderland: Sinauer Associates , Inc. Grfl)nnow, B., 1985. Meiendorf and Stellmoor revisited: An analysis of Late Palaeolithic reindeer exploitation. Acta Archaeologica 56: 131-166 . Hartz , S., 1987. Neue spatpalaolithische Fundplatze bei Ahrenshoft , Kreis Nordfriesland . Offa 44:5- 52 . Holm, J., and F.Rieck, 1987. Die Hamburger Kultur in Danemark. Archdologisches Korrespondenzblatt 17:151- 168. Holm , J., and F. Rieck, 1992. lstidsjcegere ved Jelss¢erne . Skrifter fra Museumsradet for S0nderjyllands Amt 5. Haderslev : Museumsradet for S0nderjyllands . Huntley, B., and H.J.B. Birks, 1983. An Atlas of Past and Present Pollen Maps for Europe: 0- 13,000 Years Ago . Cambridge : Cambridge University Press.

Fischer , A., 1988. A Late Palaeolithic flint workshop at Egtved , East Jutland: A glimpse of the Federmesser culture in Denmark . Journal of Danish Archaeolog y 7 :7- 23 .

Iversen, J ., 1942 . En Pollenanalytisk tidsfrestelse af Ferskvandslagene ved Nfl)rre Lyngby . Meddelelser fra Dansk Geologisk Forening 10:130- 151.

Fischer , A., 1990. A Late Palaeolithic "school " of flint-knapping at Trollesgave, Denmark: Results from refitting. Acta Archaeologica 60:33-49.

Iversen, J ., 1954. The Late Glacial flora of Denmark and its relation to climate and soil. Danmarks Geologiske Unders¢gels e 11(80):87-1 19.

166

PATTERNS OF ETHNOGEOGRAPHIC VARIABILITY

Jacobi, R., 1991. The Creswellian, Creswell, and Cheddar. In the Late Glacial in North-West Europe: Human Adaptation at the End of the Pleistocene, edited by N. Barton, A.J. Roberts and D.A. Roe, pp. 128-140. CBA Research Report 77. Oxford: Alden Press.

Liljegren, R., and J. Ekstrom, 1996. The terrestrial Late Glacial fauna in south Sweden. In The Earliest Settlement of Scandinavia, edited by L. Larsson, pp. 135-139. Acta Archaeologica Lundensia 24. Stockholm: Almquist & Wiksell International.

Jessen, K., 1934. Archaeological dating in the history of north Jutland's vegetation. Acta Archaeologica 5: 185-214.

Lowe, J.J ., and J.M. Gray, 1980. The stratigraphic subdivision of the Late Glacial of NW Europe: A discussion. In Studies in the Late Glacial of North-West Europe, edited by J.J. Lowe, J.M. Gray, and J.E. Robinson, pp. 157-175. Oxford: Pergamon Press.

Jochim, M., 1987. Late Pleistocene refugia in Europe. In The Pleistocene Old World: Regional Perspectives, edited by 0. Soffer, pp. 317-331. New York: Plenum Press. Kindgren, H., 1995. Hensbacka -Hogen-Hornborgasjon: Early Mesolithic coastal and inland settlements in western Sweden. In Man and Sea in the Mesolithic: Coastal Settlement Above and Below Present Sea Level, edited by A. Fischer, pp. 171-184. Oxbow Monograph 53. Oxford: Oxbow Books. Kindgren , H., 1996 . Reindeer or seals? Some Late Palaeolithic sites in central Bohuslan. In The Earliest Settlement of Scandinavia, edited by L. Larsson, pp. 191- 205. Acta Archaeologica Lundensia 24. Stockholm: Almquist & Wiksell International. Kolstrup, E ., 1991. Palaeoenvironmental developments during the Late Glacial of the Weichselian. In the Late Glacial in North-West Europe: Human Adaptation at the End of the Pleistocene, edited by N. Barton, A.J. Roberts, and D.A. Roe, pp. 1-6. CBA Research Report 77. Oxford: Alden Press. 14

Kromer, B., and B. Becker, 1993. German oak and pine C calibration, 7200-9439 B.C . Radiocarbon 35:125-135. Lagerlund, E., and M. Houmark-Nielsen, 1993. Timing and pattern of the last deglaciation in the Kattegat region, southwest Scandinavia. Boreas 22:337-347. Larsson, L., 1991. The Late Palaeolithic in southern Sweden : Investigations in a marginal region. In the Late Glacial in North-West Europe: Human Adaptation at the End of the Pleistocene, edited by N . Barton, A.J. Roberts, and D.A. Roe, pp. 122-127. CBA Research Report 77 . Oxford: Alden Press.

Madsen, B., 1983. New evidence of Late Palaeolithic Settlement in East Jutland. Journal of Danish Archaeology 2: 12-31. Madsen, B., 1992. Hamburgkulturens flintteknologi i Jels. In Istidsjcegere ved Jelss¢erne, edited by J. Holm and F. Rieck, pp. 93-131. Skrifter fra Museumsradet for S0nderjyllands Amt 5. Haderslev : Museumsradet for S0nderjyllands. Madsen, B., 1996. Late Palaeolithic cultures of south Scandinavia: Tools, traditions, and technology. In The Earliest Settlement of Scandinavia, edited by L. Larsson, pp. 61-73. Acta Archaeologica Lundensia 24. Stockholm : Almquist and Wiksell International. Mathiassen, T., 1946 . En senglacial Boplads ved Bromme . Aarb¢ger for Nordisk Oldkyndighed og Historie 1946: 121-197. Petersen, E.B., 1993. The Late Palaeolithic and the Mesolithic. In Digging into the Past: 25 Years of Archaeology in Denmark, edited by S. Hvass and B. Storgaard, pp. 46-49 . Aarhus: Jutland Archaeological Society . Petersen, K.S., 1984. Late Weichselian sea-levels and fauna communities in northern Vendsyssel, Jutland, Denmark. In Climatic Changes on a Yearly to Millennial Basis, edited by N.-A. Marner and W. Karlen, pp. 63-68 . Dordrecht : D. Reidel Publishing Company . Petersen, K.S ., 1985. The Late Quaternary history of Denmark: The Weichselian icesheets and land/sea configuration in the Late Pleistocene and Holocene . Journal of Danish Archaeology 4:7-22.

Larsson, L. , 1996. The colonization of south Sweden during the deglaciation. In The Earliest Settlement of Scandinavia, edited by L. Larsson, pp. 141-155. Acta Archaeologica Lundensia 24. Stockholm: Almquist & Wiksell International.

Petersen, P.V., and J. Johansen, 1991. S0lbjerg I: An Ahrensburgian site on a reindeer migration route through eastern Denmark. Journal of Danish Archaeology 10:20-37.

Larsson, L. (editor), 1996. The Earliest Settlement of Scandinavia. Acta Archaeologica Lundensia 24. Stockholm: Almquist & Wiksell International.

Pr0sch-Danielsen, L., and M. H0gest0l, 1995. A coastal Ahrensburgian site found at Galta, Rennes0y, southwest Norway. In Man and Sea in the Mesolithic. Coastal Settlement Above and Below Present Sea Level, edited by A. Fischer, pp. 123-130. Oxbow Monograph 53. Oxford: Oxbow Books.

Lee, R.B., and I. DeVore (editors), 1968. Man the Hunter. New York: Aldine Publishing Company. 167

ERIKS EN

Rust, A. , 1937. Das altstein zeitliche Rentierjdg erlager Meiendorf. Neumtinster: Karl Wachholtz Verlag. Rust , A., 1943. Die alt - und mittelstein zeitlichen Funde von Stellmoor . Neumilnster: Karl Wachholtz Verlag. Schild , R., 1976. The Final Paleolithic settlements of the European plain . Scientific American 234(2):88-99. Schild , R., 1984. Terminal Paleolithic of the north European plain: A review of lost chances , potentials, and hopes. Advances in World Archa eology 3:193- 274. Schild , R., 1996. The north European plain and eastern subBalticum between 12,700 and 8,000 B.P. In Humans at the End of the Jee A ge. The Ar cheology of the Pleistocene - Holocen e Transition , edited by L.G . Straus , B .V. Eriksen , J.M . Erland son, and D.R . Yesner, pp . 129- 157. New York : Plenum Press. Schmitt , L., 1995. Th e West Swedish Hen sbacka : A maritime adaptation and a seasonal expression of the North Central European Ahrensburgian? In Man and Sea in the Mesolithic: Coastal Settlement Above and Below Present Sea Level, edited by A. Fischer, pp. 161- 170. Oxbow Monograph 53. Oxford: Oxbow Books . Schtitrumf, R., 1943. Die pollenanalythische Untersuchung der Rentierjagerfundstatte Stellmoor in Holstein . In Die alt- und mittelsteinzeitlichen Funde von Stellmoor, edited by A. Rust, pp. 6- 45. Neumtinster: Karl Wachholtz Verlag. Schwabedissen , H., 1954. Die Federmessergruppen des nordwesteuropdischen Flachlandes. Offa -Bticher 9. Neumtinster: Karl Wachholtz Verlag. Soffer, 0. , 1987. Upper Paleolithic connubia , refugia, and the archaeological record from eastern Europe . In The Pleistocene Old World: Regional Perspectives, edited by 0 . Soffer, pp . 333- 348. New York: Plenum Press. Soffer, 0. , and C. Gamble (editors), 1990. The World at 18,000 B.P. Volume 1: High Latitudes. London: Unwin Hyman. Stapert, D., 1992. Rings and Sectors: lntrasite Spatial Analysis of Stone Age Sites. Groningen: Rijksuniversiteit Groningen. Straus , L.G., 1991. Human geography of the late Upper Paleolithic in Western Europe : Present state of the

question . Journal

of

Anthropologi cal

Research

47:259- 278. Street, M. , M . Baale s, and B. Weninger , 1994. Absolute Chronologie de s spaten Palaolithikums und des Fri.ihmesolithikum im nordlichen Rhe inland . Archdolo gisches Korresponden zblatt 24: 1-28. Taute, W., 1963. Funde der spatpalaolithischen "Federmess er-Gruppen" aus dem Raum zwischen mittlerer Elbe und Weichsel. Berliner Jahrbuch fur Vorund Fruhgeschichte 3: 62-1 11. Taute, W., 1968. Die Stielspitzen-Gruppen im nordlichen Mitteleuropa. Fundamenta Reihe AS. Koln: Bohlau Verlag. Tromn au, G., 1975. Neue A usgrabungen im A hrensburge r Tunn eltal . Offa-Bilcher 33. Neumi.inster : Karl Wachholt z Verlag. Tromnau , G ., 1976. Rentierjdger der Spdteis zeit in Norddeutschland. Wegweiser zur Vorund Fri.ihgeschichte Niedersach sens 9 . Hildesheim: Verlag August Lax. Tromnau, G ., 1987. Late Palaeolithic reindeer-hunting and the use of boats. In Late Glacial in Central Europe : Culture and Environment, edited by J.-M. Burdµckiewicz and M . Kobusiewicz , pp. 95- 105. Wroclaw: Polish Academy of Science. Usinger, H., 1975. Pollenanalytische und stratigrafische Untersuchungen an zwei Spatglazialvorkommen i Schleswig-Holstein. Mitteilungen der Arbeitsgemeinschaft Geobotanik in Schleswig-Holstein und Hamburg

25:1- 183. van Geel, B., and E. Kolstrup, 1978. Tentative explanation of the Late Glacial and Early Holocene climatic changes in north-western Europe. Geologie en Mijnbow 57:87- 89. Weniger, G., 1991. Uberlegungen zur Mobilitat jagerischer Gruppen im Jungpalaolithikum. Saeculum 42:82-103 . Wobst, H.M., 1990. Afterword: Minitime and megaspace in the Palaeolithic at 18 K and otherwise. In The World at 18,000 B.P. Volume 1: High Latitudes, edited by 0. Soffer and C. Gamble , pp. 331- 343. London : Unwin Hyman .

168

13 Final Palaeolithic in the Northwest: Migrations and Seasons Marcel Otte (U niversite de Liege)

ABSTRACT The northern plains form a geographical framework that was favorable to ethnic exchanges when the glaciers melted. One can observe phenomena of demographic expansion towards an unoccupied zone, as well as processes of cultural exchange. Seen from this viewpoint, the northwest forms a "laboratory" that favors the study of models of adaptation during the Late Glacial. ENVIRONMENT Through both its position and its history, the northern region is of crucial interest for the study of the Upper Paleolithic . Boucher de Perthes ( 1857-1864) in the mid - 19th century and Victor Commont ( 1913) at the turn of this century were geographically distant from the central institutions. They were pioneers, not only because of their audacity, but also through their fertile imaginations. In the Pleistocene, the frontier zones were also the frontiers of the nascent science of prehistory. Intellectual liberty was necessary for the approach to particular phenomena, within the limits of different geographic territories. The demonstration of human antiquity and the construction of a terrestrial quaternary chronology were fundamental works because they were conceived in marginality-in the fertile fringes where the mind wonders about things. Observation does not obey dogmas. During the late Pleistocene, the profound geographical and climatic modifications brought about a succession of very varied settings. One can thus establish a coherent relationship between cultural phenomena (migrations, adaptations, contacts) and the natural conditions in which they took place . This situation allows one to not only better understand the region's history, but also to draw closer to the general behavioral mechanisms that lie at the origin of local traditions. Fluctuations in sea-level, like the expansions of periglacial areas, by turns favored or limited the possibilities for intercultural contacts. They developed or reduced the successively accessible occupation areas and resulted in different networks. The aim of the following observations is merely to propose the general outline of this scenario. ANTECEDENTS The story starts in the Gravettian, which is known in the northern plains in an archaic and peculiar form: big, stemmed tools, associated with shaping by flat flaking . Found in England , the Rhineland, Thuringia , Belgium, and the Ile-de France (Hahn 1969; Jacobi 1980; Otte 1990; Schmider 1971), it has been dated to between 27,000 and 28,000 B.P. at Maisieres and Kent's Cavern. In this inter-pleniglacial phase, it appears to be limited to this geographical area and has affinities with eastern Europe, where it may have originated .

In the following stage, the development of the "Perigordian Va" extends into southwest France, apparently linked to the climatic deterioration that begins at this time. In the north , the sites occupied in this phase occur both in caves (Spy, Trou Magrite, Kent's Cavern) and in the open air (Nemours, Huccorgne, Pulborough, Muffendorf). Thus, the recording of context benefits from two complementary sedimentary systems. This is also the first time that a tradition spans the whole of the continent, which was soon to be divided into distinct provinces. DIVISION One can observe a break in occupation of the northern areas that is probably linked to the climatic deterioration of the second glacial stage (Pleniglacial B). This situation makes any study of later phenomena particularly fruitful because it enables one to target certain mechanisms that remain confused elsewhere, such as migrations and adaptations. During the really harsh period about 20,000 to 15,000 years ago, the southern provinces become individualized through an absence of lateral contacts. The Solutrean takes form in the Iberian peninsula and southern France, while the Epigravettian continues to evolve in Italy and in the Balkan southeast. RECONQUEST (Figure 13.1) Cultural unity was restored in central Europe at the start of the Late Glacial (Dryas I), mainly in the hilly zone and through a movement that was primarily of western origin. The late phases of the Solutrean had seen the development of equipment adapted to cold steppic areas, through the adaptations of bone tools (e.g., spear-throwers) and light stone tools. Thus , the expansion to the north was made possible both by technical progress and by the retreat of periglacial areas, which, in fact, it slightly precedes. In view of the preceding demographic deficiency, the hypothesis of a migratory movement accounting for the origin of the Magdalenian remains the most plausible explanation, at least in these zones which lie outside the phyletic origins of the Magdalenian phenomenon. This Magdalenian phenomenon can be found in very homogeneous forms in Moravia (Pekarna [Valoch 1960]), in Poland (Maszycka) , in central

DRYAS

::.·_)} _

t

Ngure 13.1-Diffusion took place at the start of the late glacial, through the hilly zone . These Magdalenian migratory movements were perhaps polygenic and imply the displacement of people and of customs that were linked to new adaptational capacities more than to eventual responses to climatic demands . l, sites in the center, in Belgium (Vaucelles); 2, central Germany ; 3, Moravia; and. 4, Poland . Seasonal migrations took place in the northern plains, along the fluvial axes, in game-filled territorie s that had emerged from the North Sea (double-pointed arrow) .

170

FINAL PALAEOLITHIC IN THE NORTH - WEST

/

',

\I

\

-

.,.Jf!ifr,/.Vb

~/!: f.~?J~ •{ ~ rrt ,...,·it ., :r..,. , :

~~

--~r, ::_.. ;;,~ -,,-

:~ t~t

8

< u ~ ~ < ~ ~ uz

~ 0

2 "' ~ z iii (.')

I-

~

~

V)

~

0

0 I-

0

~

~

.,,

---·· ·--·

II)

z3

-- -- . __.....-.. - ·- -···-- · ,

z z ~ z

~

z

ID ~ ()>- It? ::> 0 ~ 0 0 ~ ~

II)

w :j

< ...., < ~ V z ;;, w

.I:

~

Ill

~ ~

I-

II)

~

ii::

~ z :Q ~ ti; Ill

;J >

~

C0

I >

- ·· -·

a::

3 C0

~

u ::,

-


·-- -·-

z t z

~> g ::, 0 ::,

~

I-

::,

z ~ ~t ii! ~ 0 ~::, ~u 0~ ~ I= ::J II)

0

~ I-

- . ·--- -

·-- ··- ·-·-· -···

0

V

Ill

0 C0

w . a::

13 0 a,

Ill

z 0< 11'1 0 ::> ::J

~ I-

~

u ::>

~ ~ I-

II)

~ II)

r

::>

..

::> u

-
0

·-- __.. ,

...., w

~

1 ~ 8lQ ±.

(I)

z

w -' w

1 iIX 0 I

w

er

CX)

UJ

0

::>

0

~ 0,:

:::> ::>

0

w

t:l w

:5 ~

~

0

8 I

g w

w

~

zw

w 0

0

~

a..

....J

0a::

::>

lU

(I)

g