Hidden Landscapes of Mediterranean Europe: Cultural and methodological biases in pre- and protohistoric landscape studies; Proceedings of the international meeting Siena, Italy, May 25-27, 2007 9781407309033, 9781407338842

Table of contents :
9781407309033.p 265.pdf
Pages from 9781407309033.t
Pages from 9781407309033.t-2
Pages from 9781407309033.t-3
Pages from 9781407309033.t-4
Front Cover
Title Page
Copyright
Table of Contents
Preface
Introduction
Problems of chronology and function in survey assemblages: the 1999 Hidden Landscape debate reviewed
Hidden Landscapes: METHODS AND METHODOLOGY
1. Making sense of a hidden Sicilian landscape: geoarchaeological and GIS investigations of field-survey data
2. CALEROS: an erosion-deposition model for landscape archaeology
3. A prehistoric hidden landscape in an alluvial plain: investigations in the Florentine area
4. Intensive survey and protohistoric settlement in the middle Guadiana basin (Badajoz, Spain)
5. From space to place or from site to landscape? Mind the gap
6. Magnetic susceptibility prospection in northern Calabria
7. Integrating digital aerial photogrammetry and geophysical prospection in the Gargano fint mining landscape (south-eastern Italy)
8. Grotta del Romito (Cosenza): latest Pleistocene landscape
Hidden Landscapes: MOUNTAINS AND UPLANDS
9. The Visible Landscape: inferring Mesolithic settlement dynamics from multifaceted evidence in the south-eastern Alps
10. Toward an understanding of archaeological visibility: the case of the Trentino (Southern Alps)
11. Buried landscapes and cultural landscapes of the mountainous areas of Lombardy in pre- and protohistory
12. The Southern French Alps Landscape Project:an archaeological and palynological study of high-altitude settlement areas in the Southern French Alps
13. Fire making water on the Ligurian Apennines
14. Shepherds of a coastal range: the archaeological potential of the Velebit mountain range (eastern Adriatic)
15. A hidden prehistoric landscape in the Region of Madrid (Spain): the significance of the mountains during the 3rd millennium BC
16. Mountains and shore: sites and communication routes of Copper Age western Liguria
17. Ritual use of a mountain landscape: Iron Age sites in the lower Valle Sabbia (Brescia – Italy)
18. A journey through ‘hidden’ or forgotten landscapes in the northern Veneto. Patterns and biases in material culture and research methods
19. Two hidden landscapes in central Portugal: Rego da Murta (Alvaiázere) and Ocreza (Mação)
Hidden Landscapes: SEDIMENTARY BASINS
20. Landscape development of the coastal plains of Rome and Grosseto between 20,000 and 3,000 years ago
21. 20,000 Years of landscape evolution at Ca’Tron (Venice, Italy): palaeoenvironment, archaeology, virtual reality webgis
22. A hidden Neolithic landscape in Apulia (southern Italy)
23. Early farming landscapes in the Friuli plain (NE Italy)
24. Remote sensing analysis in the Florentine plain: investigating wetland contexts by comparing QuickBird images and multitemporal aerial photographs
25. Soil use from the Late Chalcolithic to the Early Middle Bronze Age in the central Po plain (northern Italy): new data from buried soils
26. Micromorphology and mineralogy of the SEV core (Santarém, Portugal): evolution from a transitional to a continental sedimentary environment during the Holocene
27. Deposits and prehistoric occupation of Ribeira Ponte da Pedra (Central Portugal)
28. The rediscovery of a 5000-year-old landscape: a copper-age village on the shores of ancient lake Maccarese (Fiumicino, Rome)
29. A hidden perilacustrine settlement: a village and its fields during the Middle Bronze Age
30. A reconstruction of geomorphological and environmental transformations and the peopling of the territory south-east of Rome during recent prehistory
31. Palaeo-environmental observations on the ancient Holocene of the Florentine area
Hidden Landscapes: DISTRIBUTION PATTERNS
32. Fossil Landscapes in the Fiora Valley. Settlement development of prehistoric and late prehistoric communities read through GIS and Remote Sensing
33. A hidden urban landscape: Populonia between survey and excavations
34. ‘Background noise’ and landscape exploitation in the Late Iron Age Andalusian countryside
35. The region of Marghine-Planargia in Sardinia (Italy): digital terrain modelling and spatial analysis in archaeology using GIS software. Hidden landscapes from the Late Neolithic to the Early Iron Age

Citation preview

BAR S2320 2011 VAN LEUSEN, PIZZIOLO & SARTI (Eds) HIDDEN LANDSCAPES OF MEDITERRANEAN EUROPE

B A R van Leusen 2320 cover.indd 1

Hidden Landscapes of Mediterranean Europe Cultural and methodological biases in pre- and protohistoric landscape studies Proceedings of the international meeting Siena, Italy, May 25-27, 2007 Edited by

Martijn van Leusen Giovanna Pizziolo Lucia Sarti

BAR International Series 2320 2011 14/12/2011 11:22:36

Hidden Landscapes of Mediterranean Europe Cultural and methodological biases in pre- and protohistoric landscape studies Proceedings of the international meeting Siena, Italy, May 25-27, 2007

Edited by

Martijn van Leusen Giovanna Pizziolo Lucia Sarti

BAR International Series 2320 2011

ISBN 9781407309033 paperback ISBN 9781407338842 e-format DOI https://doi.org/10.30861/9781407309033 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix M. VAN LEUSEN, G. PIZZIOLO, L. SARTI Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi J. BINTLIFF Problems of chronology and function in survey assemblages: the 1999 Hidden Landscape debate reviewed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Hidden Landscapes: Methods and Methodology 1

M. FITZJOHN, G. AYALA Making sense of a hidden Sicilian landscape: geoarchaeological and GIS investigations of field-survey data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2

H. FEIKEN, R. VAN BEEK, T. VAN ASCH, M. VAN LEUSEN CALEROS: an erosion-deposition model for landscape archaeology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3

G. PIZZIOLO, L. SARTI A prehistoric hidden landscape in an alluvial plain: investigations in the Florentine area . . . . . . . . 17

4

V. MAYORAL HERRERA, S. CELESTINO PÉREZ, S. WALID SBEINATI Intensive survey and protohistoric settlement in the middle Guadiana basin (Badajoz, Spain) . . . . 27

5

S. CAMPANA From space to place or from site to landscape? Mind the gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

6

A. KATTENBERG, M. VAN LEUSEN Magnetic susceptibility prospection in northern Calabria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

7

A. GALIBERTI, R. SALVINI, M. TARANTINI, F. MANTOVANI, M. BOTTACCHI, M. LINO, C. ROSSI, M. MONDET Integrating digital aerial photogrammetry and geophysical prospection in the Gargano flint mining landscape (south-eastern Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

8

A.C. COLONESE, M. GHINASSI, Z. DI GIUSEPPE, L. GOVONI, D. LO VETRO, G. MALAVASI, F. MARTINI, S. RICCIARDI, B. SALA Grotta del Romito (Cosenza): latest Pleistocene landscape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Hidden Landscapes: Mountains and Uplands 9

F. FONTANA, A. GUERRESCHI, M. PERESANI The Visible Landscape: inferring Mesolithic settlement dynamics from multifaceted evidence in the south-eastern Alps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

10 F. CAVULLI, S. GRIMALDI, A. PEDROTTI, D.E. ANGELUCCI Toward an understanding of archaeological visibility: the case of the Trentino (Southern Alps) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 11 R. POGGIANI KELLER Buried landscapes and cultural landscapes of the mountainous areas of Lombardy in pre- and protohistory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 12 K. WALSH, F. MOCCI, S. RICHER The Southern French Alps Landscape Project: an archaeological and palynological study of high-altitude settlement areas in the Southern French Alps . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 13 R. MAGGI, A. DE PASCALE Fire making water on the Ligurian Apennines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 14 S. FORENBAHER Shepherds of a coastal range: the archaeological potential of the Velebit mountain range (eastern Adriatic) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 15 P. RÍOS MENDOZA A hidden prehistoric landscape in the Region of Madrid (Spain): the significance of the mountains during the 3rd millennium BC . . . . . . . . . . . . . . . . . . . . . . . . . . 123 16 N. CHIARENZA Mountains and shore: sites and communication routes of Copper Age western Liguria . . . . . . . . . 133 17 R. POGGIANI KELLER, M. BAIONI Ritual use of a mountain landscape: Iron Age sites in the lowerValle Sabbia (Brescia – Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 18 S. DE NARDI A journey through ‘hidden’ or forgotten landscapes in the northern Veneto. Patterns and biases in material culture and research methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 19 A. FIGUEIREDO, L. OOSTERBEEK Two hidden landscapes in central Portugal: Rego da Murta (Alvaiázere) and Ocreza (Mação) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Hidden Landscapes: Sedimentary Basins 20 A. ARNOLDUS-HUYZENDVELD Landscape development of the coastal plains of Rome and Grosseto between 20,000 and 3,000 years ago . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

21 P. MOZZI, A. BONDESAN, M.S. BUSANA, P. KIRSCHNER, A. MIOLA, S. PESCARIN, M.C. VILLANI 20,000 Years of landscape evolution at Ca’ Tron (Venice, Italy): palaeoenvironment, archaeology, virtual reality webgis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 22 M. CALDARA, I.M. MUNTONI, G. FIORENTINO, M. PRIMAVERA, F. RADINA A hidden Neolithic landscape in Apulia (southern Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 23 A. FONTANA, A. PESSINA Early farming landscapes in the Friuli plain (NE Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 24 R. SALVINI, G. PIZZIOLO, L. SASSARINI, A. TROTTA, I. CALLEGARI, L. SARTI Remote sensing analysis in the Florentine plain: investigating wetland contexts by comparing QuickBird images and multitemporal aerial photographs. . . . . . . . . . . . . . . . . . . . . 201 25 M. BERNABÒ BREA, M. CREMASCHI, L. BRONZONI, F. PAVIA, C. ROVESTA Soil use from the Late Chalcolithic to the Early Middle Bronze Age in the central Po plain (northern Italy): new data from buried soils . . . . . . . . . . . . . . . . . . . . . . . . 207 26 C. NICOSIA, M.T. AZEVEDO, S. FAVARETTO, A. MIOLA, P. MOZZI, E. NUNES, I. SOSTIZZO Micromorphology and mineralogy of the SEV core (Santarém, Portugal): evolution from a transitional to a continental sedimentary environment during the Holocene . . . . . . . . . . . 215 27 P. ROSINA Deposits and prehistoric occupation of Ribeira Ponte da Pedra (Central Portugal) . . . . . . . . . . . . 221 28 A. MANFREDINI, C. CONATI BARBARO, G. CARBONI, C. GIRAUDI The rediscovery of a 5000-year-old landscape: a copper-age village on the shores of ancient lake Maccarese (Fiumicino, Rome) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 29 M. ANGLE, E. SACCHI, A. ZARATTINI A hidden perilacustrine settlement: a village and its fields during the Middle Bronze Age . . . . . . 231 30 A.P. ANZIDEI, G. CARBONI, A. CELANT, R. FUNICIELLO, G. GIORDANO A reconstruction of geomorphological and environmental transformations and the peopling of the territory south-east of Rome during recent prehistory . . . . . . . . . . . . . . . . 237 31 M. CARRA, P. MAZZA, S. PERUSIN, L. SARTI Palaeo-environmental observations on the ancient Holocene of the Florentine area . . . . . . . . . . . 241 Hidden Landscapes: Distribution Patterns 32 N. NEGRONI CATACCHIO, M. RUSCONI CLERICI, A. TAGLIABUE Fossil landscapes in the Fiora Valley. Settlement development of prehistoric and late prehistoric communities read through GIS and Remote Sensing. . . . . . . . . . . . . . . . . . . . 249

33 F. CAMBI, V. ACCONCIA A hidden urban landscape: Populonia between survey and excavations . . . . . . . . . . . . . . . . . . . . . 255 34 V. MAYORAL HERRERA, A. URIARTE, T. CHAPA, J. VICENT, A. CABRERA ‘Background noise’ and landscape exploitation in the Late Iron Age Andalusian countryside . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 35 G. CATTARI, C. TOZZI, M. BISSON The region of Marghine-Planargia in Sardinia (Italy): digital terrain modelling and spatial analysis in archaeology using GIS software. Hidden landscapes from the Late Neolithic to the Early Iron Age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271

Preface

These are the proceedings of an international research meeting held at the Collegio S. Chiara of the University of Siena, Italy, on May 25-27, 2007. The conference, on the theme of ‘Hidden Landscapes of Mediterranean Europe: Cultural and methodological biases in pre- and protohistoric landscape studies’, was promoted by the Netherlands Organization for Scientific Research NWO; the Department of Archaeology and Art History of the University of Siena; and the Dottorato di Ricerca in Preistoria e Archeologia delle Province Romane of the University of Siena. The conference program consisted of two days of plenary sessions followed by a half-day of round table discussions. Poster sessions were held in the Collegio S. Chiara in the evenings. The response to our call for papers, which went out in January 2007, was excellent (28 papers and 20 posters submitted), with authors drawn from both classical and prehistoric archaeology, from environmental and earth sciences, and from state heritage management bodies as well as universities. Case studies, though centering on Italy, were drawn from most of the northern side of the Mediterranean basin, from the Atlantic to the Adriatic. Despite our best efforts, we did not succeed in preparing these proceedings for publication within a year of the conference (perhaps too optimistic a target). The multidisciplinary character of the articles and, especially, our wish to produce a high-quality English-language publication to attract the widest possible international readership, have required a lengthy editing process. We believe the result, which now lies before you, was well worth the wait: it provides a broad and exciting view of recent research into the hidden landscapes of the northern Mediterranean. We gratefully acknowledge the funding received from the University of Siena and the Netherlands Organization for Scientific Research (grant nr. 276-61-002) for the organization and hosting of the conference and for the production of this book. A special word of thanks, finally, should go to the University of Siena students who helped organize it all. The Editors

Introduction Martijn van Leusen, Giovanna Pizziolo, Lucia Sarti

Background

how to promote the preservation of archaeological resources that are not yet known (proven by evidence) to exist. Pizziolo, joining Sarti in the late 1990s as a landscape archaeologist compiling all the available geographical and archaeological evidence in a GIS, piloted the use of remote sensing and historical cartographic sources for palaeogeographical reconstructions, which greatly helped the threedimensional understanding of the archaeological record and was instrumental in shifting attention to the surrounding hills and mountains, whose history and exploitation are closely linked to that of the sedimentary basin (Sarti 2005; Martini, Pizziolo & Sarti 2007). Van Leusen, doing research into ‘predictive’ modeling first at the University of Birmingham Field Archaeology Unit in the UK and later at the University of Groningen in the Netherlands, had been continually faced with the strong systematic biases present in regional and national archaeological records (Van Leusen 1996, 2001, 2002). At first, these were focused on visibility and accessibility biases caused by differences in land use/land cover, but it soon became evident that similar biases were operating as a consequence of geological processes, the history of 19th and, especially, 20th century land use, and the research interests and disinterests of students and researchers of archaeology themselves. Clearly, then, it was the ‘push’ given to large-scale spatial archaeological analysis by the general adoption of GIS from the early 1990s onwards that forced us to pay much more attention to gaps and biases in our data, and to study formation processes of the archaeological record.

The editors of this volume first met back in December 1995, at an international summer school on Information Systems in Archaeology, organized by the University of Siena and the Italian National Research council CNR in a monastery at Pontignano (Siena). Geographical Information Systems had only recently begun to be applied to landscape archaeology and to the mapping and analysis of the archaeological record at regional scales, but already the participants were aware of certain persistent methodological problems. Using GIS, we were able to put together large volumes of archaeological and environmental data from many different sources, and to combine and visualize these data in ways different from the traditional cartographic approach (typified in Italy by the sheets of the Forma Italiae archaeological map series). Looking for spatial patterns, and explanations for those patterns, researchers discovered that their data were marred by significant and systematic biases. In particular, we and others became aware that the data sets derived from the regional archaeological field surveys en vogue from the early 1980s onwards could not be taken at face value (Van Leusen 1996; Pizziolo 1997). Pizziolo and Sarti, working on the prehistory of the Florentine part of the Arno sedimentary basin, were faced by the twin problems of sedimentation and urban expansion, which rendered large sections of the archaeological record inaccessible. Since 1982 without interruption, Sarti has collaborated with the Superintendency of the Archaeological Heritage of Tuscany to study the stratigraphical nature of the prehistoric archaeological record in the Florentine Plain. She was one of the first prehistorians in Italy to agree with local city councils to investigate deeply buried landscapes through test pits and archaeological rescue excavations. The resulting synthesis of the Mesolithic to protohistoric peopling process, stressing man-environment relationships (Martini & Sarti 1993; Martini, Poggesi & Sarti 1999), constituted a new vision of the history of the plain, because until then nothing older than Etruscan had been known. This highlights a particular heritage management aspect of the problem of hidden landscapes: namely,

The Hidden Landscapes conference In 2005 Van Leusen received funding from the Netherlands Organization for Scientific Research to organize an international conference on the theme of ‘Hidden Landscapes’, as part of a wider methodological study of the biases introduced into regional archaeological records by both geological and anthropogenic causes. With their long-standing interest and tradition of pluridisciplinary research in landscape archaeology, the Doctoral school and Department of Archaeology at the University of Siena xi

Netherlands), we thought it fitting to invite him to open the conference by revisiting the 1999 ‘Hidden Landscapes’ debate then raging in the pages of the Journal of Mediterranean Archaeology. We include Bintliff’s keynote paper here as part of the introduction to this volume.

have a strong interest in man-environment interaction and in the historical approach to culture studies, and Van Leusen therefore found Pizziolo and Sarti enthusiastically prepared to help organize and host the conference. The Hidden Landscapes conference filled an immediate need. Within Italy, no conferences had yet been organized to discuss biases relating to pre- and protohistoric landscape analysis (Pizziolo & Sarti 2005), partly because ‘landscape archaeology of the pre-roman period’ was not yet a mainstream field, partly because it required the crossing of several interdisciplinary boundaries. Thus the conference presented a welcome opportunity to break the traditional research mould: to focus on methodological problems relating to the investigation of the Italian landscape in pre- and protohistory in the context of other countries around the Mediterranean, and to do this in a fruitful context of multidisciplinary interaction between students of the physical landscape, archaeologists and heritage managers. The conference further promised to bring together researchers and heritage managers who had been confronted by the same issues and problems in their different regions or countries, to make visible the underlying themes. For example, large-scale geological processes link the local landscape histories being studied in mountain areas with those in basins, river valleys and coastal plains of the Mediterranean. Likewise, problems of visibility are encountered by every field archaeologist, but are rarely if ever the main focus of study or discussion; the Hidden Landscapes conference promised to look at the problem of visibility and other systematic biases at various spatial scales through the lens of case studies from a wide range of Mediterranean regions. The aim of the conference was to discuss the current imbalance in landscape archaeological research, especially in the Mediterranean and especially for the pre- and protohistoric periods, in favor of the relatively easily accessible coastal plains and valley floors. ‘Hidden’ landscapes, such as mountainous areas and sedimentation basins, remain underinvestigated for all periods and therefore play at best a marginal role in current archaeological narratives as well as in cultural resource management. In other words, the conference was to assess the archaeological record that we researchers and heritage managers have created in the past, are now creating, and will be creating in the future. As the title of the conference and this volume refers to a phrase coined by Prof. John Bintliff (now at Leiden University in the

This volume This collection of papers presents a wide range of case studies and methodological reflections on the problem of Mediterranean hidden landscapes, with important consequences not just for the study of the archaeological record, but also for its management. Though focusing, in terms of the number of papers, on Italy, we want to emphasize that common underlying themes connect these case studies from distant parts of countries bordering the northern Mediterranean, using very different disciplinary approaches, and operating at widely different spatial scales. Awareness of these underlying similarities will, we hope, help to define and diffuse a large(r) set of tools with which to study, describe and manage the diversity of Mediterranean landscapes. This volume has accordingly been divided into four parts: Methods and methodology, Mountains and uplands, Sedimentary basins, and Distribution Patterns. On the theme of methods and methodology, the first issue being addressed is that of the presence of systematic biases in regional archaeological data sets. Caused by a variety of natural and anthropogenic processes acting in and on the landscape, as well as by the selective abilities and interests of the researchers themselves, these biases cause problems in assessing the reliability of the current archaeological record, and therefore in deciding on appropriate management policies. On the largest spatial scale, it means that our knowledge and understanding of mountain and upland environments is very much inferior to that of the main basins, valleys and coastal plains beloved of classical archaeologists. In the context of modern rescue archaeology, it means that we must be concerned with the question of how we should use the available historical, remotely sensed, and excavated sources of information. Whereas pre- and protohistorians have traditionally understood the need to study the past landscapes, and have therefore worked closely with earth and environmental scientists, a second issue that emerges is the need for a much more thorough geoarchaeological understanding of landscape history. This will also help us deal with issues of spatial and xii

cause the conditions of preservation and modern disturbance tend to be better than in the lowlands, they can provide a high-quality ‘window’ into large-scale change in past societies; and finally, • they offer an environment conducive to a totally different conceptions of hiddenness: that of caves and other remote, secluded (on purpose) and often sacred places. More prosaically, as the papers in this volume demonstrate once more, there is a dearth of what one might call ‘erosion studies’ in mountain and upland areas, to complement the ‘sedimentation studies’ already taking place in basins and lowlands. The geological and archeological stratigraphies which are so important for our understanding of landscape history in sedimentary environments must perforce be complemented by equally important but erosive landscape histories in the mountains and uplands. How can, and should, these be studied? Beyond some finger-pointing to general causes (‘deforestation’, ‘overgrazing’) the implications for the archaeological record in the mountains have barely been studied.

chronological scale. Since the environment is structured, and the archaeological record patterned, at a range of spatial scales, we must learn to integrate (for example) the large-scale geological processes and generic vegetation types occurring in a given region with the purely local reasons why a particular location was selected for settlement in the past and its remains were in fact detected by us so many centuries later. Similarly, both environmental and cultural change happens over a range of timescales, and we therefore need to make sure that our methods are able to detect change across this range. Both of the above issues combine to produce the problem of ‘offsite’ archaeology — effectively, those parts of the archaeological record that cannot be interpreted as ‘sites’ but which archaeologists today no longer think are ‘meaningless background noise’. Human activities that we know must have occurred in the landscape but which leave behind no direct archaeological evidence tend to be neglected in publications and analyses, and so become ‘hidden’ or even totally ‘lost’. The earth and environmental sciences can clearly help out the archaeologist here, but much potential for collaboration remains unexplored.

For sedimentary basins, the papers in this volume speak of the large impact that post-WWII urban expansion and agricultural development has had. Urban expansion and infrastructural works have, on the one hand, impeded our ability to make archaeological observations and have, on the other hand, led (in recent decades at least) to the discovery of increasing amounts of archaeological remains. To the extent that deep ploughing in sedimentary basins touches the more recently and less deeply buried archaeological landscapes, it produced a bonanza of surface archaeological remains in the 1960s through early 1980s — damaging or destroying the archaeological record while making it available for recording by archaeologists. Nowadays, the best preserved remains will be those buried below the typical depth of human interventions, especially if this is also below the groundwater table; at the same time, these remains will be the least easily accessible for scientific study, and will become blank spots on the archaeological map. It is therefore of fundamental importance to reconstruct the broad palaeo-environmental framework characterising these sedimentary basins, and the collaboration with geologists and other natural scientists is crucial to identify, for example, the features of the Last Glacial Maximum, the various key episodes of

Mediterranean mountains and uplands today tend to be zones of low accessibility and visibility, but also, outside the main valleys, zones of relatively low recent human impact. They are seen as marginal, unimportant areas from the point of view of large-scale agricultural production, especially it has to be said by classical archaeologists who are still very much influenced by the classical paradigm of civilized, urban, life in the coastal plains and large river valleys. Mountains deserve much more attention from us for very good reasons: • they stimulate rather than prevent cultural contacts, by providing trade routes and shared or contested resources such as summer grazing and raw materials for mining; • for long periods of archaeological time they provide the context for the pastoralist or mixed subsistence economy (see, for example, the international round table on the archaeology of pastoralism in southern Europe; Maggi et al. 19901991). ‘Vertical’ transhumance also provides the link between lowland and upland archaeologies, which are not often studied together; • as a marginal environment they are sensitive to the Braudelian ‘conjunctural’ economic and demographic expansions and contractions, and bexiii

sedimentation, and any changes in the hydrological setting (rivers and coastlines) of the study area.

Pizziolo G. 1997 [published 2000]. L’uso dei GIS per la valutazione dell’attendibilità dei dati archeologici ai fini di un’analisi distributiva. Bias analysis nella valle del Biferno, in Panzeri M. & Gastaldo G. (eds), Sistemi Informativi Geografici e Beni Culturali, Torino, pp. 101-107. Pizziolo G. 1999. Geographic Information System e patrimonio archeologico di superficie: comprendere la trama del contesto insediativo del passato, in Lenzi F. (ed.), Archeologia e Ambiente, Atti del Convegno Internazionale, Documenti/30, Forlì, pp. 261-266. Pizziolo G. & Sarti L. 2005. Landscape archaeology in Sesto Fiorentino: a GIS analysis for investigating settlement strategies in wetland area, in Berger J.F., Bertoncello F., Braemer F., Davtian G. & Gazenbeek M. (eds), Temps et espaces de l’homme en société, analyses et modèles spatiaux en archéologie, XXV Rencontres Internationales D’archeologie et D’histoire D’Antibes, Editions APDCA, Antibes, pp. 443-452. Sarti L. & Martini F. 1993. Costruire la memoria: archeologia preistorica a Sesto Fiorentino, 1982-1992, Garlatti e Razzai, Montelupo Fiorentino. Sarti L., in press. Archeologia nel Valdarno e aree limitrofe, in Atti 2° Seminario di Studi “Città di Montefiascone”, 28 maggio 2005 Montefiascone, Viterbo. Van Leusen P.M. 1996. Unbiasing the Archaeological Record, in Archeologia e Calcolatori 7, pp. 129-136. Van Leusen P.M. 2001. Marginal Landscapes: Survey and interpretation biases in low fi nds density regions in Italy, in Darvill T. & Gojda M. (eds), One Land, Many Landscapes (BAR International Series 987), pp. 71-73. Oxford, BAR Publishing. Van Leusen P.M. 2002. Pattern to Process: Methodological investigations into the formation and interpretation of spatial patterns in archaeological landscapes, PhD thesis, University of Groningen, http://www.ub.rug.nl/eldoc/dis/arts/p.m.van.leusen.

A final section of the volume looks again at regional archaeological distribution patterns, showing how important it is to identify and in a self-critical manner explain density variations in finds and sites maps at all spatial scales, up to and including ‘voids’. As the saying goes, the absence of evidence does not imply evidence of absence. We need robust ways of studying landscape history and its varying effects on the preservation and visibility of the archaeological record.

Conclusion The slow process of implementation of the so-called ‘Malta agreement’ (the 1992 EU Council of Ministers agreement on the preservation of the European cultural heritage) in the different countries of Europe has in 2005 led to the adoption of a new Italian law on preventive archaeology, and in 2009 national criteria were set to select assessment experts. Similar laws and procedures have been adopted in other countries fringing the northern Mediterranean. This means that the study of the landscape and context of the archaeological remains is now finally included in the formal process of assessment of archaeological risk and potential in areas transformed by large urban or infrastructural interventions. Clearly, in order to preserve the archaeological heritage particular attention can and must now be paid to the problems and peculiarities of the ‘hidden’ pre- and protohistoric evidence. Our first goal should be to identify and if necessary develop methods and tools adequate for the different environmental contexts; and secondly we should probably adapt the university curriculum to include these competences for future archaeologists and heritage managers, widening their awareness of both the problems and the potential of ‘hidden’ landscapes.

References Maggi R., Nisbet R. & Barker G. (eds) 1990-1991. Atti della Tavola rotonda internazionale della pastorizia nell’Europa meridionale, Chiavari 22-24 september 1989 (2 vols). Bordighera, Istituto internazionale di studi Liguri. Martini F., Poggesi G. & Sarti L. 1999. Lunga memoria della piana. Catalogo della mostra maggio ottobre 1999, Sesto fiorentino, Firenze. Martini F., Pizziolo G. & Sarti L. 2007. Paesaggi nell’antichità: uomo e ambiente in area fiorentina nella preistoria, in Conti S., Scardigli B.& Torchio M.C. (eds), Geografia e viaggi nel mondo antico, Affinità Elettive, Ancona pp. 123-147.

xiv

Problems of chronology and function in survey assemblages: the 1999 Hidden Landscape debate reviewed John Bintliff Faculty of Archaeology, Leiden University, PO box 9515, 2300 RA Leiden, The Netherlands – [email protected]

The Siena Hidden Landscapes conference was focussed on landscapes which are difficult to survey, in contrast to the lowland open agricultural fields typical for older generations of Mediterranean survey. The conference title actually derives from an article (Bintliff, Howard et al. 1999) in which we dealt with prehistoric finds from a standard cultivated lowland landscape in Central Greece. However the paper itself raised general issues to do with our ability to discover, map and interpret certain periods of human occupation which are relevant to all field surveys, in whatever terrain they operate. I would like here to review the subsequent and ongoing debate which has grown out of our 1999 paper and its challenge for prehistoric landscape studies, and then to move on to consider other dimensions of hidden landscapes which relate to more typical surface finds from historic periods.

Hidden prehistoric landscapes of Boeotia and in the wider Aegean

scatter of Late Bronze Age sherds. As the site was merely of farm size throughout its existence, we had to conclude that these Mycenaean finds probably marked the vestigial traces of a much earlier settlement of similar scale, later reduced to a mere fraction of the historic assemblage. Would we have found this Late Bronze Age site without the team gridding the much more obvious Classical high density focus? We began to doubt that this would have happened unless the surrounding landscape had been totally free from offsite pottery finds. The area — the Valley of the Muses — had in fact a high general spread of offsite ceramics. Dramatic confirmation of how well a longoccupied agricultural heartland could ‘hide’ lowdensity prehistoric sites came from our survey in the immediate hinterland of the ancient city of Thespiae. Some 17 sites of historic age, mainly Greek and Roman, were recorded, but one, LSE2, was a special case. All Boeotian cities are surrounded by dense offsite ceramics; in this sector alone we counted the equivalent of almost 1.5 million pieces. Still the surface density of historic rural settlement sites generally created visible local ‘peaks’ for field teams to spot. But as one approaches the City the offsite density soars to confusing levels, and the field team could not make up their minds if one locality here, LSE2, was a site or not. It was duly gridded, and no focus emerged, just a uniform spread of Classical sherds; in density terms, LSE2 is not distinct from all other fields at this distance from the City. When we sampled the periods that made up the massive offsite finds around the ancient city through sample dating, surprisingly only one period accounts for 70-80% of the offsite — Classical Greek. Note

In our survey region of Boeotia, we were long aware that the increase in new prehistoric sites as a result of intensive survey was well below that for historic periods. Given the much longer timescale for the Neolithic to Early Iron Age — some 6000 years compared to a total of 2600 for all historic phases combined — we suspected that some bias was operating against our recovery of evidence for farming prehistory. Moreover, John Cherry had already argued (1979) that although survey could find phases with many small farm sites such as the Early Bronze Age, most of these were in use for a couple of generations at the most. Thus any site distribution maps were illusory because few sites were likely to be in contemporary use. In 1985 I had tentatively suggested to explain this prehistoric site deficit in terms of loss: a model of progressive degradation of surface finds and whole surface sites with time, due to cultivation and natural weathering eroding or burying findspots. A second factor that has perhaps been restricting knowledge of prehistoric sites was the common occurrence in the Mediterranean of multi-period sites, where the older periods were likely to have their surface finds obscured by a heavier overlay of later periods. One example published in 2000 may look familiar to many of our readers (Bintliff et al. 2000): VM2, a farm site of only a fraction of one hectare in size, appeared as a well-defined density focus after gridded survey. Although the diagnostic collection was small, a Classical farm proved to be the main cause for the localised high finds density. But the site collection also turned up a much smaller xv

2. Some claim (e.g. Cavanagh 2004) that they found all the sites from prehistoric times. My response: This may be true of large nucleated sites, when at least a part is likely to be exposed with variable vegetation or land use, but is not at all likely for small sites. To my knowledge, the few intensive surveys which practice resurvey have invariably found that from season to season small sites appear and then disappear, or to quote Graeme Barker “go on and off like traffic lights” (Barker & Symonds 1984). And he was referring to hard Roman ceramics in Italy. 3. The discovery of prehistoric finds through careful work at more easily recognisable historic sites only occurs, according to Davis (2004), because the historic site provides a window into older use of the same location. My response: We believe that vestigial sites are widely scattered over the landscape; it is when we look more closely, through gridding later sites, that we see this wider settlement pattern. The fact that, when studied closely, around 40% of survey sites in the Nemea Valley Survey include prehistoric finds seems far too great a coincidence to me. 4. More originally, Pettegrew (2001; see reaction by Osborne & Foxhall 2001 and Bintliff et al. 2002) has claimed that well-defined Greco-Roman farm sites only belonged to the rich, those belonging to the poor being the main cause of the extensive offsite carpets of severely eroded sherds we found in Boeotia. My response: If offsite carpets are farms we would have no farmland left, as the scatters dominate the whole landscape for some 20 sq.km around the City. And we have plenty of well-defined ancient farms with clearly impoverished assemblages of pots and tiles to represent peasants. The implication of the Hidden Landscape Theory for Farming Prehistory in the Mediterranean, up to the inception of Iron Age complex societies, is that in many regions and many periods we should look into the fine details of intensive survey results to see if the recorded landscape is missing a whole dimension of settlement below the town and village level. The potential of such an approach can be seen for example in the Methana Survey, where Chris Mee (Mee & Forbes 1997) is professional enough to show us for the Early Bronze Age that the rural sites with more than 5 pieces present that he is happy to treat as permanent settlements, are accompanied by at least as many locations where fewer pieces mark some kind of presence. We believe many of these others are vestigial settlements.

that the Prehistoric period is extremely low as offsite, so its occurrence takes on much more significance. Effectively LSE2 was not a Classical site, since its Classical finds density was normal for the offsite emanating (we argue) from urban manuring. But the gridded collection revealed something entirely unexpected — a tiny prehistoric focus of finds. Given the very rare occurrence of offsite prehistoric, and given our theory that small rural farms could degrade to small scatters of finds as at site VM2, we argued that LSE2 was a vestigial Early Bronze Age farm site of no great occupational duration. On this basis, we looked again carefully at all the other Greek and Roman rural sites in the district, and our gridded surveys revealed a remarkable fact — almost every one, though only located due to their high levels of historic sherds, gave up small numbers of prehistoric sherds and/or lithics. These were mainly Early Bronze Age, a period when most local pottery was coarse and does not survive well in the ploughsoil. In 2004 we observed similar coarseware near ancient Tanagra city (although here of Archaic date), disintegrating into soil after deep ploughing in a newly-disturbed Classical cemetery site. After careful study of these small prehistoric collections found unintentionally by gridding later sites, and also of sporadic prehistoric finds noted in fieldwalking between sites, we published our controversial theory that these small assemblages were not casual discard but genuine activity foci and, in a minority of cases, former small rural farm sites eroded into vestigial scatters. Lying between easily identifiable prehistoric villages several kilometres apart, we were revealing what we termed a ‘hidden prehistoric landscape’. It was through comparison with very similar evidence and interpretations in survey data provided by colleagues from Prague that I became convinced that we were onto something important for Mediterranean landscape archaeology. Our 1999 paper has given rise to a series of responses and offshoots (cf. for example Barker, Mee et al. 2000), and I can here briefly summarize key points: 1. Maybe Boeotia is a special case; other regions have lots of prehistoric sites. The Kythera Survey for example (Bevan 2002) has a high density of well-preserved Bronze Age farm sites. My response: The susceptibility of ceramics to degradation is clearly a significant variable. Minoan ceramics on Kythera are very finely-made compared to the poorer coarsewares common in much of the Mainland Helladic cultures. xvi

The ways we hide landscapes from ourselves I have become increasingly convinced that the lack of clear experimental data about the effects of varying survey practices, and hence the poverty of guidelines, means that we are often creating major biases in what we see, where it is, and how much of it there really is. I shall here present some preliminary results on the possible effects of different methods of site survey on the data we recover, and on the interpretations we infer from them. I want to suggest that surveyors themselves often hide particular period landscapes from sight, through what are actually widely-favoured methodologies. Let me begin with a common survey catalogue item, site 64 from the Kea Project (Cherry & Davis 1991). A large site, some 2 ha, but not a large collection of dated pieces brought back. The decision to limit occupation to two of the periods present leaves us wondering what the almost equally well-represented sherds of other periods were doing there. This leads to the following questions: 1. How many sherds make up a ‘representative sample’ from a surface site? I think no-one knows or has given us experimental guidelines yet… 2. How do we explain variable numbers of finds from different periods on the same site? Do we make an arbitrary cut-off and say: below this number it is casual visits, temporary use, or just not significant? Most surveys do not have a formal methodology for the less common periods… 3. Does the survey method used onsite make any difference to the picture we obtain for the site’s history of use? I think no-one knows or at least has published experimental guidelines yet… Let me cite another example, rather extreme but the results are widely reproduced: the urban survey made for the city of Keressos, also from the Kea Project (Cherry & Davis 1991). Based on a total of some 100 sherds collected across this 30 ha site, the Kea Project gave a summary occupational history of the town. The team admitted that this was perhaps dangerously small for a sample, but who is to say it is not sufficient? Is the picture derived from the sample reliable? We need to test it out, and that is exactly what we have now done with the much larger database from the survey of the city of Thespiae. To give the Kea small sample approach a fair chance we took some 30 ha out of Thespiae (a 100 ha town in total), and resampled this for 267 sherds. Our original total for these 30 ha was around 2000, so about 7 times more. What difference does sample

size make? Where material from a given period is widespread but dispersed, the small sample makes it appear totally irrelevant, as with Early Bronze and generic Prehistoric finds. In the large sample these are revealed as a series of short-lived farms over the whole site. In the case of the Classical Greek city, it is hard to find the shape of the town from the subsample compared to the large sample. Where the site has a dense clear focus, however, as with the Early Roman town, small samples perform well. Small samples perform dramatically differently depending on the properties of the occupational period being sampled. They are not reliable for mapping a long-term sequence of occupations. An unintentional experiment makes the case even clearer. By mistake five Thespiae survey units were surveyed twice. The first year the students were not so hard-working and made small sherd collections, but the following year the students collected at least three times as much from the same units. When I graphed the results according to the periods present, I was genuinely shocked! One was struck how consistently the prehistoric, protohistoric and medieval eras suddenly appeared from nowhere as the sample size multiplies from the small collections to the larger: more hidden landscapes! It is too early to give definite sample size guidelines here, and I would love to have more data from other surveyors to enlarge the experiment, but the following guesstimates are a first attempt: if our ‘poor year’ level of 5-10 sherds per unit is hiding key periods, then based on our 594 grid-units, even a total of 6000 sherds for the city is not reliable for several phases of the city’s history. The sample of 30 per unit performs much better, but would mean that we need 18,000 sherds for the entire city. Luckily we are not too far off with our actual sample of 14,500 sherds, reached more by accident than planning apart from our belief, 20 years ago, that ‘a lot would be safer than a few’ sherds. More recent reanalysis using subsampling argues for Thespiae that, on a large multiperiod urban settlement, one might need some 300 sherds for each occupational period across the whole site to represent the size and location of contemporary activity with some confidence (and Thespiae has 27 mapped phases).

The significance of common and rare periods in a site collection We saw earlier that small samples, like those on the Kea Project, often give us sites with some clear occupation phases, plus small amounts of several

xvii

other periods. How could we deal with this? It may be possible to filter out genuine casual discard from different levels of genuine focussed activity, but we need elaborate recording systems to achieve this (Bintliff & Howard 1999). And we have already seen that only large numbers will give us a chance to see the real underlying patterns. At our site LSE1 in the Thespiae hinterland the sample size is reasonably large, and we saw some mysterious contrasts in period presence. Just to make the analysis more complicated, our site lies close to the city of Thespiae and clearly has massive interference from the urban discard halo. We therefore first need to clarify how much pot could have been introduced locally by urban offsite onto the site, rather than signifying localised site use. The technique we have developed for this we call Residual Analysis: it compares site totals with the general level of offsite finds for each period throughout the district. The use of the Residual Analysis and related methods of rural survey analysis is exposed in Bintliff, Howard & Snodgrass 2007. At LSE1, despite the massive Classical Greek offsite, our location remains even for this period an abnormal high density focus. The Roman finds too are well above local expectations, since Roman material is rare in the offsite. As for Late Roman, also rare in offsite, the site finds numbers are not very elevated but something is happening here. Yet how do we account for the gross differences between the period frequencies? Here the necessity of full gridding becomes clear: we can argue for a large farm in Classical times, then a much smaller Roman farm, followed by at most a shed or temporary field base for Late Roman. A combination of elaborate methods appears to allow us to attempt such a ‘cultural biography’ of surface sites.

the experiment shows that, compared with the large area samples and all our historical source evidence, small area samples give a generally poor and unrepresentative picture of the overall occupation in most periods. For example, the Hellenistic City appears far too small, whilst one of our two medieval villages ceases to exist as a focus altogether.

Does size matter?

References

Surveyors use a number of different shapes and sizes of site sample for collecting density data and sherd collections onsite. What difference does this make? Who knows? Well, we recently did some experiments on the Thespiae City data. For several of our Boeotian cities we had collected sherds at two spatial sample levels: large units of around 3000 sq.m and smaller ones of around 300 sq.m. Similar finds numbers were brought back in both types of samples, but the coverage was either small and isolated, or large and total. Only this year did we begin to compare the finds from the two sample series for the whole Thespiae city area, and this time

Barker G. & Symonds J. 1984. The Montarrenti Survey, 198283. Archeologia Medievale, pp. 278-289. Barker G., Mee C. et al. 2000. Responses to “The Hidden Landscape of Prehistoric Greece”. Journal of Mediterranean Archaeology 13, pp. 100-123. Bevan A. 2002. The rural landscape of Neopalatial Kythera: A GIS perspective. Journal of Mediterranean Archaeology 15, pp. 217-255. Bintliff J.L. 1985. IV Greece: The Boeotia Survey. In: S. Macready and F.H. Thompson, Archaeological Field Survey in Britain and Abroad. London, Society of Antiquaries, pp. 196-216. Bintliff J. & Howard P. 1999. Studying needles in haystacks – Surface survey and the rural landscape of Central Greece in Roman times. Pharos 7, pp. 51-91.

Conclusions • Prehistoric and Protohistoric small sites are often missed altogether or misinterpreted, even by high-intensity survey. Elaborate methods of total landscape recording and filtering procedures are required to find and evaluate them. • But even historic periods with plentiful, wellmade pottery, can become ‘hidden’ through negative filtering by many of the survey methods in common use. • How many sherds make up a representative sample? We need more experimental data, but a guesstimate for a large multi-period site is 300 for each period, and thus for a small to large town site 6000-18,000 sherds in total. • How can we deal with sites with multi-period use and with highly variable sherd numbers from one period to another? Quantified and dated offsite statistics are required to control offsite interference; as full a gridded survey as possible, associated with a large sample and additional tools (for example geophysics and soil analysis), can bring us close to writing a realistic ‘cultural biography’ of such a surface site. • Does it matter if we survey sites by discrete sample units or total gridded collection? It seems that only total cover has a chance of recovering a reasonably reliable long-term site history.

xviii

Bintliff J.L., Howard P. et al. 1999. The hidden landscape of prehistoric Greece. Journal of Mediterranean Archaeology 12.2, pp. 139-168. Bintliff J. et al. 2000. Deconstructing “The Sense of Place”? Settlement systems, field survey, and the historic record: A case-study from Central Greece. Proceedings of the Prehistoric Society 66, pp. 123-149. Bintliff J. et al. 2002. Classical farms, hidden prehistoric landscapes and Greek rural survey: a response and an update. Journal of Mediterranean Archaeology 15, pp. 259-265. Bintliff J., Howard P. et al. (eds) 2007. Testing the Hinterland. The work of the Boeotia Survey (1989-1991) in the southern approaches of the city of Thespiai. McDonald Institute Monographs. Cambridge, McDonald Institute, Cambridge University. Cavanagh W. 2004. WYSIWYG: Settlement and territoriality in Southern Greece during the Early and Middle Neolithic periods. Journal of Mediterranean Archaeology 17(2), pp. 165-189. Cherry J.F. 1979. Four problems in Cycladic prehistory. Papers in Cycladic Prehistory. J. Davis and J.F. Cherry. Los Angeles, University of California, pp. 22-47. Cherry J.F., Davis J.C. et al. (eds) 1991. Landscape Archaeology as Long-Term History. Los Angeles, Institute of Archaeology, University of California Los Angeles. Davis J.L. 2004. Are the landscapes of Greek Prehistory hidden? A comparative approach. In S.E. Alcock and J.F. Cherry (eds), Side-by-Side Survey. Comparative Regional Studies in the Mediterranean World. Oxford, Oxbow Books, pp. 22-35. Mee C. & Forbes H. (eds) 1997. A Rough and Rocky Place. The Landscape and Settlement History of the Methana Peninsula, Greece. Liverpool, Liverpool University Press. Osborne R. & Foxhall L. 2001. Response to David K. Pettegrew, “Chasing the Classical Farmstead”. Journal of Mediterranean Archaeology 14, pp. 212-222. Pettegrew D.K. 2001. Chasing the family farmstead: assessing the formation and signature of rural settlement in Greek landscape archaeology. Journal of Mediterranean Archaeology 14, pp. 189-209.

xix

Hidden Landscapes:

METHODS AND METHODOLOGY

1 Making sense of a hidden Sicilian landscape: geoarchaeological and GIS investigations of field-survey data Matthew FitzjohnI, Gianna AyalaII I School of Archaeology, Classics and Egyptology, The University of Liverpool, Hartley Building, Brownlow Street, Liverpool L69 3GS, United Kingdom – [email protected] II Department of Archaeology, University of Sheffield, Northgate House, West Street, Sheffield S1 4ET, United Kingdom – [email protected] Abstract Using material from a survey and excavation project in a hidden landscape of central Sicily, this paper discusses the complexities of recovering and understanding the archaeological record in Mediterranean landscapes. Although it is often the case that archaeologists construct landscape narratives of Mediterranean history from field-survey data, the authors suggest that to understand this material it is necessary to identify and model how post-depositional processes affect the visibility and preservation of the archaeological record of different periods across a landscape.

1. Introduction: Sicily’s hidden landscapes

The lack of interest in the interior of the island was furthered by the apparent difficulties of travelling. Travellers emphasised the extreme hardships that they had to endure even when travelling along the coastal roads, which were seen as highly treacherous due to the poor quality of the surface and the threat of bandits (Brydone 1773). As a consequence of these hardships, travellers presented the interior of the island as both inaccessible and uninteresting: “… the historian [and archaeologist] is not unlike the traveller. He tends to linger over the plain, which is the setting for the leading actors of the day, and does not seem eager to approach the high mountains nearby.” (Braudel 1972, p. 29). Until recently, landscape archaeology has largely been confined to the relatively easily accessible coastal zones and valley floors rather than to the mountains and upland valleys of the Mediterranean, for a variety of historical and cultural reasons. These under-examined regions are often seen as having played a marginal role in archaeological narratives of long-term Sicilian and Mediterranean history. This notion of Sicily as being composed of two distinct cultural zones, a cultured coastal zone and an uninteresting mountainous interior periphery, has contributed towards the establishment of particular perceptions of the island that are evident in archaeologists’ interpretations of the past. For example, explanations of the Sicilian Iron Age and Archaic periods have been dominated by the idea of a physical separation of colonial and indigenous landscapes in contrasting regions of the island. This concept has been further supported by the ‘core-periphery’ model, which has dominated recent studies of the Iron Age. The main tenet of

Sicily has historically been perceived in terms of a duality between the coast and the mountainous interior. The coast has often been seen as both a physically and culturally fertile zone with contacts with the wider Mediterranean world; in contrast, the interior of the island has been regarded as a less productive and isolated region. Consequently, Sicily has often been presented as an island with a culturally superior coastal region and a culturally inferior, more peripheral, interior. This perception of Sicily has been nurtured since the first Northern European travellers started to record their experiences on the island. In the second half of the eighteenth century the island became an increasingly popular destination for foreign visitors, and it often marked the climax of an extended Grand Tour. Educated ‘gentlemen’ and aristocrats including Riedesel (1773), Brydone (1773), Houel (1780), and Goethe (1778) journeyed to Sicily to see the wonders of Mount Etna and the picturesque classical sites that included Taormina, Syracuse and Agrigento (Blanchard 1998; Leighton 1999; Mozzillo 1993). These travellers often recorded their experiences in letters and paintings that were presented as mimetic representations of the island: attention was directed towards the accurate recording and description of temples and cities that were often discussed with reference to Greek and Roman myths (Riedesel 1773; Brydone 1773; Russell 1815). In contrast, regions of the island that were not deemed to have been important in antiquity were left unvisited. There was a prioritisation of sites to visit and investigate, ultimately resulting in the valuation of particular areas of the island to the detriment of others.

3

Hidden Landscapes

that have often barely experienced archaeological analysis. The Troina Project joins only a very small handful of comparably intensive regional archaeological projects conducted on the island (Alliata et al. 1988; Johns 1985, pp. 215-223; 1992, pp. 407-420; Kolb 2007, pp. 171-185; Wilson 1981, pp. 249-260; Wilson & Leonard 1980, pp. 219239) and is one of only five in eastern and central Sicily (Albanese Procelli et al. 2007, pp. 35-48; Leone et al. 2007, pp. 49-58; Giannitrapani & Pluciennik 1998, pp. 59-69; Pluciennik et al. 2004, pp. 27-65; Thompson 1999). The material that will be discussed in this paper was collected during a programme of intensive, systematic archaeological survey and excavation carried out within an area centred on the town of Troina. The techniques adopted and developed during this project, both within the field and in the computer laboratory, will contribute to the established and often lively debate among Mediterranean prehistorians and Classical archaeologists, concerning appropriate survey methodologies and the potentials and limitations of the resulting data for archaeological and historical interpretation. The idea of hidden landscapes not only corresponds to geographical zones that have been regarded as uninteresting, marginal or difficult to investigate. Certain periods of history and even regions of occupation have also been hidden due to post-depositional processes. A range of geomorphic processes, historic land use practices and modern agricultural development, have radically transformed the archaeological record and have diversely affected the level of preservation as well as visibility of archaeological material of different periods. In this paper we explore the possible contributions of the integration of geoarchaeological analysis and erosion modelling to revealing the partially hidden long-term archaeological record of rural occupation and land use practices in central Sicily. The central purpose of this article is not to offer a preliminary report on the findings of the Troina survey; rather, it is to focus on the importance of the hidden archaeological record and how we may start to reveal some of the problems with our data.

this theory as applied to archaeology is that the classical world acted as a dynamic and dominant ‘core region’ which made contact with and civilised the peripheral areas of Europe (Champion 1989, p. 2; Stoddart 1989, p. 88; Whitehouse & Wilkins 1989, pp. 102-126). The coastal ‘core’ and interior ‘periphery’ of Sicily have been regarded in this sense as a microcosm of the wider European region during the Iron Age. Perhaps the most important development in Mediterranean archaeology in the last twenty years has been the acknowledgement that mountains are a valid area of research. In the light of Braudel’s (1972) thesis on the Mediterranean world in the sixteenth century, an increasing number of archaeologists and historians have become interested in understanding the importance of the changing relationship between the uplands and the lowlands (Malone & Stoddart 1994, pp. 1-208; Barker 1995). Although the majority of historical studies have been written about the coasts of Sicily, the mountains and hills dominate the landscape of the island. Over sixty percent of the island is hilly and a little over twenty-five percent mountainous, and only fourteen percent is flat ground. Thankfully, archaeological interest in so-called hidden landscapes is spreading (Kolb 2007, pp. 171-185; Leone et al. 2007, pp. 4958). The Troina Project is one of a handful of recent archaeological landscape projects that have attempted to redress this imbalance of archaeological research; it has focused its attention away from the coast and to an upland area of the Nebrodi Mountains. The body of data presented in this paper provides our first attempt to redress the imbalance in perspectives of Sicily’s hidden uplands. The fieldwork on which this paper is based, forms part of the Troina Project, an interdisciplinary research project directed by Dr Malone and Dr Stoddart (Malone & Stoddart 2000a and b; Malone et al. 2003). Despite the long history of Classical archaeological and historical research on the town of Troina, this project stands out as the first sustained effort to carry archaeological enquiry beyond the mountain town, and the first to provide a detailed view of the long-term development of this part of the Nebrodi Mountains (figure 1). In Sicily more generally, archaeologists’ attention has centred on necropoli, settlements and the structures within them; in comparison, the systematic study of settlement distributions has tended to be ignored (Wilson 1990, p. 313). Archaeological field survey offers the possibility of insights into rural settlement history within regions

2. A hidden landscape of Central Sicily The Troina Project has aimed to elucidate multiperiod landscape histories within an upland region of central Sicily. The project has taken place in a region around Troina, a small mountain town in the Nebrodi Mountains of the Enna province. 4

Fitzjohn – Ayala, Making sense of a hidden Sicilian landscape…

Troina occupies a position on the Monte di Troina, nestled between the twin peaks of Monte Panteon to the west and Monte Mugana to the east, and at about 1120 m it is the highest inhabited town in Sicily (fig. 1). Despite the known importance of the Nebrodi Mountains and the significance of Troina as attested in historical documents (Giuliano 1901; Idrisi 2001; Wilson 1990), neither have ever been studied systematically. More generally, the mountains of Sicily have tended to remain ‘hidden’, and on those rare occasions when research has been undertaken at higher elevations it has tended to be directed to sites documented in the classical literature or to artefacts from the classical past. Troina itself has been the subject of detailed historical and archaeological research that has documented its long history of occupation and invasion from its existence as a Hellenistic town through to the consequent invasions of the Romans (Militello 1961), Saracens and Normans. Attempts have been made to identify the town from references in the classical literature: local tradition identifies Troina with the Imachera or Imakara of which Cicero spoke, and other studies

have used a passage from Diodorus Siculus who stated that Troina could be identified with Engyon, the site of a famous pre-hellenic temple of Meteres. Troina has, however, not really been placed within a broader dialogue of inter-regional contact and historical political relationships. The first archaeological research at Troina took place between 1958 and 1960 by the superintendent of Syracuse. The project attempted to understand the process of hellenisation in the centre of Sicily, and focused on the excavation of the Hellenistic city walls to the south of the town, together with the extensive Hellenistic and Roman necropolis just outside of the walls on the southern slopes of Monte Mugana (Militello 1961, pp. 362-400). One hundred and forty one burials of diverse funerary type were found, one hundred and thirteen of which were inhumations and twenty-eight were cremations; these date from the end of the 4th century to the first twenty years of the 2nd century BC (Militello 1961, pp. 362-400; Scibona 1980, pp. 377-386; Ragusa 1995, pp. 33-37). Further nearby excavations during the early 1970s revealed the full extent of the necropolis (Scibona 1980). The two main periods of excavations

Figure 1 – View from valley of Troina.

5

Hidden Landscapes

Figure 2 – Map of the Troina region with survey fields indicated.

relatively well-hidden; it had only been revealed in a single thesis (Ragusa 1995) that had documented the rock-cut tombs on the slopes of Monte Troina and Monte Mugana, and recorded sporadic finds of obsidian and ceramics of Finocchito and Licodia Eubea type from the vicinity of the Hellenistic walls. These forms of evidence were interpreted as an indication of an ‘earlier’ pre-Classical settlement on the mountain, possibly extending back to the EBA (Ragusa 1995). None of this earlier research moved beyond the walls and environs of the town and into the wider region. This is surprising, as the importance of Troina has historically lain in its geographical location at a natural point of passage across the steepest part of the southern limits of the Nebrodi mountain range. It is likely that movement from prehistory to World War II has always followed a series of interconnecting river valleys, including the Fiume di Sotto di Troina which flows into the valley of the Simeto River which in turn flows around Etna across the Catania plain and into the Ionian sea. This mountain-river route, which has been documented as early as the 1850’s (Biondi 1991), even now acts as the major road system. It was not until the inception of the Troina survey that any form of detailed analysis of Troina’s hinterland was undertaken. The survey was

in the late 1950s and 1970s, carried out on the south side of the Monte di Troina, revealed evidence for a Hellenistic and Roman occupation on the mountain from the late 4th century BC through to the 3rd century AD. The remains detailed by the excavations included numerous structures, comprising houses, bath houses with mosaic flooring (2nd-1st century BC), Hellenistic walls up to a height of 9.30 m (Militello 1961, pp. 323-326, 329; Scibona 1980, pp. 351-354; Ragusa 1995, pp. 43-45), a Roman road, a large cistern, a 3rd century AD necropolis (Scibona 1980, p. 387) and Hellenistic fortifications. During the Hellenistic and Imperial periods, Troina is believed to have flourished as the result of the proximity of the Fiume di Sotto di Troina which connected the town to Catania in the east, and to the Roman centre of Halasea on the northern coast. Although prehistoric rock-cut tombs were visible on the slopes of Monte Mugana, Troina and Monte Panteon, the excavators claimed that the town did not have a significant prehistoric occupation (Militello 1961, pp. 399-400). The first occupation of the town was believed to have been by hellenised people in the fourth century BC. Interpretation of the settlement has focused upon the origins and identity of these ‘earliest’ inhabitants as possible Campanian mercenaries from Syracuse. Until the inception of the Troina Project, the prehistoric occupation of Troina had remained 6

Fitzjohn – Ayala, Making sense of a hidden Sicilian landscape…

on the surface of the slopes and that the area had been extensively affected by erosion. As a direct response to the observations made in the 1999 archaeological field season and in the spring 2000 season of the geomorphological and pedological survey, the archaeological field survey in the autumn of 2000 focused on the southern upper plateau. A contiguous block in this southern plateau (c. 1 km x 7 km) was intensively surveyed in the hope that the depth of soil preservation (c. 45-60 cm) might reveal more cultural material. Field-walking was carried out in units (farmers’ fields) and the entire field was systematically investigated. Fields were traversed in parallel lines 10 metres apart. Field walkers called out all archaeological artefacts that were visible within a meter-wide strip in front of them, as they systematically covered the field. Due to the relatively small quantities of material, a practice of total collection was followed. Tile was collected initially, but this strategy was later modified; if there were standing structures in a field, these were noted and only a small sample of tile was collected for typological comparison. Extraordinary finds (coins, obsidian, etc.) were given small-find numbers and located on the general survey map. When concentrations were discovered all collection stopped and an initial investigation was undertaken by the team to establish its location and rough dimensions. All concentrations that were considered to be in primary context were then gridded and systematically collected on 10 m grid squares. Team members walked the grid square, a meter apart, and made one pass collecting everything in their individual line of vision, not spending more than 2 minutes per pass. Documentation of the landscape was undertaken, during field-walking, on standardised field cards. The field card required the systematic recording of information, not only the location of the field and archaeological material discovered within it, but also other variables, including: time of day; time spent walking the field; weather conditions; field-walking methods; vegetation type; and the visibility of ground surface. The intensive program of field-walking has concentrated mainly in the rural south-western and central portions of the survey universe. Material has been recovered that can be attributed to periods of history from the 4th millennium BC to the modern period. Most of the earlier material is confined to ‘sites’ or surface concentrations; such background scatter as there is sometimes comprises areas with Roman pottery, but more extensively material

initiated in 1997 as part of the Troina Project, which aimed to redress this imbalance and to document the initial occupation of the Sicilian uplands. The project was directed by Drs. C. Malone and S. Stoddart who began survey in the area of Troina and subsequently undertook excavation of the prehistoric site of Casa Sollima, which was found during the first year of survey (Ashley 2007, pp. 59-80; Malone & Stoddart 2000a and b; Malone et al. 2003). The Troina Project developed a flexible field-survey strategy in order to respond to the demands of working in this particular upland environment. In response to the apparent absence of cultural material on the surface of fields, we were constantly evaluating our methodology, developing and redefining it at different stages during the four seasons of fieldwork. The pilot season and general reconnaissance of the area in 1997 was followed by a season in 1998 of survey between the mountain town and the river valley to the south. The initial survey universe followed a rough transect along an aqueduct from the town to the southern river valley, as well as revisiting prehistoric sites that had been identified in earlier archaeological investigations (Scibona pers. comm.) (fig. 2). The study universe was modified in 1999 from being purely cultural and pragmatic in the initial year of survey, to focus on the catchment basin of the Fiume di Sotto di Troina. The 1999 field survey focused on the southern river valley, following a sampling strategy of 10% of the 10 km tract of river and valley sides under geomorphological study (in total 20 km2). The sample was established by the creation of a system of transects 100 m wide and 2 km long, placed at 1 km intervals along the 10 km of the study area. These transects, perpendicular to the river, crossed the valley and intersected the river terraces identified in the geomorphological mapping. The transects were used as guideline for field-walking and hence it was the farmers’ fields within and overlapping the designated areas that were surveyed. In situations where it was not possible to fieldwalk effectively within the transects, either because fields were overgrown and thus visibility greatly impaired or there was no access, comparable areas adjacent to the original transect were investigated instead. Over all four seasons of fieldwork the visibility of the ground and the accessibility of the fields were taken into account prior to actually surveying a field. As the 1999 season developed, it became clear that there was surprisingly little cultural material 7

Hidden Landscapes

from the last 200 years. We have also recorded rural structures dating from the last two centuries (Walker 2007, pp. 115-142). Further to the survey work, attention was directed between 1998 and 2000 to the excavation of one multi-period site at Casa Sollima (Ashley et al. 2007, pp. 59-80; Malone & Stoddart 2000a and b; Malone et al. 2003).

has influenced the preservation and movement of both site and ‘non-site’ archaeology (Ayala & Fitzjohn, 2002b; 2007, pp. 99-113; James et al. 1994; Terrenato & Ammerman 1996; Pluciennik et al. 2004, pp. 27-65). The majority of the survey universe is on hill-slopes and comprises large landholdings that are dedicated to cereal cultivation, supported with limited maintenance of orchards. Furthermore, agricultural practices which leave the fields ploughed and thus free from vegetation at the time of the autumn rains are likely to intensify the destruction, movement, and reburial of the archaeological record (fig. 3). Observations during field survey and the results of the geomorphological research have made it apparent that the archaeological record of the Troina valley is highly disturbed. We appear to be losing much of the background scatter that has been identified in other areas of the Mediterranean. Clearly, in order to understand the distribution of material across our survey universe, we needed to acquire better knowledge of the geomorphology and known land use practices which have created the current landscape. The geomorphological investigation of archaeological sites encountered during field survey, acting as windows into the history of the catchment basin, has shown that there have been several episodes of increased localised soil movement. The most pertinent example of this are the sites on the Casa Sollima plateau where the stratigraphical sequence shows at least four phases of slope disturbance and colluvial aggradation

3. The Hidden Troina landscape When attempting to understand the archaeological record of the Troina Project, we felt that it was important to try to compare quantities of material that we were recording with those discovered in other surveys. For this reason, we chose to use the density scale adopted in the Biferno valley for the definition of a site: 1-5 sherds as ‘find spots’, 6-20 as ‘probable sites’, and 21 and over for ‘sites’ when found in isolation or when found within site assemblages of later periods (Barker 1995, p. 138). In the light of discussion concerning the visibility of prehistoric sites (James et al. 1994; Bintliff et al. 1999; Barker 2000; Mee & Cavanagh 2000; Thompson 2000), we decided that we would lower those figures for prehistoric sites to compensate for poor preservation of impasto wares. Irrespective of this differential scaling of sites, we found that in comparison with other Mediterranean surveys we recorded far fewer sites, with a much lower density of material per hectare (Malone & Stoddart 1994; Barker 1995; Thompson 1999). The Troina landscape, along with other parts of Sicily and areas of the Mediterranean, has experienced a dramatic change in cultivation practices since the 1950s. An expansion of cereal production supported by increased mechanisation, which has allowed for larger fields to be created and ploughing to take place to deeper depths,

Figure 3a, b – Examples of erosion on the slopes of the Fiume di Sotto di Troina valley.

8

Fitzjohn – Ayala, Making sense of a hidden Sicilian landscape…

of sites was representative of the long-term occupational history of the region. We would then have had to assume that the land around Troina was not occupied during much of history. The combination of the geoarchaeological work and the erosion modelling, however, can provide an alternative explanation for the site pattern that we have identified from the field survey. In reality our archaeological record is only a relic of past activity: the extensive processes of erosion and deposition have potentially hidden much of the traces of past activity, leaving us with only a partial record of the past. Overall, the majority of sites were found in a less eroded, only recently exposed part of the valley to the south of the river. A number of situations that are free from deep ploughing or that have only relatively recently been subjected to such ploughing (in particular within the southern upper plateau) have revealed dense concentrations and scatters of archaeological material from a range of periods. There is, as one would expect, a strong correlation between site preservation and soil preservation. In the more degraded areas the sites that were visible appear to have been preserved by burial with either colluvium or subsequent construction, and were only visible due to fortuitous circumstances, and often only for a short amount of time (Ayala 2004; Fitzjohn 2003). Whilst the erosion modelling has provided a tool with which we can begin to understand landscape history we have learnt that it must be used with a certain amount of caution. For example, it would be problematic to use the output of such models as a predictive map of the probable location of archaeological sites. We located some sites that had remained buried until just before the field survey in areas that were predicted to show, and in fact exhibited, signs of extensive erosion due to the topography and recent land-use. If we had used our erosion model to predict the areas that we should have walked in, these sites would never have been recorded.

followed by periods of stabilisation in pre-Copper Age, post Late Copper Age/pre-Hellenistic, postHellenistic and recent times (Ayala 2004, Ayala & French 2003, pp. 229-239). Furthermore, the soils found in the valley floor attest to the frequency with which colluvial aggradation has been followed by stabilisation. This tendency for soil movement represents the vulnerability of this landscape to erosion and the consequent disturbance of the archaeological record.

4. Erosion modelling Since the beginning of the 1990s, GIS has become an essential tool to record all aspects of fieldwork in the hope of undertaking spatial analysis of the distribution of all archaeological material on most Mediterranean projects (Ayala & Fitzjohn 2002a and b; Gillings et al. 1999). It goes without saying that the popularity of GIS stems from the ability for its users to record, store, manipulate, analyse and present multiple forms of data. Numerous papers published over the last twenty years have extolled the virtue of GIS analysis; some archaeologists have even discussed the use of GIS to examine erosion (Conolly & Lake 2006, p. 41) but few have tried to relate this to the issue of artefact visibility. In the context of Troina it seemed essential to examine our survey data in relationship to GIS models of landscape change. Even though there is a wide variety of models for assessing the risk of erosion in the Mediterranean (CORINE 1992; Kirkby et al. 1996; Thornes 1993), the availability of input data has been an important limiting factor in choosing a particular model for the study area (Pistocchi et al. 2002; van der Knijff et al. 2002). In view of the objectives of the project and nature of the data, our model was based upon the concept of the Universal Soil Loss Equation (USLE; Wischmeier et al. 1978). Utilising data from the geoarchaeological survey (Ayala 2004; Ayala & French 2003, pp. 229-239), the Troina Erosion Model involved the adaptation and application of the USLE to the terrain of the catchment area. Our aim was to investigate the spatial pattern of erosion potential and the relationship between this pattern and the archaeological record. This kind of modelling does not allow for the evaluation of actual sediment dynamics and yield, and variations of these patterns through time, but it does provide an indication of the relative strength or intensity of the phenomena. Prior to the use of the erosion modelling, it would have been possible to assume that the distribution

5. Conclusions The project at Troina is one of the first to reveal aspects of a landscape that has, until relatively recently, remained hidden from discussions of prehistoric and historic Sicily. The combination of field-walking and excavation has started to complement earlier excavations in the town. We now have a much clearer understanding of the relationship between town and countryside as well as of the relationship between this landscape and other regions of Sicily. 9

Hidden Landscapes

Geoarchaeological research has revealed evidence within the Fiume di Sotto di Troina catchment which would suggest highly mobile soil types and an aggradational sequence that saw at least one distinct phase of Holocene aggradation. Site-based investigations have indicated more frequent phases of localised mass movement throughout the span of human occupation of the river valley (Ayala 2004, Ayala & French 2003; 2005). Of further interest is the integration of geoarchaeological data, field-survey data and erosion modelling. The work in Troina has illustrated the necessity of examining field-survey data in relation to the geomorphological development of a region, prior to interpretation of the archaeological record as evidence of settlement patterns and regional history. One immediate advantage is that it we were able to understand some of the reasons why our survey data was so different from many other projects in Italy and other parts of the Mediterranean. Rather than looking at our region as a blank space to be filled with archaeological data, and then characterizing an absence of data as absence of activity, the erosion modelling has helped us to understand the peculiarities of the Troina landscape. This research has also directed us to carry out more detailed studies of specific sites within the region, which has allowed for more nuanced understanding of the various periods and regions of the survey universe.

researchers. The University of Liverpool, 23-25 February 2001. British Archaeological Reports International Series, BAR Publishing, Oxford, pp. 12-19. Ayala G. & Fitzjohn M. 2002b. Seeing is believing: Questions of Archaeological Visibility in the Mediterranean, Antiquity 76, pp. 337-8. Ayala G. & Fitzjohn M. 2007. To be seen or not to be: interpretations of survey data and questions of archaeological visibility in upland Sicily, in M. Fitzjohn (ed.), Uplands of Ancient Sicily and Calabria: The Archaeology of Landscape Revisited. Accordia Specialist Study on Italy, pp. 99-113. London, Accordia Research Centre, University of London. Ayala G. & French C. 2003. Holocene landscape dynamics in a Sicilian upland river valley, in Howard A.J., Macklin M.G. & Passmore D.G. (eds), The Alluvial Archaeology of North West Europe and the Mediterranean, Balkema, Rotterdam, pp. 229-239. Ayala G. & French C. 2005. Erosion Modeling of Past Landuse Practices in the Fiume di Sotto di Troina River Valley, North-central Sicily. Geoarchaeology 20(2), pp. 149-167. Barker G. (ed.) 1995. A Mediterranean Valley. Landscape archaeology and Annales History in the Biferno Valley Leicester, Leicester University Press. Barker G. 2000. Hidden Prehistoric Landscapes: An Italian Perspective. Journal of Mediterranean Archaeology 13(1), pp. 100-102. Bintliff J., Howard P. & Snodgrass A. 1999. The Hidden Landscape of Prehistoric Greece. Journal of Mediterranean Archaeology 12(2), pp. 139-168. Biondi C. 1991. Troina Mediovale: Filippo de Samona, miles. Archivio Storico per la Sicilia Orientale Fasicolo I-III, pp. 7-145. Blanchard P. 1998. Sicily and the Aeolian Islands. London, A&C Black. Braudel F. 1972. The Mediterranean and the Mediterranean World in the Age of Philip II. London, Collins. Brydone P. 1773. A tour through Sicily and Malta: In a series of letters to William Beckford. London, W. Strahan & T. Cadell. Champion T.C. 1989a. Introduction, in T.C. Champion (ed.) Centre and Periphery: Comparative Studies in Archaeology, pp. 1-21. Cambridge, Unwin Hyman Ltd. Champion T.C. (ed.) 1989b Centre and Periphery: Comparative Studies in Archaeology. Cambridge: Unwin Hyman Ltd. Conolly J. & Lake M. 2006. Geographical information systems in archaeology. Cambridge manuals in archaeology. Cambridge, UK, Cambridge University Press. CORINE 1992. Soil Eroison Risk and Important Land Resources in the Southern Regions of the European Community. EUR 13233, Luxembourg. Fitzjohn M.P. 2003. Spatial investigation of interaction between the indigenous and colonial populations in Sicily during the First Millennium BC. PhD thesis, The University of Cambridge. Giannitrapani E. & Pluciennik M. 1998. La seconda campagna di ricognizione (Settembre 1997) del progetto “Archeologia nella Valle del Torcicoda.” Sicilia Archeologica 96, pp. 59–69. Gillings M., Mattingly D. & van Dalen J. (eds) 1999. Geo graphical Information Systems and Landscape Archaeology. Oxbow Books, Oxford. Giuliano M.F. 1901. Memorie Paesane ossia Troina dai tempi antichi sin oggi. Catania, Nicolo Giannotta Editore.

References Albanese Procelli R.M., Alberghina F., Brancato M., Procelli E., Sirena G. 2007. The project and first results of the Gornalunga and Margi Valleys Survey, in Fitzjohn M. (ed.), Uplands of Ancient Sicily and Calabria: the archaeology of landscape revisited. Accordia Research Institute, London, pp. 35-48. Alliata V., Belvedere O., Cantoni A., Cusimano C., Marescalchi P. & Vassallo S. 1988. Himera III. Prospezione archaeologica nel territorio. “L’Erma” di Bretschneider, Roma. Ashley S., Bending J., Cook G., Corrado A., Malone C., Pettitt P., Puglisi D., Redhouse, D. & Stoddart S. 2007. The resources of an upland community in the fourth millennium BC, in M. Fitzjohn (ed.), Uplands of Ancient Sicily and Calabria: The Archaeology of Landscape Revisited. Accordia Specialist Study on Italy, London, Accordia Research Centre, University of London, pp. 59-80. Ayala G. 2004. Landscape/landuse change in north-central Sicily: a geoarchaeological approach. Unpublished PhD thesis, University of Cambridge. Ayala G. & Fitzjohn M. 2002a. Prehistoric Troinese LandscapesGIS Representations of Field Survey, in Muskett G. (ed.), SOMA 2001: Symposium on Mediterranean archaeology; proceedings of the fifth annual meeting of postgraduate

10

Fitzjohn – Ayala, Making sense of a hidden Sicilian landscape…

conservation practices, www.iemss.org/iemss2002/proceedings/ pdf/volume%20tre/331_pistocchi.pdf.

Idrisi, Abu ‘Abd ‘Allah Muhammad ibn ‘Abd ‘Allah ibn 2001. Il libro di Ruggero. Excerpt in Amari M., Viaggiatori arabi nella Sicilia medievale, Biblioteca storica del viaggio in Sicilia, Palermo, p. 55. James P., Mee C. & Taylor G. 1994. Soil Erosion and the Archaeological Landscape of Methana, Greece, Journal of Field Archaeology 21, pp. 395-416. Johns J. 1985. The Monreale Survey: indigenes and invaders in Medieval west Sicily, in Malone C. & Stoddart S. (eds), Papers in Italian Archaeology IV, British Archaeological Reports International Series 246, British Archaeological Reports, Oxford, pp. 215-223. Johns J. 1992. Monreale Survey. L’insediamento umano nell’Alto Belice dall’eta paleolitico al 1500 d.C, in Biondi L., Corretti A., De Vido S., Gargini M. & Vaggioli M.A. (eds), Atti delle giornate internazionali di studi sull’area elima, Scuola Normale Superiore di Pisa, Pisa and Ghibellina, pp. 407-420. Kolb M.J. 2007. The Salemi Survey Project: long-term landscape change and political consolidation in interior western Sicily 3000 BC – AD 600, in Fitzjohn M. (ed.), Uplands of Ancient Sicily and Calabria: The Archaeology of Landscape Revisited. Accordia Specialist Study on Italy, London, Accordia Research Centre, University of London pp. 171-185. Kirkby M.J., Baird A.J., Diamond S.M., Lockwood J.G., McMahon M.L., Mitchell P.L., Shao J., Sheehy J.E., Thornes J.B. & Woodward F.I. 1996. The MEDALUS Slope Catena Model: a physically based process model for hydrology, ecology and land degradation interactions, in Brandt C.J. & Thornes J.B. (eds), Mediterranean Desertification and Land Use, John Wiley & Sons, Chichester, pp. 303-354. Leighton R. 1999. Sicily before history: An archaeological survey from the Palaeolithic to the Iron Age. Duckworth: Trowbridge. Leone A., Witcher R.E., Privitera F. & Spigo U. 2007. The Upper Simeto Valley Project: An Interim Report on the First Season, in M. Fitzjohn (ed.), Uplands of Ancient Sicily and Calabria: The Archaeology of Landscape Revisited. Accordia Specialist Study on Italy, London, Accordia Research Centre, University of London, pp. 49-58. Malone C. & Stoddart S. (eds) 1994. Territory, Time and State: The archaeological development of the Gubbio Basin, Cambridge, Cambridge University Press. Malone C. & Stoddart S. 2000a. A House in the Sicilian Hills, Antiquity 74, pp. 471-472 Malone C. & Stoddart S. 2000b. A contribution towards the understanding of Serraferlicchio. Sicilia Archeologica 33, pp. 97-103. Malone C., Ayala G., Fitzjohn M. & Stoddart S. 2003. Under the Volcano. Accordia Research Papers. London, Accordia Research Centre, University of London. Mee C. & Cavanagh B. 2000. The Hidden Landscape in Ancient Greece: A View from Laconia and Methana. Journal of Mediterranean Archaeology 13, pp. 102-107. Militello E. 1961. Troina- Scavi effettuati dall’istituto archeologia dell’Università di Catania negli anni 1958 e 1960. Notizie Degli Scavi XV, pp. 331-404. Mozzillo A. 1993. Passaggio a Mezzogiorno: Napoli e il Sud nell’immaginario barocco e illuminista europeo, Milano, Leonardo. Pistocchi A., Cassani G. & Zani O. 2002. Use of the USPED model for mapping soil erosion and managing best land

Pluciennik M., Mientjes A. & Giannitrapani E. 2004. Archaeologies of Aspiration: Historical Archaeology in Rural Central Sicily. International Journal of Historical Archaeology 8(1), pp. 27-65. Ragusa G.M.R. 1995. Insediamenti antichi in territorio di Troina. Unpublished Undergraduate Dissertation Università degli Studi di Catania, Catania. Riedesel J.H. von 1773. Travels through Sicily and that part of Italy formerly called Magna Graecia: And a tour through Egypt, with an accurate description of its cities, and the modern state of the country. [Trans. J.R. Forster]. London, Edward and Charles Dilly. Russell G. 1815. A tour through Sicily, in the year 1815. London, Sherwood, Neely & Jones. Scibona G. 1980. Troina 1: 1974-1977, Nuovi dati sulla fortificazione ellenistica e la topografia del centro antico. Estratto da Archivo Storico Messinese XXXI(Serie III), pp. 349-389. Stoddart S. 1989. Divergent trajectories in central Italy, 1200500 BC, in T.C. Champion (ed.) Centre and Periphery: Comparative Studies in Archaeology, pp. 88-101. Cambridge, Unwin Hyman Ltd. Terrenato N. & Ammerman A.J. 1996. Visibility and site recovery in the Cecina Valley Survey, Italy, Journal of Field Archaeology 23, pp. 99-109. Thompson S.M. 1999. A Central Sicilian Landscape: Settlement and Society in the Territory of Ancient Morgantina (5000 BCAD 50). Unpublished PhD thesis, The University of Virginia. Thompson S.M. 2000. The Still Hidden Landscape. Journal of Mediterranean Archaeology 13(1), pp. 111-115. Thornes J.B. (ed.) 1993. Medalus I. Final Report. Mediterranean Desertification and Land Use. Commision of the European Communities, European programme on climate and natural hazards, London. Van der Knijff J.M., Jones R.J.A. & Montanarella L. 2002. Soil erosion risk assessment in Italy, in Rubio J.L., Morgan R.P.C., Asins S. & Andreu V. (eds), Proceedings of the 3rd International Congress Man and Soil at the 3rd Millennium, pp. 1903-1917. Walker L. 2007. Troina and its medieval landscape from the 9th – 15th centuries: evidence of a complex upland environment in the published documentary sources, in Fitzjohn M. (ed.), Uplands of Ancient Sicily and Calabria: The Archaeology of Landscape Revisited, Accordia Specialist Study on Italy, London, Accordia Research Centre, University of London, pp. 115-142. Whitehouse R.D. & Wilkins J.B. 1989. Greeks and natives in south-east Italy: approaches to the archaeological evidence, in Champion T.C. (ed.), Centre and Periphery: Comparative Studies in Archaeology, Cambridge, Unwin Hyman Ltd., pp. 102-126. Wilson R.J.A. 1981. The hinterland of Heraclea Minoa (Sicily) in classical anitiquity, in Barker G. & Hodges R. (eds), Archaeology and Italian Society: Prehistoric, Roman and Medieval Studies. British Archaeological Reports International Series, Oxford, pp. 249-260. Wilson R.J.A. 1990. Sicily under the Roman Empire: The Archaeology of a Roman Province 36 BC-AD 535. Aris and Phillips, Warminster. Wilson R.J.A. & Leonard A. 1980. Field survey at Heraclea Minoa (Agrigento), Sicily. Journal of Field Archaeology 7, pp. 219-239.

11

Hidden Landscapes

Wischmeier W.H. & Smith D.D. 1978. Predicting Rainfall Losses: a guide for conservation. The USDA Agricultural Research Service Handbook 537, Washington D.C.

Acknowledgements We would like to thank the organisers of the conference Dr. Martijn van Leusen, Dr. Giovanna Pizziolo and Prof. Lucia Sarti for inviting us to participate in what was a stimulating and rewarding event. We are also grateful for their patience during the production of this paper. The research for this paper would not have been possible without the financial support provided by a number of bodies: Matthew Fitzjohn was on a NERC PhD research studentship at the University of Cambridge and Gianna Ayala was funded by a studentship from Magdalene College, Cambridge and the L’Oasi Maria SS. di Troina. Matthew Fitzjohn also received grants from St. John’s College, Cambridge. The survey formed part of the Troina Project under the directorship of Dr Caroline Malone and Dr Simon Stoddart. We would like to thank the Comune di Troina, L’Oasi Maria SS. di Troina, the Soprintendenza BB.CC.AA. Enna - Servizio Beni Archeologici and the McDonald Insitute for all of their support. We are grateful to Caroline Malone and Simon Stoddart for allowing us to run two seasons of field survey. Thanks must also be extended to Charly French for his support and advice in the field in Troina and as Gianna Ayala’s PhD dissertation supervisor. We would like to thank all of the students from the Universities of Cambridge, Catania and Rome who have invested their time and energy working with us on the field survey. In particular, we would like to thank Stephen Ashley, Ceri Boston, Andy Clarke, Matthew Brudenell, Valentina Copat, Annalisa Costa, Helen Farr, Italo Giordano, Paolo Guarino, Andy Hall, Philip Lloyd, Paolo Piccione, Jonathon Pragg, Fraser Sturt, Archondia Thanos, Lucy Walker, Nick Whitehead, Sarah Wolferstan, and Violetta Zefki. Whilst in Sicily, we benefited greatly from the unselfish assistance provided by a number of people. We are especially appreciative of the help extended by the Commune di Troina, the OASI Institute, and Padre PietroAntonio. Special thanks must go to Vito for helping us to organise logistics in Troina. Many thanks to David Redhouse and Mark Berger for their assistance with the initial organisation of the GIS and to James Brassington who was incredibly helpful with all of our questions regarding GIS and erosion modelling.

12

2 CALEROS: an erosion-deposition model for landscape archaeology Hendrik FeikenI, Rens van BeekII, Theo van AschII & Martijn van LeusenI I Groningen Institute of Archaeology, Poststraat 6, 9712 ER Groningen, The Netherlands – [email protected]; [email protected] II Department of Physical Geography, University of Utrecht, Heidelberglaan 2, 3584 CS Utrecht, The Netherlands – [email protected]; [email protected]

1. Introduction

and on the inclusion of the major anthropogenic processes: agricultural terracing and plough erosion.

Since 2000 the Groningen Institute of Archaeology (GIA) has conducted annual intensive and systematic archaeological field walking surveys in the Raganello catchment in northern Calabria (Italy). The primary result of these surveys is a series of highresolution pottery density maps, known to be distorted by the action of natural and anthropogenic processes on the landscape. However, the effects of these processes on the recovery of archaeological evidence could never be modelled in detail. In 2005 work began to remedy this and other methodological problems encountered by landscape archaeologists: the Hidden Landscapes project. In a collaboration with physical geographers from the University of Utrecht, the PC-Raster software is being used to construct the CALEROS dynamic erosion model. This poster presents our preliminary results in modelling the most important natural process: erosion. Future work will focus on refinement of the model,

2. The pilot area One of the areas studied by the Hidden Landscapes project is the Raganello river catchment in northern Calabria (fig. 1, inset map). A detailed archaeological field survey was conducted in three transects crosscutting the catchment. One of these transects runs through the catchment of a tributary stream to the Raganello: the Maddalena. Figure 1 shows the landscape in the Maddalena catchment, which we selected for the CALEROS pilot model because of the presence of interesting archaeological phenomena. The Maddalena area is made up mainly out of easily weathering schists (foreground) and hard limestones (background). A detailed geomorphological mapping was carried out in order to define landscape units in which geomorphological processes occur more or less homogeneously (Van Leusen & Feiken 2007).

Figure 1 – View of the Contrada Maddalena (northern Calabria, Italy), site of the CALEROS pilot erosion/deposition model.

13

Hidden Landscapes

Figure 2 – Landscape classification of an E-W transect through the Contrada Maddalena, with simplified results of the systematic field walking survey. Pale areas: surveyed fields. Yellow dots: archaeological sites. Background relief shade based on Lidar elevation model.

Figure 3 – Flow chart for the CALEROS model.

14

Feiken – van Beek – van Asch – van Leusen, CALEROS: an erosion-deposition model for landscape archaeology

Archaeological remains dating to the late Bronze Age were found to concentrate in one landscape unit only: the ‘undulating sloping land’ visible as lighter fields in the center of the photograph. Surface archaeology in the pilot area was mapped through intensive and systematic field walking (fig. 2, light areas). Large agricultural fields were first subdivided into 50 by 50 meter units, then walked at ten meter intervals by teams of five persons. All archaeological material within reach of the walker was collected for cleaning and detailed study at the team’s base. Descriptive information about the field, the unit, and any finds within it was recorded in the field using handheld and GPS-enabled computers. In order to understand whether the spatial pattern of protohistoric sites (yellow dots) is ‘real’ or due to systematic visibility biases, the role of slope processes within the Undulating Sloping Land unit (reddish brown) has been studied by a combination of computer simulations and field studies.

3. CALEROS Erosion modeling for archaeological purposes is not new (see, for example, Ayala & French 2005, Ebert & Singer 2004), but is mostly done to obtain an understanding of the erosion potential of an area and is often performed using the generic RUSLE algorithm. CALEROS is a dynamic model of landscape development in hilly terrain, and incorporates the soil production model of Cox (1980), the soil erosion model of Morgan, Morgan & Finney (1984) and the soil creep model of Carson & Kirkby (1972). An impression of the full model is given by the flow chart in figure 3. One of the model inputs is a digital elevation model (DEM), which we derive from ~1.5 m resolution airborne LIDAR data supplied by the Catholic University of Leuven and the Flemish Institute for Technological Research VITO (Belgium). Other inputs include climate, soil and vegetation data derived from various sources. Some inputs are not available; initial (mid-Holocene) soil thickness, for example, must be is estimated. The CALEROS model runs within PC-Raster environmental modeling software developed at the University of Utrecht (pcraster.geo.uu.nl). This software implements a computer language for the construction of iterative spatio-temporal environmental models, and provides an interactive raster GIS environment supporting immediate pre- or post-modeling visualization of spatio-temporal data. In other words, the model will work in time and space dimensions, and produces maps, time-series and tabular

Figure 4 – Time slice showing the potential depth of burial (in m) of archaeological remains in the Contrada Maddalena after 3000 years.

information. One significant restriction imposed by the PC-Raster software is its inability to run ‘backwards’ in time. We therefore have to estimate initial conditions such as Bronze Age soil depths. The output of CALEROS, a series of ‘time-slices’ showing simulated soil thicknesses (an example is shown in fig. 4), can be exported to ArcGIS where it can be analyzed in conjunction with the archeological distribution maps for the same area. The final CALEROS model will also include modern land use and land cover data (obtained by airborne hyperspectral scanning, and processed by our Belgian colleagues) and ancient land use data reconstructed from the analysis of pollen data.

References Ayala G. & French C. 2005. Erosion modeling of past land-use practices in the fiume di Sotto di Troina river valley, NorthCentral Sicily. Geoarchaeology 20(2), pp. 149-167. Carson M.A. & Kirky M.J. 1972. Hillslope form and process. Cambridge University Press. Cox N.J. 1980. On the relationship between bedrock lowering and regolith thickness. Earth Surface Processes 5, pp. 271-274. Ebert D. & Singer M. 2004, GIS, predictive modelling, erosion, site monitoring. Assemblage 8, www.assemblage.group.shef. ac.uk/issue8/ebertandsinger.html.

15

Hidden Landscapes

Leusen M. van & Feiken H. 2007. Geoarcheologie en Landschapsclassificatie in Midden- en Zuid-Italië. Tijdschrift voor Mediterrane Archeologie 37, pp. 6-16. Morgan R.P.C. 2001. A simple approach to soil loss prediction: a revised Morgan-Morgan-Finney model. Catena 44, pp. 305-322. Morgan R.P.C., Morgan D.D.V. & Finney H.J. 1984. A predictive model for the assessment of soil erosion risk. Journal of Agriculture Engineering Research 30, pp. 245-253.

16

3 A prehistoric hidden landscape in an alluvial plain: investigations in the Florentine area Giovanna Pizziolo, Lucia Sarti Dipartimento di Archeologia e Storia delle Arti, Università degli Studi di Siena, via Roma 56, Siena, Italy – pizziolo@ unisi.it; [email protected] Abstract This paper, using the case study of the Florentine area, explores bias and potentialities of exploring prehistoric context in an alluvial plain. Since 1982 ongoing rescue excavations in the Florentine area, and in particular in the territory of Sesto Fiorentino, reveal unexpectedly that the plain has been extensively inhabited since prehistoric times. In a highly sedimentary context such as the Florentine alluvial plain, our chances of finding prehistoric evidence are strongly conditioned by rescue work associated with urban expansion. In fact the archaeological heritage policy adopted by the Municipality of Sesto Fiorentino in accordance with the Soprintendenza per I Beni Archeologici requires that all areas subject to building or infrastructure construction have to be assessed through systematic preliminary archaeological test pits. This requirement has enabled archaeologists from the University of Siena to discover a substantial body of evidence, and to identify a broad temporal and spatial range of archaeological contexts. These investigations highlight that the prehistoric landscape is constantly hidden by the sedimentation or compromised by the erosion typical of the alluvial context. Which are the means that we can use to investigate such archaeological contexts? How can we manage the lack of information? The weak traces of prehistoric structures are not detectable on the surface but are revealed by test pits and subsequent excavations: these evidence remains hidden until we explore it stratigraphically. From this research perspective the integration of different sources of evidence and different scales of analysis is seen to be a powerful tool for the understanding of the prehistoric landscape. Different sources of information, such as data derived from historical cartography and different kinds of aerial photographs, have been analysed and integrated with stratigraphical information derived from pits and drillings into a GIS environment. The comparison and overlay of these data allows us to study, through the ‘lens’ of archaeological biases, the prehistoric settlement and landscape.

1. Introduction

area is part of the larger system of the Firenze-PratoPistoia alluvial plain, which was an intermontane lacustrine basin developing over a tectonic depression after the Late Pleistocene era (Conedera & Ercoli 1973; Ghinassi & Tangocci 2008). The main watercourse, the Arno River, drained this basin and gradually covered it with alluvial sediments; its tributaries created further alluvial fans along the margins of the plain (See Salvini et al., this volume). Geomorphological and palaeobotanical studies (see Carra et al., this volume) suggest that the inner basin of the Florentine plain was a wetland area during the prehistoric period, and that streams, marshes

This paper, starting from the case study of the Florentine area, illustrates some methodological procedures and discusses some issues regarding prehistoric hidden landscapes in alluvial plain contexts. A synthesis of the geographical and archaeological background is presented in order to highlight the constraints of investigating in an urban area in a sedimentary basin. The analysis of historical sources and of remote sensing data constitutes the general framework in which archaeological and geological information is compared through means including threedimensional visualisation. We highlight the benefits of a multisource and multitemporal approach in the reconstruction of the ancient landscape. The paper proposes a research approach which aims to identify biases and obtain a basic version of reliability maps which could help in exploring the prehistoric hidden landscape.

2. A complex study area: the Florentine plain A long term project related to the Florentine area (Florence, Italy) is the case study that we present as an example of methodological approach to study prehistoric landscape in an alluvial context. The study area corresponds to the Municipality of Sesto Fiorentino, a town located to the north-west of Florence in the foothills of Mount Morello (fig. 1). This

Figure 1 – The Florentine alluvial Plain; particular attention has been devoted to Sesto Fiorentino Municipality.

17

Hidden Landscapes

as well as the interactions between the prehistoric inhabitants and the wetland environment. In particular we need to analyse site distribution in terms of proximity to hypothetical shorelines or rivers, and to individuate the micro relief of the plain. In the last twenty-five years the urban expansion of Sesto Fiorentino, now stretching across the Florentine plain, has enabled archaeologists to discover a substantial body of evidence and to identify a wide temporal and spatial range of archaeological contexts through systematic test pitting within future construction areas, and to carry out several rescue excavations in the Sesto territory. Due to these circumstances, dense clusters of small trenches (with an average dimension of 1x3x3 m.) are disseminated irregularly in the study area; more than 400 of these test pits have provided positive results. In an area of 20 sq. km, 60 systematic excavations have revealed prehistoric contexts (Sarti et al. 2008) (fig. 2). Unfortunately, since rescue archaeology is severely limited by the rules and regulations of the municipality and its infrastructure, and consequently by the dimensions and plan of the constructions, it is generally not possible to investigate sites in their totality, i.e. in relation to their surroundings. In most cases it is even impossible to determine their extent (Pizziolo & Sarti 2005; 2008). Moreover, urban expansion governs the authorization to dig test pits and preliminary trenches and determines the position and size of the eventual archaeological excavations. The location of prehistoric sites suggests possible settlement trends and highlights how the area was increasingly occupied from the end of the 5th to the 2nd millennium BC (Sarti & Martini 1993). The general direction of expansion during this period was towards the southeast and somewhat toward the inner part of the plain. We can immediately observe that specific zones of the Plain were frequently chosen as settlement areas during Holocene prehistory; apparently these exerted the larger attraction with respect to other parts of the Plain. The settlement evidence is mainly composed of artefacts located in sunken structures, on palaeosurfaces located near the levees of palaeo-streams, or in the beds of abandoned palaeo-channels. Often prehistoric occupation is multiperiod, with a stratigraphic succession that indicates continuity and a prolonged use of the same morphological setting. In other cases the dwellings show a chronological spread but are not strictly part of the same stratigraphical sequence; this rather indicates the reoccupation of contiguous areas as represented by the horizontal stratigraphy.

a

b Figure 2 – A detail of the distribution of archaeological test pits (white dots). Background: a. aerial photo of 2004; b. aerial photo of 1954. The comparison highlights the landscape transformation of the Florentine plain occurred in the last decades.

and bogs dominated the landscape. Today, these watercourses are mostly regularized and transformed into channels especially where they cross the plain. The present-day landscape is characterized by evidence of the various different reclamation activities undertaken in the past to drain a territory with a natural wetland predisposition. One of the first systematic attempts, led by the Roman town Florentia, was carried out through centuriation, a process of subdivision and reclamation of land which was very effective and which had a lasting effect on the organization of the Florentine plain. This process imposed a regular grid of roads and channels following the direction of steepest slope, and its traces are still recognizable today. Since then, the hydrographic control of the alluvial plain has been a key element in the Man-Environment relationship. In fact the different reclamations carried out until the last century resulted in a network of channels in accordance with the roman grid. In comparison with these historical hydrographical networks, it is particularly interesting to investigate the prehistoric hydrography 18

Pizziolo – Sarti, A prehistoric hidden landscape in an alluvial plain: investigations in the Florentine area

The research project on Florentine Plain, started in 1982, has gradually grown over time and now involves different institutions with different roles. The Superintendecy of the Archaeological Heritage of Tuscany is in charge of preserving the archaeological heritage, the University of Siena conducts archaeological research, and the Municipality of Sesto Fiorentino takes urban planning decisions according to the archaeological evidence, in order to avoid as much as possible a negative impact on archaeological evidence. The research of the University of Siena is focused on exploring the prehistoric peopling process and settlement strategies in the area through the contextualisation of prehistoric evidence within landscape evolution. In particular we attempt to reconstruct the palaeohydrological setting and to outline the different biases caused by geomorphological processes and by the conditions under which archaeological excavations take place. L.S.

sion phases occurred throughout passive and active processes. The sedimentation cycles led to the filling of lowland plain, generating surfaces which trend to a topographical levelling characterised by a low gradient of slope. In the Holocene period this process led to the present-day landforms; in some cases, at a detailed scale of analysis (compare, for example, Valloni & Baio 2008), it is possible to observe the Holocene sedimentary system with lateral migration of channels and flooding of the inter-channel areas. However on these layers we can also observe the erosive process undertaken by streams, channels and rivers. The uniform topographical levelling of these surfaces formed by fine sediments often produces optimal conditions exploited during roman period to establish a centuriation grid, which in itself constituted a drainage surface. These brief observations lead us to remark that the formation processes of the alluvial plain cover and erode the archaeological layers and in particular they affect the prehistoric ones. Generally, prehistoric evidence is therefore not easily visible in this geomorphological context, and field survey activities will not discover a large number of prehistoric artefacts on the surface. A peculiar exception to this general trend is represented by the Pleistocene ridges which sometimes are not covered by recent alluvial sediments, standing above the plain surface (see the fluvial ridges in Veneto Plain, Bondesan et al. 2004). Other Pleistocene land units, even if not topographically prominent, may sometimes also remain visible on the surface because they were never covered by subsequent natural or anthropogenic fills (see for example Bellotti et al. 2004; ArnoldusHuyzendveld 2007; Pizziolo, in press). Since prehistoric time they have exerted a sort of attraction for human settlement choices due to their soil characteristics and morphology. Typically, however, thick layers of sediments overlaying the prehistoric evidence cause a low visibility of artefacts on the present-day surface. To have an approximate idea of the speed of this process, it is worth mentioning that an increase in sediment of 2 mm per year has been estimated for some parts of the Po Plain in the last 12,000 years (Valloni & Baio 2008, p. 34). This gives us an idea of the possible ‘hiding’ effect on prehistoric surfaces due to successive deposit formation. Moreover, prehistoric evidence was lost by the eroding effect due to local movement of streams, channels and rivers which may have worn away the archaeological layers; in this case the erosion of possible prehistoric palaeo-

3. General methodological framework Since landscape archaeology deals with the reconstruction of landscape transformation observed in a present-day setting, our effort is directed to understanding, diachronically, a palimpsest of evidence which is draped on a synchronic surface (Leonardi 1992). We observe that in alluvial contexts, unfortunately, the signs of the prehistoric evidence are often hidden. For this reason our research aims to highlight and explore continuity and change of the archaeological landscape, trying to identify features and signs related to prehistoric phases, and to connect these up into meaningful complexes with which to ‘clothe’ the buried prehistoric landscape. We have therefore adopted a multidisciplinary approach which is based on sharing archaeological and geographical data in a single framework integrating different scales and different sources into a GIS environment. Looking from a wider perspective we need to identify the different processes which gave shape to the present day settings and try to understand: a) which natural processes or human actions occurred in the past b) in which parts of the landscape analysed, c) how they affect site preservation and visibility of prehistoric contexts. In our case study data are affected by several constraints which influence the possibility to discover prehistoric finds. G.P.

4. Potentialities and biases Alluvial plains are broadly characterised by a dynamic environment where the deposition and ero19

Hidden Landscapes

these different classes of data should constitute a good basis on which to evaluate bias effects.

surfaces leads to a data deficiency which hides the consistency of prehistoric settlement or at least causes its underestimation. In these circumstances the individuation of prehistoric artefacts throughout field survey activities becomes ineffective. It is well known that in order to investigate ancient landscapes it is important to have not only dots on a map (i.e. single information derived by archaeological excavations), but continuous data which cover the whole territory under study. In alluvial plain contexts prehistoric evidence could be better individuated if we were able to investigate extensively beneath the soil surface, but this opportunity is not available especially where the area under study is densely populated and is part of the present-day settlement system. We must therefore take opportunistic advantage of the circumstances: often stratigraphies excavated for non-archaeological purposes allow us to investigate, unexpectedly, under the soil surface. However we are left with a problem of data interpretation. How can we evaluate prehistoric evidence (and subsequently prehistoric landscape) in such a difficult context? We assume that awareness of all the biases that affect our data is the first step to take. The procedure that we propose aims to gain the most information from two different approaches: a) analysis of the land surface in order to assembly information on landscape features of the past and of the present; b) sub-surface investigations in order to assemble information on archaeological contexts. Subsequent cross-tabulation and overlay analysis of

4.1 Analysis ‘on’ and ‘over’ the landscape surface

This analysis starts from the acquisition, in a GIS framework, of multitemporal sources which provide information on topographical and landscape settings of our study area for a wide time range. The first set of sources used for this analysis consists of modern cartographic data (technical and topographical maps at different scales which describe the urban/ rural characteristics and expansion trends) and historical cartography (historical series of topographical maps dating back to 1889 and historical cadastres dating back to 1820); other historical sources dating back to the 15th century have been examined but not input into the GIS. These cartographic data have been analysed, and information on landscape settings has been extracted and converted into vector layers. In this way it is possible to isolate and date single or thematic territorial settings. Another very effective data source is constituted by remote sensing (Pizziolo 2006, 2010; Salvini et al., this volume). Particular attention has been dedicated to recent and historical aerial photographs (2004, 1997, 1965, 1954, 1943), and to the Near-Infrared colour photos produced during the construction of the high speed railway which crosses the alluvial plain (1997). The photointerpretation, which was initially aimed at the identification of archaeological features, provides its most interesting information when used to outline landscape features and in particular palaeo-hydrographical features. Aerial photos have also been compared with QuickBird satellite images (May, 2006) in order to enhance with different sources the palaeo-hydrography of the plain (see Salvini et al., this volume). The archaeological record put into the GIS (artefact scatters, structures, standing features, infrastructures elements) and observable on the landscape surface is mainly related to the Etruscan, Roman and Medieval evidence. A very large archaeological feature is the centuriation grid imposed by the centre of Florentia in the 1st century AD on the Florentine Plain. In order to input the traces of centuriation into the system we use a module based on squares with 355m sides, which has been oriented and virtually ‘anchored’ to the roman evidence discovered in the plain (Cresci & Zannoni 1999), obtaining the most probable model of the centuriation network of our area. To give an example of the benefit of comparison and overlay of data, we may observe that the field

Figure 3 – An aerial photo of 1954 shows the continuity of the orientation of the field system and of the centuriation grid (dashed lines).

20

Pizziolo – Sarti, A prehistoric hidden landscape in an alluvial plain: investigations in the Florentine area

a

b

Figure 4 – An example of integration of sources: data derived by archaeological excavation and aerial photos. a. The palaeoriver bed of Cilea occupied by settlement activities during the Bell Beaker and Bronze Age period (white polygons). The sinuous shape of the palaeoriver (black line) contrasts with the regular field system (aerial photo of 1954). b. In the aerial photo of 2004 the field system has been transformed into an urban context.

management, discontinuous from the less impacting prehistoric setting, which can be revealed by a detailed reconstruction of the palaeo-hydrography situation especially on the basis of anomalies detected within the regular roman field system. If remote sensing data and cadastral maps provide information that can be used to individuate hydrological anomalies, this needs to be related with data regarding the underground settings.

system drawn on the historical cadastre confirms the general orientation of fields and follows the organisation of the territory imposed through the centuriation system. This continuity is confirmed by the historical aerial photographs of 1954 which shows the same field pattern (fig. 3). The greatest impact on the archaeological landscape is due to this reclamation system. In recent aerial photos the signs of this centuriation network are less evident, and the peculiar field system is not visible anymore. The roman grid therefore had a strong impact on the territorial organization, lasting until the 20th century when the plain was still characterised as a traditional rural area. In the last decades, with the expansion of the main city centres, the plain became mostly an urban territory. Today we can recognise just relict traces of past settings at the surface, surrounded by urban structures or infrastructures or enclosed in mechanized farming fields. Nevertheless the possibility to visualise the documented landscape history in a GIS framework constitutes a good starting point for exploring the prehistoric settings. In effect the roman territorial organisation has produced a new approach to land

4.2 Analysis ‘beneath’ the land surface

The analysis beneath the land surface starts with the collection of data derived from different sources. First of all, the archaeological data produced by rescue excavations have been input into the system and visualised in their three-dimensional position. Particular attention has been paid to the reconstruction of prehistoric palaeosurfaces, which also have been visualised three-dimensionally beneath the land surface. In several cases the palaeosurfaces are sufficiently extensive to be analysed in their spatial relationship with the present day land surface. This is the case, for example, with the Bell Beaker 21

Hidden Landscapes

correlation of lithological units has been undertaken with the aim of recognising geomorphological features. The acquisition into the GIS of this information, beyond the archiving utilities offered by its database structure, supports the formulation of geological hypotheses on the underground settings within a landscape archaeological perspective. Unfortunately this type of information is available only in the part of territory affected by the construction of buildings and infrastructure according to the rescue archaeology policy adopted in the area. Dense clusters of small trenches are spread irregularly in the plain. Obviously, in the historical urban zones only a few test pits have been dug because they have not been urbanized recently. The same unbalance may be observed in the rural area. Looking at the density distribution of archaeological test pits it appears clear that it is biased by urban expansion; if the discovery of prehistoric evidence is strictly related to the opportunity of examine stratigraphies beneath the surface, then the prehistoric site distribution must be similarly biased. A third type of information derives from the analysis of stratigraphies excavated for non-archaeological purposes. From municipality archives we have collected the descriptions of geognostic drillings executed over the last three decades, mainly in the territory of Sesto Fiorentino municipality. These cores, which have been drilled to obtain information on the subsoil for constructing foundations of buildings, are generally available as reports where each stratum is briefly described. These reports are very heterogeneous in terms of accuracy, detail and criteria adopted for the description of stratigraphic units, and the depth of drilling may vary from 3 to 30 m. We standardised them and managed them into the GIS using an application developed by the Geographical Information System Unit (Direzione Generale Territorio e Urbanistica – Sistema Informativo Territoriale) of the Region of Lombardy. The DB is effectively structured to standardise data in terms of lithological composition and the application, named urca.avx, allows one to create and visualise cross sections in a correct spatial way within the GIS (fig. 5). Even if these drillings are not detailed enough for archaeological interpretation, they do provide useful information on underground settings, for instance the presence of gravel deposits, which can be used to reconstruct the geomorphological setting. The geognostic drillings partially fill the gaps left by the irregular distribution of archaeological test pits.

palaeosurfaces which occupied the abandoned beds of palaeorivers (Sarti & Martini 2001): here we can observe the morphology of the palaeoriver through the three-dimensional visualisation of its contour lines (fig. 4) and compare this morphology with the present-day surface. Secondly, much information also derives from the acquisition into the GIS of all the archaeological test pits dug in the area as a consequence of the rescue excavation regulations adopted by some municipalities of the Florentine plain. Every single trench has been georeferenced, and its stratigraphy has been acquired stratum by stratum three-dimensionally into the system along with its lithological composition and chronological and cultural attribution, where available (Pizziolo 2007). As a result of this work, the sedimentary units of the top 3 m of some zones of the plain are well known. This information is very useful for better defining the nature of palaeohydrological settings, in particular as concern palaeorivers. For instance, the distribution of gravel deposits is the first clue for detecting possible river beds or areas affected by ancient flooding. The

a

b Figure 5 – The three-dimensional software environment allows us to visualise, beneath the land surface on which we drape an airphoto, both geognostic drillings (a) and archaeological test pits (b). In (b) we have also displayed the stratigraphic position of a Bell Beaker structure (‘floating’ horizontal lines).

22

Pizziolo – Sarti, A prehistoric hidden landscape in an alluvial plain: investigations in the Florentine area

In general the ability to associate archaeological and chronological information to lithological units is an important tool with which to assign chronological ranges to prehistoric land surfaces. G.P.

5. Investigating prehistoric contexts in alluvial plains The results obtained by investigating ‘over’, ‘on’ and ‘beneath’ the land surface need to be compared. Our first consideration is how to integrate archaeological evidence into the landscape framework: we need to use all the different sources that can provide information on prehistoric settlement settings, shifting dynamically from the micro to the macro scale and vice versa. The integration of the two scales (global and local) is fundamental to the study of the prehistoric landscape, and intriguing results are coming from the analysis of palaeosurfaces and test pits stratigraphies. In particular the reconstruction of settlements which occupied ancient riverbeds (Sarti 1997; Sarti & Martini 2001) provides information regarding the morphology and sinuosity of these features, which are confirmed by photo interpretation (Pizziolo & Sarti 2006) and by geomorphological analysis of cross sections (see section 4.2). This analysis shows how it is important to compare sources in order to interpret these small relict segments of the prehistoric landscape: in fact in the case of palaeostreams large- and small-scale sources are consistent and allow us to link individual features and define a more detailed hydrological framework in which the prehistoric settling occurred. The reconstruction of the morphological context needs to be accurate because in the Florentine Plain the prehistoric evidence is characterised by the absence of marked structures and large settlement, as well as by the presence of a range of ‘light’ evidence, probably related to a seasonal occupation of the area which has left weak traces. If prehistoric evidence does not constitute a strong structural contrast compared to the morphological background then we really need to know the natural settings in detail so that we can perceive this ‘low impact’ appropriation of nature (sensu Ingold 1986) occurring during Holocene prehistory. Unfortunately these ephemeral characteristics give to our prehistoric landscape a further bias and determine moreover its ‘hidden’ nature. The adoption of methodologies which are sensitive to offsite and near-site evidence plays an important role. If such evidence could be connected one with the other in an integrated system it would become sig-

Figure 6 – An example of a set of nearby archaeological areas (Lastruccia 1, 2a, 2b, and 3) that need to be inter-connected and may form a ‘settlement unit’.

nificant, because part of a group, and outlined as features of the prehistoric landscape. We have to remember that the definition of the extent of a site during rescue excavation is determined by the needs of the building site or by urban factors that sometimes take into account little of the true extent of the prehistoric structures (fig. 6). Furthermore the assessment of the areas adjacent to the site is not easily achievable through archaeological test pits, making it even more difficult to identify areas dedicated to agricultural production and animal breeding. However even if the Florentine Plain is not characterised by complex structured prehistoric settlements, a lot of evidence testifies to a sense of attraction exerted by this area which often is focussed on the same zones. It is important to understand if these observations can be related and interpreted through spatial relationships. Instead of looking at single site or off-site evidence we can try to connect archaeological finds and delineate ‘settlement units’ (Pizziolo & Sarti 2008, pp. 45-49) that is to say a complex of evidence which in a portion of territory may be characterised by: • dwellings contemporaneous or compressed in a narrow chronological band; • dwellings in stratigraphic succession that indicate continuity and a prolonged use of the same morphology or the reoccupation of the same area but with new settings; • dwellings in chronological succession but not stratigraphically superimposed; the dwellings indicate the reoccupation of areas very close by. This perspective, which tries to link single observations to settlement processes, when analysed and related to morphological features may provide new hints in the reading of prehistoric landscape. It also 23

Hidden Landscapes

leads us to pay attention to the areas around the archaeological digs as being potential ‘hidden’ areas of frequentation that escape the preliminary emergency investigations. G.P., L.S.

Bell Beaker settlement strategies) but in others as negative because they may have eroded prehistoric evidence. Clearly it is fundamental in such cases to provide a chronological attribution to the geomorphological features, in order to assess them as positive or negative. If we overlay the map layers derived from the positive classification of these two criteria, we can produce a new layer characterised by fragmented polygons in which large parts of the rural landscape are excluded even if they may still preserve prehistoric evidence. Indeed the classification of rural areas in terms of bias detection is quite complex: modern agricultural activities (heavy movement of soil for levelling the agricultural surface, landuse and in general the trend to obtain homogeneous characteristics in the soil production) may affect in different ways the underground settings and they have been recorded in our data set as negative zones. The building up of a reliability map, even if simplified, is a first step to deal with biases (fig. 7). Overlaying it with settlement distribution maps, we can use it to calibrate our spatial analysis and settlement interpretations, avoiding under- or overestimations of prehistoric site densities and

6. Reliability maps to mitigate bias effects The integration of the sources and data mentioned above allows us to create a ‘reliability’ map to delimit portions of the plain where the interpretation of the prehistoric landscape is less affected by biases; this map should show both (1) areas effectively investigated using means adequate for individuating prehistoric contexts, and (2) areas that may still preserve prehistoric evidence. We have classified as ‘positive’ all the zones including test pits and as ‘negative’ the historical urban areas and the areas covered by buildings built up to the 1980s. For the second criterion, we have excluded again as ‘negative’ all the historical urban areas and all the area affected by erosive processes or strong deposition processes, such as alluvial fans or debris. Specific attention has been paid to palaeorivers which in some cases may be classified as highly positive because they constitute an attractive area for prehistoric settlement choices (see for instance

Figure 7 – The reliability map.

24

Pizziolo – Sarti, A prehistoric hidden landscape in an alluvial plain: investigations in the Florentine area

able to delineate these features in the GIS, we will be able to update our reliability map and to assess in the field if some parts of the prehistoric landscape are indeed less ‘hidden’ than before. G.P., L.S.

patterns. Nevertheless we have to go further than binary classification and Boolean operators, introducing other variables and procedures. Some valuable suggestions about methodologies derive from predictive modelling experiences developed in America (Kvamme 1990) and spread in the last decade in Europe (Van Leusen & Kamermans 2005). Here we do not wish to discuss the different predictive approaches but simply observe that the scale of map sources and the level of detail of data collected in the field strongly affect the utility of the outcome. For the Florentine context the work is in progress and our reliability map is conceived as a descriptive tool to handle macro biases. Our biggest problems concern rural or abandoned areas in the central part of the plain, where it is difficult to estimate the erosion and deposition impact and to outline the palaeo-hydrological settings. The scale of map sources is a key element in the analytical process (see Saile & Lorz 2005), and at this stage of the project the 1:10,000 geological maps are not detailed enough. Our analysis of historical sources suggests that the inner part of the plain was subject to frequent alluvial events and water stagnation but we do not know yet how large the swamps and wetlands in the inner plain were during Holocene prehistory. When such information becomes available we can better define our reliability map. G.P.

References Arnoldus-Huyzendveld A. 2007. Le trasformazioni dell’ambiente naturale della pianura grossetana, in Citter C. & Arnoldus-Huyzendveld A. (eds), Archeologia urbana a Grosseto. Origine e sviluppo di una città medievale nella “Toscana delle città deboli”. Le ricerche 1997-2005. Vol. I: la città di Grosseto nel contesto geografico della bassa valle dell’Ombrone, All’Insegna del Giglio, Firenze, pp. 41-62. Bellotti P., Caputo C., Davoli L., Evangelista S., Garzanti E., Pugliese F. & Valeri P. 2004. Morpho-sedimentary characteristics and Holocene evolution of the emergent part of the Ombrone River Delta (southern Tuscany), Geomorphology 61, pp. 71-90. Bondesan A., Meneghel M., Rosselli R. & Vitturi A. (eds) 2004. Geomorphological Map of the Province of Venice, scale 1:50.000. LAC, Firenze, 4 sheets. Conedera C. & Ercoli A. 1973. Elementi geomorfologici della piana di Firenze dedotti da fotointerpretazione. L’Universo 53, pp. 255-262. Cresci M. & Zannoni M. (eds) 1999. Itinerari nella storia del territorio sestese, Archeoprogetti, Firenze. Ghinassi M. & Tangocci F. 2008. Middle-Late Holocene deposits of the Sesto fiorentino area (Florence, Italy): preliminary data, in Baioni M., Leonini V., Lo Vetro D., Martini F., Poggiani Keller R., Sarti L. (eds), Bell Beaker in Every day life, Proceedings of the 10th Meeting “Archéologie et Gobelet” (Florence – Siena – Villanuova sul Clisi, May 12-15, 2006), Millenni, Firenze, pp. 33-37. Ingold T. 1986. The appropriation of nature: essays on human ecology and social relations. Manchester University Press, Manchester. Kvamme K.L. 1990. The Fundamental Principles and Practice of Predictive Archaeological Modeling, in Voorrips A. (ed), Mathematics and Information Science in Archaeology: A Flexible Framework. Studies in Modern Archaeology 3, Holos-Verlag, Bonn, Germany, pp. 257-295. Leonardi G. 1992. Assunzione e analisi dei dati territoriali in funzione della valutazione della diacronia e delle modalità del popolamento, in Bernardi M. (ed), Archeologia del paesaggio, IV Ciclo di lezioni sulla ricerca applicata in archeologia, Certosa di Pontignano 14-26 gennaio 1991, Firenze 1992, pp. 25-66. Pizziolo G. 2007. Towards prehistoric landscape interpretation: GIS analysis of stratigraphical and remote sensing data in Sesto Fiorentino plain, in Figueiredo A. & Leite Velho G. (eds), The world is in your eyes - Proceedings of the XXXIII Computer Applications in Archaeology Conference: Tomar March 2005, CAA Portugal, Tomar, pp. 441-446. Pizziolo G. 2010. Landscape Archaeology at Sesto Fiorentino: the contribution of aerial photographs to the study of archaeological contexts within an integrated approach, in Niccolucci F. & Hermon S. (eds), Beyond the artifact, digital interpretation of the past. Proceedings of CAA 2004, Prato 13-17 April 2004. Budapest, Archaeolingua, pp. 479-483.

7. Further work As we have already suggested it would be very important to describe with more precision the geomorphological processes that have occurred in the area. Other future work will be focused on a detailed analysis of series of test pits and trenches to define palaeohydrological settings. Special attention will be paid to the interpretation of suspected erosion surfaces through the integral three-dimensional analysis of pits, drillings and of information derived from archaeological excavations. The erosive processes may affect the preservation of prehistoric evidence. A sedimentological approach has been undertaken during archaeological excavations and the elaboration of data is in progress. Ongoing geomorphological studies will be able to furnish important data on the presence of possible ‘high’ morphologies, which are useful in the definition of structural constants in the analyses of the archaeological landscape at a local scale. In some cases these morphological features may have exerted a specific attraction for prehistoric people. Once we are 25

Hidden Landscapes

Pizziolo G. in press. The prehistoric peopling process in the Holocene landscape of the Grosseto area: how to manage uncertainty and the quest for ancient shorelines, Proceedings of LAC2010, the first international Landscape Archaeology Conference, Amsterdam: Free University of Amsterdam. Pizziolo G. & Sarti L. 2005. Landscape archaeology in Sesto Fiorentino: a GIS analysis for investigating settlement strategies in wetland area, in Berger J.F., Bertoncello F., Braemer F., Davtian G. & Gazenbeek M. (eds), Temps et espaces de l’homme en société, analyses et modèles spatiaux en archéologie, XXV Rencontres Internationales d’Archeologie et d’Histoire d’Antibes, Antibes, Editions APDCA, pp. 443-452. Pizziolo G. & Sarti L. 2006. Exploring the archaeological landscape through a local perspective: spaces and places in the prehistory of the Florentine plain, in Campana S. & Forte M. (eds), From Space to Place, 2nd International Conference on Remote Sensing in Archaeology, Proceedings of the 2nd International Workshop, CNR, Rome, Italy, December 4-7, 2006 (BAR International Series 1568), Oxford, BAR Publishing, pp. 261-269. Pizziolo G. & Sarti L. 2008. Prehistoric Landscape. Peopling process and Bell Beaker settlements in the Florentine Area, in Baioni M., Leonini V., Lo Vetro D., Martini F., Poggiani Keller R., Sarti L. (eds), Bell Beaker in Everyday life, Proceedings of the 10th Meeting “Archéologie et Gobelet” (Florence – Siena – Villanuova sul Clisi, May 12-15, 2006), Millenni, Firenze, pp. 39-58. Saile T. & Lorz C. 2005. An Essay on Spatial Resolution in Predictive Modelling, in Van Leusen M. & Kamermans H. (eds), Predictive Modelling for Archaeological Heritage Management: A research agenda (Nederlandse Archeologishce Rapporten 29), National Service for Archaeological Heritage, Amersfoort, pp. 139-147. Sarti L. & Martini F. 1993. Costruire la memoria: archeologia preistorica a Sesto Fiorentino, 1982-1992, Garlatti e Razzai, Montelupo Fiorentino. Sarti L. (ed.) 1997. Querciola. Insediamento Campaniforme a Sesto Fiorentino, Ed. Garlatti e Razzai, Firenze. Sarti L. & Martini F. 2001. Strategie Insediative del Campaniforme nell’Italia Centrale Tirrenica, in Nicolis F. (ed.), Bell Beakers Today. Pottery, people, cultures, symbols in Prehistoric Europe. Proceedings of the International Colloquium, Riva del Garda (Trento, Italy) 11-16 May 1998, pp. 187-198. Sarti L., Balducci C., Brilli P., Fenu P., Leonini V., Martini F., Pizziolo G. & Zannoni M. 2008. Catalogue of Bell Beaker Settlements in Sesto Fiorentino, Florence, in Baioni M., Leonini V., Lo Vetro D., Martini F., Poggiani Keller R. & Sarti L. (eds), Bell Beaker in Every day life, Proceedings of the 10th Meeting “Archéologie et Gobelet” (Florence – Siena – Villanuova sul Clisi, May 12-15, 2006), Millenni, Firenze, pp. 23-32. Valloni R. & Baio M. 2008. Sedimentazione altoquaternaria nel tratto emiliano del tracciato Alta Velocità, in Bernabò Brea M. & Valloni R. (eds), Archeologia ad alta velocità in Emilia. Indagini geologiche ed archeologiche lungo il tracciato ferroviario, Atti del Convegno Parma, giugno 2003 (Quaderni di Archeologia dell’Emilia Romagna 22), All’Insegna del Giglio, Firenze, pp. 21-40.

26

Van Leusen M. & Kamermans H. (eds) 2005. Predictive Modelling for Archaeological Heritage Management: A research agenda (Nederlandse Archeologische Rapporten 29), National Service for Archaeological Heritage, Amersfoort.

4 Intensive survey and protohistoric settlement in the middle Guadiana basin (Badajoz, Spain) Victorino Mayoral HerreraI, Sebastián Celestino PérezI, Sabah Walid SbeinatiII I Instituto de Arqueología–Mérida (CSIC–Junta de Extremadura), Plaza de España 15, 06800 Mérida, España – [email protected]; [email protected] II UNDERGROUND Arqueología y Gestión del Patrimonio S.C., España – [email protected] Abstract Although agriculturally marginal, the Guadiana river basin has been an axis of connection between the Mediterranean and Atlantic shores of the Iberian peninsula. Nevertheless its archaeological landscape remains mostly hidden. Little effort has been put into regional-scale survey in comparison to other peninsular regions. In this paper we show recent work carried out by the Institute of Archaeology of Merida in this direction. Intensive survey results from the area surrounding the archaeological site of Cancho Roano (Zalamea de la Serena, Badajoz) are analyzed. Our aim is to find a balance between flexible and quick recording methods, and the detailed study of artefact distributions. A wide range of archaeological finds was detected, including very low density scatters. Problems related to ground visibility in pastoral areas, site definition and dating of materials will be discussed. Finally we will consider interpretation of the results in terms of land exploitation during Protohistoric and Roman times.

1. Introduction

(see a comprehensive sample of these activities at www.ujaen.es/centros/caai/) and the coastal plains of Levant (Bonet Rosada & Mata Parreño 2001) and Catalonia (Sanmartí 2001). In contrast, there are still few initiatives to penetrate the mountainous and economically marginal areas, in which fieldwork poses more difficulties and provides fewer rewards in terms of finds (Ruiz del Arbol Moro 2005). This causes an unequal understanding of the historical evolution of the various peninsular landscapes. We are still far from being able to sketch broad diachronic developments in extensive areas.

In a recent synthesis of the current state of archaeological survey (Ruiz Zapatero 2004), it was noticed that, despite the good health of this type of studies at a European scale, the Spanish panorama presented numerous deficiencies and uncertainties. The 1980s and 1990s saw a stage of enthusiastic embrace of field surveying; a generation of researchers critical of the academic tradition of their teachers formed high expectations regarding its potential for the understanding of people-environment interactions. This was accompanied by a growing awareness of the need of exhaustive documentation on a regional scale for the development of protection and management policies. However, the increase of field surveys over the years has not been as strong as was hoped at the time. The autonomous regional governments have developed data bases and procedures for the recording of new sites, but have not backed a sustained effort outside the areas affected by infrastructural and urban development. Cases like that of the Community of Madrid, with a systematic and complete survey coverage of its territory, are exceptional (Velasco 2000). Although there have been important academic research projects in which field surveys played an important role, most of these have focused on particular periods, especially prehistory and protohistory. Another important constant is the limited concern with the specification of methodology, depriving us of the possibility to weigh the validity of the data presented. In consequence, strong regional imbalances persist. There is, for example, abundant information for areas like the Guadalquivir valley

2. Looking to the far west: survey projects in Extremadura We want to illustrate this theme here by presenting the situation in one specific area: the peninsular southwest, and in particular the region of Extremadura. Although this region has been the object of several systematic field surveys, it cannot be said that there were strong links between the different projects, or that they have been properly published. The most intensively studied part of the region has been the middle course of the Guadiana river. Prehistorians interested in Chalcolithic settlement patterns developed systematic work along the valley and across a series of tributary streams (see a description of the survey at www.departamento.us.es/dpreyarq/ web/vhp1.htm). More selective work has been done looking for the location of protohistoric tumuli (see below; Duque Espino 2007). By contrast, a great part of the region is occupied by mountainous terrain covered by Mediterranean oak woodlands (the so called dehesa). In these areas only selective surveys were conducted, aiming to identify characteristic set27

Hidden Landscapes

tlement types like Iron Age hillforts. This work has been mainly guided by toponymy, aerial photography and local informations. An interesting sampling experiment was developed by Martín Bravo (1994) in order to assess the probability of finding sites in flat areas. However, the surroundings of some of these settlements have provided evidence of intense occupation. This concern for the off-site record is

reflected by the undertaking of intensive surface surveys within a radius that, more or less explicitly, is identified with the natural area of resource catchments. This kind of survey was carried out around the Middle Iron Age settlement of Villasviejas del Tamuja (Botija, Cáceres; Hernández, Bravo & Galán 2009) and the early Iron Age building of La Mata (Campanario, Badajoz; Rodríguez Díaz 2004). Within this context, from the end of the 1990s, a research program has been developed by the Merida Institute of Archaeology that focuses on the archaeological survey and analysis of the Serena region in Extremadura (fig. 1). This work departs from a longterm study of a very singular archaeological site: the protohistoric building of Cancho Roano (Celestino Pérez & Jiménez Ávila 1993, Celestino Pérez 1996, 2003). Detailed knowledge of this enclave, excavated in its totality, unavoidably led to the need to explore its territorial context by means of intensive field survey. In the face of the fierce debate over its function, we hoped to obtain a complementary view from the surrounding territory. We bounded the survey using a 3 km buffer around Cancho Roano (fig. 2), departing from the assumption that there would have existed some kind of interaction between the building and the nearby contemporary sites. We also assumed that the distribution of the latter could tell us something about the productive and social logic that governed this relationship. The overall objective of this work was, however, to study the evolution of the agrarian landscapes of the area from prehistoric times up until the most recent past. Indeed, a key aspect of research into the history of human occupation in the region is the apparent collapse of the population in the 5th century BC, following the violent destruction of Cancho Roano and other monumental buildings of the area.

Figure 1 – Location of the Serena Region in the Iberian Peninsula, Autonomous Region of Extremadura.

3. Intensive survey in the Serena region 3.1 Methodology

The methodology followed in the intensive survey tried to achieve a balance between quality of data and work investment. We were fully aware of the limitations that would arise from ignoring the fuzzy boundaries between the archaeological sites and the ‘background noise’ detected between them. The survey thus set out to record the density of surface items throughout the entire landscape. Our specific procedures employed the current agrarian field boundaries as the basic working units. The location of materials was marked and recorded with GPS

Figure 2 – Limits of the intensive survey area around Cancho Roano.

28

Mayoral Herrera et alii, Intensive survey and protohistoric settlement…

receivers. Clusters of up to five fragments and/or materials that showed no clear concentration were recorded as point features. When it was possible to define a clear scatter, the finds were recorded as an area feature. As a complementary measure, the entire survey area was subjected to random sampling, intended to verify the reliability of density estimations produced by this type of study. For this work we used a recording system that enables the easy location and delimitation of the areas to be sampled, and the fast recording of the requisite information. Sampling units were outlined as circles of 15 m radius, their coordinates pre-loaded in a GPS receiver and then located in the field by navigation. Once found, the sampling area was defined with the help of a tape measure. A fixed time was established to search and recover sherds inside. Forms were designed to obtain a detailed description of the sampled area (land use, ground visibility, topography…). We took away every sherd, and description and quantification of the materials was carried out in the laboratory. An additional approach based on remote sensing techniques proved to be very successful. The fusion of Landsat TM and Spot Pan images of the survey area pointed to several places of interest: NDVI values showed abnormal moisture in nine locations, seven of which were confirmed as protohistoric sites by ground verification (Nieto Masot et al. 2003).

Figure 3 – Survey methods employed: a. full coverage survey; b. sampling survey.

of the hydraulic network, drainage capacity of the soil, slope, potential sun exposure, and underlying geology). The strategy followed in the search for patterns first required the characterisation of point distributions. Density maps and tests such as Nearest Neighbour analysis were carried out to enable us to observe the existence of significant clustering. Tests were then applied to independent samples in order to explore the possible role of particular landscape parameters in the formation of these clusters. The distribution of variables describing the contents of the material scatters (weight and number of fragments) was analysed, as well as the characteristics of the terrain in which they were located. For each period, the population was set to the totality of locations with archaeological finds. However, the differences between the more diffuse and the more consistent scatters were also analysed. In both cases, the actual archaeological distributions were compared against randomly generated samples. Finally, we compared the information belonging to protohistoric and Roman chronologies. Beyond the confirmation or negation of the existence of differences in the spatial distribution of settlements and potential resources, we wanted to calibrate the intensity of the correlations, between them, although this part of the study has yet to be elaborated.

3.2 Data analysis: general procedures

Although for various reasons the analysis of the ceramic materials has not yet been completed, we have enough information to begin our analysis of the spatial distribution of the finds. The first step has been to define variables that in our opinion may indirectly reveal patterns of location and occupation of the landscape. A central aspect has been to verify the hypothesis that Cancho Roano was placed in a very special location. To explore this hypothesis, the degree of accessibility between the building and the surrounding sites was explored, and topographic prominence and cost weight in time units were quantified. The visual relationship between Cancho Roano and its surroundings was analysed by calculating of cumulative viewshed values for all of the sites. With regard to the distribution of the evidence throughout the entire survey area, we have analysed variables relating to the potential productivity of the terrain. A scale of suitability was designed on the basis of available parameters (distance to elements 29

Hidden Landscapes

3.3 Protohistoric settlement around Cancho Roano

eas of maximum density in the tributary streams of the Cagancha and along this river within a radius of approximately 1km around Cancho Roano. We can define at least three foci that should be identified as small ‘sites’. The chronology of these groupings has been determined with some degree of certainty. Materials such as amphorae and grey ware are well dated from the early 7th century to the early 5th century thanks to excavation contexts such as at Medellín and at Cancho Roano itself. It cannot however be excluded that other materials of less diagnostic value, documented in some of these places, may in fact correspond to later occupation phases, particularly

As a result of the survey numerous traces of Iron Age settlement were identified that we were careful, at first, not to interpret as ‘sites’. In some cases these formed clear clusters, while in others it is more likely that the materials are evidence of the exploitation of the territory in the form of places of restricted activity or field manuring. An initial visual approach to the protohistoric colonization of the study area can be performed through the creation of density maps. As can be observed (fig. 5), once this calculation is carried out for the totality of finds, we can define a series of ar-

Figure 5 – Density map of protohistoric finds around Cancho Roano.

1200

1000

CRP3

800 CRP5

600

400

200

0 N=

Figure 4 – GIS based analysis of the area surrounding Cancho Roano: a. topographic prominence; b. cost surface (time units).

28

32

Area

Point

Figure 6 – Box-plot graphic of distances to water from protohistoric finds, comparing isolated point and area entities.

30

Mayoral Herrera et alii, Intensive survey and protohistoric settlement…

of Roman date. But as we have seen, there exists a much larger number of sites, up to 11, that can also be framed within the same chronology. This number does not take into account the important concentration of materials within 200-250 m from Cancho Roano itself, whose assessment is problematic. On the one hand, it is reasonable to interpret this material as the ‘halo’ of residues that a continued occupation produces at most habitation sites. However, it cannot be ruled out that much of it is caused rather by the numerous disturbances and alterations to the area, including the transport of earth from the excavation of the building itself. The global distribution of protohistoric materials at a five kilometres range from Cancho Roano displays little intensity, and is mostly associated with the fluvial system. This obviously points to a tendency for the location of settlements to prefer the least prominent areas with the greatest drainage capacities. A look at the distribution of distance-to-watercourses indicates that this tendency is stronger in the case of the more important scatters (fig. 6). There is, in any case, little topographic contrast in the landscape, and human occupations did not stand out visually in their immediate surroundings. In fact quite the opposite turned out to be the case: considering the average elevation of the locations, a non-random behaviour can be suggested which is confirmed by the results of independent sample tests. The building of Cancho Roano itself, for example, is camouflaged in the terrain, right next to the riverbank of the Cagancha.

Figure 7 – Detail of sherd distributions (black dots) around the protohistoric building of Cancho Roano.

ent on the authority of the ‘palaces’. The dispersed nature of this population could, indirectly, enable us to delimit the aristocratic domains. Agricultural colonization in this area was achieved by these rural centres of power, which have been considered as the centres of great fundi. This constitutes a remarkable difference with the settlement dynamics in other parts of the peninsula, like the upper Guadalquivir valley during the Early Iron Age, which are characterized by a strong process of nucleation in urban settlements. At the end of the 5th century this vigorous rural flourishing was interrupted by the abandonment of the monumental buildings. The excavation at Cancho Roano shows evidence of a ritualized and wellplanned destruction of the building by fire, and the nearby site of La Mata was also destroyed by fire. Following this collapse, the visibility of the Middle Iron Age occupation in the La Serena region is very low. We know about a small number of fortified villages identified by selective survey in strategic places along the main rivers and mountains, but evidence for single farms is nearly absent. The preliminary examination of the pottery collected during our survey suggests, however, that a small number of the small sites around Cancho Roano survive during this period. This evidence of continuity should widen our perspectives on the study of scattered rural habitats after the end of the monumental buildings. Research carried out in Central Alentejo (Portugal) show a similar phenomenon of humble continuity, after a peak in rural settlement during the first Iron Age (Mataloto 2004).

3.4 Land exploitation and rural power in the Guadiana valley

Constructions similar to Cancho Roano (although with a much less clear religious component) can be found throughout the middle course of the Guadiana river (Rodríguez Díaz et al. 2004b, Jiménez Avila 1997, Rodríguez & Ortiz 1998). These have been interpreted as rural palatial residences, where the monumentality of the buildings contrasts with the low visibility of a great number of surrounding sites with protohistoric pottery (fig. 8). The latter are usually located in flat, open places close to small streams, and do not show evidence of defensive works or other large structures. Quite often their surface record includes sets of saddle querns and storage vessels. Excavations of some of these show very small, domestic-scale agricultural activities (Rodríguez Díaz et al. 2007), and they are currently interpreted as small farms managed by peasant families depend31

Hidden Landscapes

The Serena landscape seems to remain quite empty during the 4th and 3rd centuries BC, or at least there is a lower population density, mostly concentrated in hillforts. But at the end of this period, with the Roman conquest, we can observe again two clear and new occupational phenomena. First, a series of large and fortified settlements were developed, that can be identified as oppida. These seem to be related to the exercise of strategic control over the territory; the

one closest to Cancho Roano is Magacela (Rodríguez Díaz & Ortiz Romero 2003). Second, selective survey has identified an ensemble of buildings of large stone masonry whose function is disputed (Ortiz Romero 1995, Ortiz Romero & Rodríguez Díaz 1987, Rodríguez Díaz & Ortiz Romero 2003). These very small structures consist of quadrangular towers surrounded by one or more walled enclosures. Their occupational sequences begin in the 2nd century BC

Figure 8 – Distribution map of late orientalizing tumuli in the Middle Guadiana Basin (according to Rodríguez Díaz et al. 2004)

Figure 10 – Roman walled enclosure near Cancho Roano.

Figure 9 – Fortifications of the Magacela oppidum (photo: V. Mayoral).

32

Mayoral Herrera et alii, Intensive survey and protohistoric settlement…

and continue through to the Imperial period; one such site is located at a short distance (700 m) from Cancho Roano. We cannot expand much on the behaviour of the Roman finds distributions. Independent sample tests do not point to the existence of significant changes in topographic prominence values compared with protohistoric finds. The R values of nearest neighbour analysis are quite similar in both periods. However, the average distances between settlement traces indicate a greater degree of clustering than in protohistoric times, while the less well-defined scatters are more abundant and more widely dispersed than in the previous period.

4. Discussion We will close this paper with a brief commentary on some of the problems encountered and on future research lines. Regarding the field survey, a recurring problem was the existence of strong contrasts in ground visibility. Most of the area was occupied by open oak woodland, often left fallow or under pasture. To avoid this bias, we opted to revisit fields that had not been ploughed during the first reconnaissance. However, some non-cultivated areas could not be surveyed at all. No corrective index was adopted for land use or ground visibility. The issue of site definition is also problematic. Traditionally arbitrary thresholds tend to be set to establish ‘significant’ groupings in the density of finds, but we feel that its hardly possible to set clear boundaries, especially in areas with some nearby small clusters. A quantitative approach based on raw counts per surface unit must be complemented with a more detailed recording. Sometimes what is considered a ‘site’ according to the number and density of sherds, reveals a pattern of size, roundness and fragmentation that rather suggests off-site activities like manuring. The need for geoarchaeological studies is equally pressing, if we want to refine our understanding of the significance of the surface scatters and to calibrate the magnitude of the alteration processes. Another fundamental need is for a more intensive study of the materials recovered in order to date the scatters. The lack of typological parallels causes great chronological ambiguity over long periods, and the sequence at Cancho Roano spans approximately 300 years. Even more urgent is the problem of the low visibility of particular periods, such as the Second Iron Age which finds itself ‘squashed’ between two phases characterised by a much more obstrusive and monumental record.

Figure 11 – Density map of Roman finds around Cancho Roano.

Finally, we would like to repeat that the difficulty of establishing comparisons with other nearby survey projects is a major determining factor in the persistence of our ‘hidden’ landscapes. Differences in the theoretical and methodological backgrounds have led to the use of very different recording criteria and categories of analysis. This is, for example, the case with the intensive field survey in the surroundings of the building of Mata de Campanario, with a similar chronology to Cancho Roano and located only 18 km away (Rodríguez Díaz et al. 2004a). It is remarkable that, even when applying the same ‘magnifying glass’, the use of different methodological and analytical filters perpetuates the imbalances in our understanding of the historical evolution of the landscape. To conclude, intensive survey around the site of Cancho Roano has provided us with a detailed picture of the archaeological distribution in a limited space. This task was conditioned by the previous development of research, and it is now imperative to consider a broader view by means of extensive field surveys. This is currently our main objective. Recent campaigns are offering new results. The aim is on the one hand to calibrate the representativity of the archaeological record in different landscape units. On the other hand, we hope to cover sufficiently large spaces in order to assess the existence of territorial structures throughout time. 33

Hidden Landscapes

References

Rodríguez Díaz A., Pavón Soldevilla I. & Duque Espino D. 2004a. “La Mata” y su territorio, in Rodríguez Díaz A. (ed.), El edificio protohistórico de “La Mata” (Campanario, Badajoz) y su estudio territorial. Cáceres, Universidad de Extremadura. Rodríguez Díaz A., Pavón Soldevilla I. & Duque Espino D. 2004b. “La Mata”: macroespacio y contexto histórico, in Rodríguez Díaz A. (ed.), El edificio protohistórico de “La Mata” (Campanario, Badajoz) y su estudio territorial. Cáceres, Universidad de Extremadura. Rodríguez Díaz A., Pavón Soldevilla I., Duque Espino D. & Ortiz Romero P. 2007. La “señorialización del campo” postartésica en el Guadiana Medio: el edificio protohistórico de La Mata (Campanario, Badajoz) y su territorio, in Rodríguez Díaz A. & Pavón Soldevilla I. (eds), Arqueología de la tierra. VI Cursos de Verano Internacionales de la Universidad de Extremadura. Cáceres, Universidad de Extremadura. Rodríguez Díaz A. (ed.) 2004. El edificio protohistórico de “La Mata” (Campanario, Badajoz) y su estudio territorial, Cáceres, Universidad de Extremadura. Ruiz Del Arbol Moro M. 2005. La Arqueología de los espacios cultivados. Terrazas y explotación agraria romana en un área de montaña: La Sierra de Francia (Salamanca), Madrid, Consejo Superior de Investigaciones Cientificas. Ruiz Zapatero G. 2004. La prospección arqueológica de superficie en los inicios del siglo XXI. Arqueología Espacial 2425, pp. 17-31. Sanmartí J. 2001. Territoris i escales d’integració política a la costa de Catalunya durant el període ibéric ple (segles IV-III a.C.), in Martín Ortega A. & Plana Mallart R. (eds), Territori polític i territori rural durant l’ edat del Ferro a la Mediterrània Occidental. Actas de la Taula Rodona celebrada a Ullastret. Gerona, Museu d’Arqueologia de Catalunya Ullastret. Velasco F. 2000. La carta arqueológica de Madrid: concepto, desarrollo, aplicaciones y perspectivas de futuro. Boletín de la Asociación Española de Amigos de la Arqueología 39-40, pp. 385-394.

Bonet Rosado H. & Mata Parreño C. 2001. Organización del territorio y poblamiento en el País Valenciano entre los siglos VII al II a.C., in Berrocal L. & Gardes P. (eds), Bibliotheca Archaeologica Hispana. Actas de la Mesa Redonda “Entre Celtas e Íberos. Las poblaciones protohistóricas de las Galias e Hispania”, Madrid, 1998. Madrid, Real Academia de la Historia Casa de Velazquez. Celestino Pérez S. & Jiménez Ávila J. 1993, El Palacio-Santuario de Cancho Roano IV. El sector norte, Badajoz, B. Gil Santacruz. Celestino Pérez S.E. 1996. El Palacio-Santuario de Cancho Roano V-VI-VII. Los sectores oeste, sur y este, Junta de Extremadura, Publicaciones del Museo Arqueológico Provincial de Badajoz 3. Celestino Pérez S.E. 2003. Cancho Roano VIII. Los materiales arqueológicos, Mérida, Junta de Extremadura Instituto de Arqueología de Mérida CSIC Bartolomé Gil Santacruz. Duque Espino D. 2007. La colonización agraria orientalizante en la Cuenca Media del Guadiana, in Rodríguez Díaz A. & Pavón Soldevilla I. (eds), Arqueología de la tierra. VI Cursos de Verano Internacionales de la Universidad de Extremadura. Cáceres, Universidad de Extremadura. Hernández P., Martín Bravo A. & Galán Domingo E. 2009. A la vista de las murallas: Análisis arqueológico del entorno del castro prerromano de Villasviejas del Tamuja (Cáceres). Complutum 20(1), pp. 109-132. Jiménez Avila J. 1997. Cancho Roano y los complejos monumentales post-orientalizantes del Guadiana. Complutum 8, pp. 141-160. Martín Bravo A. 1994. Metodología de prospección para la Edad del Hierro en la zona de Alcántara. Zephyrus 46, pp. 183-194. Mataloto R. 2004. Um “monte” da Idade do Ferro na Herdade da Sapatoa: ruralidade e povoamento no I milénio a.C. do Alentejo Central, Lisboa, Ministerrio de Cultura, Instituto Portugués de Arqueologia. Nieto Masot A., Hernández Carretero A. & Leco Berrocal F. 2003. Metodología de prospección arqueológica mediante imágenes Landsat TM y Spot PAN en Zalamea de la Serena (Badajoz). IX Conferencia Iberoamericana de Sistemas de Información Geográfica. Cáceres. Ortiz Romero P. 1995. De recintos, torres y fortines: usos (y abusos). Homenaje a la Dra. Milagro Gil-Mascarell Boscá. Extremadura Arqueológica V, pp. 177-193. Ortiz Romero P. & Rodríguez Díaz A. 1987. Problemática general en torno a los recintos-torre de La Serena, Badajoz. XIX Congreso Nacional de Arqueología, I. Castellón. Zaragoza. Rodríguez A. & Ortiz P. 1998. La Mata de Campanario (Badajoz): un nuevo ejemplo de “Arquitectura de prestigio”, in Rodríguez Díaz, A. (ed.), Extremadura Protohistórica: Paleoambiente, economía y territorio. Cáceres, Universidad de Extremadura. Rodríguez Díaz A. & Ortiz Romero P. 2003. Defensa y territorio en la Beturia: castros, oppida y recintos ciclópeos, in Morillo Á. Cadiou F. & Hourcade D. (eds), Defensa y territorio en Hispania. De los Escipiones a Augusto. (Espacios urbanos y rurales, municipales y provinciales). Coloquio celebrado en la Casa de Velázquez, 19-20 de marzo de 2001. Salamanca, Casa de Velázquez / Universidad de León.

Acknowledgements The work presented in this paper has been partially developed in the framework of the research projects ‘Entre el Atlántico y el Mediterráneo: Contraste de dinámicas en la evolución histórica del paisaje’ (HAR 2009-10666) and ‘Paisaje, territorio y cambio social en el suroeste peninsular: De la protohistoria al mundo romano’ (HAR 2008-1973). We would like to thank Enrique Cerrillo for his kind advice in the GIS analysis of land suitability. We are also grateful to Antonio Uriarte and Carlos Fernandez Freire for the careful and attentive reading of the paper and the useful suggestions on statistical treatment of the data.

34

5 From space to place or from site to landscape? Mind the gap Stefano Campana University of Siena, Department of Archaeology and History of Arts, Landscape Archaeology, Via Roma 56, Siena, Italy – [email protected] Abstract In recent years the University of Siena has been building up experience in the use of integrated prospection approaches in the study of archaeological ‘landscapes’. We have been aware from the outset that our research would have to focus initially on settlements rather than on landscapes as such. Inevitably, the intensity of archaeological research would be less in the intervening spaces, which for the most part would therefore remain areas of ‘emptiness’. We have also started to apply the highest available level and intensity of archaeological prospection methods on a large, complex and stratified site, occupied from the Etruscan period through Roman times and into the Middle Ages. At the same time we have also started to explore the surrounding landscape using the same basic methods. The archaeological objectives and outcome of this approach will be influenced by the critical impact of the kinds of information that are available for recording and assessing the potential and interpretation of the landscape. In attempting this it will be at least as important to appreciate what is not visible as to know what is. If we can achieve success in our sample area through the application of the highest level of research intensity and through conscious attempts to raise the level of visibility, and hence our understanding of the ‘emptinesses’ between identifiable ‘sites’, we will be able to make contributions to better practice in a variety of fields: conservation, academic issues and practice, future archaeological research and the development of research strategies.

1. From space to place: steps towards an improved integration of survey methods in the reconstruction of surface and subsurface archaeology

conduct of ‘aerial reconnaissance’ and for mapping archaeological features. But it must be understood that the vertical coverage was collected for nonarchaeological purposes, without any consideration of the factors that influence the visibility of archaeological evidence from the air. We do not wish to imply that we should altogether reject the study of vertical air photographs, but we must acknowledge their limitations and look for effective solutions. To overcome the problems with vertical photography, we started a programme of oblique aerial survey and related mapping in 2000; today we undertake an average of 45 hours of flight per year and have so far collected more than 38,000 oblique aerial images. The flexibility of the method, which allows us to respond quickly to day-to-day changes in archaeological traces, is of great importance (Campana et al. 2006a). Applied properly, this technique offers an extraordinary contribution to the search for new sites and for the continuous monitoring of the cultural heritage. It can be also be extremely helpful in the documentation of standing monuments surviving in built-up areas or in the countryside, such as castles, monasteries, churches or urban landscapes (Musson et al. 2005). The main limitations of oblique air photography lie in its dependence on the environmental conditions of the moment and other factors such as the local soils, geology and land-use. Aerial survey is also influenced by the individual archaeologist’s choice of which part of the landscape to document, and therefore involves a degree of serendipity if its full benefit is to be gained in any particular area.

There are few fields of scientific research which make simultaneous use of so many and so widely varying methods and instruments as archaeology, and in particular that section of archaeology that aims to understand past landscapes. Our work in the field of archaeological cartography has been based on remote observation systems, along with the enhancement of surface collection techniques through the use of new instruments and methods of collecting and documentating both archaeological and environmental data. Though this is a ‘work in progress’ we are already putting into effect a new research strategy. This is flexible, open-ended and based on the conviction that only through the integrated application of a wide range of research methods and information technology will we be able to confront the complexities inherent in the study of the landscapes of the past. So far, we have put in train the approaches summarized in fig. 1. Our experience with landscape research started almost thirty years ago with the use of vertical air photographs. With their wide temporal range, these represent an irreplaceable source for the analysis of the Tuscan landscape (Cosci 2005). Anyone interpreting the photographs through a digital photogrammetric workstation will see a 3D model of the landscape as it was in 1938, 1954, 1976, etc., until the present day. In addition to their historical content, vertical photographs are of course an important source for the 35

Hidden Landscapes

Figure 1 – Flow-chart of archaeological mapping processes.

Initially we tried to offset these problems by studying very high resolution Ikonos and Quickbird satellite imagery. This was aimed in the first place at providing a total, continuous and objective view of a whole area at a specific time of the year chosen by the archaeologist. Another advantage of this kind of imagery is its capacity to provide multispectral data to monitor plant health and to detect waterstress in vegetation where it cannot be seen by the naked eye (Campana 2004). Later we decided also to address our inability to deal in any significant way with the 50% of Tuscany (but also of Italy) that is covered in woodland. In this context there have been promising results in the United States, Great Britain, Germany and Austria from experiments in the use of lidar technology (or ‘airborne laser scanning’), which can record with great accuracy the surface morphology; it can even ‘see through’ woodland cover and revealing in great detail the underlying micro-relief (Holden et al. 2002; Doneus & Brise 2006). With the help of a Culture 2000 project entitled ‘European Landscapes: past, present and future’ and colleagues at the Natural Environment Research Council (UK) and at the Unit for Landscape Modelling of the University of Cambridge, we were able in 2005 to gain our first experience in lidar data acquisition, processing and interpretation for four sample areas in the provinces of Siena and Grosseto. The results allowed us finally to see archaeological sites protected by the

woodland from destruction by ploughing or other human activities (Campana et al. forthcoming). We believe that we have so far seen only the tip of the iceberg in the development of this technique, which will undoubtedly have a very significant impact on our knowledge of the landscape. In addition, we have been working on the implementation of ground-based sensing methods, and particularly of geophysics. Only a few years ago geophysics lay mainly in the domain of geophysicists, but now archaeologists can undertake some parts of the work themselves and can extend the scale of research from the individual site to the broader landscape (Powlesland 2006; Campana et al. 2006b). In the past few years we have been testing new tools for the large-scale acquisition of geophysical data, in particular the multiple antenna radar system GSSI Terravision (Finzi et al. 2005), the fast electrical imaging method ARP® developed by the French CNRS (Dabas 2008) and the magnetic survey system Foerster MULTICAT (Campana 2006). The results have been extremely significant, revealing archaeological features totally invisible to traditional archaeological research methods. Moreover the combination of — for instance — magnetic and electrical resistivity tomography (ERT) data allows us to see new evidence. Without mapping all of the information collected by this wide range of remote sensing techniques we will not be able to integrate different aspects 36

Campana, From space to place or from site to landscape? Mind the gap

2. From site to landscape: the Aiali case study

of our knowledge or obtain the full value of our work. Mapping, i.e. fixing the archaeological features in geographical space, is the first condition for understanding and conserving the cultural heritage. The archaeological evidence may be well interpreted, understood and described, but if it is not reproduced in map form it is merely another site that cannot be protected or fully understood in relation to the landscape (fig. 2). Remote sensing techniques are a very important part of the archaeological process, but they are not the only approach which we should use in landscape studies. Field-walking survey is a traditional but still extremely productive technique in the search for archaeological evidence in the Mediterranean. It has, like all others, some limitations but in combination with mobile technology (GPS and PDA) or test trenching the quality of the results is greatly improved (Campana & Francovich 2007). Just as with remote sensing, the mapping of field-walking survey data facilitates both the conservation of the discovered sites and the progressive accumulation of data sets. At the end of the archaeological process we are left with a huge amount of information. We have drawn attention to the mapping process because only through this is it possible to recover pieces of the jigsaw puzzle of historical traces that survive scattered in the modern landscape (Guaitoli 2003).

At this early stage of our search for a better understanding of ‘landscape’ problems we inevitably have to focus on a smaller scale. We must start from a site-based approach, while at all times highlighting aspects to be borne in mind when we move back up to the landscape scale. The villa site of Aiali is located on low-lying land between the medieval town of Grosseto and the Roman town of Roselle in central Italy. The site was first detected from the air during the Aerial Archaeology Research School organized by the University of Siena in 2001 (Campana et al. 2006b). Aerial survey allowed us to recognize an area within which the differential growth of the wheat revealed the plan of a 4 hectare complex of structures interpreted as a Roman villa. In the following years, Aiali became the most important test site for our laboratory. Since 2001 we have collected, processed and interpreted many different kinds of data: Quickbird-2 satellite imagery, historical and recent vertical coverage (from 1954 to 2001), oblique air photographs for various years, seasons and lighting conditions; fieldwalking and geophysical survey (magnetometry, GPR, EM, ERT and ARP©). One aim of the Aiali project is to apply the highest available level and intensity of archaeological prospection methods on a large, complex and stratified site that has produced

Figure 2 – Integrated landscape mapping.

37

Hidden Landscapes

survey has revealed the existence of a large medieval mound or castle (Campana et al. 2006b).

material from the Etruscan, Roman and Medieval periods. At the same time Aiali forms the starting point for a more wide-ranging attempt to return to the ‘landscape’ perspective by applying the same level of research intensity to the open countryside between Grosseto and Roselle.

2.2 Vertical air photography

As usual we began our work by examining the oldest available aerial photographs, in this case from the national coverage of 1954. Unfortunately no features were visible at Aiali because the area was at that time used for olive cultivation. While, the land-use changed to grain cultivation between the 1950s and 1970s, we still saw no useful information on vertical photographs from 1976, 1996 and 2001 (Campana et al. 2006b).

2.1 Satellite imagery

At the end of spring 2002 we planned the acquisition of 70 km2 of Quickbird-2 imagery around Aiali, ordering both multispectral and panchromatic data. Unfortunately Digital Globe only acquired the data on 13 June, at the very end of the intended time window. For this reason, perhaps, satellite imagery failed to make a meaningful contribution at Aiali itself, leading us to feel that without a multitemporal approach the acquisition of satellite imagery (and of airborne remote sensing data generally) will not necessarily be effective for intra-site analysis. However, it may still be useful for large-scale landscape studies. On the same image set, for instance, we can identify many other archaeological features, sometimes very close to Aiali. One such site lies only 800 m to the southeast, where field-walking

2.3 Aerial survey and oblique photography

As noted above, Aiali was first detected in early summer 2001 during oblique aerial survey. Because of logistical problems there was no opportunity to survey the site again from the air until the spring of 2004 (Campana et al. 2006b). In that year, however, the site was monitored at intervals of between two and four days from the air from the end of May to the middle of June — that is, throughout the cropripening season — to record the development of the

Figure 3 – GPS mapping of very high density artifact scatters.

38

Campana, From space to place or from site to landscape? Mind the gap

different morphologies. There are also some vessels that imitate ARSW forms, but their distribution at Aiali coincides exactly with that of the ARSW itself. The pottery analysis reveals the abandonment of the villa at the end of the 6th or the start of the 7th century, followed by a gap of about three centuries. During the late 9th and into the 10th century there are some vessels which reveal a re-occupation of the site. This new phase does not last beyond the 12th century and seems to be concentrated in the area of the villa structures. In summary, we can say that the systematic collection within georeferenced grids has been extremely useful in seeking to understand long-term topographical transformations within the site.

cropmarks. This procedure allowed us to identify traces that had not been visible before. Photography was continued into the late summer, and revealed new evidence of an abandoned river-course and two structures in the northwestern part of the field, adjacent to the main complex of buildings (fig. 4). The site was monitored from the air again in 2005 but the maize crop proved unhelpful; a similar situation applied in 2006 when the field was not under cultivation. The field returned to grain in 2007 and allowed the recovery of some further information in the western and northern parts of the area. 2.4 Field-walking survey and pottery analysis

During 2004 we carried out field-walking survey and surface collection within a 10 m grid. This confirmed the character and interpretation of the archaeological site as seen from the air, and demonstrated a high level of correspondence between the aerial evidence and concentrations of archaeological finds (fig. 3). Analysis of the large amount of pottery recovered from the site permitted us, at this stage of the work, to define the chronological range of the villa. Some 1100 fragments of pottery were catalogued by Dr. M. Ghisleni. The artefacts recovered from the grid collection were catalogued and drawn to identify typological parallels with which to assign a date to each single fragment. Each pottery fabric has been analysed under the optical microscope to identify characteristic mineral inclusions pointing to the shorter or longer distance trade in products. This analysis confirmed, for instance, the presence at Aiali of medieval pottery from Pisa, which contains sand rich in micro-fossils from the Arno. For the Imperial period it was possible to define pottery from Africa, in particular amphora body-sherds, by its characteristic mineral inclusions. The settlement began as an imperial Roman villa in the 1st century AD. The best marker for the early imperial period is italic red slip ware. A substantial presence of African red slip ware of types A, C and D allowed us to estimate that the villa continued until the 6th century AD, though with reduced amounts of African imports from the last quarter of the 3rd century. An interesting feature is the total absence of North African pottery from the 4th and 6th centuries within two adjacent rectangular structures in the north-western part of the field. The study of the pottery assemblage as a whole, however, shows that fine red-painted pottery (ARSW) produced locally or in the region in this same period was very much present, with a minimum of 101 vessels of many

2.5 Magnetic investigations

In the autumn of 2004 we collected 2 hectares of gradiometer data at intervals of 0.5 m along profiles set 1 m apart. One series of magnetic anomalies closely match the traces visible on the oblique air photographs (fig. 4); another group of anomalies fills in many of the ‘gaps’ in the main building complex. The central part of the villa consists of a rectangular structure measuring about 70 × 25 m, oriented northeast/southwest, at each end of which are four square rooms of 10 × 10 m across. The magnetic signal from a disused iron pipe obscures the archaeology along its length without reducing the general readability of the data. On the evidence of the aerial photographs, which do show continuity across the pipeline, we can assume that the belowground archaeological deposits are essentially undisturbed. This is confirmed by the GPR survey data discussed below. It is fair to suggest that in the absence of the pipe the gradiometer data would have produced equally positive results. Further magnetic anomalies can be seen in various other previously blank parts of the field. Some tens of metres to the northeast and southeast of the main complex, a series of linear anomalies, more or less aligned with the main structure, probably represent an enclosure, perhaps with an entrance-way. To the north there is another weaker anomaly (from –10 to +10 nT/m), approximately rectangular but not aligned on the villa and thus presenting problems of interpretation. Both the magnetic and aerial surveys produced poorer results in the northwestern part of the field, where the ground survey and gridded surface collection yielded considerable amounts of structural material and ceramics. 39

Hidden Landscapes

Figure 4 – Top left: archaeological interpretation of the features visible from the air. Top right: archaeological interpretation of three different sets of magnetic data — fluxgate, overhouser and caesium. Bottom left: archaeological interpretation of the GPR survey. Bottom right: interpretation of the archaeological features visible in the fast electrical imaging data set.

2.6 GPR investigation: data acquisition and processing

the data can be processed to generate horizontal maps (so-called time slices) of the recorded radar amplitudes at various time/depth windows across the recorded radargram dataset. These time slices can provide information regarding the size, shape, location and depth of subsurface archaeological structures. Time slice data at Aiali were created using the spatially averaged square wave amplitudes of the return reflections, which were then gridded using a Kriging routine (Piro et al. 2003, Goodman et al. 2001). The spatially averaged time slice parameter minimizes the effects of line striation parallel to the profile direction. The larger time window helped to create an understandable visualization of the subsurface features. In particular, when structures are not level within the ground or when there are significant speed variations across the site, the larger time window can capture continuous features at variable depths (Conyers & Goodman 1997). Other line noises, parallel to the profile collection direction, were removed using a moving filter with customized

In 2006 a high-resolution GPR survey was conducted (by Salvatore Piro, ITBC CNR Rome and Dean Goodman, Archaeometry Laboratory, Los Angeles, USA) over four areas at Aiali to test the potential of this technique. A GSSI SIR3000 instrument equipped with a 400 MHz bistatic antenna with constant offset was used for the measurements. In each area radar profiles at 0.5 m spacing were collected alternately in reversed and unreversed directions across the survey grids. Radar reflections along the transects were recorded continuously across the ground at 40 scans per second. Markers were spaced every 1 m along each profile to provide spatial reference. The gain control was manually adjusted to be more effective. The data were later corrected for speed variations to a constant 30 scans/m (or 1 scan per approximately 0.03 m). All radar reflections within a 50 ns (two-way travel time) time window were recorded digitally in the field as 16-bit data and 512 samples per scan. When GPR profiles are collected along closely spaced profiles 40

Campana, From space to place or from site to landscape? Mind the gap

threshold settings. Filter thresholds were set to signal levels just below the average reflections from buried Roman walls.

reveals all of these characteristics. In our current case, for instance, three groups of buildings are readily distinguishable on the geoelectrical map (fig. 6): west (a), south-east (b) and north (c). The availability of magnetic data for the whole of the field allows us to establish that the large building complexes (a) and (b), which have a much lower magnetic signal, are made of different materials from the large building (c). Further differences are revealed by a comparison between the radar data, the magnetics and the geoeletrics. In the case of complex (b) the geoelectrical data, so effective in uncovering the geometry of complexes (a) and (c), does not produce clear results. Similar difficulties can be seen in the magnetometer measurements, while a much stronger response is provided by the radar signals (fig. 4, interpretation of radar maps). These different responses perhaps reflect different structural characteristics in the buried deposits. It is worth observing in this context that the overall plan clearly shows three building complexes situated close to one another but apparently with quite different plan-forms and with variable orientations.

2.7 Automatic Resistivity Profiling ARP©

In autumn 2007 a 12 hectare area around the site was surveyed using the fast electrical imaging system ARP© developed in France by Michel Dabas (Dabas, forthcoming). The ARP© system was first designed for agricultural applications in 2001 (by the GEOCARTA company, a spin-off from CNRS) and relies on the standard galvanic electrical method, used widely for different applications since its discovery by Marcel and Conrad Schlumberger in the 1930s. It was not until 2004, however, that the system was released for archaeological surveying, which requires higher positional and measurement accuracy (Dabas 2008; fig. 5). The results are striking: new and extremely significant information was added to the western side of the surveyed area, and could be integrated with the existing results from other techniques (fig. 5, bottom right). The machine also produced a high resolution DTM of the surveyed area. 2.8 GIS-based archaeological interpretation

Bearing in mind the preliminary stage of the research we do not want to attempt a definitive interpretation but we are confident enough to outline a few archaeological and methodological issues. Firstly, we want to stress the very clear difference in the map between the amorphous concentrations provided by the field-walking surface-collection data and the integrated remote sensing results (in white in fig. 6). The integration of the various techniques enriches the resulting information both quantitavely and qualitatively. Figure 4 shows how the comparison between the various methods makes it possible in some cases to recognize elements that are not (or not clearly) revealed by other survey techniques. In these cases the inherent differences between the various techniques produce a quantitative enrichment in the representation of the buried evidence as seen in the composite map (fig. 6). It is important to appreciate that, where the anomalies seem to duplicate one another, the ‘redundancy’ between the different sources of information is more apparent than real. The integration of the geoelectrical methods, radar and magnetometry makes it possible to acquire information on the geometrical pattern, depth and even some of the chemical or physical properties of the buried features, where no single technique

Figure 5 – The Automatic Resistivity Profiling system at work on the Aiali site.

Figure 6 – Integrated interpretation. Notice the striking difference between the GPS survey of the main artifact scatters (grey areas) and the integrated remote sensing datasets.

41

Hidden Landscapes

In conclusion we should recognize that each technique has produced a remarkable increase in the quantity and quality of the available archaeological information. In particular the use of geophysical survey represents an advance of great importance. It could also be argued that without field-walking and grid collection it would not have been possible to achieve the feedback necessary to create a convincing archaeological interpretation of the site in terms of its dating and development over time.

Secondly, if it is difficult, from this discussion, to deny the obvious contribution of the ground-based and remote-sensing techniques, the importance of the archaeological field data is equally undeniable. As we have seen, the surface material was collected and recorded in this case within a predetermed grid, making it possible to overlay on the remotely sensed data a distribution map of the finds concentrations that offer clues to the chronological and even functional phasing of the buildings. The finds distribution demonstrates the occupation of the site as a whole, and of the three building complexes, from the early imperial period to the high medieval period, with an apparent hiatus (perhaps attributable to problems in the recognition of ceramic material from this particular phase) in the 7th and 8th centuries. The gridded collection and the resulting quantitative analysis of the finds distribution, however, allow us to suggest a more intensive early Imperial occupation of building complex (a), while from the 2nd-3rd century AD the centre of greatest concentration falls on complex (c) and in the medieval period on complex (b). This differentiation is supported by the differing plan-forms of the various complexes as outlined by the remotely-sensed data. The layout of complex (a) has a local parallel with buildings of the late republican/early imperial period (Campana et al. 2009). The four square corner structures at building complex (c) might perhaps be interpreted as towers of exactly the kind found in some ‘fortified’ late antique italic and thracian residential complexes. Without resorting to excavation it is impossible to ascribe a specific data to the four towers, which could in reality form part of a single building project along with the rest of the villa. At this stage in the research, given their identical dimensions, it seems fairly clear that they were built within a single phase, but we do not know whether they formed part of the original layout of the villa or were added at a later date. Villas with towers are already present in the early Imperial period but became more widespread in late antiquity, not only, perhaps, in response to militarisation within the countryside but also as an indication/symbol of the social rank of the owner (Sfameni 2005). Complex (b) is most clearly defined in the radar data which, amongst a variety of other indications, reveal the only instance on the site of a possible apsidal structure. The association between the prevalence of medieval material in this area and the presence of an ‘apse’ suggests the existence of a religious structure.

3. Mind the gap In conclusion we believe that on a large, complex and stratified archaeological site such as Aiali the contribution of an integrated approach is extraordinarily effective. The contrast between the data gained from field-walking alone and that achieved through the combination of various prospecting methods within a GIS environment is extraordinary, demonstrating — even, we hope, to the most skeptical of archaeologists — the extremely important role of remote sensing in the archaeological process (fig. 6). The unavoidable question arising from this case study appears to be: where would archaeology be without remote sensing when one needs to know about a buried site? But today we believe the question and the challenge should really be: when one needs to know about buried landscapes, where would archaeology be without remote sensing? Our next challenge should be to extend this kind of research to the broader landscape in the valley between the towns of Grosseto and Roselle (Gaffney & Gaffney 2006; Powlesland 2008). We have just began a PhD study in this area, with the aim of collecting about 600 hectares of magnetic data. The same strategy has been followed during the ARP© survey, connecting Aiali to a location about 800 m away, where aerial reconnaissance and field-walking survey allowed us to identify a medieval mound or castle. Between the two sites the ARP© survey shows many features invisible on the ground and in some cases also invisible (so far) through aerial photography. It is easy to understand, however, that it will never be possible to extend this approach to areas of the scale covered by the University of Siena’s archaeological mapping projects for the region as a whole. For instance, the archaeological map of the province of Grosseto covers an area of 4030 km2 and that of Tuscany as a whole about 22,990 km2. The archaeological objectives and outcome of the Aiali project depend on a critical assessment of the 42

Campana, From space to place or from site to landscape? Mind the gap

kinds of information that are available for recording: in assessing the potential or interpretation of a landscape it is at least as important to know what may not be visible as to appreciate what is visible. Selecting suitable sample areas for the application of the highest level of research intensity and archaeological visibility will, in our view, produce better returns: • Conservation: increased awareness of the archaeological resource as a whole, so as to create more effective and better-adapted policies for landscape monitoring and conservation. • Academic issues: the recognition of ‘emptiness’ (that is, the absence of evidence) as being equally important as the presence of archaeological evidence, in what we hope will be a new approach to research into the development of settlement patterns and landscape history. • The future: the search for means to achieve better visibility in our present ‘emptinesses’ will hopefully produce new and perhaps different kinds of data, in turn creating new kinds of feedback in the investigative and interpretative process.

The challenge we now face is fundamentally linked to the scale of visualization, but also to a changing approach to landscape studies. The huge archive collected in recent decades by the University of Siena, combined with these recent developments in improving the variety and intensity of research, are still not enough to answer our long-term needs because they will allow us to achieve only unconnected ‘islands’ of knowledge (fig. 7). We will probably have to reduce the size of our study areas, moving from the regional scale to more ‘human’ and continuous landscapes of limited size, without abandoning our broader interpretative and theoretical framework. We have to accept and respond to the inherent complexity of the record, not only in our research designs but also in the composition of our research teams, in the methods that we use in the field and in our data processing in the office. We have to develop new strategies aimed at implementing a ‘total archaeology’ approach to our chosen study area, so as to understand through real archaeological evidence the meaning of the

dot dot

dot dot

Figure 7 – Landscape between Grosseto and Roselle and related areas surveyed with very high intensity and integrated methods by the University of Siena. It can be seen that there is a huge loss of information in the landscape between ‘dots’ (sites); lacking this ‘connective tissue’ we lose any evidence of relationships.

43

Hidden Landscapes

‘emptinesses’ caused by limitations of visibility or accessibility within the framework of the landscape. Our current research strategy already includes, in addition to traditional approaches and multiple prospecting methods, both minimalist and openarea archaeological excavation. These are aimed in the first instance at obtaining a better understanding of ‘central places’ but (we trust) in the future also of field patterns, communication systems and the like. Our research must attempt to reproduce the complexity of past landscapes with their central places, settlements, sacred sites, cemeteries and industrial areas, etc. but also with their ‘connective tissue’, their spatial relationships, their physiographic peculiarities and their patterns of development over time.

Dabas M., forthcoming. Theory and practice of the new fast electrical imaging system ARP©, in Campana S., Piro S. & Francovich R. (eds), Seeing the unseen archaeology. Geophysical prospection for landscape archaeology, proceedings of the XVth International Summer School in Archaeology, Taylor and Francis. Doneus M. & Brise C. 2006. Full-waveform airborne laser scanning as a tool for archaeological reconnaissance, in Campana S. & Forte M. (eds), From Space to Place, Proceeding of the IInd International Conference Remote Sensing Archaeology, Rome, 4-7 December 2006, BAR International Series 1568, BAR Publishing, pp. 99-105. Finzi E., Francese R.G. & Morelli G. 2005. High-resolution geophysical investigation of the archaeological site “Le Pozze” in the surroundings of the town of Lonato (Brescia, Northern Italy), in Piro S. (ed.), Proceedings of the 6th International Conference on Archaeological Prospection, Rome, pp. 215-219. Gaffney C. & Gaffney V. 2006. No further territorial demands: on the importance of scale and visualization within archaeological remote sensing, in From artefacts to anomalies: Paper inspired by the contribution of Arnold Aspinall (University of Bradford 1-2 December 2006), www.brad.ac.uk/archsci/conferences/aspinall/presentations/. Goodman D., Nishimura Y. & Piro S. 2001. High resolution ground-penetrating radar, in Gaffney et al. (eds), Forum Novum-Vescovio: studying urbanism in the Tiber valley, Journal of Roman Archaeology 14, pp. 5-8, 20. Guaitoli M. (ed.) 2003. Lo sguardo di Icaro. Le collezioni dell’Aerofototeca Nazionale per la conoscenza del territorio, Campisano Editore, Rome. Holden N., Horne P. & Bewley R. 2002. High resolution digital airborne mapping and archaeology, in Bewley R.H. & Rączcowski W. (eds), Aerial Archaeology: Developing Future Practice (NATO Science Series I, Life and Behavioural Sciences 337), pp. 173-180. Musson C., Palmer R. & Campana S. 2005. In volo nel passato. Aerofotografia e cartografia archeologica, All’Insegna del Giglio, Florence. Piro S., Goodman D. & Nishimura Y. 2003. The study and characterization of Emperor Traiano’s Villa (Altopiani di Arcinazzo, Roma) using high-resolution integrated geophysical surveys. Archaeological Prospection 10(1), pp. 1-25. Powlesland D. 2006. Redefining past landscapes: 30 years of remote sensing in the Vale of Pickering, in Campana S. & Forte M. (eds), From Space to Place, Proceeding of the IInd International Conference Remote Sensing Archaeology, Rome, 4-7 December 2006, BAR International Series 1568, BAR Publishing, pp. 197-201. Powlesland D., forthcoming. Why bother? Large scale geomagnetic survey and the quest for “Real Archaeology”, in S. Campana, S. Piro & R. Francovich (eds), Seeing the unseen archaeology. Geophysical prospection for landscape archaeology, proceedings of the XVth International Summer School in Archaeology, Taylor and Francis. Sfameni C. 2005. Le villae-praetoria: i casi di San Giovanni di Ruoti e di Quote San Francesco, in Volpe G. & Turchiano M. (eds), Paesaggi e insediamenti rurali in Italia meridionale fra tardoantico e altomedioevo, Bari, pp. 609-622.

References Campana S. 2004. Le immagini da satellite nell’indagine archeologica: stato dell’arte, casi di studio, prospettive, in Archeologia Aerea. Studi di Aerotopografia Archeologica 1, Istituto Poligrafico e Zecca dello Stato, Roma, pp. 279-299. Campana S. 2006. DGPS e mobile GIS per l’archeologia dei paesaggi, in Laser scanner e GPS: paesaggi archeologici e tecnologie digitali 1, proceedings of the workshop of 3 March 2005 at Grosseto, All’Insegna del Giglio, Firenze, pp. 201-225. Campana S., Francovich R., Vaccaro E., Frezza B. & Ghisleni M. 2005. Il popolamento tardoromano e altomedievale nella bassa valle dell’Ombrone. Progetto Carta Archeologica della Provincia di Grosseto, in Archeologia Medievale XXXII, pp. 461-480. Campana S., Francovich R., Corsi M. & Pericci F. 2006a. Aerial Survey Project in Tuscany: years 2000-2005, in Campana S. & Forte M. (eds), From Space to Place, Proceedings of the IInd International Conference on Remote Sensing in Archaeology, Rome, 4-7 December 2006, BAR International Series 1568, BAR Publishing, pp. 497-503. Campana S., Felici C., Ghisleni M. & Piro S. 2006b. From Space to Place: the Aiali project (Tuscany-Italy), in Campana S. & Forte M. (eds), From Space to Place, Proceeding of the IInd International Conference Remote Sensing Archaeology, Rome, 4-7 December 2006, BAR International Series 1568, BAR Publishing, pp. 131-136. Campana S., Donoghue D. & Galiatsatos N. (forthcoming). Undercanopy archaeology using airborne laser scanner to overcome the Mediterranean vegetation, Archeologia Aerea III. Campana S. & Francovich R. 2007. Understanding archaeological landscapes: steps towards an improved integration of survey methods in the reconstruction of subsurface sites in South Tuscany, in Wiseman J.,el Baz F. (eds), Remote Sensing in Archaeology, Springer, Boston, pp. 239-261. Conyers L.B. & Goodman D. 1997. Ground penetrating radar. An introduction for archaeologists. AltaMira Press. Cosci M. 2005. Ricerche aeree in Toscana, in Musson C., Palmer R. & Campana S. (eds) 2005. In volo nel passato. Aerofotografia e cartografia archeologica, All’Insegna del Giglio, Florence, pp. 263-271.

44

Campana, From space to place or from site to landscape? Mind the gap

Acknowledgments The author owes an enormous debt of gratitude to the late Professor Riccardo Francovich, of the University of Siena, for his interest, help and criticism throughout all stages of the project. Special thanks are also due to Chris Musson for help with the English version of this text and for his valuable comments on the practice of archaeological research. Bob Bewley (English Heritage), Daniel Donoghue (University of Durham), Dean Goodman (GAL-USA), Darja Grosman (University of Ljubljana), Salvatore Piro (ITABC - CNR) and Dominic Powlesland (Landscape Research Centre, UK) have helped in a variety of ways. Helmut Becker (formerly of the Bavarian State Department of Historical Monuments, Germany) and Iacopo Nicolosi (National Institute of Geophysics and Vulcanology, Italy) contributed greatly in the field but also helped me to understand the best configuration for our magnetometer work. Salvatore Piro (ITABC-CNR Rome) and Dean Goodman (Geophysical Archaeometry Laboratory, Los Angeles, USA) did great work in GPR data acquisition and processing. Many researchers and students have collaborated, and are still collaborating, in the Siena and Grosseto archaeological mapping projects. Special thanks are also due to the team of the Laboratory of Landscape Archaeology and Remote Sensing at the University of Siena: Anna Caprasecca, Cristina Felici, Barbara Frezza, Mariaelena Ghisleni, Francesco Pericci and Emanuele Vaccaro.

45

6 Magnetic susceptibility prospection in northern Calabria Alette KattenbergI & Martijn van LeusenII I ArGeoLogic Geophysical and Geoarchaeological Consultancy, Stangobreck, Holm, Orkney KW17 2RX (UK) – [email protected] II Groningen Institute of Archaeology, Poststraat 6, 9712ER Groningen (NL) – [email protected]

1. Introduction

In 2006, geophysical field surveys were conducted both to establish the response on and around known sites, and to create a data set representative of conditions in the hinterland.

Archaeologists have investigated large swaths of the Mediterranean landscape, mainly using ‘topographical’ and field walking approaches. These methods work well where the archaeological record is easily accessible, but they fail in areas that are difficult of access, not ploughed, or covered by recent sedimentation. In the Groningen Institute of Archaeology (GIA) surveys, conducted since 2000 in the watershed area of the Raganello river (northern Calabria, Italy), field walking has indeed been very successful in the foothill zone, with mostly arable land use, but has met with trouble both in the coastal plain and in the uplands and mountains of the hinterland. In the coastal plain marine and alluvial sedimentation have deeply covered remains of the Roman period and before: the famous Greek city of Sybaris, for example, now lies some 6 meters below the surface. In the hinterland, arable fields are scarce, large areas are covered by forest, and the mountains cannot be surveyed by standard methods.

2.1 Magnetic susceptibility survey using the Bartington MS2D

Figure 2a shows the magnetic susceptibility of a set of samples that was taken on a transect across one of the archaeological sites under investigation (the small Bronze Age site ‘4120’). Samples 1, 2, and 7 are taken off-site, samples 3, 4 and 5 on-site, and sample 6 in the site halo. There is a clear enhancement in the magnetic susceptibility of the on-site soil samples. Based on these and similar results from other archaeological sites, magnetic susceptibility was expected to be a suitable candidate variable for the geophysical detection of small archaeological sites. Figure 2b shows the results of the topsoil magnetic susceptibility survey on site ‘4120’. Clearly, the observed distribution of archaeological material in the site core (dense hatching) and the site halo

2. Geophysical survey Poor visibility in the hinterland makes the study area often unsuitable for field walking surveys. An investigation into the application of geophysical methods in this area was started in 2006, in order to find a method that could map archaeological sites in areas with poor visibility, and that could be used alongside the field walking survey at a similar speed. The field methodology aimed to resolve two issues. First, can the known small archaeological sites be reliably detected, using geophysical prospection methods similar in speed to the field walking survey? Second, can ‘false positives’ caused by local geological features be reliably distinguished from true archaeological sites? In preparation for the geophysical field surveys, soil samples were collected in 2005 and measured on an Agico KLY-2 magnetic susceptibility bridge in the Palaeomagnetic Laboratory of the University of Utrecht (the Netherlands). A clear magnetic susceptibility contrast was observed between on- and off-site samples from the uplands and the highlands (fig. 2). Samples from the marine terraces area, on the other hand, did not appear to have an archaeologically relevant magnetic contrast.

Figure 1 – Location of the study area, divided into coastal and alluvial plains (a), the foothill zone (b) and up- and highlands (c).

47

Hidden Landscapes

a

b

Figure 2 – Topsoil magnetic susceptibility survey of the small Bronze Age site ‘4120’ compared with archaeological survey. a. Magnetic susceptibility profile across the archaeological distribution: samples 1, 2, and 7 are taken off-site, samples 3, 4 and 5 on-site, and sample 6 in the site halo. b. Distribution of archaeological material in the site core (dense hatching) and halo (wide hatching).

a test, surveying fields totaling 3 hectares in the vertical dipole mode, with an interline spacing of 5 meters; 6600 data points were collected in 4 hours. Unfortunately the variations in magnetic susceptibility (in ppt) do not appear to reflect the archaeological situation as it is known from the fieldwalking survey, but are likely to be dependent on the geological background instead. The peak sensitivity of the EM38B in the ‘horizontal’ mode is assumed to be at 25 cm below the instrument, with a total penetration of 75 cm. In ‘vertical’ mode, the highest sensitivity is achieved at 40 cm depth, with an overall penetration of 75 cm (Berle Clay 2006). This is much deeper than the Bartington MS2D, where 90% of the signal is obtained from the first 6 cm below the instrument (Lecoanet et al. 1999), and suggests that the two techniques should be viewed as complementary. The magnetic susceptibility enhancement in the center of the Bartington MS2D dataset (figure 3, top) coincides with the location of the superficial material that makes up the archaeological site. In the EM38B response (figure 3, middle and bottom), the contribution of this enhanced material to the signal is clearly minimal, suggesting that the enhancement is limited to a thin plough-soil layer at the very surface: the EM38B cannot ‘see’ it. Without further (intrusive) archaeological investigations the magnetic susceptibility variations in the EM38B data cannot be interpreted.

Figure 3 – Schematic representation of the three ‘depth slices’ that were investigated at site 5104. Top: Bartington MS2D (6 cm), range 13.188.1 SI units; middle: EM38B horizontal dipole (25 cm), range 0.5-0.9 ppt; bottom: EM38B vertical dipole (40 cm), range 1.0-1.5 ppt.

(wide hatching) is reflected in the topsoil magnetic susceptibility. In general the Bartington MS2D data corresponds very well to the archaeological record as it is known from the fieldwalking surveys. Both the MS2D survey and field walking rely on the presence of superficial archaeological material. When used in conjunction a realistic picture can be obtained of archaeological sites at the present surface, both in areas with good and with poor visibility.

3. Conclusions and further work

2.2 Covering larger areas at speed: Geonics EM38B

The Bartington MS2D magnetic susceptibility meter was very successful in detecting those small archaeological sites that had already been discovered during the field walking survey. The information on

Large areas can be covered relatively quickly when using the EM38B in conjunction with the Allegro datalogger and a GPS unit. We have conducted 48

Kattenberg – van Leusen, Magnetic susceptibility prospection in northern Calabria

the location, size and shape of the archaeological sites under investigation appears to be almost identical in both datasets, and in some cases coincides with the darker soil colour associated with ploughedup humic layers. Used in conjunction, topsoil magnetic susceptibility surveys covering areas with poor surface visibility can supplement traditional field walking in areas with good or fair visibility, and a comprehensive picture of the superficial archaeological record within the study area can be obtained. The speed of investigation of the EM38B at present is much higher than that of the Bartington MS2D. A GPS based datalogger for the Bartington MS2D is currently being developed, and is expected to greatly improve coverage in future experiments. The tests with the EM38B have yielded magnetic susceptibility data from larger soil volumes than the Bartington MS2D. As the field walking surveys yield only information about the archaeology of the ploughsoil, the relation between subsoil magnetic susceptibility variations and the archaeological record can not be established until permission has been obtained for on-site coring and test trenches. Further study is also required to understand the effects of the highly variable local geology and geomorphology on all data collected in 2006.

References Berle Clay R. 2006. Conductivity survey: a survival manual, in Johnson J.K. (ed.), Remote sensing in archaeology: an explicitly North American perspective. The University of Alabama Press, Tuscaloosa. Lecoanet H., Lévêque F. & Segura S. 1999. Magnetic susceptibility in environmental applications: comparison of field probes, in Physics of the Earth and Planetary Interiors 115, pp. 191-204.

49

7 Integrating digital aerial photogrammetry and geophysical prospection in the Gargano flint mining landscape (south-eastern Italy) Attilio GalibertiI, Riccardo SalviniII, Massimo TarantiniI, Fabio MantovaniII, Marta BottacchiII, Manuela LinoII, Claudio RossiIII, Massimiliano MondetIII I Department of Archaeology and Arts History, Prehistory Section, University of Siena, via Roma 56, 53100 Siena, Italy – [email protected]; [email protected] II Earth Sciences Department and Centre of Geotechnologies, University of Siena, via Vetri Vecchi 34, 52027 San Giovanni Valdarno, Italy – [email protected]; [email protected]; [email protected]; [email protected] III SGG – Studio di Geologia e Geofisica srl, Strada Massetana Romana 56, 53100 Siena, Italy – [email protected]; [email protected] Abstract The aim of this paper is to illustrate the results of a national research project devoted to studying both features of a prehistoric flint mining landscape and underground mines not visible, or partially visible, on the surface. The geographical context of the project is the promontory of Gargano (south-eastern Italy), where an extraordinary abundance of flint is associated with evidence of numerous extraction complexes (underground mines). According to the available data, the mines cover a time-span of almost three thousand years; the most important of them is the Defensola A mine, the oldest in Europe and active from roughly 5800 BC onwards. Unlike other European prehistoric mining complexes, mine entrances in Gargano are generally located on convex slopes, and have often produced extremely large cone-shaped spoil heaps outside. This suggested a method of investigation based on digital aerial photogrammetry. A DGPS survey was executed to measure ground control points, to be used for the external orientation of the photographs, as well as to obtain precise positions of the main known mines. Thanks to the stereo-interpretation it was possible to recognize key geological beds, such as limestone rich in flint, and possible ancient entrances to mines. Sometimes anthropogenic covers (dumps and alluvial fans) near the mines were also recognized and interpreted as evidence of quarrying activity. A second level of investigation used geophysical surveys. These can determine the internal layout of each mining structure of which only the entrance is known. Moreover, geophysical prospection allowed us to verify the presence of new mining structures in cases where one or a few vertical shafts are present, which generally indicates the presence of larger mining complexes. The archaeological features of the area suggested to perform additional surveys using other geophysical methods: resistivity and ground penetrating radar. This multi-method approach allowed us to check the data from several independent measurements and to increase the amount of geometrical and physical information that can be used for the interpretation.

1. The geological and chronological context

ic deformation has been weak, the flint contained in these layers is often unjointed. Hitherto, 18 complexes for flint extraction have been identified on the basis of their original entrances or in the course of earth removal for building or road construction purposes. Many mines still present problems with regard to their chronological and/or cultural attribution; however, in recent years their context has been defined more precisely (Galiberti 2005). Nine complexes have been attributed to specific cultural or chronological contexts. We can now date the beginnings of extraction activities to the early 6th millennium BC (Defensola A: Utc 1342 6990±80 BP, cal. 2σ 5993-5652 BC), coinciding with the neolithization of south-eastern Italy); mining terminates at the end of the Copper Age (Defensola B: Beta 171597 4050±40 BP, cal. 2σ 2850-2820, 2670-2470 BC).

The promontory of Gargano is located in the southern part of the foreland area of the Apennine-Dynaric system (Funicello et al. 1992; Patacca & Scandone 2004) in eastern Apulia (south-eastern Italy). The geology of the area (fig. 1) is characterized by a cyclical sequence of carbonatic platform sediments (tidal and subtidal flats) and, in the eastern peripheral part, by basin carbonatic limestone that underwent a sedimentation process compensated by tectonic subsidence (Bosellini et al. 1993). From the Neogene, the area was affected by an extended and continuous raising (Doglioni et al. 1996; Gambini & Tozzi 1996) caused by changed tectonic conditions in the Apennine thrusting phase. These processes were responsible for the present geo-structural setting, and also caused extensive karst phenomena (Pieri & Ricchetti 1999). At the top of the mesocenozoic geologic formations of the stratigraphic sequence, the carbonatic-selciferous lithological units of the Cretaceous-Eocene formations (particularly the carbonatic limestones called ‘Maiolica’, ‘Scaglia’ and ‘Peschici Limestones’) outcrop. Because tecton-

2. Mining types and methods Two categories of extraction sites can be identified in the Gargano area, both appearing to be closely linked to the geological setting of the formations 51

Hidden Landscapes

Figure 1 – Geological map with flint mines. 1. Quaternary; 2. Mio-Pliocene; 3. M. Saraceno sequence; 4. Scaglia; 5. Fucoidi Schists; 6. Mt. Acuto and Mt. S. Angelo Formations; 7. Cagnano Breccias; 8. Maiolica; 9. Mt. Sacro sequence. Geological data from Bosellini & Morsilli 2002.

ogy, and have often created large cone-shaped spoil fans outside. Our first level of investigation concerned the identification of specific geological features from the digital aerial photogrammetry stereo-restitution. Multiperiod aerial photos dating from 1954, 1992, 1996 and 2001 were analyzed within ERDAS™ Imagine 9.2 LPS and StereoAnalyst modules. A DGPS survey was conducted to measure both ground control points (GCPs, to be used to determine the external orientations of the photographs) and the precise positions of the main known mines. The GPS survey was executed with GPS Leica™ 1200 receivers both in the static and in the Real-Time Kinematic (RTK) differential modes; for the former, the IGM95 vertex 157701 located at Vieste was used (fig. 2), whose coordinates and description card were made available by the Italian Geographic Military Institute (IGMI). The RTK GPS survey was carried out through a mobile phone by connecting the rover receivers to a new reference station, purpose-built for the project.

involved. Nevertheless, the fact that sub-horizontal excavations in compact formations are the only type attested during the Neolithic, whereas vertical excavations and those in formations affected by tectonic activity belong exclusively to the height of the Copper Age, clearly suggests that social and cultural factors also played a role (Tarantini 2006). 2.1 Sub-horizontal quarries

Where the formations are compact, excavation proceeded by removing individual limestone layers in their entirety. Depending on the thickness of the layers, a sufficient number of these were removed to obtain a floor-ceiling height allowing excavation to proceed. This height rarely exceeds 60 cm. This excavation method uses the seams between layers as guidelines for their removal, thereby creating structures with a characteristic flat roof. Therefore, unlike other European prehistoric mining complexes, mine entrances in the Gargano (e.g. Defensola A) are often located on slopes with a convex morphol52

Galiberti et alii, Integrating digital aerial photogrammetry and geophysical prospection in the Gargano…

The survey produced more than 80 GCPs. Using VERTO™ software the WGS coordinates were reprocessed from ellipsoidal to orthometric elevation and projected onto the national Gauss-Boaga reference system. To these points all available remotely sensed data were referenced. On the basis of their stereo-interpretation it was possible to recognize key geological beds in limestones rich in flint, which were suspected to represent the locations of ancient mine entrances. Rarely, anthropogenic covers (dumps and alluvial fans) nearby those mines were also recognized and interpreted as evidence of quarrying activity (fig. 3). Because mines consist of one or more communicating chambers, with pillars of untouched rock and/ or heaps of excavation debris probably having been used as support structures, a second level of investigation used geophysical surveys to determine the internal layout of each mining structure for which the entrance was the only known part (e.g. Arciprete, Principe). The depth to which the interior of the hill has been excavated ranges from about 10 m in some complexes to over 100 m in Defensola A, and follows a ‘chamber and pillar’ layout. Their underground surface area may be as large as the 3500 sqm currently mapped for the Defensola A mine 1. The application of non-excavation methods is recognized as an important element in mining site research (Budziszewski 1990), and geophysical methods in particular have been successfully used in other Neolithic mining complexes (Borkowski et al. 1991; Herbich 1997; Leopold & Völkel 2004). The archaeological features of the area also suggested to perform a geophysical survey, in this case using different methods: resistivity and ground penetrating radar (GPR). This multi-method approach allowed us to check the data from several independent measurements and to increase the amount of geometric and physical information that can be used for the interpretation. Two-dimensional apparent resistivity investigations were performed by an automatic device (Abem™ Terrameter SAS 1000) that supports 64 electrodes. The Schlumberger acquisition method with a high n factor1 was chosen because it is sensitive to both horizontal and vertical structures. From this acquisition a pseudo-section, that is a two-dimensional image of the apparent resistivity variation subsurface, can be obtained. Subsequent computer modelling helped to interpret geo-electrical data by using inverse methods (Res2Dinv Geotomo™ software) to produce images of the subsurface resistivity

Figure 2 – DGPS measurement at IGM95 vertex 157701 (Vieste, S. Francesco church).

from the pseudo-sections. Two-dimensional Electrical Resistivity Tomography (ERT) survey provided information about the distribution of the resistivity in the ground (Loke 2004). Resistivity is a physical property of materials, which approximately describes the nature of each material; the resistivity of soils and rocks strongly depends on their water content, on the presence of certain minerals, and on porosity. For this reason the same rock type may display a wide range of resistivity values. The GPR survey was executed using a SIR-20 GSSI™ device with a 400 MHz antenna. GPR produces results similar to those from seismic reflection but at higher resolutions (Martinez et al. 1996). GPR is most frequently used to acquire 2D reflection profiles, by keeping the two GPR antennas an equal distance apart while taking measurements at equal spacing along a traverse (Davis & Annan 1989). The success of GPR is based on electromagnetic (EM) waves operating in the frequency range where displacement currents dominate and losses associated with conduction currents are minimal (Annan 1996). Since soil materials rarely have a magnetic permeability that is appreciably different from the unity (except for few magnetic minerals), the electromagnetic wave moves with a velocity depending 53

Hidden Landscapes

on the resistivity or the dielectric constant. Consequently, for many earth materials the velocity of an EM wave at high frequencies or high resistivity is determined only by the relative dielectric constant. The GPR data presented here were post-processed with RADAN software (GSSI™), and the results (in terms of intensity signal) are represented by contiguous wiggle trace curves.

which are mainly visible in roadside sections. However, on the basis of comparisons with other mining contexts in the area (Di Lernia & Galiberti 1993; Weisgerber 1999), this exploitation strategy was found to involve the digging of multiple shafts, often adjacent to one another, and sometimes providing access to single structures. The debris from these vertical quarries was partly used to fill nearby shafts or structures, partly distributed over the land surface in ways only observed in roadside sections, making its recognition difficult. The size of the areas covered by these complexes is difficult to determine by surface survey. Therefore, the aim of the project was to determine the extent of these areas by geophysical prospection and to verify the presence of new mining structures in cases where one or a few vertical shafts were present, generally as a part of larger mining complexes (e.g. Defensola B shaft, Vallone Due Ulivi).

2.2 Vertical quarries

Vertical access to flint-bearing formations is the only option at locations where the morphology is flat, as is the case in large areas of central Europe. Extraction using vertical shafts is also the only option in formations which are strongly affected by tectonic activity, where it would be impossible to safely carry out sub-horizontal excavations. Little information is available about this extraction method, in part because of the difficulty of interpreting structures

a

b Figure 3 – Ortophoto mosaic of the northern Gargano. Rectangle: Defensola mining area. a. Key beds in limestones; b. Debris fans identified at Defensola A mine 2 (Vieste).

54

Galiberti et alii, Integrating digital aerial photogrammetry and geophysical prospection in the Gargano…

Figure 4 – Defensola A. The four lines indicate the geophysical profiles. At top right, a sketch of the hillside and of the fan formed by dumped material. At the bottom, electrical tomography profile number 4 with the anomalous signal DA_T1 in dark grey.

3. Results

ERT line 3 starts from the upper terrace, perpendicular to lines 1, 2, the GPR profile and the fan, and measured down to a depth of 10 m. Whereas the terraces form a horizontal surface, the same homogeneous layer of limestone with high resistivity values (ρ > 200 Ωm) is present, and lower resistivity values are registered in the lower part of the section. In this tomography the presence of excavation rubble is indicated by higher resistivity values (ρ > 300 Ωm); probably the rubble fan is situated on top of a wider layer of limestone and is composed of the same material. All the data collected along line 3 can be interpreted in terms of the local geological framework, and there is no evidently anomalous signal that could indicate mining activity. ERT line 4 shows the same resistivity pattern: under a 3 m thick limestone layer (ρ > 200 Ωm), the presence of water in the porous material reduces the resistivity values for the investigated section up to a depth of 16 m (ρ < 60 Ωm). The only interruption in this homogeneous context is a spot of very high resistivity (ρ > 1000 Ωm) located 15 m underground (visible at the bottom of fig. 4). This confined area is

The results from three different sites investigated with resistivity and GPR techniques are here presented: Defensola A and B (municipality of Vieste), and Principe (municipality of Mattinata). 3.1 Defensola A (mine 2)

Figure 4 shows the resistivity and GPR profiles through the site. On the terraces at the top of the hill, three 2D electrical imaging surveys (length of line 1 = 50.4 m, line 2 = 63 m and line 4 = 94.5 m) and one GPR profile (along line 2 for a length of 97 m) were carried out.. Another 2D resistivity survey was performed along the hill’s slope, across the rubble fan (length of line 3 = 63 m). The ERT and GPR results on lines 1 and 2 show a homogenous layer of limestone rocks (ρ > 200 Ωm) up to a depth of about 3 m. From this layer up to the maximal depth (about 10 m), another homogeneous zone with low resistivity values (ρ < 60 Ωm) can be detected; this is probably caused by a larger water content in the porous material. On the basis of these data the presence of any significant cavities or chambers can be excluded. 55

Hidden Landscapes

located 56 m north-west of the area without vegetation, at a depth of 8 m from the terrace where the fan is situated.

aging surveys (length of line 1 = 31.5 m, line 2 = 31.5 m, line 3 = 31.5 m, line 4 = 31.5 m). In all the data sets a homogenous (humid) soil clearly appears; in the ERT data it shows up as a thin layer with resistivity values lower than 100 Ωm. The big flint nodule represented a good opportunity to test the instruments on this kind of rare target: it produced an impressive GPR signal (P1), and the resistivity exceeds 104 Ωm (DB_T1) on ERT E1 and E4. This extremely compact rock is set in a limestone layer of about 2 m thick, present in all of the investigated area.

3.2 Defensola B

The Defensola B site includes in one location a vertical shaft and a small tunnel that encloses a huge flint nodule with an estimated diameter of about 1.5 m (fig. 5). The area is fully vegetated and sometimes the bushes were a hindrance to the surveys. An area of about 800 m2 was investigated using eleven GPR profiles of various lengths and four 2D electrical im-

Figure 5 – Defensola B. Positioning of 6 GPR and 4 ERT profiles. DB_T1 indicates flint near the mine entrance; DB_T2 is the target confirmed by GPR profiles P4 and P6 and by electrical tomography line E4; DB_T3 is the complex target detected by GPR profiles P2 and P4 and by electrical tomography lines E2 and E4; DB_T4 indicates the target obtained by data from GPR profiles P3 and P5 and electrical tomography line E3; DB_T5 is an evidently anomalous signal in electrical tomography line E3.

56

Galiberti et alii, Integrating digital aerial photogrammetry and geophysical prospection in the Gargano…

Figure 6 – Pit at the Principe site. The vertical access shaft was recorded with lines 2, 3 and 4. The third electrical tomography line (bottom) shows the evident signal of the tunnel (in dark grey) departing from the pit.

DB_T4, at a depth of about 1 m, is a target confirmed by GPR profiles P3 and P5; its defined shape and its high resistivity value (ρ > 8000 Ωm) are difficult to interpret. Finally, at DB_T5 a peculiar target located 1 m underground was found, which produced the lowest resistivity value of the whole area (ρ < 60 Ωm); its geological interpretation is problematic because of its high and sharp contrast with the surrounding resistivity values.

The analysis of data from both methods concentrated on four anomalous zones (fig. 5). In DB_T2 a particular target located 1 m underground could be recognized that produced the highest GPR signal and had a resistivity value greater than 104 Ωm. Such a signal could be produced by a cavity, but as we have seen flint nodules also demonstrated very high resistivity, making their identification difficult especially when in contact with empty or partially filled tunnels. DB_T3 appeared as a more complex target: the intense reflector recorded on the crossing point of GPR profiles P2 and P4 is confirmed by high resistivity values near the surface (ρ > 3000 Ωm, 0.5 m) exactly on the crossing point between ERT survey lines E2 and E4. Below this, ERT revealed a spot with anomalously low resistivity value (80 < ρ < 100 Ωm). This problematic combination could be interpreted in many anthropogenic ways, but almost certainly it represents a significant geologic anomaly.

3.3 Principe

The geophysical surveys at the Principe site concentrated on the vertical entrance to an ancient mine. A square grid of 22 GPR profiles covering an area of about 400 sqm was created, and a 3D model (20 x 20 x 3 m) representing the most important reflectors was produced. The pit is located between the profiles 6, 7 and 18, 19. The four 31 m long 2D electrical imaging sections range to a depth of about 4.5 m (fig. 6). Results from both methods confirmed the hypothesis 57

Hidden Landscapes

Here, additional fieldwork and archaeological studies are needed.

of the archaeologist who explored the first part of the tunnel: there is a good reflector with an east-west orientation between 1.2 and 1.8 m below the surface, and the resistivity models show high values (ρ > 2000 Ωm) where the section intercepts this cavity. The mine seems not to keep to the same depth further from the shaft: at its eastern end the depth is almost 2 m. The geological context for this site is heterogeneous: the limestone is characterized by resistivity values higher than 100 Ωm and appears to be extremely fractured; there are areas where the water reduces the resistivity to 10 Ωm. These geophysical surveys allowed us to reconstruct the near-surface geology and the first part of the tunnels, but they did not give a complete overview of the network of cavities. Neither did the digital aerial images show any trace on the surface of the course of the tunnels as revealed by geophysical prospecting. Probably the tunnels continue beyond the investigated area, but to confirm this hypothesis additional geophysical studies will be necessary.

References Annan A.P. 1996. Transmission dispersion and GPR, J. Environ. Eng. Geophys 0(2), pp. 125-136. Borkowski W., Migal W., Salacinski S. & Zalewski M. 1991. Possibilities of investigating Neolithic flint economies, as exemplified by the banded flint economy, Antiquity 65, pp. 607627. Bosellini A., Neri C. & Lucani V. 1993. Guida ai carbonati cretaceo-eocenici di scarpata e di Bacino del Gargano (Italia meridionale). Annali Università di Ferrara, Sezione Scienze della Terra IV, p. 77. Bosellini A. & Morsilli M. 2002. Il Promontorio del Gargano. Cenni di geologia e itinerari geologici, Quaderni del Parco Nazionale del Gargano. Budziszewski J. 1990. Remarks on methods of studying prehistoric areas of flint exploitation, in Séronie-Vivien M.R. & Lenoire M. (eds), Le silex de sa genèse à l’outil. Cahiers du Quaternaire 17, Paris, pp. 217-223. Davis J.L. & Annan A.P. 1989. Ground-penetrating radar for high-resolution mapping of soil and rock stratigraphy, Geophysical Prospecting 37, pp. 531-551. Di Lernia S. & Galiberti A. 1993. Archeologia mineraria della selce nella preistoria. Definizioni, potenzialità e prospettive della ricerca. Quaderni Dipartimento di Archeologia e Storia delle Arti, Sezione archeologica, Università di Siena. Doglioni C., Tropeano M., Monelli F. & Pieri P. 1996. Middle Late Pleistocene uplift of Puglia: an “anomaly” in the Apenninic foreland. Mem. Soc. Geol. It. LI (1), pp. 101-117. Funiciello R., Montone P., Salvini F. & Tozzi M. 1992. Caratteri strutturali del Promontorio del Gargano, Mem. Soc. Geol. It. XLI, pp. 1235-1243. Galiberti A., 2005. Defensola. Una miniera di selce di 7000 anni fa, Siena. Gambini R. & Tozzi M. 1996. Tertiary geodynamic evolution of the Southern Adria microplate, Terranova VIII, pp. 593-602. Herbich T. 1997. The electro-resistivity survey of the “Krzemionki” mining complex, Poland. Problems and results, in Ramos-Millan A. & Bustillo M.A. (eds), Siliceous Rocks and Culture: Granada: Editorial Universidad de Granada, pp. 303-314. Leopold M. & Völkel J. 2004. Neolithic flint mines in Arnhofen, southern Germany: a ground-penetrating radar survey, Archaeological Prospection 11(2), pp. 57–64. Loke M.H. 2004. Electrical imaging surveys for hydrogeological, environmental and engineering studies. Geofluid, Piacenza. Martinez A., Feldman H.R., Kruger J. M. & Beaty D.S. 1996. Three-dimensional characterization of a fluvial sandstone reservoir analog in northeast Kansas using high-resolution ground-penetrating radar, Kansas Geological Survey Openfile Report 96-38, p. 30. Bosellini A. & Morsilli M. 2002. Il Promontorio del Gargano. Cenni di geologia e itinerari geologici, Quaderni del Parco Nazionale del Gargano. Patacca E. & Scandone P. 2004. The Plio-Pleistocene thrust belt-foredeep system in the southern Apennines and Sici-

4. Conclusions This project sprang from the conviction that an understanding of the mechanisms governing the process of acquiring raw materials, often difficult to document, represents a key element in the study of the economic, social and behavioural dynamics of prehistoric societies. In order to attain this purely archaeological objective of reconstructing aspects of human activity, different archaeological, geological and applied geotechnologies have been adopted. The use of an integrated approach represented a subsidiary methodological objective of the research project. The results from aerial digital photogrammetry and geophysical prospection demonstrate that the primary objective can only be achieved through an interdisciplinary approach, which has been tested for the first time in this area. The applied technologies have allowed us to increase the geomorphological know-how necessary to identify new sites, and to determine the extent of individual tunnel complexes whose archaeological aspects are still under investigation. Nevertheless, many ambiguities in the interpretation of the data remain, especially with regard to the geophysical results. It often proved impossible to interpret resistivity anomalies as either cavities, natural karst, or even filled-in tunnels. Targets that produced the lowest resistivity values could hardly be explained as geological phenomena, because they evidently contrast with the surrounding resistivity values. 58

Galiberti et alii, Integrating digital aerial photogrammetry and geophysical prospection in the Gargano…

ly (Italy), Geology of Italy, special vol. of the SGI for the 32 IGC, Florence. Pieri P. & Ricchetti G. 1999. Guide geologiche regionali, Puglia e Monte Vulture: Società Geologica Italiana, I parte. Tarantini M. 2006. Le miniere neolitiche ed eneolitiche del Gargano. Tecniche estrattive e dinamiche diacroniche, Atti della XXXIX Riunione Scientifica. IIPP, Firenze, vol. I, pp. 343353. Weisgerber G. 1999. 5000 Jahre Feuersteinbergbau. Die Suche nach dem Stahl der Steinzeit, Deutsches Bergbau-Museum, Bochum.

Acknowledgement The research presented here was part of the Progetti di Ricerca di Interesse Nazionale 2005: Geotecnologie applicate alla ricostruzione paleomorfologica e allo studio delle modalita di approvvigionamento e circolazione della selce. National co-ordinator of this program: Prof. Attilio Galiberti.

Note 1

For a Schlumberger acquisition the n factor is defined as: 1 AB 1 n
 2 MN

where AB and MN is the distance between the current’s electrodes and the voltage electrodes respectively. For these surveys n = 6 was used.

59

8 Grotta del Romito (Cosenza): latest Pleistocene landscape André Carlo ColoneseI, Massimiliano GhinassiII, Zelia Di GiuseppeI, Lisa GovoniIII, Domenico Lo VetroI, Giulia MalavasiIII, Fabio MartiniI, Silvia RicciardiIII, Benedetto SalaIII I Dipartimento di Scienze dell’Antichità ‘G. Pasquali’, Università degli Studi di Firenze, via S. Egidio 21, 50122 Firenze, Italy – [email protected]; [email protected]; [email protected]; [email protected] II Dipartimento di Geoscienze, Università degli Studi di Padova, via Giotto 1, 35137 Padova, Italy – [email protected] III Dipartimento di Biologia ed Evoluzione, Università degli Studi di Ferrara, Corso Ercole I d’Este 31, 44100, Ferrara, Italy – [email protected]; [email protected]; [email protected]; [email protected] Abstract The paper presents the results of the application and combination of several non-archaeological disciplines, such as sedimentaryfacies analysis and malacology, to analyze two separate phases in the Late Pleistocene sequence at Grotta del Romito (Papasidero, Cosenza). The lower part of the sedimentary sequence is divided into three main intervals (RM1-3) documenting two main episodes of underground streams deactivation. The first phase was influenced by the development of arid climatic conditions during the last glacial maximum. The second deactivation was probably induced by a rapid choking of the drainage channel caused by cave-wall collapses. This event allowed a prolonged human use of the cave, and finally stopped fluvial sedimentation. The vegetation structure near the cave during the Late Glacial episodes of human presence was reconstructed through analysis of terrestrial mollusc shell assemblages. Data suggest that, quantitative and qualitative changes in terrestrial mollusc assemblages were likely associated with a varying intensity of human impact on the vegetation near the cave during the deposition of the Late Epigravettian layers.

1. Introduction

2. The Romito cave: morphological and archaeological context

Most data on the Upper Palaeolithic and Mesolithic cultures of Southern Italy derive from cultural remains preserved in cave deposits. Due to the peculiarity of these environments, multidisciplinary studies attach great importance to the reconstruction of the landscape and of human impact on the environment. In order to show how non-archaeological disciplines such as sedimentary-facies and terrestrial mollusc analysis were applied and combined, the Late Pleistocene sequence of the Grotta del Romito (Papasidero, Cosenza) will be discussed, with particular emphasis on two main goals. The first goal was to achieve a better understanding of the landscape history of this region during the deposition of the first 3.5 m of the sedimentary succession (Gravettian and early Late Epigravettian). The second goal was to present an example of human interference with the environment during the accumulation of the topmost sedimentary sequence related to the Late Glacial (Late Epigravettian). Clastic sediments deposited in caves and rock shelters have peculiar sedimentological characteristics and have seldom been considered as a high-resolution proxy record of climatic or environmental change. In this study we demonstrate that local mechanisms can be also responsible for the observed sedimentary patterns at Grotta del Romito, which complicates the correlation with regional palaeoclimatic events. Similarly, the terrestrial mollusc assemblages from the cave may be affected by local environmental changes pointing to local anthropogenic interventions.

The Romito cave is located in Southern Italy, at a distance of 25 km from the Tyrrhenian coast of Calabria, in the Lao valley (fig. 1a,b). The cave entrance is at 275 m above sea level, at the base of a rocky cliff to the south of a narrow creek that is a tributary of the Lao river. Huge boulders lie on the talus in front of the entrance (fig. 1c) indicating that there was originally an overhang (rock shelter) that has now partially collapsed. The cave has been gouged out in a distinctly stratified, darkish Jurassic limestone. The deposits at Grotta del Romito contain one of the most important Upper Palaeolithic sequences of the Italian peninsula (from Gravettian to Late Epigravettian). The site has a long history of excavation (both in the cave and in the rock shelter) which has produced a detailed 13,000 year record of Palaeolithic occupation (Martini 2002; Martini et al. 2004 and 2007; Martini & Lo Vetro 2005a and 2005b). Based on the stone tool assemblages contained in the deposits, the Palaeolithic cave-fill stratigraphy can be divided into two main sequences. The lower sequence (ca. 3.5 m thick) mainly consists of natural cave-fill deposits that contain several layers of anthropogenic origin. These are the archaeological layers E, F, G, H and I. The upper sequence (ca. 2.5 m thick) is entirely made up of anthropogenic deposits and contains the archaeological layers A, B, C and D. 61

Hidden Landscapes

3. The lower interval: a Final Pleistocene climate record based on coarse-grained cave deposits

RM1b and RM1d include anthropogenic deposits in their upper part. Sub-unit RM1d contains the most complete stratigraphical sequence, consisting of channel fill and crevasse sediments covered by the archaeological layers I and I1. Charcoal from a fireplace in level I yielded an uncalibrated radiocarbon date of 23,475±190 yrs BP (fig. 2). The faunal assemblages from the anthropogenic deposits at the top of the RM1d sub-unit are dominated by ibex, followed by chamois. Woodland species, such as red deer and wild-boar, are rare. Three small mammal species have been found in level I: Talpa europea, Talpa caeca and Microtus arvalis. Unit RM2 has three sub-units (RM2a-c) and consists of 1.2-1.4m of channelized deposits (fig. 2). The channels of unit RM2 are wider and shallower than those of unit RM1. The channel fill is represented by silty sand, which has been deposited by short-lived flows and covers gravelly lag deposits. Cobble- to boulder-sized stones from the collapsed cave roof are common (fig. 2). Sub-unit RM2b is composed of the series of archaeological layers H3. Charcoal from level H4 provided a radiocarbon date of 24,200 ± 750 cal. yrs BP (fig. 2). Sub-unit RM2c consists of channelized deposits and related crevasse deposits,

3.1 Methods

The deposits forming the lower sequence are located close to the entrance of the cave. The archaeological excavations created two main vertical sections, called East and North, that were up to 6 m deep (fig. 1c). Two detailed sedimentological logs in the East profile were measured and correlated with a third measured log in the North section. In sedimentological parlance, a log is a schematic representation of a sedimentary succession obtained by documenting the thicknesses and sedimentological features (grain size, sedimentary structures, etc.) of superimposed sedimentary units (beds). 3.2 Sedimentary succession

The lower sequence can be divided into three main units (RM1-3), usually consisting of a basal alluvial portion followed by an anthropogenic deposit (fig. 2). Interval RM1 comprises four sub-units (RM1a-d) and mainly consists of 1.2-1.3 m of channelized gravel (fig. 2). The channel fill consists of channel lags and side-bar gravels (fig. 2). Sub-units RM1a and RM1c are entirely made up of fluvial gravel, whereas

Figure 1 – Site location of the Grotta del Romito (a, b); excavation area and cave geomorphology (c, d).

62

Colonese et alii, Grotta del Romito (Cosenza): latest Pleistocene landscape

Figure 2 – Geological logs and facies interpretations of the Grotta del Romito.

(fig. 2). In the SE part of the study area the channelfill deposits change abruptly into rock-fall accumulations. Isolated cobble- to boulder-sized stones, fallen from the cave roof, are commonly associated with alluvial deposits. The channels within sub-units RM3ab are characterized by a low depth/width ratio; they are filled with gravelly lags and silty sand deposited by short-lived flows. The topmost channel-fill deposits mainly consist of anthropogenic sediments. The channel within sub-unit RM3c is slightly deeper, filled with gravelly side-bars (fig. 2) and gradually changing into the related crevasse deposits (fig. 2). Sequence RM3 contains the archaeological layers E1-E16, which provided radiocarbon dates ranging from 19,270 ± 210 to 16,250 ± 500 cal. yrs BP (fig. 2). Sub-sequence RM3 is followed by 1.5-2.5 m thick anthropogenic deposits which contain the archaeological layers A-D. The lowest part of layer D (layers D35-D29) produced radiocarbon dates ranging from 15,400 ± 500 to 14,600 ± 400 cal. yrs BP (fig. 2). As was the case in sub-sequence RM2, the faunal assemblages from the anthropogenic deposits of

succeeded by an anthropogenic sub-sequence (fig. 2). This sub-sequence is the thickest one in the entire stratigraphy and contains the archaeological layers F1-F4. Several samples taken from layers G and F produced radiocarbon dates that range from 23,025 ± 600 to 20,550 ± 350 cal. yrs BP (fig. 2). The faunal assemblages from the anthropogenic deposits of unit RM2 are similar to those from layers I-I1, although a slight but significant increase of red deer occurs in layers H and G, where roe-deer appears for the first time. The only carnivore is Crocuta crocuta in H3. In the F-series of layers roe-deer and chamois increase considerably, but ibex is still the dominant species. The presence of Marmota marmota in F4, F3 and F2 is considerable, while wolf and fox appear in F4. In the G-series of layers remains are found of Arvicola terrestris, Microtus savii, Microtus arvalis, Microtus gr. multiplex-subterraneus, Microtus agrestis. In the F-series, Apodemus gr. sylvaticus-flavicollis and Clethrionomys glareolus are found. Unit RM3 includes three sub-units (RM3a-c), composed of channel fill gravel that gradually changes into rock-fall deposits 63

Hidden Landscapes

dry stage is marked by a resumption of underground drainage, as well as by a progressive growth of the competence of the watercourses (fig. 2). By contrast, the drainage deactivation which occurred after the deposition of sub-unit RM3 (fig. 2), was an abrupt event. It has been dated to around 15,000-16,000 cal. yr BP. This latter event interrupted a progressive increase of the competence of the watercourses. It was not connected with any climatic change and was probably induced by a rapid choking of the drainage channels, caused by collapses of the cave wall. Nevertheless, this event allowed prolonged human settlement of the cave, and finally put an end to the fluvial sedimentation.

sub-sequence RM3 show an increase of roe-deer and chamois, but ibex remains the dominant species. The deposit contains also remains of wild cat and bear. Small mammals are represented by species of open landscapes as well as a few woodland species, such as Apodemus gr. sylvaticus-flavicollis, Clethrionomys glareolus and Moscardinus avellanarius. 3.3 Discussion

The channel-fill deposits of RM1, with their wellstratified bars, indicate confined and constant flows during major flood events, according with the scarcity of anthropogenic deposits. The transition from the confined and persistent flows of sub-unit RM1 to the short-lived and weakly confined ones of sub-unit RM2 marks a progressive deactivation of the channels that drained the cave. The minimum level of this progressively diminishing flow is reflected in the top level of sub-unit RM2c. The channel reactivation at the base of unit RM3 did not cause significant erosion of the upper anthropogenic layers of unit RM2, as is shown by the radiometric age of the deposits. Sub-units RM3a-c mark the gradual transition from low-competence watercourses to persistent streams, as well as the resuming of drainage conditions similar to those present in sub-unit RM1. Nevertheless, the overall coarse-sediment load was strongly reduced, as is indicated by the presence of numerous stones deriving from cave walls. The abrupt occurrence of thick anthropogenic deposits above sub-unit RM3 suggests immediate channel deactivation and a final decline of water-transported sediment. The changes in assemblages of small and large mammals in the RM1-3 sequence indicate that the environment initially opened up almost completely (RM1 and RM 2) but subsequently became moderately wooded (top of RM2 and RM 3). However, the information provided by the fauna remains was not as detailed as that disclosed by the hydrological changes recorded in the clastic sediments. Particularly the gradual deactivation and reactivation of underground drainage between ca. 23,000 uncalibrated yr BP and 16,000 cal. yr BP correlates very well with the dry climatic conditions observed in records for southern Italy at the last glacial maximum (e.g. Giraudi & Frezzotti 1997; Huntley et al. 1999; Ramrath et al. 1999; Chondrogianni et al. 2004). The development of these arid conditions (fig. 2) is indicated in the Romito cave deposits by the gradual decrease in the competence of the watercourses, which led to their final deactivation and consequently enabled human occupation. The wet trend that followed this

4. The upper sequence: Late Glacial human impact on surrounding vegetation, data from terrestrial mollusc shells Terrestrial molluscs are known to be very sensitive to several environmental parameters (e.g. Colonese et al. 2007) and to human impact on the landscape (Magnin et al. 1995). Moreover, terrestrial molluscs form communities of numerous individuals, and can therefore be analyzed statistically (Rousseau 1990). Thanks to these characteristics, terrestrial mollusc shells from anthropogenic deposits can be useful tools for reconstructing past environmental conditions and human impact on the landscape. Assemblages of shells from the Late Glacial sequence of Grotta del Romito were analyzed for the purpose of elucidating the dynamics of the vegetation history surrounding the cave during the Late Epigravettian settlement episodes. 4.1 Material and methods

The shell remains come from the upper-Late Epigravettian stratigraphic sequence located near the entrance of the cave. This sequence contains the archaeological layers B, C and D, further subdivided into palaeosurfaces consisting of anthropogenic settlement layers that are characterized by structured fireplaces and horizontal accumulation of ecofacts and artefacts (bones and lithic assemblages) (Martini et al. 2004). Shells were collected during archaeological excavation by sieving the bulk of the sediment from layers C (palaeosurfaces C1-C4) and D (palaeosurfaces D-D13). Statistical analyses were performed using the software programme C2 (Juggins 2003). This anthropogenic sequence (layers C-D13) has a detailed radiocarbon chronology, ranging from 14,876-14,107 to 13,170-12,890 cal yrs BP (2ƫ; Colonese et al. 2007). 64

Colonese et alii, Grotta del Romito (Cosenza): latest Pleistocene landscape

4.2 Results and discussions D7

Terrestrial mollusc assemblages are mainly composed of woodland- and hygrophilous species (Acicula sp., Argna sp., Cochlodina cf. laminata, Clausilia sp., Discus rotundatus, Vitrea sp., Retinella cf. olivetorum, Oxychilus sp., Oxychilus draparnaudi, Helicodonta obvoluta, Testacella cf. Haliotidea, Limax sp., Milax sp.), followed by xerothermic species living on the rock face (Cochlostoma cf. montanum, Chondrina a. avenacea, Medora sp., Medora cf. dalmatina, Marmorana fuscolabiata) and open places (Xerotricha conspurcata), and finally by mesophilous species (Pomatias elegans and Helix cf. ligata). Frequencies of species and specimens are not distributed homogeneously across the stratigraphy. A significant increase of both (r = 0.66; P  0.001) is observed from the bottom (D13) to the top layers (C) (fig. 3a). Using correspondence analysis, species with similar ecological requirements have been grouped (fig. 3b). The first axis (r = 0.64) represents the environmental gradient from open to woodland environments, whereas the second axis (r = 0.12) discriminates the rock-wall species from the hygrophilous ones. The overlapping Correspondence-Analytical results of species and layers clearly indicate a transition from open-dry to moist-shady environment, in which the first (opendry) is characterised by either mesophilous species or those living in open places and on the rock wall (MR), and the second (moist-shady) by hygrophilous and woodland species (HW), located near the origin of the graph (i.e. at zero along all axes). By combining the taxonomic, quantitative and ecological data, three main palaeoenvironmental episodes can be recognized in stratigraphic succession (fig. 4): • Semi-forested I – from layer D13 to D9. This interval is characterized by a generally low frequency of species and specimens, in which MR species are dominant while HW species are few. • Semi-forested II – from layers D8 to D. In this phase the previous environmental conditions continue, but an increase of both species and specimen frequencies is observed. In this interval HW species also increase (e.g. Discus rotundatus, Oxychilus draparnaudi, Cochlodina cf. laminata and Helicodonta obvoluta) while MR species decrease. • Forest – from layer C4 to C. In this phase HW species are dominant, whereas MR species are rare. Discus rotundatus is the main element and indicates an increase of vegetation and moistshady places near the cave.

D8

4.8

SP

b

14 12 10 8

a

6 4 2 0

3.6

r = 0.66

0

100

Rock wall

200

SPM

R20 2.4

AXIS 2

D2 R15

1.2

0.0

R12

R18 D3R7 R22 D4 R1 R10 R5 C2 D1 R9 C3R19 R6R3 CR2 R11 R17 R4 C1 C4 R14 D

D11 R8 D10 R16 D9 R21 D12

-1.2

Woodland

R13

Open

Moist -2.4 D13

-3.6 -1.0

0.0

1.0

2.0

3.0

4.0

AXIS 1

Figure 3 – a. Linear regression between species and specimen frequency. b. Biplot of correspondence analysis (CA) between species (open triangle) and layers (full circle): R1 – Cochlostoma cf. montanum; R2 – Acicula sp.; R3 – Pomatias elegans; R4 – Chondrina avenacea avenacea; R5 – Argna sp.; R6 – Discus rotundatus; R7 – Vitrea sp.; R8 – Retinella cf. olivetorum; R9 – Oxychilus sp.; R10 – Oxychilus draparnaudi; R11 – Testacella cf. haliotidea; R12 – Medora sp.; R13 – Medora cf. dalmatina; R14 – Cochlodina cf. laminata; R15 – Clausiliidae gen. sp. ind.; R16 – Clausilia sp.; R17 – Limax sp. and Milax sp.; R18 – Xerotricha conspurcata; R19 – Helicodonta obvoluta; R20 – Helicidae gen. sp. ind.; R21 – Helix cf. ligata; R22 – Marmorana fuscolabiata.

The taxonomic variations in terrestrial mollusc assemblages don’t show any transition from cold to warm conditions; remarkably, no cold-resistant species were found in the assemblages but only temperate ones still living in the region. Assemblages of small mammals from the same layers (Colamussi 2004; Malavasi 2005-2006) do not record any significant change in environmental conditions at that time either. Rather, they indicate a continuous woodland environment, which agrees with palaeoclimatic records from southern Italy for this period 65

Hidden Landscapes

MR

d.

um

. in

nic

. sp

am

en

cf.

eg da

um

idi

lici

Pis

He

Ac icu Arg la sp n . Cla a sp. usi liid Cla ae g usi e Co lia s n. sp ch p . in d. Dis lodin . cus a c rot f. la un mi d a na tus ta

Vit re R e a sp. tin Ox ella ych cf. ilu oli Ox s sp. veto ru m ych ilu sd rap arn He au lic di Nu odon de ta (Li ob Te max volu sta sp ta Co cella ., M i ch c los f. h lax s p t Ch oma aliot .) on cf id Me drin . mo ea nt d a Me ora s aven anum d p. ace aa Xe ora c v rot f. d ric alm enac h ea Ma a co atin n a rm ora spur na cata fus Po col ma ab He tias iat lix ele a cf. ga lig ns ata

HW

SPF

SPMF

AXIS 1

C C1 C2

F

C3 C4 D D1 D2 D3

SF II

D4 D7 D8 D9 D10 D11

SF I

D12 D13 0

0

0

20 0

0

0

20

40

60

80

0

0

0

20 0

20

40 0

0

20 0

0

20 0

0

0

0

20 0

20

0

0

20

40

60

80 0

20 0

20 1 3 5 7 9 11 0

80

160 240-1,0 0,0 1,0 2,0 3,0

Figure 4 – Stratigraphic species (SPF) and specimen (SPMF) frequencies. Environmental gradient values are displayed in column ‘Axis 1’, and palaeoenvironmental intervals in the right-most column.

(Huntley et al. 1999 and references therein). Similarly, oxygen and carbon isotopic analysis on terrestrial mollusc shells do not show any significant changes in climatic conditions, except for the brief climatic deterioration recorded in layer 8D (Colonese et al. 2007). This is significant, because neither abrupt nor subtle palaeoclimatic changes could explain the observed taxonomic variations in the terrestrial mollusc remains. The observed changes may therefore have been caused by environmental changes at a local scale, in the immediate surroundings of the site. Terrestrial mollusc shells become deposited in cave sediments mainly by water- and wind transport after death (e.g. Girod 1997). In order to become deposited they must therefore have lived relatively close to the cave. By contrast, remains of small mammals in cave deposits are mainly the result of nocturnal predators, such as owls. They can therefore become deposited in caves even if they lived at some distance from them. The discrepancy between malacological and small-mammal records at Grotta del Romito could be related to these different depositional process as well as to differences in their natural habitats. The terrestrial mollusc assemblages suggest an initial phase of dominant open spaces and exposed calcareous rock surrounding the cave. This period was followed by the expansion of vegetation and moistshady places. Such environmental transitions, without a climatic origin that might have cause them, could be explained by the deterioration-restoration of the surrounding vegetation due to the intense use

by humans of the cave during Late Glacial period. Human impact on landscapes is well-known in prehistoric times (e.g. Berglund 2003) and previous studies have demonstrated that terrestrial mollusc assemblages can record such impact due to variations of local ecological conditions (Magnin et al. 1995; Martin et al. 2003). In Palaeolithic sites where intensive and continuous settlement occurred, like Grotta del Romito, deteriorations of the vegetation near the habitation zone can reasonably be expected (e.g. deforestation and fire), and this could presumably be reflected in the assemblages of local terrestrial molluscs (Magnin et al. 1995). In the Late Glacial stratigraphic succession of Grotta del Romito, changes in terrestrial mollusc assemblages could be explained by variations in human impact on the vegetation near the cave during the Late Epigravettian settlement. This suggests that layer D should be interpreted as an interval of very intensive human frequentation of the cave, or that its immediate environment was used differently than during the episodes reflected in layer C. This may have brought about a substantial deterioration of the vegetation, as a result of deforestation on a previously unknown scale. This human impact consequently changed the local malacological community, favouring the presence of MR species that live in open places. At the bottom of layer C a decrease in the use of the cave or the area immediately surrounding it by human groups was probably responsible for the restoration and expansion of woodland 66

Colonese et alii, Grotta del Romito (Cosenza): latest Pleistocene landscape

vegetation and associated species in this area. However, assemblages of bones and lithics do not show any substantial quantitative difference during these intervals. This may support the hypothesis that the area outside the cave was used differently, rather than a hypothesis of varying intensity of its use, but further research is required to shed light on this matter.

Palaeogeography, Palaeoclimatology, Palaeoecology 254, pp. 550-560 [doi 10.1016/j.palaeo.2007.07.005]. Giraudi C. & Frezzotti M. 1997. Late Pleistocene glacial events in the Central Apennines, Italy. Quaternary Research 48, pp. 280-290. Girod A. 1997. L’uso dei molluschi continentali come indicatori paleoambientali: problemi collegati all’ambiente “Grotta”. Società per la Preistoria e Protostoria della regione FriuliVenezia Giulia, pp. 213-219. Huntley B., Watts W.A., Allen J.R.M & Zolitschka B. 1999. Palaeoclimate,chronology and vegetation history of the Weichselian Lateglacial: comparative analysis of data from three cores at Lago Grande di Monticchio, southern Italy. Quaternary Science Reviews 18, pp. 945-960. Juggins S. 2003. C2 User guide. Software for ecological and palaeoecological data analysis and visualisation. Newcastle: University of Newcastle Press. Magnin F., Tatoni T., Roche P. & Baudry J. 1995. Gastropod communities, vegetation dynamics and landscape changes along and old-field succession in Provence, France. Landscape and Urban Planning 9, pp. 249-257. Malavasi G. 2005-2006. Micromammiferi dei livelli I-D di Grotta del Romito (Papasidero, CS) – Analisi paleontologiche e ricostruzioni ambientali. Università degli Studi di Ferrara, unpublished thesis. Martin S., Magnin F. & Kiss L. 2003. Land snail and human impact: temporal resolution of Holocene assemblages, in Fouache (ed), The Mediterranean world environment and History. Paris, Elsevier SAS, pp. 239-250. Martini F. 2002. Grotta del Romito. Guide del Museo e Istituto Fiorentino di Preistoria. Museo e Istituto Fiorentino di Preistoria, Firenze. Martini F., Cattani L., Colamussi V., Colonese A.C., Martino G., Mallegni F., Noto F., Ricciardi S., Rickards O. & Rolle R. 2004. Primi risultati delle nuove ricerche nei livelli epigravettiani di Grotta del Romito a Papasidero (scavi 2000-2002). Atti della XXXVII Riunione Scientifica I.I.P.P. Preistoria e Protostoria della Calabria 1, pp. 35-58. Martini F. & Lo Vetro D. 2005a. Grotta del Romito (Papasidero, Cosenza): recenti risultati degli scavi e degli studi. Atti delle giornate di studio sulla Preistoria e Protostoria della Calabria. Scavi e ricerche 2003, pp. 5-15. Martini F. & Lo Vetro D. 2005b. Il passaggio Gravettiano– Epigravettiano a Grotta del Romito (scavi 2003-2004). Prime osservazioni, in Martini F. (ed.), Askategi, miscellanea in memoria di Georges Laplace, Rivista di Scienze Preistoriche 55, suppl. 1, pp. 151-176. Martini F., Cilli C., Colonese A.C., Di Giuseppe Z., Ghinassi M., Govoni L., Lo Vetro D., Martino G. & Ricciardi S. 2007. L’Epigravettiano tra 15.000 e 10.000 anni da oggi nel basso versante tirrenico: casi studio dell’area calabro-campana, in Martini, F. (ed.), L’Italia tra 15.000 e 10.000 anni fa. Cosmopolitismo e regionalità nel Tardoglaciale, Atti della tavola rotonda, Firenze 18 Novembre 2005, pp. 157-207. Ramrath A., Zolistschka B., Wulf S. & Negendanck J.F.W. 1999. Late Pleistocene climatic variations as recorded in two Italian maar lakes (Lago di Mezzano, Lago Grande di Monticchio). Quaternary Science Review 18, pp. 977-992. Rousseau D.D. 1999. Statistical analyses of loess molluscs for paleoecological reconstructions. Quaternary International 7-8, pp. 81-89.

5. Conclusion The work described here shows how both local natural and anthropogenic mechanisms may affect the climatic and environmental ‘archives’ preserved in archaeological records. Especial care should be taken to distinguish between the effects of climatic changes and those caused by local geomorphic development and human interventions. The abrupt deactivation of stream drainage at the Grotta del Romito between about 16,500 and 15,500 cal. yr BP, for example, allowed another major phase of human settlement to take place but was probably caused by the insulation of the cave system due to surface forestation, fracture/ sink clogging and collapses of the karstified rock that may have been triggered by earthquakes. By contrast, in the uppermost part of the cave succession, quantitative and qualitative variations in terrestrial mollusc suggest intensive human impact on the vegetation cover close to the cave between about 14,800-13,300 cal. yr BP. A lowering of that impact between about 13,400-12,800 cal. yr BP (perhaps due to a drop in the cave occupation) resulted in the expansion of forested areas. The combination of analytical approaches employed by us has only emphasized the need to consider complex local factors if we want to recognize the hidden features of prehistoric landscapes.

References Berlung B.E. 2003. Human impact and climate changessynchronous events and a causal link?, Quaternary International 105, pp. 7-12. Chondrogianni C., Ariztegui D., Rolph T., Juggins S., Shemesh A., Rietti-Shati M., Guilizzoni P., Lami A., Niessen F., McKenzie J.A. & Oldfield F. 2004. Millennialto interdecadal climate variability in the Mediterranean during the LGM–The Lake Albano record. Quaternary International 122, pp. 31-41. Colamussi V. 2004. I micromammiferi della sequenza epigravettiana di Grotta del Romito. Atti XXXVII Riunione Scientifica IIPP II, pp. 641-644. Colonese A.C., Zanchetta G., Fallick A.E., Martini F., Manganelli G. & Lo Vetro D. 2007. Stable isotope composition of Late Glacial land snail shells from Grotta del Romito (Southern Italy): palaeoclimatic implications.

67

Hidden Landscapes:

MOUNTAINS AND UPLANDS

9 The Visible Landscape: inferring Mesolithic settlement dynamics from multifaceted evidence in the south-eastern Alps Federica Fontana, Antonio Guerreschi, Marco Peresani Università degli Studi di Ferrara, Dipartimento di Biologia ed Evoluzione, Sezione di Paleobiologia, Preistoria e Antropologia, Corso Ercole I d’Este 32, 44100 Ferrara – [email protected]; [email protected]; [email protected] Abstract This paper focuses on the high visibility of the Mesolithic record in the Eastern Italian Alps and discusses the factors that have favoured the collection, in the last thirty years, of multifaceted evidence which allowed one of the most successful models of prehistoric European mountain landscapes to be defined. It also contains a critical review of the main approaches and criteria which were adopted over time in order to delineate the current settlement pattern. In particular, the latter was analysed by considering different levels of resolution, from intra-site organisation to the identification of mobility routes. The reason why the current model no longer seems to completely fit the variability of the available record may be the existence of more complex dynamics than was previously believed. These still await a closer study, at both a synchronic and a diachronic level. The authors argue that recent work suggests that a more realistic vision of the matter will depend on enlarging the range of methodologies used by applying a truly multidisciplinary large-scale approach to the data.

1. Introduction

While discussing these aspects, special attention will be given to a few particular geographical contexts of the region in question. These constitute the specific areas investigated by the Authors: the Mondeval de Sora terrace in the Belluno Dolomites and its surrounding territory, and the Cansiglio plateau in the Venetian Prealps. In both these areas research carried out in the last 20 to 25 years has brought to light some notable evidence which has added significant elements to the reconstruction and revision of the model. It should be noted that due to the impressive amount of publications produced during several decades of research in the area, the cited works can only represent a selection.

Data provided by over thirty years of intensive research and study on the settlement dynamics of the last hunter-gatherers of the south-eastern Alps have allowed the development of what may be called a successful model of prehistoric hidden mountain landscapes. There is an exceptional amount of information available, especially if we consider that the presentday morphology of the Alpine chain, like that of other active mountain systems, is the result of interplay between several complex morphogenetic and tectonic processes which have caused and are still causing rapid and intense change. However, some factors have almost paradoxically favoured the preservation, visibility and recovery of the archaeological evidence in this sector of the chain. Although in most cases the available record appears ephemeral, its abundance, variety and variability across the territory has offered the possibility to outline a fairly detailed framework, the reconstruction of which is strictly connected to the adoption of a specific array of methodological approaches. To illustrate this case-study three main issues will be addressed: 1. an analysis of the factors that may have favoured the preservation and visibility of the evidence and the recovery of what can be considered a rich archaeological record; 2. the main methodological approaches and criteria that have been applied to reconstruct the current pattern along with the different levels of analysis which have been undertaken, and 3. some possible bias factors and future research perspectives.

2. Factors favouring the visibility of the record From a geological viewpoint the southern slope of the Eastern Alps can be divided into two sectors, a southern and a northern one, separated by the Linea insubrica running along the Valtellina and the Venosta, Isarco and Pusteria valleys. The southern sector includes the Prealps, the Dolomites and other massifs mainly formed from volcanic and sedimentary deposits which originated between the Permian and the Cenozoic. The southern margin of the prealpine fringe itself articulates in a series of karst plateaux reaching altitudes between 1775 m (Monte Grappa) and 2300 m (Cima Dodici, Altopiano dei Sette Comuni). The Dolomites are delimited by the Valsugana Valley (valley of the river Brenta) to the south, the Piave Valley to the east, the Pusteria Valley (valley of the river Rienza) to the north and the Adige-Isarco Valley to the west. They are mainly characterised by alternating dolomite and volcanic rocks that reach 71

Hidden Landscapes

Grotta di Ernesto in the Altopiano di Asiago; Awsiuk et al. 1991). Several factors seem to have favoured the visibility and recovery of evidence and, at the same time, to have biased the preservation of the record.

altitudes of between 2996 (Pale di San Martino) and 2998 m (Tre Cime di Lavaredo). The main river courses that cross the southern slopes of the Eastern Alps are, from west to east, the Adige, Brenta, Piave, Tagliamento and Isonzo rivers. These all run north to south except for the Adige which, as it reaches the plain, turns east to flow into the Adriatic Sea. If we consider all the available data for the investigated period (the Early Holocene; Orombelli & Ravazzi 1996) the evidence appears richer and more varied in highland areas, both on the prealpine belt and on the Alps, than along valley bottoms (Dalmeri & Pedrotti 1992). In the latter, alluvial sediments represent the main agents responsible for obliterating possible open-air settlements, while slope-waste deposits have hidden most shelters that open at the base of the main rock-walls. Nonetheless, some important sites of the latter type contain deep stratigraphic series and have good preservation when not destroyed by the quarrying that led to their identification in the first place. These sites are known especially along the Adige river valleybottom (Romagnano, Vatte di Zambana, Pradestel, Soman, Gaban, and Dos de la Forca rock-shelters) and, more rarely, along other valley bottoms such as the Natisone (Biarzo rock-shelter) and the inner Alpine valley of the Isarco river (Stuffles; Battaglia et al. 1992; Broglio 1980 1992b; Broglio & Improta 1994-1995; Kozlowski & Dalmeri 2000; Guerreschi 1996; Bazzanella et al. 2004). Only a few open-air settlements have been identified so far, such as that of La Vela near Trento (Bazzanella et al. 1997). Conversely, the number of sites located in highland territories amounts to some hundreds (Dalmeri & Pedrotti 1992; Broglio 1992a, 1992b). Their present distribution reflects higher densities in the sector located directly east of the Adige River valley including, from south to north, the pre-alpine massifs of Lessini and Altopiano di Asiago, the Lagorai chain, the Latemar, Sella and Alpe di Siusi groups, and the Pusteria Valley. Further east the record appears more scattered, with some concentrations of sites in the Belluno Dolomites (Alciati et al. 1992; Fontana et al. 2002; Fontana & Pasi 2002) and the Friuli Alps (Pessina 2005). In most cases the record consists solely of lithic scatters found at specific open-air locations (saddles, passes, small lake basins, dominating points) and at different altitudes which may reach over 2200 m above sea level. More rarely the sites are located under rock-shelters usually formed by large boulders and at the entrance of small caves (there is only one significant example, at

2.1 The anthropic factor

This mainly refers to the intensive occupation of these territories by the latest Palaeolithic and Mesolithic hunter-gatherers, following the deglaciation process, due to the improved climatic conditions that took place from the Late Glacial interstadials onwards (Ravazzi et al. 2007). Mountain territories became ice-free, favouring human seasonal occupation and the exploitation of available local resources (Bertola et al. 2007). The low impact of anthropisation processes in the area during the succeeding periods may have contributed to preventing the destruction of the most ancient evidence. 2.2 The environmental factor

This can be analysed from a twofold perspective. Following the alpine deglaciation from 14.7 to 14.3 ky Cal BP, the eastern Alps underwent general conditions of medium- and short-term stability at middle altitudes which supported pedogenetic processes and the 800 m upward shift of the tree line, which remained at 2000-2100 m for the next five millennia (only in historical times did the artificial tree-line shift 300-400 m downwards due to human impact; Vescovi et al. 2007; Tinner & Vescovi, in press). As a consequence highland territories, especially at the altitudes which were occupied by the Mesolithic groups, underwent only minor changes. This is exemplified by the disappearance of some small pro-glacial lakes which had constituted points of attraction for the last hunters and gatherers that occupied these territories, such as in the case of the Mondeval de Sora terrace (Belluno Dolomites; Alciati et al. 1992). The second perspective involves formation processes of the record when passing from the systemic to the archaeological context, and the role of postdepositional factors. Especially at open-air sites and in conditions of low or absent sedimentation, traces of human occupation are progressively incorporated into the grass cover, where they remain buried inside the litho-soil (fig. 1). Consequently, their discovery usually follows either erosion processes which occur at specific locations such as slopes and lake shores, or low-impact anthropic and animal practices (construction of ski installations, tracks and pathways, marmots’ dens, etc.). 72

Fontana – Guerreschi – Peresani, The Visible Landscape: inferring Mesolithic settlement dynamics…

At the same time, depositional and post-depositional processes have some important consequences for the preservation of the archaeological record. These processes produce a dearth of perishable remains along with the destruction of any existing traces of ‘evident’ structures. Therefore, in most sites, lithic assemblages are scattered throughout the thin soil profile as a consequence of strong postdepositional disturbance mainly by long-term soil formation processes, and they represent the only preserved evidence of past human occupation. The horizontal and vertical distribution of artefacts along the soil profile suggests that the processes acting at some of these sites were rather complex (Angelucci & Peresani 2000). One problem to solve consists of establishing whether or not these processes operated contemporaneously. Furthermore, in many cases it is unclear whether the archaeological contexts analysed are the result of one or many occupation phases. In addition to soil formation, another very important post-depositional factor is cryoturbation. This can produce strong effects especially at the top of the profile, and bias the orientation patterns of artefacts. The same holds true for trampling, a dynamic and important process at the moment of occupation, and bioturbation, an intensively active and sometimes ongoing phenomenon. Graviturbation is another important agent, particularly at sites located on a limestone bedrock and affected by karst dissolution: it works as a sort of conveyor belt, often starting from the moment of site occupation and continuing until the present day, slowly burying finds at differential rates depending on weight throughout the soil profile.

Conversely, anthropic signatures are often more visible at rock-shelters, which appear much less represented than open-air sites, and in some cases persistent occupation produces thin stratigraphic sequences in which organic remains and dwelling structures are preserved. Some case studies have proven the potential of geoarchaeological approaches for the investigation of these questions. In particular, soil micromorphological techniques have been applied at some sites, revealing the impact of the aforementioned processes in the formation of the record (Angelucci et al. 1996; Angelucci & Peresani 2000). 2.3 The research factor

Following the first occasional discoveries made in the 1970’s, the territory became the object of several survey projects for over thirty years. The modality of research, involving both official groups and local amateurs, produced a huge amount of data but has resulted in a typical ‘leopard-spot’ distribution pattern of the evidence. Several factors contribute to this uneven distribution, but especially a lack of discussion and sharing of a common research strategy. Survey programmes were therefore adopted that differed on several important points, including the use of reference grids that varied according to the different areas explored, failure to record prospected archeologically sterile areas, and the absence of preventive programmes to define archaeological risks. In salvage archaeology factors relating to the visibility of the Mesolithic record, especially where this has an ephemeral character, are unfortunately not part of the specific training of specialised personnel, who are therefore not well-prepared to rec-

Figure 1a, b – View of the same fire-place taken three years apart (2000 and 2003) at the Alpe Veglia (Lepontine Alps, 2000 m) showing that, at open-air highland locations, traces of human occupation are progressively incorporated into the grass cover, where they remain buried inside the litho-soil (Photos: A. Guerreschi).

73

Hidden Landscapes

ery, where cores and the larger — often not laminar (blade-shaped) — products and tools have a high incidence. Moreover, the latter are often characterised by antithetic distribution patterns in comparison to microliths which, along with residues deriving from microlith manufacture (i.e. microburins) and burnt artefacts, mostly concentrate in central areas. The distribution of burnt artefacts seems to testify to the presence of hearths rather than to reflect zones where combustion residues were discarded. The location of economic activities within these contexts, mostly considered as resulting either from single occupations or from multiple occupations with partial superposition of similar patterns (for example Colbricon 2A and 2B, 8A and 8B, 8A’ level 3, 8A’ level 4-5), is highlighted by the differential distribution of the main functional categories. Refitting of débitage products and residues shows the relationships between different areas within the sites (Bagolini & Dalmeri 1987; Peresani & Angelini 2002). Conversely, sites located under small rock-shelters (Plan de Frea, Broglio et al. 1982; Mondeval de Sora – US8, Fontana & Vullo 2000) result in more complex situations with paved areas, structures delimiting habitation surfaces and hearths, and superimposed distribution patterns of the main functional categories. These probably indicate complex multifunctional areas for the development of different tasks such as blank production, microlith manufacture, and raw material processing, and seem to result from multiple and repeated occupations over time. These shelters, which represent landmarks in the landscape and which show evidence of occupation from prehistory to the present, have also been referred to as ‘persistent’ places (Angelucci et al. 1998). Some simpler situations, i.e. with features similar to open-air sites, are also documented, such as at Lago delle Buse, site 1 (Dalmeri & Lanzinger 1992). Concerning site function, the approach used was based on an integration of different data sets: a) structural analyses of the lithic assemblages, namely evaluation of ratios between tools and microliths and between microliths and their technical residues (each of these categories is considered a ‘functional indicator’ and therefore these ratios respectively reflect maintenance versus hunting practices and offsite versus on-site production of microliths); b) the characteristics of the topographic location of sites; c) and — when available — their intra-site spatial organisation. Combining these indexes showed that at locations scattered on ridges, peaks and saddles, microliths

ognise this type of evidence. Nonetheless, given the limits imposed by the diverse research strategies used and the irregular morphology of the investigated territory, the great bulk of collected data as well as its distribution across the territory can be considered to at least partially reflect the exploitation modality of these areas by Mesolithic groups.

3. Levels of analysis and criteria adopted to define the current pattern Several hypotheses on settlement dynamics have been elaborated on different scales, starting from the first ‘pioneer’ works which were carried out in the 1970’s and 1980’s (Broglio 1973, 1980; Bagolini 1972; Bagolini et al. 1983; Bagolini & Dalmeri 1987). These range from studies of the internal spatial organisation and function of specific sites to the elaboration of exploitation models for definite territories and the definition of the regional settlement pattern. As is typical of the Italian tradition, the approach adopted is mostly based on the archaeological evidence, with occasional reference to specific ecological contexts. It is not always an easy task to distinguish the different scale levels, since they are closely connected and all contribute to the definition of the general model. Nonetheless, establishing such scale levels can effectively illustrate the complexity of the current framework. 3.1 Intra-site spatial organisation and site function

The analysis of evident structures and the study of spatial data for recognising latent structures have been the basic means of investigating intra-site organisation, while site function has mostly been determined by the analysis of the composition of lithic assemblages and the topographic location of sites. Both analyses were developed mainly for mountain camps rather than valley-bottom sites. Generally speaking, an increasing degree of spatial complexity is observed from open air settlements to rockshelters (Fontana & Guerreschi 2003). Comparisons of discrete spatial distribution maps of finds either at a 1:1 scale or by grids of 33 and 50 cm consider both the total quantity of artefacts and the basic functional categories of cores, tools, microliths, residues and burnt artefacts. Concerning open-air sites (for example Colbricon, Bagolini & Dalmeri 1987; Casera Davià II, Peresani & Angelini 2002; Cima Dodici 3 and Cima Dodici 4, Peresani & Perrone unpublished), distribution patterns appear almost systematically homogeneous, with a radial decrease from one or more central areas towards the periph74

Fontana – Guerreschi – Peresani, The Visible Landscape: inferring Mesolithic settlement dynamics…

widely dominate over common tools in association with high incidences of microburins. This suggests that these locations were mostly used for manufacturing and restoring microliths, an expedient task intimately connected to hunting parties. In contrast, lithic assemblages recovered at rock shelters (both in mountain areas and along valley bottoms) and at some open mountain camps located near lakes produced more balanced ratios of common tools to microliths, revealing multifunctional tasks (preparation of hunting sets, food processing, transformation of perishable and non-perishable materials) and therefore indicate a residential function (Lanzinger 1985; Broglio & Lanzinger 1990). To enlarge the variability observed, an anomalous case of a transient station at Grotta di Ernesto, located at an altitude of 1130 m above sea level on the Eastern side of the Altopiano dei Sette Comuni, was made the object of detailed interdisciplinary analyses. This site was interpreted as the result of a single short-term occupation within a longer journey towards the summer highland camps (Awsiuk et al. 1991).

Figure 2 – Inter-site relations in the area of Colbricon. Sites 1, 3 and 4 are considered subsistence camps while sites 2, 9, 8A, 8 and 6 fall into the category of hunting stands. Possible ‘exportation’ of armatures from site 1 (identified as a lithic workshop) to sites 8 and 8A is proposed (modified after Bagolini & Dalmeri 1987)..

3.2 Exploitation models for limited territories

The definition of exploitation models for definite areas was one of the first results achieved by research in the area (Bagolini & Dalmeri 1987). The paradigm proposed is strictly connected to the research strategies used (see section 2.3) and which, in many cases, have led to the identification of concentrations of sites with different characteristics within specific delimited territories, often corresponding to areas with homogeneous geological and environmental features (karstic plateaus, mountain river basins delimited by crests, saddles and passes, valley systems). As already mentioned, the approach to intersite variability was strongly based on analyses of site geo-topographic features, but much emphasis was also placed on the structural characteristics of the lithic assemblages that constitute the only omnipresent evidence in all of the deposits (see section 3.1). Several examples may be cited, starting from the pioneering study carried out in the Colbricon region of the Trentino Dolomites (Bagolini & Dalmeri 1987), which is an exemplary case of inter-site variability analysis (fig. 2). Other notable studies are those of the Seiser Alm-Alpe di Siusi, Dolomites (Lanzinger 1985) and Lago delle Buse, Lagorai Chain (Dalmeri & Lanzinger 1992) and the most recent work on the Cansiglio Plateau (sites of Casera Lissandri I, Casera Lissandri 17, Casera Davià II, and Palughetto; Peresani & Angelini 2002; Peresani

et al., in press), Mondeval de Sora, Belluno Dolomites (Cusinato et al. 2003; Fontana & Guerreschi 2003; Fontana & Pasi 2002), and Cima Dodici on the Asiago Plateau (Broglio et al. 2006). All these situations testify to the association, within the same limited geographical context, of one or more possible residential sites with several ephemeral off-site stations which usually occur in panoramic positions and whose distribution frequently superimposes on current pathways. Interpretation of these mosaics of sites as complementary parts of the same settlement systems has obviously depended on either the supposed synchronism of site occupation based on their homogeneous lithic techno-typological features, or the presumed stability of occupation patterns through time, at least within each single cultural phase recognised for the Mesolithic (i.e. Sauveterrian, Castelnovian). 3.3 The regional settlement pattern

The current pattern established to describe Mesolithic regional settlement strategies, with special reference to the most ancient Sauveterrian phase, is visualised as a ‘vertical seasonal nomadic system’. Its definition is mainly due to the synthesis carried out by Broglio in the 1980’s and 1990’s (Broglio 1980, 1992a; Broglio & Improta 1994-1995). Using data 75

Hidden Landscapes

late teeth eruption. However, the application of this method has so far only been possible at one valley-bottom site (Biarzo rock-shelter, RowleyConwy 1996) and a few mountain deposits located under rock-shelters, where the chemical nature of soils has likely favoured the preservation of organic remains (Mondeval de Sora, Plan de Frea and Riparo La Cogola; (Alciati et al. 1992; Fontana et al. 2009; Angelucci et al. 1998; Fiore & Tagliacozzo 2004); • palaeoenvironmental studies undertaken on long sequences support the proposed model, especially concerning its diachronic evolution. Specifically, environmental processes from the Late Glacial to the early Holocene involved a reduction of mountain pastures and a corresponding increase in the altitude of the tree-line (Tinner & Vescovi, in press) as well as of forest density. To these phenomena seems to correspond a progressive increase in the altitude of territories occupied by human groups from the Late Glacial (Late Epigravettian) to the early Holocene (Sauveterrian) along with an apparent decrease in the intensity of frequentation of highland areas during the midHolocene Atlantic chronozone. Also, as recently suggested by Clark (2000), the disappearance of ibex and chamois from the bone assemblages of valley-bottom sites (Romagnano, Pradestel, Vatte di Zambana rock-shelters) during the late Mesolithic period (Castelnovian) could be due to the fact that high altitude hunting was no longer taking place on the scale previously recorded. Instead, there is a greater reliance on deer and wild boar, which reflect the increasingly dense forest conditions. According to the same author, the contemporaneous increase in the range of recorded anatomical elements from other large mammals and of younger animals, particularly of red deer, could reflect a change in hunting strategies towards increasingly localised subsistence and, consequently, a reduction in the size of the subsistence territories with increasing sedentism and intensification of certain resource types.

from more than 250 sites spanning from the watershed to the pre-alpine foreland, this author developed a model based on a bi-relational differentiation of valley-bottom sites on the one hand, mostly located under rock-shelters and/or in proximity to wet zones that were repeatedly used over long periods as residential camps for multiple functions, and mountain camps on the other hand, characterised by a much higher functional variability (see sections 3.1 and 3.2). It has thus been advanced that human occupation in these two main ecological contexts (valley bottoms vs. highlands) took place according to a seasonal rhythm. In particular, highland sites would have been occupied in the most favourable season of the year, and their economy mainly based on hunting carried out around the ecotone, in areas located in close proximity to the timber-line. In contrast, the valley bottoms would have been settled over the remaining part of the year. Here the presence of wet zones would also have favoured the exploitation of a more varied range of resources from ungulates to fish (Dos de la Forca, Wierer & Boscato 2006), fowl, birds, freshwater molluscs, and plants. In contrast, the later Mesolithic period (Castelnovian) would have seen a decline in the occupation of mountain areas and probably a shift of human groups towards the pre-alpine hill belt and the coast along with an increase in sedentism (Broglio et al. 1987). Four main sources contributed to the definition of this model: • the specific topographic setting of sites along with their extent, finds density and the whole set of data extracted from the variably-preserved archaeological record are the basic elements for the reconstruction of this scenario; • the relations between highland and valley-bottom sites were revealed by the analysis of lithic raw material provenance. Therefore petroarchaeology has proven to be a key tool for the reconstruction of settlement dynamics by allowing exploited sources to be mapped (flint from the Mesozoic formations of the Southern Alps, flint from the Dolomite area – Marne del Puez and Livinallongo Formation, and quartz from the area located north of the Val Pusteria) and proving the existence of strict relationships between lowland and highland sites; • the seasonality of sites was assessed mainly through archaeozoological studies, especially with data obtained from the remains of foetal and new-born individuals, as well as from ungu-

4. Concluding remarks: bias factors and future perspectives As we have tried to illustrate, both the high visibility of the record and the intensity of research in the last thirty years have favoured the collection of a great bulk of data on the early Holocene occupation of the Italian Eastern Alps. Consequently an articulated settlement model for the Mesolithic groups set76

Fontana – Guerreschi – Peresani, The Visible Landscape: inferring Mesolithic settlement dynamics…

occur in densely forested environments, yield evidence of short and repeated displacements by the entire group or by specialised hunting parties (Peresani et al. 2000; Avigliano et al. 2000). Therefore, contrary to what was originally believed, these areas were points of attraction and occupation for the Mesolithic groups and must have fallen within the articulated framework of regional occupation. Secondly, the analysis of the bone assemblages from Mondeval de Sora and Plan de Frea, in which the role of red deer was emphasized since the Preboreal, has contradicted the original hypothesis that Mesolithic hunting in the highlands was ibex-specialised (Fontana et al. 2009, Govoni 2006, Angelucci et al. 1998). • Last but not least, one of the main limits of the currently favoured settlement model (Broglio 1980, 1992b) is a lack of definition, other than in very general terms (identification of the ‘vertical seasonal nomadic system’), of the mobility routes followed by the groups settled in the area. A recent study that analysed survey data from the Adige River basin (Kompatscher & Hrozny Kompatscher 2007) has begun to shed new light on this aspect. It used a methodological approach strictly based on topographical and ecological observations and took into account several factors, including the general position of sites within the surrounding territory (topography of the area, foraging radius calculated within 2 hours walk from the camp, and strategic position of the site in connection with the possibility of communicating with the surrounding territories) and criteria which may have influenced the choice of the camp location (availability of water sources, visibility over the surrounding territory, soil fitness, and wind direction). The data acquired are coherent with a type of mobility characterised by short and frequent displacements along the main ridges and mountain sides, following routes which are essentially defined by orographic barriers (fig. 3, fig. 4). This work is essentially in agreement with observations from the territory of the upper Piave Valley, where the location of identified sites corresponds to large passes and some smaller saddles connecting the main valleys, according to a distribution pattern which could reflect the possible routes covered by Mesolithic hunters (Fontana et al. 2002). To complete this model, a more refined definition of raw material provisioning areas would be needed. This requires very sophisticated systems due to the provenance of most docu-

tled in the area was developed, and this has enjoyed great success as probably one of the most detailed developed to date in European prehistory. Now, some decades later, this paradigm will need to be revised in the light of the most recent research results, which show that the variability of the evidence is much higher than was previously believed. This multifaceted evidence now seems to be a consequence of more complex dynamics which still await a more refined definition both synchronically and diachronically. 4.1 Synchronic aspects

At a synchronic level, the main issues identified are the following: • Concerning the definition of intra-site organisation and function, it is likely that too much emphasis has been placed on structural analyses and the use of ratio computation. Their reliability depends more on a detailed evaluation of all possible bias factors (does the analysed assemblage truly represent the total amount of tool-kit abandoned by the group? Does it correspond to one or more occupation phases? etc.) and on its integration with the widest possible set of available data. In particular, the application of micro-wear analyses to both lithic and, when available, bone artefacts can highlight the specific functions of some categories of tools, armatures and blanks (Lemorini 1992) and therefore the type and intensity of activities carried out at the sites (Peresani et al. in press). When associated with spatial analyses these may also support the interpretation of specific functional zones (Mondeval de Sora, Siviero & Ziggiotti unpublished; Fontana et al. 2009). • On the matter of inter-site variability, recent reconsideration of data from some contexts along with the availability of a new set of radiocarbon dates (Colbricon) appears to contradict the hypothesis that sites located within the same delimited territory are contemporaneous and should be considered to be part of the same occupation system. Therefore they suggest possible changes in settlement dynamics over time even within the Sauveterrian itself (Grimaldi 2006; Peresani & Miolo unpublished). • At the level of definition of the regional settlement pattern, two main points may be underlined. Firstly, recent research has shown that areas located around 1000 m on the pre-alpine plateaux (Altopiano del Cansiglio), which therefore 77

Hidden Landscapes

mented rocks from the same geological formations both at valley bottom and high altitude sites. We believe the social context should also be considered, given the fact that hunter-gatherer mobility strictly depends on the definition of territorial boundaries which, given a probable demographic increase in this period, would have been particularly relevant.

of the archaeological record as a consequence of post-depositional processes (destruction of the record due to the succeeding occupations), anthropic factors (possible change in mobility patterns), and the shorter duration of the Castelnovian (5960-4500 BC or 8653-7285 BP cal) vs. the Sauveterrian (7950-5765 BC or 11,009-8421 BP cal) (Broglio & Kozlowski 1983; Broglio & Improta 1994-1995). A shift from residential mobility in the Sauveterrian to a more logistical mobility was recently proposed (Fontana 2006), with only specialised task groups moving to mountain areas in the Castelnovian. Nonetheless, a greater quantity of data along with a deeper insight into the evidence from this most recent phase of the Mesolithic in Northern Italy is needed in order to shed light on the definition of behavioural patterns during this period. • Some current studies show evidence for possible changes in settlement dynamics within the Early Mesolithic itself. In addition to the data already mentioned for the context of the Colbricon area (see above and section 4.1), some technical indicators show a higher standardisation in the manufacture of armatures in relation to the topo-

4.2 Diachronic aspects

At a diachronic level two main issues are discussed: • Evidence for a change of settlement pattern from early (Sauveterrian) to late Mesolithic (Castelnovian): if the available data show clear evidence for displacement of the latter Mesolithic groups towards coastal areas, the role of highland occupation has probably been underestimated. Evidence from the site of Mondeval de Sora, particularly the discovery of a burial with excellent preservation found at site 1, located at an altitude of 2150m, indicates the important role that these areas may have continued to play (Fontana 2006). Therefore the apparent reduction of highland occupation during the Castelnovian could alternatively be explained by the lower visibility

Figure 3 – The topographic distribution of Mesolithic highland sites in the Dolomite area, within the Adige-Isarco river basin, corresponds to different types of logistical contexts. Saddles, passes, wide basins and ridges such as the Alpe di Luson/Lüsner Alp, represented in this picture, preserve traces of temporary camps which, according to Kompatscher & Hrozny Kompatscher (2007), are not distributed at random intervals, but reflect a precise strategy for optimising resource exploitation.

78

Fontana – Guerreschi – Peresani, The Visible Landscape: inferring Mesolithic settlement dynamics…

graphic and ecological location and the functional orientation of sites. During the first part of the Sauveterrian, evidence from specialised huntingrelated tasks is strictly connected to a marked presence of ‘personal indicators’ in microlith manufacture. This was presumably done by a selected group, probably of the most skilled and experienced hunters, who were also the bearers of ‘normalised handicraft practices’. In the succeeding phases a decrease of this phenomenon is recorded independently of the location and the typology of the site (Peresani & Miolo unpublished): technological indicators of manual skill become less polarised, suggesting either a lower degree of ‘normalisation’ of handicraft practises or a different group composition both in qualitative and quantitative terms. Therefore in spite of the highly visible available archaeological record we can conclude that no appropriate and unique model accounts for the great variability observed in the Mesolithic record over the investigated territory. In our opinion, the main path to follow in order to reach an increasingly realistic view of the matter consists in enlarging the range of methodologies used by embracing a truly multidisciplinary large-scale approach to data analysis, favouring more dynamic and detailed investigations of known and unknown records. We would like to underline that this conclusion does not deny the value of the results attained so far, but is simply aimed at providing a further impulse towards the improvement of the current explanatory framework.

Figure 4 – Evaluation of the settlement potentiality of a highland location may be inferred from the interpolation of two parameters, the estimate of the surrounding territory which may be exploited within one day (i.e. 2 hours’ walk from the camp) and its strategic position which could allow the Mesolithic hunters and gatherers to profit from a specific ecotonal area; the latter is estimated to cover 38 sq. m. in the case of Passo Sella (Dolomites) (after Kompatscher & Hrozny Kompatscher 2007).

anical and archaeological record in the eastern Pre-Alps: the Palughetto basin (Cansiglio Plateau, Italy). Journal of Quaternary Science 15(8), pp. 789-803. Awsiuk R., Bartolomei G., Cattani L., Cavallo C., Dalmeri G., D’Errico F., Giacobini G., Girod A., Hercman H., JardonGiner P., Nisbet R., Pazdur M.F., Peresani M. & Riedel A. 1991. La Grotta d’Ernesto (Trento): frequentazione umana e paleoambiente. Preistoria Alpina 27, pp. 7-160. Bagolini B. 1972. Primi risultati delle ricerche sugli insediamenti epipaleolitici del Colbricon (Dolomiti). Preistoria Alpina 11, pp. 201-35. Bagolini B., Broglio A. & Lunz R. 1983. Le Mésolithique des Dolomites. Preistoria Alpina 19, pp. 15-36. Bagolini B. & Dalmeri G. 1987. I siti mesolitici di Colbricon (Trentino). Analisi spaziale e fruizione del territorio. Preistoria Alpina 23, pp. 7-188. Battaglia L., Broglio A., Castelletti L., Lanzinger M. & Maspero A. 1992. Abri Soman. Preistoria Alpina 28(2), pp. 291-298. Bazzanella M., Betti L. & Wierer U. 2004. Un nouveau site sauveterrien dans la vallée de l’Adige (Bozen/Bolzano, Italie). Actes du XIVème Congrès UISPP, Université de Liège, Belgique, 2-8 septembre 2001, Section 7: The Mesolithic, BAR International Series 1302, pp. 215-226. Bazzanella M., Lanzinger M., Moser L. & Pedrotti, A. 1997. I livelli mesolitici de La Vela (Trento). Campagne di scavo 1987-88. Preistoria Alpina 33, pp. 15-16. Bertola S., Broglio A., Cassoli P.F., Cilli C., Cusinato A., Dalmeri G., De Stefani M., Fiore I., Fontana F., Giacobini G., Guerreschi A., Gurioli F., Lemorini C., Liagre J., Malerba G., Montoya C., Peresani M., Rocci Ris A., Rossetti P., Tagliacozzo A. & Ziggiotti S. 2007. L’Epigravettiano recente

References Alciati G., Cattani L., Fontana F., Gerhardinger E., Guerreschi A., Milliken S., Mozzi, P. & Rowley-Conwy, P. 1992. Mondeval de Sora: a high altitude Mesolithic camp-site in the Italian Dolomites. Preistoria Alpina 28(1), pp. 351-366. Angelucci D.E., Alessio M., Bartolomei G., Cassoli P.F., Improta S., Maspero A. & Tagliacozzo A. 1998. The Frea IV rockshelter (Selva Val Gardena, BZ). Preistoria Alpina 34, pp. 99-109. Angelucci D.E. & Peresani M. 1996. The micromorphology of some Palaeo-mesolithic Living-Floors in the Southern Alps: Preliminary Data. Extended Abstract, in Castelletti L. & Cremaschi M. (eds.), Micromorphology of deposits of anthropogenic origin. Proceedings XIII International UISPP Congress, Forlì: ABACO, Colloquia 3, pp. 161-174. Angelucci D.E. & Peresani M. 2000. The open-air sites of Val Lastari and Malga Lissandri (Venetian Pre-Alps, Northern Italy): preliminary results on spatial analysis and the study of postdepositional processes. Quaderni del Civico Museo Storico-Archeologico, Savona, pp. 47-55. Avigliano R., Di Anastasio G., Improta S., Peresani M. & Ravazzi C. 2000. A new late glacial – early Holocene palaeobot-

79

Hidden Landscapes

Epigravettiano e Mesolitico antico. Preistoria Alpina 40, pp. 99-105. Fontana F. 2006. La sepoltura di Mondeval de Sora (Belluno). Differenziazione sociale e modalità insediative degli ultimi popoli cacciatori e raccoglitori dell’Italia nordorientale, in Martini F. (ed.), La cultura del morire nelle società preistoriche e protostoriche italiane. Studio interdisciplinare dei dati e loro trattamento informatico. Dal Paleolitico all’età del Rame. Istituto Italiano di Preistoria e Protostoria. Origines, Progetti 3, Firenze pp. 269-292. Fontana F., Govoni L., Guerreschi A., Padoanello S., Siviero A., Thun Hohenstein U. & Ziggiotti S. 2009. L’occupazione sauveterriana di Mondeval de Sora 1, settore I (San Vito di Cadore, Belluno) in bilico tra accampamento residenziale e campo da caccia. Preistoria Alpina 44, pp. 207-226. Fontana F. & Guerreschi A. 2003. Highland occupation in the Southern Alps, in Larsson L. (ed.), Mesolithic on the move. Proceedings of the 6th International Conference on the Mesolithic in Europe, Stockholm, Oxbow Books, Oxford, pp. 96-102. Fontana F., Guerreschi A. & Reberschak M. 2002. Nuovi dati sul popolamento dell’alta valle del Cordevole nel Mesolitico, Quaderni di Archeologia del Veneto 17, pp. 15-21. Fontana F. & Pasi E. 2002. Risultati delle ultime prospezioni nell’area di Mondeval de Sora (San Vito di Cadore, Belluno). Quaderni di Archeologia del Veneto 18, pp. 21-30. Fontana F. & Vullo N. 2000. Organisation et fonction d’un camp de base saisonnier au coeur des Dolomites: le gisement mésolithique de Mondeval de Sora (Belluno, Italie), in Richard A., Cupillard C., Richard H. & Thévenin A. (eds), Les derniers chasseurs-cueilleurs d’Europe occidentale. Annales Littéraires 699, Environnement, sociétés et archéologie 1, pp. 197-208. Govoni L. 2006. Le associazioni faunistiche a grandi mammiferi della Grotta del Romito (Papasidero, CS) e del sito VF1 di Mondeval de Sora (Val Fiorentina, BL) indicatrici delle variazioni climatiche del Tardoglaciale e dell’Olocene antico. Tesi di Dottorato di Ricerca in “Sistemi biologici: struttura, funzione ed evoluzione”, Ciclo XIX, Università degli Studi di Ferrara. Grimaldi S. 2006. Radiocarbon dating of the early Mesolithic Colbricon site. Journal of Anthropological Sciences 84, pp. 137-145. Guerreschi A. (ed.) 1996. Il sito preistorico del Riparo di Biarzo, Valle del Natisone, Friuli. Museo Friulano di Storia Naturale 39. Kompatscher K. & Hrozny Kompatscher N.M. 2007. Dove piantare il campo: modelli insediativi e di mobilità nel Mesolitico in ambiente alpino. Preistoria Alpina 42, pp. 137-161. Kozlowski S.K. & Dalmeri G. 2000. Riparo Gaban: the Mesolithic layers. Preistoria Alpina 36, pp. 3-42. Lanzinger M. 1985. Ricerche nei siti mesolitici della Cresta di Siusi (Auf der Schneide, siti XV e XVI dell’Alpe di Siusi) nelle Dolomiti. Considerazioni sul significato funzionale espresso dalle industrie mesolitiche della regione. Preistoria Alpina 21, pp. 33-48. Lemorini C. 1992. Etude fonctionnelle des industries mésolithiques de Lago delle Buse 1 et Lago delle Buse 2 (Lagorai, Trentino) par la méthode des traces d’utilisation. Preistoria Alpina 28(1), pp. 51-59.

nell’area prealpina e alpina orientale, in Martini F. (ed.), L’Italia tra 15.000 e 10.000 anni fa. Cosmopolitismo e regionalità nel Tardoglaciale. Millenni, Studi di Archeologia Preistoria, Museo Fiorentino di Preistoria “Paolo Graziosi”, Firenze, 5, pp. 39-94. Broglio A. 1973. L’Epipaléolithique de la Vallée de l’Adige. L’Anthropologie 77(1-2), pp. 5-34. Broglio A. 1980. Culture ed ambienti della fine del Paleolitico e del Mesolitico nell’Italia nord-orientale. Preistoria Alpina 16, pp. 7-29. Broglio A. 1985. Ecologia del Mesolitico Atesino, in Moroni A., Anelli A. & Ravera O. (eds), Atti del II Congresso Nazionale Società Italiana di Ecologia, Zara Edizioni, pp. 939-941. Broglio A. 1992a. Le Mésolithique des Dolomites. Preistoria Alpina 28(1), p. 311-316. Broglio A. 1992b. Mountain sites in the context of North-East Italian Upper Palaeolithic and Mesolithic. Preistoria Alpina 28(1), pp. 293-310. Broglio A., Corai P. & Lunz R. 1982. Risultati preliminari delle prospezioni nei siti mesolitici della Val Gardena e degli scavi al Plan de Frea. Bullettin d’Études Préhistoriques Alpines XV, pp. 19-53. Broglio A., De Stefani M. & Peresani M. 2006. I siti mesolitici di cima XII (Altopiano dei Sette Comuni), in Pessina A. & Visentini P. (eds), Preistoria dell’Italia settentrionale. Studi in ricordo di Bernardino Bagolini. Museo Friulano di Storia Naturale, pp. 43-58. Broglio A., Bavero V. & Marsale S. 1987. Ritrovamenti mesolitici intorno alla laguna di Venezia. Istituto Veneto di Scienze, Lettere ed Arti, Rapporti e Studi 10, pp. 195-231. Broglio A. & Improta S. 1994-1995. Nuovi dati di cronologia assoluta del Paleolitico superiore e del Mesolitico del veneto, del trentino e del Friuli. Atti dell’Istituto Veneto di Scienze, Lettere ed Arti, Classe di Scienze Fisiche, Matematiche e Naturali CLIII, pp. 1-45. Broglio A., & Kozlowski S.K. 1983. Tipologia ed evoluzione delle industrie mesolitiche di Romagnano III. Preistoria Alpina 19, pp. 93-148. Broglio A., & Lanzinger M. 1990. Considerazioni sulla distribuzione dei siti tra la fine del Paleolitico superiore e l’inizio del Neolitico nell’Italia nord-orientale, in P. Biagi (ed.), The Neolithisation of the Alpine Region. Monografie di Natura Bresciana 13, pp. 53-69. Clark R. 2000. The Mesolithic hunters of the Trentino. A case study in hunter-gatherer settlement and subsistence from Northern Italy. British Archaeological Reports, International Series 832. Cusinato A., Dalmeri G., Fontana F., Guerreschi A. & Peresani M. 2003. Il versante meridionale delle Alpi durante il Tardiglaciale e l’Olocene antico: mobilità, sfruttamento delle risorse e modalità insediative degli ultimi cacciatori-raccoglitori. Preistoria Alpina 39, pp. 129-142. Dalmeri G. & Lanzinger M. 1992. Risultati preliminari delle ricerche nei siti mesolitici del lago delle Buse, nel Lagorai (Trentino). Preistoria Alpina 28(1), pp. 317-349. Dalmeri G. & Pedrotti A. 1992. Distribuzione topografica dei siti del Paleolitico superiore finale e Mesolitico in Trentino Alto-Adige e nelle Dolomiti Venete (Italia). Preistoria Alpina 28(2), pp. 247-267. Fiore I. & Tagliacozzo A. 2004. Riparo Cogola: il contesto paleoecologico e lo sfruttamento delle risorse animali tra

80

Fontana – Guerreschi – Peresani, The Visible Landscape: inferring Mesolithic settlement dynamics…

Orombelli G. & Ravazzi C. 1996. The late glacial and early Holocene: chrononology and paleoclimate. Il Quaternario 9(2), pp. 439-444. Peresani M. & Angelini A. 2002. Il sito mesolitico di Casera Davià II sull’Altopiano del Cansiglio (Prealpi Venete). Rivista di Scienze Preistoriche LII, pp. 197-230. Peresani M., Di Anastasio G. & Bertola S. 2000. Épigravettien récent et Mésolithique ancien dans un contexte préalpin: les données du haut Plateau du Cansiglio (Italie du Nord), in Crotti, P. (ed.), MESO ‘97, Actes de la Table ronde Épipaléolithique et Mésolithique, Cahiers d’Archéologie Romande 81, pp. 267-276. Peresani M., Ferrari S. & Ziggiotti S., in press. Primi dati sul sito di Casera Lissandri 17 (Altopiano del Cansiglio) e sull’industria mesolitica in selce e cristallo di rocca. Quaderni di Archeologia del Veneto. Pessina A. 2005. Pramollo, Doss Confine: ricerche 2004-2005. Un accampamento stagionale di cacciatori preistorici. Gortania 27, pp. 49-67. Ravazzi C., Peresani M., Pini R. & Vescovi E. 2007. Il Tardoglaciale nelle Alpi e in Pianura Padana. Evoluzione stratigrafica, storia della vegetazione e del popolamento antropico. Il Quaternario 20(2), pp. 163-184. Rowley-Conwy P. 1996. Resti faunistici del Tardiglaciale e dell’Olocene antico, in Guerreschi A. (ed.), Il sito preistorico del Riparo di Biarzo, Valle del Natisone, Friuli. Museo Friulano di Storia Naturale 39, pp. 61-80. Tinner W. & Vescovi E., in press. Ecologia ed oscillazioni del limite della foresta nelle Alpi dal pleniglaciale al presente. Studi Trentini di Scienze Naturali – Acta Geologica. Vescovi E., Ravazzi C., Arpenti E., Finsinger W., Pini R., Valsecchi V., Wick L., Ammann, B. & Tinner, W. 2007. Interactions between climate and vegetation during the Lateglacial period as recorded by lake and mire sediment archives in Northern Italy and Southern Switzerland. Quaternary Science Reviews 26(11-12), pp. 1650–1669. Wierer U. & Boscato P. 2006. Lo sfruttamento delle risorse animali nel sito mesolitico di Galgenbühel/Dos de la Forca, Salorno (Bz): la macrofauna, in U. Tecchiati & B. Sala (eds), Studi di Archeozoologia in onore di A. Riedel, Bolzano, pp. 85-98.

81

10 Toward an understanding of archaeological visibility: the case of the Trentino (Southern Alps) Fabio Cavulli, Stefano Grimaldi, Annaluisa Pedrotti & Diego E. Angelucci Laboratorio di Preistoria ‘B. Bagolini’, Dipartimento di Filosofia, Storia e Beni Culturali, Piazza Venezia 41, 38122, Trento (Italy) – [email protected]; [email protected]; [email protected]; [email protected] Abstract This article deals with the topographic setting of late Upper Palaeolithic and early Mesolithic sites (corresponding to the final Epigravettian and Sauveterrian cultures, respectively) in relation to three environmental characteristics: slopes, modern land use, and the presence of slope, alluvial or glacial deposits. Together with a pair of other papers (see Cavulli & Grimaldi 2007 and 2009), this work is the virtual extension of two other studies that concentrated mainly on the differences in elevation and site typology (Bagolini & Pedrotti 1992; Dalmeri & Pedrotti 1992). Whereas we then considered a large region made up from the mountain areas of Veneto, Alto-Adige/Südtirol and Trentino, we now consider a more restricted area (Trentino) because of the limited availability of basic environmental digital data, but also because we hope to extend our current method of analysis to a larger region in the future.

1. General setting

be due to the morphology of the landscape instead of to conscious decision-making. As already suggested by previous authors, materials for Epigravettian lithic industries were mainly knapped from local flint (Broglio et al. 1995) that, eventually, might be hoarded and stored in-site (Peresani 2003); the off-site use of products made from local flint is also evident (Broglio & Improta 1995). Sauveterrian (early Mesolithic) human groups peopled the alpine environment from the beginning of the Preboreal to the end of the Boreal zones, and seem to have occupied the whole of the Trentino region. The early Mesolithic site distribution seems denser than that of previous Palaeolithic groups, and the elevation distribution of archaeological evidence in the Trentino — taking into account both excavated sites and surface finds of lithic artefacts — is characterized by the presence of sites on the Adige river valley floor, in the foothills, and in many areas situated at around 2000m asl (see review in Bagolini 1980; Bagolini & Pedrotti 1992; Dalmeri & Pedrotti 1992; Dalmeri et al. 2001). While the former are mainly rock shelters characterized by thick deposits and frequently showing complete Mesolithic stratigraphic sequences, the latter consist of small concentrations of lithic artefacts, with little or no clear evidence of settlement features. Furthermore, faunal remains have been rarely preserved at high altitude sites. S.G.

Northeastern Italy represents one of the archaeologically richest areas in the Alps. Several hundreds of archaeological sites (excavations in well-stratified deposits as well as surface finds), attributed to the late Upper Palaeolithic/early Mesolithic, are known. Field data have been gathered over decades and there is therefore an abundant bibliography (see, for example, references in Leonardi 1963; Broglio 1972, 1973, 1982, 1994a, b, 1995a, b; Broglio & Improta 1995; Broglio & Lanzinger 1990, 1996; Alessio et al. 1977, 1984; Bagolini 1980a, b, 1982; Bagolini & Broglio 1985; Bagolini & Pasquali 1984; Bagolini et al. 1984; Biagi 1995; Dalmeri & Lanzinger 2001; Dalmeri et al. 2001; Lanzinger 1987, 1991). After the Last Glacial Maximum in the Alps, the first presence of human groups in the Trentino, attributable to the Epigravettian, is dated to about 11,200 uncal. BP (Riparo Dalmeri; Bassetti et al. 2002; see also Broglio & Improta 1995). It has been suggested that the entry of human groups into the mountain area took place in the south as a response to the climatic change brought about by the lateglacial (Bagolini & Broglio 1985). This climate change caused a movement towards the northern alpine pastures, the characteristic habitat of gregarious ungulates. Ibex in particular was presumably the staple of subsistence of these prehistoric groups, as testified by the faunal remains from sites such as Riparo Dalmeri (Fiore et al. 2002). Epigravettian rock-shelters are mainly situated in the pre-Alps, while most of the Epigravettian open-air sites in mountain environments are situated near small lakes. Moreover, the elevation distribution of the Epigravettian sites suggests a settlement pattern characterized by the occupation of increasingly higher altitudes. It is worth mentioning that this may

2. Methods Considering the Pleistocene-Holocene transition and our earlier study which explains the distribution of sites by reference to an exploitation strategy, we asked ourselves if other factors could have influenced the archaeological record (fig. 1). 83

Hidden Landscapes

e Territorio (SIAT) of Provincia Autonoma di Trento. • The CORINE Land Cover map (reference scale 1:10,000), simplified to seven main land use classes: Agriculture, Forest, Open uncultivated, Urban areas, Water basins, Glaciers and Marshy areas. • The Lithological formation group map, from which the Quaternary sediments of ‘slope, alluvial and glacial deposits’ (here after named ‘SAG’) were selected. Lacking better cartography this is the only map showing the position and extent of recent covering sediments (Bosellini et al. 1999). Although shown as a single legend unit, from an archaeological point of view this has to be split up into three different polygon types: Slope, Alluvial and Glacial deposits, respectively. While the former two are relevant to post-depositional processes affecting archaeological visibility, the glacial deposits precede the formation of archaeological sites. Conscious of this limitation,

The landscape information available for a study of site location choice is limited to modern geographical data such as land use, slopes, elevation, groups of geological formations, and a few others. Recent soil and geomorphological maps are only available for parts of the region. Land use, slope morphology, and the presence of Holocene surface sediments seem to us to be the features most relevant in relation to post-depositional processes and the (in-) visibility of evidence. These aspects were studied in relation with the three main elevation ranges (0-500m, 500-1500m, above 1500m), which other authors had already found to have a peculiar relation to the distribution of these sites (Bagolini & Pedrotti 1992; Dalmeri & Pedrotti 1992). The following cartography (in Gauss-Boaga projection) was used for the analysis: • A slope map of the region (raster and vector) created from the Digital Terrain Model 10m, which has a reference scale of 1:10,000. The DTM has been Created by Sistema Informativo Ambiente

Figure 1 – Late Upper Palaeolithic (Epigravettian) and early Mesolithic (Sauveterrian) sites of the Trentino region.

84

Cavulli – Grimaldi – Pedrotti – Angelucci, Toward an understanding of archaeological visibility…

tom to over 3000m, often with abrupt transitions between altitudinal belts. The geographical setting of the Trentino is closely linked to its rather varied geological and structural layout. Three main geological domains may be recognized (Bosellini et al. 1999): • Southern Alpine terrains extend along the central and eastern sectors of the region and comprise pre-Permian crystalline basement materials (such as the metamorphic rocks outcropping next to the Cima d’Asta granite) and Permian to Tertiary covers. These range from the volcanic rocks of the so-called Adige porphyric platform to the sedimentary ones, mostly Mesozoic carbonates, which form the bulk of Dolomites; • Austroalpine terrains are mostly metamorphic and outcrop in the NW corner of the region; and • The Adamello batholith is formed of granitoid lithotypes dating to the Cenozoic and spreads out along the eastern border of the Trentino. The relief is mainly controlled by geological factors, even if the present landscape strongly depends on dynamics relating to the last glacial – interglacial cycle. Structural and geological factors are responsible for the high relief energy and orientation of the main hydrographic and orographic axes. The two major drainage systems (Sarca and Adige rivers) were formed by the intense erosion triggered by the Miocene Messinian crisis. Nevertheless, the most obvious landscape features today are due to glacial (and periglacial) morphogenesis, which intensely modelled the landscape during the coldest phases of the Pleistocene when glaciers were flowing from the upland through the valleys and into the Po Plain. Interglacial morphogenesis was controlled by fluvial and slope activity, with the formation of relict slopes and terraces often still recognizable along valley flanks. The last glacial cycle is the best known. The preWürmian drainage pattern was significantly different from the present one, with the palaeo-Adige flowing through what is today the Sarca valley. At the Alpine Last Glacial Maximum (ALGM, circa 20,000 BP), Trentino was almost completely subject to glacial and periglacial conditions, with glacial bodies thicker than 1 km occupying the main Adige, Sarca and Brenta valleys. At this time the Alpine margin was experiencing steppe-like conditions, and main reliefs (including prealpine plateaux) were subject to periglacial environments. The pre-ALGM surface inside the Adige valley lies some 200m below the present surface, as detected by coring at Trento (Felber et al. 2000). The importance of these ice-

we tested the relevance of this variable for the distribution of archaeological records. The area considered in this study is restricted to Trentino province; the results of future studies will be exported to the regions of Alto-Adige/Südtirol and Veneto. Studying a limited area provides the opportunity of testing our exploratory analysis in a homogeneous morphology; at a later stage this can be adapted to a larger area considering only the specific morphologies (i.e. minimum/maximum elevations of valley bottoms and peaks. In the tables and histograms below, the relative site densities of areas (such as forested areas above 1500m asl or slopes with a gradient of between 15° and 35° below 500 m asl, etc.) are considered more important than the absolute densities or frequency. Nonetheless we provide the frequencies and percentages for the interested reader. In the present analysis, all bibliographical evidence relating to the late Upper Palaeolithic and early Mesolithic periods is considered, irrespective of whether they are made up by a deposit or by a few scattered finds. The flint artefacts attributed generically to the Mesolithic (89 examples) or to the ‘Palaeo-Mesolithic’ period (7 examples) were considered only towards the end of the work when comparison with chronologically precisely attributed finds was possible. 22 records can be related to the Epigravettian culture (Upper Palaeolithic period), and 48 to the Sauveterrian (early Mesolithic); the late Mesolithic Castelnovian has not been considered in this work. F.C.

3. A brief overview of the geology and geomorphology of the Trentino region The Trentino region is located in the eastern sector of the southern Alps and mostly coincides, together with South Tirol, with the mountain reaches of the Adige river drainage basin. Trentino’s eastern part (e.g., the Sarca valley) drains directly into the Po river further south. The physiography of the region is rather diverse, as a result of its geological complexity and the polygenetic Quaternary landscape evolution. From a geographical point of view, the Adige and Sarca valleys form the ‘backbones’ of the region. Both valleys are oriented roughly NorthSouth and cross the whole southern Prealpine and Alpine ranges, linking the area to the northern slope of the Alps as well as (through East-West oriented valleys such as Valtellina, Valpusteria, and Valsugana) to other segments of the Alpine chain. Altitude ranges from 200m in the Adige valley bot85

Hidden Landscapes

the Holocene. Following intense slope activity in the early Holocene (recorded at several sites located in footslopes, like Romagnano or Vatte di Zambana), geomorphological surfaces progressively stabilized and the drainage systems re-established equilibrium by infilling of depressions such as those at Trento and Valle dei Laghi (‘Valley of the Lakes’ in the Sarca drainage). Stability and biostasy up to circa 2000m have been observed during this time span, matching the abundant archaeological evidence from the Mesolithic in mountain areas. Soil formation continued throughout the Atlantic until it was interrupted by the beginning of ‘Neoglaciation’, the effects of which were often magnified by the first human impact on the landscape (Cavulli et al. 2002). As a result of the complex geological layout and late Quaternary morphodynamics, the Trentino region is composed of a mosaic of juxtaposed and superimposed landscape units. In general, the region presents dramatic landscape contrasts and high relief energy. Floodplains (less than 10% of the total surface of the Trentino) are limited to the main valley bottoms; the latter are often constrained by steep slopes with compound profiles reaching mountain areas. Talus scree and cones are found at the foot of these slopes. Land over 1500m amounts to almost 50% of the surface and abrupt periglacial-related morphologies are common, as well as structural plateaux or depressions derived from the still unattained hydrological equilibrium. The upland areas have been subject, during the Holocene, to erosion or non-sedimentation, explaining the general shallowness and near-surface position of the archaeological evidence. By contrast, the lower-lying areas were subject to intense sedimentation through the lateglacial and early Holocene, with consequent accumulation of thick successions that may embed significant archaeological sequences but, at the same time, have buried and hidden them from view. D.E.A.

slope ° SAG

Soil Use Type

0-5

present agriculture

0-5

present

0-5

absent forest

0-5

absent

5-15

absent agriculture

5-15

open uncultivated open uncultivated

Upper Palaeo. Early Meso. sites

related dynamics means that the ALGM, for most of the Trentino territory, has to be regarded as time ‘zero’ for most of the archaeological record. The deglaciation was a time-transgressive process that took place between circa 16,000 and 14,500 BP at the Alpine margin, and around 12,000 BP in mountain areas, the mid-altitude belt being nearly deglaciated around 14,000 BP (Casadoro et al. 1976; Avigliano et al. 2000; Pellegrini et al. 2005). At this time, intense discharge from the glacial fronts led to the accumulation of thick sedimentary prisms in the main valleys and in the Po Plain. Other relevant processes of the lateglacial include loess deposition in the Prealps (probably before GI1), soil formation between circa 11,500-11,000 BP (during GI1, former Bølling and Allerød zones), and the shifting of biological communities towards mid-altitude uplands. Human occupation reached the mid-altitude Prealpine plateaux during GI1. The GS1 (i.e. Younger Dryas) cold shift caused an interruption of these trends, another possible phase of wind-blown dust sedimentation, the reduction of Prealpine human occupation to ‘refugee’ areas, and intense vertical aggradation due to fluvial activity in valley bottoms (Fuganti et al. 1998; Angelucci & Bassetti 2009). The climatic and environmental trends observed during GI1 were renewed at the onset of

1

1500 1500

1

1

1

2

2

4

2

2

3

1

2

absent forest

13

2

7

4

5-15

open absent uncultivated

11

2

4

5

15-35

present forest

5

2

3

15-35

present urban area

1

15-35

absent agriculture

1

1

15-35

absent forest

8

1

5

2

15-35

absent

35-90

absent forest

35-90

absent

35-90

absent urban area

open uncultivated open uncultivated

4. Palaeolithic and Mesolithic finds and the environment (table 1)

1

17

How can we explain the observed distribution of finds/deposits: as a direct reflection of exploitation strategies, or as the result of a combination of research limitations and such exploitation strategies? It is possible that the landscape itself, its transformation through time and post-depositional processes, has affected the archaeological record. It is probable that vegetation and slope and alluvial sediments now cover archaeological deposits and evidence; in other

17

1

1

1

1

1

1

Table 1 – Summary characteristics of Upper Palaeolithic and early Mesolithic sites of the Trentino region.

86

Cavulli – Grimaldi – Pedrotti – Angelucci, Toward an understanding of archaeological visibility…

talus cones. Today, because of extraction activities, they are on vertical walls.

words, the possibility of finding traces of them today may be severely limited. Therefore, we would like to verify to what extent the spatial distribution of the archaeological record corresponds to a valid picture of the past. We will first describe the general elevation, slope, SAG and land-use properties of our site sample, then go into more detail regarding Late Upper Palaeolithic and early Mesolithic sites in sections 4.1 and 4.2.

Slope, alluvial and glacial (SAG) deposits (fig. 4)

The few sites that are located on SAG deposits are located on cone and scree deposits above the main valley floor such as La Vela, Romagnano Loch; on high elevation saddles such as Passo S. Barbara, Malga delle Buse del Sasso; or close to ancient elevated lakes such as Viotte. Combining the data on slopes and those on the presence of SAG deposits,

Elevation (fig. 2)

Taking the whole sample of Epigravettian and Sauveterrian sites together, the site count increases with higher altitudes (almost 50% of sites are found above 1500 m), but the highest density is located between 90 and 500 m asl.

0-500 m 0.013 0.021

Slope (fig. 3)

sites/km2

With regard to the variability of slope inclination in the region and the classification of site frequency into three main categories (0-15, 15-35, and 35-90 degrees), it comes as no surprise to find a higher site density in flat areas. What is more surprising is to find three sites above the angle of repose (35 degrees): the early Mesolithic sites of Vatte, Pradestel and Bus de la Vecia are shelters placed on vertical rock walls a few metres above the Adige Valley, and linked to it by means of steep

0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

500-1500 m

0.008 > 1500 m

Elevation m 0-500 500-1500 > 1500

Area km2

Area %

559.3 2955.3 2693.8

9.0 47.6 43.4

Freq. of sites 12 24 34

Sites % 17.143 34.286 48.571

Density (Sites/km2) 0.021 0.008 0.013

Figure 2 – Site density (per km2) per elevation zone.

0-500 m 500-1500 m > 1500 m

0-15

15-35

35-90

degree

Elevation m

Slope °

Area km2

Area %

# Sites

Sites %

Sites/km2

0-500 0-500 0-500 500-1500 500-1500 500-1500 > 1500 > 1500 > 1500

0-15 15-35 35-90 0-15 15-35 35-90 0-15 15-35 35-90

369.5 147.0 42.8 891.8 1741.3 322.0 469.1 1810.1 414.0

6.0 2.4 0.7 14.4 28.0 5.2 7.6 29.2 6.7

6 3 3 17 7 0 12 22 0

8.571 4.286 4.286 24.286 10.000 0.000 17.143 31.429 0.000

0.016 0.020 0.070 0.019 0.004 0.000 0.026 0.012 0.000

Figure 3 – Site density (per km2) per elevation zone for each slope class.

87

Hidden Landscapes

0.045 0.040 0.035

SAG sediments

absent

0.030 0.025 0.020 0.015 0.010 0.005 0.000 0-500 m

500-1500 m

> 1500 m

Elevation m

SAG deposits

Area km2

Area %

# Sites

Sites %

Sites/km2

0-500 500-1500 > 1500 0-500 500-1500 > 1500

present present present absent absent absent

324.6 561.3 203.1 234.7 2394.0 2490.7

5.2 9.0 3.3 3.8 38.6 40.1

2 2 4 10 21 30

2.899 2.899 5.797 14.493 30.435 43.478

0.006 0.004 0.020 0.043 0.009 0.012

2500 Agriculture

km2

2000 Forest

1500

Urban area

1000 500

open uncultivated

0 0-500

500-1500

> 1500

m asl

0-5

present

open uncultivated

1

0-5

absent

forest

2

2

0-5

absent

open uncultivated

2

2

5-15

absent

agriculture

1

1

5-15

absent

forest

6

5

5-15

absent

open uncultivated

5

15-35 present

forest

1

1

absent

forest

4

3

15-35

Figure 5 – Area (in km2) of modern landuse per elevation zone in the Trentino region.

>1500 m

Upper Palaeo. sites

Land use

500500 m

Slope SAG (°) deposits

Modern Land Use in Trentino

1500 m

0.01 0 Agriculture

Urban area

Forest

Open uncultivated

Land use

Area km2

Area %

# Sites

Sites %

Sites/km2

Agriculture Forest Open uncultivated Urban area Water basins Glaciers Marshy area

696.3 3159.4 1876.9 152.3 38.8 55.4 0.3

11.2 50.9 30.2 2.5 0.6 0.9 0.0

5 29 34 2 0 0 0

7 42 48 3 0 0 0

0.007 0.009 0.018 0.013 0.000 0.000 0.000

Figure 6 – Site density (per km2) per elevation zone for each land use type.

88

Cavulli – Grimaldi – Pedrotti – Angelucci, Toward an understanding of archaeological visibility…

Colle dei Meneghini

Viotte

Colle Val d'Antenne

Malga Artillione

Pian dei Laghetti II

Pian dei Laghetti I

Passo delle Fittanze

Malga Scura

Albi Val Coperte

Palù Echen

Colle dei Colombi

Riparo Dalmeri

Laghetto delle Regole LR3

Laghetto delle Regole LR2

Laghetto delle Regole LR1

Augustello

Passo S.Barbara

Riparo La Cogola

Andalo

1500

Carbonare-Scomp

2000

Madonna della Neve

2500

Terlago

1000

500

0 Figure 7 – Elevations of late Upper Palaeolithic sites. White symbols indicate sites that continue into the Sauveterrian.

the sites turn out to be generally located on sub-flat surfaces without SAG sediment. 0.001

Land use (figs 5, 6)

A large part (more than 50%) of the study region is covered by forest, but the main sites density is recorded in open uncultivated areas (30% by area) and urban areas (town, villages, industrial and extraction areas, roads, etc., altogether 2.5% of the area). However, the number of sites in forested land is still quite high, demonstrating the possibility of recovering finds/sites even in difficult visual conditions. Considering the differences in elevation, it is clear that sites above 1500m asl are in pastures or in forests; areas occupied by agriculture or villages are scarce at this altitude. Sites between 500 and 1500m asl are mainly in forested areas – forest being the most common type of land use at these elevations – and secondarily in cultivated or uncultivated open areas. Below 500m asl the sites seem to have no ‘preference’ for a particular land use.

0.002

0-500 m 500-1500 m

0.006

Elevation m 0-500 500-1500 > 1500

Area km2 559.3 2955.3 2693.8

> 1500 m

Area % 9.0 47.6 43.4

# Sites 1 18 3

Sites % 5 82 14

Sites/ km2 0.002 0.006 0.001

Figure 8 – Density of late Upper Palaeolithic sites (per km2) per elevation zone.

ably because of the strategic position between two mountains without a high altitude passage: Monte Bondone, where the site of Viotte is located, and Monte Paganella with the site of Andalo. Indeed, this south-north line appears to mark a significant route into the Southern Alps, from Monte Baldo to the end of the Non Valley (Le Regole site). Evidently, Palaeolithic sites preferred to be located in flat areas at elevations between 1000 and 1600m asl (figs 8, 9). The case of Terlago at 450m asl shows that it is possible to find Epigravettian

4.1 Late Upper Palaeolithic sites (table 2)

Once the whole sample is analysed it is worth having a look at how sites divided by chronology act. As mentioned above, the Epigravettian sites are located at middle elevations, between 1000 and 1600m asl (fig. 7). The site of Terlago is exceptional, prob89

Hidden Landscapes

4.2 Early Mesolithic sites (table 3)

evidence even below 1000m. In spite of two cases of sites in SAG sediments (Passo S. Barbara and Viotte), the most commonly selected locations are SAG-free and are often in forested or pasture areas (figs 10, 11).

Five early Mesolithic finds are located between 1000 and 1238m asl. This group is important because it has not been studied previously and because there appear to be clear ‘gaps’ in between the finds be-

0.018 SAG present

0.014

0-500 m

0.008

0.012

500-1500 m

0.007

0.010

SAG absent

0.006

> 1500 m

0.008

sites/km2

Sites/km2

0.016

0.006 0.004 0.002

0.005 0.004 0.003 0.002

0.000

0.001 0°-15°

15°-35°

35°-90°

0.000

degrees

0-500 m

Elevation m 0-500 0-500 0-500 500-1500 500-1500 500-1500 > 1500 > 1500 > 1500

Slope ° 0-15 15-35 35-90 0-15 15-35 35-90 0-15 15-35 35-90

Area km2 369.5 147.0 42.8 891.8 1741.3 322.0 469.1 1810.1 414.0

Area % 6.0 2.4 0.7 14.4 28.0 5.2 7.6 29.2 6.7

# Sites 1 0 0 14 4 0 2 1 0

Sites % 5 0 0 64 19 0 10 5 0

Sites / km2 0.003 0.000 0.000 0.016 0.002 0.000 0.004 0.001 0.000

0-5

present

1

0-5

absent

5-15

absent

agriculture

2

1

1

5-15

absent

7

2

2

5-15

absent

forest open uncultivated

6

1

15-35

present

forest

4

15-35

present

urban area

1

1

15-35 15-35

absent absent

1 4

1 1

15-35

absent

agriculture forest open uncultivated

35-90

absent

forest

1

1

35-90

absent

open uncultivated

1

1

35-90

absent

urban area

1

1

0.025

2

2

324.6 561.3 203.1 234.7 2394.0 2490.7

Area %

# Sites

Sites %

5.2 9.0 3.3 3.8 38.6 40.1

0 1 1 1 17 2

0 5 5 5 77 10

Sites/ km2 0.000 0.002 0.005 0.004 0.007 0.001

< 500 m

0.020

500-1500 m

0.015

> 1500 m

0.010 0.005

3

0.000 Agriculture

5 1

17

present present present absent absent absent

Area km2

> 1500 m

Figure 10 – Density of late Upper Palaeolithic sites (per km2) per elevation zone for areas with/without SAG (slope, alluvial and glacial) deposits.

1

2

0-500 500-1500 > 1500 0-500 500-1500 > 1500

sites/km2

agriculture open uncultivated

Early Meso. sites

SAG deposits

>1500 m

SAG Land use deposits

500-1500 m

Slope (°)

1500 m 0.040 0.020 0.000

500-1500

0.020

0°-15°

15°-35°

> 1500 Elevation m

0.002

Elevation m 0-500 500-1500 > 1500

Area km2 559.3 2955.3 2693.8

Area % 9.0 47.6 43.4

# Sites 11 6 31

Sites % 23 13 65

35°-90°

degree

0-500 0-500 0-500 500-1500 500-1500 500-1500 > 1500 > 1500 > 1500

Sites/ km2 0.020 0.002 0.012

Slope ° 0-15 15-35 35-90 0-15 15-35 35-90 0-15 15-35 35-90

Area km2

Area %

# Sites

Sites %

369.5 147.0 42.8 891.8 1741.3 322.0 469.1 1810.1 414.0

6.0 2.4 0.7 14.4 28.0 5.2 7.6 29.2 6.7

5 3 3 3 3 0 10 21 0

10 6 6 6 6 0 21 44 0

Sites/ km2 0.014 0.020 0.070 0.003 0.002 0.000 0.021 0.012 0.000

Figure 14 – Density of Mesolithic sites (per km2) per elevation zone for each slope class.

Figure 13 – Density of early Mesolithic sites (per km2) per elevation zone.

91

Hidden Landscapes

(fig. 15). The high density of sites in low-lying urban areas is of course due to modern construction and extraction activities during the 1970s and 1980s, which led to the discovery of many sites (fig. 16). The largest group of finds generically attributed to the Mesolithic or Palaeo-Mesolithic period are from above 1500m asl (fig. 17). These scattered flints, usually bladelets, without any diagnostic characteristics (formal tools) are interpreted as Mesolithic finds only because they were found on high altitude mountains. Even if in this zone it is not easy to collect finds from other periods, there is still no proof that this attribution is correct. The difference between the high site frequency in the eastern part of the region and the lower frequency to the west (fig. 18) appears to be due only partially to morphology (except for the Non valley in the north-east, the rest of east Trentino presents steep and high slopes) and mainly to a lack of research. F.C.

high density between 35 and 90 degrees is due to the low altitude shelters already discussed above. There is a high density of Mesolithic finds at low elevations without SAG sediments, but there is another group at altitudes over 1500m asl where these sediments are present — probably because many of these are located about 2000m where erosion and scree are abundant

0.035

SAG present

density (sites/km2)

0.030 SAG absent

0.025 0.020 0.015 0.010 0.005 0.000 0-500 m

Elevation m 0-500 500-1500 > 1500 0-500 500-1500 > 1500

SAG deposits present present present absent absent absent

500-1500 m

Area km2 324.6 561.3 203.1 234.7 2394.0 2490.7

Area % 5.2 9.0 3.3 3.8 38.6 40.1

> 1500 m

# Sites 2 1 3 7 5 28

Sites % 4 2 6 15 10 58

Sites/ km2 0.006 0.002 0.002 0.030 0.002 0.011

5. Conclusions Late Upper Palaeolithic and early Mesolithic sites in the Trentino are usually found in flat or sub-flat

2500

Figure 15 – Density of Mesolithic sites (per km2) per elevation zone for areas with/without SAG (slope, alluvial and glacial ) deposits.

2000 1500

< 500 m

0.045

Generic “Mesolithic” sites

> 1500 m

0.035 Sites/km2

1000

500-1500 m

0.040

500

“Palaeomesolithic” sites

0.030 0.025

0

0.020 0.015

Figure 17 – Altitude distribution of sites generically attributed to the Mesolithic and to the Palaeolithic or Mesolithic.

0.010 0.005 0.000

Land use Open uncultivated Forest Agriculture Urban area Water basins Glaciers Marshy areas

Agriculture

Area km2 1876.9 3159.4 696.3 152.3 38.8 55.4 0.3

Area % 30.2 50.9 11.2 2.5 0.6 0.9 0.0

Forest

# Sites 26 16 4 2 0 0 0

Open uncultivated

Sites % 54 33 8 4 0 0 0

km2

Urban area

Sites/ km2 0.014 0.005 0.006 0.013 0.000 0.000 0.000

Figure 16 – Density of early Mesolithic sites (per km2) per elevation zone for each land use type.

4000.000 3500.000 3000.000 2500.000 2000.000 1500.000 1000.000 500.000 0.000 0°-15°

15°-35°

35°-90°

WEST TN

836.227

1995.645

468.397

tot

1731.782

3698.823

778.903

EAST TN

895.555

1703.178

310.506

Figure 18 – Comparison of areas (km2) of each slope class between East and West Trentino.

92

Cavulli – Grimaldi – Pedrotti – Angelucci, Toward an understanding of archaeological visibility…

graphical basins, because the differences of valley bottoms, slopes and top areas are very similar between the regions of Trentino-Alto Adige, Veneto, and Friuli. Taking into account our earlier remarks about archaeological visibility and post-depositional processes, we are now better placed to investigate the exploitation strategies of the last hunter-gatherer groups of the southern Alps. A.P., F.C., S.G., D.E.A.

areas (fig. 3), free of a significant overburden of alluvial or glacial slope deposits (fig. 4). Elevation is a very important variable (fig. 2): while Palaeolithic sites (fig. 7) are located with one exception at middle elevations (1000-1600m asl), the Mesolithic ones are located at three different elevation ranges (fig. 12): low (between 95 and 594m asl), mid (1000-1238m) and high (1773-2000m). Currently, there are no finds attributed to these chronological phases in the ‘gaps’ between these ranges. Because of the wet, marshy conditions of the valley floor of the Adige River during the early Holocene (Fuganti et al. 1998), the shelters are located above the valley bottom on footslopes of talus cones (fig. 15). That they are currently found on vertical walls is due to the well-known and intensive exploitation of these talus on footslopes for gravel extraction from the 1970s onwards. More is known today about the middle elevation range because of important new research (La Cogola, Dalmeri et al. 2000; Dalmeri 2005; and Laghetto della Regola, Dalmeri et al. 2005). These Sauveterrian sites, excepting Grotta d’Ernesto, have a previous Epigravettian occupation (fig. 7). Finds from the middle elevations (both Mesolithic and Palaeolithic) are located in specific places such as rock shelters/ caves (Riparo La Cogola, Riparo Dalmeri and Grotta d’Ernesto), smooth morphologies such as flat open areas or plateaus (Passo S. Barbara, Madonna della Neve, Colle dei Colombi, Albi Val Coperte, Malga Scura, Malga Artillione, Colle Val d’Antenne, Colle dei Meneghini, Augustello), and close to small lakes (Viotte, Andalo, Terlago, Carbonare, Laghetto della Regola LR1-3, Pian dei Laghetti I–II, Palù Echen). The middle altitudes in this region are mostly unstable (almost 60% has an inclination between 15 and 35 degrees, and more than 10% is steeper than 35 degrees) and of low archaeological visibility. More than 50% of the SAG deposits are found at this elevation range, and dense coniferous forests hide the potential finds (more than 50% of the region and more than 70% of the middle elevations are forested). Although we must take the presence of 24 finds of the two periods in this elevation range seriously, they occupy only those morphologies that have a high survey visibility and are preserved by surface dynamics. All these considerations must be taken into account when promoting the exploration of the site gaps in western Trentino and at middle elevations. The success of this analysis means that we can extend the method to neighbouring regions, using an approach based on multivariate analysis and hydro-

References Alessio M., Allegri L., Bella F., Broglio A., Calderon G., Cortesi C., Improta S., Preite Martinez M., Petrone V. & Turi B., 1984. 14C datings of three mesolithic series of Trento Basin in the Adige Valley (Vatte di Zambana, Pradestel, Romagnano) and comparisons with Mesolithic series of other regions, Preistoria Alpina 19, pp. 245-254. Alessio M., Allegri L., Bella F., Improta S., Belluomini G., Calderoni G., Cortesi C., Mandra L. & Turi B. 1977. University of Rome carbon-14 dates, Radiocarbon 19, pp. 79-104. Angelucci D.E. & Bassetti M. 2009. Humans and their landscape from the Alpine Last Glacial Maximum to the Middle Holocene in Trentino: geoarchaeological considerations, Preistoria Alpina 44, pp. 59-78. Avigliano R., Di Anastasio G., Improta S., Peresani M. & Ravazzi C. 2000. A new late glacial - early Holocene palaeobotanical and archaeological record in the eastern PreAlps: the Palughetto basin (Cansiglio Plateau, Italy), Journal of Quaternary Science 15(8), pp. 789-803. Bagolini B. & Pasquali T. 1984. Le Mésolithique dans la Chaine du Lagorai, Preistoria Alpina 19, pp. 197-200. Bagolini B. 1980a. Il Trentino nella preistoria del mondo alpino. Temi, Trento. Bagolini B. 1980b. Riparo Gaban: preistoria ed evoluzione dell’ambiente. Edizioni didattiche del Museo Tridentino di Scienze Naturali, Trento. Bagolini B. 1982. Valle dell’Adige: modello interpretativo della colonizzazione postglaciale di un territorio alpino fino all’introduzione dell’agricoltura, in Atti del convegno “Uomo e agricoltura”, Seminario di Scienze Antropologiche, suppl. 1, Trento, pp. 53-58. Bagolini B., Broglio A. & Lunz R. 1984. Le mésolithique des Dolomites, Preistoria Alpina 19, pp. 15-36. Bagolini B. & Broglio A. 1985. Il ruolo delle alpi nei tempi preistorici (dal Paleolitico al Calcolitico). Studi di Paletnologia in onore di S.M. Puglisi. Roma, Università “La Sapienza”, pp. 671-676. Bagolini B. & Pedrotti A. 1992. Vorgeschichtliche Hohenfunde im Trentino-Südtirol und im Dolomitenraum vom Spätpaläolithikum bis zu den Anfangen der Metallurgie, in Hopfel F., Platzer W.& Spindler K. (eds), Der Mann im Eis. Band 1. Bericht über das Internationale Symposium. Innsbruck, pp. 359-377. Bassetti M., Dalmeri G., Kompatscher K., Kompatscher Hrozny M. & Lanzinger, M. 2002. Research on the Epigravettian site of Riparo Dalmeri on the Sette Comuni plateau (Trento), Preistoria Alpina 34, pp. 139-154. Biagi P. 1995. Mesolithic exploitation of the highland zone: a case study for the southern Alps, Preistoria Alpina 28 (1992), pp. 367-372. Bosellini A., Castellarin A., Dal Piaz G.V. & Nardin M. (eds), 1999. Carta litologica e dei lineamenti strutturali del

93

Hidden Landscapes

Trentino. Servizio geologico della Provincia autonoma di Trento, Trento, 1999. Broglio A. & Lanzinger M. 1990. Considerazioni sulla distribuzione dei siti tra la fine del Paleolitico superiore e l’inizio del Neolitico nell’Italia nord-orientale. Natura Bresciana 13, pp. 53-69. Broglio A. & Lanzinger M. 1996. The human population of the southern slopes of the eastern Alps in the Wurm Late Glacial and early Postglacial, Il Quaternario 9(2), pp. 499-508. Broglio A. 1972. I più antichi abitatori della Valle dell’Adige, Preistoria Alpina 8, pp. 157-176. Broglio A. 1973. La preistoria della valle padana dalla fine del Paleolitico agli inizi del Neolitico: cronologia, aspetti culturali e trasformazioni economiche, Rivista di Scienze Preistoriche 28, pp. 143-155. Broglio A. 1982. Culture e ambienti della fine del Paleolitico e del Mesolitico nell’Italia nord-orientale, Preistoria Alpina 16 (1980), pp. 7-29. Broglio A. 1994a. Il Mesolitico, in Il veneto nell’antichità, vol. 1: Preistoria e protostoria, Verona, pp. 281-311. Broglio A. 1994b. Man and environment in the Alpine region (Paleolithic and Mesolithic), Preistoria Alpina 26, pp. 61-69. Broglio A. 1995a. Mountain sites in the context of the northeast Italian Upper Paleolithic and Mesolithic, Preistoria Alpina 28 (1992), pp. 293-310. Broglio A. 1995b. Le Mésolithique des Dolomites, Preistoria Alpina 28 (1992), pp. 311-316. Broglio A., Castelletti L., Frigo G., Martello G., Maspero A. & Peresani M. 1995. Le site épigravettien de Val Lastari sur l’haut plateau d’Asiago (Préalps de la Vénétie), Preistoria Alpina 28 (1992), pp. 207-225. Broglio A. & Improta S. 1995. Nuovi dati di cronologia assoluta del Paleolitico superiore e del Mesolitico del Veneto, del Trentino e del Friuli. Atti dell’Istituto Veneto di Scienze, Lettere ed Arti 153, pp. 1-45. Casadoro G., Castiglioni G.B., Corona E., Massari F., Moretto M.G., Paganelli A., Terenziani F. & Toniello V. 1976. Un deposito tardowürmiano con tronchi subfossili alle Fornaci di Revine (Treviso). Bollettino del Comitato Glaciologico Italiano 24, pp. 22-63. Cavulli F., Angelucci, D.E. & Pedrotti A. 2002. La successione stratigrafica di Lugo di Grezzana (Verona), Preistoria Alpina 38, pp. 89-107. Cavulli F. & Grimaldi S. 2009. Raw material and settlement strategies at the Pleistocene/Holocene boundary in Trentino (northeastern Italian Alps): a GIS approach, in Proceedings of the 7th International Conference on The Mesolithic in Europe, Belfast, Irlanda del Nord (U.K.), August 29 – September 2, 2005. Cavulli F. & Grimaldi S. 2007. To see or not To see. Archaeological data and surface visibility as seen by an AIS (Archaeological Information System) approach, in A. Figueiredo & G. Velho (eds), The World is in your eyes, Proceedings of the XXXIII Computer Application and Quantitative Methods in Archaeology Conference (March 2005 – Tomar, Portugal), 2005, pp. 413-420. Colecchia A., Casagrande L., Cavulli F., Mura L., Nebbia M. 2011. Paesaggi medievali del Trentino (progetto APSAT). Post Classical Archaeologies, 1/2011, pp. 245-274. Dalmeri G. (ed.), 2005. Studi sul Riparo Cogola (Carbonare di Folgaria – Trento). Frequentazione umana e paleoambiente, Preistoria Alpina 40 (2004), pp. 89-200.

Dalmeri G. & Lanzinger M. 2001. The evolution of the environment and human population of the Adige basin at the end of the late Ice Age and in the early Holocene, Preistoria Alpina 34, pp. 15-18. Dalmeri G. & Pedrotti A. 1995. Distribuzione topografica dei siti del Paleolitico superiore finale e Mesolitico in Trentino, Preistoria Alpina 28 (1992), pp. 247-267. Dalmeri G., Bassetti M., Cusinato A., Degasperi N., Kompatscher K. & Kompatscher Hrozny M. 2000. “La Cogola”, nuovo sito in riparo sottoroccia presso Carbonare di Folgaria (Trento), Preistoria Alpina 31 (1995), pp. 53-59. Dalmeri G., Bassetti M., Cusinato A., Kompatscher K. & Kompatscher Hrozny M. 2006. La frequentazione umana a Palù Echen (Folgaria, Trento) alla fine del Tardiglaciale. Nota sul saggio di scavo nel settore 1, Preistoria Alpina 41 (2005), pp. 147-151. Dalmeri G., Kompatscher K., Hrozny Kompatscher M., Bassetti M., Cusinato A. & Piazzi O. 2005. Dinamiche comportamentali degli ultimi cacciatori raccoglitori in area alpina. Il caso di studio del sito L3 del Laghetto delle Regole (Castelfondo, Trento), Preistoria Alpina 40 (2004), pp. 5-26. Dalmeri G., Grimaldi S. & Lanzinger M. 2001. Il Paleolitico e il Mesolitico, in Lanzinger M., Marzatico F. & Pedrotti A. (eds), La preistoria e protostoria – Storia del Trentino, volume I, pp. 15-118. Trento, Istituto Trentino di Cultura. Felber M., Veronese L., Cocco S., Frei W., Nardin M., Oppizzi P., Santuliana E. & Violanti D. 2000. Indagini sismiche geognostiche nelle valli del Trentino meridionale (Val d’Adige, Valsugana, Valle del Sarca, Valle del Chiese). Studi Trentini Scienze Naturali, Acta Geologica 75, pp. 3-52. Fiore I., Tagliacozzo A. & Cassoli P.F. 2002. Ibex exploitation at Dalmeri rockshelter and “specialized hunting” in the sites of the eastern Alps during the Tardiglacial and the early Holocene, Preistoria Alpina 34, pp. 173-184. Fuganti A., Bazzoli G. & Morteani G. 1998. The Quaternary evolution of the Adige Valley near the city of Trento (Northern Italy) as deduced from wells and radiocarbon dating. Preliminary results. Studi Trentini Scienze Naturali, Acta Geologica 73, pp. 93-97. Lanzinger M. 1987. Il popolamento del versante meridionale delle Alpi orientali fra Paleolitico superiore e Mesolitico. Tesi di Dottorato di Ricerca in Paleoantropologia, Firenze. Lanzinger M. 1991. Popolamento e strategie di caccia nella preistoria delle Dolomiti Ladine. Mondo Ladino 3-4, pp. 273-307. Leonardi P. 1963. Il Paleolitico nel versante meridionale delle Alpi. Rendiconti della Società di Cultura Preistorica Tridentina 1, pp. 62-85. Pellegrini G.B, Albanese D., Bertoldi R. & Surian N. 2005. La deglaciazione nel Vallone Bellunese, Alpi Meridionali Orientali. Supplementi Geografia Fisica e Dinamica Quaternaria 7, pp. 271-280. Peresani M. 2003. 12000 anni fa al Bus de la Lum. Società Naturalisti Silvia Zenari, Pordenone.

Acknowledgements The research was initially financially supported by the Fondo Progetti di Ricerca Post Doc, Provincia Autonoma di Trento (Project STRIM 2003-2005). Funding was then continued with the Alpinet European project (Culture 2000) and APSAT ‘Ambiente e paesaggi d’altura trentini’, Provincia Autonoma di Trento – bando ‘Grandi progetti 2006’ delibera G.P. 2790-2006 (Colecchia et al. 2011).

94

11 Buried landscapes and cultural landscapes of the mountainous areas of Lombardy in pre- and protohistory Raffaella Poggiani Keller Soprintendenza per i Beni Archeologici della Lombardia, via E. De Amicis 11, I-20123 Milano – [email protected] Abstract Targeted field research directed since 1985 by the Soprintendenza per i Beni Archeologici in trial areas, systematic mapping, valley by valley, of finds and sites together with emergency excavations, have contributed in recent years to an understanding of settlement strategies and approaches to the exploration and resource exploitation in the mountainous part of Lombardy from the late Neolithic until the arrival of the Romans. No clean break is evident between protohistoric and historical times. Discoveries include permanent and seasonal settlements, craft workshops, funerary areas, and cult and ceremonial sites. Of particular note are megalithic sanctuaries founded in the 4th and 3rd millennium BC which served as visible and long-lasting territorial markers along routes used for transhumance and mineral extraction. The examples given below demonstrate how the identification and comprehension of these prehistoric and protohistoric activities is inseparable from an overall appreciation of mountain landscapes through time: cultural landscapes in which ancient signs and survivals of extraordinary immediacy persist next to buried remains.

1. Introduction

This has thrown much light on settlement distribution and strategies regarding exploration and resource exploitation in the Lombardy mountains between the Late Neolithic (the period in which the area became densely populated) and the time of Romanization. It has served to identify the various kinds of buried landscapes which are described below, using case studies that are emblematic with regard both to territorial context and epoch.

The Lombardy Alps and Pre-Alps include the provinces of Sondrio, Bergamo and Brescia, a vast area of contact and interchange between the Po Valley and the heart of the central Alps, home to the river systems of the Inn and the Danube. This part of Lombardy possesses several distinctive features: the valley heads are generally closed off by the high Alps, and are therefore not suitable as important transAlpine communication routes (with the exception of the Ticino Valley to the east, on the border with Piedmont); raw materials — flint, copper and iron — are abundant; and it is renowned for the presence of important and extensive examples of rock art (in Valle Camonica and Valtellina). From 1985 onwards, the Soprintendenza per i Beni Archeologici della Lombardia has conducted in the Alpine and Pre-Alpine areas systematic exploration in the field, mapping sites in each valley (fig. 1); these were published (1989-1992) in a series of studies of the territory accompanied by archaeological maps (Poggiani Keller 1989 and 1992; Rossi 1991; Fortunati & Poggiani Keller 2007). More recently, survey work1 and targeted studies were undertaken, including the creation of a GIS record of the (over 180) Valle Camonica rock-art sites and of a Digital Elevation Model of the Chiese Valley, elaborated in collaboration with the CNR-IREA.2 Other work carried out includes targeted field exploration in selected areas or with regard to certain contexts (high altitude prehistoric bivouacs, sites with engravings, sites involving the use of minerals, Copper Age sites marked by steles); extensive archaeological work in advance of large public construction projects;3 and other rescue and research excavations.

2. Settlement strategies Settlement strategies in Alpine and Pre-Alpine Lombardy privileged the foundation of villages on high ground equipped with natural defences, or on slope terraces that gave control over widespread territories, or on transit routes along or between valleys from which paths led off to the high pasture, or that were used in the search for mineral resources. This is well illustrated by the example of Tirano in Val-

Figure 1 – Location map. In grey the territory of the Sondrio, Bergamo and Brescia provinces.

95

Hidden Landscapes

a

b

c

d

Figure 2 – a. Sites around Tirano in Valtellina, controlling the trade route through the Poschiavina valley and over the Bernina pass. In the area rocks with schematic engravings and a 2nd or 1st millennium BC metal hoard have been found. b. Three settlements in dominant positions near Grosio. At the start of the Middle Bronze Age the territory was organised in a new way, which lasted until the Roman period. c. Parra (Oromobiorum oppidum, 1st millennium BC), located on a fluvial terrace overlooking the confluence of the Serio and Nossana rivers. d. Valle Camonica, Cemmo: settlement and sanctuary (3rd millennium BC).

tellina (fig. 2a), at the end of the Poschiavo Valley which connects the Adda Valley with Graubünden (Switzerland) by means of the Bernina Pass. The settlement at San Perpetua, which controlled the road that ran over this pass — already in use during the Early Bronze Age — lasted until the end of the Iron Age and long coexisted with a second settlement located on a hillock at the centre of the fertile valley of the River Adda. The Early Bronze Age date is provided by a hoard containing two Alpine-type daggers found at Crotto di Piattamala (De Marinis 1989). The second settlement on Dosso di Tirano is only known from ex situ finds on the hill slopes, since it was on the site of the present village, occupied since the medieval period (Poggiani Keller 1999b). Further confirmation is seen in the arrangement found from the onset of the Middle Bronze Age until the arrival of the Romans in another part of the same valley near Grosio, where three settlements in dominant positions were present (Dosso dei Castelli, Dosso Giroldo and Dossa (fig. 2b). The fact that

in the Middle Iron Age all were in use suggests the existence of a hierarchy and of economic and defensive specialization (Poggiani Keller 1999a, p. 112). In the valleys near Bergamo too, a pattern of settlements controlling the river and mineral resources can be discerned. The system is exemplified by Parra, the oppidum of the Orobic Alps (Pliny, Nat. Hist. III 17, 124-125), located at an altitude of 580 m on a fluvial terrace dominating the confluence of the Serio and Nossana rivers (fig. 2c), in an area rich in minerals (calamine and ores of silver, lead and iron) that were already exploited in the Late Bronze Age (Poggiani Keller 2006 and 2007a). In Valle Camonica, with respect to other valleys, settlements appear to have been more stable and long-lasting, surviving (on the basis of current knowledge, see Poggiani Keller 2010) from the time of their foundation (the Late Neolithic characterized by advanced VBQ (square-mouthed pottery) and Lagozza cultural features), up until the time of Romanization. This is found not just in the case of high-altitude set96

Poggiani Keller, Buried landscapes and cultural landscapes of the mountainous areas…

tlements such as Luine, Breno Castle, Dos dell’Arca and Seradina, but also in valley-bottom settlements which controlled river traffic, such as Cividate Camuno-Malegno, which maintained its function during the Roman period. In this valley, furthermore, the area studied contained a network of permanent and seasonal settlements, with areas of craft activity, funerary and cultic use. Figure 2d shows the relationship between the Seradina settlement (4th-1st millennium BC), located on a hill where the valley narrows and in a symmetrical position with respect to the contemporary Doss dell’Arca settlement on the opposite side, and the sanctuary of Cemmo which was founded during the Copper Age over Mesolithic and Late Neolithic remains.

tions were conducted. Of the 29 localities identified, 18 were of ancient foundation. It may be seen that the distribution of the sanctuaries is clearly concentrated in two areas, both at an altitude of 600-800 metres: the group of hills near Teglio in the Adda Valley and the Ossimo-Borno upland in the Oglio Valley. Rarely was a valley-floor location chosen; two examples are the Corni Freschi site at Darfo Boario Terme, and the sanctuary at Cemmo. The latter was established in the 3rd millennium BC in a small glacial valley at the foot of a high rock face, and centred on an ephemeral water pool; it was marked by two fallen rock masses, engraved, and a N-S line of steles the principal faces of which looked eastwards. In the Late Bronze Age this sanctuary, after a period of disuse, was given a more monumental aspect with the construction of a semicircular wall, enclosing the sacred area around the two rock falls and incorporating (and thus hiding from view) several of lines of steles. The site was subjected to a further series of modifications and lasted until late antiquity, when it was destroyed by Christians; a later substitution being the nearby Romanesque church of S. Siro. In Valtellina, where the settlement sites remain unknown, the megalithic sanctuaries are all located in the hills near Teglio, which overlook the junction

3. Ritual landscapes and sanctuaries At the cusp of the 4th and 3rd millennia BC, a few centuries after the great increase in settlement in mountainous areas that was associated with the foundation of many long-lasting settlements on the valley floors, there grew up a system of cult and ceremonial megalithic sites which functioned as long-term territorial markers along routes used for transhumance and mineral prospecting. In five of the latter (Caven di Teglio in Valtellina; Passagròp, Pat, Anvòia di Ossimo, and Cemmo di Capo di Ponte in Valle Camonica4) area excavations have been conducted. These have revealed the sites’ organization, based on lines of engraved standing stones associated with other structures, such as platforms and votive offering enclosures (Caven; Anvòia and Pat) and/or walls delimiting the sacred areas (Caven and Cemmo). The presence of metal slag (Ossimo-Passagròp and Pat) and the creation of extensive areas of meadow by controlled burning prior to the establishment of the sanctuaries (fig. 3a) suggest that mineral prospecting, grazing and stock-rearing first generated interest in these areas. The occurrence of sanctuaries in the Lombard valleys is limited to just the two interconnected valleys of Valle Camonica and Valtellina (Poggiani Keller 2006, fig. 1), at both of which the Lombardy Cultural Heritage Superintendency has since 1985 conducted systematic explorations. This has included mapping of sites with steles and engraved rocks and an examination of the characteristics — morphological, geographical, interrelational and functional — of each locality, in an attempt to identify the parameters according to which these sites were selected and to understand the underlying reasoning. At the same time, whenever new discoveries were made, area excava-

a

b Figure 3 – a. The Ossimo-Borno plateau: a ritual landscape marked by four sanctuaries of the 4th and 3rd millennium BC. b. Valtellina, the complex of Teglio: locations of megalithic cult sites on terraces overlooking the valley and near historical roads.

97

Hidden Landscapes

between the Adda and Oglio valleys. All the sites are found between altitudes of 600 and 700 metres near hamlets around Teglio, on terraces overlooking the valley and on ancient trackways leading up from the valley floor (fig. 3b). The sanctuaries were purposely designed to be visible form many directions, along visual corridors defined by the local morphology, as may be clearly seen at Caven and Somasassa di Teglio (fig. 3a). In Valle Camonica these visual corridors were created ad hoc through the deliberate deforestation of portions of woodland, as was demonstrated for the plateau at Ossimo. Here, a singular ritual landscape was discovered at an altitude of 800 m, made up of four closely-spaced cult sites, each of which could be seen from the others. The forest clearings in which these are located were created by controlled burning when the sanctuaries were established, an operation of long-term ‘landscape construction’ which is still visible today (fig. 2d). The archaeological identifica-

tion of these ‘buried’ Chalcolithic landscapes was due — in addition to the morphology of the local terrain and indirect evidence including place-names, signs of the Christianization of certain localities, and ancient roadways — to direct evidence such as the discovery of steles and of menhirs eroding out onto hill slopes or still in place (e.g. the Pat 2 menhir, still visible above ground level). The monuments’ relationship with the most suggestive components of the surrounding environment, especially its mountains and water, appears to have been of decisive importance. An example is provided by the cult and ceremonial site at Pat di Ossimo, founded about the mid-4th millennium BC on a large terrace overlooking the Inferno Valley. This was visible from an imposing mountain called Cimon della Bagozza, towards which pointed a line of engraved standing stones set up in the late 4th or early 3rd millennium BC (fig. 4); subsequently the site remained the object of repeated ritual activities, such as the lighting of fires and the deposition of artefacts throughout the Iron Age.

4. The buried landscape Indirect indications of the existence of buried landscapes come from features of the present-day cultural landscape. During archaeological work in advance of large-scale public construction projects such as the 380 KV electricity line between Valle Camonica and Switzerland, which involved the construction of 119 pylons in locations about which nothing was known, several sites were found thanks to morphological features and indirect evidence. For example, the presence of an early 3rd millennium BC settlement (perhaps a cult site) on Dosso Poglia, Cedegolo (Brescia), which consists of a small hillock artificially covered with rocks and enclosed by a wall and ditch, was initially suggested by the presence of numerous Christianizing crosses engraved on the surrounding rocks, religious shrines and a cross-roads of historic routes: all clearly signs of a long-standing memory of the place and its function. At times it may be observed in mountain areas that there exists a sharp break between protohistoric and historical epochs; in order to understand the hidden prehistoric and protohistoric landscapes one must be able to ‘read’ the current cultural landscape diachronically, since it consists of a palimpsest of events spanning numerous centuries. This is still more evident at high altitude, where the mountain — more than other types of territory — conserves traces of a past unwinding throughout

a

b Figure 4 – a. Ossimo pat: a general view of the site. The stele alignments are located on a large terrace overlooking the Inferno valley. b. A view of the alignment of engraved stones at Pat di Ossimo. The inset shows a detail of the engravings of stele Pat.

98

Poggiani Keller, Buried landscapes and cultural landscapes of the mountainous areas…

the Holocene up to the present, with strong traditions derived from the activities that have long been carried out there, and indeed still are (fig. 5). Recent European studies (e.g. Palet Martinez 2006) have found that the sites of Mesolithic bivouacs were often reused in cyclical fashion, mainly for pastoral activities, in the Late Neolithic, the Bronze Age, the Roman period, and beyond. Research under way in the Lombard Alps shows that the mountains were re-occupied in the Late/Recent Neolithic, and that subsequently, in the Middle Bronze Age, groups of high-altitude seasonally-occupied wooden dwellings were built, such as that found in Valle Camonica at Cevo-Dos Curù, at 2000 m above sea level (archaeological investigations in the mountain pastures of the Valle Brembana, 2003-present: Poggiani Keller 2007b). The remains resemble, with regard to construction style and date, those of the settlement of Storo-Dosso Rotondo in Trentino (Marzatico 2005; Bassetti et al. 2008). The continued reuse of sites may at times be reflected in continuity in the use of ancient structures, in some cases exceptionally well preserved, so much so that ancient stone-built walls at first sight seem to be of recent construction. An example is the 6th-5th century BC high-altitude mining village at CevoDos del Curù mentioned above. The chance find, in 2000, of an inscribed text in ‘Camuno’ alphabet led to further investigation, and the excavation begun by the Soprintendenza in 2004 (and still in course) revealed the presence — in an area of mines dug in the 18th and 19th centuries AD — of an extensive Iron Age mining village with rectangular houses measuring 5 by 12 metres (fig. 6). This village grew up on top of older remains (a mountain hut from the onset of the Middle Bronze Age) and was associated with numerous stones with Camuno inscriptions, at times accompanied by figured motifs. We have yet to understand their significance. On the stone-slabbed floor of one of the houses (House D), together with anvils, mortars and crucibles for metallurgical activities, a bronze S. Giacomo-type fibula (for the Von Eles typology, see Masi 1986) and a Ciaslir di Monte Ozol-type jug were found. They date to the late 6th – first half of the 5th century BC, a period distinguished in this area by the presence of Rhaetic cultural features (Breno-Dos de l’Arca facies, part of the more widespread Fritzens-Sanzeno group). The protohistoric remains, together with examples of schematic art (cup-marks carved into rock and stones), are surrounded by clear evidence of the historically recent mining activity conducted on the site.

Figure 5 – Alpe Azzaredo (Mezzoldo-Bergamo): stockyard and pre-historical camp (marked by the arrow, on the left).

Figure 6 – View of the excavation at the high-altitude mining village of Cevo-Dos del Curù: house F (Middle Iron Age).

5. Conclusion The examples outlined above serve to emphasize the conclusion that the identification and comprehension of prehistoric and protohistoric sites cannot be separated from an overall view and understanding of the mountain landscape through time, a cultural landscape in which ancient signs and survivals of extraordinary immediacy persist alongside buried remains. In particular, there exists a dense and well-preserved network of historic paths of ancient origin, which still today connect numerous archaeological and rock-art sites. These constitute the fabric within which developed the structured settlements, the exchange, trading and other activities, and they deserve the archaeologists’ careful attention.

References Bassetti M., Dalmeri G., Mottes E. & Nicolis F. 2008. La frequentazione delle alte quote nell’età del Bronzo. Il sito di StoroDosso Rotondo, Atti 1° Convegno interregionale Archeologia lungo il Chiese, Storo 24-25 Ottobre 2003, pp. 107-127.

99

Hidden Landscapes

Boschetti M., Gallo I., Brivio P.A. & Binaghi E. 2002. Metodologie per l’analisi delle relazioni tra caratteristiche ambientali e distribuzione di siti archeologici: mappe di plausibilità di sviluppo per l’epoca romana, Proceedings VI Conferenza Nazionale ASITA Geomatica per l’ambiente, il territorio e il patrimonio culturale, Perugia 5-8 November 2002, pp. 511-516. De Marinis R. 1989. Tirano, Piattamala-località al Crotto, in Poggiani Keller R. (ed.), Valtellina e mondo alpino nella preistoria, Catalogo mostra, Milano 1989, Modena, pp. 52-53. Fedele F. 2006. Asinino-Anvòja. Il Parco Archeologico, Cerveno. Fedele F. 2007. Monoliths and human skeletal remains: ritual manipulation at the Anvòia cerimonial site, Ossimo (Val Camonica, Italy), Atti Congresso internazionale Le pietre degli dei. Statue-stele dell’Età del Rame in Europa. Lo stato della ricerca, Brescia 16-18 September 2004, Notizie Archeologiche Bergomensi 12-2004, pp. 49-66. Fortunati M. & Poggiani Keller R. (eds) 2007. Storia economica e sociale di Bergamo. I primi millenni. Dalla preistoria al Medioevo, Bergamo, Vol. I. Marzatico F. 2005. Lo sfruttamento dei pascoli montani nella pre-protostoria, Atti del Convegno Ars Venandi. Malghe. Tra gestione ambientale integrata e cultura del territorio, S. Lorenzo in Banale 27 August 2004, pp. 29-37. Masi P. 1986. Le fibule dell’Italia settentrionale, Prähistorische Bronzefunde 5, München. Palet Martinez J.M. 2006. Stratègies de la recherche archéologique en haute montagne: l’expérience des projets Champsaur (Alpes) et Madriu-Cadì (Pyrénées), in Alpis Graia. Archéologie sans frontières au col du Petit-SaintBernard, Aosta, pp. 381-385. Poggiani Keller R. (ed.) 1989. Valtellina e mondo alpino nella preistoria, Catalogo mostra, Milano 1989, Modena. Poggiani Keller R. (ed.) 1992. Carta archeologica della Lombardia. II La Provincia di Bergamo, voll. I-III, Modena. Poggiani Keller R. 1995. Grosio (SO), Dosso dei Castelli e Dosso Giroldo. Un insediamento protostorico sotto i castelli e altri resti dell’età del Bronzo e del Ferro, Quaderni del Parco delle Incisioni Rupestri di Grosio 2, Sondrio. Poggiani Keller R. 1999a. Dinamica dell’insediamento nell’età del Ferro in Valtellina (Italia), in Della Casa Ph. (ed.), Prehistoric alpine environment, society and economy, Papers of the international colloquium PAESE 97, Zurich 1997, Universitätsforschungen zur prähistorischen Archäologie 55, Bonn, pp. 109-118. Poggiani Keller R. 1999b. Ricerche e scavi in Valtellina, Atti II Convegno Archeologico Provinciale, Grosio 20 e 21 ottobre 1995, Quaderni del Parco delle Incisioni Rupestri di Grosio 3, Sondrio, pp. 41-62. Poggiani Keller R. 2004. Siti di culto megalitici e occupazione del territorio nell’età del Rame in Valtellina e Valcamonica, in Actes Xe Colloque international sur les Alpes dans l’Antiquité. Implantations rurales et économie agro-pastorale dans les Alpes de la Préhistoire au Moyen Age, Cogne-Vallée d’Aoste 12-14 September 2003, Aosta, pp. 143-160. Poggiani Keller R. (ed.) 2006. L’oppidum degli Orobi a Parre (BG), Milano. Poggiani Keller R. 2006. Santuari megalitici nelle valli lombarde, in Preistoria dell’Italia settentrionale, Atti del Convegno in ricordo di Bernardino Bagolini, Udine settembre 2005, Udine, pp. 245-268.

Poggiani Keller R. 2007a. L’età del Ferro. Dall’oppidum degli Orobi alla formazione della città sul colle, in Fortunati M. & Poggiani Keller R. (eds), Storia economica e sociale di Bergamo. I primi millenni. Dalla preistoria al Medioevo, Bergamo, Vol. I, pp. 146-189. Poggiani Keller R. 2007b. Le “piccole regioni” alpine: testimonianze senza tempo di un paesaggio culturale da indagare, in Fortunati M. & Poggiani Keller R. (eds), Storia economica e sociale di Bergamo. I primi millenni. Dalla preistoria al Medioevo, Bergamo, Vol. I, pp. 218-227. Poggiani Keller R. 2010. Aspetti dell’insediamento e abitati d’altura nell’età del Bronzo e del Ferro in Lombardia (con Appendici di M. Baioni e A. Massari), in Dal Ri L., Camper P. & Steiner H. (eds), Höhensiedlungen der Bronze und Eisenzeit. Kontrolle der Verbindungswege über die Alpen/ Abitati dell’età del Bronzo e del Ferro. Controllo delle vie di comunicazione attraverso le Alpi [Atti Convegno di studi Ganglegg. Die befestigte Siedlung am Ganglegg, Schluderns 22-25 nov. 2000], Forschungen zur Denkmalpflege in Südtirol/Beni Culturali in Alto Adige-Studi e Ricerche VI, Bolzano, pp. 164-231. Rossi F. (ed.) 1991. Carta archeologica della Lombardia. I La Provincia di Brescia, Vols. I-II, Modena.

Notes 1

In Valtellina-SO (1995-1996), Valtrompia-BS (1999) and the Chiese Valley (1999-2000). This work is still unpublished, except for brief descriptions in the Notiziario della Soprintendenza Archeologica (NSAL 1999-2000) and general accounts (Poggiani Keller 2010, with appendices by M. Baioni and A. Massari on the Chiese Valley and Valtrompia). 2 The project Production of a GIS model of the middle Chiese Valley for the purpose of archaeological studies. Creation of a DEM and archaeological database for the municipal areas of Gavardo and Villanuova sul Clisi, concluded in 2002, was conducted by Mirco Boschetti and P. Alessandro Brivio of the CNR-Reparto di Telerilevamento IREA of Milan and, with regard to the archaeological part, by Marco Baioni of the Museo Archeologico della Valle Sabbia in Gavardo and the present author. With regard to the methods employed, previously applied to the Lake Garda morainic hills for the Roman and Early Medieval periods, see Boschetti et al. 2002. 3 To cite just a few (reports in NSAL, see note 1): the SS 45 bis Gardesana road in the Chiese Valley; the trunk road SS 42, del Tonale e della Mendola in Valle Camonica; the gas pipeline between Valtrompia and Valle Camonica that passes at high altitude along mountain crests; and the 380 KV electricity cable between Lombardy and the Swiss border that passes through Valle Camonica and Valtellina. 4 The first investigation of Passagròp di Ossimo and the excavation of Anvòia di Ossimo were conducted by F. Fedele of Università degli Studi ‘Federico II’ in Napoli (Fedele 2006 and 2007), while later work at Passagròp and the area excavations at Caven di Teglio, Cemmo di Capo di Ponte and Pat di Ossimo, still in course, were conducted by the present writer for the Soprintendenza (Poggiani Keller 2006).

100

12 The Southern French Alps Landscape Project: an archaeological and palynological study of high-altitude settlement areas in the Southern French Alps Kevin WalshI, Florence MocciII, Suzi RicherI I Department of Archaeology, The King’s Manor, University of York, York YO1 7EP, UK – [email protected]; [email protected] II Centre Camille Jullian, UMR 6573 CNRS, Université de Provence, M.M.SH., 5 rue du Château de l’Horloge, BP 647, 13094 Aix-en-Provence, cedex 2, France – [email protected] Abstract Until the start of our work, the high altitude (sub-alpine and alpine) zones of the southern French Alps were largely ignored by archaeologists. Whilst a good amount of research into the prehistory of the northern French Alps, and parts of the Italian Alps, has already taken place, our understanding of settlement in the high altitude zones of the southern French Alps is quite limited. Previous research in our study area had concentrated on valley bottoms and lower slopes at altitudes between 1000 and 1500 m. Our research comprises archaeological survey and excavation combined with the use of palaeoecological evidence, especially the interpretation of pollen indicator species and anthracological data from the excavated archaeological sites. Whilst the high altitude zones (2000 m and above) in both areas have yielded few sites for these periods, the Ubaye Valley is famous for the large number of Iron Age burials and metal-work finds. At lower altitudes in the Ubaye Valley, small Roman settlement sites (villas and farms) have been found, whilst such evidence is rare in the Ecrins.

Drac

VALENCE

Drôm

BRIANCON

Die

e

Ecrins GAP

Ubaye BARCELONNETTE

Nyons

DIGNE-LES-BAINS Verdon

The high peaks and plateaus of the southern French Alps are topographically potent and ever present. Despite this, they remained ‘hidden’ in terms of research until the start of the Southern French Alps Archaeological Project in 1998, which has aimed to investigate the history and archaeology of human activity in the Ecrins and Ubaye areas of the French Alps. Thus far, our research covers all periods from the Mesolithic through to the post-medieval period. It is rare for such a chronological range of archaeological sites to be studied at this altitude (between 2000 and 2400 m above sea level). Until the start of our work, the high altitude (sub-alpine and alpine) zones of the southern French Alps were largely ignored by archaeologists. Whilst a fair amount of research into the prehistory of the northwestern French Alps, and parts of the Italian Alps, has already taken place, our understanding of settlement in the high altitude zones of the southern French Alps is quite limited. Previous research in our study area concentrated on lower altitudes, i.e. towards the valley bottoms and the lower slopes between 1000 and 1500 m. In the Ecrins, we have concentrated on research in the zones above 2000 m; in the Ubaye Valley, we have tried to study all zones from the valley bottoms up to the alpine zone above 2200 m. Whilst the Ubaye Valley is characterised as an open, easily accessible valley system, the Ecrins in comparison are characterised by relatively narrow valleys, steep mountainsides and high plateaux.

The aim of this project is to explain the various phases of settlement expansion and contraction as well as associated changes to the alpine environment. Moreover, we wish to examine the range of socioeconomic, cultural and ideological processes that may have affected the use of high-altitude areas in specific periods, both in terms of small-scale local changes, and wider regional-scale events. In addition, we attempt to examine the physical and perceptual relationships between people and their environments. Our research comprises archaeological survey and excavation combined with the use of palaeoecological evidence, especially the interpretation of pollen indi-

Rhône

1. Introduction

Carpentras Forcalquier

Var

AVIGNON Castellane

Apt NICE

Dur

anc

Arles

Etang de Berre

Grasse

e

Draguignan

Aix-en-Provence

Istres Brignoles

MARSEILLE TOULON

MER MEDITERRANEE

Figure 1 – Location of the study areas.

101

Iles d'Hyères

Hidden Landscapes

Figure 2 – An example of a ‘Bronze Age’ enclosure on the Faravel Plateau at 2200 m, in the Ecrins.

V. Dumas/CNRS/CCJ/2004

Prehistoric

Indet.

30%

Alpine 2100-3000 m

Sub alpine 1600-2100 m

Sub-montane (forest) zone 1000-1600 m

Gallo-Roman

Protohistoric

Medieval / Modern

Protohistoric

Gallo-Roman

Medieval / Modern

10%

Prehistoric

Protohistoric

Indet.

Prehistoric

Medieval / Modern

20%

Colline and planar 0-1000 m

Figure 3 – Chronological and vertical distributions of sites in the Ubaye Valley.

cator species and anthracological data from the excavated archaeological sites (Walsh & Richer 2006). We hope that through the combination of our extensive archaeological data set and the palaeoecological record, a fuller diachronic picture of ‘human-environment’ relationships in the high-altitude areas will emerge.

2. Results The surface inspection and excavation of archaeological sites within both study areas has been ongoing since 1998. The project has thus far located and

recorded many sites (c. 250 in the Ecrins and c. 120 in the Ubaye), ranging from Mesolithic flint scatters to the remains of medieval stone-built structures (Walsh et al. 2007; Walsh 2005; Garcia et al. 2007). In both areas, the Mesolithic period is represented by flint scatters which indicate that people were coming up to these high altitudes to hunt during the summer months. This type of activity continued into the Neolithic period. Towards the end of this period, and into the Bronze Age (the late 3rd and 2nd millennia BC), we see the construction of the first stone structures at this altitude, which are usually identified as pastoral enclosures. Once we come to the Iron Age, the nature of the evidence in the Ubaye and the Ecrins diverges. Whilst the high altitude zones (2000 m and above) in both areas have yielded few sites for this period, the Ubaye Valley is famous for the large number of Iron Age burials and metal-work finds at lower altitudes. One of the most important sites that we have excavated is the Sagnes burnt mound, which comprises evidence for feasting and the burial of people. No such sites are known in the Ecrins. At lower altitudes in the Ubaye Valley, small Roman settlement sites (villas and farms) have been found, and these are rare in the Ecrins. It seems that Iron Age and Roman settlement was concentrated at the lower altitudes. It is wellknown that the Romans were wary of the high-altitude zones that were difficult to control. During the Medieval period, especially from the 13th century onwards, there was an enormous increase in activity across the Alps. In the Ecrins, this is represented by a large number of structures at all altitudes, including the alpine zone (above 2200 m). It seems that entire families lived and worked in the high altitude zone during summer, and were involved in both pastoral and silver-mining activities. Despite climatic deterioration during the socalled Little Ice Age (16th–19th centuries AD) there was a lot of activity in the alpine zone. One of the main observations from our high altitude study in the Ecrins is therefore that climate does not seem to have a direct influence on the ebb and flow of settlement. Though the landscapes in our study area are relatively open and tree-less today, the pollen and charcoal evidence suggests that in the past these areas would have contained coniferous trees and birch. Palynological and non-pollen palynomorph evidence (NPP, mostly spores from dung fungi) has been used to illustrate that grazing by sheep

102

Walsh – Mocci – Richer, The Southern French Alps Landscape Project…

Faravel XIII Mesolithic Late Mesolithic

Faravel XVIII Faravel XII

middle/late Neolithic.

Faravel XXII

middle/late Neolithic.

Faravel XV

middle/late Neolithic.

Faravel XVII

middle/late Neolithic.

ChichinIV ChichinVI

middle/late Neolithic. middle/late Neolithic.

Chichin III (US 207) 3970±35BP Chichin III 3895±35BP Chichin IIIa (US 104-C4) 3845±35BP Faravel XIX 3670±45BP Faravel VIIId 3665±40BP Lac des LauzonsII 3470±100BP Jujal I 3275±40BP ChichinII (US 109) 3220±35BP Lac de Lauzons IIa 3180±60BP Jujal Ia 2945±35BP Col du Palastre 2770±95BP Faravel XIIIb 2460±50BP Faravel XIV 1985±50BP Col du Palastre 1915±80BP Vallon de la Vallette 1900±30BP Cheval de Bois I 1550±40BP Vallon de la Vallette 1210±60BP Fangeas VIIa 1180±80BP Clot Lamiande II 1145±50BP Cabane de la Barre 1105±35BP Chapeau Roux 995±40BP Jas du Cros 955±70BP Cheval de Bois II 895±35BP Faravel VIIIb 857±40BP ChichinII (US 108) 685±35BP Faravel XIIa 680±30BP Jas des Provencaux 660±50BP Faravel XIIb 650±40BP Fangeas VIIb Fangeas VI FangeasVIIb 500±40BP Col du Palstre 240±30BP Chapeau Roux 200±30BP 10000CalBC

8000CalBC 6000CalBC 4000CalBC 2000CalBC CalBC/CalAD

2000CalAD

14

C date

date from ceramic material

Single structures with more than one occupation phase

date from lithic typology periods of climatic deterioration

Grey text = Champsaur Black Italics = Freissinères

Figure 4 – Plot of securely dated archaeological sites against phases of climatic deterioration. Atmospheric data from Stuiver et al. (1998); OxCal v3.9 Bronk Ramsey (2003); cub r:4 sd:12 prob usp[chron].

103

Hidden Landscapes

and goats began in the high altitudes in the Bronze Age, with peaks in grazing activity occurring in the Iron Ages and the medieval/modern periods. The large-scale increase in medieval and modern pastoral activity in the Ubaye valley is associated with the change from local pastoralism to largescale transhumance, with animals brought from the plains of Provence to the mountains for summer pasturing.

work by J.-L. de Beaulieu (Institut Mediterraneen d’Ecologie et de Paleoecologie). The Parc Nationale des Ecrins and the Service Régionale pour l’Archéologie (PACA) must also be thanked for their continued support of this project. The research in the Ubaye is co-directed with Dominique Garcia of the Centre Camille Jullian, CNRS/Université de Provence, Aix-en-Provence.

3. Conclusions Despite the remote and ‘hidden’ nature of this landscape, the relative lack of disturbance to these upland zones means that we have inherited an almost pristine archaeological record. It is clear that the history of activity in the high altitude zones is not one of constant increase over time, but one of movement to and from these areas influenced by complex local and regional socio-economic and cultural processes.

References Garcia D., Mocci F. & Walsh K. 2007. Archéologie de la vallée de l’Ubaye (Alpes-de-Haute-Provence, France): présentation des premiers résultats d’un Projet collectif de Recherche. Preistoria Alpina 42, pp. 23-48. Walsh K. 2005. Risk and marginality at high altitudes: new interpretations from fieldwork on the Faravel Plateau, HautesAlpes. Antiquity 79, pp. 289-305. Walsh K., Mocci F., & Palet-Martinez J.M. 2007. Nine thousand years of human/landscape dynamics in a high altitude zone in the southern French Alps (Parc National des Ecrins (Hautes-Alpes (05)). Preistoria Alpina 42, pp. 9-22. Walsh K. & Richer S. 2006. Attitudes to altitude: changing meanings and perceptions within a “marginal” Alpine landscape – the integration of palaeoecological and archaeological data in a high altitude landscape in the French Alps. World Archaeology 38, pp. 436-454.

Acknowledgments The palaeoecological work is carried out with the support of the Institut Méditerranée d’Ecologie et Paléoécologie, Université Paul Cezanne, Aix-Marseille III (France). The following specialists have made essential contributions to our research in the Alps. The study of flint material was undertaken by S. Tzortzis (SRA, PACA) and C. Bressy, J.-P. Bracco and A. D’Anna (Economies, Sociétés, Environnements Préhistoriques, UMR 6636, Aix-en-Provence), and that of the ceramic material by L. Vallauri (Laboratoire d’Archéologie Méditerranéenne Médiévale, UMR 6572, Aix-en-Provence). Charcoal analysis was undertaken by A. Durand (Laboratoire d’Archéologie Méditerranéenne Médiévale, UMR 6572, Aix-en-Provence) and B. Talon (Faculté de St Jérome, Université de Provence), Vanessa Py (l’Université d’Aix-Marseille I) and palynological

104

13 Fire making water on the Ligurian Apennines Roberto MaggiI, Andrea De PascaleII Direzione regionale per i beni culturali e paesaggistici della Liguria, via Balbi 10, 16126 Genova, Italy – [email protected] II Museo Archeologico del Finale, Istituto Internazionale di Studi Liguri, sez. Finalese Chiostri di S. Caterina, 17024 Finale Ligure Borgo (SV), Italy – [email protected] I

Abstract Research at jasper quarries and copper mines in the eastern Ligurian Apennines has pointed to the exploitation of substantial amounts of raw materials there since the mid-4th millennium BC, and it has been suggested that pastoralism was the main economic base supporting such activities. Multidisciplinary investigations of buried soils and peat bogs at various altitudes also suggest that widespread exploitation of the vegetative resources can explain widespread transformations of the soils. Cases such as the twenty years old example of Prato Mollo (1480 m asl; late 4th – 3rd millennium BC), Piana Damisa (600 m asl; Bronze Age: late 2nd millennium BC) and others show that the clearance of the early Holocene Abies forest also generated the waterlogging of basins. Therefore the authors envisage a model for late Neolithic-Bronze Age integrated construction of pastures and water meadows, suitable for both feeding and watering herds. Research currently ongoing at the Mogge di Ertola basin (1115 m asl; 8th millennium BC till late Roman) also suggests more articulated practices of woodland management.

1. The 4th-3rd millennium BC in the Ligurian Apennines: low population, few archaeologists, or both? The territory of eastern Liguria (NW Italy) is very rugged, with steep valleys and mountains that fall straight into the sea. The prehistoric archaeology of this area is very poor as far as settlement and burial are concerned. For the whole of the 4th and 3rd millennium BC, only one burial site yielded more than ten individuals (the collective cave burial Grotta dei Colombi on Palmaria island; Facchini & Veschi 1994). Villages with features such as huts, hearths, and walls are unknown. Ten boxes are enough to

Figure 1 – Burial cave (black circles), residential site (black square) and scattered finds (white squares) of the 4th and 3rd millennium BC in eastern Liguria.

store all of the artefacts from burials, settlements and scattered finds, and less than ten metal objects have been found so far (fig. 1). Does this indicate a low population number? According to the number and size of burials and residential places the answer is yes, but a different scenario arises when production sites are taken into consideration as well (fig. 2).

2. Production sites 2.1 Red jasper quarries

Several outcrops of red jasper (Radiolarite) occur in eastern Liguria. Excavations in the large open-air

Figure 2 – Production sites and biostratigraphical archives of the 4th and 3rd millennium BC in Eastern Liguria: 1. Valle Lagorara red jasper quarry, 2. Libiola copper mine, 3. Monte Loreto copper mine, 4. Piana Damisa and Pianaccia di Suvero buried soils, 5. Cian di Tenenti buried soil, 6. Prato Mollo peat bog, 7. Lago Riane peat bog, 8. Mogge di Ertola peat bog, 9. Casanova di Rovegno peat bog, 10. Pian del Lago peat bog.

105

Hidden Landscapes

Figure 3 – Aerial view of the open-air red jasper quarry at Lagorara valley (Maissana, La Spezia).

red jasper quarry at Lagorara valley (Maissana, La Spezia; fig. 3) revealed that, from 3500 to 2500 cal. BC, jasper was mainly exploited for the production of small, oval chipped bifaces. These are the pre-forms for foliate arrowheads of the type found scattered on the mountains, in settlements and in burials (Maggi et al. 1996; Campana & Maggi 2002). The dimensions of the extraction niches resulting from the quarrying suggest that at least 1000 cubic meters of material has been exploited. Since the exploited material was mainly used to produce arrowheads, it has been calculated that some 5-10 millions of these must have been produced (Maggi 2002). Where have all of these artefacts gone, if less than 50 arrowheads have been found so far? 2.2 Copper mines

The ophiolithic area located around Sestri Levante (Genova), just a few kilometres away from the coast and not far from the Lagorara valley, hosted several copper mines in the 19th century. Some of these, such as Monte Loreto and Libiola, provided evidence of medieval and prehistoric exploitation. Between 1996 and 2005 the Soprintendenza per i Beni Archeologici della Liguria and the Department of Archaeology of the University of Nottingham

investigated the conspicuous remains left at Monte Loreto (367 m asl), where prehistoric copper mines and ore-dressing facilities have been recorded in various places along the slopes of the mountain. Twelve radiocarbon dates demonstrate that early mining and ore exploitation occurred from the mid4th to the late 3rd millennium cal. BC (Maggi & Pearce 2003, 2005). It appears that the technique employed was able to quarry almost exclusively the ore, leaving the embedding rock largely untouched, except where it was necessary to widen the works for access. The prehistoric mining features include dumps and shafts, as well as evidence of fire-setting practices. In a fissure investigated between 1999 and 2003 (fig. 4) there were layers very rich in charcoal dated 4090±60 BP (2875-2475 cal. BC - 2ı; Beta-135159); the fill also contained hammer stones and Copper Age pottery. A total of seven hundred complete and fragmented hammer stones, many showing damage from use, were collected during the fieldwork campaigns. They are made of basalt, dolerite, sandstone, diorite and gabbro, with sources sometimes a few kilometres away from the site. The hammers have single or double notches or are grooved for hafting. It is interesting to note that the same particularly hard types of rock were also used at the contemporary jasper quarries of Valle Lagorara, located near

106

Maggi – De Pascale, Fire making water on the Ligurian Apennines

Monte Loreto (Cortesogno et al. 2006; De Pascale 2004a, 2004b). The investigation of in-situ ore treatment areas and of exploited veins allowed us to calculate the total amount of quarried material. Keeping in mind that other prehistoric mines existed at the same time, such as at Libiola, it has been suggested that the Sestri Levante mining district produced an amount of copper ore suitable for casting up to perhaps 5000 artefacts per year (Campana et al. 2006). These indicators of a quite significant production of arrowheads and of copper ore contrast with the low demography suggested by the scant remains of villages and cemeteries, which in themselves would not allow us to argue for an extensive exploitation of the environmental resources. Therefore a question arises automatically: is there evidence of subsistence practices to support a level of population consistent with the quarrying and mining activities calculated above? The answer is yes, such evidence exists, lying hidden in the landscape (figs. 5-6).

Micromorphology also indicates the occurrence of shallow waterlogged features. The Bronze Age transformation was so strong that it seems the area never went back to woodland (Ottomano et al. 2006). A similar situation occurs at the nearby site of Pianaccia di Suvero (600 m asl), with a long Middle Palaeolithic, Mesolithic, Neolithic and Bell Beaker history. During the late Bronze Age a shallow terrace was built, that promoted the accumulation of an extended soil similar to that of Piana Damisa (Macphail 1987; Gernone & Maggi 1998). Bubble flotation of a significant soil sample yielded (despite

3. Subsistence (in the hidden landscape) 3.1 Buried soils

In Eastern Liguria buried soils occur widespread at different locations and altitudes. The most extensive date to the late Bronze Age. The main archaeological features of this period are the so-called ‘Castellari’, open-air sites located strategically on rocky hilltops, which probably functioned as control sites for both pastures and watershed routes of a short-distance transhumance system. Piana Damisa (600 m asl) is a broad field located about 1200 m south of the middle to late Bronze Age site of Castellaro di Zignago (950 m asl) and 1100 m north of Castellaro di Vezzola (500 m asl, also late Bronze Age). Test trenches revealed that what is now a large flat field, before the Bronze Age was in fact a valley, filled by colluvium that began in the late Bronze Age. In fact this is the date of two extensive buried soils, stratigraphicaly located at the bottom of the infill and dated to 3180±50 BP (1525-1380 cal. BC - 2ı; Beta 100354) and 3160±50 BP (1515-1305 cal. BC - 2ı; Beta 100355). The absence of artefacts and features such as pottery, huts and fireplaces suggests that this site did not have a residential function. Micromorphological analyses (carried out by C. Ottomano) suggest that the buried soil was formed as a consequence of activities such as woodland clearance by fire. The considerable organic enrichment of the Bronze Age soil suggests that the site was intensively visited by herds.

Figure 4 – A shaft of the prehistoric copper mine at Monte Loreto (Castiglione Chiavarese, Genova). Photo: Centro Studi Sotterranei, Genova.

Figure 5 – A weighted comparison between the archaeology of artefacts versus the archaeology of hidden landscapes.

107

Hidden Landscapes

an abundance of charcoal) no seeds of domesticated plants, suggesting that the area was not used for agriculture. Another extended late Bronze age-early Iron age buried and terraced soil, again rich in charcoal and organic matter but lacking any residential domestic evidence, has been found 6 km from the Castellaro di Uscio, at Case Cordona (Cremaschi et al. 1992). Among several late Bronze Age soils, at least two belong to the 3rd millennium BC: Giridello 2 (Campana et al. 1998) and Cian di Tenenti (San Colombano, Genova; Maggi 2000,

2004a). Here too, bubble flotation shows that no seeds of domesticated plants were associated with the charcoals and organic matter. Furthermore, two buried soils that do include tiny fragments of pottery go back to the middle and early Neolithic (Giridello 1 and Isolalunga 1; Maggi 2003, 2004a). The long use-history, the location away from habitations, the evidence for forest clearing/ colluvium, the dark brown color, the abundance of charcoal and organic matter, the absence of cultivated seeds, and the occurrence of features such as terracing, all suggest that these extensive

Figure 6 – Organising the Ligurian chronology of the 4th and 3rd millennium BC.

108

Maggi – De Pascale, Fire making water on the Ligurian Apennines

buried soils are the archaeological remains of productive practices probably related to the management of herds. 3.2 Peat bogs

Prato Mollo (1480 m asl) is one of several peat sites in eastern Liguria where palynological and sedimentological studies were carried out by R. I. Macphail and G. M. Cruise (Baffico et al. 1987; Courty et al. 1989, pp. 305-309). The site is a shallow depression (possibly a wurmian glacial cirque) located on the southern slopes of M. Aiona and M. Nero, where Mesolithic and Copper Age flint artefacts have been recovered. The Abies alba forest around Prato Mollo underwent a major decline during the period 4300±60 BP (3079-2642 cal. BC - 2ı; Bln 3132) to 4130±60 BP (2889-2472 cal. BC - 2ı; Bln 3131), in response to multiple widespread fires. In the examined stratigraphy the grey-blue serpentinite gravel below the peat shows that, after the ice melted, the basin was permeable perhaps up to the 4th millennium BC; micromorphological analysis of the sediment show that the layer of serpentinite gravel was then sealed and made impermeable by clay infilling as a result of the erosion caused by the disturbance of the surrounding forest soil. This caused waterlogging conditions favouring the accumulation of peat. A large amount of charcoal provides the dark colour for two bands located at the base of the peat (fig. 7); micromorphological analysis shows that each dark band is multilayered and that the grass burned together with the trees (Courty et al. 1989, pp. 305-309). Most researchers agree that the invasive human exploitation during the late 4th and 3rd millennia cal. BC was intended to improve pastures. As the Abies forest was opened up locally, some pollen taxa which were transported up from lower altitudes, e.g. Quercus and Ulmus, became statistically more important. Open Fagus woodland expanded greatly and became dominant after the Abies forest had been cleared. It is well known that beech wood-pastures provide the best fodder for cows, and the locally manufactured San Ste cheese is still very much appreciated. The pollen diagram shows that the marked fall of the primary Silverfir from 40% to 20% occurred together with the evidence for fire, erosion, waterlogging, and peat formation. In contrast, as already written, there is an increase of grasses and of beech, which today is dominant (Baffico et al. 1987). It can therefore be argued that, at the beginning of the 3rd millennium

Figure 7 – Prato Mollo (Borzonasca, Genova): two multilayered bands full of charcoal, indicating clearing by fire (modified after Baffico et al. 1987).

Figure 8 – Mogge di Ertola (Rezzoaglio, Genova), waterlogged trees buried in the peat, sealed by an early medieval clay-silty layer.

BC, fire-based practices were used to manage the woodland cover and the soil to create both pastures and water reservoirs (for watering herds?). Pian del Lago (830 m asl) is a small peat-filled basin with a maximum depth of 6 m. It is situated on the seaward slopes of Mount Roccagrande at Casarza Ligure, only 9 km inland from the coast of Sestri Levante, in what is now a highly eroded

109

Hidden Landscapes

Figure 9 – Comparison between Prato Mollo and Mogge di Ertola.

heath landscape. The stratigraphy spans the whole Holocene through to the Middle Ages: the bottom dates to 10870±90 BP (11040-10640 cal. BC - 2ı; GnR 21307) and the top to 700±60 BP (1220-1400 cal. AD - 2ı; GnR 21308). The pollen diagram, which is being studied by G. M. Cruise, shows the presence of domesticated cereals at levels dated two to four centuries before 5000 BC. As it is well known that cereal pollen is heavy and the wind could not transport it more than a few hundred meters, this suggests local agriculture. The pollen could also have been transported by sheep coming to drink (Cruise et al. 1998; Cruise & Maggi 2000; Maggi 2004b), but these would then predate the rise of pastoralism. These earliest domesticated cereals are closely followed by the first appearance of Vitis pollen, which are therefore also definitely earlier than 5000 BC (Cruise & Maggi 2000). Charred pips of Vitis vinifera L. are known from various settlements of the earliest Neolithic in Italy (Castelletti & Rottoli 1998). In the Late Neolithic, around 4300-4200 BC, Olea pollen become a bit more frequent: this could indicate some kind of management. The walnut tree (Juglans) appears in the early centuries of the Copper

Age, between 3500 and 3000 BC (Cruise et al. 1998; Cruise & Maggi 2000; Maggi 2004b), also suggesting possible productive practices. Preliminary results of the research therefore suggest that the vegetation and the soils of the area around Pian del Lago were disturbed by grazing during the Neolithic and the Copper to Early Bronze Age. The intensive use of fire for controlling the vegetation was unequivocally attested for the late Iron Age and the Roman period, when there is evidence for a greater impact of deforestation and heathland spreading (Maggi 2000, 2004b; Cruise et al. 2009). Another peat bog near Rezzoaglio (Genova), locally called Mogge di Ertola (1115 m asl), was subjected to a preliminary investigation in 2001 (Guido et al. 2003). In 2004-2006 three campaigns of environmental archaeological research were then carried out jointly by the Direzione regionale per i beni culturali e paesaggistici della Liguria and the University of Genova, Dipartimento di Storia Moderna e Contemporanea and Dipartimento per lo studio del Territorio e delle sue Risorse. Coordinated by D. Moreno, C. Montanari, M.A. Guido and R. Maggi, this study focused on the historical ecology of the site (Moreno et al. 2005; De Pascale et al. 2006; Bellini et al. 2007; Menozzi et al. 2007). The remains of a stone structure close to a clay edge of the bog, the surface of which shows evidence of burning, and which is buried by peat dated to 4660 ± 50 BP (3540–3350 cal. BC - 2ı; LTL2215A), is interpreted as a dam. Several waterlogged trees are buried in a peat that contains charcoal (fig. 8). The chronology of this layer of buried silver firs at Mogge di Ertola agrees with that of the multilayered charcoal bands at Prato Mollo, while the peat bog itself dates from as early as 8912 ± 100 BP (8300-7700 cal. BC - 2ı; LTL547A) up to the Roman period: 1977 ± 50 BP (110-130 cal. AD - 2ı; LTL776A). The two sites are located some 7 km apart in the eastern Apennines, and both are close to the watershed: Prato Mollo at 1480 m altitude and Mogge di Ertola at 1115 m. Despite the similarity in location, the ecological history of the two sites is different (figs 9 and 10). The most visible difference occurs perhaps in the 3rd millennium BC, when Prato Mollo shows evidence of the clearing by fire of the silver fir forest, while Mogge di Ertola contains the actual silver firs fallen down. At the present stage of research we do not know if, how and why the two phenomena might correlate. At Prato Mollo the signs of environmental practices detected so far place the shift from a ‘natural’ to

110

Maggi – De Pascale, Fire making water on the Ligurian Apennines

Figure 10 – Comparison of the local history of Prato Mollo and of Mogge di Ertola against the background of the regional history.

a ‘man-made’ landscape in the late 4th – early 3rd millennium BC; at Mogge di Ertola the felled silver firs are part of a perhaps more complex history, beginning with the formation of a colluvial clay that sealed the basin around 7200 BC, if not earlier. It is important to note that this clay formation below the peat contains several charcoals (mainly of silver fir), suggesting that human practices might also here be involved in the lake/wetland formation at a surprisingly early time (Mesolithic) and apparently with more ‘hidden’ practices. We do not know yet the level of detail that the ongoing multidisciplinary research at Mogge di Ertola will reach, nor if and to what degree it will be able to answer the questions posed in fig. 10. However it is certain that the site, together with the previously investigated Prato Mollo and the other sites mentioned, is emblematic of the explanatory potential hidden in the landscape.

References Baffico O., Cruise G.M., Macphail R.L., Maggi R. & Nisbet R. 1987. Monte Aiona – Prato Mollo, in Melli P. & Del Lucchese A. (eds), Archeologia in Liguria III.1 – Scavi e scoperte 1982 – 86. Genova, Soprintendenza Archeologica della Liguria, pp. 57–66. Bellini C., Cevasco R., Moreno D., Guido M.A. & Montanari C. 2007. Mogge di Ertola, Aveto valley, Ligurian Apennines: evidence of past cultural landscapes, in Krzywinski K.,

O’Connell M. & Küster H. (eds), Cultural Landscapes of Europe, Fields of Demeter, Haunts of Pan. Delmenhorst, Aschenbeck & Oeljeschläger, pp. 108-109. Campana N. & Maggi R. (eds) 2002. Archeologia in Valle Lagorara. Diecimila anni di storia intorno a una cava di diaspro. Firenze, Istituto Italiano di Preistoria e Protostoria. Campana N., Colombi N. & Maggi R. 1998. Giridello (Rocchetta Vara – SP), in Del Lucchese A. & Maggi R. (eds), Dal diaspro al bronzo. La Spezia, Luna editore, pp. 164-166. Campana N., Maggi R., Pearce M. & Ottomano C. 2006. Quanto rame? Stima della produzione mineraria del distretto di Sestri Levante fra IV e III millennio BC, in Materie prime e scambi nella Preistoria Italiana, Atti IIPP 39. Firenze, Istituto Italiano di Preistoria e Protostoria, pp. 1339-1348. Castelletti L. & Rottoli M. 1998. L’agricoltura neolitica italiana. Una sintesi delle conoscenze attuali, in Pessina A. & Muscio G. (eds), Settemila anni fa il primo pane. Ambienti e culture delle società neolitiche. Udine, Museo Friulano di Storia Naturale, pp. 15-24. Cortesogno L., De Pascale A., Gaggero, L., Maggi R. & Pearce M. 2006. Strumenti litici per estrazione mineraria: il caso di Monte Loreto, in Materie prime e scambi nella Preistoria Italiana, Atti IIPP 39. Firenze, Istituto Italiano di Preistoria e Protostoria, pp. 683-695. Courty M.A., Goldberg P. & Macphail R.I. (eds) 1989. Soils and Micromorphology in Archaeology. Cambridge: University Press. Cruise G.M., Macphail R.I., Maggi R., Engelmark R., Linderholm J., Haggart B.A. & Moreno, D. 1998. New approaches to old problems: Neolithic to Medieval land-use at “Lago” di Bargone, Eastern Liguria, Italy, in De Marinis R., Bietti Sestieri A.M., Peroni R. & Peretto C. (eds), Atti del XIII Congresso UISPP. Volume 1. Forlì – Italia 8 – 14 settembre 1996. Forlì, Abaco, pp. 401-413.

111

Hidden Landscapes

Cruise G.M. & Maggi R. 2000. Pian del Lago (Bargone), paesaggio costruito e paesaggio naturale tra la fine della glaciazione ed il medioevo, in Figone F., Franceschini I. & Stagnaro A. (eds), Museo Parma Gemma, ventenni di attività culturale e di ricerche. Recco, Comunità Montana Val Petronio, pp. 10-13. Cruise G.M., Macphail R.I., Linderhom J., Maggi R. & Marshall P.D. 2009. Lago di Bargone, Liguria, N. Italy: a reconstruction of Holocene environmental and land-use history, The Holocene 19(7), pp. 987-1003. De Pascale A. 2004a. Studio preliminare dei mazzuoli litici della miniera preistorica di Monte Loreto. Analisi formale e classificazione, in Giannichedda E. (ed.), Metodi e pratica della Cultura Materiale: produzione e consumo dei manufatti (Atti della Scuola Interdisciplinare delle Metodologie Archeologiche – S.I.M.A.). Bordighera, Istituto Internazionale di Studi Liguri, pp. 53–58. De Pascale A. 2004b. «Hammerstones from early copper mines»: sintesi dei ritrovamenti nell’Europa e nel Mediterraneo orientale e prime considerazioni sui mazzuoli di Monte Loreto (IV millennio BC - Liguria). Rivista di Studi Liguri 69 (2003), pp. 5-42. De Pascale A., Maggi R., Montanari C. & Moreno D. 2006. Pollen, herds, jasper and copper mines: economic and environmental changes during the 4th and 3rd millennia BC in Liguria (NW Italy). Environmental Archaeology 11.1 (2006), pp. 115-124. Facchini F. & Veschi S. 1994. I reperti preistorici dell’Isola Palmaria (La Spezia): revisione critica dei ritrovamenti e nuove osservazioni antropologiche. Rivista di Antropologia 72, pp. 163-195. Gernone G. & Maggi R. 1998. Lavorazione della steatite alla Pianaccia di suvero (Alta Val di Vara, La Spezia), pp. 9597 in Mottes E., Nicolis F. (eds), Simbolo ed Enigma. Il bicchiere campaniforme e l’Italia nella preistoria europea del III millennio a.C. Riva del Garda, Provincia Autonoma di Trento. Guido M.A., Menozzi B.I., Montanari C. & Scipioni S. 2003. Il sito Mogge di Ertola come potenziale fonte per la storia ambientale del crinale Trebbia/Aveto. Archeologia Postmedievale 6 (2002), pp. 111-116. Macphail R.I. 1987, Soil report, pp. 24–28 in Melli P. & Del Lucchese A. (eds), Archeologia in Liguria III.1 – Scavi e scoperte 1982 – 86. Genova, Soprintendenza Archeologica della Liguria. Maggi R. 2000. Aspetti di archeologia del territorio in Liguria: la formazione del paesaggio dal Neolitico all’Età del Bronzo. Ambienti e storia della Liguria. Studi in ricordo di Emilio Sereni - Annali Istituto Alcide Cervi 19 (1997), pp. 143–162. Maggi R. 2002. La storia olocenica di Valle Lagorara, in Campana N. & Maggi R. (eds), Archeologia in Valle Lagorara. Diecimila anni di storia intorno a una cava di diaspro. Firenze, Istituto Italiano di Preistoria e Protostoria, pp. 367-373. Maggi R. 2003. Suoli sepolti e paesaggio sull’Appennino ligure, in Chiaramonte Trerè C. (ed.), Antichi liguri sulle vie appenniniche tra Tirreno e Po. Milano, Quaderni di Acme 61, pp. 161-173. Maggi R. 2004a. Alle radici delle trasformazioni del paesaggio, in Spinetta R. (ed.), Neirone. Natura, Storia, Arte. Genova, Comune di Neirone, pp. 99-105.

Maggi R. 2004b. I monti sun eggi: the making of the Ligurian landscape in prehistory, in Balzaretti R., Pearce M. & Watkins C. (eds), Ligurian Landscapes, studies in archaeology, geography & history in memory of Edoardo Grendi. London, Accordia Research Institute, pp. 71-82. Maggi R. & Pearce M. 2003. Excavations at the Fourth Millennium cal BC Copper Mines at Monte Loreto (Liguria – NW Italy), Archaeometallurgy in Europe (International Conference – Proceedings, vol. 1). Milano, A.I.M., pp. 587-596. Maggi R. & Pearce M. 2005. Mid fourth-millennium copper mining in Liguria, north-west Italy: the earliest known copper mines in Western Europe. Antiquity 79 (2005), pp. 66-77. Maggi R., Campana N., Negrino F. & Ottomano C. 1996. The quarrying and workshop site of valle Lagorara (Liguria – Italy). Accordia 5 (1994), pp. 73-96. Moreno D., Cevasco R., Guido M.A. & Montanari, C. 2005. L’approccio storico-archeologico alla copertura vegetale: il contributo dell’archeologia ambientale e dell’ecologia storica, in Caneva G. (ed.), La biologia vegetale per i beni culturali, conoscenza e valorizzazione, Vol. II. Firenze, Nardini Editore, pp. 463-498. Menozzi B.I, Bellini C., Cevasco A., Cevasco R., De Pascale A., Guido M.A., Maggi R., Moe D., Montanari C. & Moreno D. 2007. The archaeology of a peat bog in context: contribution to the study of biodiversification processes in historical time (Ligurian Apennine, NW Italy), in pre-prints of the 4th International Congress of Medieval and Modern Archaeology – Topic 7, Environmental Archaeologies. Paris. Ottomano C., Giannichedda E., Del Soldato M. & Maggi R. 2006. Piana Damisa: un’unità produttiva territoriale dell’Età del Bronzo costruita col fuoco, in Cocuzza N. & Medri M. (eds), Archeologie, Studi in Onore di Tiziano Mannoni. Bari, Edipuglia, pp. 257- 264.

112

14 Shepherds of a coastal range: the archaeological potential of the Velebit mountain range (eastern Adriatic) Stašo Forenbaher Institute for Anthropological Research, Gajeva 32, HR-10000 Zagreb, Croatia – [email protected] Abstract The limestone summits of Velebit, a long coastal range overlooking the Eastern Adriatic, reach an elevation of around 1700 m just a few kilometers away from the shore. They define a sharp boundary between contrasting maritime and continental environments. Different environmental zones, stacked one above another, enhance the variability of the landscape, making the area attractive for seasonal pastoralists. For a long time Velebit was neglected by archaeologists. Rugged terrain, difficulty of access and a notion that nothing was to be found in those remote mountains kept researchers away. Common problems of archaeological detection of seasonal pastoralists are compounded by intensive erosion which has obliterated or buried most of the open-air sites in this heavily karstified landscape. The archaeological potential of Velebit is nevertheless great. Transhumant sheepherding has survived into modern times, providing the possibility of ethnoarchaeological research. Rich archival records containing legal documents regulating access to pasture and water rights, extend from Austro-Hungarian times back to the era of Roman Imperial administration. Thick stratigraphic sequences in karstic caves provide the opportunity to explore long-term change in herding practices since the time of the first Neolithic farmers.

1. Introduction Velebit is a 130 km long mountain range stretching along the coast of the north-eastern Adriatic (Poljak 1969). It rises abruptly from the sea and reaches its highest elevation just a few kilometers inland. Numerous peaks of this remarkably compact range reach elevations between 1600 and 1800 m, while only three passes lying below a thousand meters above sea level offer obvious transversal routes of communication (fig. 1). The mountain forms a sharp climatic and environmental barrier between the Mediterranean coast and its continental hinterland (Penzar & Penzar 1995). As a consequence, its maritime and continental aspects differ greatly. The maritime slope (Rogić 1958) is heavily karstified, especially in its lower reaches. The limestone bedrock of which the range is composed is everywhere evident, creating a rugged maze of crags and knolls. In this extreme erosional environment, sub-Mediterranean vegetation is degraded by long-term human impact and the only patches of soil worth mentioning are terra-rossa-filled depressions at the bottom of dolines (fig. 2). With increasing altitude, the landscape becomes more wooded (Forenbacher 1990). First there are groves of Mediterranean pine, above which follow mixed forests of beach and fir that give the highest part of the range a sub-Alpine character. Patches of grassland, varying in size from a few hectares to a couple of square kilometers, are scattered across those elevations. The largest areas of open grassland extend along the ridges, at altitudes above 1500 m, and in the slightly lower-lying karstic depressions between the rocky, forested peaks (fig. 3). From there,

the heavily wooded continental slope drops down to the Lika highlands, at 600 m above sea level. The Velebit range receives most of its precipitation during winter (Perica & Orešić 1995). Depending on the altitude, it comes as rain or snow. Along the summits and the continental slope, snow often lingers until early summer. Because of the karstic terrain there are no permanent surface streams of any significance: water is immediately lost underground through extensive sinkhole and cave systems, eventually to reappear at submarine springs along the coast. Summers tend to be dry, with intensive insolation, and desiccation is aggravated by the bura, a dominant cold and dry wind which sweeps down the maritime slope on every other day of an average year (Tutis & Ivančan-Picek 1998). Consequently, in spite of heavy winter snow cover and abundant rainfall, water (or the lack of it) is the critical limiting factor for exploitation (Rukavina 1977).

Figure 1 – Map of the Velebit mountain range.

113

Hidden Landscapes

Figure 2 – Dolac poda Tulom, a mid-altitude pasture at 800 m a.s.l., photographed in late spring.

Figure 3 – Javornik, a high altitude pasture and a seasonal dwelling at 1300 m a.s.l., photographed in summer 1976.

A rich variety of wild flora and fauna inhabits the Velebit range, but the total quantity of exploitable wild foods is not great. Mountain vegetation (Forenbacher 1990) includes numerous plant species with edible bulbs, tubers, seeds, nuts, and fruits, but most of them are available only briefly and in quantities which can not support a large human population. Those more generally available, such as beech nuts, have low nutritional value (Grlić 1986). Large mammals, including red and fallow deer, mountain goat, wild boar, brown bear, wolf, lynx, marten, badger, fox and rabbit, are still relatively common due to the wildness and remoteness of the area (Forenbacher 2002), but their populations are fairly small. In the not so distant past, the valleys and lowlands would have been richer hunting grounds than the mountains. This would have been even more true during the Last Glacial period, when the highest reaches of the Velebit range were glaciated (Belij 1985; Bognar et al. 1991) and the northern Adriatic lowlands, exposed by the lowered sea level, would have offered much more productive hunting grounds (van Andel & Shackleton 1982, pp. 451; Miracle & O’Brian 1998, pp. 44-45). It is therefore not surprising that

the earliest, still ambiguous evidence of forager presence on the mountain dates from the early Holocene, a time when most of the Adriatic lowlands were flooded by marine transgression, and local populations were compelled to seek subsistence in the surrounding highlands (Shackleton et al. 1984, p. 311; Forenbaher 2002, pp. 362-366). The evidence consists of a handful of flaked stone artifacts recovered from a context underlying the earliest Neolithic levels at one of the most informative prehistoric sites on Velebit, Vaganačka Cave (Forenbaher & Vranjican 1985, pp. 7-8 and plate 1:1-3). Although marine resources are available along the coast, represented by a variety of fish and mollusks, again they occur in limited quantities. The Adriatic sea is relatively poor in plankton, and unlike some other seas does not support large populations of anadromous fish or marine mammals. Nor is the agricultural potential of the mountain range significant (Rogić 1958). Along its maritime slope, less than 2% of the land is suitable for cultivation. At high altitudes the growing season is short due to low temperatures and snow cover, while lower down it is cut short by summer drought. Extreme daily temperature swings and poor soil quality allow the growth only of hardy cultivars, with low and uncertain yields. The most important cultivar is the potato, which was introduced to the area in the late 18th century. Some cereals and vegetables are also grown, but not enough to support the local population which traditionally imports wheat in exchange for locally grown figs. More sensitive Mediterranean cultivars such as vine and olive can survive only by exception, in a few sheltered areas along the coast. The scattered grasslands of the Velebit mountain range are inferior to the Alpine pastures in quality and productivity, but they cover a third of the total land area and are distributed at all altitudes, throughout the different environmental zones. Some pastures are thus available for grazing at any particular time of the year, and these are best fitted for raising sheep and goats. Optimal exploitation of these grasslands can be accomplished by moving the herds up and down the mountain, following the best grass as its location changes by the season. The Velebit range thus provides a natural setting which invites ‘vertical’ transhumant sheep and goat herding as the soundest subsistence strategy.

2. Ethnographic and historical sources Transhumant shepherds move their herds and their households in regular annual cycles between fixed

114

Forenbaher, Shepherds of a coastal range: the archaeological potential of the Velebit…

bases located in different environmental zones. In our case, these zones are stacked tightly one above the other on the slopes of the Velebit range. Until a few decades ago, before tourism developed along the coast, sheep and goat herding was the main economic activity on and around the mountain (Rogić 1958; Marković 1980). It is still practiced to a small extent, but the last traditional cycles of seasonal mobility were shattered by the war some twenty years ago. Seasonal herding on Velebit was at its peak during the second half of the 19th century. It is estimated that, at that time, some 50,000 sheep and as many goats grazed the Velebit grasslands (Rogić 1958, p. 70). Herds were marshaled up and down the mountain by three different groups of shepherds, each based on a different side of the mountain range, and each following a specific annual mobility cycle (fig. 4) (Marković 1980, pp. 8-10 and pp. 53-64). Shepherds from the littoral spent winters with their herds near the coast. In spring, as grass dried out at lower altitudes, the staged ascent to ever higher pastures began. These one-step-at-a-time moves were short, each one accomplished in less than a day. As a rule, every household possessed two equally well-built houses, one near the coast and the other halfway up the mountain (fig. 5). In late spring, the entire household moved with all its belongings to the higher house. In addition to these there were other, less solidly built shelters at various altitudes. Caves and rock shelters were often used as well, sometimes by shepherds, but mostly by their herds (fig. 6). By late summer some of the herds from the littoral reached the highest pastures along the summit ridge; their retreat would begin in early September. Animals and shepherds reversed their spring movements, finally to overwinter near the coast. Shepherds from the interior were based in the Lika highlands, around 600 m above sea level, where adequate pasture is available through much of the year. They too drove their herds to mountain pastures in early summer, primarily to keep them away from the maturing crops. The short move from the highlands to the summit ridge was accomplished in a single day. As opposed to the shepherds from the littoral, only a few youths usually accompanied the herds, while most people remained at home for agricultural work. The sojourn on the mountain was relatively brief, the mountain dwellings simple, and only the most essential equipment was brought. By mid-summer the herds descended back to Lika, vacating the highest pastures which then became

available for use by the shepherds from the littoral. Some of the herds remained at their bases on the highlands during the winter months, fed on the hay harvested during summer. Others were driven over the Velebit passes before the first snows, overwintered in the northern Dalmatian lowlands, and returned to the Lika highlands in spring after the passes had reopened.

Figure 4 – Diagrams illustrating seasonal herding cycles on Velebit Mountain (adapted from Marković 1980).

Figure 5 – Reljinovac: substantial mid-altitude houses at 550 m a.s.l., photographed in spring 2005.

115

Hidden Landscapes

The third group of shepherds was based in the northern Dalmatian lowlands. In summer, when the grass in the lowlands was used up and dried out, the Dalmatian shepherds moved to the highland pastures of the southern Velebit, which are the largest and the best on the whole mountain range. Their seasonal moves traversed longer distances, usually took several days to complete, and sometimes involved ferrying the herds by boat across the deep inlets of the Adriatic. In a massive communal effort, several villages would combine to organize and equip a summer herding ‘company’ that would drive thousands of animals onto the mountain. The organ-

ization of the company was strictly hierarchical, consisting of a headman, his assistants, and senior and auxiliary shepherds, and it was supported by a pack mule train, which in late spring transported equipment and supplies to the seasonal highland settlements (figs 7 and 8) and in the fall brought back the cheese and other pastoral products. These interlocking cycles of mobility, which included occasional sharing of pasture lands, provided opportunities for interaction but also harbored potential for conflict. In the late nineteenth century the Austro-Hungarian authorities drafted elaborate legislation that specified the size, location and

Figure 6 – Stolačka peć: a cave used as a sheepfold at 900 m a.s.l., photographed in 1989.

Figure 7 – Dušice: a seasonal highland dwelling of Dalmatian shepherds at 1300 m a.s.l., photographed in summer 1985.

Figure 8 – Malovanski stanovi: seasonal highland dwellings of Dalmatian shepherds at 1600 m a.s.l., photographed in summer 1976.

116

Forenbaher, Shepherds of a coastal range: the archaeological potential of the Velebit…

ownership of all pasture lands, regulated passage of herds across other owners’ pastures, and set grazing fees (Marković 1980, pp. 13-23). Mandatory grazing permits were issued to shepherds, and reciprocal needs of different pastoral groups were resolved by reciprocal access to each other’s pastures. Particular care was paid to access rights to the scarce water resources. While earlier historical evidence is far less abundant, a constant trickle of legal documents regulating ownership of pasture, grazing rights, and passage of shepherds can be traced back to the 12th century AD (Rogić 1958, pp. 99-102; Marković 1980, pp. 13 and 125). These were issued by whichever authority happened to control the area, be it the Habsburg monarchy, the Republic of Venice, or Croatian feudal lords. Of particular interest, however, is one unusual monument from the Roman period, hidden away in a remote corner of the Velebit range, and known as ‘the inscribed stone of Begovača’ (Brunšmid 1901a, pp. 99-101). Engraved on a massive boulder near a perennial spring, a Latin inscription (fig. 9) cites a decision by a Roman magistrate settling a dispute over water rights between two autochthonous communities: the Ortoplini, based on the coast, and the Parentini from the interior. The conflict is resolved by confirming that the spring is located within the territory of the Parentini, while allowing the Ortoplini access to the water along a narrow footpath. The inscription does not explain what these people were doing high up on the mountain, far away from their bases, and why the spring was so important to them. It seems reasonable to assume, however, that the dispute broke out between two groups of shepherds who were bringing their herds to high summer pastures, and who both depended on the only reliable spring in the area.

3. Prehistoric herders Archaeological investigations of Velebit’s prehistory have been of a relatively low standard and quite limited in extent. Small test trenches were excavated at a few caves (Forenbaher & Vranjican 1982, 1985; Forenbaher 1991) and at even fewer open-air sites (Glavičić 1984, pp. 7-10; Faber 2000). Several researchers gathered information through nonsystematic field survey, especially in the southern part of the range which is now beginning to receive more focused attention (Glavičić 1966, 1968, 1982, 1984; Forenbaher & Vranjican 1990; Dubolnić 2006a, 2006b, 2006c). One may add to this a small number of chance finds collected by non-professionals over

Figure 9 – Latin inscription on ‘the written stone of Begovača’: EX CONVENTIONE FINIS | INTER ORTOPLINOS ET PARE | NTINOS ADITUS AD AQUAM | VIVAM ORTOPLINIS PASUS | D LATUS I

a century ago, or acquired more recently through illicit excavation (Brunšmid 1901b; Glavičić 1968, p. 14; Težak-Gregl 1984; Glogović 2000, p. 11). As a consequence, only a preliminary assessment of Velebit’s prehistory can be offered. The soundest diachronic information comes from caves that are scattered across the mountain’s heavily carstified maritime slope, from the shoreline up to about a thousand meters above sea level. Many caves have yielded surface-collected evidence of prehistoric use, while a few that were testexcavated have revealed deep stratigraphies that span most of the Holocene. The most informative of the latter is Vaganačka Cave, where a 1 × 2 m test trench was excavated to a depth of some four meters (Forenbaher & Vranjican 1985). Its regular and easily distinguishable strata yielded finds that cover the period from around 6000 BC up to the last centuries BC (fig. 10). As was already mentioned, extremely modest evidence has been recovered so far that may predate Neolithic times. The earliest substantial traces of human presence on the mountain coincide with introduction of farming into the region (Forenbaher & Miracle 2005). This intensive occupation begins with herders who appear rather abruptly sometime during the sixth millennium BC. From the very outset, faunal assemblages associated with their activities are heavily dominated by sheep and goat remains, a feature that will persist throughout prehistory. These earliest Velebit shepherds are linked to the early farming villages of northern Dalmatia by several formal characteristics of their artifactual assemblages, best visible in the pottery assemblages. Early Neolithic Cardial Impressed Ware (Batović 1979, pp. 498-509; Müller 1994) is scarce, but nevertheless present; both ‘Danilo-style’ Middle Neolithic pottery (Batović 1979, pp. 540551; Forenbaher et al. 2004, pp. 87-89) and ‘Hvarstyle’ Late Neolithic pottery (Batović 1979, pp. 592-

117

Hidden Landscapes

603; Čečuk & Radić 2005, pp. 149-160) are well represented and clearly related to contemporary sites in the northern Dalmatian lowlands. Pottery and flaked stone artifacts, faunal remains, as well as the character of the stratified deposits which, like those in the caves of the Trieste karst (Boschian & Montagnari-Kokelj 2000), appear to be composed primarily of fossilized sheep and goat droppings, indicate that the caves were mainly used as sheepfolds and shelters. Most likely these were seasonally occupied sites of the kind that Brochier (Brochier 1991) would call habitat bérgerie or grotte bérgerie. Their deep stratigraphic sequences testify to their continued use during the Copper Age, Bronze Age and Iron Age. It is tempting to see this evidence as residue left behind by prehistoric herders based in the northern Dalmatian lowlands, working their annual rounds in ways analogous to those known from historic times. For the moment, however, this interpretation must remain hypothetical, due to the limitations of the available data. Obviously, cave sites can provide only a partial picture of a complex seasonal herding strategy. We would like to know much more about the related

open-air sites, but this presents a serious problem in the extreme erosional environments in question. In most places, the soil cover that may have contained archaeological deposits is minimal or nonexistent, while in the rare locations where soil accumulates, prehistoric sites are likely to be buried under many meters of colluvium and are therefore not recorded by standard field survey procedures. Widely spaced isolated finds and rare surface scatters indicate that such sites did indeed exist, and that the wider landscape was used during prehistory, but have very limited interpretive value as the non-diagnostic lithics and heavily weathered potsherds can not be ascribed to a specific prehistoric period. Consequently, we have a very poor idea how they might fit into a Neolithic seasonal pastoral round. The situation is somewhat more favorable for later prehistoric periods, from which monumental architectural remains have survived. Related to major social changes that marked the beginning of the Bronze Age, hillforts and burial cairns appear along the eastern Adriatic littoral (Marović & Čović 1983, pp. 201-205) as well as in Velebit’s continental hinterland (Drechsler-Bižić 1983, pp. 244-248). On the

Figure 10 – A stratigraphic section and a selection of diagnostic pottery finds from Vaganačka cave.

118

Forenbaher, Shepherds of a coastal range: the archaeological potential of the Velebit…

Figure 11 – Mirovo: an abandoned seasonal highland settlement at 1350 m a.s.l., photographed in 2007.

mountain range itself, most of those structures and monuments lie close to the coast, below an altitude of 300 m. A few burial cairns and a couple of openair settlement sites have been recorded as high as 900 m above sea level, all of them positioned close to major pastures. This would suggest a settlement pattern similar to that of recent historic times, with permanent settlements located low along the coast and substantial seasonal dwellings at mid-altitudes.

4. Research potential In spite of the obvious limitations imposed by the natural environment and the scantiness of information collected mainly through non-systematic research, the archaeological research potential of the Velebit mountain range is considerable. First, numerous caves with deeply stratified sequences offer an opportunity to explore continuity and change in pastoral strategies from their introduction some eight thousand years ago up to the present day. Second, the extensive and detailed ethnohistoric record, if used judiciously, may guide our attempts to model prehistoric pastoral systems, explain variability among the sites, and determine how these might fit

into a larger regional picture. Furthermore, ethnohistoric information can be combined with archaeological exploration of recently abandoned seasonal settlements and mountain trails, remains of which abound throughout the mountain range (fig. 11), to help us refine our search for much older evidence of pastoralist activities and enrich our interpretations. Third, we may explore the Velebit range as a sharp natural barrier that, at the same time, is an area of contact. Nowhere else along the eastern Adriatic is the hinterland so close to the coast, nor is the transition so abrupt. During the early stages of the Neolithic the range would have been a frontier zone between foragers and farmers (Zvelebil & Lillie 2000, pp. 60-67; Forenbaher & Miracle 2005, pp. 523-524), with seasonal herders possibly playing a crucial role in their interactions. Unfortunately, with the exception of a single Danilo-style Neolithic sherd from Golubnjača Cave in Lika (DrechslerBižić 1970, plate 3:1), we know nothing about the earlier prehistory of Velebit’s continental hinterland, where research has focused entirely on rich Bronze Age and later remains. It seems unlikely, however, that the region was uninhabited prior to those pe-

119

Hidden Landscapes

to spot prehistoric burial cairns. Though separated by millennia, many must have recognized the view from that particular spot as somehow special.

References

Figure 12 – Mirila at Ljubotić.

riods, which leaves pre-Bronze Age Lika as another ‘hidden landscape’ waiting to be discovered. Finally, while it is clear that eight millennia of pastoral use and abuse are largely responsible for the current eroded and degraded character of Velebit, the details of the timing and intensity of human impact upon this fragile ecosystem remain to be worked out.

5. Post scriptum Most of those who visit the mountain today would agree that the tortured landscape of Velebit is quite spectacular. We may never know for sure whether it was perceived as such in the remote past, but there may be a few hints. Coming down any of the mountain trails, one often comes across places with particularly fine views of the next valley, the next hamlet, or the coast and the sea beyond. When you reach such a place, look around you, and you may notice dressed and carved stones lined up in rows along the trail. These are mirila, unusual burial monuments that were still erected as recently as fifty years ago (fig. 12). They mark places where those who had died on the mountain were set on the ground for a short rest while they were being carried down to be buried next to the church, in the village cemetery (Krajač 1934; Gavazzi 1961). Now, look around a bit more carefully, and somewhere nearby you are likely

Batović Š. 1979. Jadranska zona, in Benac A. (ed.), Praistorija jugoslavenskih zemalja, 2. Sarajevo, Akademija nauka i umjestnosti Bosne i Hercegovine, pp. 473-635. Belij S. 1985. Glacijalni i periglacijalni reljef Južnog Velebita. Beograd, Srpsko geografsko društvo. Bognar A., Faivre S. & Pavelić J. 1991. Tragovi oledbe na Sjevernom Velebitu. Geografski glasnik 53, pp. 27-39. Boschian G. & Montagnari-Kokelj E. 2000. Prehistoric Shepherds and Caves in the Trieste Karst (Northern Italy). Geoarchaeology: An International Journal 15(4), pp. 331371. Brochier J.-E. 1991. Géoarchéologie du monde agropastoral, in J. Guilaine (ed.), Pour une archéologie agraire. Paris, A. Colin, pp. 303–322 . Brunšmid J. 1901a. Arheološke bilješke iz Dalmacije i Panonije IV. Vjesnik Hrvatskog arheološkog društva, n.s. 5, pp. 87168. Brunšmid J. 1901b. Groblje bronsanog doba na Klačenici kod Jablanca (Kotar Senj). Povijest mjesta Jablanca. Vjesnik Hrvatskog arheološkog društva, n.s. 5, pp. 53-62. Čečuk B. & Radiý D. 2005. Vela spila: višeslojno pretpovijesno nalazište – Vela Luka, otok Korÿula. Vela Luka, Centar za kulturu. Drechsler-Bižić R. 1970. Zaštitna iskopavanja pećine Golubinjače kod Kosinja. Vjesnik Arheološkog muzeja u Zagrebu, 3. serija 4, pp. 111-117. Drechsler-Bižić R. 1983. Srednje brončano doba u Lici i Bosni, in B. Čović (ed.), Praistorija jugoslavenskih zemalja 4. Sarajevo, Akademija nauka i umjestnosti Bosne i Hercegovine, pp. 242-270. Dubolniü M. 2006a. Prapovijesna nalazišta na području Starigrada Paklenice. Radovi Zavoda za povijesne znanosti HAZU u Zadru 48, pp. 1-55. Dubolniü M. 2006b. Rezultati rekognosciranja Južnog Velebita (okolica Starigrada-Paklenice). Obavijesti Hrvatskog arheološkog društva 38(2), pp. 51-64. Dubolniü M. 2006c. Gradina Kneževiüi na Malom Libinju. Obavijesti Hrvatskog arheološkog društva 38(3), pp. 54-63. Faber A. 2000. Život na velebitskoj visoravni u pretpovijesno doba, Veliko Rujno. Senjski zbornik 27, pp. 15-44. Forenbacher S. 1990. Velebit i njegov biljni svijet. Zagreb, Školska knjiga. Forenbacher S. 2002. Kompendij velebitske faune. Zagreb, Veterinarski fakultet. Forenbaher S. 1991. Tragovi brončanodobnog služenja pećinom Separovačom kod Donje Klade. Senjski zbornik 18, pp. 217225. Forenbaher S. 2002. Prehistoric Populations of the Island of Hvar, an Overview of Archaeological Evidence. Collegium Antropologicum 36, pp. 361-378. Forenbaher S., Kaiser T. & Miracle P.T. 2004. Pupićina Cave Pottery and the Neolithic Sequence in Northeastern Adriatic. Ati della Società per la Preistoria e Protostoria della regione Friuli-Venezia Giulia 14, pp. 61-102.

120

Forenbaher, Shepherds of a coastal range: the archaeological potential of the Velebit…

Forenbaher S. & Miracle P.T. 2005. The Spread of Farming in the Eastern Adriatic. Antiquity 79, pp. 514-528. Forenbaher S. & Vranjican P. 1982. Pećina u Pazjanicama – Velika Paklenica, prilog pretpovijesti Hrvatskog primorja. Senjski zbornik 9, pp. 5-14. Forenbaher S. & Vranjican P. 1985. Vaganačka pećina. Opuscula archaeologica 10, pp. 1-21. Forenbaher S. & Vranjican P. 1990. Velebit, Survey of Caves and Rock-shelters. Arheološki pregled 29, pp. 237-239. Gavazzi M. 1961. Totenreststeine. Schweizerisches Archiv für Volkskunde 57. Glavičić A., 1966. Arheološki nalazi iz Senja i okolice I. Senjski zbornik 2, pp. 383-418. Glavičić A. 1968. Arheološki nalazi iz Senja i okolice II. Senjski zbornik 3, pp. 8-11. Glavičić, A. 1982. Nalazi kamenih gromila na Velebitu (I dio). Senjski zbornik 9, pp. 33-42. Glavičić A. 1984. Arheološki nalazi iz Senja i okolice VI. Senjski zbornik 10-11, pp. 7-28. Glogović D. 2000. Brončanodobne ostave iz Dalmacije. Opuscula archaeologica 23-24, pp. 11-20. Grlić Lj. 1986. Enciklopedija samoniklog jestivog bilja. Zagreb, August Cesarec. Krajač I. 1934. Mirila. Zbornik za narodni život i običaje Južnih Slavena 29(2), pp. 161-168. Marković M. 1980. Narodni život i običaji sezonskih stočara na Velebitu. Zbornik za narodni život i običaje Južnih Slavena 48, pp. 5-139. Marović I. & Čović B. 1983. Cetinska kultura, in Čović B. (ed.), Praistorija jugoslavenskih zemalja 4. Sarajevo, Akademija nauka i umjestnosti Bosne i Hercegovine, pp. 191-231. Miracle P.T. & O’Brien C.J. 1998. Seasonality of Resource Use and Site Occupation at Badanj, Bosnia-Herzegovina, Subsistence Stress in an Increasingly Seasonal Environment? In Rocek T.R. & Bar-Yosef O. (eds), Seasonality and Sedentism: Archaeological Perspectives from Old and New World Sites. Harvard, Peabody Museum, pp. 41-74. Müller J. 1994. Das Ostadriatische Frühneolithikum: Die Impresso-Kultur und die Neolithisierung des Adriaraumes. Berlin, Volker Spiess. Penzar B. & Penzar I. 1995. Velebit – klimatska prekretnica. Paklenički zbornik 1, pp. 11-15. Perica D. & Orešić D. 1995. Klimatska obilježja Južnog Velebita. Paklenički zbornik 1, pp. 17-24. Poljak Ž. 1969. Velebit. Zagreb, Planinarski savez Hrvatske. Rogić P. 1958. Velebitska primorska padina. Zagreb, Sveučilište u Zagrebu. Rukavina A. 1977. Voda na ličkom kršu. Veterinarska stanica 8(5-6), pp. 50-60. Shackleton J.C., van Andel T.H. & Runnels C.N. 1984. Coastal Paleogeography of the Central and Western Mediterranean during the Last 125,000 Years and Its Archaeological Implications. Journal of Field Archaeology 11, pp. 307-314. Težak-Gregl T. 1984. Novi prethistorijski nalaz iz Jurjeva. Senjski zbornik 10-11, pp. 1-6. Tutis V. & Ivančan-Picek B. 1998. Strong Bora Wind – a Risk Factor in Traffic. Promet 10(3), pp. 107-111. van Andel T.H. & Shackleton J.C. 1982. Late Paleolithic and Mesolithic Coastlines of Greece and the Aegean. Journal of Field Archaeology 9, pp. 445-454.

Zvelebil M. & Lillie M. 2000. Transition to Agriculture in Eastern Europe, in Price T.D. (ed.), Europe’s First Farmers. Cambridge, Cambridge University Press, pp. 57–92.

Acknowledgements Work presented in this paper was supported in part by the Ministry of Science, Education and Sports of the Republic of Croatia, project #196-962766-2740.

121

15 A hidden prehistoric landscape in the Region of Madrid (Spain): the significance of the mountains during the 3rd millennium BC Patricia Ríos Mendoza Dpto. de Prehistoria y Arqueología, Facultad de Filosofía y Letras, Universidad Autónoma de Madrid, Ciudad universitaria de Cantoblanco – [email protected] Abstract The Copper Age archaeological record (from the 3rd millennium BC until the beginning of the 2nd millennium BC) in the region of Madrid shows a clear difference between two distinct lithological zones: the mountains of the Central Range and the basin of the Tagus valley. Most of the known sites (settlements, burials and mines) are located in the sedimentary basins, revealing a high occupation density during the 3rd and 2nd millennia BC. Among them, two kinds of settlements can be distinguished. Large settlements encircled by ditched enclosures present clear signs of stable and long occupation, whereas many small sites with simple structures and ephemeral architecture do not. Some years ago, only two funerary sites and a group of rock art sites with schematic paintings were known and identified as belonging to the chalcolithic period in the mountain area. Recently, several megalithic structures, some rock art sites as well as some other places with a possible occupation in this period (mainly rock shelters and caves, but also two open air sites) have been discovered . In order to know if the chalcolithic settlement system and land use strategies can be understood on the basis of the current archaeological record, we need to study the circumstances surrounding the formation of the archaeological record.

1. Introduction The scientific archaeological record of the 3rd millennium BC in the central Iberian Peninsula, and in particular of the Region of Madrid, is still far from good. This situation stems mainly from the lack of fieldwork and analysis tackling a comprehensive study of the data (archaeological evidence and remains). Given this background, some PhD theses done in the last decade (like Garrido 2000 and Díaz del Río 2001) have contributed new theoretical approaches which have substantially enriched the state of research, but again had to face the problem of limited data which often had little or no context. We know that the Chalcolithic in the region of Madrid is defined by the occurrence of important cultural changes related to the development of farming. There is a consolidation of agriculture and cattle-breeding reinforced by the incipient development of metallurgy. The chalcolithic archaeological record for this region shows a lot of sites around sedimentary basins that are characterized by the accumulation of excavated structures, called pits. This density of sites and structures inside them constitutes a notable difference with respect to the previous (Neolithic) phase, which is characterized by few and small settlements without a sedentary occupation both in open air sites in basins and in caves or shelters in the mountains. For the Chalcolithic, very few sites are known in these mountains to date, with a very specific use of caves and a few high places (fig. 1). Consequently, the current settlement distribution appears to reflect a duality between the two significant geological land-

scapes of the Madrid region: mountains and sedimentary basins. While sites are rare in the former, in the latter they are plentiful. Can the chalcolithic settlement patterns be understood on the basis of the current archaeological record? Although the differences in the archaeological record mentioned above has been traditionally attributed to a lack of research in the mountains, some authors claim that the results of systematic and intensive survey provide strong enough indications to be sure that this duality reflects the true prehistoric occupation (Garrido 1994, p. 73; Díaz del Río 2001, p. 47; Rubio 2000). Similarly, although the few available palaeoenvironmental data also suggest this interpretation (Lopez 1997) — basins with highly an-

0 5 10 15 20 25 km

Figure 1 – Distribution of Copper Age sites in the Region of Madrid.

123

Hidden Landscapes

thropised landscape while the contrary occurs in the mountains — this should not be used to play down the significance of the occupation of the mountains. From our point of view, as happens in other prehistoric horizons, the fact that almost every site is located in the southern part of the region provides us with an unbalanced picture which cannot be corrected sufficiently due to the concurrence of several different limiting factors related to the archaeological search, the physical environment, the kind of investigation performed, etc. Moreover, there exists a group of direct and indirect data, some of it recently reviewed (Lucas et al. 2006; Jiménez 2000), which documents chalcolithic occupation of the mountains. From this we may infer that this area was neither a stopover nor as marginal as it might seem. Consequently, the mountains should not be ignored in the study of settlement patterns, not only as a potential area for the catchment of natural resources but also as a region where farming societies could be located. The same argument can be applied for the Neolithic and Bronze Age in this region, and for nearby mountainous areas like Guadalajara or Avila (Bueno et al. 2006; Fabián 2006) in the Chalcolithic. Therefore, we think that the mountains as a ‘marginal landscape’ in the study of the Chalcolithic in the Madrid region have been underestimated. Their role goes beyond being a source of raw materials, although this fact by itself is a strong argument for settling in similar areas. In this sense, the mountains of the Madrid region are a hidden landscape which has been virtually neglected in the archaeological record and in settlement patterns proposed to date. It is clear that the occupation models for each human group should be determined considering a set of interrelated factors without any exclusion (physical, economical, social, political, religious, military, historical or traditional factors, etc.). However, there are two conditioning factors: on one hand the issues of the theoretical interpretation adopted by the archeologists and on the other hand the partiality of the data due to conservation problems; all these factors lead to an overestimation of the peopling process (Bermúdez 2000).

2. The known archaeological record 2.1 In basins

We observe, within the densely populated areas documented for the basins (fig. 1), and taking into account the different relevance existing between these points, that this intense population can be mislead-

ing. The data coming from most of the excavated sites indicate that they are small settlements with few domestic structures of single occupation which can be interpreted as seasonal and transitory, probably used by small groups or even by a single family (Blasco et al. 1994). This means that, in many cases, a significant diachrony between the different sites should exist, even though the domestic equipment is relatively similar, and therefore, the density of settlement, although important, does not seem to be excessively high. Up to now, all or most of the points included both in the official list of sites, excavation reports and publications have been taken into account in the studies of chalcolithic settlement in the region of Madrid. Many of those locations have only been identified from the presence of materials that can be ascribed to the period. This circumstance makes that, for instance, isolated burials have been included equivalently to settlements in the distribution maps, producing again a distorted image of the settlement. Obviously, we consider that the existence of different types of sites should be taken into account in the study of settlements. In order to perform this analysis, we select only those sites where ‘habitation structures’ securely identified as belonging to the chalcolithic period have been reported. Furthermore, we neglect those locations that, although they could correspond to habitations, it is not possible to catalogue them firmly within this period on the basis of the available data (Ríos 2006). Thus, we study the known sites more properly and probably remain closer to the prehistoric situation with this type of analysis. We observe in the resulting distribution map (fig. 2) the existence of at least two kinds of habitation sites, two types of mines and some burial sites: • Large settlements encircled by ditched enclosures that present clear signs of stable and long occupation. These sites, still scarce in the archaeological record, are situated on medium and high river banks, at some distance from the main river courses (Díaz del Río 2003; Ríos 2006). • Large numbers of small sites with simple structures and ephemeral architecture. These are often situated on low river banks, close to the river beds. Although their domestic assemblages are quite similar, probably they were not all occupied at the same time. These habitations can represent seasonal or temporary occupation. • Burials. These refer to the presence of burials which are not associated with settlements —

124

Ríos Mendoza, A hidden prehistoric landscape in the Region of Madrid (Spain)…

mostly Bell Beaker single burials, as well as a Bell Beaker necropolis located in Ciempozuelos (Cuesta de la Reina) and a collective cave burial found in Tielmes (Juan Barbero’s cave). • Mines. Two types of mineral exploitation (salt and flint) are known. Among salt mines, the concentration of springs in the surroundings of Aranjuez stands out, where the exploitation of salt is

established at least from the Chalcolithic period (Bell Beaker) onwards in the ‘Salinas de Espartinas’ and in the area of Ciempozuelos (Puche et al. 2001; Valiente et al. 2002). The salt would have been of crucial importance both for the cattle and for food conservation. Other known locations with salt outcrops are spread around the sedimentary basins in the region of Madrid

Figure 2 – Distribution of the various site types (classification is definite in the basins, probable in the mountains).

125

Hidden Landscapes

(Blasco et al. 1994). Flint outcrops are located around the basins, and derive from silica levels in claystones and rolled flint nodules from the river beds and bluffs. One of the outcroppings, where an intense activity during Recent Prehistory has been documented, is the flint mine of Casa Montero, located in a high river bluff on the right bank of the Jarama river. Casa Montero to date is one of the earliest Neolithic flint mines in Europe, and the earliest in Iberia. It has more than 2500 shafts, whose characteristics suggest an Early Neolithic chronology (c. 5400-5000 cal BC) confirmed by radiocarbon dates from two shafts (Consuegra et al. 2004; Díaz del Río et al. 2006). Although with the current data concerning the excavated structures it is not possible to determine if there was a chalcolithic occupation here, the evidence provided by the lithic assemblages from chalcolithic sites close to this mine points to the exploitation and use of the flint extracted from this area (Blasco et al. 2007). • In summary, the big settlements would be playing a fundamental role in the cohesion of the territory, forming the first large and stable concentrations of settlement around the middle of the 3rd millennium BC which corresponds to sites like Fuente de la Mora or Camino de las Yeseras. These extensive sites, some of them including a Bell Beaker level in their stratigraphy, would be contemporaneous with a number of smaller and probably temporary ones, located preferentially on low terraces. These new data form a break with the traditional interpretation of this period and also with the characteristic settlement. Until now, the most widespread hypothesis was that during the 3rd and 2nd millennia BC the sedimentary environment of Madrid was occupied by small human groups, even by nuclear families, with an economy based primarily on cattle farming and complemented with agriculture, collecting and hunting. The documentation concerning the first stable settlements does not show this kind of economy and does not support the idea that these areas were occupied by reduced groups. The data extracted from these recently excavated sites point to the relevance of a more stable domestic stockbreeding (bovines and pigs) and also to the noticeable presence of ovicaprines. Furthermore, the results of palaeobotanic studies show a rather anthropized landscape although they do not point to an aggressive or destructive type of farming. Hence, the evidence does not support the movements of the groups that would determine the distribution of

the small known settlements. Contrary to this rough interpretation, all of these new data confirm a longterm permanence of the groups in the territory and suggest the existence of more or less long-distance expeditions — in which probably not all members of the group were involved — the goal of which were the exchange or procurement of raw material and, in general, the knowledge of the territory (further away from their own settlements). Concerning the exploitation of natural resources, the evidence found in the analyzed sites in basins indicates that the main guideline is the proximity to water courses, usually fluvial basins, which constitute also the most important routes and natural paths. Control and exploitation of the river bank seems to be rather usual in all of the locations studied, probably because of the potential richness of the river basin that makes possible the development of farming activities. Most of the sites are located very close (within 3 km) to the best quality flint outcrops (Ríos 2006). There are also abundant flint nodules rolling down from every bluff, as well as a large number of locations with salt that would be used both for farming and food conservation. In addition, a large quantity of clay is found in the sedimentary basins. The lithic objects of the different sites studied in this area show that clay is certainly abundant there, not only in more or less deep levels but also in numerous outcrops (fig. 3). However, there exist a number of mineral raw materials that are naturally absent in the basin environment but which do appear normally like finished objects in the sites. Precisely these resources are found in the mountains. 2.2 In mountains

The mountains until a few years ago were contemplated only as a natural route and a catchment area for certain materials that were not available in the basins. Only two funerary sites were known: a collective burial in a cave (El Rebollosillo) and a megalithic tomb (dolmen of Entretérminos). The other known sites include some rock paintings and a small number of caves or shelters with materials on the surface, both ascribed to chalcolithic period, which were interpreted as traces of specific trips to the mountains. In recent years more data has been found and compiled. Although not yet conclusive, they reveal a higher presence of chalcolithic groups in this territory and increase significantly the number of known sites. However, the main difference between the basins and mountains is still the lack of excavations in the latter region.

126

Ríos Mendoza, A hidden prehistoric landscape in the Region of Madrid (Spain)…

Prehistoric groups found some important rocks and minerals in the mountains: plutonic rocks like granite and diorite for making millstones, metamorphic rocks like gneiss, schist, graphite, amphibolite or fibrolite for making axes and other polished tools, and metallic minerals like copper and tin to produce metal objects (fig. 3). It was precisely with the search for exploitation areas that the mountains started to be considered as a region of interest for chalcolithic studies (Blasco & Rovira 1992; Blasco et al. 1994). We observe in the resulting distribution map (fig. 2) the existence of at least five different kinds of sites documented in mountains up to now: megalithic monuments, burial caves, open air settlements, caves and shelters with domestic materials on the surface, and rock art sites. • Megalithic sites. Until recently there was only one example of this kind of site in Madrid — the Dolmen of Entretérminos (Collado Villalaba) which is a megalithic burial, excavated a long time ago, that yielded Bell Beaker vessels among its findings. Recent work (Jimenez 2001, p. 100) has not been able to determine whether these Bell Beaker materials are related to the foundation of the monument or to its reuse. Another corridor tomb with circular chamber, El Rincón, was found and partly excavated in El Escorial in 1997 (Jimenez 2000, p. 101). It, too, was found to be partially destroyed and we do not have any details about the materials collected from the survey. In recent years, the presence of other megalithic monuments such as the mounds of Zorreras (El Escorial) and Cerca de las Hachas and Tomillar (Alpedrete) has been reported, as well as standing stones at El Cañal near Entretérminos (Jimenez 2000, p. 100). Garcia Gilabert (1996) documented the presence of sporadic findings of handmade pottery and flint in the same places close to the dolmen Entretérminos, and claims that these materials could come from similar megalithic burials. Also in 1997 Jimenez (2000, p. 101) documented a possible cromlech on a hill in Mataelpino, overlooking the surroundings and the pathways. • Open-air settlements. Up to now, there are only two known sites of this kind. The first one, El Dehesón (El Escorial), is a small settlement found on a surface with pits, and situated in a plain holm oak forest close to a stream and a huge granite stone. Pottery and some lithic objects were found on this site which has been interpreted as a seasonal settlement. The second one, El Jaralón

Figure 3 – Environmental resources within the study area in use during Recent Prehistory (from Blasco et al. 1994, fig. 7).

(Collado Mediano), is a settlement fortified with a double stone wall located at the foot of a hill in the Guadarrama mountains. Chalcolithic materials were found on the surface in this site. On the top of the hill there is a livestock path related to the Cañada Real Segoviana. In addition, there is an ancient copper mine nearby. • Burial caves. The collective burial of El Rebollosillo (Torrelaguna) is the only example of this kind and it is located in a small natural cave close to a tributary of the Jarama river. Human bones with no anatomical connections were found here, in association with ceramic vessels and stone beds (some made of variscite), both ascribed to the pre-Beaker Chalcolithic phase. This kind of site is not uniquely found in the mountains because similar burials have also been reported in the basins. In particular, there are two such caves situated in Tielmes (Juan Barbero) and Estremera (Pedro Fernández). • Rock art sites. There are some 18 caves and rock shelters recorded which present manifestations of schematic parietal post-palaeolithic art. In at least six of these, materials ascribed to the chalcolithic period have been collected (Bell Beaker and non Bell Beaker) (Lucas et al. 2006). Unfor-

127

Hidden Landscapes

tunately it is still difficult to link, both culturally and chronologically, the artistic manifestations with the archaeological sites, mainly due to the lack of excavations in this area. These sites are located in both geological units which constitute the mountains — the granitic and more occidental region and the calcareous and more northern one – but most sites are found in the latter. • Caves and shelters. There are a number of known caves, namely Cueva del Sifón, Cueva de la Cascada, Cueva de la Flecha among others, in which materials ascribed to the chalcolithic period have been collected (Lucas et al. 2006). Most of the sites located in the mountains have not been excavated and only a few of them have been the subject of research other than the usual and superficial analysis derived from their own discovery. However, there exist a number of works in the bibliography that consider these sites within the framework of Recent Prehistory, showing a set of hypotheses about their dating and their role in the territory (Jiménez 2000; Rubio 2006). Concerning the megalithic monuments, all the examples given here are from the granitic region, which is crossed by the upper course of the Guadarrama river and by other small streams. Due to their situation and orientation, they have been interpreted as symbols of control of the territory and the seasonal cycles (Jiménez 2000, p. 101). They are also understood as a part of a more complex megalithic territory related to pathways used by shepherds in the Middle Ages and surely also used in prehistoric periods (Jiménez 2000, p. 102). Although the known data do not allow us to establish a clear chronology, some authors ascribe these sites to the Late Neolithic or Chalcolithic periods (Jiménez 2000, p. 103) based on the existence of habitation sites and chalcolithic burials in their surroundings. Another possibility pointed out by other authors is that these monuments were built in the Neolithic and were reused later during different chalcolithic phases, as has been shown to occur in similar areas in the southern and northern Mesetas (Bueno et al. 2006; Rubio 2006). However, the chronology which corresponds to the cave burial in El Rebollosillo is clearly established to be Chalcolithic. With respect to the settlements, the existence of El Dehesón (small settlement with pits) and El Jaralón (fortified settlement) would indicate a similar occupation to the one reported for the basins. The deposits located in certain caves and shelters – whether or not they are associated with schematic art or the

occurrence of chalcolithic materials – probably also fit with those at the small habitation sites found in the basins. The fortified site of El Jaralón is exceptional in this region, in the same manner as others found in the Meseta which belong to this period (El Pedroso in Zamora and Alto del Quemado in Ávila).

3. Factors which determine current knowledge of the Chalcolithic in the region of Madrid From our point of view, there are two factors that have essentially determined in the past and still determine now the difference in the knowledge of the Chalcolithic existing between the mountains and the basins: first, the archaeological field work developed in each case, and second the specific physical characteristics of each region. The urban and industrial development of Madrid — situated in a sedimentary environment — has had a major influence on the knowledge of the archaeological record. Both the continual urban expansion and the exploitation of sand quarries from the beginning of the 20th century around the city have resulted in the documentation of a large number of sites in the sedimentary basins. In contrast, such developments rarely occur in the mountains due to their little urbanization and, therefore, have had little effect on archaeological activities. It is important to note that the systematic and general prospections performed in the whole region are not effective enough in localizing certain sites that do not show perceptible evidence on the surface. Hence, we must assume that these surveys by themselves are insufficient to judge the usefulness of archaeological research in the study of a particular period. If we draw global conclusions based not only on what we know — little in the case of the mountains — but also on what we do not know (neglecting the possibility of applying a different research methodology in this region), we will end up constructing models to describe the settlement and the occupation of the territory that will not reproduce past reality. The excavation of large open areas performed in the last years provides a paradigmatic example of the aspects discussed above, and has allowed the assessment of the entire chalcolithic sites. The origin of this kind of excavation lies in the development of town planning and linear infrastructure. The first documented cases of Copper Age ditched enclosures in the region are one of the most relevant results coming out of these works, because these kinds of structures had not previously been taken into ac-

128

Ríos Mendoza, A hidden prehistoric landscape in the Region of Madrid (Spain)…

count by any type of survey; nor had excavations allowed a definite identification. In spite of the performance of general prospections, without the excavation of big areas that results from large-scale infrastructural developments it is quite probable that most of the archaeological heritage in the mountains remains unnoticed. Furthermore, as has been already pointed out earlier, excavations derived from Cultural Resource Management and academic investigation projects are scarce in the mountains. The singular physical characteristics of the mountain environment have to be added to the great disparity in the origin of the data; this is indispensable to localize the sites and also very relevant to the analysis of the conditions and natural resources implied in the prehistoric occupation. However, other than in relation to the exploitation of resources, those landscape characteristics have been sometimes ignored. They have rarely been used to determine the type of settlement and, perhaps more relevant, to find evidence for occupation. In addition to the small number of excavations, prospections in the mountains are much more constrained by the natural environment. This region shows high elevations with marked slopes, dense vegetation in some areas, surfaces which are currently suffering erosion, etc. In contrast, the basins are characterized by almost flat relief and open vegetation. Although certain zones near the river beds have been eroded, this fact is taken into account when interpreting the post-depositional movement of materials in Palaeolithic sites. The specific characteristics of the mountains make more difficult not only the discovery of archaeological remains, but also their conservation. Hence, neither should the same kind of remains be expected in the mountains and in the basins nor should the same research strategy be utilized because this in itself could bias the outcome of the investigations. For instance, erosion causes the movement of materials in the mountains, and this should be analyzed in order to determine the location of the sites themselves (Planas 2000). The large number of sites excavated in the basins in recent years, and the necessity of providing proper interpretations for them, have led to a lack of research of Chalcolithic in the mountains. Unavoidably, the processes of Cultural Resources Management and (indirectly) the current development policy for the Region, produce a better knowledge of the basins. However, it could be argued that the process of archaeological discovery in both zones by

chance leads to an accurate image of past reality, i.e., that the basins really were densely occupied while the mountains were not. In this article we want to reflect — from the known data and these determining factors — on a number of direct and indirect indications for occupation in the mountains. Although the broad river valleys were undoubtedly found more attractive for the development of an economy based on farming, and therefore supported a higher density of settlements, we know that the mountains could also have been occupied.

4. Evidence for the occupation of the mountains: final considerations Taking into account all the circumstances which are involved in our current knowledge of the research, we consider that there is a number of complex factors that should be contemplated in the arguments used in the interpretation of the occupation of the mountains in the 3rd millennium BC — and also probably in Late Prehistory. We have the following arguments to support settling in the mountains: • The natural conditions are suitable for a more or less permanent settlement. Valleys appropriate for farming, stockbreeding and forest exploitation are found in this region. • The occupation of caves in the region very probably follow the pattern of Neolithic occupation continuing into the Bronze Age, observed in the neighbouring region of Guadalajara. • Many of the caves and shelters are situated opposite fertile plains in which settlements can be found, similar to what is observed in Guadalajara (Bueno et al. 2006) and corroborated in the nearby Valle del Amblés in Avila (Fabián 2006). • The possible open-air sites documented in the area of El Escorial could certainly support a stable settlement, as is also observed in the nearby Valle del Amblés (Fabián 2006) which is located in the mountains and which presents similar characteristics to many mountain valleys in the Region of Madrid (Lozoya, Guadarrama). • The existence of chalcolithic burials is not by itself proof of occupation, but the presence of assorted funerary manifestations ascribed to the Neolithic and the Bronze Age signals the continuous presence of humans throughout Late Prehistory. • This presence could be connected to the control of certain raw materials whose utilization has been proven in sites found in the basins. Furthermore, it may be possible to discover settlements

129

Hidden Landscapes

in the surroundings of the mines exploited during this period, although we do not yet have any evidence for this type of sites. • The density of sites would surely be lower than the one documented for the basins, and their distribution would probably be dependent on the characteristics of the terrain. The most suitable places for settling would be the edges of the mountains, the foothills and the high intermontane valleys. These arguments lead us to believe that there is a more complex occupation pattern to be found in the mountains than is traditionally considered. Even if it is not possible to determine exactly what kind of pattern this is, some general assumptions may be proposed. Possibly, a settlement system similar to the one documented for the basins, with many small, short-term habitations and a few bigger settlements with stable occupations was present during the 3rd millennium BC. In that case it would be possible to establish links between mountains and basins through exchanges of raw materials. Alternatively, if large and stable settlements did not exist in the mountains, it is possible to envisage a system of small settlements geared more to livestock farming than to agriculture. This would explain the close relationship between the few known sites and the pathways. It is also probable that the mountains would have been frequented or visited by chalcolithic groups from the basins and from the other side of the Central Range for raw materials supply or the seasonal use of pastures. The mountains would then be a contact area for different seasonally mobile groups. In any case, it seems clear that there was some contact between these two environmental zones during the 3rd millennium BC, although we do not know if there existed any dependency relationship between the settlements in the mountains and the basins, or even in other neighbouring regions, because this relationship could be based on exchange among groups of similar development and cultural complexity or, on the contrary, on nothing more than frequent visits between them. Although we cannot draw any firm conclusions on the basis of this study, we do believe there was important settlement in the mountains, as also occurred in other nearby mountain areas with similar features, especially in the presence of significant exploitable resources. At this point, there is not enough chronological precision to demonstrate any evolution in the use of the mountains, but (on the basis of the data

from the sites located in basins) it is quite probable that an intensification in the use of the mountain resources occurred since the Bell Beaker phase. This hypothesis is based on the increasing use of granites and metallurgical evidences, besides other materials with distant origin (Blasco et al. 2007). We do not know whether the origin of this intensification lies in exchange among people exploiting the mountain resources or in a more intensive use of the resources by the groups from the basins. In order to clarify these hypotheses, we need data from excavations in the mountains. It would be important to develop wider research projects in mountains considering the possibility of studying new sites, analysing palaeobotanical data, materials, etc. in order to obtain good information about the relation between occupation and exploitation of the territory during this period.

References Baena Preysler J., Blasco Bosqued C. & Recuero V. 1995. The spatial analysis of bell beaker sites in the Madrid region of Spain, in Lock G. & Stancic Z. (eds) Archaeology and Geographical Information Systems: a European perspective. London: Taylor and Francis, pp. 101-106. Bermúdez J. 2000. La aplicación de los sistemas de información geográfica a la arqueología. Doctoral thesis at Universidad Autónoma de Madrid (unpublished). Blasco C., Delibes G., Baena J., Liesau C. & Ríos P. 2007. El poblado calcolítico de Camino de las Yeseras (San Fernando de Henares, Madrid): un escenario favorable para el estudio de la incidencia campaniforme en el interior peninsular, Trabajos de Prehistoria 64(1), pp. 151-163. Blasco Bosqued C., Baena J. & Recuero V. 1994. Los asentamientos, El horizonte Campaniforme de la Región de Madrid en el Centenario de Ciempozuelos, Universidad Autónoma de Madrid, pp. 47-74. Blasco C. & Rovira S. 1992-93. La metalurgia del cobre y del bronce en la región de Madrid, Tabona: Revista de prehistoria y de arqueología VII(2), pp. 397-416. Bueno P., Barroso R. & Balbín R. de 2006. Agricultores y metalúrgicos en la Meseta Sur, in Pereira J. (ed.), Prehistoria y Protohistoria de la meseta Sur (Castilla–La Mancha). Ciudad Real, pp. 57-94. Consuegra S., Gallego Mª.M. & Castañeda N. 2004. Minería neolítica de sílex de Casa Montero (Vicálvaro, Madrid). Trabajos de Prehistoria 61(2), pp. 127-140. Díaz del Río P. 2001. La formación del paisaje agrario: Madrid en el III y II milenios BC, Arqueología, Paleontología y Etnografía 9. Díaz del Río P. 2003. Recintos de fosos del III milenio AC en la Meseta peninsular. Trabajos de Prehistoria 60(2), pp. 61-78. Diaz del Rio P., Consuegra S., Castañeda N., Capote M., Criado C., Bustillo M.A. & Pérez-Jiménez J.L. 2006. The earliest flint mine in Iberia. Antiquity Project Gallery, March 2006, www.antiquity.ac.uk/projgall/diazdelrio307/index.html.

130

Ríos Mendoza, A hidden prehistoric landscape in the Region of Madrid (Spain)…

García Gelabert Mª. P. 1996. Carta arqueológica del término de Alpedrete. Reunión de Arqueología Madrileña (25-26 enero), Madrid, pp. 269-272. Garrido Pena R. 1994. El fenómeno campaniforme en la región de Madrid: actualiación de la evidencia empírica y nuevas propuestas teóricas. Estudios de Prehistoria y Arqueología madrileñas 9, pp. 67-90. Garrido Pena R. 2000. El Campaniforme en la Meseta Central de la Península Ibérica (c. 2500-2000 a.C.). Bar Int. Series 892. Jiménez J. 2000. Megalithic tombs and chalcolithic settlement in the Guadarrama mountains: following ancient roads, marking out territory. Journal of Iberian Archaeology 2, pp. 99-109. Fabián J.F. 2006. El IV y el II milenio AC en el Valle del Amblés (Ávila). Monografías Arqueología en Castilla y León, 5. Junta de Castilla y León. López P. 1997. El paisaje vegetal de la Comunidad de Madrid durante el Holoceno Final, Arqueología, Paleontología y Etnografía 5. Lucas Mª.R, Cardito L.M & Gómez J. 2006. Inventario de yacimientos, Arqueología, Paleontología y Etnografía, 11. Arte Rupestre en la Comunidad de Madrid, pp. 151-262. Planas M. 2000. Aplicación de un Sistema de Información Geográfica a la prospección del yacimiento de Cancho Gordo (La Cabrera, Madrid), Cuadernos de Prehistoria y Arqueología de la Universidad Autónoma de Madrid (CuPAUAM) 26, pp. 25-45. Puche O., Ayarzagüena M. & Mazadiego L. 2001. Salinas históricas de la Comunidad de Madrid: caso particular de Espartinas. Huelva. Congreso Geológico y Minero. Universidad de Huelva, Huelva. Ríos P. 2006. Espacio doméstico y colectivo en el calcolítico madrileño: el yacimiento de Camino de las Yeseras (San Fernando de Henares, Madrid). Análisis espacial apoyado en los SIG. Trabajo de Investigación de Doctorado (unpublished). Universidad Autónoma de Madrid. Rubio I. 2000. Las primeras sociedades agrícolas en Madrid. Neolítico y Calcolítico Precampaniforme, in Ruano E. (ed.), La Arqueología madrileña en el final del Siglo XX: desde la Prehistoria hasta el año 2000, BAEAA 39-40, pp. 105-126. Rubio I. 2006. Pastores de ovejas y cultivadores de trigo: El color rojo domestica la naturaleza. Mirando las paredes. Contextualización cultural de los abrigos con Arte Postpaleolítico (Pintura Esquemática). Discusión cronológica, in Lucas M.L., Cardito L. & Gómez J. (eds), Arqueología, Paleontología y Etnografía, 11. Arte Rupestre en la Comunidad de Madrid, Madrid, pp. 263-309. Valiente Cànovas S., Ayarzagüena Sanz M., Moncò Garcia C. & Carvajal Garcia D. 2002. Excavaciones arqueológica en las salinas de Espartinas (Ciempozuelos) y prospecciones en su entorno, Archaia 2, pp. 33-45.

131

16 Mountains and shore: sites and communication routes of Copper Age western Liguria Neva Chiarenza Dipartimento di Scienze Archeologiche, Università degli Studi di Pisa, via S. Maria 53, Pisa, Italy – [email protected] Abstract This article discusses the character and location of Copper Age sites in western Liguria province, especially with regard to regional communication routes. Western Liguria is characterized by high valleys near a rocky shore; these valleys are continuous with the near alpine structures: the resulting environment is a mixed one that is influenced by the short distance from the sea and that enables people to make use of pastures and alpine passes. Archaeological sites, mainly in caves, are located along the principal valleys and were used as burial sites or as supporting areas for movements of different lengths and durations; it has been observed that higheraltitude sites are concentrated in the western area, while the Finale Ligure area includes lower-lying sites: this is in part due to the geomorphological character of the territory. Communication routes were located along the highest parts of the valleys: a large share of human frequentations is documented by sites located en route to the passes. Such passes allowed communications with neighbouring valleys and with valleys located in the Piedmont, such as those of the Bormida and Tanaro rivers. But it is also possible to reach the Nervia and Roia basins, while to the west it is possible to reach the Valle delle Meraviglie and the Valle del Monte Bego.

1. Introduction Connections between the Western Ligurian valleys and their archaeological sites are quite difficult be-

cause of the impenetrable nature of these valleys, continuing up until the Alps. All of these places share a steep aspect and the closeness of alpine watershed

Figure 1 – Western Liguria with principal rivers and passes; the black dots indicate Copper Age sites.

133

Hidden Landscapes

to the sea, but the eastern zone (region near Finale) has more open spaces and lower elevations than the valleys near the French border. The site distribution seem to identify some paths toward the Piedmont and the French valleys. The existence of these contacts is supported by the nature of the finds: ogival, rhombic and willow leaf shaped arrow points, pearls à ailettes and drop or cross shaped (sites of Valle Argentina). Some materials like steatite, and finds of pots with a particular shape and decoration, however show contacts between Western and Eastern Liguria as well as with Peninsular Italy (in particular the Vecchiano facies) (Cocchi Genick & Grifoni Cremonesi 1989).

2. Site locations Except for occasional open-air finds, the Copper Age sites of this region mainly consist of cave or rock shelter deposits. The remains are mainly those of slaughtering sites and transit stations for sheep farming and hunting. The Finale region, between Borgio Verezzi and Finalpia, is characteristic because of its steep calcareous headlands on the sea, surrounded by a quite open area. The Copper Age sites occur from the coast toward the inland and from west to east. Sites start in fact from the coast near Verezzi, with Arma delle Arene Candide (arma is a ligurian term for ‘cave’; Bernabò Brea 1946, 1956, Maggi & 7

6

5

4

3

2

1

Aquila_Sciusa

Pora_Aquila

Verezzi

Maremola

Varatella

Pennavaira

Impero

Argentina

Armea

Nervia

Bevera

Roia

0

Figure 2 – Number of sites in each of the western ligurian valleys.

Starnini 1997), Armorari Cave (Richard 1932, 1939, Del Lucchese & Odetti 1998), Antenna Cave (Lamberti 1971), all between 100 and 300 m asl. Further inland up to Perti, on the left bank of the Aquila river, a little north from its confluence with the Pora river, there are Anime Cave at 350 m (Giuggiola et al. 1966), Pollera Cave at 284 m (Odetti 1972, Tiné 1972, Tiné & Traverso 1991-92), Sanguineto Cave at 175 m (Fusco & Soffredi 1963) and Morto Cave at 267 m (Del Lucchese & Vignolo 1993). At the same altitude, but between Aquila and Sciusa, there are the caves of Vacché (Odetti 1987a, 2003, Del Lucchese & Odetti 1988) and Fascette I (Maggi & Pastorino 1984) at 280 m. At the confluence between the Sciusa stream and Rio Cornei there is the Pipistrelli Cave at 280 m (Almagro 1955, Almagro & Ripoll 1957, Almagro et al. 1957a, 1957b). All of these inland sites are situated at the same distance from the sea and not very far from each other. The territory between the area of Albenga and the area of Finale is characterized by ridge-lines perpendicular to the coast, separating the deep and steepsided valleys dug by the Maremola, Nimbalto and Varatello streams. The first documented Copper Age site in the Maremola valley, starting from the mouth of the river, is a small cave, now destroyed, not very far from the coast near Rocca delle Fene (180 m asl; Barocelli 1933, 1979, Odetti 1996). A little further north than the confluence between the Giusténice and Maremola streams, on Monte Bruxacrava, there is Ponte di Vara Cave (100 m), and to the north of the confluence between the Lavezzino and Maremola streams there is the burial site of Tana (‘Lair’ or ‘Den’) dell’Armusso (300 m; Odetti 1980, 1984, 1987-88). Sites in the Varatella valley cluster in the Toirano area. Most of them were in fact found in caves in the steep calcareous side of the Rio Vero valley (left affluent of the Varatella): S. Lucia Superiore Cave at 215 m (Tozzi 1962, 1963), Basura Cave at 186 m (Chiappella 1955, Maggi & Starnini 1984), and Ulivo Cave at 250 m (Muñoz Amilibia 1958, Maggi et al. 1998). The only isolated site is Gera Cave, located at 590 m at the foot of Monte Ravinet, at the top of Vallone di Servaira, which is one of the highest confluents of the Varatella (Lamberti 1984, Del Lucchese et al. 1994). Going west, behind the Albenga plain, the Maritime Alps quickly reach elevations beyond 2000 m, at which altitude there are some pastures. Among the valleys that follow this sort of natural barrier, the steepest and narrowest is carved by the Rio Penna-

134

Chiarenza, Mountains and shore: sites and communication routes of Copper Age western Liguria

vaira, merging in the Neva stream. The first Copper Age site we find when entering the Pennavaira valley and walking away from the Neva, is Arma del Nasino, located in a flat area at 350 m in the middle reaches of the stream (Leale Anfossi 1967, 1974). Tana del Barletta at 1000 m shows signs of high altitude activities (mostly pasture but also hunting and trading; Del Lucchese et al. 1987, Barker et al. 1990, Maggi & Nisbet 1991), while the Camere Cave, located at the top of the valley at an elevation of 900 m, is a burial site (Leale Anfossi 1957, 1962). The Pertusello Cave (570 m) is instead located on the right bank and not far from Camere (Leale Anfossi 1958-61a, 1958-61b, 1962, Maggi 1998). On the north west side of the Pertusello, at the confluence between Rio Ferraria and Pennavaira, there is Arma du Cuppâ, at 810 m (Leale Anfossi 1958-61c). Closer to the French border, the Argentina valley is characterized by a more open landscape near the coast and by narrower and steeper walls in its middle and upper part. The first site we encounter going up toward the mountain, a chance find of an arrow point, is at a few kilometres from the sea at Entrà (630 m; Ricci 1991, Del Lucchese & Ricci 1998). On the left side of the valley, the first real site is found at 1050 m: Tana del Bertrand is a burial site not far from the top of Monte Faudo (1149 m; Crowfoot 1926, Barocelli 1933, Ricci 1988a). Above Triora, in a particularly steep area near Monte Trono, where the stream passes through a gorge, there are the Loreto Hideout (Ricci 1964, Lorenzelli & Ricci 1988), Arma della Vigna (Ricci 1988, 1998) and Tana della Volpe sites (Lorenzelli & Ricci 1980, 1988), all between 600 and 700 m. Near the village of Realdo, almost at the top of the valley, there is Arma della Grà di Marmo (985 m), a funerary site located on a calcareous protrusion over the stream (Ricci & Lanteri Motin 1963, 1964, 1965). In the inland of Sanremo, finds are mainly located on the surface and are clustered in the Tana della Ratapena site, on the Monte Caggio, at 1000 m. In the surroundings, ruins of what looks like a dolmen were found; a similar find was made on the nearby Monte Carparo (Del Lucchese & Ricci 1998). The Nervia valley, located on the southern fringes of Monte Toraggio and Monte Pietravecchia, is flanked by the high pre-alpine ridges, and its lower slopes are generally steep-sided. There are two sites, both located near the northern boundary of the watershed: Tuvetto Cave (565 m) near Pigna (Barocelli 1979, Calandri et al. 1979), and the Grotta Grande sotto la Cava della Diga (1250 m) near the artificial lake of Tenarda (Maifredi 1962).

a 9

8 7 6 5 4 3 2 1 0

b 1300 1200

Nervia

1100 1000 900 800 700 600

Argentina Varatella

500

Aquila

400

Sciusa

Pennavaira 300 200 100

Maremola Verezzi

0

Figure 3a, b – General altitude distribution of western ligurian sites (a); altimetric distribution of western ligurian sites by watershed (b).

3. Site distribution These sites are not homogeneously distributed along the valleys (see fig. 2), which might be caused either by the original site location choice or by biases in the research progress, often impeded by the steepness of the terrain and agricultural modifications. In the western part of the region, the certain Copper Age sites are mainly located in the Argentina valley, while sporadic older sites are mainly found along

135

Hidden Landscapes

the French-Italian border. There are no finds along its confluent streams the Roia and Bevera, probably because no search has been made. In the Savona region, some minor valleys (Varatella, Maremola and, above all, the faraway and narrow Pennavaira) yielded a lot of finds. In the Finale region, there are Copper Age finds mainly on the low crests separating the Fine (Verezzi), Pora, Aquila and Sciusa valleys. A particular case is Tana Rapatena, on the Monte Caggio in the Imperia countryside, which is situated on the main watershed in the Sanremo Region. When we look at the site altitudes, these seem more to be a consequence of the nature of the territory than of a real choice (see infra for the highest altitudes). Most sites are located (fig. 3a) between 200 and 300 m, and then between 100 and 200 m; the remaining sites are distributed quite homogeneously between 300 and 400 m and between 500 and 700 m, and sometimes beyond 800 m. Of particular interest are the site altitudes over 900 m, because these are usually located on passes. Until now, only one such site has been found at an altitude below 100 m: the tomb of Castellari, with Bell Beaker pots, which is still being studied by its discoverers (Del Lucchese & Odetti 2001). We have no sites in saddle locations between 400 and 500 m or between 700 and 800 m. Looking at the data distribution for each separate valley (fig. 3b), it appears that the highest altitude sites are present in the Argentina valley, while most finds of the Finale region are located at medium altitude (300-350 m). The situation in the Pennavaira valley looks more varied: Arma del Nasino is located at only 350 m, while Barletta reaches 1000 m and Camere 900 m.

1100

C.le di Melogno

1000 900

C.la di S.Giacomo

800 700 600 500 400

Anime Antenna

300 200

Vacché

Armorari

Fascette

Pipistrelli

Morto Matta

A. Candide

100 0

Pollera

a

mare

1200 1100 1000

Giogo Giustenice

Colle Melogno

900 800 700 600 500 400 300 200

Fene

Armusso

100

mare

Vara

0

1000

Bric Pratello

Colle Preglia

900 800

b

crinale

Giogo Toirano 700 600

Gera

500 400 300

S. Lucia Sup.

200

Ulivo

100

Basura

c

0

1600 Caranche 1400 1200

Prale

1000

Camere

Barletta

800 600

Pertusello

4. Paths

400

Nasino

200

d

0

1800 Colle Ardente 1600

Passo Guardia

1400 M. Faudo 1200

Bertrand Passo Merlo

1000

Pratipiani

Grà di Marmo 800 600

Volpe

Vigna

Loreto

Entrà

400 200 0

mare

e

Figure 4a-e – Altimetric distribution of sites and passes. Finale region (a), Maremola valley (b), Varatella valley (c), Pennavaira valley (d), Argentina valley (e).

Communications through the study area depend on the altitude and the morphology of the valleys. Crests in the Finale countryside (fig. 4a) are actually passable so the watershed boundaries represent a connection rather than a barrier. From the sites in this region it is possible to reach Piano dei Corsi to the north, and from there via the Colla di Cravezza (958 m) and Colla di S. Giacomo (796 m) continue toward Bormida di Mallare. This latter crest can also be followed towards the east, where it leads to the valleys of the Savona region. The Colle di Melogno (1028 m) represents a real junction, as it is possible to reach Pian dei Corsi and the Finale region or the Maremola valley to the south, to reach the Giusténice (1183 m) and Toirano (801 m) yokes when following the direction of the Varatella valley, and to

136

Chiarenza, Mountains and shore: sites and communication routes of Copper Age western Liguria

arrive in the river basin of Bormida di Millesimo, thus opening a path toward the north. In view of their locations, the sites of Tana dell’Amusso (Maremola valley) and Gera Cave (Varatella valley) seem to benefit from this northerly path. Gera Cave especially is near to the crest, from where it is not far to the Toirano junction which allows easy access in other directions to nearby valleys (Maremola through Giusténice, Neva through Scravaion) or northern valleys (Bormida). Nimbalto valley is also reachable through Passo Preglia (889 m asl), north of Boissano, while Maremola valley is reachable by passing Bric Pratello (911 m), located southeast of Giogo di Giusténice (fig. 4b and c). It is also possible to arrive in the northern part of the region through the Passo di Prale, which connects the Pennavaira and Tanaro valleys (fig. 4d). Links toward the sea are also important in the network of communications, as witnessed by finds of seashells in most of the sites, as well as by the location of some sites near the sea. An important find in this regard has been made at Entrà (fig. 4e; Argentina valley, Imperia province). This is a quite open location which suggests movements toward the coast and toward the nearby Merlo pass (904 m) which leads to the Armea valley. From the Tana del Bertrand it is possible to reach the small valleys located in the west of the region, such as the valleys of the S. Lorenzo and Prino streams. A path along the northern crest follows the Carpasina stream (a left tributary of the Argentina) to the coast, until it reaches the pastures of Pratipiani and Colle d’Oggia. From here the valley of Giara di Rezzo (tributary of the Arroscia, which is itself a tributary of the Pennavaira) or the valley of Impero to the west can be reached, or one could even turn back toward the east to the Argentina valley. Going north, on the left bank of the river, Monte Croce Castagna (1344 m) in the region of Creppo can be reached; continuing north, it is then possible to reach the passes of Guardia (1461 m) and Garlenda (2021 m), which lead to the Tanaro basin. From the highest site in the Argentina valley, Arma della Grà di Marmo, Colle Ardente (1617 m) can be reached with a short trip, and from there it is easy to reach other valleys: going south one arrives in the Nervia and Roia basins, going west toward the Mercantour Park one reaches the Meraviglie valley and Monte Bego, and going north there is the river basin of Tanaro. However, all these routes are quite long and until now searches have not given any evidence that they were used.

Finally, the sites in the Nervia valley are also significantly located with respect to communication routes. Tuvetto Cave, being located near the Muratone valley, is right on the route leading to the Roia river basin. From the area of Tenarda it is possible to reach mountain passes to the east and west. Both sites are therefore located near routes leading to mountain passes which are still used today, and both sites are on the routes leading to the valleys of the Argentina and the Roia. The position of these sites is even more interesting since this area, in spite of its distance from the coast (35-40 km) and high elevations, has in fact a completely Mediterranean climate. In fact, while to the west the Monte Toraggio, Pietravecchia and Grai have steep slopes where snow usually remains until spring, on the contrary the southern side of Colle Langan is completely covered until 1200 m by heather, arbutus, pinaster, rock roses and, above all, holm-oaks. The finding of seashells in the site near the dam of Tenarda again proves movement toward and/or links with the coast. It still remains to be proven if the lower part of the valley (frequented in the Roman and Medieval periods) is in fact lacking prehistoric sites, or if this is just because no systematic research has yet been conducted.

5. Final remarks It appears from the foregoing sections that the characteristics of this rough and steep territory have not discouraged people from settling, although it did influence their choice of locations. Somehow it even seems that the peculiarities of the region, which might seem to make it inhospitable (e.g. the proximity of the alpine watershed to the coast, with only a small and narrow space left to settle), have instead made it suitable for this population, since in a small space there are a lot of routes, junctions and passes. The lower watersheds, perpendicular to the coast, often function as access routes to pastures or to higher watersheds, from where passes provide access to routes into Piedmont, the Roia region and the Mercantour region. Probably the sea itself was also used as an easy travelling route, permitting links between the eastern and western regions.

References Almagro M. 1955. Escavaciones de 1954 en la “Caverna dei Pipistrelli” (Finale Ligure). Rivista di Studi Liguri XXI, pp. 5-31. Almagro M. & Ripoll E. 1957. Campañas de excavaciones de 1954 y 55 en la “Caverna dei Pipistrelli”. IV Congressio Arqueologico Nacional. Zaragoza, Spagna, pp. 75-79.

137

Hidden Landscapes

Almagro M., Ripoll E. & Muñoz A.M. 1957a. Excavaciones en la caverna dei Pipistrelli. Cuadernos de Trabajos Esquela Española Historia y Arqueologia Roma IX, pp. 169-222. Almagro M., Ripoll E. & Muñoz A.M. 1957b. Gli scavi italo spagnoli nella caverna dei Pipistrelli (Finale Ligure, Italia). Rivista Ingauna Intemelia XII, p. 76. Barker G. et al. 1990. From hunting to herding in the Val Pennavaira (Liguria – northen Italy), in P. Biagi (ed.), The Neolithisation of the Alpine Region. Monografie di “Natura Bresciana” 13, pp. 99-121. Barocelli P. 1933. Nuove ricerche di preistoria nel territorio degli Ingauni. Collana storico-archeologica della Liguria occidentale II, fasc. 5, pp. 5-24. Barocelli P. 1979. Il pugnale litico di Pietra Ligure. Appunti sulle industrie litiche preistoriche della Liguria occidentale Italiana. Rivista di Studi Liguri, anno XXXX, pp. 5-24. Bernabò Brea L. 1946. Gli scavi nella Caverna delle Arene Candide, Gli strati con ceramiche, I. Bordighera: Istituto Internazionale di Studi Liguri. Bernabò Brea L. 1956. Gli scavi nella Caverna delle Arene Candide, Gli strati con ceramiche, II. Bordighera: Istituto Internazionale di Studi Liguri. Calandri G., Grippa C. & Ramella L. 1979. Lo sgarbu di Barraico in Val Nervia (Provincia di Imperia). Riviera dei Fiori, giugno, pp. 3-13. Chiappella G. 1955. Grotta della “Basua”a Toirano (Savona), Quaernaria II, pp. 282-284. Cocchi Genick D. & Grifoni Cremonesi R. 1989. L’età del Rame in Toscana. Viareggio. Crowfoot G.M. 1926. Note on Excavation in a Ligurian Cave – 1907-09. Man, pp. 83-88. Del Lucchese A. & Odetti G. 1998. Il Finalese, in A. Del Lucchese & R. Maggi (eds), Dal diaspro al Bronzo, L’Età del Rame e L’Età del Bronzo in Liguria: 26 secoli di storia fra 3600 e 1000 anni avanti Cristo. La Spezia: Luna Editore, pp. 90-94. Del Lucchese A. & Odetti G. 2001. I vasi campaniformi in Liguria, in F. Niccolis (ed.), Bell Beakers today. Pottery, people, culture, symbols in prehistoric Europe, Proceedings of the International Colloquium vol. II, Riva Del Garda 1998, pp. 625-627. Del Lucchese A. & Ricci M. 1998. Altri ritrovamenti della provincia di Imperia, in A. Del Lucchese & R. Maggi (eds), Dal diaspro al Bronzo, L’Età del Rame e L’Età del Bronzo in Liguria: 26 secoli di storia fra 3600 e 1000 anni avanti Cristo. La Spezia: Luna Editore, pp. 68-69. Del Lucchese A., Franceschi E. & Rossi G. 1994. Analisi archeometriche di alcuni bronzi preistorici e nuove conoscenze sulla prima metallurgia della Liguria. Bollettino dei Civici Musei Genovesi, anno XVI, n. 47-4849, pp. 15-23. Del Lucchese A., Maggi R. & Nisbet R. 1987. La Tana del Barletta. Archeologia in Liguria III.1, Scavi e Scoperte 19821986, Preistoria e Protostoria, Genova: Tormena Editore, pp. 165-168. Fusco V. & Soffredi A. 1963. Grotta del Sanguineto o Grotta della Matta (Finale Ligure, prov. di Savona). Rivista di Scienze Preistoriche XVIII, Not., p. 321. Lamberti A. 1984. Grotta della Giara. Archeologia in Liguria II, Scavi e scoperte dal 1976 al 1981, Genova: Tormena Editore, pp. 185-186.

Leale Anfossi M. 1957. Ricerche preistoriche in Val Pennavaira: “Le Camere” Grotta sepolcrale neolitica (Scavi 1954-55). Rivista Ingauna Intemelia XII, n. 1-3, pp. 22-30. Leale Anfossi, M. 1958-61a. Revisione dei materiali fittili e faunistici provenienti dagli scavi nella Grotta del Pertusello (Val Pennavaira-Albenga). Quaternaria V, pp. 318-320. Leale Anfossi, M. 1958-61b. Conclusione degli scavi al “Pertusello” (Val Pennavaira, Albenga). Quaternaria V, pp. 347-348. Leale Anfossi M. 1958-61c. Scavi e ricerche in Val Pennavaira. Quaternaria V, pp. 357-359. Leale Anfossi M. 1962. Ritrovamenti archeologici e giacimenti preistorici nelle Grotte di Val Pennavaira. Rassegna Speleologica Italiana XIV(2), pp. 190-196. Leale Anfossi M. 1967. Vasi di tipo campaniforme nell’arma di Nasino (Val Pennavaira – Albenga), XI and XII Riunione Scientifica dell’Istituto Italiano di Preistoria e Protostoria, Florence, Italy 11-12 february 1967, Palermo and Lipari, Italy 22-26 october 1967, pp. 249-321. Leale Anfossi M. 1974. L’Arma del Nasino (Savona). Gli strati con ceramica. XVI Riunione Scientifica Istituto Italiano di Preistoria e Protostoria, Genova and Bordighera, Italy 3-5 november 1973, pp. 131-140. Lorenzelli S. & Ricci M. 1980. Quattro anni di scavi nella Tana della Volpe. Nota preliminare. Bollettino Gruppo Speleologico Imperiese, C.A.I. X(15), pp. 49-56. Lorenzelli S. & Ricci M. 1988. La Tana della Volpe e il vaso campaniforme del riparo della Cava di Loreto (Imperia). Rassegna di Archeologia 7, pp. 598-599. Maggi R. (ed.) 1997. Arene Candide: a functional and environmental assessment of the Holocene sequence (excavations Bernabò Brea – Cardini 1940-50), Memorie dell’Istituto Italiano di Paleontologia Umana, n.s. 5, Roma. Maggi R. 1998. Val Pennavaira. In Del Lucchese, A. & Maggi, R. (eds), Dal diaspro al Bronzo, L’Età del Rame e L’Età del Bronzo in Liguria: 26 secoli di storia fra 3600 e 1000 anni avanti Cristo. La Spezia: Luna Editore, pp. 70-76. Maggi R. & Nisbet R. 1991. Prehistoric pastoralism in Liguria. Rivista di Studi Liguri LVI, pp. 265-296. Maggi R. & Pastorino M.V. 1984. Riparo Fascette I. Archeologia in Liguria II, Scavi e scoperte dal 1976 al 1981, Genova: Tormena Editore, pp. 171-173. Maggi R. & Starnini E. 1984. Materiali preistorici olocenici depositati presso il Museo della Val Varatella, Toirano. Rivista Ingauna Intemelia XXXIX, nn. 1-2, pp. 52-60. Maggi R., Biasotti M. & Giovinazzo R. 1998. Toirano – Grotte dell’Olivo e della Gera, in Del Lucchese, A. & Maggi, R. (eds), Dal diaspro al Bronzo, L’Età del Rame e L’Età del Bronzo in Liguria: 26 secoli di storia fra 3600 e 1000 anni avanti Cristo, La Spezia: Luna Editore, pp. 77-78. Maifredi P. 1962. Una grotta sepolcrale presso Tenarda (Pigna). Rivista Ingauna Intemelia XVII, nn.1-4, pp. 50-52. Muñoz Amilibia A.M. 1958. Prospecciones y escavaciones arqueologica en la region de Toirano: La Grotta dell’Olivo (Savona, Italia). Cuadernos de Trabajos de la Escuela Española de Historia y Arqueologia en Roma X, pp. 173-200. Odetti G. 1972. Cento anni di scavo nella grotta Pollera. Atti e Memorie della Società Savonese di Storia Patria, n.s. VI, pp. 11-32. Odetti G. 1984. Val Maremola. Archeologia in Liguria II, Scavi e scoperte 1976-82, pp. 183-184.

138

Chiarenza, Mountains and shore: sites and communication routes of Copper Age western Liguria

Odetti G. 1987a. Grotta del Vacchè. Archeologia in Liguria III.1. Scavi e scoperte 1982-1987, Genova: Tormena Editore, pp. 129-131. Odetti G. 1987b. Val Maremola. Grotta di Ponte di Vara o Varè. Archeologia in Liguria III.1. Scavi e scoperte 1982-1987, Genova: Tormena Editore, pp. 137-140. Odetti G. 1987-88. La grotticella sepolcrale dell’Armusso in Val Maremola (Savona). Rivista di Scienze Preistoriche XLI, pp. 315-330. Odetti G. 1988. Corredi sepolcrali di una cavernetta ligure: Tana dell’Armusso (Savona). Rassegna di Archeologia 7, pp. 600-601. Odetti G. 1996. Pietra Ligure e la Val Maremola dalla Preistoria alla storia, Pietra Ligure. Odetti G. 2003. La Grotta I del Vacché (SV): una cavità sepolcrale del Finale. Ligures I, Bordighera: Istituto Internazionale di Studi Liguri, pp. 215-221. Ricci M. 1964. Un vaso campaniforme nell’Alta Valle Argentina (a Loreto, presso Triora). Rivista Ingauna Intemelia, N.S. XIX , nn. 1-4, pp. 56-59. Ricci M. 1988. Le grotte sepolcrali della Valle Argentina (Imperia) con “perles à ailettes”. Rassegna di Archeologia, 7, pp. 596-597. Ricci M. 1998. Alta Valle Argentina, in Del Lucchese A. & Maggi R. (eds), Dal diaspro al Bronzo, L’Età del Rame e L’Età del Bronzo in Liguria: 26 secoli di storia fra 3600 e 1000 anni avanti Cristo, La Spezia: Luna Editore, pp. 58-62. Ricci M. 1991. Il calcolitico e le grotticelle sepolcrali dell’Età dei metalli nell’estrema Liguria di Ponente (Prov. IM). Le Mont Bégo, una montagne sacrée à l’Age du Bronze, prétirage du Colloque International de Tende, Alpes Maritimes, vol. I, pp. 52-56. Ricci M. & Lanteri Motin E. 1963. Una cavernetta sepolcrale eneolitica a Realdo (Alta Valle Argentina). Rivista Ingauna Intemelia XVIII, nn. 1-4, pp. 93-95. Ricci M. & Lanteri Motin E. 1964. Nuovi scavi nella cavernetta di Realdo. Rivista Ingauna Intemelia, n.s. XIX, nn. 1-4, pp. 60-61. Ricci M. & Lanteri Motin E. 1965. La terza campagna di scavo nella Grotta Sepolcrale eneolitica di Realdo. Rivista Ingauna Intemelia XX, nn. 1-3, pp. 66-69. Richard C. 1932. La caverna degli “Armorari” presso Verezzi. Bollettino della Società Piemontese di Belle Arti e Archeologia XVI, pp. 185-192. Richard C. 1939. Nuovi scavi nella Caverna degli « Armorari» o «Parmorari» (Borgio-Verezzi). Bullattino di Paletnologia Italiana III, pp. 11-24. Tiné S. 1972. Gli scavi nelle caverne delle Arene Candide e della Pollera. Atti della Società Savonese di Storia Patria VI, pp. 7-9. Tiné S. & Traverso A. 1991-92. Gli strati dell’età del Bronzo della Grotta Pollera (Finale Ligure – Savona). Rassegna di Archeologia 10, pp. 335-339.

Acknowledgement This article is a part of a doctoral thesis supervised by R. Grifoni Cremonesi (Pisa University) and D. Binder (Nice University). I would especially like to thank Prof. R. Grifoni Cremonesi for her support for this article.

139

17 Ritual use of a mountain landscape: Iron Age sites in the lower Valle Sabbia (Brescia – Italy) Raffaella Poggiani KellerI, Marco BaioniII I Soprintendenza per i Beni Archeologici della Lombardia, via E. De Amicis 11, I-20123, Milano, Italy – [email protected] II Museo Civico Archeologico della Valle Sabbia, piazzetta San Bernardino 2, I-25085, Gavardo (BS), Italy – [email protected] Abstract A mountainous area in the lower Valle Sabbia (Province of Brescia) has been the subject of numerous archaeological surveys intended to find sites, as part of a project on the valley’s prehistory that started in 2001. This research aims to analyse land use throughout prehistory. This paper focuses on the data regarding a specific type of presence in the territory, connected with the ritual and religious environment. Recent research has identified a genuine sacred landscape, perhaps related to rites of passage celebrated by the whole community, by individual families or by clans. A group of woodland ritual sites is discussed, dating from the Early Iron Age to the Roman period and situated in correspondence to the presence of rock outcrops. The rites involved the lighting of fires, the sacrifice of animal parts, and the ritual fragmentation and dispersal of ornamental objects. These sacred sites were not particularly structured, but made use of natural features such as gullies, caves and suggestively-shaped rocks. Hilltops with a series of terraces were preferred. Wooden structures such as posts and floors were probably present, and in a few cases traces of constructions such as dry-stone walls have been found.

1. Background This paper presents the results of a preliminary investigation aimed at understanding the significance of a series of scattered finds, often metal artefacts, from mountainous areas which are sometimes wooded and difficult to reach. The research was conducted by the Soprintendenza per i Beni Archeologici della Lombardia in collaboration with the Civico Museo Archeologico della Valle Sabbia, and involved the southern part of the Valle Sabbia, the name given to the Lombard portion of the valley of the Chiese River. The shape of the valley — winding and at times narrow — led to the development of a number of alternative routes to the valley-floor road as lines of communication between the Po Plain and the Giudicarie valleys. These roads ran part-way up the hillside or down side valleys, often with easily accessible passes, and were used from prehistoric times until the modern epoch, being abandoned only after the construction of new roads in the 19th century. Several parts of this extensive area were chosen for targeted surveys in the ‘Valle del Chiese’ project, which consisted of a critical re-evaluation of the known prehistoric and protohistoric finds and a renaissance of archaeological investigations in the valley, with the aim of understanding the pre- and protohistoric settlement dynamics of this territory (Baioni et al. 2000; Poggiani Keller & Baioni 2008). The project included an experimental exercise carried out in collaboration between the Milan CNR-IREA Department of Remote Sensing and the Soprintendenza per i Beni Archeologici, entitled

Creation of an archaeological GIS for the middle valley of the Chiese. Creation of a DEM and an archaeological database for the municipalities of Gavardo and Villanuova sul Clisi. This work, coordinated on behalf of the CNR-IREA by P.A. Brivio and conducted in collaboration with the Soprintendenza, was carried out in 2002 by Mirco Boschetti, with the participation of M. Baioni for the study of the archaeological records. The purpose was to combine a GIS based on the DEM with statistical models to explore the relationship between archaeological sites and their environmental characteristics, and thus to provide an explanation for the distribution of

Figure 1 – Locations of ritual sites dated from the Early Iron Age to the Roman period, located in the mountainous area around the lower Valle Sabbia (Province of Brescia): 1. Villanuova – Cima Monte Covolo; 2. Sabbio Chiese – Dos de la Rocca (or Rocchetta); 3. Vallio Terme, Oriolo – Dos de le Preghiere; 4. Gavardo, Monte Magno.

141

Hidden Landscapes

known sites as well as generate ideas for the search for new ones. The identification of parameters describing the geographic locations of known sites (the orientation, steepness, visibility etc. of natural localities), conducted in well-understood trial areas, lays the foundation for predictive models, of particular promise in the case of large-scale public construction projects. Within this extensive and largely mountainous territory two areas were selected which had not suffered a high degree of modern human impact nor been the object of intense research (fig. 1). The first of these was the wooded northern part of Monte Magno, in the municipal areas of Gavardo and Sabbio Chiese, which contains various hills, at times steep, and deeply incised valleys which could have afforded valley-floor routes. The spacious upland known as ‘Monte Magno’ was crossed by one of the roads most used in prehistory and the Roman period (and indeed afterwards), and which by means of a natural saddle joined the lower and middle valley, by-passing the weirs in the valley at Tormini. From the 1960s onwards, numerous Iron Age and Roman finds have come to light in this area. The second selected zone was to the west of Monte Magno and continued (in the Vallio Terme municipal area) up to the watershed between the valleys of the Vallio and the Garza. This area is known by the place-name ‘Ere’, which often indicates open, slightly sloping meadows part-way up a hill. Nowadays it is bare and un-wooded, with numerous small relief features and ancient trackways. The surveys in these two areas led to the identification of (amongst other features) several examples of a type of site previously little known in Brescia’s hinterland. These upland locations may be inter-

Figure 2 – The small rectangular area surrounded by rock outcrops on the top of Dos de la Rocca – Sabbio Chiese.

preted as Iron Age religious sites which sometimes continued into the early Roman Empire. Previously, the only site of this type known was that found in 1973 by G. Bocchio on top of Monte Covolo and later excavated (Barfield et al. 1995). Here there were at least two phases of use, involving a square, walled construction (8 by 8 m) founded in the 7th-6th century BC which remained in use until Roman times, as may be deduced from the objects found in part within the structure and in part — the older material, redeposited — on the hill slopes, cut into narrow terraces in antiquity. Although the structure has not been dated securely, its importance is clear, especially in view of its geographical position between the Chiese Valley and Lake Garda, not far from the Rocca di Manerba, which — to judge from the presence of a notable hoard of Early Iron Age dress pins (Salzani 1982; De Marinis 1989) — could have had a religious function. The newly discovered localities not only bring to an end the isolation of the Monte Covolo site, but are clearly themselves interrelated, creating a sort of sacred landscape. Several characteristics, such as a preference for hilltops and the terracing of the slopes, have been found elsewhere, but the stonebuilt structure on Monte Covolo remains unique.

2. Dos della Rocchetta The first discovery in the northern part of Monte Magno was that of Dos della Rocca (or Rocchetta) in the area known as Rasine in the territory of Sabbio Chiese, found in winter 2003 and currently investigated in part (2004, 2005). It is situated on exposed rock in woodland (fig. 2) that slopes down from Monte Magno towards the hollow containing the modern village of Sabbio (Poggiani Keller & Baioni 2005). The northern and southern slopes of the outcrop are precipitous; the eastern side consists of a rocky spur which offers the only easy access, whereas to the west there is a gentler slope with natural terraces. A fairly soft limestone is exposed throughout the zone; it shows karst erosional features including cave-like hollows, small shelters, sinkholes and large blocks separated from the main body of rock. During the first visit it was already evident that the site exhibited striking features, such as the presence of metal – mostly bronze – artefacts, whole or fragmentary and occasionally showing signs of exposure to heat, and the widespread occurrence of small pieces of burnt bone. The good state of preservation of the finds, their position, the complete lack of visible burial features (grave-cuts etc.) immediately suggested an area of cult practices similar to

142

Poggiani Keller – Baioni , Ritual use of a mountain landscape: Iron Age sites in the lower Valle Sabbia…

the Alpine-type Brandopferplätze (Gleirscher 2002; Steiner 2010). The goal of the 2004 and 2005 excavation campaigns was to try and understand the nature of this archaeological discovery, to assess its preservation state, and to determine its chronology. Prior to the 2004 excavation several preliminary tasks were accomplished. A detailed topographic survey of the area was made by Total Station, permitting the identification and measurement of a series of terraces. Using aerial photography a DEM was prepared, which allowed the discoveries to be placed in their natural context. Subsequently, with the aid of a metal detector, the rough limits of the zone in which metal (bronze and iron) finds were present on the ground surface was established and added to the general plan. It was clear straight away from this survey that there was a concentration of metal objects on the central and western portions of the upper flat area and the natural steps on the west side. Clearly visible on the uppermost area was a small rectangular area surrounded by rock outcrops, which at some points seemed to have been shaped by human hands so as to create a sort of room with a regular floor. This bounded area was oriented N-S and situated near the highest point of the rock. The sides and floor of limestone were pierced by a series of vertical holes, some of which were probably intended to hold wooden supports, perhaps part of a standing structure with a wooden floor. In one corner of this rectangular area there was a deep crevice, a sort of natural cistern filled with deposits containing many archaeological finds. This small area was surrounded by a floor partly formed by rock which had been smoothed and partly paved with small slabs of the same limestone. Archaeological finds were present everywhere, both in the rectangular area, in the rocky area to the west, and on the lower terraces. Often they were found in cracks and crevices, as well as under small overhangs where measures had been taken to direct the flow of rainwater, for example in drainage channels. Although the study of the abundant material has not yet been completed, the following general picture of the site’s occupation phases has emerged: • The area was already in use — perhaps as a hunting station — in the late Early Bronze Age, from which epoch flint tools such as tanged knife blades, tanged and winged arrowheads, and scarce pottery fragments were found. • This first occupation phase seems to have been followed by a considerable pause, since the finds

a

b Figure 3 – Finds from the top of Dos de la Rocca – Sabbio Chiese. a. Certosa type fibula; b. three-ringed pendant with notched edges and three hollows in correspondence to the rings.

from the second phase date to the middle of the Iron Age, when the site acquired its cult function and underwent the first important transformations. The artefacts from this phase appear to date from the 5th century BC, such as a Certosa fibula that resembles Teržan type VII, and a three-ringed pendant with notched edges and three hollows in correspondence to the rings (fig. 3). • A Po Plain silver drachma dates to the 1st or 2nd century BC, while the majority of the metal objects, both bronze and iron – brooches, spirals, pendants etc. – are from the 1st century BC to the 1st century AD. This more recent material includes various fibulas of Rhaetian and Celtic tradition, together with Timoline-type brooches, perhaps better referred to as Valsabbia fibulas. Particularly interesting is the presence of a disc-shaped part of a fibula portraying a helmeted head. Alongside these items of local tradition there are objects more expressive of Romanization (period coin, keys, etc.). The occurrence of several hundred small stud-like

143

Hidden Landscapes

nails scattered throughout most of the area is still awaiting interpretation. The Dos della Rocchetta site was an area devoted to cult practices, where particular precious objects were perhaps dedicated to a divinity. Although the older artefacts are complete, the 1st century BC and AD material is deliberately broken and in rare cases shows signs of exposure to high temperatures. This phenomenon recalls the Brandopferplatz sites, as does the presence of numerous, often tiny, pieces of burnt bone and relatively few scattered charcoal fragments. The usual blackish carbon-rich layers and hearths are lacking, although the heavy rain runoff to which the area is subject could have completely removed features (which would not in any case have been very robust). The finds recovered attest to the existence of various rituals: • the offering of unbroken objects that were probably first deposited in a place of storage and then scattered over the hillside (similar to the situation found on Monte Covolo); • the offering of ritually broken objects that were deposited in crevices;

• the offering of objects and parts of animals on a fire capable of reaching high temperatures. In the woods surrounding the main site several other sites were later discovered. These were smaller, but with the same characteristics: location on areas of rocky relief with a sheer drop on at least one side, the presence of scattered pottery and metal objects, and unusual rock formations. These sites are ranged around the principal site on routes leading to the valley bottom and are always in highly visible locations. Stratigraphic data from the principal locality suggests that the area may previously have been less densely wooded and more open, a situation compatible with the lighting of fires; visibility in general, and intervisibility between these sites, seems to have been required. Next to larger sites that were probably used by an entire community, more restricted areas, often containing few finds and probably associated with family or clan cults, occur all over Monte Magno. A recently discovered example is the three Timoline or Valsabbino-type fibulas and several potsherds in Palazzina (Gavardo). This development makes it necessary to

Figure 4 – The hill known locally as ‘Dos de le Preghiere’ (Prayer Hill) occupies a commanding position on the slope to the north of the village of Oriolo, Vallio Terme.

144

Poggiani Keller – Baioni , Ritual use of a mountain landscape: Iron Age sites in the lower Valle Sabbia…

reclassify all the finds from the Valsabbina and Garda areas that were considered sporadic or funerary. For instance, objects found not long ago in Gardesina (Salò) and published as grave goods might well have come from this other type of site (Simonotti 2005).

3. Conclusion The majority of the minor localities which make up this network of sacred sites spread through woods and over hillsides were detected by means of systematic survey and by the re-assessment of old sporadic finds. We have so far investigated completely only two mountain sites characterized by a limited number of ceramic and metal finds. Monte Mizzigolo, Agnosine, lies outside the category discussed above for reasons of geography and typology, but the other site which – significantly – is known locally as ‘Dos de le Preghiere’ (Prayer Hill) provides an excellent example (fig. 4). This modest hill occupies a commanding position on the slope to the north of the village of Oriolo (Vallio Terme). In this case no constructions have been recognised, just a surface deposit partially covering an erosion channel. In the basal soil a few stone artefacts, probably Mesolithic in date, were found, while the single overlying layer (covered directly by grass) yielded pottery and metal fragments attributable to the Early and Late Iron Ages. Amongst the limited number of offerings made of bronze in this small sacred area, the presence of pendants (boot and lance-shaped), generally thought to indicate burials of women or young people, might suggest initiation rites which marked passage into the adult world (Casini 1994, p. 187). A sacred landscape is thus coming to light in this area, a use of this mountainous zone not tied to subsistence, in which principal sanctuaries are accompanied by smaller sacred localities, perhaps connected with family cults.

Casini S. 1994. Gli amuleti nella cultura di Golasecca tra il V ed il IV sec. a. C., Notizie Archeologiche Bergomensi 2, pp. 187-197. De Marinis R. 1989. Preistoria e protostoria della Valcamonica, Valtrompia e Valsabbia: aspetti della cultura materiale dal Neolitico all’età del Ferro, in Poggiani Keller R. (ed.), Valtellina e mondo alpino nella preistoria, Modena, pp. 101-119. De Marinis R. 1992. Il territorio prealpino e alpino tra i laghi di Como e di Garda dal Bronzo Recente alla fine dell’età del Ferro, in Metzger I.R. & Gleirscher P. (eds), I Reti. Die Räter, Bolzano, pp. 145-174. De Marinis R.C. 1999. Il confine occidentale del mondo protoveneto/paleo-veneto dal Bronzo Finale alle invasioni galliche del 388 a.C., in Protostoria e storia del “Venetorum Angulus”, Atti del XX Convegno di Studi Etruschi ed Italici 16-19 ottobre 1996, Pisa-Roma 1999, pp. 511-564. Gleirscher P. 2002. Brandopferplätze in den Ostalpen, in Gleirscher P. , Nothdurfter H. & Schubert E., Das Rungger Egg (Römisch-Germanische Forschungen 61), Mainz am Rhein, pp. 173-264. Poggiani Keller R. 2001. L’area prealpina e alpina lombarda nell’età del Ferro. Alcuni aspetti del popolamento e relazioni culturali, in Atti 3° Convegno Archeologico Regionale La protostoria in Lombardia, Como 22-24 ottobre 1999, Como, pp. 415-437. Poggiani Keller R. & Baioni M. 2005. Sabbio Chiese (BS), Dos de la Rocchetta – Rasine. Sito dell’antica età del Bronzo e area di culto dell’età del Ferro, Notiziario della Soprintendenza per i Beni Archeologici della Lombardia, pp. 83-84. Salzani L. 1982. Un probabile ripostiglio sulla Rocca di Manerba (Brescia), in Studi in onore di Ferrante Rittatore Vonwiller, Parte I, Vol. II, Como, pp. 665-677. Simonotti F. 2005. Salò (BS), Località Serniga, Indagini archeologiche, Notiziario della Soprintendenza per i Beni Archeologici della Lombardia, pp. 84-85. Steiner H. (ed.) 2010. Alpine Brandopferplätze-Roghi votivi alpini, Forschungen zur Denkmalpflege in Südtirol V, Trento.

References Poggiani Keller R. & Baioni M. 2008. La Valle Sabbia nella preistoria e protostoria: ricerche e lavori in corso, in Mottes E., Nicolis F. & Zontini G. (eds), Archeologia lungo il Chiese. Nuove indagini e prospettive della ricerca preistorica e protostorica in un territorio condiviso tra Trentino e Lombardia, Storo, pp. 91-106. Baioni M., Bocchio G. & Spinelli P. 2000. Il progetto “Valle Sabbia” e le altre attività del Museo, Annali del Museo 18, pp. 109-127. Barfield L.H., Buteux S. & Bocchio G. 1995. Monte Covolo: una montagna e il suo passato, Birmingham. Bocchio G. 2004. Nuovi reperti dal luogo di culto di Monte Covolo, Annali del Museo 19, Gavardo, pp. 148-149.

145

18 A journey through ‘hidden’ or forgotten landscapes in the northern Veneto. Patterns and biases in material culture and research methods Sarah De Nardi Department of Geography, University of Hull, United Kingdom – [email protected] Abstract This paper focuses on problems concerning the extent of knowledge about Iron Age and/or Roman votive deposits and sanctuaries in the Po plain (very well known) and votive sites in the Alps and mountains (virtually unknown) in the Veneto. I identify the problems and barriers that occur when trying to establish an overall understanding of cult in the region in light of this bias. The paper touches on both methodological issues (biases in chance find/excavation/survey) and the problematic cultural and spatial processes pertaining to place, cult and identity.

1. Introduction My research seeks to understand and interpret the extent of people’s engagement with their landscape in the region known as the Veneto in northeast Italy, in a time span encompassing most of the Iron Age (from the 8th century BC) to Romanisation (2nd century BC). During my PhD research I became increasingly aware of the imbalance in knowledge and understanding of not only the socio-cultural implications of the landscape interaction for pre-Roman peoples inhabiting the Veneto, but also of the stark inequality in what was- and is- known about the various parts of the region. Patterns in the distribution, extent and chronology of sacred places, settlements and cemeteries are uneven in this vast region encompassing a variety of milieus, ranging from the Alpine Cadore in the north to the Adriatic coastline, the Venetian lagoon and the Po plain (fig. 1). In this paper I will limit my geographical scope to the northern sector of the Veneto — the modern provinces of Treviso and Belluno (TV and BL for short). I will bring to the fore issues concerning the interpretation of Venetic landscapes, identify existing biases, theoretical and methodological shortcomings in Venetic studies, and suggest alternative ways of approaching the material culture of the region in a way that takes into account its ‘hidden’ and ‘forgotten’ corners. I will do so in four ways: in section 2 I identify biases and inconsistencies in our knowledge of Venetic material culture pertaining to cult and burials, and identify the likely reasons why some parts of the region (i.e. the Pedemontana and the mountains) seem to have fewer votive deposits, cemeteries and particularly settlements than the Po plain. In section 3 I will present four case studies that illustrate the situation: Monte Altare and I Frati – Vittorio Veneto (TV), Monte Nenz – Trichiana

and Monte Dolada – Staol di Curago (BL). In section 4 I will briefly highlight one issue that differs widely — and is quite puzzling — in the region: the distribution and extent of literacy during the Iron Age. Finally in section 5 I will address the issue of journeys and movement of people and ideas across ‘frontiers’.

2. Where are the sites? The extent to which archaeological evidence is published and disseminated throughout the Veneto is uneven. We know much more about the Po plain and, increasingly, about the Venetian lagoon than about the Pedemontana and the Alps. There are fewer research excavations, next to no survey projects, and fewer museums displaying grave goods and votive objects from sites in these ‘peripheral’ sectors of the region than elsewhere (De Nardi forthcoming a).

Auronzo Lagole

Monte Dolada Mel

Monte Nenz

Monte Altare Ai Frati Castello Roganzuolo

Villa di Villa

Santorso Montebelluna

Oderzo Altino

Lova

Montegrotto Este

0

10 20

50 km

Figure 1 – The Veneto study area.

147

Venice

Hidden Landscapes

Site name

Context

Wellknown locations

Sufficiently Little known known locations locations

*Lova

Lagoon

×

*Este Baratella

Po plain

×

*Este Meggiaro

Po plain

×

*Este Caldevigo

Po plain

×

Este Benvenuti

Po plain

×

*Montegrotto

Po plain

×

*Padua – Pozzo Dipinto

Po plain

×

*Padua – San Daniele

Po plain

×

*Altichiero

Po plain

*Mortise

Po plain

×

*Mandriola

Po plain

×

*Musile di Piave

Lagoon

*Altino La Fornace

Lagoon

*Altino Canevere

Lagoon

*Asolo

Piedmont

Montebelluna

Piedmont

*Fossalta di Portogruaro

Lagoon

*Vicenza

Piedmont

*Santorso

Piedmont

Ai Frati

Piedmont

*Castello Roganzuolo

Piedmont

*Prà di Stalla

Piedmont

×

*Montereale Valcellina

Piedmont

×

*Scomigo

Piedmont

*Monte Altare

Piedmont

*Villa di Villa

Piedmont

Monte Dolada

Alps

*Vallesella

Alps

×

I would like here to draw attention to four relatively underrated sites, and ask how they can inform our knowledge of the wider development of cult places or cemeteries in the Veneto. In a research arena still widely dominated by knowledge of sites in the Po plain and — to a lesser extent — the lagoon, the upland sector of the Veneto is not a ghost country: if one knows where to look, the traces are there. Most of what we know about cult and funerary data in the Veneto, however, comes from Este and the Po plain (Padua, Montegrotto etc.), and this may lead to inaccurate generalisations and the imposition of ‘templates’ onto areas that culturally and geographically had little in common with these centres of the Po plain (De Nardi forthcoming a). Perhaps we can reconstruct a more faithful, comprehensive picture of ritual and funerary processes in the Veneto by including the alternative locations of a. Monte Altare, Vittorio Veneto; b. I Frati cemetery, Vittorio Veneto (TV); c. Monte Nenz, Trichiana; d. Monte Dolada, Staol di Curago (BL).

× × × × × × × × ×

× × × × ×

Alps

*Lagole

Alps

×

*Auronzo

Alps

×

*Valle di Cadore

Alps

Mel – cemetery

Alps

of which votive deposits

3. Case studies: the search for equal representation

×

*Monte Nenz

Total sites

ology that privileges parts of the region — such as Este and its environs — which therefore yields the most evidence. The local topography of the Veneto is also partly to blame: it is easier to identify sites and objects in a plain or lagoon area than on the wooded hills and mountains of Pedemontana and the Alps. The imbalance in the number of prehistoric archaeological sites known in upland regions may therefore be due to visibility constraints as well as to an actual lesser extent of occupation by local peoples during later prehistory.

Precise location unknown

×

× × 10

10

7

6

9

8

7

5

3.1 Monte Altare

Table 1 – Extent of accuracy pertaining to the location of Venetic cult places and cemeteries and their topographical context. Cult places are marked by an asterisk (*).

Chance finds are summarily and sketchily published in local journals and rarely get a chance of entering nation-wide publications or bulletins. Table 1 shows the variation in the extent to which we can identify and locate Iron Age votive deposits or cemeteries. Este and sites in the Po plain are wellknown, widely studied and have been a prominent feature of Venetic studies since the dawn of archaeology in the region in the 19th century. This could point to a well-established tradition of doing archae-

This site, a votive deposit located on the top of a hill near the town of Vittorio Veneto in the province of Treviso and originally excavated in 1989, was published in 2005 as part of a volume entitled Stipi Votive delle Venezie (Gambacurta & Gorini 2005). This monograph provides no contextualisation of the site in its local milieu, no social interpretation of the finds associated with the votive deposit, and the selective catalogue of objects published in the volume is not exhaustive (Gorini 2005). The locality is an evocative hill in the piedmont, a very special place attracting votive dedications more or less without interruptions from the 6th century BC to the 4th century AD. Figure 2 shows the hill and

148

De Nardi, A journey through ‘hidden’ or forgotten landscapes in the northern Veneto…

its environs, and the typical votive object dedicated by the Iron Age worshippers, the bronze lamina (fig. 3). I have extensively studied this locale from a variety of perspectives: it stands out from other Venetic meaningful locales for its long-term significance. Looking at the view in figure 2a, the morphology of the Monte Altare environs is reminiscent of a type of votive object dedicated at the site from the 5th century BC down to the 2nd BC circa: the ‘geometric’ or ‘crenellated’ laminae. In the photograph, Monte Altare is on the left, Monte Piai in the middle, and San Paolo (site with Roman burials and late Roman fortifications) on the right. Their shape resembles the laminae from Monte Altare, which elsewhere I interpreted as stylised depictions of the sacred hills (De Nardi 2006, p. 137 ff.; De Nardi forthcoming b).

of the Veneto know the ‘star find’: the Trichiana Key (fig. 4d). This is a clear example of a fairly wide-spread scholarly attitude to votive deposits, which accepts the indisputable value of the artefact as absolute material culture object, but denies the need for clear spatial contextualisation. The context and circumstances of deposition of meaningful objects such as the Trichiana key is lost unless we try to investigate and interpret the act of deposition itself, and the first way to do this is to move out of the display room and visit the actual hill. I did so, exploring the spatial relationship of the hillock with its environs, looking at what would have been visible or hidden from this ‘one-off’ location that many hesitate to call ‘votive’ tout court.

a 3.2 I Frati

Positioned in the plain south of Monte Altare, this is a ‘forgotten’ cemetery: excavated in the 1880s during works for a new theatre and never documented, with the ‘worthy’ grave goods duly entering private collections, thereby destroying evidence of a burial site with a story to tell. The history of this lost burial site was published by Giorgio Arnosti (Arnosti 1996a) in the local journal Il Flaminio: the nature of the finds, the likely chronology of the cemetery and its social implications. Figure 4a-c shows a selection of grave goods from this cemetery, the chronology of which encompasses a staggering seven to eight centuries (from the 7th century BC to the Republican period; De Nardi forthcoming a). The site was never until now published in a national or international forum (De Nardi forthcoming a) and the objects are dispersed in a number of local museums and private collections: the finds are ‘hidden’. Thus, any publication dealing with Venetic mortuary practices and patterns until now has left out I Frati, and has therefore painted an inaccurate picture of Venetic funerary material culture. This led to the inaccurate impression that the funerary typology from the Po plain also applied to other parts of the region, despite the fact that the types of grave goods found in the northern and eastern parts of the Veneto are almost never found in the south.

b

Figure 2 – a. Monte Altare, Vittorio Veneto. Photograph courtesy of the GAC; b. a cumulative vista of hills: the model for the votive laminae?

3.3 Monte Nenz, Trichiana (Belluno)

Monte Nenz is one of those sites that few can place geographically (the generic locality of the nearby village of Trichiana is given in the scarce publications devoted to the site), but of which all scholars

Figure 3 – Sample of bronze votive laminae from Monte Altare. Photos courtesy of Giorgio Arnosti and the GAC.

149

Hidden Landscapes

stations for travellers to pay their dues to local deities (De Nardi forthcoming a), especially if we accept the interpretation of the key as a symbol of passage and transition, if not of the ultimate act of passage: death (Bonomi & Ruta Serafini 1994, p. 12). The diversity of material culture and mode of deposition at Monte Nenz remind us that we have to interpret and attempt to understand Venetic rituals from a local perspective, not by applying criteria created for urbanised areas of the Po plain. 3.4 Monte Dolada, Pian dell’Agnella

Figure 4 – a-c. Selection of Iron Age grave goods from the Ai Frati cemetery in Vittorio Veneto. All photos courtesy of the Gruppo Archeologico del Cenedese; d. The Trichiana key. Courtesy of Gruppo Archeologico Bellunese.

Figure 5 – Monte Nenz, Trichiana (Belluno).

Figure 6 – The Pian dell’Agnella cemetery. a. positioning of the graves on the edge of the escarpment. b. looking up towards the cemetery from the dip below.

However we interpret the scatter of objects found on the hilltop and slopes, it must be borne in mind that the agents of the act of deposition of the key and the pottery found in the same pit (Padovan pers. comm.) would have been aware of the topography of the area, and would have selected particular locations according to whether they were significant, special, sacred or tied to legends, ancestral tales and cosmologies. The fact that the key seems to belong to a Hallstatt rather than a Venetic cultural horizon (Gambacurta 2003, p. 51) may mean that traffic and commerce to and from Austria was thriving and necessitated ‘sacred’

This site, situated on the south-facing slope of the Monte Dolada, is a relatively recent discovery, and at 800 metres a.s.l. one of the highest-altitude Venetic cemeteries. There were two separate burial places at the same elevation but about 150 m apart, the Iron Age site (Pian dell’Agnella) and a Roman cemetery (known as the ‘Dolada’ cemetery). What distinguishes this location from other Alpine cemeteries is the isolated but evocative nature of the place. As can be seen from fig. 6, the burials, dating from the 6th to the 4th–3rd centuries BC (circa), are distributed on a plateau overlooking an escarpment. Visiting the location, my colleagues and I had the distinct impression of a dramatic positioning of the graves as if on display, on a ledge, overlooking the dip below (De Nardi forthcoming b). The grave goods, from what little information I could gather from the internet (www.archeoagordo.it/12/alpago.htm), are consistent with a high-rank burial place, with the richest graves positioned towards the edge of the escarpment. This suggests the site was established by aristocratic families to bury prominent (mainly male) individuals, but the lack of evidence for any nearby settlements leads me to ask whether this was perhaps a regional high status burial lot rather than a local cemetery for local people. Again, the unique context and nature of this burial place demands that it be included in national publications on Venetic cemeteries, to really convey the idea of the variations happening across the region in the Iron Age, closely mirrored by the irregularity of settlement, sanctuary and cemetery pattern typical of the Roman period in the northeast.

4. What about literacy? In the introduction we saw how inequalities in the weight of evidence about Venetic sites tended to place greater importance on the south, in particular the Po plain. Making comparisons between this area and other parts of the region is not only desirable, but indispensable to an overall understanding of

150

De Nardi, A journey through ‘hidden’ or forgotten landscapes in the northern Veneto…

Venetic culture(s): however this task is not easy, and pan-regional patterns are hard to identify in light of the disparity of available knowledge and the different extent of research and fieldwork going on in the different parts of the Veneto. This disparity is best expressed by different quantities of known sites in the region and by the diversity of cult expressions across the land, both in terms of landscape context, longevity, relationship with settlements and material culture. Writing plays an important role in the construction of local identities across the Veneto, and is implicated in both the ritual and funerary spheres of Venetic culture. Inscriptions, mostly from ritual contexts, are present in the northwest (Vicenza, Santorso), at the Po plain site of Este (Baratella) and in Lagole, but what about the area in between? To go from the eastern Po plain and the rest of peninsular Italy to the Alps, you had to go via Val Lapisina, the Vittorio Veneto area, and the Val Belluna: what was later to become the Via Alemagna. It seems likely that the stimulus to write was less strongly felt in the Pedemontana area apart from a few exceptions — mostly cult contexts (Villa di Villa and Castello Roganzuolo, see fig. 7). My interpretation of this is that in the area between the great urbanised centres of the Po plain and the La Tène area of Noric Austria, the local groups and communities were not as concerned about making a statement of independence or ownership of the land because before Romanisation there were no pressures from neighbouring groups that could have threatened their livelihood and peace.

5. Journeys in marginal(ised) landscapes The pattern of presence (and absence) of archaeological traces in the Veneto can aid in the reconstruction of the patterns of dwelling and movement of people, goods and ideas in the ancient land. There is ample evidence for the passage and existence of a number of ‘cultural groups’ in the territory of the Veneto during the Iron Age and the period of Romanisation. This is not always brought to the fore in the archaeological literature, largely because most of it focuses on the southern part of the region, where people came in contact with Etruscans, Greeks and other peninsular peoples by way of trade and travel (e.g. Capuis 1993). The extent to which movement happened in the east and north is of course different and more difficult to reconstruct based on the scantier archaeological evidence. Movement of people, ideas and material culture (including the practice of writing) could be detected by identifying objects or practices characteristic of a cultural group out-

side of the bounds of that group; by drawing parallels, or lines of development across neighbouring or faraway areas; and by tracing actual ‘journeys’: commercial, processional ways or symbolic tracks in the landscape, signalled by markers such as artefact depositions and more or less seasonal occupation. One typical case study is that of Gurina in the Gailtal, Austria. This site was most probably a sanctuary producing bronze laminae bearing Venetic inscriptions (Capuis 1993; Arnosti pers. comm.) and fibulae of Northern Venetic type (Arnosti pers. comm.). The dating of the objects is not clear, but they can be ascribed to the middle Iron Age due on a typological basis. The area did not yield any traces of a settlement, suggesting that the site of Gurina could have been an en route cult place founded or frequented by Venetic peoples to safeguard the passage across the Alps on their journeys. This demonstrates a certain extent of mobility and of awareness of people’s surroundings. Whether the presence of a cult place beyond the traditionally Venetic bounds can indicate an actual occupation of the area by Venetic peoples is much disputed, but the presence of a cult place apparently not linked to any settlements or cemeteries fits the pattern of Veneto’s piedmont. There, we have sites such as Castello Roganzuolo and Villa di Villa positioned along north-south and west-east routes to and from the Austrian Noric and the Carnic Friuli, that were established as symbolic landmarks, points of rest and worship catering to travellers, merchants and pilgrims, rather than as territorial sites used year-round. The presence of Noric-type armour and northerntype votives at cult sites such as Lagole, Monte NenzTrichiana, Vallesella di Domegge (Belluno), and the so-called ‘Noric obols’ at piedmont sites such as Monte Altare and Villa di Villa also indicates the move-

Figure 7 – Inscribed cippus (votive?) from the cult place of Castello Roganzuolo, Treviso.

151

Hidden Landscapes

ment of people and the practice of dedicating objects to local deities to appease them and ensure one’s safe journey through often hostile landscapes. For brevity’s sake I will simply suggest here that many sites in the northern part of the Veneto should be interpreted as en route locales established for the safe passage of people in a dynamic landscape: as such they ought to be understood differently from large-scale cult sites in the south and east of the region, where settlements and territorial units were responsible for their establishment and maintenance; these were local people’s shrines, protecting the community or catering for certain parts of society rather than ‘public’ pilgrimage shelters and travellers’ oases across the land.

and essential to the building of archaeological knowledge anywhere, and the dialogue should be more consistent and constructive, geared towards collaboration rather than simply being ‘optional’. There is also the need to contextualise northern Venetic culture(s) in the Alpine and Central-European milieu, for it is clear that they share many features with transalpine populations, especially as reflected in cult and mortuary rituals. In future I will endeavour to bring the Veneto, but especially the lesser known parts of it, out of the local shell and onto the wider European archaeological debate.

References 6. Conclusion I have tried to convey a sense of what is missing, or under-represented, in the archaeological literature pertaining to the pre-Roman Veneto, but do not wish to imply that the publication process is necessarily biased per se. An element of choice and preference towards sites and areas one knows best, or most, about, is inbuilt in the academic way of doing things, especially when time is restricted and there are no funds to publish extensively about every site. However, this effect should be kept to a minimum, and scholars should be made aware of the consequences. By choosing what to represent and what to publish one can in fact ‘hide’ landscapes, places and meaningful contexts from the archaeological scene, and prevent interpretation. The ‘umbrella approach’ tends to ignore local features (sites established, used and perceived in different ways and for different reasons) in favour of generalisations that hinder rather than help a holistic and reliable interpretation of Venetic material culture and people-landscape interaction. The fact that Venetic archaeology has, until very recently, been mostly concerned with the social-spatial dynamics of settlements and of ritual sites connected to settled space has led to the bias that ignores or marginalises rural landscapes, minor isolated sites and artefact scatters. These in themselves can tell much about cultural patterns across the region if included in the wider scope of Venetic studies. If these marginal or hidden locations continue to lurk in the local amateur’s realm we will continue to miss out on much of the bigger picture. By working alongside local Archeoclubs and university-based scholars I have sought to bridge the gap and facilitate the dialogue between these two kinds of specialist. As noted by PoggianiKeller (this volume), local knowledge is paramount

Arnosti G. 1996a. Per Cenetam gradiens. Appunti sulle vie della romanizzazione con riferimento all’antico Cenedese, Il Flaminio – Rivista di Studi Storico-Archeologici della Civiltà Montana delle Prealpi Trevigiane 9, pp. 59-105. Arnosti G. 1996b. L’età del ferro nell’antico Cenedese, Circolo Vittoriese di Ricerche Storiche, Vittorio Veneto, pp. 49-81. Bonomi S. & Ruta Serafini A. 1994. Una “chiave di Penelope” dal territorio Bellunese, in Quaderni di Archeologia del Veneto 10, pp. 11-13. Capuis L. 1993. I Veneti: società e cultura di un popolo dell’età Preromana. Longanesi, Milan. Capuis L. 1998. Per una rilettura dell’iconografia/iconologia dei dischi, in Gambacurta G. & Capuis L., Contributi: dai dischi di Montebelluna ai dischi di Ponzano, Quaderni Di Archeologia del Veneto 14. De Nardi S. 2006. Il Monte Altare nel suo contesto paesaggistico: aspetti di un’analisi interpretativa “umanistica”, in Il Flaminio – Rivista di Studi Storico-Archeologici della Civiltà Montana delle Prealpi Trevigiane 15, pp. 133-144. De Nardi S. forthcoming a. Landscapes of the prehistoric Veneto, Italy. A plurality of local identities reflected in cult and landscape perception, in PIA, Papers from the Institute of Archaeology, UCL (2008). De Nardi S. forthcoming b. How natural are natural places? Challenging stereotypes in the interpretation of landscape in Iron Age Veneto, Italy, Proceedings of “Place as material culture (materiality, methodology and metaphor)”, session at the EAA meeting in Zadar, 20/09/07. Gambacurta G. 2003. I circoli di Mel (BL) e la chiave di Trichiana (BL), in Malnati L. & Gamba M. (eds), I Veneti dai bei cavalli. Treviso: Canova, pp. 50-51. Gambacurta G. & Gorini G. 2005. Il deposito votivo di Monte Altare (Treviso), in Gorini G. & Mastrocinque A. 2005 (eds), Stipi votive delle Venezie: Altichiero, Monte Altare, Musile, Garda, Riva. Giorgio Bretschneider Editore, pp. 105-231. Gorini G. 2005. Catalogo H – monete, in Gambacurta G. & Gorini G. (eds), Stipi votive delle Venezie: Altichiero, Monte Altare, Musile, Garda, Riva. Giorgio Bretschneider Editore, pp. 173-231. Poggiani-Keller R. 2007. Buried landscapes and cultural landscapes in the mountain areas of Lombardia between pre- and protohistory. Paper presented at the Hidden Landscapes conference, Siena 25-27 May 2007.

152

19 Two hidden landscapes in central Portugal: Rego da Murta (Alvaiázere) and Ocreza (Mação) Alexandra Figueiredo, Luiz Oosterbeek Department of Territory, Archaeology and Heritage, Polytechnic Institute of Tomar, Tomar, Portugal – [email protected]; [email protected] Abstract Since the 1990s, the TEMPOAR programme team has been developing several research projects on the prehistoric settlement of Alto Ribatejo (central Portugal). These interdisciplinary studies have contributed to a better understanding of the life and interrelationships of these populations. Based upon two ‘hidden landscapes’ located along the northern edge and in the south-east of the region (Rego da Murta in Alvaiázere and Ocreza in Mação), we will present the main conclusions reached so far, focusing on the range of options and behaviours of prehistoric populations that led us to a new approach and to go beyond the hidden-open dichotomy.

1. Introduction Since Christopher Tilley (1994) published his phenomenological approach, a concern with the landscape problematic has been growing within the archaeological community. Tilley’s approach, however, is not the only possible one: others including Thomas (1993, 1996), Bradley (1998) and Ingold (2000) have also offered reflections. These approaches also had an impact on the development of spatial analyses employing Geographic Information Systems (e.g. Llobera 1996, 2001). This theoretical shift towards a phenomenological approach (despite the huge differences between Tilley’s, Ingold’s and Thomas’s works: see Ingold 2005) gave rise to the perception of landscape not as a determinative element (as according to an environmental determinism), but as an element interpreted according to the concept of being in the world. This shift has put the emphasis on the sensory aspect as an approach to understanding: concepts such as visibility, soundscapes and others have grown in importance. Actually, the topic of landscape visibility is a recent concern (see, for instance, the dedicated issue 39.1 of the World Archaeology journal, published in 2007). So, what can we theoretically understand? Can we establish an understanding about what a ‘hidden landscape’ is or was? The introduction to the call for papers referred to the apparent diversity within the archaeological record of the Mediterranean region. It suggested that coastal sites were reasonably tracked, whereas mountain and valley sites were poorly recorded. This raises a methodological problem: are there more archaeological sites in coastal areas? Are these more visible than those in inland areas? Or are inland areas insufficiently studied? Now, this characterization of mountain and valley landscapes as closed landscapes refers us again to issues such as landscape understanding, interpreta-

tion of its morphological characteristics and its relationship with the type of sites found therein. We do not want, however, to fall back into the fallacy of environmental determinism which dominated archaeology before, and even less to define cultural circles (which contain fallacies that are already identified and do not fall within the scope of this small paper). When we consider the Portuguese case, and particularly the Alto Ribatejo region (inner central region of Portugal), it becomes clear that the idea of a hidden landscape is much more related to the presence of gaps in the archaeological record than to any access or visibility constraints of the archaeological record itself (in the sense that some regions have attributes that would make archaeological sites more visible than others). Inescapably, the work developed by our team in the Alto Ribatejo, particularly that of the current authors (Oosterbeek 1994; Figueiredo 2006) and Ana Rosa Cruz (1997), has attempted to fill pre-existing research gaps allowing the construction of a visible record rather than a hidden one as in Heidegger’s Hervorbringen approach. The path proposed here is closely related to the hermeneutic approach to landscape perception, not only using modern visualisation tools such as GIS, but also the experience acquired during the process. Thus, in the study region two cases are worth mentioning as interesting examples of possible hidden landscapes. The first is the significant number of cave burials and initial unawareness of other practices in artificial monuments in the Nabão region. This made us realise that in limestone areas the type of occupation found in prehistoric burials would be different from that of neighbouring geological regions, when we compare these practices with invisible areas of limited access within closed landscapes or in more restricted social circles. The second case includes technical means in understanding all these

153

Hidden Landscapes

Figure 1 – Location of the Megalithic Complex of Rego da Murta in the northern part of the Alto Ribatejo region. To the east lies the Zêzere river.

different factors, where using more complex methodologies is necessary for a more reliable approach to the analysis of the archaeological record. This is the case with the prehistoric art of Ocreza Valley, which is mainly located in steep areas or on the banks of the Ocreza river and dispersed throughout its typical closed, mountain-fringed valley, which contributes to a greater invisibility. Despite the differences between them, these two examples help us explain how this hidden-open dichotomy can be overcome through systematic fieldwork.

2. Caves and megalithic monuments of the Nabão region Until the last decade, the Nabão region was interpreted according to different patterns from those of neighbouring regions and practices in use in the Tagus or Zêzere regions. The structure, cults, materials and types of monuments characterised it as a unique region with its own characteristics, probably related with different social groups or different relationships/contacts, producing over time a unique, almost invisible archaeological case. Therefore, studies developed in the nineties (Oosterbeek 1994; Cruz 1997) reflected an approach in which the Alto Ribatejo was split into two different areas: the Nabão linked to the cave world and originating from the cardial and the Tagus and Zêzere linked to the constructors of the megalithic monuments.

The earliest papers on cave contexts, such as that at Cadaval (Oosterbeek 1985), recognised contacts with the ‘megalithic world’, but still suggested on the basis of the archaeological record that the organization of the first Neolithic landscape in the Nabão region followed different rules from those found in the South. For example, in the Nabão region cave depositions were used instead of artificial monuments or nuclei standing out from the landscape as in the Tagus and Zêzere regions. Ritual features combined with artificial structures in cavities funnelling the entrances, the lack of data for bone deposition comparison, differences between the exhumed artefacts and those recorded in known megalithic monuments and the invisibility of these sites could be sufficient reasons for this region to be considered as a hidden landscape in the living sense of a prehistoric community that was different from the remainder. A more systematic interpretation of parallel articulated rituals between the caves and the megaliths in the Later Neolithic was proposed in the mid-1990s (Oosterbeek 1995). However, the discovery (by father Jacinto from Pussos) of two megalithic monuments in the complex of Rego da Murta (including at least 3 dolmens and 7 menhirs located near the caves; see fig. 1) and further studies of the Tagus and Zêzere regions allowed the formulation of new theoretical models (Figueiredo 2005; 2007; in press). The data obtained enabled these researchers to establish

154

Figueiredo – Oosterbeek, Two hidden landscapes in central Portugal…

connections between both regions and the different types of monuments from at least the middle Neolithic onward, and to start viewing this region as a dominant element of a broader society incorporated in a broad contact network and possessing diverse internal ritual practices. It was no longer viewed as the result of hidden landscapes (inaccessible areas or closed groups), but as a variety of social explorations and meanings. The rituals observed in the Rego da Murta dolmens (figs 2, 3) were similar to those observed in the caves and, although cults here have started at an earlier stage, these spaces continued to be used alongside the constructions and depositions observed further east in the dolmens. Also, the material culture seems to follow the same pattern observed in burial sites all over the Middle Tagus region and therefore can be included in the same social-symbolic system. The GIS analyses of the visibility, slope and appearance combined with the remains observed, some of them absolutely exogenous, either in the caves with the cardial presence or in the megalithic monuments with the record of external raw materials, allowed us to understand that these structures were closely

related with the society and imposed on landscape an extraordinary openness. It should be noted that dolmens, for instance, are located in areas of easier access. For this case, what was initially considered as a hidden landscape (due to the presence and effective use of caves without a predominant role in landscape marking and that seemed to be somewhat distinct from the patterns and rituals used in neighbouring regions, thus characterizing it as a closed space or at least having typical cultural traditions) is now revealed to be a matter of perception of the archaeological record. All elements man interacted with were an integral part of a structured landscape which was familiar to the community living in it and incorporated wellknown symbolic meanings that superseded material actions (Figueiredo 2006). As our research group proposed since 1985, caves should not be considered as different-meaning elements or as sites having a minor significance in space restructuring. Like megalithic monuments they act as symbolic items for communities, with an active role in the landscape and in the relationship with other people. In the words of Fontijn (2007), ‘invisible places could

Figure 2 – Plan and photographs of dolmen I of Rego da Murta.

155

Hidden Landscapes

have been just as important in ritual landscapes as the visible monuments that we usually study’.

3. The prehistoric art of Ocreza Valley The prehistoric art of the Ocreza Valley is another example of how the archaeological record interferes with our perception. In the 1970s, during the impact assessment survey for the construction of the Fratel dam, which would eventually submerge the Tagus rupestrian complex, a number of carved panels of relative interest have been identified in the Ocreza valley, a tributary of the Tagus. Further surveys thirty years later enabled us to identify not only more carvings in the Ocreza and another nearby valley (Carvalheiras), but also a number of red paintings (at Pego da Rainha, a spring in the Zimbreira stream, a tributary to the Ocreza). The carved representations on the banks of the Ocreza can be included in two sequences: the upper Palaeolithic, represented by figurative dotted gravures, some of them thread-like, the most important being a headless horse (fig. 4), and a later stage corresponding to recent prehistory (Early Neolithic or Chalcolithic and Early Bronze Age) represented by dotted and thread-like gravures, predominantly schematic, as well as the paintings men-

tioned above which, however, are found in higher, steep areas and sheltered zones (fig. 5). The artificial lakes resulting from the Pracana and Fratel dams led to the submersion of most rock art panels; only some are uncovered again each summer as the water level decreases. In addition, the difficult access caused by the area’s morphology (steep and abrupt slopes) renders survey of, and comparison between, these remains difficult. An incomplete understanding of the space and distribution of its cultural elements prevents an unequivocal interpretation. The presence of these elements in places of difficult access or even hidden by water hampers the use of common technologies, as for instance, in surveys where underwater archaeology methods have to be used to collect the maximum number of remains and obtain a more complete picture of these actions.

4. Conclusion Archaeology is concerned to a large extent with reconstructing past landscapes, i.e. perceptions of the territory projected by extinct communities, which is always a challenge. As a matter of fact, hidden landscapes at this level couldn’t exist because a landscape has to be known in order to be considered as such.

Figure 3 – Plan and photographs of dolmen II of Rego da Murta.

156

Figueiredo – Oosterbeek, Two hidden landscapes in central Portugal…

However, certain portions of the territory are likely to have been subject to differentiated and probably selected settlements. The fact that, from the Middle Neolithic, burial spaces (in caves or dolmens) tended to be organised in ‘necropoles’ outside the populated areas suggests the construction of a landscape where the land of the dead and the land of the living were dichotomized (in the fashion of Lévi-Strauss). In this context, certain spaces such as burial areas have a double role as hidden landscapes: possibly hidden from some members of those communities in case of limited access (the low, narrow entrances in some monuments suggest that at least their interiors would remain hidden from most community members) and, for sure, hidden from modern communities, who do not recognise them as specific spaces, i.e. who tend not to link them to significant sets of their landscapes/perceptions. Something similar happens with the rupestrian contexts in Vale do Tejo where a more accessible art engraved at the bottoms of the valleys (what we could call ‘public art’) contrasts with an art painted in exiguous shelters of difficult access. If the former can be hidden from the modern untrained eye, the latter is twice hidden since, due to its difficult access, only a few individuals can enjoy it at a time. In fact, the concept of hidden landscape seems to be more related to the individual who perceives it than to the object that is perceived. Considering a landscape as ‘hidden’ or ‘open’ from the prehistoric point of view is rather complex because we would have to perceive it with the forever lost mental framework of those past communities. We have nothing to rely on but a number of remains, which more or less conjugated, enable us to draw conclusions on these problems and require supplementary effort in the reconstruction of interpretations and perception of actions. Cult areas such as those containing megalithic art or monuments serve mostly as landmarks of an organized landscape. Even if they were intentionally hidden, they would still be known by society and integrated in the everyday practices of these populations. As we have been suggesting ‘the understanding of a space comprises everything existing therein, always having in mind the set of meanings within it’. From this perspective, the relationship established between man and landscape is always dynamic, intentional and complex. It is dynamic because space functions as a social construction in which man operates benefiting from its own properties or the properties he conferred to it in order to take social,

Figure 4 – Engraving of a headless horse. Palaeolithic, Vale do Ocreza.

Figure 5 – Painting in red of anchor motives. Neolithic, Pego da Rainha.

symbolic or economical advantage of all its constitutive elements. It is intentional because all actions are directed to the organization of that space, even if all elements within it are merely natural, since their meanings may be as important as those pertaining to any artificial architectural structure (Bradley 1998; Fowler & Cummings 2003). It is complex because it is subjective: each individual contribution (Ingold 2000, p. 179) interrelates with others and produces new concepts that shapes the landscape itself: the Dasein in Heidegger’s words (Inwood 2000, p. 22; Figueiredo 2006). Therefore, to consider any of these spaces as a hidden landscape, further tests will be necessary to confirm that this could have been the perspective of those ancestral communities.

References Bradley R. 1998. The significance of monuments: on the shaping of human experience in Neolithic and Bronze Age Europe. London, Routledge.

157

Hidden Landscapes

Cruz A.R. 1997. Vale do Nabão: do Neolítico à Idade do Bronze, ARKEOS 3, Perspectivas em diálogo, CEIPHAR, Tomar. Figueiredo A. 2005. Contributo para a análise do megalitismo no Alto Ribatejo. Complexo Megalítico do Rego da Murta, Alvaiázere, Almadan, II série, n. 13, pp. 134-136. Figueiredo A. 2006. Complexo megalítico de Rego da Murta. Pré-História recente do Alto Ribatejo (IV-IIº milénio a.C.): Problemáticas e Interrogações, PhD thesis, Faculdade de Letras, Universidade do Porto. Figueiredo A. 2007. Entre as grutas e os monumentos megalíticos: problemáticas e interrogações na pré-história recente no Alto Ribatejo, Almadan 15, electronic edition: www.almadan.publ.pt/15adendaIII.pdf. Figueiredo A. in press. Contributo para o estudo e compreensão do megalitismo no Alto Ribatejo: Anta I do Rego da Murta, Alvaiázere, Leiria. Actas do IV Congresso Peninsular de Arqueologia (2004). Fontijn J. 2007. The significance of “invisible” places, in Viewing Space, World Archaeology 39(1), pp. 70–83. Fowler C. & Cummings V. 2003. Places of transformation: building monuments from water and stone in the Neolithic of the Irish Sea. Journal of the Royal Anthropological Institute 9, pp. 1-21. Ingold T. 2000. The Perception of the Environment: Essays on Livelihood, Dwelling & Skill. Routledge, London. Llobera M. 1996. Exploring the topography of mind: GIS, social space and archaeology. Antiquity 70, pp. 612-622. Llobera M. 2001. Building Past Landscape Perception With GIS: Understanding Topographic Prominence. Journal of Archaeological Science 28, pp. 1005-1014. Oosterbeek L. 1985. A Facies Megalítica da Gruta do Cadaval (Tomar). Actas da 1ª Reunião do Quaternáro Ibérico, vol. II. Lisboa, Instituto Nacional de Investigação Científia. Oosterbeek L. 1994. Echoes from the East: The western network. North Ribatejo (Portugal): an insight to unequal and combined development, 7000–2000 B.C. PhD thesis, University College London. Oosterbeek L. 1995. Megalitismo e Necropolizaçao no Alto Ribatejo – o IIIº milénio. O Megalitismo no Centro de Portugal. Actas do Seminário. Mangualde, Centro de Estudos PréHistóricos da Beira Alta, pp. 137-149. Thomas J. 1993. The politics of vision and the archaeologies of landscape, in Bender B. (ed.), Landscape: Politics and Perspectives. Oxford, Berg, pp. 19-48. Thomas J. 1999. Time, Culture and Identity. An Interpretive Archaeology. London, Routledge. Tilley C. 1994. A Phenomenology of Landscape: Paths, Places and Monuments. Oxford, Berg.

158

Hidden Landscapes:

SEDIMENTARY BASINS

20 Landscape development of the coastal plains of Rome and Grosseto between 20,000 and 3,000 years ago Antonia Arnoldus-Huyzendveld Digiter s.r.l., via della Fortezza, 58, Rocca di Papa (RM), Italy – [email protected] Abstract This contribution is based on reference study in combination with field work as a geoarchaeologist. Its aim is to confront the landscape evolution of the closely related deltaic plains of Rome and Grosseto, in the light of interpreting the distribution of the known prehistorical sites on late Quaternary land units, and of formulating a probability pattern for the unknown ones. The two areas have in common their wave-dominated deltaic character and their position along the Tyrrhenian sea. This means the presence of one or more large rivers, a gently sloping sea floor, a tombolo type shoreline and a narrow tidal excursion. In the last 20,000 years these areas have undergone surprisingly strong landscape changes. Until 6,000 years ago, the landscape development and more specifically the position of the coastline, was regulated strictly by the rising post-glacial sea level. When the sea reached more or less its present level the landscape was modelled mainly by the opposing forces of river sediment load and strength of the marine currents. The coastal landscape still underwent strong changes in historical times, both of natural and anthropical origin, leading to the burial and erosion of former landscapes. Apart from the similarities, the differences between the two areas are analyzed. The landscape map of the Grosseto area is proposed as a useful thematic map for analyzing the distribution of prehistorical finds and for archaeological survey planning.

1. Introduction Concerning the distribution of prehistoric sites, we suppose a definite post-depositional bias, which through the use of earth sciences would be so predictable as to avoid gross interpretative errors. For that purpose, we will analyze in the first place the three layers of modern filters that hamper our vision of the prehistorical coastal landscape. In reverse chronological order these are 1) the events after 1950; 2) the land reclamations of the late 19th and 20th century; and 3) the landscape evolution of the last 3000 years. For the Rome area reference is made principally to an overview of the late Pleistocene and Holocene landscape evolution presented in Arnoldus-Huyzendveld (2005a), and for the Grosseto area to Arnoldus-Huyzendveld (2007 a, b). The present shoreline of the coastal plains of Rome and Grosseto is composed of tomboli or sandy coastal barriers, whose shape is influenced by rock outcrops and river outlets (fig. 1). The two most recent filters are best illustrated by comparing the 1823 Leopoldino cadastre of Grosseto with modern aerial photographs (see Caprasecca 2007). The general situation in the 19th century is separated from the present one by a series of decisive events. Specifically, several historical trends inverted in the 20th century, such as the expansion of the shoreline, the cutting of the forests and the relative stability of the sea level. Moreover, land reclamations have strongly transformed the wetlands of the coastal plains. Anthropical factors can

be considered the most significant in the recent well documented retreat of the Ombrone outlet (Bellotti et al. 1999), the drastic reduction of the sediment load of the fluvial system having been caused by an increase in forest cover, hydraulic works, and extraction of sand from the river bed. A subordinate cause would be the sea level increase of about 12-15 cm during the last century, related to the global temperature rise.

2. Changes of the landscape during the last 3,000 years The third filter concerns the major changes to the coastal landscape during the last 3,000 years

Figure 1 – The tombolo coast of the Grosseto plain, extending between Castiglione della Pescaia, the river Ombrone and the Uccellina Mountains.

161

Hidden Landscapes

(excluding the last century): the seawards expansion of the shoreline, increased river flooding, and the change from brackish to sweet water lagoons. Historical coastline advance has been reconstructed for Rome by Segre (Dragone et al. 1967; Segre 1986) and for the Grosseto plain by Bellotti et al. (1999; based partly upon Mori 1935). In both areas the shift was initially relatively slow, but became successively more rapid: about half the growth occurred from the 16th century AD onwards. For the Rome area, Bellotti et al. (1989) calculated an average shore line advance of 0.8 m/year for the period 540-1420 AD, and of 7.5 m/year between the 15th century and 1950. The intensification of the delta growth should be related to the increased sediment load of the rivers, caused by a change of land use in the catchments. Such a change is known to have occurred in Central Italy during the late Medieval and Renaissance period (Sereni 1987): ever steeper slopes were deforested and put under cultivation or pasture, causing strong erosion. A contribution to the increase in sediment load may have been provided by the low-

Figure 2 – Comparison of the historical coastline progradation of the Rome and Grosseto coastal plains with the frequency and level of the known major Tiber floods.

er average temperatures, heavier winter rains and lighter summer rains that occurred from the 13th century onwards. Such a change must have led not only to heavier rain impact on the soil, but also to less favourable conditions for a protective vegetation cover, and thus potentially to stronger erosion. The increase in fluvial sediment load meant not only a seawards migration of the beach, but also more intense flooding. It is not a coincidence that 16th century Rome was characterized by the highest floods registered in historical times (Di Martino & Belati 1980; Le Gall 2005). On the occasion of the severe flood of 1557 the Tiber meander near Ostia was naturally cut off, leaving behind a characteristic oxbow lake which still existed a hundred years ago, as is shown on the 1884 maps of Amenduni. The first documented flood in Grosseto, during which the Ombrone moved its course away from the city for about 1 km (Arnoldus-Huyzendveld 2005b), occurred in 1318 or 1333 (Venerosi Pesciolini 1925); other major documented floods occurred in 1557, 1758, 1944 and 1966. Figure 2 offers a graph confronting the registered Tiber floods with the historical coastline shift of Rome and Grosseto, with an evident high coincidence between the phenomena. Thus, together with the coastline advance, the valley floor levels of the coastal plains were raised through the deposition of flood sediments. At the same time the lagoons became smaller, shallower and less salty, due to the combined influence of the floods and the growth of the coastal barriers that isolated them more and more from the sea. For the Grosseto plain this development has been reconstructed for historical times (Arnoldus-Huyzendveld 2007b). The coastal barrier growth has not been completely unidirectional: along the roman coast Giraudi (2004) has recognized two major marine incursions during the last 3,000 years, the first between 910 and 800 BC and the second in the 1st century AD. For the Grosseto plain, phases of coastal retreat are considered by Mori (1935), Innocenti & Pranzini (1993), Stea (1995) and Bellotti et al. (2004). The environmental changes of the lagoons have induced a continual shift of the salt extraction industry towards the coast: in Grosseto from an area to the west of the town (probably Squartapaglia) first towards the Trappola (functioning from 1386 to 1758) and then to Castiglione della Pescaia. It should be stressed that the shallow tidal excursion of only 20-30 cm of the Tyrrhenian sea does not offer optimal conditions for salt extraction. For that reason, under a palaeo-environmental perspec-

162

Arnoldus-Huyzendveld, Landscape development of the coastal plains of Rome and Grosseto…

tive, the history of the Tyrrhenian salt pans can be read as a regular human adaptation to the changing landscape, in the search for lagoon borders in a position such as to allow the influx of salt water through natural ways or canals but at the same time free of inundation risk from the sea and rivers.

3. Landscape changes during pre- and protohistory Having considered the historical filters that hide the prehistorical landscape, we will now look at the landscape changes that occurred during preand protohistory. Our starting point will be 20,000 years ago, the cold peak of the Würm III glaciation, when the sea was about 125 m below the present level. This event is thought (although precise data are lacking for the areas in consideration) to have caused such strong inland erosion as to erase many former morphological features, whereas the successive sea level rise has laid the foundation for the present coastal morphology. The landscape evolution of this period is well documented. For the Rome coastal area we refer in particular to Belluomini et al. (1986), Bellotti (1992, 2000), and Bellotti et al. (1989, 1999). For the Grosseto area we refer to Motta (1969), Curri (1978), Bravetti & Pranzini (1987), Innocenti & Pranzini (1993), Bellotti (2000), Bellotti et al. (1999, 2001, 2004), Stea (1995), Stea & Tenerini (1996), and Biserni & van Geel (2005). Some publications refer specifically to the Mediterranean sea level variations in this period, partly based upon archaeological data (Antonioli & Leoni 1998, Lambeck et al. 2004). The publication most directly related to our theme is the morpho-sedimentary model of the major Italian Tyrrhenian deltas - the Volturno, Tiber, Ombrone and Arno plains - as proposed by Bellotti (2000). According to Bellotti (1994, 2000, Bellotti et al. 2004) the evolution of the Tyrrhenian coastal plains over the last 20,000 years was guided by two factors: the glacio-eustatic rising of the sea level with higher temperatures (melting of polar glaciers; expansion of sea water) and the position of the main river with respect to the coastal barrier. Three evolutionary phases are distinguished: 17,000-6000 years ago; 6000-3000 years ago; and afterwards. During the first phase (see fig. 3, referring to the Tiber delta) the river outlet was inside the lagoon and the sea level rose with a discontinuous and ever decreasing rhythm. The fluvial sediments were not able to compensate the volumetric increase of the lagoon induced by the sea level rise, hence the coastal

Figure 3 – The development of the Tiber delta between 12,000 and 6000 years ago, modified from Bellotti (1994). Arrows indicate the direction of coastline shifting.

barrier system and the lagoon migrated landwards. It is worthwhile realizing that in this period the average sea level rise was close to 1 meter per century. At the beginning of the second phase (6000 years BP), the sea level was almost stable and close to the present level, though still rising very slowly. The great rivers (Tiber, Ombrone) started constructing a delta within the lagoon, thus splitting it up into several parts. A continuous peat layer marking a stagnation in the sea level rise (Morzadec-Kerfourn 2005) was deposited in the Rome plain, and is now found at an average depth of 4 m. According to Bellotti’s model at the end of this phase, about 3000 BP, the main rivers reached open sea: their depocenter had moved definitively from within the lagoon to the open sea.

4. The Rome and Grosseto plains in the final Pleistocene - Holocene period Figure 4 offers a reconstruction of the Grosseto plain about 5000 years ago, designed by Studio Inklink in preparation for the panel exhibited since 2004 in the MuseoLab of the University of Grosseto. One notices the two rivers that have an outlet in the lagoon, the Ombrone and the Bruna. It is an essential point that the Ombrone river was large enough to reach the seashore about 3000 years ago (which is according to Bellotti’s model) but that

163

Hidden Landscapes

Figure 4 – A 3D reconstruction of the Grosseto plain about 5000 years ago, with a large and rather deep salt water lagoon and a weakly developed discontinuous coastal barrier. Legend: G) location of Grosseto; B) Bruna river; O) Ombrone river; U) Uccellina mountains. Design Studio Inklink – Florence.

the smaller Bruna river apparently never succeeded in doing that. Precisely this is considered the main reason for the persistence of the Prile or Castiglione lake (see fig. 5); a secondary reason may be differential subsidence (Innocenti & Pranzini 1993). It is also important that in the Grosseto plain there are two or three younger Pleistocene relicts extending into the lagoon, and they are radially developed. Thus, several lagoon partitions were created by these former ridges. The Pleistocene terraces not only extend into the lagoon but also follow the river valleys land inwards. Several levels are distinguished: the early Quaternary relicts overlying tertiary formations (indicated as on the geological map of 1969), the late Quaternary terraces (), distributed on at least two different levels, and the most recent ones () which, according to Bravetti & Pranzini 1987, also date to the Pleistocene and not to the Holocene as argued by Motta (1969). The situation is different for the Rome coastal plain. Here we have one large river, the Tiber, which has divided the original lagoon into two parts, the northern Stagno di Maccarese and the southern Stagno di Ostia. The Arrone and Galeria rivers had a minor role, forming small alluvial fans. The late Pleistocene terraces developed parallel to the coastline, whereas terraces along the river valleys are completely lacking. Older terrace levels are interwoven with the middle Pleistocene volcanic deposits of the Sabatino and Latium centres, though partially

remaining parallel to the coastline (Giordano et al. 2003). The best preserved coastal terrace north of the Tiber, at about 30 m a.s.l., is generally assigned to isotopic stage 5 (or 5e; about 125,000 years BP; Conato et al. 1980). There is a discussion about the existence and the number of younger terraces north of the Tiber. Belluomini et al. (1986) recognize three terrace levels between 2 and 12 m a.s.l., ArnoldusHuyzendveld et al. (1991) propose two different levels, and Giraudi (2002) recognizes at least one level, whereas the geological map of 1969 (sheet 149 Cerveteri) shows no lower terraces at all. South of the Tiber the isotopic stage 5 terrace is also present, as well as the stage 7 and 9 terraces (Giordano et al. 2003) and traces of younger terraces, though these are partly covered with recent dune sands (Dai Pra & Arnoldus-Huyzendveld 1984). Finally, only in the area of Rome there has been recent volcanic activity at least between 6000 and 2400 years BP, originating from the Albano crater (Vulcano Laziale) in the form of mud streams or lahar (Funiciello et al. 2003; Gioia 2008). We still do not know the influence of this activity on the coastal area, if ever there has been. The mud streams and related deposits may have reached the coastal plain through the river system. Anyway, there are clear traces of recent hydrothermal activity, linked to the final activity of the Latium Volcano, in the Tiber valley near the coast (Tuccimei et al. 2007, ArnoldusHuyzendveld et al. 2009).

164

Arnoldus-Huyzendveld, Landscape development of the coastal plains of Rome and Grosseto…

We can now summarize the main evolutionary differences between the Rome and Grosseto plains for the final Pleistocene–Holocene period as follows: 1. Grosseto: two rivers feeding the lagoon: one larger, one smaller; Rome: one main river; 2. Grosseto: partitioning of the lagoon in at least three compartments by the late Pleistocene relicts; Rome, no partitioning in that sense; 3. Grosseto: Pleistocene terraces all around the lagoon, extending landwards into the river valleys and as relicts upon the tertiary reliefs; Rome: late Pleistocene terraces parallel to the coastline, along the internal lagoon borders; older terraces interrelated with middle Pleistocene volcanic deposits, though still partially parallel to the coastline; almost complete lack of tertiary reliefs; 4. Rome: potential influence of recent volcanic activity in the coastal plain, although up to now only hydrothermal traces are found.

bution of soils in their lithological and geomorphological context, with less detail than the original soil maps. In compensation, this kind of map contains more information on other landscape characteristics, since Land is a wider concept than Soil, and therefore can be considered more useful for a series of applications, among which landscape archaeology and potential land evaluation (for the latter see

5. The Land Units Map of the Grosseto area For the Grosseto area there is available a Landscape Map in scale 1:50,000, covering an area of about 800 km2 (Arnoldus-Huyzendveld 2007b). This map is derived from two soil maps: Sevink et al. (1986) and Van Berghem et al. (1991). Generally, a Landscape Map or Land Units Map describes the spatial distri-

Figure 5 – Historical map by Pallari 1781, with the bipartition of the Grosseto plain still existing at that time: a southern part under the influence of the Ombrone river and a northern part under the (less potent) influence of the Bruna river, leaving space for the Castiglione Lake. Near the Ombrone outlet are visible the salt pans of the Trappola and the channel connecting them to the sea. Map north towards one o’clock.

Code

Description

A

valley floors, river terraces

B

Ombrone valley and reclaimed plain

C

dunes, coastal terrace

D

recent and older alluvial fans

E

terrace of Grosseto

F

travertine plain (part of the Grosseto terrace)

G

recent and older colluvial deposits

H

conglomerates, limestone, marls, clay, sandstone (Macigno), schists (Verrucano)

I

rhyolite and ultra basic rocks

Table 1 – Systems of the Landscape Map of Grosseto

Figure 6 – Landscape map of Grosseto divided according to the main Systems. Legend, see text above; systems A and B are Holocene, systems D, E, F, G mainly Pleistocene, systems H and I are older; arrows indicate the Pleistocene ridges extending into the former lagoon; 1) partition of the lagoon delimited to the south by the ridge of the Canale Diversivo / Marruccheto; 2) idem, delimited to the east by the Pleistocene ridge of Querciolo / Poggiale; both 1 and 2 mainly under the influence of the Bruna river; 3) idem, part of the lagoon under the influence of the Ombrone river.

165

Hidden Landscapes

Attema et al. 2002; Van Joolen 2003). In the present context it appears to be the best map available for interpreting the distribution of prehistorical sites (and non-sites) in relation to the environment and the physical landscape, and for the derivation of archaeological probability maps. The criteria generally applied in the creation of a Landscape Map depend on the characteristics of the area and on the aims. For the Grosseto area, the selected criteria have been: 5. drainage class, a parameter that expresses, in the case of a flat or undulating landscape, the combined effect of texture, permeability and groundwater influence, and which can be interpreted in terms of agricultural productivity, terrain accessibility and risk of flooding; 6. slope class; 7. physiography and/or lithology. Nine Land Systems have been distinguished (see figure 6 and table 1). Each Land System has been divided in Subsystems and then in Land Units, for

a total of 46 map units. In figure 7 is given an extract of the map for the area south of the town of Grosseto. For a more detailed legend, see ArnoldusHuyzendveld 2007b.

6. Conclusions Now we can have a look at the distribution of prehistoric finds in both coastal areas, referred to the period between 20,000 and 3000 years BP. In the first place, from 17,000 to 6000 years BP all the lagoonal areas have been subject to quick sedimentation, afterwards slowing down but still active in historical times. Therefore all traces of prehistoric frequentation along the contemporary lagoon margins are supposedly buried at the depth of the corresponding sea levels. Secondly, most Pleistocene terraces have been exposed during that time span and therefore mainly run a risk of erosion (removal) of the artefacts; however, middle Palaeolithic flint was found on the late Pleistocene terraces in the Rome area (Bietti Sestieri &

Figure 7 – Extract of the Landscape Map of Grosseto, divided by Land Units. Legend: A2) valley floors, flat, moderately well drained; B1) Ombrone valley and reclaimed plains, slightly excessively drained; B2) idem, well drained; B3) idem, moderately well drained; B4) idem, imperfectly drained; C1) dunes, excessively drained; D3) recent (Pleistocene) alluvial fans, undulating, moderately well drained; D4) idem, hilly; G3) colluvial deposits, undulating, imperfectly drained; G4) idem, well drained; H3) limestone, hilly; H6) clays, undulating, well drained; H8) sandstone - Macigno -, undulating, well drained; H9) idem, hilly; H10) schist - Verrucano - , hilly.

166

Arnoldus-Huyzendveld, Landscape development of the coastal plains of Rome and Grosseto…

Figure 8 – The three levels of late Pleistocene terraces north of the Tiber, according to the Soil Map of Rome (Comune di Roma 1981; Arnoldus-Huyzendveld 2003); T1 is the terrace assigned to isotopic stage 5, age about 125,000 years, T2 and T3 are younger terraces; white dots refer to Middle Palaeolithic finds, black dots symbolize Neolithic to Middle Bronze age sites.

Figure 9 – Grosseto, physiographical situations analogous to the Rome late Pleistocene terraces with prehistoric finds. The Grosseto terrace (E, F) is delimited by a continuous black line; Pleistocene terraces are indicated with the letters D, E, F.; circles indicate schematically the presumably best preserved parts of these, and dotted lines the terrace margins towards the lagoonal infill.

Sebastiani 1986; Arnoldus-Huyzendveld et al. 1993). The richest areas for recent prehistorical sites (from Neolithic to Middle Bronze) are the margins of the lower terrace north of the Tiber, which are just sufficiently buried by lagoonal deposits to prevent disturbance by modern ploughing (see fig. 8; ArnoldusHuyzendveld et al. 1998; Giraudi 2002; Manfredini 2002). In figure 9 are proposed the comparable physiographical situations in the Landscape of Grosseto, i.e. the presumably best preserved coastal terrace surfaces and their margins towards the lagoon. The scheme given in figure 9 does not pretend to be anything other than a guideline for future field survey in the coastal area of Grosseto, to be tested and corrected. Neither does the proposed model refer to prehistoric finds in caves or on older parts of the landscape, or to the Lower Palaeolithic (for an overview of prehistorical finds around Grosseto see Balducci et al. 2007).

References Amenduni G. 1884. Sulle Opere di Bonificazione della Plaga litoranea dell’Agro Romano. Roma, Tipografia Eredi Botta. Antonioli F. & Leoni G. 1998. Siti archeologici sommersi e loro utilizzazione quali indicatori per lo studio delle variazioni recenti del livello del mare. Il Quaternario, Italian Journal of Quaternary sciences 11(1), pp. 53-66. Arnoldus-Huyzendveld A. 2007a. La carta delle Land Units del centro storico di Grosseto, ed alcune delle sue elaborazioni

167

in chiave archeologica, in Citter C. (ed.), Archeologia urbana a Grosseto. Origine e sviluppo di una città medievale nella “Toscana delle città deboli”. Le ricerche 1997 – 2005. Vol. II: edizione degli scavi urbani 1998-2005. Dall’insediamento stabile al grande cantiere mediceo (ca. 600 – 1625), All’Insegna del Giglio, Firenze, pp. 41-61. Arnoldus-Huyzendveld A. 2007b. Le trasformazioni dell’ambiente naturale della pianura grossetana, in Citter C & Arnoldus-Huyzendveld A. (eds), Archeologia urbana a Grosseto. Origine e sviluppo di una città medievale nella “Toscana delle città deboli”. Le ricerche 1997 – 2005. Vol. I: la città di Grosseto nel contesto geografico della bassa valle dell’Ombrone, All’Insegna del Giglio, Firenze, pp. 62-71. Arnoldus-Huyzendveld A. 2003. I suoli di Roma, due passi sulle terre della città; Carta dei Suoli del Comune di Roma in scala 1:50.000 con Note Illustrative. Comune di Roma, Dip. to X – IV° U.O. Sviluppo Sostenibile. Arnoldus-Huyzendveld A., Crovato C. & Zarlenga F. 1991. Analisi paleoambientale dei depositi “intrawurmiani” ed olocenici della Piana di Maccarese. ENEA/ RT/ AMB/91/26. Arnoldus-Huyzendveld A., Gioia P., Mussi M., Toniutti P., Zampetti D. & Zarlenga F. 1993. Paleosuperficie esposta, concentrazione di industria litica e stratigrafia nella piana costiera di Maccarese (Roma). Il Quaternario 6(2), pp. 213-226. Arnoldus-Huyzendveld A., Morelli C., Carbonara A. & Ceracchi C. 2009. La viabilità nel territorio portuense, Atti del Convegno Suburbium II – Il Suburbio di Roma dalla fine dell’età monarchica alla nascita del sistema delle ville (V-II sec. a.C.), Roma 2005, pp. 599-619. Arnoldus-Huyzendveld A., Pascucci P. & Mineo M.. 1998. Il sito costiero dell’età del Bronzo di Le Cerquete-Olivetello (Fiumicino). Bollettino della Comissione Archeologica Comunale di Roma XCIX, pp. 393-411.

Hidden Landscapes

Arnoldus-Huyzendveld A. 2005a. The natural environment of the Agro Portuense, in Keay S., Millett M., Paroli L. & Strutt K. (eds), Portus, an archaeological survey of the port of imperial Rome, the British School at Rome, pp. 14-30. Arnoldus-Huyzendveld A. 2005b. Il paleoambiente storico di Grosseto, in Citter C. (ed.), Lo scavo della Chiesa di S. Pietro a Grosseto, Nuovi dati sull’origine e lo sviluppo di una città medievale, All’Insegna del Giglio, Firenze 2005, pp. 59-68. Attema P., Burgers G., van Joolen E., van Leusen P.M. & Mater B. (eds) 2002. New Developments in Italian Landscape Archaeology (BAR International Series 1091). Balducci C., Lo Vetro D. & Volante N. 2007. La preistoria nel territorio di Grosseto, in Citter C. & Arnoldus-Huyzendveld A. (eds), Archeologia urbana a Grosseto. Origine e sviluppo di una città medievale nella “Toscana delle città deboli”. Le ricerche 1997 – 2005. Vol. I: la città di Grosseto nel contesto geografico della bassa valle dell’Ombrone, All’Insegna del Giglio, Firenze, pp. 119-130. Bellotti P. 1994. Sedimentologia ed evoluzione olocenica della laguna costiera presente un tempo alla foce del Tevere. Atti del 10° Congresso A.I.O.L., Alassia 4-6 novembre 1992, pp. 633-642. Bellotti P. 2000. Il modello morfo-sedimentario dei maggiori delta tirrenici italiani. Boll. Soc. Geol. It. 119, pp. 777-792. Bellotti P., Belluomini G., Bergamin L., Carboni M. G., Di Bella L., Improta S., Letunova P. P., Manfra L., Potemkina T. G., Valeri P. & Vesica P. 2001. Nuovi dati cronostratigrafici sul sottosuolo della piana deltizia del Fiume Ombrone (Toscana meridionale), Studi Costieri 4, pp. 31-40. Bellotti P., Caputo C., Davoli L., Evangelista S. & Valeri P. 1999. Lineamenti morfologici e sedimentologici della piana deltizia del Fiume Ombrone (Toscana Meridionale). Boll. Soc. Geol. It. 118, pp. 141-148. Bellotti P., Caputo C., Davoli L., Evangelista S., Garzanti E., Pugliese F. & Valeri P. 2004. Morpho-sedimentary characteristics and Holocene evolution of the emergent part of the Ombrone River Delta (southern Tuscany). Geomorphology 61, pp. 71-90. Bellotti P., Carboni M.G., Milli S., Tortora P. & Valeri P. 1989. La Piana deltizia del Tevere: analisi di facies ed ipotesi evolutiva dall’ultimo low stand glaciale all’attuale. Giornale di Geologia 51, pp. 71-91. Belluomini G., Iuzzolini P., Manfra L., Mortari R. & Zalaffi M. 1986. Evoluzione recente del delta del Tevere. Geologica Romana 25, pp. 213-234. Bietti Sestieri A.M. & Sebastiani R. (eds) 1986. Preistoria e Protostoria nel territorio di Roma. Modelli di insediamento e vie di comunicazione, in Il Tevere e le altre vie d’acqua del Lazio antico, Archeologia Laziale VII, pp. 30-70. Biserni G. & van Geel B. 2005. Reconstruction of Holocene palaeoenvironment and sedimentation history of the Ombrone alluvial plain (South Tuscany, Italy). Review of Palaeobotany and Palynology 136, pp. 16-28. Bravetti L. & Pranzini G. 1987. L’evoluzione quaternaria della pianura di Grosseto (Toscana): prima interpretazione dei dati del sottosuolo. Geografia Fisica e Dinamica Quaternaria 10 (1987), pp. 85-92. Caprasecca A. 2007. Il contributo dell’archeologia alla ricostruzione della storia della città e del territorio, in Citter C. & Arnoldus-Huyzendveld A. (eds), Archeologia urbana

a Grosseto. Origine e sviluppo di una città medievale nella “Toscana delle città deboli”. Le ricerche 1997 – 2005. Vol. I: la città di Grosseto nel contesto geografico della bassa valle dell’Ombrone, All’Insegna del Giglio, Firenze, pp. 109-119. Comune di Roma, Assessorato per gli interventi in Agricoltura, Ripartizione XIII, 1981. Carta dei Suoli 1:50.000. Cooperativa C.U.P. Roma. Litostampa Nomentana, Roma. Conato V., Esu D., Malatesta A. & Zarlenga F. 1980. New data on the Pleistocene of Rome. Quaternaria 22, pp. 131-176. Curri G.B. 1978. Forma Italiae: Regio VII, volumen IV: Vetulonia, Firenze. Dai Pra G. & Arnoldus-Huyzendveld A. 1984. Lineamenti stratigrafici, morfologici e pedologici della fascia costiera dal Fiume Tevere al Fiume Astura (Lazio). Geologica Romana 23, pp. 1-12. Di Martino V. & Belati M. 1980. Qui arrivò il Tevere. Le inondazioni del Tevere nelle testimonianze e nei record storici. Multigrafica Editrice, Roma Dragone F., Maino A., Malatesta A. & Segre A.G. 1967. Note illustrative della C.G.I. alla scala 1:100.000. Foglio 149 (Cerveteri). Ministero dell’Industria, del Commercio e dell’Artigianato, Direzione Generale delle Miniere, Servizio Geologico d’Italia. Funiciello R., Giordano G. & De Rita D. 2003. The Albano maar lake (Colli Albani Volcano, Italy): recent volcanic activity and evidence of pre-Roman Age catastropic lahar events. Journal of Volcanology and Geothermal Research 123, pp. 43-61. Gioia P. (ed.), 2008. Torre Spaccata, Roma S.D.O. Le indagini archeologiche, Rubettino, Roma-Soveria Mannelli. Giordano G., Esposito A., De Rita D., Fabbri M., Mazzini I., Trigari A., Rosa C. & Funiciello R. 2003. The sedimentation along the roman coast between middle and upper Pleistocene: the interplay of eustatism, tectonics and volcanism new data and review. Il Quaternario, Italian Journal of Quaternary Sciences 16(1Bis), pp. 121-129. Giraudi C. 2002. Il territorio, in Manfredini A. (ed.), Le dune, il lago, il mare, una comunità di villaggio dell’età del Rame a Maccarese, pp. 25-35, Origines, Roma. Giraudi C. 2004. Evoluzione tardo-olocenica del delta del Tevere. Il Quaternario, Italian Journal of Quaternary Sciences 17(2/2), pp. 477-492. Innocenti L. & Pranzini E. 1993. Geomorphological Evolution and Sedimentology of the Ombrone River Delta, Italy. Journal of Coastal Research 9(2), pp. 481-493. Lambeck K., Antonioli F., Purcell A. & Silenzi S. 2004. Sea-level change along the Italian coast for the past 10,000 yr. Quaternary Science Reviews 23, pp. 1567-1598. Le Gall J. 2005. Il Tevere fiume di Rome nell’antichità. Mocchegiano Carpano C. & Pisani Sartorio G. (eds), Quasar. Manfredini A. (ed.) 2002. Le dune, il lago, il mare, una comunità di villaggio dell’età del Rame a Maccarese, Origines, Roma. Mori A. 1935. L’evoluzione della costa grossetana dal Pliocene ai giorni nostri. Annuario 1932-1935 del R. Liceo-Ginnasio di Grosseto. La Maremma, Grosseto. Morzadec-Kerfourn M.-T. 2005. Interaction between sea-level changes and the development of littoral herbaceous vegetation and eutrophic dinoflagellates. Quaternary International 133-134, pp. 137-140. Motta S. 1969. Note Illustrative della C. G. I. d’Italia alla scala 1:100.000, foglio 128, Grosseto. Ministero dell’Industria,

168

Arnoldus-Huyzendveld, Landscape development of the coastal plains of Rome and Grosseto…

del Commercio e dell’Artigianato, Direzione Generale delle Miniere, Servizio Geologico d’Italia. Segre A.G. 1986. Considerazioni sul Tevere e sull’Aniene nel Quaternario. Quaderni del Centro di Studio per l’Archeologia etrusco-italica 12: Il Tevere e le altre vie d’ acqua del Lazio Antico; CNR, pp. 9-17. Sereni E. 1974. Storia del paesaggio agrario italiano. Biblioteca Universale Laterza, no. 69. Servizio Geologico d’Italia 1963. Carta Geologica d’Italia alla scala 1:100.000, Foglio 149 Cerveteri. II edition. Servizio Geologico d’Italia 1969. Carta Geologica d’Italia alla scala 1:100.000, Foglio 128 Grosseto. II edition. Sevink J., Beemster T. & van Stiphout T. 1986. Soil Survey and Land Evaluation of the Grosseto area. Rapporto interno dell’Università di Amsterdam per la Regione Toscana. Stea B. & Tenerini I. 1996. L’ambiente naturale della pianura grossetana e la sua evoluzione dalla preistoria alla cartografia rinascimentale, in Citter C. (ed.), Grosseto, Roselle e il Prile. Mantova. pp. 13-24. Stea B. 1995. Il forte di San Rocco: l’ambiente geologico, in Bueti S. (ed.), Il forte di San Rocco. Una struttura militare nel litorale maremmano del sec. XVIII, Grosseto, pp. 113-125. Tuccimei P., Soligo M., Arnoldus-Huyzendveld A., Morelli C., Carbonara A., Tedeschi M. & Giordano G. 2007. Datazione U/Th di depositi carbonatici intercalati ai resti della via Portuense antica (Ponte Galeria, Roma): attribuzione storico-archeologica della strada e documentazione cronologica dell’attività idrotermale del fondovalle tiberino. www.fastionline.org/docs/FOLDER-it-2007-97.pdf. Van Berghem J., Meijvogel T. & Windmeijer P.N. 1991. La Valle dell’Albegna, Pedologia e valutazione del territorio. Giunta regionale toscana, Marsilio Editori. Van Joolen E. 2003. Archaeological land evaluation. A reconstruction of the suitability of ancient landscapes for various land uses in Italy focused on the first millennium BC. PhD thesis, Rijksuniversiteit Groningen, Netherlands. Venerosi Pesciolini G. 1925. Mura e casseri di Grosseto nell’Evo Medio, Maremma. Bollettino della Società Storica Maremmana II, pp. 215-236.

169

21 20,000 Years of landscape evolution at Ca’ Tron (Venice, Italy): palaeoenvironment, archaeology, virtual reality webgis Paolo MozziI, Aldino BondesanI, Maria Stella BusanaII, Paolo KirschnerII, Antonella MiolaIII, Sofia PescarinIV, Maria Cristina VillaniIII I Dipartimento di Geografia, Università di Padova, via del Santo 26, 35123 Padova, Italy – [email protected]; [email protected] II Dipartimento di Archeologia, Università di Padova, piazza Capitaniato 7, 35139 Padova, Italy – [email protected]; [email protected] III Dipartimento di Biologia, Università di Padova, via Ugo Bassi 58/B, 35100 Padova, Italy – [email protected]; [email protected] IV VHLab CNR ITABC, via Salaria, km 29,300 Monterotondo St. Roma, Italy – [email protected] Abstract Ca’ Tron is located in the lower part of the Venetian plain, at the inner margin of the Lagoon of Venice. Archaeological evidence in the 11 sq.km -large study area ranges from the Late Mesolithic/Ancient Neolithic Age to the Middle Ages. The most prominent archaeological feature is a stretch of the Roman via Annia, which was built in the 2nd century BC, but which in the study area may have followed an earlier, protohistoric track. The geoarchaeological research project, started in 2000, aims to understand how past human settlements and infrastructures interacted with the dynamic environmental conditions of this coastal alluvial plain. The research strategy of the multidisciplinary investigation is presented. Methods of data acquisition range from archaeological survey and excavation to pollen analysis, remote sensing, geomorphological mapping, sedimentology and palaeopedology. The reconstruction of landscapes in periods of interest (e.g. Last Glacial Maximum, Bronze Age, Iron Age, Roman Age, Middle Ages) is discussed, as part of an ongoing attempt to use palaeoenvironmental data to build robust and immersive web-based 3D virtualreality landscape reconstructions.

1. Introduction Since the year 2000 a multidisciplinary team of archaeologists, geologists and palynologists has been carrying out joint investigations at Ca’ Tron, an 11 sq.km estate owned by the Fondazione Cassamarca. A major aim of these investigations has been the reconstruction of the palaeoenvironmental evolution during the last 20,000 years of this area, located in the low Venetian alluvial plain at the inner margin of the lagoon of Venice (fig. 1). The most prominent archaeological feature is a stretch of the Roman via Annia, the consular road to the Latin Colony of Aquileia. It was constructed in the 2nd century BC, but the earliest archaeological finds date back to the Late Mesolithic/Early Neolithic. The analysis of the response of ancient human activities to changing environmental conditions in this critical coastal setting has been another major goal (Ghedini et al. 2002). The research carried out so far indicates that a major environmental change took place between the end of the last glaciation and the beginning of the Holocene, due to the adjustment of the fluvial system and of the vegetation to temperate climatic conditions (Bondesan et al. 2004; Fontana et al. 2008). Another important change was related to the Holocene sea-level rise, which led to the formation, after about 6000 BP, of the lagoon of

Venice (Serandrei-Barbero et al. 2005; Amorosi et al. 2008). Changes in the direction and morphology of minor groundwater-fed rivers crossing the study area, as well as continuous land subsidence, also led to important variations of the local environmental conditions. Furthermore, pollen records indicate that human activities have induced changes in the late Holocene vegetation at least since the Iron Age (Miola & Valentini 2004).

Figure 1 – Location of the study area.

171

Hidden Landscapes

The possibility of integrating available palaeoenvironmental and archaeological data in time-frame reconstructions of Ca’ Tron landscapes has been discussed in Mozzi et al. (2005). More recently, new data have been gathered. Since 2004, a systematic survey and new excavations took place inside the Ca’ Tron estate, with the specific purpose of understanding land use and settlement distribution in the past (Busana 2005; 2007; 2008a; 2008b). These studies have for the first time produced a realistic, even if not complete, estimate of territories and their populations in an area that in the past has produced only little evidence (Busana 2002). The present paper represents a step forward in the integrated analysis of palaeoenvironmental and archaeological data for the reconstruction of ancient landscapes at Ca’ Tron. Following an updated overview of the archaeological information available in the area, problems concerning the determination of changes in the physical landscape are discussed, with a focus on the crucial task of deriving the spatial distribution of vegetation types from pollen records. A first attempt to use the geoarchaeological dataset for building 3-dimensional, immersive, web-based virtual-reality reconstructions is also presented. At present there is a clear understanding of the potentiality of virtual reality systems for research purposes, and not only for cultural heritage information and communication (Forte & Campana 2006). This is particularly true when dealing with archaeological landscapes. Their complexity, due to their multidimensionality and also to the dynamism of processes and relations, requires their treatment as ‘virtual ecosystems’. In fact, only with a VR realtime and three-dimensional approach it is possible to manage multidimensionality, changes, connectivity etc. Moreover, the use of open-source software and of an ‘open’ methodology allows a clear and transparent interpretation and validation of ancientlandscape reconstructions (Pescarin 2006; 2007). The goal of the virtual reality approach in this case is therefore to enable the analysis, the processing and the interpretation of the Ca’ Tron archaeological landscape; it is not limited to the typical purposes of scientific communication and education. On the basis of all studies five specific periods have been identified, which significantly differ in geomorphology, vegetation and archaeology. These five time-frames are: 1. the Last Glacial Maximum (LGM), of about 20,000 BP 2. the Bronze Age (2nd millennium BC)

3. the Iron Age (8th century BC) 4. the Roman period (1st century AD) 5. the late Medieval period (14th–15th century AD) Here, only the landscape reconstructions associated with the first two time-frames are presented, as work on the others is still in progress. The decision to start with the earliest time-frames respects the concept of landscape evolution through time, allowing for the use of inherited landscape features in the development of reconstructions for the later time-frames.

2. Archaeology of Ca’ Tron: general background and new data 2.1 Methods

The first archaeological research projects in the Ca’ Tron area were focused on the via Annia, which crosses the south-east part of the territory. Air photo interpretation showed two branches of the consular road, one by the coast near the lagoon and the other further inland. These two branches have been investigated by means of ground surveys and trenches. These trenches extended as far as 50 m and involved the road, the two side ditches and a part of the surrounding area, reaching the bottom of the ditches. Their purpose was to understand the dimensional and technical features of the routes and to place them in a chronological framework. Two large trenches were opened at the point where the inner and the outer routes crossed the Canna palaeochannel, in order to determine the connection between the palaeoenvironmental aspects and the two roads and to identify the crossing points in the ancient watercourse. From 2004 a systematic survey has taken place inside the Ca’ Tron estate, and archaeological trenches were opened at points where remains had been identified, in order to understand the past types and forms of land use and occupation of the area (Busana 2007). The gridded survey has so far covered an area of 786 ha, using the modern field pattern as a reference. 81% of this area corresponded with zones of high (ploughed soils) or medium (uncultivated soils) visibility. Each field (about 30 × 200 m) was traversed four times in straight lines, and a file was created in which were registered the topographic coordinates, the soil visibility relative to land use (uncultivated, ploughed, trimmed) and also meteorological conditions, geomorphic features, anthropic anomalies detected during the preceding remote sensing analysis, and the collected artefacts. Identified sites were demarcated on the basis of the areas of concentration

172

Mozzi et alii, 20,000 Years of landscape evolution at Ca’ Tron (Venice, Italy)…

and dispersion of the archaeological materials, photographed and georeferenced. Next, the total number of artefacts was calculated, and the more important ones were individually georeferenced and picked up. In fields that produced only sporadic materials, these were georeferenced and collected. The study of the artefacts allowed us to establish some macrophases, assisted also by archaeological excavations at three different sites: one prehistoric, one Roman and one post-Roman, the latter documented by Renaissance pottery. 2.2 Results 2.2.1 Pre- and protohistory

The prehistoric phase is attested by the find of twelve worked flints found at separate locations in the central and south-west zones. Very interesting are particularly four flintstones from adjacent fields in the south-west zone, not far from the Sile river. Their significance and chronological context are not yet clear: they are waste material, except for a blade dated to the Late Mesolithic/Ancient Neolithic Age that was found near the Canna palaeochannel. More reliable data have come from archaeological excavation. During the early phase of the Middle Neolithic a settlement probably existed near the north bank of the Canna palaeochannel, as is suggested by some worked flint (a scraper and a sickle) and some wells or underground silos, probably connected to storage activities. Fragments of ceramic vessels that provide further information about pre- and protohistoric human presence in the area were also found in the Canna palaeochannel. These fragments are older than the 7th century BC, because of the absence of turning grooves. The investigations (Basso et al. 2004; Mozzi et al. 2005) along this route, which probably corresponds to the first Roman consular road, have revealed that it was constructed on top of a protohistoric track possibly dating back to the end of the Bronze Age (13th-10th century BC), and used continually during the Iron Age (9th-5th century BC). This is suggested by the radiocarbon dates of two wood fragments from the remains of a wooden bridge that crossed the Canna palaeochannel (Martinelli 2004). The construction of the first via Annia obliterated this pre-Roman track, which may have resembled the Bronze Age road of the Valli Grandi Veronesi, which was a straight 6-7 m wide track on a sandy embankment, lined by ditches about 3 m wide (Balista et al. 2005). The preliminary state of our researches makes the reconstruction of the orig-

inal features of the pre-Roman bridge and its subsequent transformations only hypothetical. 2.2.2 Roman period

The first Roman consular road dates from the middle of the 2nd century BC. It was a dirt road, 70 Roman feet (about 21 m) wide at ground level, lined by two ditches of respectively 4.80 and 7 m wide and 0.75 and 0.55 m deep. Between the end of the 1st century BC and the beginning of the 1st century AD, the road and the bridge were covered by lagoonal muds as a consequence of the relative rise of the sea level (Basso et al. 2004; Bondesan et al. 2004). Therefore the Romans built a new route on a dry location further inland, adopting new techniques (a 17 m wide foundation, on an embankment, with a roadbed made of cobblestone and solid, durable gravel, lined by ditches 9 m wide and 1.5 m deep). A new bridge over the Canna palaeochannel was built as well, made of stone and bricks on wooden foundations, with a width of 6 m, a total length of 9.2 m, a span of 2 m, and a related ford upstream (Basso et al. 2004). The collaboration with researchers of the Engineering Faculty of Università di Udine produced a 3D reconstruction of this Roman bridge, on the basis of planimetric and structural data and an estimation of the curve of the (depressed) arch and of the thickness of the embankments. The stratigraphic sequence, finds and radiocarbon analysis of the wooden fragments made it possible to reconstruct the most important phases in the road’s life: a first phase, of intense use, during the 1st and 2nd century AD and a second one during the 4th century AD (Martinelli 2004). As far as the human presence is concerned, nine sites have been dated to the Roman phase. Pottery was found at all sites. The areas with greater amounts of material were classified according to their size: ‘small’ from 1000 to 2000 sq.m, ‘medium’ from 2000 to 5000 sq.m and ‘large’ from 5000 to 8000 sq.m. The spatial distribution of sites of different sizes attests to a stable and selective territory occupation. The sites are concentrated in the northern and eastern parts of the area, on higher and drier locations with a sandy substratum, near palaeochannels or on the banks of active rivers such as the Canna palaeochannel and the Vallio River. In contrast, no sites are located south of the via Annia. This site distribution suggests that environmental factors may have determined human territorial choices. The apparent absence of sites in the narrow area by the Sile river, a very important watercourse, is probably connected with visibility problems of the surface, because the Roman surface

173

Hidden Landscapes

was covered by later deposits of the Sile. The via Annia would certainly have been a very important element in the population increase in this region. 2.2.3 Medieval and modern periods

For the medieval and modern periods, when human impact became overwhelming, information was obtained through archaeological research, which at the moment focuses on the inner branch of the via Annia. After five centuries of ‘silence’, in the 10th century AD and again in the 13th century AD a little wooden track was built on the location of the Roman bridge; both were linked to a local route mapped by Francesco Tron in 1613 (Martinelli 2004). The archaeological results can be integrated with historical data in the form of maps and written records, in order to reconstruct complex phenomena such as infrastructure, agriculture and buildings from the Middle Ages until the present (Basso et al. 2004).

3. Methods of elaboration of palaeoenvironmental data 3.1 Geomorphology

Geomorphological data provided the physical base for the virtual landscape reconstruction. Two GIS layers were produced for each period: a DTM of the specific morphological situation, and a palaeohydrographical map. Both are based on the interpretation of landform evolution at Ca’ Tron, which has been presented and discussed in previous papers (Bondesan et al. 2002; Bondesan et al. 2004; Mozzi et al. 2005; Miola et al. 2006). The DTMs of the separate time frames are modifications of a LIDAR DTM acquired in 2005. These modifications were carried out using the software Virtual Terrain Project (see also section 5), which allows the insertion of relevant morphological changes (i.e., erase or enhance fluvial ridges, cut incised valleys, excavate river beds, etc.). The palaeohydrographical networks of the different periods follow the geometries detected through remote sensing. The timing of activation and deactivation of each fluvial channel is based on radiocarbon dates and on interpretation of the sedimentary sequences. 3.2 Vegetation

The reconstruction of virtual landscapes requires the estimation of many data about past environments. Pollen analysis can provide information about plant communities. Plant macrofossils and non-pollen palynomorphs (NPP) also contribute to the definition of the past environment, because they are generally

less dispersed than pollen and they can therefore be considered direct evidence of the local presence of producers (Birks & Birks 1980; van Geel 2001). For a quantitative reconstruction of past vegetation on the basis of the pollen record, theoretical models have been proposed and tested in the last three decades (Prentice 1988; Sugita 1993). Given the specific pollen parameters and the size of the sedimentation basin, pollen records can be transformed into past vegetation data (Sugita 1998), but the Prentice-Sugita model cannot be applied to the Ca’ Tron pollen record because the features of the sampling sites do not agree with those assumed by the model, and because specific pollen parameters for vegetation in the Italian plain are unknown. As a consequence the reconstruction of the vegetation landscape can only be qualitative, and relies on both indirect and direct evidence: • Indirect evidence of the structure of the vegetation is given by the AP/NAP (arboreal pollen/ non arboreal pollen) ratio. This can be calculated from pollen percentages, pollen concentrations or even raw numbers of pollen counts (Tables 1, 2). Because of the generally low range of variation of NAP concentrations compared with the range of AP concentrations, AP/NAP ratio is considered an approximate index of the temporal variation of the tree density in a landscape (Magri 1994). • Indirect evidence of plant assemblages that lived together in the same space is formed by pollen assemblages that consistently occur in sediments of the same age and that have been sampled in a delimited area (Birks & Birks 1980). It has been shown that highly similar pollen assemblages — characterized by a very low percentage of Pinus and Betula and high percentages of herbs like Poaceae and Cyperaceae as well as by the presence of many different types of helophytes and hydrophytes — occur in many Last Glacial Maximum (LGM) pollen spectra from different cores in the low Venetian plain (including those at Ca’ Tron; Miola et al. 2006). This result has been used in this paper. • Direct evidence of the local presence of plant species (Birks & Birks 1980). Peat formations in waterlogged environments (as those of the LGM) preserve micro- and macro-remains of some plants that lived in them. Their occurrence indicates the local presence of these plants. We refer particularly to pollen of hydrophilous and hygrophilous plants, and to NPPs that generally have a very local dispersion. Macrofossil analysis enhanced the pollen record, allowing the identification of plant

174

Mozzi et alii, 20,000 Years of landscape evolution at Ca’ Tron (Venice, Italy)…

species unidentifiable by means of pollen morphology. Therefore, plant species identified by macrofossil analysis of LGM sediments cored 25 km away from Ca’ Tron, with very similar pollen assemblages (Miola et al. 2006), have been used together with the LGM pollen and NPP records of Ca’ Tron. NPP data have been presented by Miola et al. (2006) and used mainly in the Last Glacial Maximum virtual landscape reconstruction. • On the basis of the list of plant species derived from pollen and macrofossil identification, modern natural vegetation types that include the identified plant species were selected from among the rare modern natural plant communities that survive in the low plain of north-eastern Italy. Assuming that vegetation types have not changed during the last 20,000 years, the ecological value of the modern vegetation types was used for the reconstruction of environmental features. • Environmental conditions suggested by geomorphological and lithological evidence were used as well in the selection of the modern vegetation types and their spatial distribution in the virtual landscape. Pollen and NPP data were obtained from ten cores taken in the Ca’ Tron area, and were published in Miola et al. (2006). Pollen data of the Bronze Age are not available from the area of Ca’ Tron, but relevant archaeological evidence and environmental changes help us to reproduce the virtual landscape. Therefore we used data from Fiorentina (Venice, unpublished) and from Concordia Sagittaria (Venice, core San Gaetano S3, in Favaretto & Sostizzo 2006), located in the North Adriatic coastal plain respectively 25 and 50 km east of Ca’ Tron, as well as pollen data from other alluvial plain sites in Northern Italy taken from the literature. The sediments from Fiorentina have been radiocarbon dated to the early Bronze Age (2125-2075 cal BC, in Bondesan & Meneghel 2004) and those of Concordia Sagittaria to the Final Bronze Age on the basis of archaeological finds (Bondesan et al. 2005). For sampling, pollen extraction from sediments, and identification methods, see Miola et al. (2006).

4. Results of the processing of the palaeoenvironmental data 4.1 The landscape of the Last Glacial Maximum 4.1.1 Geomorphology

Most of the Ca’ Tron estate consists of an LGM alluvial plain built by the glaciofluvial Piave river system, which is fed by the piedmont Piave glacier (Bondesan & Meneghel 2004; Mozzi 2005; Fontana

et al. 2008). It is characterized by silty-clayey sediments with fine sands along the palaeochannels. As evidenced by well-preserved palaeochannels, the fluvial style in this distal section of the LGM Piave megafan was characterized by low-sinuosity channels, locally separated by longitudinal bars and islands. The building-up of natural levees led to the formation of shallow alluvial ridges, rising 1 to 2 m above the surrounding floodplain. Several extensive (1 ha – 10 sq.km) peaty layers detected in the LGM sedimentary sequence at Ca’ Tron suggest that the inter-ridge depressions hosted large mires, due to the near-surface ground water table (Miola et al. 2006). Other water-logged areas correspond to the abandoned river beds. By contrast, inactive alluvial ridges had better-drained soils thanks to their elevated morphology. In the LGM reconstruction (fig. 2a), we considered as active only the larger and better preserved Canna palaeochannel which crosses the study area; minor channels are interpreted as relict features with marshy conditions, which could be reactivated during major floods. The present alluvial ridge of the Sile river did not exist and therefore the corresponding area has been flattened. This area corresponded to the depression at the boundary between the LGM megafans of the Piave and Brenta rivers (Mozzi 2005). 4.1.2 Vegetation landscape

The vegetation landscape was probably almost exclusively herbaceous, as indicated by low values for the AP/NAP ratio (less than 0.3) and the AP concentrations (2000-10,000 grains per cm3 dry sediment) in Table 1 (Magri 1994). The high occurrence of pollen of hygrophilous plants and helophytes, of algal NPP types, and also of Poaceae, mostly Phragmites australis (Miola et al. 2006), suggests that the landscape was mainly characterized by herbaceous wetlands (fig. 2a). The interpretation of the pollen spectra on the basis of these criteria suggests the presence of the following plant communities (between brackets the indicative fossil types): 1. Sedge meadows developing in the wide mires between abandoned fluvial ridges. Open water may be present and the wetland water level fluctuates. The plant communities were characterized by Cyperaceae, Poaceae and brown mosses. Indicator species of past plant communities are Carex fusca (seeds, rootlets; leaf remains and pollen of Cyperaceae), Menyanthes trifoliata (pol-

175

Hidden Landscapes

Figure 2a, b – Maps of the LGM (a) and Bronze Age (b) landscape reconstructions.

len), Scorpidium scorpioides (leaf remains), Calliergon trifarium (leaf remains) and Drepanocladus exannulatus (leaf remains) (Miola et al. 2006). The physiognomy of the landscape was probably similar to a vegetation mosaic composed of tussocks of Cyperaceae and, in the lower and wetter stands with frequent open water, fringes of Menyanthes trifoliata communities. Geographic location, water chemistry, absence of Sphagnum species and palynological data all suggest a similarity with relict fens of the Eastern Po plain (Poldini 1973; Sburlino & Ghirelli 1994), related to Caricetalia davallianae Br.-Bl. 1949 (Scheuchzerio-Caricetea fuscae Tx. 1937). Additional studies on plant macrofossils could allow us to zoom in on a particular phytosociological association, on the basis of floristic composition and, above all, on the basis of species identification of Cyperaceae. 2. Marshes in the abandoned channels. The plant communities were characterized by the dominance of helophytes such as Phragmites australis (pollen of Poaceae, leaves and rhizomes of Phragmites australis), Typha latifolia (pollen) and Typha angustifolia (pollen). These species

normally grow in dense individual stands, often extensive and very poor in species (Rodwell et al. 1995). The reed beds occur in periodically waterlogged habitats. P. australis is widespread in riparian habitats along sluggish rivers, channels and dykes, in floodplain mires and in estuarine zones on salt marshes (Haslam 1987). Phragmition communis Koch 1926 can be considered to be the phytosociological reference for these communities (Balátová-Tuláková et al. 1993), as the association is recorded in Venetian plain (Marchiori & Sburlino 1997). During the LGM phase vegetation canopy was probably composed of patches with different physiognomy, with alternating clumps of Phragmites australis, Typha latifolia, or Typha angustifolia, depending on water depth, frequency and extent of flooding and competitive interaction among species. On occasionally waterlogged soils, wet meadows of the Magnocaricion Koch 1926 could grow as they are considered to be the next stage in a dynamic sequence from waterlogged to dry soils (Sburlino & Marchiori 1985; 1987). 3. On the banks of the Canna palaeochannel a perennial herbaceous riparian vegetation dominated

176

Mozzi et alii, 20,000 Years of landscape evolution at Ca’ Tron (Venice, Italy)…

by Typhoides arundinacea or Carex sp. (pollen of Cyperaceae and Poaceae) could grow, as on the modern banks of the Adige and Po rivers (Brac-

co & Villani, in press) The riverbank vegetation could tolerate the cyclical actions of running waters, with its alternation of periodic submersion,

Taxa

Pollen

Macro remains

Pollen % range min – max

n. of samples in which pollen type occurs

Pinus undiff.

+

–

1.7 – 20.2

19

Betula

+

–

0.0 – 2.6

16

Larix type, Picea, Corylus, Ephedra fragilis type, Hyppophae rhamnoides, Alnus, Salix

+

–

0.0 – 0.6

1–6

2.9 – 21.2

Trees and shrubs Poaceae

+

leaves

10.3 – 39.5

19

Artemisia

+

–

1.0 – 6.1

19

Chenopodiaceae

+

–

0.0 – 2.5

14

Scrophulariaceae

+

–

0.0 – 1.8

14

Saxifragaceae

+

–

0.0 – 1.4

15

Galium type

+

–

0.0 – 1.2

13

Aster type, Asteraceae Cichorioideae, Scorzonera type, Lactuca sativa type, Helianthemum, Plantaginaceae, Apiaceae, Caryophyllaceae, Thalictrum, Primula veris type, Urtica dioica, U. pilulifera, Ranunculus arvensis, Potentilla type, Pinguicola, Labiatae B, Astrantia major, Gentiana.

+

–

0.0 – 0.9

1 – 13

16.5 – 43.5

Herbs Potamogeton subg. P. type

+

leaves

0.0 – 11.0

16

Callitriche

+

–

0.0 – 4.1

5

Menyanthes trifoliata, Myriophyllum verticillatum, Nymphaea alba type, Potamogeton subg. C. type

+

sclereids of Nymphaea

0.0 – 3.0

3–7

0.0 – 15.4

 

38.3 – 71.5

19

0.0 – 2.2

5–7

Obligate aquatics Cyperaceae

+

leaves, roots

Carex fusca

–

seeds

Phragmites australis

–

leaves, rhizomes

Sparganium erectum, Typha angustifolia type, T. latifolia type

+

–

38.3 – 71.9

Helophytes Scorpidium scorpioides, Drepanocladus exannulatus, Calliergon trifarium

–

leaves

Brown mosses 256 – 862

Total pollen

1.728 – 10.52

AP concentration gr/cm3 AP/NAP

0.0 – 0.3

AP/NAP (only Herbs)

0.1 – 1.1

Table 1 – Pollen percentage range, AP/NAP and AP concentrations of Ca’ Tron LGM sediments (19 samples). The litho- and biostratigraphic correlation of the samples is discussed in Mozzi & Bondesan (2004) and Miola et al. (2006). Macro-remains are recorded in LGM sediments at three sites located 25 km from Ca’ Tron (Miola et al. 2006).

177

Hidden Landscapes

Fiorentina 1.07m

Concordia Sagittaria S3 2.73 m

17.7

23.6

Quercus ilex

0.0

0.6

Corylus avellana

3.8

8.1

Fagus

3.4

6.1

Carpinus type

1.9

5.5

Ostrya type

0.0

2.6

Ulmus

2.6

0.9

Tilia

0.0

0.6

Fraxinus

0.0

0.3

Non-opercolate Rosaceae

0.0

1.4

Viburnum

0.0

0.6

Pinus

3.0

4.6

Picea abies

1.1

4.0

Abies alba

6.8

3.2

Juniperus type

0.0

2.6

Betula

1.5

0.6

Larix type

0.4

0.0

Alnus

19.9

13.5

Salix

1.9

1.4

Populus

3.4

0.0

Juglans

0.0

0.6

Castanea sativa

0.0

0.3

Arboreal pollen (AP)

67.3

81.0

Poaceae

16.9

11.5

Artemisia

2.6

1.7

Chenopodiaceae

5.6

1.7

Asteraceae undiff

1.5

0.9

Papaver rhoeas

0.0

0.6

Centaurea nigra type

0.0

0.3

Plantaginaceae

0.0

0.3

Rhinanthus type

0.0

0.3

Rumex

0.4

0.0

Other herbs

5.6

1.7

32.7

19.0

AP/NAP

2.1

4.3

Total land pollen

266

347

Total pollen

342

371

35,400

37,000

Pollen types Quercus robur group

Non arboreal Pollen (NAP)

AP concentration gr/cm3

related to sedimentation and soil erosion during flood events, and of summer drought. Typhoides arundinacea survives the disturbance connected with flooding by producing aerial shoots and adventitious roots at the nodes after the stems have been flattened, and thus starts a new growth phase. 4. On the alluvial ridges, where good drainage created arid places, the plant landscape was characterized by dry grassland formations. Poaceae, Artemisia and Asteraceae Chicorioideae pollen suggest the presence of plant communities similar to the modern associations with Chrysopogon gryllus, Bromus erectus, Onobrychis arenaria, Anthyllis vulneraria of the Scorzonero-Chrysopogonetalia Horvatic 1963, recorded in Friuli (Poldini 1989; Feoli Chiapella & Poldini 1993) and Veneto (Sburlino et al. 1995). These dry meadows today thrive on soils with good drainage and coarse soil texture. They are considered to be in catenal relations with the vegetation of the peaty areas (Sburlino et al. 1995). 4.2 The Bronze Age landscape 4.2.1 Geomorphology

Table 2 – Pollen percentages, AP/NAP and AP concentrations of Fiorentina (4075-4025 cal. BP) and Concordia Sagittaria (3000 BP).

The most outstanding geomorphic feature in the Bronze Age landscape was the incised valley which existed along the present-day Sile river. This valley formed within the LGM interfandepression mentioned earlier, probably during the general entrenching of the hydrographic network that occurred in the whole Venetian-Friulian plain at the end of the last glaciation as a consequence of environmentally-induced changes in fluvial dynamics (Fontana et al. 2008). Though the overall incision is ca. 13 m deep, the radiocarbon date of a piece of wood embedded in the alluvial valley-fill (4090-3860 cal BP, Bondesan et al. 2004) indicates that, at the beginning of the 2nd millennium BC, the valley floor was about four metres lower than the surrounding plain. On the basis of these data the DTM has been modified accordingly. The Sile river ran through this 500 m-wide valley (fig. 2b), possibly receiving occasional outflow from the Piave river. In the reconstructed morphology, the minor rivers that crossed the Ca’ Tron estate are influenced in their drainage directions by the relict LGM landforms (e.g. alluvial ridges, palaeochannels). Soils in the terraced plain were probably relatively well-drained. Wet areas persisted in the elongated depressions left by the LGM palaeochannels and in the valley floor, where flooding by the Sile/Piave river could occur.

178

Mozzi et alii, 20,000 Years of landscape evolution at Ca’ Tron (Venice, Italy)…

4.2.2 Vegetation landscape

In the Bronze Age the AP/NAP ratio rose to 2.14.3 (table 2). The AP concentrations were 3-10 times greater than the AP concentrations of the LGM period. The pollen spectra of Fiorentina and Concordia Sagittaria present a pollen flora quite similar to what has been documented in the literature for the same period in the Veneto and Emilia Romagna plains (Accorsi et al. 1999; Mercuri et al. 2006; Valsecchi et al. 2006). According to these authors the landscape during the Sub-Boreal, the end of which corresponds to the Final Bronze Age – Early Iron Age, was characterized in the Emilia Romagna and Veneto plains by a patchwork of forested and open areas. The components of mixed oak forest were Quercus (deciduous type), Carpinus betulus, Ulmus, Ostrya carpinifolia/Carpinus orientalis, Fraxinus excelsior, F. ornus, Corylus avellana, Tilia and Acer campestre. Some rare Fagus sylvatica, Betula, Pinus and Abies, that today live at higher altitudes in the hills or in the mountains, also grew in the plain (Mercuri et al. 2006). Hygrophilous plants such as Alnus glutinosa, Salix and Populus grew in depressed waterlogged areas or along rivers. During the Middle and Late Bronze Ages the forest was fairly open or cleared near human settlements (Mercuri et al. 2006; Valsecchi et al. 2006). From the Final Bronze Age a new expansion phase of the forest has been recorded around the Lago di Garda (Valsecchi et al. 2006). On the basis of present archaeological knowledge no Bronze Age human settlements have been reliably documented in the area of Ca’ Tron, and we therefore tentatively represent the landscape as a fairly forested, mixed oakwood on dry soil, substituted by a hygrophilous wood in humid depressed areas and along the river banks (fig. 2b).

5. The virtual-reality (eco)system The purpose of the multimedia project of Ca’ Tron is to enhance, through the use of specific technologies, the understanding of geoarchaeological data by inserting them in a dynamic way into the landscape, thus highlighting connective and spatial relations between the archaeological record and the environment. These relationships become clear only in their context or, even better, in the dynamic development of the landscape, because this enhances the dynamic rules and the interactions between sites and landscape. The application of virtual reality (VR) relies on the geographic contextualization of the available geomorphological, palynological and archaeological data. These are placed in their spatial and

chronological context in a dynamic and multilayered way. It will thus be possible to emphasize the dynamics of transformation inherent in the evolution of landscapes. In this complex connective texture, virtual reality techniques allow an increase of information and perceptions through innovative methodologies, thus making accessible new maps of ‘mind and foot’ routes to specialists and the general public. The goal of the project is the reconstruction of an archaeological landscape using a desktop system of virtual reality or a virtual web-GIS system, available on simple PC workstations. This system is set up as an innovative project of scientific communication and education, anchored to the territory, and able to modulate the degree of information according to the target user. Thanks to a bird’s-eye view, which allows the user to virtually move around in an interactive and subjective way within the territory of Ca’ Tron (i.e., choosing a path independently and not being limited to strict, preconfigured routes), the user can appreciate the reality as a whole. He or she can check present and ancient spatial relationships and get information on specific phenomena (the ‘monographic realities’) such as plant communities, archaeological sites, architectural features and landscape data in a general sense (fig. 3). In ancient landscape reconstruction we follow a work flow that enables an easy exchange of data and information between different teams of experts. Sharing interpretations is especially important for a multi-disciplinary team, where each team member has to process the same information in different ways and produce different results that can be used in the 3D reconstruction. As a first step, a common spatial digital archive was created. Pre-processed information (aerial photographs, satellite imagery, raster and ‘paper’ cartography) was georeferenced and overlaid — as explained above — in a GIS project. During ‘interpretation sessions’ information provided by geologists, archaeologists and palynologists on different periods was georeferenced and added to the GIS in the form of interpretative vector layers that could be used for further processing. We kept all these interpretation steps in order to achieve transparent reconstructions. Our approach is, indeed, to deal with archaeological and geological reconstructions not as finished products of our researches, but rather as interactive, updatable, and open geospatial processes. The various vector information layers were then added to a terrain - ecosystem generator. It

179

Hidden Landscapes

proved necessary to employ a specialised software for handling geospatial characteristics, importexport functions as well as 3D formats (such as *.3ds) and 3D terrain-export functions. We have chosen Visual Nature Studio 2 (VNS: 3dnature. com) for the ecosystem processing, which is a lowcost and powerful tool that also allows the export of terrain data in open source formats such as Virtual Terrain Project (VTP: www.vterrain.org). VNS cannot handle tiled and paged 3D terrain, while VTP Enviro gave us a perfect 3D interface to test our working hypotheses, but with VNS we could export data in GIS formats such as Digital Elevation Models and GeoImages (Geotiff). This was particularly important as we wanted to have the possibility to publish 3D reconstructed landscapes over the web. We then used OSG4web, a plug-in for Mozilla Firefox and Explorer browsers based on the OpenSceneGraph library (www.openscenegraph. org) and developed by CNR-ITABC and CINECA, to publish and interact dynamically with 3D terrains and vector information. The goal of this specific output, still in progress, is to create a web-based 3D shared working environment for archaeological and geological purposes (www.vhlab.itabc.cnr.it/ openheritage).

6. Conclusions The results of this first attempt to reconstruct ancient landscapes at Ca’ Tron are encouraging, though we faced several problems. The most critical point with regard to data acquisition has been the lack of continuous palynological series or other palaeoenvironmental proxies for the early and middle Holocene, due the sedimentary characteristics of the study area. With regard to the processing environmental data, the application of current models of interpolation of single-site pollen sequences to the reconstruction of the vegetation cover was sometimes problematic, particularly when human impact on vegetation had to be evaluated in terms of woodland-grasslandcropland proportions. In order to overcome such limitations, a helpful element of this research has been the possibility to integrate vegetation data with geomorphic information. This has been crucial for the production of the landscape maps of figure 2, where the polygons are traced on the basis of landforms and soil geography, while the attributes derive from palynological data. The application of VR techniques in our project and the complete integration of archaeological data in the model are still at an early stage. Their potential use in the reconstruction and visualization of ancient landscapes is, of course, enormous. An important aspect of our tests has been the possibility to introduce very detailed georeferenced information into the VR model, at the scale of a single river branch or plant association, while preserving the accuracy of the data set in the visualization. Another important point to be underlined is that such applications force a direct interaction between data producers (geologists, palynologists, archaeologists) and data modellers. This creates interesting feedbacks in the understanding of the 3D reality of past landscapes, which can be directly translated into the model through real-time modifications and updates.

References

Figure 3a, b – VR snapshots of the LGM landscape.

Accorsi C.A., Bandini Mazzanti M., Forlani L., Mercuri A.M. & Trevisan Grandi G. 1999. An overview of Holocene forest pollen flora/vegetation of the Emilia Romagna region – northern Italy. Archivio Geobotanico 5(1-2), pp. 3-27. Amorosi A., Fontana A., Antonioli F., Primon S. & Bondesan A. 2008. Post-LGM sedimentation and Holocene shoreline evolution in the NW Adriatic coastal area. GeoActa 7, pp. 41-67. Balátová-Tuláková E., Mucina L., Ellmauer T. & Wallnöfer S. 1993. Phragmiti-Magnocaricetea, in Mucina L., Grabherr G.

180

Mozzi et alii, 20,000 Years of landscape evolution at Ca’ Tron (Venice, Italy)…

& Ellmauer T. (eds), Die Pflanzengesellschaften Österreichs Teil II, pp. 79-130. Gustav Fischer, Jena. Balista C., De Guio A., Vanzetti A., Betto A., De Angeli G. & Sartor F. 2005. Paleoidrografie, impianti terramaricoli e strade su argine: evoluzione paleoambientale, dinamiche insediative e organizzazione territoriale nelle Valli Grandi Veronesi alla fine dell’età del Bronzo. Padusa, n.s. 41, pp. 97-138. Basso P., Bonetto J., Busana M.S. & Michelini P. 2004. La via Annia nella Tenuta di Ca’Tron. Percorso, tecnica, cronologia, in Ghedini F. & Busana M.S. (eds) 2004, pp. 41-98. Birks H.J.B. & Birks H.H. 1980. Quaternary Palaeoecology. Edward Arnold, London. Bondesan A., Mozzi P., Primon S. & Zamboni C. 2002. Antiche acque e antiche genti: l’indagine geomorfologica tra passato e presente, in Ghedini F., Bondesan A. & Busana M.S. (eds), pp. 15-74. Bondesan A. & Meneghel M. (eds) 2004, Geomorfologia della provincia di Venezia. Esedra Padova. Bondesan A., Finzi E., Fontana A., Francese R., Magri S., Mozzi P., Primon S. & Zamboni C. 2004. La Via Annia a Ca’ Tron: nuovi contributi della geomorfologia, della geofisica e del telerilevamento, in Busana M.S. & Ghedini F. (eds) 2004, pp. 109-146. Bondesan A., Asioli A., Favaretto S., Fontana A., Gobbato D., Lubiani A., Miola A., Sostizzo I., Toffoletto F. & Valentini G. 2005. Paleoambienti tardo quaternari nella bassa pianura friulana: ricerche multidisciplinari nell’area di Concordia sagittaria (VE). Atti del Convegno nazionale A.I.Geo. Montagne e Pianure. Padova (15-17 febbraio 2005), Materiali del Dipartimento di Geografia, Università di Padova 28, pp. 44-46. Bracco F. & Villani M., in press. Syntaxonomy of the Typhoides arundinacea (L.) Moench Vegetation of the Adige River (Northern Italy). Fitosociologia. Busana M.S. 2002. La documentazione archeologica: l’area di Ca’Tron nel quadro dei rinvenimenti tra Sile e Piave, in Ghedini F., Bondesan A. & Busana M.S. (eds) 2002, pp. 151-160. Busana M.S. (ed.) 2005. La Tenuta di Ca’Tron (RoncadeTreviso/Meolo-Venezia): le indagini su un edificio rustico. Quaderni di Archeologia del Veneto 21, pp. 62-72. Busana M.S. (ed.) 2007. Progetto Ca’Tron (Roncade-Tv/MeoloVe): indagini nell’agro orientale di Altino. Quaderni di Archeologia del Veneto 23, pp. 49-59. Busana M.S. (ed.) 2008a. Progetto Ca’ Tron (Roncade–Treviso/ Meolo–Venezia): indagini su due insediamenti rustici di età romana nell’agro orientale di Altino. Quaderni di Archeologia del Veneto 24, pp. 41-51. Busana M.S. 2008b. Indagini nell’agro orientale di Altino: il popolamento in età romana tra Sile e Piave, in Spazi, forme e infrastrutture dell’abitare, Atlante Tematico di Topografia Antica 18. Erma, Roma, pp. 27-47. Favaretto S. & Sostizzo I. 2006. Vegetazione e ambienti del passato nell’area di Concordia Sagittaria (VE). Quaderni del Dottorato. Università di Padova 1, pp. 57-70. Feoli Chiapella L. & Poldini L. 1993. Prati e pascoli del Friuli (NE Italia) su substrati basici. Studia Geobotanica 13, pp. 3-140. Fontana A., Mozzi P. & Bondesan A. 2008. Alluvial megafans in the Veneto-Friuli Plain (north-eastern Italy). Evidence of sedimentary and erosive phases during Late Pleistocene and Holocene. Quaternary International 189, pp. 71-90.

Forte M. & Campana S. (eds) 2006. From Space to Place. Proceedings of the 2nd International Conference on Remote Sensing in Archaeology, Rome, Italy, Decembre 4-7, 2006 (BAR International Series 1568). Ghedini F., Bondesan A. & Busana M.S. (eds) 2002. La tenuta di Ca’Tron. Ambiente e storia nella terra dei Dogi, Cierre Edizioni, Verona. Ghedini F. & Busana M.S. (eds) 2004. La via Annia e le sue infrastrutture. Atti delle Giornate di Studio (Ca’Tron di Roncade, 6-7 novembre 2003). Grafiche Antiga, Cornuda (Treviso). Haslam S.M. 1987. River plants of western Europe. Cambridge University Press. Magri D. 1994. Late-Quaternary changes of plant biomass as recorded by pollen stratigraphical data: a discussion of the problem at Valle di Castiglione, Italy. Review of Palaeobotany and Palynology 81, pp. 313-325. Marchiori S. & Sburlino G. 1997. Present vegetation of the Venetian Plain. Allionia 34, pp. 165-180. Martinelli N. 2004. Il ponte della Via Annia. Esame filologico e datazione assoluta delle palificazioni di fondazione e dei resti lignei venuti in luce a Ca’ Tron, in Ghedini F. & Busana M.S. (eds) 2004, pp. 99-108. Mercuri A.M., Accorsi C.A., Bandini Mazzanti M., Bosi G., Cardarelli A., Labate D., Marchesini M. & Trevisan Grandi G. 2006. Economy and environment of Bronze Age settlements – Terramare – on the Po Plain (northern Italy): first results from the archeobotanical research at the Terramare di Montale. Vegetation History and Archaeobotany 16, pp. 43-60. Miola A., Bondesan A., Corain L., Favaretto S., Mozzi P., Piovan S. & Sostizzo I. 2006. Wetlands in the Venetian Po Plain (northeastern Italy) during the Last Glacial Maximum: Interplay between vegetation, hydrology and sedimentary environment. Review of Palaeobotany and Palynology 141, pp. 53-81. Miola A. & Valentini G. 2004. La via Annia a Ca’ Tron: il contributo dell’analisi palinologica, in Ghedini F. & Busana M.S. (eds) 2004, pp. 147-162. Mozzi P. 2005. Alluvial plain formation during the Late Quaternary between the southern Alpine margin and the Lagoon of Venice (northern Italy). Suppl. Geogr. Fis. Dinam. Quat. 7, pp. 219-230. Mozzi P., Bondesan A., Busana M.S., Francese R., Miola A. & Valentini G. 2005. Setting archaeological landscapes within palaeoenvironmental dynamics in the Ca’Tron area, Venice (Italy): a geo-archaeological approach. Proceedings of the 2nd Italy-United States Workshop, Rome (BAR International Series 1379), pp. 35-52. Pescarin S. 2006. Open source in archeologia: nuove prospettive per la ricerca. Archeologia e Calcolatori 17, pp. 137-155. Pescarin S. 2007. Interpretazione e ricostruzione del paesaggio antico: una proposta di formulazione teorica esplicita, in La Villa di Livia; un percorso di ricerca di archeologia virtuale. Ed. Erma, Roma. Poldini L. 1973. Die Pflanzendecke der Kalkflachmoore in Friaul (Nordostitalien). Ber. Geobot. Inst. ETH 51, pp. 166-178. Poldini L. 1989. La vegetazione del Carso isontino e triestino. Ed. Lint, Trieste. Prentice I.C. 1988. Records of vegetation in time and space: the principle of pollen analysis, in Huntley B. & Webb T. III (eds), Vegetation History, pp. 17-42. Kluwer Academic, Dordrecht.

181

Hidden Landscapes

Rodwell J.S. (ed.) 1995. British Plant Communities. Vol. 4. Aquatic communities, swamps and tall-herb fens. Cambridge University Press. Sburlino G. & Marchiori S. 1987. La vegetazione idroigrofila del medio corso del fiume Brenta (Veneto. Italia settentrionale). 5 Jornadas de Fitosociologia, Univ. de La Laguna. Ser. Informes 22, pp. 297-304. Sburlino G. & Ghirelli L. 1994. Le cenosi a Schoenus nigricans del Caricion davallianae Klika 1934 nella Pianura Padana Orientale (Veneto-Friuli). Studia Geobotanica 14, pp. 63-68. Sburlino G., Bracco F., Buffa G. & Ghirelli L. 1995. Rapporti dinamici e spaziali nella vegetazione legata alle torbiere basse neutro-alcaline delle risorgive della Pianura Padana orientale (Italia Settentrionale). Colloques Phytosociologiques 24, pp. 285-294. Serandrei-Barbero R., Lezziero A., Albani A. & Zoppi U. 2001. Depositi Tardo-Pleistocenici ed Olocenici nel sottosuolo veneziano: paleoambienti e cronologia. Il Quaternario, Italian Journal of Quaternary Sciences 14, pp. 9-22. Sugita S. 1993. A model of pollen Source Area for an Entire Lake Surface. Quaternary Research 39, pp. 239-244. Sugita S. 1998. Modelling pollen representation of vegetation. Paläoklimaforschung / Palaeoclimate Research 27, pp. 1-16. Valsecchi V., Tinner W., Fisinger W. & Ammann B. 2006. Human impact during the Bronze Age on the vegetation at Lago di Lucone (northern Italy). Vegetation History and Archaeobotany 15, pp. 99-113. Van Geel B. 2001. Non-Pollen Palynomorphs, in Smol J.P., Birks H.J.B & Last W.M. (eds), Tracking Environmental Change Using Lake Sediments. Volume 3: Terrestrial, Algal, and Siliceous Indicators, pp. 99-119. Kluwer Academic Publishers, Dordrecht, The Netherlands.

Acknowledgments Financial support by the Fondazione Cassamarca (Treviso), the Regione del Veneto, Interreg IIIA (Interadria Project), ARCUS (via Annia Project) and the Italian Ministry of University and Research MIUR (PRIN Project: Environmental crisis and human communities in the Holocene: a geoarchaeological approach, national coordinator M. Cremaschi) is gratefully acknowledged. Sincere thanks are due to Giovanni Leonardi for the study of the pre- and protohistoric materials and to Francesco Ferrarese for preparing the figures. The geomorphology and sedimentology were undertaken by P. Mozzi and A. Bondesan; palynology and phytosociology by A. Miola and C. Villani; archaeology by M.S. Busana; virtual reality by S. Pescarin and P. Kirschner.

182

22 A hidden Neolithic landscape in Apulia (southern Italy) Massimo CaldaraI, Italo M. MuntoniII, Girolamo FiorentinoIII, Milena PrimaveraIV, Francesca RadinaV I Dipartimento di Scienze della Terra e Geoambientali, Università degli Studi di Bari ‘Aldo Moro’, via Orabona 4, 70125 Bari, Italy – [email protected] II Soprintendenza per i Beni Archeologici della Puglia, via De Nittis 7, 71121, Foggia, Italy – [email protected] III Dipartimento di Beni Culturali, Università degli Studi del Salento, Via D. Birago 64, 73100 Lecce, Italy – [email protected] IV Dottorato di Ricerca in ‘Geomorfologia e Dinamica Ambientale’ (XX ciclo), Università degli Studi di Bari ‘ Aldo Moro’, via Orabona 4, 70125 Bari, Italy – [email protected] V Soprintendenza per i Beni Archeologici della Puglia, Palazzo Simi, Strada Lamberti, 70122 Bari, Italy – [email protected] Abstract Hidden Neolithic landscapes of the Adriatic coastal region in Apulia have been reconstructed by comparing geomorphological, palaeoenvironmental, archaeobotanical and archaeological data. Ancient landscapes are hidden and/or transformed by natural processes (rising sea-level, erosion and/or accumulation processes) and by recent human activity (intensive land use and urban expansion). Recent research on archaeological landscapes in the Tavoliere, Ofanto and Murge areas has highlighted several less well-known aspects of hidden Early to Late Neolithic landscapes.

1. Introduction: hidden or transformed landscapes? Within the study area it is possible to identify different types of ‘hidden landscapes’ spanning the Holocene to the beginning of the 4th millennium BC and coinciding with the decline of the Neolithic world. The Neolithisation of the region is characterised by settlements that are often surrounded by walls and ditches, by the use of natural shelters and caves that are often artificially modified for funerary or cult practices, and by the use of marshlands and waterways for transport between coastal and inland territories. All this significantly defined and redefined the natural landscape, making it the oldest ‘stratum’ of the cultural landscape. This Neolithic landscape has largely been altered by subsequent natural and man-made modifications. The slow but steady rise in sea levels has been the main factor affecting the relationship between coastal area and settlement; further inland, natural processes of erosion as well as ploughing severely affected the preservation of the archaeological record. Heavy urbanization in recent times has brought about a radical transformation of the coastal landscape. Finally, the typical karstic caves and dolines of the Apulian landscape can be considered naturally hidden landscapes, often used by Neolithic communities for ritual and/or funerary activity (Radina in press). While these factors have complicated the process of aquiring data, initial results from recent multidisciplinary studies have allowed us to outline a preliminary reconstruction of the hidden landscapes of the

Adriatic coastal region of the Murge Plateau over a period of more than two thousand years (from Early to Late Neolithic) through the comparison of geological, palaeoenvironmental, archaeobotanical and archaeological data from the area of Ariscianne (Barletta), Pulo di Molfetta and Scamuso (Torre a Mare) in the provinces of Barletta-Andria-Trani and Bari (fig. 1).

2. The study area The Tavoliere is a broad plain gently sloping seaward, bordered by the Gargano promontory, the Murge hills and the Subapennine mountains. During the Quaternary a regional discontinuous uplift, still active and complicated by a sequence of glacioeustatic sea-level fluctuations, took place and resulted in a sequence of marine and alluvial terraced deposits. The slight slope being the main morphological feature of the area, the Holocene sea-level rise, characterised by small fluctuations, produced effects both in the coastal area and deep into the alluvial plain (Caldara et al. 2002; 2005b). The Murge, by contrast, is a highplain extending between the Ofanto valley and the Salento hills, formed from Mesozoic limestone bordered by Bradanic (Pliocene-Lower Pleistocene) and Quaternary terraced marine deposits. The morphology of this massif (fig. 2) is influenced also by karst phenomena and by a relic fluvial network (the socalled lame) carved into the calcareous rocks. The Holocene sea level rise caused the drowning of the lower fluvial valleys, which became rias. Several rias

183

Hidden Landscapes

were barred by longshore drifting sand, creating small coastal lakes (Caldara & Pennetta 2002). Both the Tavoliere plain and the Murge plateau, although characterised by different landscapes, were

densely settled in the Neolithic period between the end of the 7th and the beginning of the 4th millennium BC. The two areas were the scene of systematic agricultural exploitation and animal husbandry.

Figure 1 – Reconstruction of the NW Murge coastal plain during the Early Neolithic, and localization of the geological cross sections of fig. 2 (modified from Caldara & Pennetta 2002).

Figure 2 – Schematic geological cross sections and localization of Neolithic villages.

184

Caldara – Muntoni – Fiorentino – Primavera – Radina, A hidden Neolithic landscape in Apulia…

Since the 1960s the archaeology and environment of the Tavoliere have been intensively studied, and the dynamics of the Neolithic landscape and human settlement are sufficiently well known (Bradford 1957; Tinè 1983; Cassano & Manfredini 1983; 2005). Since the 1990s archaeological research into the Neolithic has turned to the Murge plateau, with extensive surveys, settlement excavations (Pulo di Molfetta, Balsignano) and cave excavations (Grotte di Santa Croce) that have also shed light on the process of adoption of new forms of economic production (Geniola 1987; Fiorentino et al. 2000; Radina 2002; 2006; Muntoni 2003). Closely linked to natural processes (mainly sea level changes) and human activity, the palaeoenvironment of the Apulian coast has been reconstructed through the multidisciplinary analysis of sediment cores drilled in various coastal areas. Geomorphological, palaeobotanical, archaeobotanical, palaeontological and sedimentological studies have allowed us to define the main features and dynamics of mid-Holocene environmental change. Different types of proxy data, such as molluscs, ostracods and foraminifera, macrophytes and charophytes, seeds/fruits, pollen and spores have been compared with archaeological evidence for the Neolithic occupation of the Murge and Tavoliere derived from field survey, excavation, and underwater prospecting, and the results have been tied to the radiocarbon chronological framework.

3. Results 3.1 Ariscianne

The Ariscianne-Belvedere wetland is located between Trani and Barletta near the mouth of the Camaggi stream at the foot of Murge plateau. This area was inhabited during the Neolithic and some Early and Middle Neolithic villages (S. Antonio a Callano, Pezza delle Rose, etc.) were identified by archaeological surveys on the surrounding hills (Muntoni 2002). Neolithic pottery fragments, lithic industries and faunal remains have also been collected along the coastline. A multidisciplinary research project carried out in 2002-2003 aimed to identify the primary contexts of these archaeological remains. Archaeological excavations and underwater investigations (fig. 2) have also shown the presence of archaeological deposits below the present sea level. These seem linked with settlements originally located near the coastline but now completely submerged and eroded by wave action (Caldara et al. 2005a). The environmental evolution of this

coastal area, reconstructed on the basis of sediment cores (Simone 2003; Caldara et al. 2005a; Primavera 2008), shows that, during the Neolithic (between 5910 BC and 5105 BC), the area was probably the inner part of a coastal basin characterised by freshwater ponds, as shown by molluscs assemblages of the cores closest to the sea. The palaeoenvironment near the Pezza delle Rose Neolithic settlement can be inferred from the plant macrofossil assemblage in the ARI24 core (the one farthest from the sea). Analysis indicates no relevant environmental changes (fig. 3) during the mid-Holocene (5190-2610 BC), which was characterised by the presence of fens, a particular wetland type fed by freshwater springs. An erosional surface separates these levels from recent alluvial sediments. The study carried out on the Ariscianne wetland has allowed us to distinguish at least three periods of elevated water levels, occurring between 49304390 BC, around 3240 BC, and before 2610 BC. For the period 4930-4390 BC the fen-alkaline water body is characterised by Great fen-Sedge (Cladium mariscus) and the temporary presence of submerged vegetation (Characeae, Potamogeton sp., Nuphar luteum) together with wet-meadows plants (Sagittaria sagittifolia, Potentilla palustris). Between 4390-3240 BC the plant macrofossil record is characterised by decreases of charophytes and the absence of submerged macrophytes, suggesting a reduction in water level. The renewed spreading of marsh vegetation, highlighted by high concentrations of rush (Juncus sp.) and fen-sedge (Cladium mariscus) seeds, suggests a return to a greater fresh-water availability around 3240 BC and shortly before 2610 BC. 3.2 Pulo di Molfetta

The surroundings of the Pulo di Molfetta doline as well as caves inside this oval depression were densely settled from the Early to the Late Neolithic period (from the 7th to the 5th millennium BC). Systematic research on the eastern border of the doline (Fondo Azzollini locality), carried out from 1997, have revealed a heavily eroded Early Neolithic settlement (6100-5880 BC [2ı]) enclosed by a wide dry-stone wall located on a calcareous plateau above the sink-hole (Radina 2007; fig. 2). Topographical and morphological features show eroded sediments flowing into the adjacent natural basin (Fiorentino 1999; Muntoni 2007). Preliminary archaeobotanical analysis has been carried out on these sediments to reconstruct how environmental resources were exploited during the Mid-Holocene. Our research

185

Hidden Landscapes

Figure 3 – Ariscianne core 24, plant macrofossil diagram and faunal content. The highlighted phases indicate the major water flow periods of the marsh.

Figure 4 – Palaeoenvironment and landscape reconstruction from off-site and in-site archaeobotanical analyses: on the left data from core S1-bis; on the right data from the Early Neolithic settlement of Fondo Azzollini.

186

Caldara – Muntoni – Fiorentino – Primavera – Radina, A hidden Neolithic landscape in Apulia…

(fig. 4) compares archaeobotanical data from the archaeological site itself with plant remains from the S1-b core drilled in the Pulo (off-site) (Fiorentino & Primavera 2007; Primavera et al. 2007). The on-site archaebotanical record comprises charcoals and a few seed/fruit remains. Two different catchment areas for wood exploitation, mainly connected to domestic use, have been identified from anthracological analysis of Early Neolithic layers. The first comprises mixed deciduous wood and river- and lakeside vegetation (Corylus/Alnus; Carpinus betulus; Hedera helix, Vitis vinifera subsp. sylvestris); the second consists of sclerophyllous trees and shrubs of different macchia types (Olea europaea, Pistacia lentiscus, Erica sp., Cistus sp.). Despite the low frequency of cereal remains (mainly Triticum dicoccum and Hordeum vulgare sp.), agricultural activity can also be inferred from weeds growing over cultivated fields such as Chenopodium album; Galium aparine; Polygonum sp., Malva sp. and from spikelets, chaff remains and seed casts on pottery (Fiorentino et al. 2000, fig. 5b). Preliminary off-site archaeobotanical analysis was carried out on the S1-b core, drilled in the floor of the doline to reconstruct the changes in the environment from the Neolithic to the Bronze Age period. Three Radiocarbon AMS dates provide the correlation between settlement and cave occupation phases and the sediment stratigraphy in the core. The large amount of cereal grains and chaff remains (Triticum dicoccum and T. monococcum) preserved in the core sediments suggests that the Murge plateau was intensively cultivated from 5545 to 4770 BC. The anthracological record, made up of river- and lakeside vegetation and mixed deciduous taxa (Populus/ Salix, Corylus/Alnus, Cornus sp., Quercus pubescens type), suggests wet conditions during the Early Neolithic and more easily available water resources for Neolithic farmers. Changes in vegetation and agricultural practices can be identified after 4770 BC and during the Bronze Age (2050 BC). Drought conditions are suggested by the spread of Olea europaea together with other sclerophyllous taxa (Myrtus communis, Cistus sp., Erica sp.), and the absence of cereal remains coincides with the abandonment of the settlement and the start of the ritual use of caves in the doline. 3.3 Scamuso

The Neolithic settlement of Scamuso, excavated from 1985 to 1988 by Rome’s ‘Tor Vergata’ University (Biancofiore & Coppola 1997), is lo-

cated near Torre a Mare along the Adriatic coast to the south-east of Bari, at the foot of the Murge plateau (fig. 2). Systematic archaeological excavation revealed a continuity in the stratigraphic sequence spanning from the Early (6400-5320 BC [2ı]) to the Middle (5480-4500 BC [2ı]) and Late Neolithic phases. Habitation structures with stone footings and wooden frames with a daub covering and functional zones (structured firing places, cobble structures, postholes…), dated to the three different phases of the site, were identified. Despite the erosion linked to the Holocene sea level rise, and the submersion of some structures, the site is still important for the palaeoenvironmental reconstruction of the Adriatic Murge. Palynological analysis (Renault-Miskovsky & Bui-Thi-Mai 1997) has been somewhat limited for the archaeological context, with many taphonomical problems involving the preservation of pollen and spores. The pollen diagram (fig. 5), which has here been plotted without the over-represented Cichorioideae, shows that the arboreal pollen percentage is less than 5%. This suggests an open landscape with characteristic coastal features. The presence of Olea, Phillyrea, Pistacia and Cupressaceae reveals the existence of a dune belt during the Early Neolithic, later destroyed by the rising sea level. Concentrations of Chenopodiaceae pollen are evidence for increasing salinity, while the absence of these taxa points to a crisis in the dune system during the Middle Neolithic.

4. Discussion: environmental dynamics in the Tavoliere and Murge during the Neolithic The post-glacial climate improvement resulted in a global sea level rise, causing some late-glacial coastal areas to be submerged. The sea level rise was, however, characterised by several phases of stability; the most marked of these occurred at the beginning of the Neolithic. Along the Tavoliere coast, streams and rivers draining soft erodible terrain were characterized by a large turbid discharge. On the other hand, streams draining the Murge, flowing on a calcareous substrate, transported a very low amount of sediment. There was therefore a greater availability of material to be distributed alongshore on the Tavoliere coastline (Caldara et al. 2002). At the beginning of the Neolithic (fig. 6), sea levels from Manfredonia to Bisceglie stood at a depth of about 10-15 m. A narrow coastal barrier and a wider lagoon developed in the Tavoliere plain. About 10 km offshore, at least 3 or 4 rows of dune

187

Hidden Landscapes

Figure 5 – Scamuso, section A I Nord. Pollen diagram processed without Cichorioideae (modified from Renault-Miskovsky & Bui-ThiMai 1997).

ridges can in fact be identified at a depth of 8 meters (Segre 1969). During the Early Neolithic this coastal lagoon was open to the sea through wide tidal inlets. The results presented here confirm that during this phase (5910–5105 BC) the Ariscianne area was probably the inner part of a shallow coastal basin characterized by freshwater environments. Underwater features suggest a sand barrier-lagoon system at a depth of about 10-15 m. Following climate change and warmer temperatures, the wide plains between Manfredonia and Bisceglie and, in particular, the inner shores of the lagoons and the riverine areas were colonised and settled by Neolithic communities. As suggested by the research at Pulo di Molfetta (Fiorentino & Primavera 2007; Fiorentino 2002a), the palaeovegetation was probably characterised by mixed deciduous woods. At the beginning of the Middle Neolithic (fig. 6) the sea level in the Mediterranean appears to have risen to –3 m (Antonioli et al. 1998), where it probably then stood for several millennia. Neolithic settlement in both areas reached its maximum extent in this period (Cassano & Manfredini 2005; Radina 2006). While the sea level appears to have been close to the present level, the coastline of Tavoliere none-

theless took on quite a different shape, retreating on the north where it is close to the Gargano Headland while extending out many kilometres offshore in the middle. The inner lagoon shoreline seems to have been quite indented due to river-mouths and streams flowing into the lagoon, and water-flow appear to have been sufficient to guarantee fresh water throughout the year for human settlement. The formation and development of dune ridges, and the subsequent maximum expansion of the lagoon, suggest that the climate, though temperate, was wetter than at present; this is also reflected in the Ariscianne wetland phase 4930-4390 BC (fig. 3). Palaeovegetation is characterised by trees and macchia shrubs (Fiorentino 2002a). Towards the end of the Middle Neolithic (fig. 6) there appears to have been a phase of de-population in the Tavoliere and the coastal areas of the Murge (Cassano & Manfredini 2005; Radina 2006), set in motion by a deep environmental crisis of aridification. The coastal area of Tavoliere experienced predesert conditions (Boenzi et al. 2001) characterised by low rainfall and high temperatures. As a result, the lagoon was cut off from the sea and turned into a sabkha or supra-tidal plain, as witnessed by the pres-

188

Caldara – Muntoni – Fiorentino – Primavera – Radina, A hidden Neolithic landscape in Apulia…

Figure 6 – The Tavoliere landscape in the Early, Middle and Late Neolithic (modified from Caldara et al. 2002, 2005b).

189

Hidden Landscapes

ence of gypsum finds crystallized in different shapes (such as the ‘desert rose’; Caldara & Pennetta 1993; Caldara et al. 2002). Data from the Ariscianne wetland tend to confirm a period (4390-3240 BC) of reduced water flow (fig. 3), and the Pulo study (fig. 4) also points to the spread of schlerophyllous Mediterranean taxa in the palaeovegetation (Primavera et al. 2007).

5. Final considerations Various geomorphological characteristics of the coastal plains considered here have been highlighted by the present research. In the gulf between Manfredonia and Bisceglie, the coastal plain is characterised by a low developed dune system isolating brackish coastal basins characterised by different degrees of confinement (Caldara et al. 2002). Neolithic sites are distributed along the fluvial valleys (from the Candelaro to the Ofanto rivers) in the Tavoliere plain and along the short incised streams in the north-west of the Murge. Coastal settlements can only be found along the inner margins of the lagoons (Tinè 1983; Cassano & Manfredini 2005; Radina 2006). In the Murgia south of Bisceglie, the coastline is low-lying, rocky, and cut only by a number of small streams. The transport of solids is low and the longshore drift is considerably reduced compared to the Tavoliere coast. Submerged by the sea, the lower reaches of these streams can really be considered as rias, the mouths of several of which were barred by sediments transported by longshore currents, causing the formation of small brackish basins (Caldara et al. 2002). Two types of Neolithic settlement can be distinguished in the coastal Murge area (Radina 2002): inland sites located near agricultural terracing and incised stream, and coastal sites linked to wet areas and temporary ponds. A great variability in the production of cereal crops, which seems already well developed in terms of both cultivation and food processing techniques, is reflected in the archaeobotanical data for the whole region. Analysis of charred vegetal remains from wall daub and pottery from Neolithic sites in the Tavoliere plain and the Murgia shows the presence of hulled wheats such as Triticum diccoccum and Triticum monococcum and hulled barley (Hordeum sp.) together with the more productive naked wheats (T. aestivum/durum) (Fiorentino 2002b; Costantini & Stancanelli 1994). In the Tavoliere plain cultivation of T. spelta also seems to have been important, while in the Murgia area naked wheats and hulled

barley, particularly Hordeum distichum and H. vulgare, seem to be more widespread, as is the presence of minor cereals, legumes and even olives and grapes. These differences can probably be attributed to distinct environmental and eco-pedological conditions as well as different agricultural practices (Fiorentino 2002b). Research here has confirmed the validity of a multi-disciplinary analysis and integration of proxy data for the reconstruction of palaeoenvironmental and hidden landscapes. Preliminary results suggest a strong relationship between environmental dynamics and the distribution of Neolithic sites in the Tavoliere and Murge over more than two millennia (between the end of the 7th and the beginning of the 4th millennium BC) during the Neolithic. Further research needs to be done to understand the establishment and consolidation of agricultural communities in Apulia and so understand both the human role in shaping the landscape and the role of environmental change in affecting human behaviour.

References Antonioli F., Borsato A., Frisia S. & Silenzi S. 1998. L’uso degli speleotemi per ricostruzioni paleoclimatiche e variazioni del livello del mare. Italian Journal of Quaternary Sciences 11(1), pp. 67-78. Boenzi F., Caldara M., Moresi M. & Pennetta L. 2001. History of the Salpi lagoon-sabhka (Manfredonia Gulf, Italy). Italian Journal of Quaternary Science 14(2), pp. 93-104. Biancofiore F. & Coppola D. (eds) 1997. Scamuso. Per la storia delle comunità umane tra il VI e il III millennio nel Basso Adriatico. Roma, Dipartimento di Storia dell’Università di Tor Vergata. Bradford J. 1957. Ancient landscapes: studies in field archaeology. London, Bell & Sons. Caldara M. & Pennetta L. 1993. Ambienti aridi del tipo “sabkha” nei sedimenti olocenici della piana costiera fra Manfredonia e Zapponeta. Bonifica 8(3), pp. 73-82. Caldara M. & Pennetta L. 2002. L’ambiente fisico delle Murge durante il Neolitico, in Radina F. (ed.), Paesaggi, uomini e tradizioni di 8.000 anni fa. La Preistoria della Puglia. Bari, Mario Adda Editore, pp. 21-26. Caldara M., Pennetta L. & Simone O. 2002. Holocene Evolution of the Salpi Lagoon (Puglia, Italy). Journal of Coastal Research 36, pp. 124-133. Caldara M., Caroli I., Lopez R., Muntoni I.M., Radina F., Sicolo M. & Simone O. 2005a. I primi risultati sulle ricerche nel sito di Belvedere – Ariscianne (Barletta), in Atti del 25° Convegno Nazionale sulla Preistoria, Protostoria e Storia della Daunia. San Severo, Italy 3-5 December 2004. Foggia, Stab. Lit. Centrografico Francescano, pp. 99-138. Caldara M., Pennetta L. & Simone O. 2005b. L’ambiente fisico nell’area dell’insediamento, in Cassano S.M. & Manfredini A. (eds), Masseria Candelaro. Vita quotidiana e mondo ideo-

190

Caldara – Muntoni – Fiorentino – Primavera – Radina, A hidden Neolithic landscape in Apulia…

logico in un comunità neolitica del Tavoliere. Foggia, Claudio Grenzi Editore, pp. 25-38. Cassano S.M. & Manfredini A. 1983. Studi sul Neolitico del Tavoliere della Puglia, indagine territoriale in un’area campione. Oxford, British Archaeological Reports. Cassano S.M. & Manfredini A. (eds) 2005. Masseria Candelaro. Vita quotidiana e mondo ideologico in un comunità neolitica del Tavoliere. Foggia, Claudio Grenzi Editore. Costantini L. & Stancanelli M. 1994. La Preistoria agricola dell’Italia centro-meridionale: il contributo delle indagini archeobotaniche. Origini XVIII, pp. 149-243. Fiorentino G. 1999. Gli scavi all’ex fondo Azzolini: risultati delle recenti campagne e prospettive di ricerca. Studi Molfettesi 9-11, pp. 25-32. Fiorentino G. 2002a. Il paleoambiente e le variazioni della vegetazione in Puglia all’inizio dell’Olocene, in Radina F. (ed.), Paesaggi, uomini e tradizioni di 8.000 anni fa. Bari, Mario Adda Editore, pp. 27-32. Fiorentino G. 2002b. I più antichi agricoltori ed i processi di sfruttamento delle risorse vegetali, in Radina F. (ed.), Paesaggi, uomini e tradizioni di 8.000 anni fa. Bari, Mario Adda Editore, pp. 221-225. Fiorentino G. & Primavera M. 2007. Analisi archeo- e paleo-botaniche al Pulo di Molfetta: il villaggio neolitico e la dolina, in Radina F. (ed.), Natura, Archeologia e Storia del Pulo di Molfetta. Bari, Mario Adda Editore, pp. 109-112. Fiorentino G., Muntoni I.M. & Radina F. 2000. La neolitizzazione delle Murge baresi: ambienti, insediamenti e attività produttive, in Pessina A. & Muscio G. (eds), La Neolitizzazione tra oriente e occidente. Udine, Museo Friulano di Storia Naturale, pp. 381-412. Geniola A. 1987. Il Neolitico della Puglia centrale, in Atti della XXV Riunione Scientifica dell’Istituto Italiano di Preistoria e Protostoria “Preistoria e Protostoria della Puglia Centrale”. Monopoli, Italy 16-19 October 1984. Istituto Italiano di Preistoria e Protostoria, Firenze, pp. 55-86. Muntoni I.M. 2002. Gli insediamenti del basso corso ofantino, in Radina F. (ed.), La preistoria della Puglia. Paesaggi, uomini, tradizioni di ottomila anni fa, Bari, Mario Adda Editore, pp. 43-49. Muntoni I.M., 2003. Modellare l’argilla. Vasai del Neolitico antico e medio nelle Murge pugliesi. Firenze, Istituto Italiano di Preistoria e Protostoria. Muntoni I.M. 2007. E la terra si riappropriò del Pulo…: processi di accumulo e classi ceramiche, in Radina F. (ed.), Natura, archeologia e storia del Pulo di Molfetta. Bari, Mario Adda Editore, pp. 207-212. Primavera M. 2008. Ricostruzione della dinamica ambientale delle aree umide costiere pugliesi durante l’Olocene attraverso i resti vegetali. PhD thesis, University of Bari. Primavera M., Radina F., Maggiore M., Caldara M. & Fiorentino G. 2007. The exploitation of environmental resources during the Neolithic period at Pulo di Molfetta (Italy) by insite and off-site archaeobotanical investigations. Quaternary International 167-168, Supplement 1, p. 332. Radina F. 2002. Paesaggi, uomini e tradizioni di 8.000 anni fa. Bari, Mario Adda Editore. Radina F. 2006. Rapporti e scambi tra le più antiche comunità neolitiche in Puglia sulla base dell’indicatore ceramico, in Atti della XXXIX Riunione Scientifica dell’Istituto Italiano

di Preistoria e Protostoria “Materie prime e scambi nella preistoria italiana”. Firenze, Italy 25-27 November 2004. Istituto Italiano di Preistoria e Protostoria, Firenze, pp. 1049-1059. Radina F. 2007. L’insediamento preistorico al Pulo di Molfetta, in Radina F. (ed.), Natura, archeologia e storia del Pulo di Molfetta. Bari, Mario Adda Editore, pp. 89-107. Radina F. (in press), Raffigurazioni simboliche e culti in grotta: una statuetta in argilla del Neolitico da Grotta della Tartaruga di Lama Giotta (Bari), in Atti della XLII Riunione Scientifica dell’Istituto Italiano di Preistoria e Protostoria “L’arte preistorica in Italia”. Trento, Italy 9-13 October 2007. Istituto Italiano di Preistoria e Protostoria, Firenze. Renault-Miskovsky J. & Bui-Thi-Mai 1997. Étude pollinique du site néolithique de Scamuso (Bari, Italie), in Biancofiore F. & Coppola D. (eds), Scamuso. Per la storia delle comunità umane tra il VI ed il III millennio nel Basso Adriatico. Roma, Dipartimento di Storia dell’Università di Tor Vergata, pp. 185-197. Segre A.G. 1969. Linee di riva sommerse e morfologia della piattaforma continentale italiana relative alla trasgressione marina versiliana. Quaternaria 11, pp. 141-154. Simone O. 2003. Evoluzione olocenica e dinamica ambientale delle piane costiere pugliesi. Ph.D. thesis, University of Bari. Tinè S. 1983. Passo di Corvo e la civiltà neolitica del Tavoliere. Genova, Sagep.

191

23 Early farming landscapes in the Friuli plain (NE Italy) Alessandro FontanaI, Andrea PessinaII Dipartimento di Geografia, Università degli Studi di Padova, via del Santo, 26, 35123 Padova, Italy – [email protected] II Soprintendenza dei Beni Archeologici dell’Abruzzo, Via dei Tintori, 1, 66100 Chieti, Italy – [email protected] I

Abstract The Friuli plain is part of the northernmost large alluvial environment in the Mediterranean – the Po basin. After the Last Glacial Maximum (LGM) many areas were not affected by later sedimentation, leading to a good visibility of Early Neolithic sites even along the coast, and permitting us to postulate some hypotheses about the associated settlement strategies. Cultural elements indicate that strong relationships existed with other cultures of the Po Plain, but there were also important links with some of the Dalmatian cultural groups and with those on the Balkans. In the upper part of the plain, the Neolithic settlements are concentrated almost exclusively in a few and small areas with particular geo-pedological characteristics (e.g. high fertility and water availability). In the lower part of the plain, where the groundwater table is high, the settlements are quite far apart and are situated on the top of LGM fluvial ridges. These are characterized by sandy-silty levees with well-drained and developed soils, contrasting with the waterlogged adjacent flood plain. Settlements show a linear pattern which is comparable to the pattern of Neolithic settlements observed in the inland sections of several other European alluvial plains. In NE Italy numerous radiocarbon datings for the first farming settlements demonstrate the contemporaneity between neolithization and lagoon formation, suggesting the important role played by this ecotonal environment.

1. Introduction The Friuli plain extends from the Livenza river to the classical Karst Plateau and corresponds to the eastern sector of the Po Plain (fig. 1). This is the northernmost alluvial plain in the Mediterranean basin and it represents the natural continental pathway for cultural and economic transactions between northern Italy and the regions of the Carpathian Balkans and of the northern side of the Eastern Alps. The Neolithic settlements in the study area belong almost all to the early phase, and their presence is generally limited to the alluvial plain. In the last decades many archaeological and geoarchaeological investigations were carried out here, and the combination of geological, palaeo-ecological and archaeological data produced interesting information on the relationships between natural environments and human communities during the Early Neolithic. Large extant surfaces of the Friuli Plain date back to the end of the upper Pleistocene and this results in a high visibility of Neolithic sites; which makes it possible to investigate the diffusion of the first farmers even along the coast and, therefore, to consider some hypotheses about their settlement strategies. The research reported here focuses on the geomorphological and pedological setting which constrained the neolithization (second half of the 5th millennium BC), on the settlement pattern of the villages and on the environmental changes induced by the new human activities (e.g.

agriculture, animal husbandry and settlement construction). The work analyzes general aspects of the Neolithic landscape, whereas the in-site situation is not discussed in detail.

2. Materials and methods This work is mostly based on a review of the available data on the archaeology and palaeo-environment of the Friuli Plain. The geomorphological characteristics have been reconstructed with a detailed survey of the distal (seaward) sector of the plain (Bondesan et al. 2004; Fontana 2006; Zanferrari 2008a, 2008b). The relations between palaeohydrography, pedology and stratigraphic position have been analyzed for each site (Ferrari & Pessina 1992, 1996; Fontana 1999, 2006). The high resolution geoarchaeological surface survey covered the entire lower part of the plain between the Tagliamento and Corno di Nogaro rivers; in the higher part of the plain selected areas have been investigated (Ferrari & Pessina 1992; Pessina 2006; Fontana 2006). Archaeological excavations were realized in the areas of Sammardenchia (Pessina et al. 1998; Ferrari & Pessina 1999; Pessina 2006), Piancada (Pessina et al. 1998; Ferrari & Pessina 1996), Fagnigola (Biagi 1975; Ferrari & Pessina 1996; Pessina et al. 1998), Valler (Fasani et al. 1994) and Pavia di Udine (Pessina et al. 2004; Pessina 2006). In the sites of Sammardenchia, Piancada and Precenicco the reconstruction of land use and ag-

193

Hidden Landscapes

ricultural practises is based also on detailed micromorphological analyses of some stratigraphical samples (Ottomano 1999; Fontana 2006). The palaeo-ecological aspects have been mainly supported by a recent review of palaeobotanical information for the whole Friuli region (Rottoli & Pessina 2007) and by archaeo-zoological investigations carried out in Piancada, Fagnigola and Nogaredo al Torre (Petrucci et al. 2005). In figure 2, the 85 radiocarbon dates available for the Early Neolithic sites between the Karst Pla-

teau and the Piave river (Improta & Pessina 1998; Pessina 2006) have been plotted. The diagram also shows the age estimates of the base of the lagoonal deposits in the NE Italy, in particular for the Marano and Caorle lagoons (Marocco 1991; Galassi & Marocco 1999; Fontana 2006) and for the sector between the Piave river and the northern Venice Lagoon (Bondesan & Meneghel 2004; Canali et al. 2007). The dates are expressed in conventional 14C years with uncalibrated years Before Present. The average standard deviation of the radiocarbon dates

Figure 1 – Geomorphological sketch of the Friuli Plain (modified from Fontana 2006); 1. rivers; 2. fluvial terraces; 3. northern limit of spring belt; 4. isobate curve; 5. 0 m a.s.l. contour line; 6. major Early Neolithic site; 7. tectonic terraces; 8. Alps; 9. moraine system; 10. LGM gravelly deposits; 11. LGM fine deposits; 12. post-LGM gravelly deposits; 13. post-LGM fine deposits; 14. Holocene coastal deposits. The rectangle corresponds to fig. 3.

194

Fontana – Pessina, Early farming landscapes in the Friuli plain (NE Italy)

is ±80 years and the statistical mode of the standard deviations is also about ±80 years.

3. Cultural settings and chronological elements of the Early Neolithic in North-East Italy Most of the Neolithic sites in the Friuli Plain belong to the early phase (7500-6800 years cal. BP), but the excavations often yield evidence for a slightly longer occupation or of phases of reoccupation until the beginning of the Middle Neolithic. The pottery assemblages from the sites show contacts between the Po Plain (Fiorano culture) and the eastern Adriatic coast (Danilo culture) (Ferrari & Pessina 1999). Analyses of the raw materials reveal the existence of large exchange networks, connecting the Friuli Plain Neolithic sites with the north-western Alps (greenstone: Pessina & D’Amico 1999), the Monti Lessini area (flint: Ferrari & Mazzieri 1998) and southern Italy (obsidian from Lipari: De Francesco et al. 1998). Paleobotanical data from all the excavated sites witness the existence of fully developed farming economies during the Early Neolithic (Rottoli & Pessina 2007). Many cultivated cereals (einkorn, emmer, spelt, barley and wheat) were available, in addition to pulses (pea, lentil, and vetch). Wild fruits (hazelnuts, acorn, sloe, cornel, blackberry and elderberry seeds) were also collected and perhaps stored after roasting, as for example hazelnuts. Degradation of the vegetation (a dense mixed oak forest) close to the Neolithic villages is suggested by the quantity of charcoal of pomaceous trees in the archaeobotanical record (Rottoli 1999).

4. Late Pleistocene and Holocene evolution and archaeological visibility in the Friuli Plain In the Friuli Plain the main phase of alluvial sedimentation took place during the Last Glacial Maximum (LGM, corresponding to the period 24,000-15,000 14 C years BP) (Fontana et al. 2008, 2010). During the first part of the LGM, the so-called pleniglacial phase (30,000-20,000 years cal BP), the fluvioglacial activity caused a differentiation between the so-called high and low plain, which correspond respectively to the proximal gravelly sector and to the distal clayey-silty plain (fig. 1). During the last period of the LGM, the so-called cataglacial phase (20,000-17,000 years cal BP), the beginning of deglaciation led the Alpine fluvial systems to entrench in the proximal sector of their megafan. Sedimentation continued only in some sectors of the distal plain, with the activation of several fluvial branches;

Figure 2 – Plot diagram of the radiocarbon datings of the Early Neolithic sites of the Friuli Plain (circles) and of the base of the lagoonal sediments in NE Italy (triangles).

along these directions in the area near the presentday lagoon, narrow fluvial ridges were built (fig. 3a). Since the beginning of the post-LGM the alluvial systems of the Cormor and Corno di S. Daniele rivers were abandoned and large sectors of the alluvial megafans of the Tagliamento, Torre and Isonzo rivers were deactivated (Fontana et al. 2008). Thus, in the last 17,000 years large areas of the plain ceased to be affected by later sedimentation, and the Holocene archaeological sites in these areas are exposed. Due to the permeability of the coarse gravelly sediments, the water table in the high plain is at least 30-50 m deep and there is almost no natural surface drainage. By contrast, at the boundary with the distal plain the presence of clayey-silty deposits forces part of the groundwater to emerge, feeding a number of springs. These are aligned along the so-called spring belt (fig. 1) and support an important network of groundwater-fed rivers which developed on the LGM surface, creating several shallow and narrow incisions. The low plain is also characterized by a rolling topography due to the alternation between cataglacial fluvial ridges and overbank floodplains (fig. 3); generally the groundwater table is at a depth of 1 to 2.5 m. The high gravelly plain is characterized by a homogeneous topography, interrupted by a few large, deep post-LGM incisions of the main rivers (fig. 1). Other geomorphological features are the raised terraces south of Udine, which are connected to the neotectonic activity of the Julian and Carnic Alps (Fontana 1999; Zanferrari 2008a). The geomorphological and hydrological differences between high and low plain led to the formation of very different soils which greatly influenced ancient settlement. The post-LGM marine transgression began to affect the Friuli area in the middle Holocene and triggered the formation of deltas and lagoons. Compari-

195

Hidden Landscapes

Figure 3 – Top: geomorphological map of the south-western sector of the Friuli Plain. 1. Pleniglacial alluvial deposits; 2. cataglacial fluvial ridges; 3. post-LGM incision of Stella river; 3a. post-LGM deposits of Stella river; 4. post-LGM deposits of Tagliamento; 5. present day ridge of Tagliamento river; 6. lagoonal deposits; 7. coastal deposits; 8. incised paleo-channel of Tagliamento; 9. Holocene abandoned channels; 10. fluvial scarp; 11. Mesolithic site; 12. major Neolithic site; 13. minor Neolithic site; 14. Eneolithic site; 15. landward limit of lagoon around 5000 BC. Letters indicate major sites: P: Precenicco; N: Piancada Nogali; L: Piancada Latteria; f: Piancada Fraida; B: Muzzana Bonifica; M: Muzzana Bosco Comunale. Figure modified from Fontana 2006. Bottom: cross-section of a Late LGM fluvial ridge of the Cormor system (for location, see above, n. 16). 1. Channel deposit (gravelly sand with cross-bedding); 2. natural levee (silty sand); 3. overbank floodplain (clayey silt); 4. LGM floodplain (clayey silt); 5. channel deposit (fine–medium sand); 6. peat, organic clay; 7. carbonate calcium concretions; 8. Neolithic dwelling structures. Figure modified after Fontana 2006.

196

Fontana – Pessina, Early farming landscapes in the Friuli plain (NE Italy)

sons between the age estimates of the base of lagoonal sediments and of the neolithization process (fig. 2) indicate that in the Veneto-Friuli Plain these two processes were partly overlapping; this meant that the first farmers could also exploit the multivariate resources of the lagoonal environment. The sea-level rise occurring after the Neolithic caused the lagoon to extend landwards, affecting also some of the areas that had been previously influenced by human occupation. Many of the prehistoric sites of the low plain have therefore re-emerged thanks to the 20th-century reclamations which took place along the lagoonal fringe.

5. Settlement patterns in the high plain In the high plain, Neolithic settlements were concentrated almost exclusively in the few and very small areas with silty deposits. These places are generally characterized by geomorphological and pedological properties that differ greatly from the homogeneity of the LGM coarse-gravelly sediments that form the rest of the high plain. The most important archaeological area is Sammardenchia (Pozzuolo del Friuli), where the archaeological finds cover about 6 sq.km without interruptions. The highest concentration of finds has been recorded on a tectonic terrace (figs 1 and 4). This is a feature raised 2-4 m above the surrounding plain, and the loamy soil on top of the terrace has a natural fertility which is almost the highest in the Friuli Plain and which

dramatically contrasts with the fairly poor soils of the adjacent zones. Along the slopes bounding the terrace some springs existed, and these were some of the rare places where natural water emerges in the area between the moraine system and the spring belt. The springs disappeared during the 20th century following extensive levelling and deep ploughing which strongly altered the original morphology of the scarps. The excavations at Sammardenchia produced evidence for a large complex of structures with many reoccupation phases and fairly large villages. Stratigraphy and radiocarbon dates demonstrate that the area remained settled for about 1000 14C years. As was demonstrated for the central Po Plain (for a review see Cremaschi 2000, pp. 292-293), the first farmers occupied a large area and returned cyclically to the same places where they found favourable conditions for agriculture. This reoccupation cycle produced an extremely large archaeological area, and a completely different settlement pattern than that in the distal sector of the plain. Hundreds of Neolithic dwellings have been recorded: not only pits, storage-structures and post-holes, but also perimeter ditches, trenches and water-reservoirs. This suggests a Neolithic landscape where isolated farmhouses coexisted with larger settlements. Micromorphological studies indicate that, even if the activities of the first farmers triggered the ero-

Figure 4 – Digital topographic model of the tectonic terrace of Pozzuolo del Friuli in the area of Sammardenchia. Black dots correspond to the Early Neolithic structures and light-grey circles to water springs. Figure modified after Fontana 1999.

197

Hidden Landscapes

sion of the soils, pre-existing pedoclimate conditions were restored after human abandonment (Cremaschi 1990; Ottomano 1999). At the site of Pavia di Udine (fig. 1) the surface is characterized by the presence of a 2 to 4 m-thick layer of silt. Its properties are comparable to the surface of the top of the terrace of Sammardenchia. The fine sediment also hosts a seasonal water table within a depth of 2-3 m, whereas in the high plain it normally lies at 30-50 m.

6. Settlement pattern in the low plain In the distal sector of the plain, the Neolithic presence is rather spread out, especially along the modern lagoon fringe. The larger settlements are those of Fagnigola, Valler, Piancada, Precenicco, Muzzana Bonifica and Muzzana Bosco Comunale (fig. 2); about 30 minor Early Neolithic sites are known. All settlements were found on top of narrow cataglacial fluvial ridges (fig. 3, top), characterized by a gravellysand channel bordered by sandy-silt levees that are elevated 1-2 m above the flood plain. The raised position of the levees and their fairly good permeability favoured pedogenesis, allowing the formation of well-developed soils (sols bruns lessivés); in contrast, the sediments on the surrounding flood plain and in the ancient channel were almost permanently waterlogged by groundwater, preventing soil formation processes. Soil fertility on the ridges varies greatly; it was excellent on top of the levees, especially for the cultivation of cereals, which prefer light soils and do not tolerate long periods with a high water table. Thanks to the detailed geo-archaeological surface survey and the high archaeological visibility of the prehistoric remains, we can be sure that no dwellings were present in the flood plain: Neolithic finds and structures perfectly overlap with the sandy-silt levees. The remains suggest that settlement zones were narrow (100-200 m wide) but had an impressive length (500-3000 m). As for Sammardenchia, the presence of large areas with settlement remains seems to be related to the cyclic migration of villages, driven by a search for new resources and good soils to exploit. The position of sites along the ridges could be described as a ‘linear dwelling system’ resembling the pattern of Neolithic settlements observed in the distal sector of several other European alluvial plains. A similar pattern has been described for the Körös culture in the Alföld region of Hungary (Makkay 1998), were the villages occupied the tops of the silty levees (largely inactive since the Lateglacial), whereas the flood plain was prone

to seasonal flooding till last century. In the RhineMeuse Delta (the Netherlands) an important linear concentration of Neolithic settlements has also been recorded on top of the Lateglacial river dunes that rise several metres above the plain and had a fairly well-developed soil thanks to the presence of sandy drained deposits (Louwe Kooijmans 1993). In the area of Piancada-Precenicco, the presence of archaeological structures such as ditches suggests the possibility that systems of irrigation or drainage existed since the Early Neolithic. In the same area micromorphological analyses carried out on Neolithic surfaces indicate large concentrations of charcoal and pedorelicts in off-site areas; this probably relates to deforestation and to agricultural practices such as slash-and-burn (Fontana 2006). Archaeozoological remains testify to the great importance of animal husbandry in the Early Neolithic (especially ovicaprines, but also pigs and cattle), having an impact on the forest due to leaf fodder harvesting and grazing (Petrucci et al. 2005). As mentioned above, the formation of the lagoons led to a new environment in the coastal plain, with many and varied ecological resources. At the site of Piancada this is testified by the occurrence of quantities of Cerastoderma glaucum, which is a typical lagoonal shell. During the Early Neolithic the landward limit of the lagoon was a few kilometres further downstream than at present (symbol 15 in fig. 3, top), but the connection between lagoon/sea and the main site at Piancada-Precenicco and Muzzana was assured by the groundwater-fed rivers. Even during the Neolithic, these rivers were navigable upstream for several kilometres thanks to the low gradient of the plain. Unfortunately, in spite of the excellent archaeological visibility in the Friuli plain and the recent emergence of some sites as a result of reclamation, it is still possible that several prehistoric settlements are submerged in the Marano and Grado lagoons. Therefore, the reconstruction of the settlement pattern that has been proposed here may be incomplete.

References Biagi P. 1975. Stazione neolitica a Fagnigola (Azzano DecimoPN). Relazione preliminare dello scavo 1974. Annali Univ. Ferrara 15 (2-6), pp. 247-269. Bondesan A. & Meneghel M. (eds) 2004. Geomorfologia della provincia di Venezia, Esedra, Padova. Bondesan A., Meneghel M., Rosselli R. & Vitturi A. (eds) 2004. Geomorphological Map of the Province of Venice, scale 1:50.000. LAC, Firenze, 4 sheets.

198

Fontana – Pessina, Early farming landscapes in the Friuli plain (NE Italy)

Canali G., Capraro L., Donnici S., Rizzetto F., Serandrei-Barbero R. & Tosi L. 2007. Vegetational and environmental changes in the eastern Venetian coastal plain (Northern Italy) over the past 80,000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 253, pp. 300–316. Cremaschi M. 1990. Pedogenesi medio olocenica ed uso dei suoli durante il Neolitico in Italia settentrionale, in Biagi P. (ed.), The Neolithisation of the Alpine Region. Monografie di Natura Bresciana 13, Brescia, pp. 71-89. Cremaschi M. (ed.) 2000. Geoarcheologia. Manuali Laterza, Editori Laterza, Bari. De Francesco A.M., Crisci G.M. & Lanzafame U. 1998. Metodo analitico non distruttivo in Fluorescenza X per risalire alla provenienza di ossidiane archeologiche nell’area mediterranea, in Pessina A. & Muscio G. (eds), Settemila anni fa il primo pane. Ambienti e culture delle società neolitiche, Catalogo della mostra, Udine, pp. 171-176. Fasani L., Biagi P., D’Amico C., Starnini E. & Voytek B.A. 1994. Stazione neolitica a Valer (Azzano Decimo – Pordenone): rapporto preliminare degli scavi 1990-91. Atti Società Preistoria e Protostoria Regione Friuli Venezia-Giulia, Trieste 8, pp. 97-113. Ferrari A. & Mazzieri P. 1998. Fonti e processi di scambio di rocce silicee scheggiabili, in Pessina A. & Muscio G. (eds), Settemila anni fa il primo pane. Ambienti e culture delle società neolitiche, Catalogo della mostra, Udine, pp. 165-169. Ferrari A. & Pessina A. 1992. Considerazioni sul primo popolamento neolitico dell’area friulana. Atti della Società di Preistoria e Protostoria Regione Friuli-Venezia Giulia, Trieste 6, pp. 23-60. Ferrari A. & Pessina A. (eds) 1996. Sammardenchia e i primi agricoltori del Friuli. Banca di Credito Cooperativo di Basiliano, Arti Grafiche Friulane, Udine. Ferrari A. & Pessina A. (eds) 1999. Sammardenchia-Cueis contributi per la conoscenza di una comunità del primo Neolitico. Monografie Museo Friulano Storia Naturale 41, Udine. Improta S. & Pessina A. 1998. La neolitizzazione dell’Italia settentrionale. Il nuovo quadro cronologico, in Pessina A. & Muscio G. (eds), Settemila anni fa il primo pane. Ambienti e culture delle società neolitiche, Catalogo della mostra, Udine, pp. 63-71. Fontana A. 1999. Geomorfologia dell’area di Sammardenchia, in Ferrari A. & Pessina A. (eds), Sammardenchia-Cueis contributi per la conoscenza di una comunità del primo Neolitico, Monografie Museo Friulano Storia Naturale 41, Udine, pp. 11-22. Fontana A. 2006. Evoluzione geomorfologica della bassa pianura friulana e sue relazioni con dinamiche insediative antiche. Monografie Museo Friulano Storia Naturale 47. Fontana A., Mozzi P. & Bondesan A. 2008. Alluvial megafans in the Venetian-Friulian Plain: evidence of aggrading and erosive phases during Late Pleistocene and Holocene. Quaternary International 189, pp. 71-90. Fontana A., Mozzi P. & Bondesan A. 2010. Late Pleistocene evolution of the Venetian-Friulian Plain. Rendiconti Lincei 21, Suppl. 1, pp. 181-196. Galassi P. & Marocco R. 1999. Relative sea-level rise, sediment accumulation and subsidence in the Caorle lagoon (Northern

Adriatic Sea, Italy) during the Holocene. Il Quaternario 12 (2), pp. 249-256. Improta S. & Pessina A. 1998. La neolitizzazione dell’Italia settentrionale. Il nuovo quadro cronologico, in Pessina A. & Muscio G. (eds), Settemila anni fa il primo pane. Ambienti e culture delle società neolitiche, Catalogo della mostra, Udine, pp. 107-115. Louwe Kooijmans P. 1993. The Mesolithic/Neolithic transformation in the Lower Rhine-Meuse basin, in Bogucki P.I. (ed.), Case Studies in European Prehistory, CRC Press, pp. 85-180. Makkay J. 1998. I primi agricoltori dell’Europa sud-orientale e il Neolitico del Bacino dei Carpazi, in Pessina A. & Muscio G. (eds), Settemila anni fa il primo pane. Ambienti e culture delle società neolitiche, Catalogo della mostra, Udine, pp. 35-54. Marocco R. 1991. Evoluzione tardopleistocenica-olocenica del Delta del F. Tagliamento e delle lagune di Marano e Grado (Golfo di Trieste). Il Quaternario 4 (1b), pp. 223-232. Ottomano C. 1999. Suoli e uso del suolo nel Neolitico dell’Italia nord-orientale, in Ferrari A. & Pessina A. (eds), Sammardenchia-Cueis contributi per la conoscenza di una comunità del primo Neolitico, Monografie Museo Friulano Storia Naturale 41, Udine, pp. 133-145. Pessina A. 2000. Il primo Neolitico dell’Italia settentrionale. Problemi generali, in Pessina A. & Muscio G. (eds), Atti del Convegno: “La neolitizzazione tra Oriente e Occidente”, Udine 22-23 aprile 1999. Museo Friulano Storia Naturale, Udine, pp. 81-90. Pessina A. 2006. Nuovi dati sugli aspetti culturali del Primo Neolitico in Friuli e sui rapporti con l’Adriatico orientale, in Pessina A. & Visentini P. (eds), Preistoria dell’Italia settentrionale. Studi in ricordo di Bernardino Bagolini. Atti del Convegno, Udine 23-24 settembre 2005, pp. 279-302. Pessina A., Ferrari A. & Fontana A. 1998. Le prime popolazioni agricole del Friuli, in Pessina A. & Muscio G. (eds), Settemila anni fa il primo pane. Ambienti e culture delle società neolitiche, Catalogo della mostra Udine 1997-98, Museo Friulano Storia Naturale, Udine, pp. 133-145. Pessina A., Fiappo G.C. & Rottoli M. 2004. Un sito neolitico a Pavia di Udine. Nuovi dati sull’inizio dell’agricoltura in Friuli. Gortania, Atti del Museo Friulano di Storia Naturale 25, Udine, pp. 73-94. Pessina A. & D’Amico C. 1999. L’industria in pietra levigata del sito neolitico di Sammardenchia (Pozzuolo del Friuli, Udine). Aspetti archeologici e petroarcheometrici, in Ferrari A. & Pessina A. (eds), Sammardenchia-Cueis contributi per la conoscenza di una comunità del primo Neolitico, Monografie Museo Friulano Storia Naturale 41, Udine, pp. 23-78. Pessina A. & Rottoli M. 2007. Neolithic Agriculture in Italy: an update of archaebotanical data with particular emphasis on northern settlements, in Colledge S. & Connoly J. (eds), The origins and spread of domestic plants in Southwest Asia and Europe, University College London, Institute of Archaeology Publications, pp. 1411-1454. Petrucci G., Pessina A., Visentini P. & Vitri S. 2005. Allevamento e caccia nei siti neolitici del Friuli, in Malerba G. & Visentini P. (eds), Atti del IV convegno di Archeozoologia. Quaderni del Museo Archeologico del Friuli Occidentale 6, pp. 161-167.

199

Hidden Landscapes

Rottoli M. 1999. I resti vegetali di Sammardenchia – Cûeis (Udine), insediamento del Neolitico antico, in Ferrari A. & Pessina A. (eds), Sammardenchia-Cueis contributi per la conoscenza di una comunità del primo Neolitico, Monografie Museo Friulano Storia Naturale 41, Udine, pp. 307-326. Zanferrari A., Avigliano R., Monegato G., Paiero G. & Poli M.E. 2008a. Note illustrative della Carta Geologica d’Italia alla scala 1:50.000 – Foglio 066 “Udine”. Graphic Linea, Tavagnacco, Udine. Zanferrari A., Avigliano R., Fontana A. & Paiero G. (eds) 2008b. Note Illustrative della Carta Geologica d’Italia alla scala 1:50.000 – Foglio 087 “San Vito al Tagliamento”. Graphic Linea, Tavagnacco, Udine.

200

24 Remote sensing analysis in the Florentine plain: investigating wetland contexts by comparing QuickBird images and multitemporal aerial photographs Riccardo SalviniI, Giovanna PizzioloII, Ludovico SassariniI, Adriana TrottaI, Ivan CallegariI, Lucia SartiII I Earth Sciences Department and Centre of Geotechnologies, University of Siena, via Vetri Vecchi 34, 52027 San Giovanni Valdarno, Arezzo, Italy – [email protected]; [email protected] II Department of Archaeology and Art History, Prehistory Section, University of Siena, via Roma 56, 53100 Siena, Italy – [email protected]; [email protected] Abstract Archaeological investigations carried out in the Sesto Fiorentino area (Florence, Italy) revealed a significant number of prehistoric finds. Landscape archaeological analyses were carried out using recent (2004) and historical (1954 and 1965) aerial photos in conjunction with near infra-red (NIR) images (1997 and 2006) and historical cartography, and several morphological anomalies connected to the palaeo-drainage of the plain were detected. In order to carry out the work, additional technologies such as DGPS and GIS were adopted. DGPS surveys in RTK modality offered autonomy and high accuracy when acquiring ground control points used in photogrammetric processing. Multilayer GIS provided a common platform to analyze all the data, highlighting its use in palaeo-channel identification. Within this research framework, one of the objectives of the study was to illustrate the potential of high-resolution QuickBird satellite images for the identification of palaeo-rivers by means of spectral enhancement. The comparison of different remote sensing data sources pointed out their advantages, characteristics and potentialities for the reconstruction of a prehistoric sedimentary landscape. The research revealed that this approach offers effective tools for the investigation of hidden prehistoric landscape features.

1. Geological and archaeological context The area of Sesto Fiorentino is located north-west of Florence (fig. 1) at the western boundary of the Inner-Northern Apennines, where, following the compressional phases that were responsible for the main tectonic thrusting, an extensional tectonic trend began in the Middle Miocene (Carmignani et al. 2001). This extensional trend led to the partial dismemberment of the previous geo-structural setting (Bartolini et al. 1983) and to the formation of some fluvio-lacustrine basins aligned parallel to the main chain (Boccaletti et al. 1987; Bossio et al. 1992). Sesto Fiorentino is located in the Firenze–Pistoia basin. This NW-SE orientated tectonic depression is around 10 km wide and 30 km long, and is limited to the north-east by a master-fault with Apennine orientation, dipping to the south-west. The fluvio-lacustrine basins, affected by subsidence, were progressively filled up by Plio-Pleistocene continental deposits and subsequent Holocene alluvial sediments (Martini & Sagri 1993). At the bottom of the secondary valleys, and interfingered with sediments of the main valley, systems of alluvial fans crop out. In these lithologically, spatially and morphologically heterogeneous deposits, prehistoric people found their raw materials. Flint fragments were and are still being produced by the slow dismantling of the Apennine chain and its geological formations, which contain, in varying percentages, flint-bearing sediments such as the diaspri of the

ophiolitic series (Bortolotti 1983), the radiolarites of the Ligurian formations, and siliceous sandstones and limestone for diagenetic substitution. People moved into this geomorphological context from the Mesolithic onwards (Sarti & Martini 1993), and between the Neolithic and the Bronze Age the occupation of the study area became increasingly dense. Prehistoric sites expanded towards the south-east, within the inner part of the Sesto Fiorentino plain.

Figure 1 – Location of the study area (panchromatic QuickBird images as background).

201

Hidden Landscapes

obtain more, as well as more accurate, information on the palaeo-drainage context. Particular attention was therefore paid to the full use of the information provided by different sources, focusing on the diachronic comparison of remotely sensed images.

2. Materials and methods

Figure 2 – Palaeo-hydrological reconstruction of the Firenze–Prato plain (from Conedera & Ercoli 1973).

The wide distribution of multiperiod sites, even if some show chronological gaps in their cultural sequence, testifies to a universal interest in specific locations (Pizziolo & Sarti 2005, 2006). Of particular interest here is the tendency of Bell Beaker people to settle in dry river beds (Sarti & Martini 1998). From an archaeological point of view it is crucial to understand the relationship between the palaeoenvironmental setting and the distribution of prehistoric finds. Palaeo-rivers may constitute an attractive resource to be exploited, or may constitute a landscape feature determining prehistoric settlement choices in other ways. Previous geomorphological studies (Conedera & Ercoli 1973; Capecchi et al. 1975) suggested that the inner basin of the Florentine plain was a wetland area during prehistory, and that creeks, marshes and bogs dominated the landscape. Conedera and Ercoli in 1973 proposed the first complete palaeo-hydrological reconstruction of the Firenze– Prato plain, at a 1:50,000 scale (fig. 2). Their interpretation is based on analogue stereo 1954 aerial photos (at a nominal scale of 1:33,000). Many factors played a role during the interpretation of these photos: slight elevation differences, darker colours within grey tones indicating dump marks, road patterns and field boundaries. Whilst their interpretation offers important information on the wetland characteristics of the area, it is not very detailed and thus does not allow an exhaustive and meticulous palaeo-environmental reconstruction. In order to identify the landscape features which may have played a role in prehistoric settlement strategies (see Pizziolo & Sarti, this volume) it is necessary to

The following data were used: • 1954 panchromatic aerial photos at a nominal scale of 1:33,000, produced by the US Army and provided by Italian Geographic Military Institute (IGMI), with camera calibration certificate. • 1965 panchromatic aerial photos at a nominal scale of 1:18,000 produced by IGMI, without camera calibration certificate. • 1997 Near-Infrared (NIR) colour aerial photos at a nominal scale 1:6,000 provided by the Consorzio Alta Velocita Emilia–Toscana (CAVET) and Fiat Engineering and produced during the construction of a high speed railway. • 2004 colour aerial photos at a nominal scale of 1:6,000 provided by Compagnia Generale Riprese aeree (CGR, Parma, IT), with camera calibration certificate. • QuickBird satellite images dated 16 May 2006 (Basic Bundle Images, separate Panchromatic and Multispectral files), with metadata describing sensor internal geometry, optics and external orientation parameters. In order to analyse these images and to enable a comparison of their interpretations, the aerial photos and the satellite images were orthorectified using Erdas™ Imagine 9.2 software. Appropriate pre-processing to deal with the different formats and scales was necessary before any further analysis could be done. The first step was to apply geometric correction, giving the images the same reference system. For this purpose 48 ground control points (GCPs) were collected by a DGPS survey in Real Time Kinematic (RTK) mode executed by a Leica™ 1200 receiver. The correction of the surveyed points was done with a GSM modem connection to the Leica™ Geosystem SpiderNet (www.gpsnet.it/spidernet) network of reference stations. Using VERTO™ software, the WGS84 coordinates were converted from ellipsoidal into orthometric elevation and projected onto the national ‘Roma 1940’ reference system. GCPs were uniformly distributed across each image and were used both in the exterior orientation process of aerial photos and in the orthorectification of QuickBird satellite images, using rigorous methods (Toutin & Cheng 2002).

202

Salvini et alii, Remote sensing analysis in the Florentine plain: investigating wetland contexts…

The parameters to be analyzed in order to obtain a reconstruction of palaeo-drainage on the basis of the image interpretation include vegetation, humidity and micro-relief, as well as changes in soil composition and texture of the fields. For vegetation and moisture, which are generally considered to be good indicators for the interpretation of palaeo-drainage systems, the Normalized Difference Vegetation Index (NDVI) was computed on multispectral satellite images and on colour NIR aerial photographs. The NDVI technique measures the radiation absorbed by chlorophyll during plant photosynthesis in the red and near-infrared sensor channels. The computed NDVI images will therefore show intensity variations that may be connected to palaeodrainage systems. In order to be able to perform the NDVI calculation on NIR images, the aerial photographs were first scanned in RGB mode. The vegetation index was then obtained by determining the ratio between Red (corresponding to the near infrared) minus Green (corresponding to the red) and Red plus Green RGB bands (Eastman Kodak Company 2003). This process does not result in values which can be compared directly with vegetation indexes obtained from multispectral satellite images, but nonetheless is very effective in the identification of palaeo-drainage features (Pizziolo 2010). Other spectral enhancements (e.g., tasselled-cap or IHS) did not lead to the identification of any new palaeo-drainage anomalies or archaeological traces, but sometimes confirmed the results of the NDVI analysis. It is also worth noticing that supervised or unsupervised classifications did not produce any useful results; in fact, palaeo-drainage anomalies are not characterised by a homogeneous signature or marker but are identified essentially on the basis of the effects of dry or wet soils. Analysis of the shape and configuration of agricultural fields may therefore be the more useful approach for identifying ancient palaeo-rivers (Salvini et al. 2006). The problem of proposing a detailed chronological scheme for the sequence and dynamics of drainage systems relative to the various archaeological phases is still unresolved. For this, the results of photo-interpretation need to be correlated with additional data derived from the analyses of historical sources, fieldwork and geophysical investigations.

components of fields. Field boundaries and field shapes, especially narrow and winding ones, were also used to support interpretations of palaeo-rivers or channels. A compilation of the interpreted data was carried out within a GIS framework using ESRI™ ArcGis 9.2 software. The observed anomalies were stored as polygon shapefiles (fig. 3), and their original shape was reconstructed by making use of the convergence of evidence derived from each separate dataset. The fragmentary nature of the evidence and the great impact of human activities during the last thirty years preclude a precise identification with the palaeochannels mapped by Conedera and Ercoli (1973), even if a good general correspondence between the

3. Results Several anomalies were recognized during the integrated study of all imagery, particularly on the basis of different values for the hue and intensity colour

Figure 3 – Map of the identified palaeo-drainage systems in the Sesto Fiorentino area.

203

Hidden Landscapes

two data sets is notable. However, the existence of some anomalies must still be verified by detailed fieldwork and/or geophysical prospection. Some anomalies found on the photos taken in March 2004 were no longer visible on the May 2006 QuickBird images. This indicates that the seasonal condition of crops, as well as other factors, strongly influences visibility. NDVI images were very useful to draw out information unobtainable from traditional photographs, and the NIR images played an essential role in this, because they provided highresolution qualitative data on soil wetness (fig. 4). Digital stereoscopic analysis of all the aerial photos did not lead to the discovery of any related micromorphological variations.

4. Conclusions The data sources used during the analysis produced interesting results. Each of our sources has its own specific strengths in revealing different characteristics of wetland features, and their complementary use is very important for the reconstruction of the prehistoric landscape in these sedimentary basins. Palaeo-rivers are in fact not easily detected by any single method, because their traces are often modified by more recent changes to the landscape. The analysis of the 1954 and 1965 panchromatic aerial photos revealed that the Sesto Fiorentino plain was then still characterised by a traditional rural landscape, not yet made uniform by modern agriculture nor radically changed by urban activities. The photo-interpretation of these datasets allowed

Figure 4 – Example of a palaeo-channel in the Quickbird near infrared band.

us to undertake a valuable look at hidden features of the past. Moreover, it has been particularly motivating to work on the historical photographs used by Conedera and Ercoli, as these allow a more detailed study of the large palaeo-drainage features that had previously been identified by geomorphologists at a scale of 1:50,000. Undoubtedly the multi-temporal approach has led to better results for anomalies on aerial photographs, and allows us to improve our understanding of territorial and landscape changes. The data collected during this project have been entered into a dedicated geodatabase for future spatial analysis, making it possible to link the past landscape to settlement choices and the exploitation of sources of raw materials as well as their circulation, for the various cultural horizons in the area. The results of aerial digital photogrammetry and remote sensing demonstrate that only a multiscale, multitemporal and multisource approach of this type will lead to interesting results. The described techniques enabled us to improve our knowledge of geomorphology and of individual palaeo-channels, but many ambiguities in the interpretation of the data remain. Additional geophysical prospection is therefore recommended and further archaeological studies will be necessary.

References Bartolini C., Bernini M. & Carloni G.C. 1983. Carta Neotettonica dell’Appennino settentrionale. Note illustrative. Boll. Soc. Geol. It. 101, pp. 523-549. Boccaletti M., Decandia F.A., Gasperi G., Gelmini R., Lazzarotto A. & Zanzucchi G. 1987. Carta Strutturale dell’Appennino Settentrionale. Note illustrative. CNR, P.F. Geodinamica, p. 203. Bortolotti V. 1983. Stratigrafia, tettonica ed evoluzione geodinamica delle ofioliti della Toscana. Mem. Soc. Geol. It. 25, pp. 63-74. Bossio A., Cerri R., Costantini A., Gandin A., Lazzarotto A., Magi M., Mozzanti R., Mazzei R., Sagri M. & Salvatorini G. 1992. I bacini distensivi neogenici e quaternari della Toscana. 76 Adun. Estiva Soc. Geol. It., Guida alle escursioni, Florence. Carmignani L., Decandia F.A., Disperati L., Fantozzi P.L., Lazzarotto A., Liotta D., Kligfield R. & Meccheri M. 2001. Inner northern Apennines, in Martini P. & Vai G.B. (eds), Anatomy of an orogen: the Apennines and adjacent Mediterranean basins. Kluwer Academic Publishers, pp. 197-214. Conedera C. & Ercoli A. 1973. Elementi geomorfologici della piana di Firenze dedotti da fotointerpretazione. L’Universo 53, pp. 255-262. Capecchi F., Guazzone G. & Pranzini G. 1975. Il bacino lacustre di Firenze–Prato–Pistoia. Bollettino Società Geologica Italiana 94, pp. 637-660. Eastman Kodak Company 2003. Aerial Data AS-77, www.kodak.com.

204

Salvini et alii, Remote sensing analysis in the Florentine plain: investigating wetland contexts…

Martini I.P. & Sagri M. 1993. Tectono-sedimentary characteristics of Late Miocene – Quaternary extensional basins of the Northern Apennines, Italy. Earth Science Reviews 34, pp. 197-233. Pizziolo G. 2010. Landscape Archaeology at Sesto Fiorentino: the contribution of aerial photographs to the study of archaeological contexts within an integrated approach, in Niccolucci F. & Hermon S. (eds), Beyond the artifact, digital interpretation of the past. Proceedings of CAA 2004, Prato 13-17 April 2004. Budapest, Archaeolingua, pp. 479-483. Pizziolo G. & Sarti L. 2005. Landscape archaeology in Sesto Fiorentino: a GIS analysis for investigating settlement strategies in wetland area, in Berger J.F., Bertoncello F., Braemer F., Davtian G. & Gazenbeek M. (eds), Temps et espaces de l’homme en société, analyses et modèles spatiaux en archéologi, XXV Rencontres Internationales D’archeologie et D’histoire D’Antibes, Antibes, Editions APDCA, pp. 443-452. Pizziolo G. & Sarti L. 2006. Exploring the archaeological landscape through a local perspective: spaces and places in the prehistory of the Florentine plain, in Campana S., Forte M. (eds), From Space to Place, 2nd International Conference on Remote Sensing in Archaeology, Proceedings of the 2nd International Workshop, CNR, Rome, Italy, December 4-7, 2006 (BAR International Series 1568), Oxford, BAR Publishing, pp. 261-269. Salvini R., Guastaldi E, Coscini N. & Del Seppia N. 2006. Ricostruzione del paleoalveo del Fiume Serchio (Lucca, Italia) tramite rilievi LIDAR, foto aeree ed immagini QuickBird. AIQUA Il Quaternario, Italian Journal of Quaternary Sciences 19(2), pp. 297-308. Sarti L. & Martini F. 1993. Costruire la Memoria. Archeologia preistorica a Sesto Fiorentino (1982-1992), Garlatti e Razzai, Firenze. Sarti L. & Martini F. 1998. Strategie insediative del Campaniforme in Italia centrale, in Atti Congr. Int. Bell Beaker today, Riva del Garda, 1998. Toutin T. & Cheng P. 2002. QuickBird. A Milestone for HighResolution Mapping. Earth Observation Magazine 11, pp. 14-18.

205

25 Soil use from the Late Chalcolithic to the Early Middle Bronze Age in the central Po plain (northern Italy): new data from buried soils Maria Bernabò BreaI, Mauro CremaschiII, Lorenza BronzoniIII, Fabrizio PaviaIII, Cristina RovestaIV I Soprintendenza per i Beni Archeologici dell’Emilia Romagna, Museo Archeologico Nazionale di Parma, Palazzo Pilotta, 43100 Parma, Italy – [email protected] II Dipartimento di Scienze della Terra ‘A.Desio’, via Mangiagalli 34, 20133 Milano,Italy – [email protected] III Ars/Archeosistemi, via Nove Martiri 11, 42100 Reggio Emilia, Italy – [email protected] IV Gea Ricerca e Documentazione Archeologica, borgo XX marzo 4, 43100 Parma, Italy – [email protected] Abstract On the basis of recent discoveries in the region of western Emilia in the course of rescue archaeology, new data on the soil use from the Chalcolithic to the early Middle Bronze age were acquired. The Late Chalcolithic and Early Bronze age communities used the slashand-burn technique, already known in the Neolithic period, to clear the forest. Both at Rubiera and at San Pancrazio, and to a lesser extent also at Cortile San Martino, burned trees were left on the ground and only parts of the cleared surfaces were used for cultivation, and the cleared area did not extend over the entire plain. At around 1700 cal yr BC the soil use changed dramatically, and cultivation was supplemented by water management in a landscape that was already fairly deforested. The countryside appears therefore to have already been significantly modified by human activities (deforestation and irrigated agriculture) that, together with pastoralism, would be further amplified during the Middle and Recent Bronze Age by the land management typical of the Terramare culture.

1. Introduction Rescue archaeological projects promoted by the Soprintendenza per i Beni Archeologici dell’Emilia Romagna, have in the last few years through an increased number of large excavations resulted in marked progress in our knowledge of the prehistory of the central Po plain. Aside from various important discoveries, large-area projects of urban development and major infrastructure have afforded the unique opportunity to study broad surfaces of the Holocene soils buried in the alluvial plain, and therefore to investigate not only the archaeological sites included in these soils, but their surroundings as well, thus providing information on soil use and agricultural practices. In the region of Parma and Reggio Emilia, some sites dating from the Late Chalcolithic to the beginning of the Middle Bronze Age were identified and excavated. These sites appear to be particularly significant because this period, almost unknown up to now in the western Emilia region, is critical in reconstructing the changes in the land use of the region — also in the perspective of the rise of the Terramare culture. The sites we are dealing with (the recently discovered ones together with S. Ilario and Rubiera which have already found their way into the literature) are all located in the Holocene alluvial plain, at the southern fringe of the alluvial fans descending from the Apennines. From east to west they are: Rubiera in the bed of the Secchia River, Buco del Signore close to Reggio Emilia, Sant’Ilario in the bed of the river Enza, San Pancrazio in the outskirts of Parma, and

Cortile San Martino in the plain south of Parma (fig. 1). In this area, as indicated by a traditional but still effective model (see for instance Walker 1984 and Bridge 2005), the alluvial plain developed through the repeated superposition of fine overbank sediments, separated by dark-coloured soils representing pauses in the sedimentation and stability of the topographic surfaces (fig. 2). These soils are some ten centimetres thick and their profiles consist of an A1 horizon, followed occasionally by a shallow Bw horizon; they may be classified as entisols, poorly developed inceptisols, and vertisols. Their low degree of development indicates that the interval between consecutive sedimentation cycles, in which the soil developed, was fairly short (Cremaschi 1987). A program of radiocarbon dating of the humic fraction of the buried soils in the area was undertaken some years ago by the Soil Survey Service of the Regione Emilia Romagna (www.Regione.Emilia-Romagna. it/geologia/). The dates so far obtained, combined with others gathered from the literature (see Cremaschi & Marchetti 1995 for references) (fig. 3), indicate that the buried soils in the investigated area (in the plain at the fringe of the alluvial fans) range from the Late Pleistocene to the historical period. The radiocarbon ages of the buried soils are clustered at the very start of the Holocene, and then again in its middle part with a peak in the Sub-boreal, which is the period of the archaeological evidence discussed in this paper. This peak of radiocarbon dates must be regarded as an index of the instability of the soil forming processes

207

Hidden Landscapes

and enhanced fluvial activity. With some uncertainty due to their low number, the radiocarbon dates may be regarded as a rough indication of a drier climate in the Po plain during the Bronze Age in comparison to earlier and later periods of the Holocene. This is consistent with the general palaeoclimatic scenario for Western Europe (Kristiansen 1998; Orombelli 1997) and with some detailed Alpine records (Magny 2004; Cremaschi et al. 2005).

2. The archaeological buried soils 2.1 The Neolithic buried soils

Geoarcheological research on the Neolithic buried soils of Emilia (Cremaschi 1990) has so far been mainly conducted in on-site locations, such as La Razza di Campegine (Bernabò Brea et al. 2008),

Ponte Ghiara near Fidenza (Bernabò Brea et al. 2000), Collecchio,1 S. Andrea Travo,2 which date back to different moments of the 5th millennium BC. Micromorphological investigation of these sites, when available, reveals discontinuous occupation as phases of dwelling alternate with phases of cultivation (Ottomano in Bernabò Brea et al. 2000). The Buco del Signore site, on the outskirts of Reggio Emilia, represents an off-site situation located at the periphery of a Early Neolithic site: the buried soil identified here was never intensively settled but mainly devoted to cultivation. The micromorphological evidence of agricultural practice consists of charcoal dispersed in the matrix, and baked soil fragments which are evidence of land clearance through fire. No charcoal concentration has been

Figure 1 – Distribution of the sites mentioned in the text in their geomorphological context.

208

Bernabò Brea et alii, Soil use from the Late Chalcolithic to the Early Middle Bronze Age in the central Po plain…

observed on the soil surface, and charcoal integration in the soil matrix, associated with coarse coating in voids, may be interpreted as evidence of soil disruption with a digging stick or hoe, and therefore of cultivation. 2.2. Chalcolithic and early Bronze Age soil use at San Pancrazio

A rather different soil use is documented in the buried soil of San Pancrazio. At this locality to the west of Parma, an excavation for building foundations exposed a stratigraphic sequence consisting of six metres of alluvial loam and sand, in which, below the Ap horizon containing Roman and Iron Age pottery fragments, three buried soils are intercalated at depths of 1, 3, and 5 m from the modern surface. The soil buried at 1 metre depth turned out upon first investigation to be rich in archaeological material, mainly dating to the Early Bronze Age, and has therefore been excavated 3 over a surface of about 7000 sq. m. It is rather homogeneous over the whole area: about eighty centimetres thick, very dark in colour, and loamy-clayey in texture. It has a well developed blocky structure and can be classified as a shallow vertisol covered, through an abrupt upper boundary, by a thin layer of white silty loam. When this layer was removed from the surface of the soil, several root casts were identified, and around them large spots of charcoal were found, which are interpreted as the result of wood clearance through slash-and-burn practices and not severely dispersed by subsequent cultivation. The root casts represent the parts of the trees that were left in place, while the charcoal spots represent the parts which were cut down and burned on the surface of the soil. The soil was then flooded and covered by the white alluvium, which penetrated in the root casts after the wood decomposed (fig. 4). About 30,000 fragments of pottery, dispersed in the top ten centimetres of the soil profile, were collected and mapped by means of a Total Station. In the southern part of the study area (area A), where most of the root casts are located, these pottery fragments are very small (from 1 to 3 cm in diameter) and have rounded edges. They are mostly distributed around the charcoal concentrations. The northern edge of the excavated area B is rather different: the pottery fragments are larger and often occur in dense clusters; they can be fitted together over short distances. Also in area A, while very small and poorly preserved, some of the pottery fragments (out of a sample of about 300 individuals) were refitted. The

Figure 2 – Block diagram representing the geomorphological and sedimentary model for the study area.

Figure 3 – Radiocarbon dates available for the buried soil in the study area.

Figure 4 – San Pancrazio. Root casts and charcoal concentrations at the top of the Early Bronze Age buried soil.

lines joining the refitted fragments run parallel to each other and trend to the north, which is the direction of the local slope. While vertic processes typical of this kind of soil may explain the fragmentation and the state of corrosion of the single fragments, suggesting long permanence inside the soil, the direction of dispersion of the refitted fragments may

209

Hidden Landscapes

Figure 5 – San Pancrazio. Soil profile (A) with graph of sherd concentration by depth (B).

be related to ploughing, there being no evidence of colluviation in the soil profile. The concentration of sherds at a depth of ten centimetres from the top of the soil, too, may, in the absence of stratigraphic discontinuities in the soil profile, be interpreted as an indication of the base of the Ap horizon (fig. 5). The thin sections were not very eloquent about the agrarian use of the soil, as the dominant features (slicken sides, birefringent fabric) are related to the vertic nature of the soil and to its high clay content. Distributed over the whole of the excavated area, three complete biconical vessels were found inserted into small pits at the base of the soil. Four pits were also identified; their content in one case consisted of a large jar broken into fragments (fig. 6), and in the others — besides archaeological material such as two coarse pottery funnels (fig. 7.5), fragments of fine ware cups, a large fragment of a jar decorated with applied cordone and faunal remains — large stones (one fragment of a grinding stone and two large unworked stones) were fitted like stele at the top of the pits. There is no apparent functional explanation for these vessels set into the ground, nor for the large stones inserted as stele in the pits. These features therefore probably are marks of ownership, installed in the course of the Early Bronze Age occupation which involved wood clearance and cultivation. While the high concentration of pottery in

area B may suggest the existence of dwelling areas, no associated habitation structures were found: only a few widely dispersed post-holes were found in the buried soil. The age and the length of the occupation of the San Pancrazio soil rely on three radiocarbon dates, and on the archaeological context. The radiocarbon dates obtained from the charcoal spots at the top of the soil are 3860±60 14C yr BP (GX 32453) and 3890±50 14C yr BP (GX 32454). An additional date was obtained by the Soil Survey Service of the Regione Emilia Romagna, on the humic fraction of the lower soil: 4205±100 14C yr BP 4. From the archaeological point of view, apart from weak evidence for the Chalcolithic tradition, represented by a lunate flint (fig. 7.1), two tanged arrow-heads in jasper and flint (fig. 7.2), and several stone pendants (fig. 7.4), the bulk of the archaeological features dates to the Early Bronze Age, from a starting phase (ca. 2200-2000 cal BC) represented by a oar-shaped pin 5 (fig. 7.3) to the beginning of the Middle Bronze Age (about 1750-1650 cal BC), which is represented by an axe-shaped handle. Apparently the distribution of the archaeological finds suitable for dating does not allow to distinguish, either stratigraphically or topographically, the different phases of occupation. The oar-shaped pin and the axe- shaped handle, for example, were both found in the same stratigraphic position at the top of

210

Bernabò Brea et alii, Soil use from the Late Chalcolithic to the Early Middle Bronze Age in the central Po plain…

the buried soil in area B. The phases and typology of the occupation will be better understood only when the rich but poorly preserved archaeological material will have been thoroughly studied.6 For the moment, the occurrence of the following pottery types may be mentioned: elbow-shaped handles beside a unique axe-handle, deep cups with sinuous profiles, several types of carinated cups with short and concave walls and sometimes with bobbin lugs, jars decorated with applied cordons, frequently with impressed marks, an unperforated tongue lug, and clay sieves. These materials may be compared with the archaeological context, also unfortunately not homogeneous, from the sites of the Panaro river at Spilamberto (Ferrari et al. 1997, fig. 153). Both radiocarbon and archaeological dates are in agreement in indicating that the soil use lasted several centuries and that it may have been cultivated (or used in some other way), abandoned, and reoccupied by vegetation several times. Charcoal spots and root casts may indicate only the final phase of clearance through fire, before the soil was buried by a flood. Wood clearance by slash-and-burn is also recorded in a slightly earlier period (4525±95 14C yr BP), in a site with scale-decorated pottery dating back to the Middle Chalcolithic at Rubiera along the Secchia river. Here, a buried entisol preserves upright sections of oak trunks surrounded by a broad scatter of a considerable amount of charcoal, over a length of 100 m of the exposed section (Cremaschi & Ottomano 1993). In thin section there are convoluted features, pedorelicts, fragments of charcoal dispersed in soil matrix, which cannot be interpreted unambiguously but may be related to artificial disruption of the soil (Cremaschi 1997). However, in the period ranging from the Late Chalcolithic to the Early Bronze Age, the deforestation did not extend to the whole of the plain; on the contrary, some sites were located in stable woodlands. This is for example the case with the two Bell-Beaker settlements at Rubiera (Bermond Montanari et al. 1982) and S. Ilario d’Enza (Barfield et al. 1975), which stood on buried soils preserving stumps of large oaks which radiocarbon dating revealed to be contemporaneous to the archaeological sites.

entisol including Roman bricks and pottery, and one metre of alluvial loam, there is a buried 20cm thick grey entisol. This is clayey in texture, flat and gently dipping to the NW, where evidence of hydromorphy is better expressed. As at San Pancrazio, root casts surrounded by charcoal concentrations were found on its surface, and a very few small, dispersed and corroded fragments of pottery, charcoal, and occasional faunal remains were collected from it. Despite the apparent scarcity of archaeological material, a complex network of archaeological features was discovered at the base of this soil (fig. 8). A palaeochannel trending to the north crosses the whole excavated area. An elbow-shaped handle from its fill

Figure 6 – San Pancrazio, area B. A biconical jar (diam. circa 25 cm) in a shallow pit.

2.3. Changing soil use during the early Middle Bronze Age: Cortile S. Martino

A rather different picture emerges from the site of Cortile S. Martino (Parma). Here, too, an area of 7000 m2 was excavated for a building project.7 Below the present-day Ap horizon, a slightly developed

Figure 7 – San Pancrazio. 1. flint lunate; 2. tanged arrowhead in jasper; 3. oar-shaped pin; 4. stone pendant; 5. coarse pottery funnel.

211

Hidden Landscapes

indicates activity during the Early Bronze Age. The clayey fill of the palaeochannel and its surroundings are cut by a network of ditches. In the largest

Figure 8 – Cortile San Martino. Irrigation/drainage ditches in the Early Middle Bronze Age soil (the base of the photograph is about 20 m wide).

of these, which has the same orientation as the palaeochannel, three minor ditches converge. These are associated with some wells, which reach a water table held by a sand layer at a depth of about three metres, and with shallower pits which may regarded as water reservoirs. Ditches and related wells can be interpreted as a hydrological system intended to extract water from the subsoil, to collect it in the main ditch, and to distribute it in the surrounding fields. Alternatively, the ditches may have been used also for drainage purposes, as temporary waterlogging of the soil is indicated by hydromorphic features. Beyond the hydrological structures, two sets of post-holes distributed along right-angled alignments were found around the more elevated parts of the excavated area. Near one of these areas some shallow pits containing complete ovoid jars were found (fig. 9.1). These structures may be interpreted as the remains of small isolated buildings, possibly storage huts. A radiocarbon date on charcoal provided the age of 3180±50 14C yr BP (LTL 1980A), but appears unreliable when compared to the archaeological context, which suggests a more probable date for the occupation of the soil at the start of the Middle Bronze Age (i.e. about 1750-1650 cal yr BC). The archaeological material collected from the buried soil and in the associated archaeological structures 8 is very scarce and mainly consists of coarse pottery fragments, not very suitable for a reliable chronological attribution. However, a few fragments of fine ware may be attributed to the ‘axe-handle phase’ dating to the BM1,9 and among these is one fragment of particular interest: it is part of a biconical cup decorated by a double incised zig-zag line, with a precise parallel in the site of Braglia/Case del Lago (Tirabassi 1997, fig. 160.9; fig. 9.2). Also a conical button in vitreous paste was found which is not in conflict with a chronological attribution to the BM1.10 While the micromorphology and pollen content of the Cortile San Martino site have not yet been studied, the evidence presently available points to intense soil use and agriculture supported by water management (irrigation and drainage).

3. Conclusion

Figure 9 – Cortile San Martino. 1. ovoid jar; 2. biconical cup decorated with a double incised zig-zag line.

Summing up the evidence, we may conclude that the Late Chalcolithic and Early Bronze Age communities used the slash-and-burn technique to perform wood clearance. Both at Rubiera and at San Pancrazio, and to a lesser extent also at Cortile San Martino, burned trees were left on the ground and only

212

Bernabò Brea et alii, Soil use from the Late Chalcolithic to the Early Middle Bronze Age in the central Po plain…

some of the cleared surfaces were cultivated. At this stage it is not possible to say if this was due to a natural hazard (a flash flood leading to the abandonment of the soil) or to a cultural choice (destination of only a portion of the cleared surface to agricultural practices). In any case, in this period (i.e. at about the end of the 3rd and the start of the 2nd millennium BP), the cleared area did not extend over the entire Po plain, as a number of other sites were located in a stable woodland environment. Around 1700 cal yr BC, there is a near chronological coincidence between the end of the San Pancrazio site and the beginning of the Cortile San Martino site, and at this time the soil use also dramatically changed. At Cortile San Martino soil cultivation is supported by water management in a landscape that, on the basis of the few pollen studies available for this period (see Cremaschi et al. 2004) and on the basis of the dendrochronological evidence from the pile-dwelling site of Chiaravalle della Colomba (Bernabò Brea & Miari 2004), was already fairly deforested. The countryside appears therefore to have already been significantly modified by deforestation and irrigated agriculture, and together with pastoralism this will have been further amplified by the land management typical of the Middle and Recent Bronze Age Terramare culture (Bernabò Brea & Cremaschi 2004; Cremaschi et al. in press).

References Barfield L.H., Cremaschi M. & Castelletti L. 1975. Stanziamento del Vaso Campaniforme a S. Ilario d’Enza (Reggio Emilia), Preistoria Alpina 11, pp. 155-199. Bellintani P. 2000. I bottoni conici ed altri materiali vetrosi delle fasi non avanzate della media età del Bronzo dell’Italia settentrionale e centrale. Padusa XXVI, n.s. pp. 95-110. Bermond Montanari G., Cremaschi M. & Sala B. 1982. Rubiera: insediamento del vaso campaniforme, Preistoria Alpina 18, pp. 79-109. Bernabò Brea M., Battiston C., Mazzieri P. & Ottomano C. 2000. Un gruppo di figurine fittili dal sito di Ponte Ghiara, in Pessina A. & Muscio G. (eds), La neolitizzazione tra Oriente e Occidente, Atti del Convegno di Udine 1999, Udine, pp. 269-287. Bernabò Brea M., Bronzoni L., Cremaschi M., Mazzieri P., Salvadei L., Trombino L., Valsecchi V., Bruni S., Gregari C. & Gulgielmi V. 2008, Lo scavo estensivo nel sito neolitico di Razza di Campegine (RE), in Bernabò Brea M. & Valloni R. (eds), Archeologia ad Alta Velocità. Indagini archeologiche e geologiche lungo il tracciato ferroviario Alta Velocità, Atti del Convegno Parma, 9.6.2003. Quaderni di Archeologia dell’Emilia Romagna 22, All’Insegna del Giglio, Firenze, pp. 41-86. Bernabò Brea M. & Cardarelli A. 1997. Le terramare nel tempo, in Bernabò Brea M., Cardarelli A. & Cremaschi M. (eds),

Le Terramare. La più antica civiltà padana, Catalogo della Mostra, Milano, pp. 295-301. Bernabò Brea M., Cardarelli A. & Cremaschi M. (eds) 1997, Le Terramare. La più antica civiltà padana, Catalogo della Mostra, Milano. Bernabò Brea M., Castagna D. & Occhi S. 1999. L’insediamento del Neolitico superiore a S. Andrea di Travo (PC). Padusa XXXIV, pp. 7-54. Bernabò Brea M. & Cremaschi M. 1997. La Terramara di S. Rosa di Poviglio: Le strutture, In Bernabò Brea M., Cardarelli A. & Cremaschi M. (eds), Le Terramare. La Più Antica Civiltà Padana, Catalogo Della Mostra, Milano, pp. 196-212. Bernabò Brea M. & Cremaschi M. 2004. La terramara di S. Rosa a Poviglio (Reggio Emilia) nel corso del Bronzo Recente, in Cocchi Genick D. (ed.), L’età del Bronzo Recente in Italia, Atti del Congresso, Viareggio, pp. 101-110. Bernabò Brea M. & Miari M. 2004. Preistoria e Protostoria del Piacentino, in Passeggiate Archeologiche Piacentine, Guide Belvedere, ed. Diabasis, pp. 36-67. Bernabò Brea M., Miari M., Bianchi P., Bronzoni L., Ferrari P., Guarisco F., Lari E., Lincetto S., Maggioni S., Occhi S. & Sassi B. 2005. Le prime fasi del sito dell’età del Bronzo di Forno del Gallo a Beneceto (Parma). Padusa XL, pp. 81-115. Bridge S.J. 2005. Rivers and floodplains. Blackwell, New York. Bronzoni L. & Cremaschi M. 1989. La terramara di Case Cocconi (RE). Padusa XXV, pp. 173-226. Carancini G.L. 1975. Gli spilloni nel’Italia continentale, PBF XIII(2), München. Cremaschi M. 1987. Paleosols and vetusols in the Central Po Plain (northern Italy). A study in Quaternary Geology and Soil development. Unicopoli, Milano, p. 316. Cremaschi M. 1990. Pedogenesi medio-olocenica ed uso dei suoli durante il Neolitico in Italia settentrionale, in P. Biagi (ed.), The Neolithisation of the Alpine region, Monografie di Natura Bresciana 13, pp. 71-89. Cremaschi M. (ed.) 1991-1992, Ambiente, insediamento, economia. 1. Italia settentrionale, in L’età del bronzo in Italia nei secoli XVI-XIV a.C., Atti del congresso, Rassegna d’Archeologia 10, pp. 145-188. Cremaschi M. 1997. Terramare e paesaggio padano, in Bernabò Brea M., Cardarelli A., Cremaschi M. (eds), Le Terramare. La più antica civiltà padana, Catalogo della Mostra, Milano, pp. 107-125. Cremaschi M. & Ottomano C. 1993. Valutazione del Rischio Archeologico nelle Cave Corradini presso Rubiera, Reggio Emilia, Rapporto alla Soprintendenza per i Beni Archeologici dell’Emilia Romagna, inedito. Cremaschi M. & Marchetti M. 1995. Changes in fluvial dynamics in the Central Po Plain (Italy) between Late Glacial and Early Holocene, in Frenzel, B. (ed.), Palaeoclimate research / Paläoklimaforschung 14, pp. 173-190. Cremaschi M., Pizzi C. & Valsecchi V. 2006. Water management and land use in the terramare and a possible climatic co-factor in their collapse. The case study of the terramara S. Rosa (Northern Italy). Quaternary International 151, pp. 87-98. Cremaschi M., Bernabò Brea M. & Pizzi C. (eds), in press. The Terramare countryside, B.A.R. International series. David-Elbiali M. 2000. La Suisse occidentale au II millénaire av. J.C. Chronologie, culture, intégration européenne, Lausanne.

213

Hidden Landscapes

Ferrari A., Morico G., Steffè G. & Spilamberto S. 1997. Cesario sul Panaro (MO): tracce di insediamenti del Bronzo antico e dell’inizio del Bronzo medio, in Bernabò Brea M., Cardarelli A. & Cremaschi M. (eds), Le Terramare. La più antica civiltà padana, Catalogo della Mostra, Milano, pp. 303-304. Kristiansen K. 1998. Europe before history. New Studies in Archaeology. Cambridge University Press, Cambridge. Magny M. 2004. Holocene climate variability as reflected by midEuropean lake-level fluctuations and its probable impact on prehistoric human settlements, in Brauer A. & Guilizzoni, P. (eds), The record of Human/Climate Interaction in Lake Sediments, Quaternary International 113, pp. 65-79. Mutti A. 1993. Caratteristiche e problemi del popolamento terramaricolo in Emilia occidentale, Bologna. Orombelli G. 1997. Le condizioni climatiche durante il II millennio a.C., in Bernabò Brea M., Cardarelli A. & Cremaschi M. (eds), Le terramare. La più antica civiltà padana, Milano, pp. 51-58. Tirabassi J. 1997. La Braglia (RE), in Bernabò Brea M., Cardarelli A. & Cremaschi M. (eds), Le Terramare. La più antica civiltà padana, Catalogo della Mostra, Milano, pp. 313-314. Walzer R.G. (ed.) 1984. Facies Models Geosciences Canada, Reprint Series 1. Geoscience. Canada St John’s Newfoundland.

10

Conical buttons in vitreous paste with a horizontal hole occur in several terramare sites dating back to the BM2; the specimen of Cortile S. Martino is the first in the Emilia region which can be dated to the BM1 period. In the Piemonte region conical buttons have been found in the Mercurago site, dating back to BM1 (Bellintani 2000).

Notes 1

Unpublished geoarchaeological report by Diego Angelucci, 1992. 2 Research in progress by Luca Trombino. The S. Andrea sites dates back both to a late Neolithic phase and to the full Squared Mouthed Pottery phase, which is exclusively represented in all the other sites recorded in this paper (Bernabò Brea et al. 1999). 3 Unpublished archaeological excavations, performed 20062007 by the Gea company of Parma under the scientific direction of the Soprintendenza Beni Archeologici Emilia Romagna (M. Bernabò Brea). Field director dr. Cristina Rovesta. 4 Laboratory ID not available. 5 See David-Elbiali 2000, pp. 140-143; Carancini 1975, p. 95 ff (pin “con capocchia foliata”). 6 The archaeological material, after only a preliminary check in the field, is being studied by Cristina Rovesta. 7 Unpublished archaeological excavations, performed 2007 by the AR/S Archeosistemi company of Reggio Emilia under the scientific direction of the Soprintendenza Beni Archeologici Emilia Romagna (M. Bernabò Brea). Field directors dr. Lorenza Bronzoni and dr. Fabrizio Pavia. 8 The archaeological material, after only a preliminary check in the field, is being studied by dr. Lorenza Bronzoni and dr. Chiara Pizzi. 9 The cultural phase indicated in the Emilia region as BM1 represents the early period of the Terramare Culture and a part of it is characterized by the axe-handle pottery (Bernabò Brea & Cardarelli 1997, p. 297). The post-dwelling site of Chiaravalle della Colomba (PC), whose archaeological context is quite similar to that of the Braglia and Cortile S. Martino sites, gave a radiocarbon date of 3430±40 14C yr BP, cal. 2ı 1880/1635 BC; central calibrated age 1714 cal yr BC (ARC 1651; Bernabò Brea & Miari 2004).

214

26 Micromorphology and mineralogy of the SEV core (Santarém, Portugal): evolution from a transitional to a continental sedimentary environment during the Holocene Cristiano NicosiaI, Maria Teresa AzevedoII, Sonia FavarettoIII, Antonella MiolaIII, Paolo MozziIV, Elisabete NunesII, Ismaele SostizzoIII I Geoarchaeological and Soil Micromorphological Consultant, via Cilento 10, 36100 Vicenza, Italy – [email protected] II Centro de Geologia, Universidade de Lisboa, Edificio C6, 3 Piso; Campo Grande, 1749-016 Lisboa, Portugal – [email protected]; [email protected] III Dipartimento di Biologia, Università degli Studi di Padova, via U. Bassi 58/B, 35121 Padova, Italy – [email protected]; [email protected]; [email protected] IV Dipartimento di Geografia, Università degli Studi di Padova, via del Santo 26, 35123 Padova, Italy – [email protected] Abstract The present work focuses on the soil micromorphology of a multi-disciplinary study performed on a 20 m-deep core taken in the alluvial plain of the Tagus river at Entre Valas, a few km upstream of Santarém (Portugal). The micromorphology was integrated with palynological, sedimentological and geochemical data. The SEV (Santarem Entre Valas) core is representative of the evolution of the Tagus River valley during the last 15,000 years, encompassing Late Glacial and Holocene sediments. The lowest part of the series corresponds to a deltaic/estuarine sedimentary system, with shallow freshwater backswamp ponds receiving occasional inputs of seawater during exceptionally high tides or storms. The upper part of the core represents the transition from the lower deltaic/ estuarine sedimentary complex towards an alluvial plain during the middle and upper Holocene. No indicators of sedimentary hiatuses accompanied by pedogenesis were observed in thin-section, suggesting a rather ‘continuous’ sedimentation. At a depth of 3.5 m pedorelicts of eroded and re-deposited former topsoil horizons were observed. These are most likely related to a phase of topsoil erosion that could be the result of deforestation and heath-land expansion, which in the region are known to begin in the Bronze Age.

1. Introduction The object of this study is a 20 m-deep core drilled in the Tagus River alluvial plain (Portugal), in Entre Valas, a few kilometres upstream of Santarém on the northern side of the valley and less than 1 km from the valley margin (fig. 1; SEV is the acronym for the site Santarém-Entre Valas). The core was studied by a multidisciplinary team; sedimentological, palynological, and soil micromorphological analyses were performed (Ramos et al. 2007). Radiocarbon dates were also obtained (see fig. 2). The full sequence can be split into three main complexes. The lowermost unit, up to ca. 15 m depth, mainly consists of fine to medium sandy alluvial deposits relating to the end of the Last Glacial Maximum and the Lateglacial. There then follows a sequence that probably corresponds to the floodplain of a deltaic/estuarine system characterized by Mediterranean climatic conditions (pollen of Quercus ilex, Juniperus type and Arbutus, together with some pollen grains of Vitis vinifera and Olea europea), with shallow freshwater ponds that seasonally dried out (Isoetes). An abundance of pyrite, often oxidized to jarosite in association with gypsum, may indicate an occasionally brackish groundwater table (fig. 3d-f). Even though the vegetation was not that

of a coastal salt marsh, the backswamp ponds in the floodplain may have received seawater. Such events were presumably related to the interplay between the floodings of the Tagus and the tides coming upriver. Sea surges and consequent spill-over of seawater into the plain may have occurred during storms

Figure 1 – Location of the Santarém Entre Valas (SEV) core (from Ramos et al. 2007).

215

Hidden Landscapes

Figure 2 – The SEV core: characteristics and location of thin sections.

with strong western winds from the Atlantic Ocean and/or exceptionally high tides. The upper part of the sequence corresponds to a transition from a deltaic/estuarine situation to an alluvial plain. This suggests a more inland position of the coastline in the middle Holocene, followed by a regression. The maximum transgression was reached at the culmination of the post-glacial eustatic sea-level rise, while the regression was caused by the seaward progradation of the alluvial wedge of the Tagus fluvial system during the upper Holocene sea highstand.

2. The top part of the Holocene sequence (0-8 m depth) Thin-section analysis (samples 1 and 2 in fig. 2) focused on finding traces of sedimentary hiatuses associated with periods of surface stability and soil development within the sedimentary sequence. Abundant redoximorphic and calcareous features were observed macroscopically, and micromorphology was employed in order to reveal if such features indicated the existence of sedimentary hiatuses and associated soil development. The results of thin-section study revealed that redoximorphic features are common, as well as micro-

crystalline calcareous ones, especially in the form of nodules and hypocoatings (fig. 3a). Nevertheless, no specific indicators of sedimentary hiatuses associated with prolonged surface stability have been observed. In none of the thin sections the typical A/C horizon alternation produced by short-lived stability surfaces in floodplains was observed. Moreover, the samples show almost no porosity, no microstructure, no traces of biological activity (e.g. calcite biospheroids, excremental features). Orthic calcite nodules without exclusion of skeletal grains and calcite hypocoatings (fig. 3a) are generally ascribed to the gradual infilling of microvoids under vadose circumstances (Wieder & Yaalon 1974; PiPujol & Buurman 1997). Thus these soil features in themselves are not necessarily proof of prolonged surface stability. Moreover, close observation of such nodules under UV light failed to spot calcareous root pseudomorphs or calcified root cell structures (see Becze-Deák et al. 1997). Their presence could have testified to the action of plant roots linked to a stable surface, having formed in response to supersaturation in dissolved calcium carbonate due to root water uptake. At a depth of 3.5 m some rounded and irregularly-shaped pedorelicts were observed (fig. 3b).

216

Nicosia et al., Micromorphology and Mineralogy of the SEV Core (Santarém, Portugal)…

Pedorelicts are fragments originally belonging to soils which were eroded and re-deposited. They are precious indicators since they testify to the existence of phases of soil erosion, but at the same time retain

information on the soils they originally represented. Pedorelicts can thus be a source of information on ‘hidden landscapes’ that were eroded away due to natural or anthropic processes. The pedorelicts ob-

Figure 3 – Thin sections of the SEV core. For location of section numbers, refer to Figure 2. a. thin-section 1: juxtaposed micritic calcite hypocoating and coarse calcite infilling in a channel. PPL 40×. Frame length approx. 3.6 mm. b. thin-section 2: relict fragment of probable topsoil horizon eroded and re-deposited (‘pedorelict’). Notice abundant charred vegetal tissue residues and comminuted charcoal. PPL 40×. Frame length approx. 3.6 mm. c. thin-section 2: detail of a pedorelict showing two phytoliths (circled). PPL 200×. Frame length approx. 0.6 mm. d. thin-section 3: aggregate of pyrite framboids in an elongated void. PPL 100×. Frame length approx. 1 mm. e. thin-section 5: B = brown phlobaphene-rich organic tissue residues; J = jarosite hypocoatings; G = gypsum crystal infillings. PPL 25×. Frame length approx. 6 mm. f. thin-section 6: G = gypsum infilling in planar void; J = jarosite hypocoating. XPL 25×. Frame length approx. 6 mm.

217

Hidden Landscapes

served in thin-section 2 (figs. 2 and 3b-c) contain abundant charred fragments of vegetal tissue along with numerous comminuted charcoal particles and rare phytoliths (siliceous plant skeletons). Dispersed in the groundmass, abundant fine charcoal and charred fragments of organic matter are also visible. The presence of charcoal, charred vegetal matter, phytoliths and burned soil fragments within the soil is usually linked to practices of woodland clearance by fire (Macphail 1992). These practices are known to increase the soil susceptibility to erosion and to trigger processes of severe soil degradation, consisting mainly in the decrease in the organic matter content of the soil and in the deterioration of its nutrient status (Macphail 1992; Gebhardt 1993). The existence of a phase of topsoil erosion in the Tagus valley, which may be man-induced (but the occurrence of natural fires should not be excluded), may hence relate to practices of deforestation and to resulting heath-land expansion. On the basis of archaeological evidence these are known to begin in the region in the Bronze Age (2nd millennium BC). The terminus post quem for such a phase of deforestation is a radiocarbon date of 2930±40 BP (3213-2959 cal. BP), obtained from a sample taken at a depth of 4.55 m.

3. The lower part of the Holocene sequence (8-15 m depth) The lower part of the sequence (thin-sections 3 to 6, see fig. 2) shows a marked change in the sedimentary environment. It consists of sulphuric materials with jarosite and gypsum (thin-sections 5 and 6) overlain by sulphidic, pyrite-rich sediments (thin-sections 3 and 4). These mineral associations are typical of coastal, deltaic or estuarine systems, where potential acid sulphate soils (undrained sulphidic materials) and actual acid sulphate soils are encountered (Pons et al. 1982, Dent & Pons 1995, Kraus 1998). Thin-section 3 has a clayey texture, with rare diatoms and rectangular phytoliths and abundant organic residues, both as brown tissue residues and as blackened fragments (Stoops 2003). In this sample pyrite is present in large quantities, mainly as single framboids or as aggregates of framboids associated with organic matter (fig. 3d), while jarosite and gypsum are absent. Fe-(hydr)oxides are absent and the groundmass appears greyish (reduced) in PPL. Thin-section 4 presents a fine-grained matrix without diatoms, with thin beds indicating the original sedimentary structure. A wedge-shaped inclusion of material was observed that was similar to the sediments pertaining to the top part of the sequence.

The weak orientation of the b-fabric and parallel alignment of micas along the boundary between the inclusion and the surrounding groundmass may suggest some shrinking and swelling. It is, however, also possible that the inclusion corresponds to a disturbance related to the extraction of the core. Thin-sections 5 and 6 consist of very fine-grained, diatom-rich sediments, deposited in mm-thick beds separated by planar voids in which abundant organic residues are found. The latter are brown in colour due to the formation of phlobaphene, an oxidation product of tannin very resistant to further alteration (Bullock et al. 1985; Poch et al. 2004). Jarosite [KFe3(SO4)2(OH)6] occurs as hypocoatings and infillings of planar voids (fig. 3e-f), both preferential pathways for the oxidation of pyrite. Jarosite pseudomorphs of pyrite framboids have been observed in close association with brown organic residues. Pyrite is in fact stable only under severely reducing conditions, and its oxidation generates sulphuric acid - with a steep drop in pH - and mobile Fe(II), which can be removed from the system. The end product of this process is the formation of jarosite, visible as yellowish mottles around pores and ped faces (Pons et al. 1982; Dent & Pons 1995). When a source of Ca, such as detrital calcareous materials or shell fragments, is present (Poch et al. 2004), or in very dry conditions, the oxidation of pyrite is accompanied by gypsum precipitation. Gypsum indeed occurs in thin-sections 5 and 6, as infillings of elongated prismatic crystals or as ‘powdery accumulations’ in close association with brown organic tissues and jarosite (fig. 3e-f). No trace of calcareous material is present in these samples, suggesting that it may have completely disappeared from the deposit. The fact that sulphidic, non-oxidized materials occur above sulphuric, oxidized sediments required additional explanation. Dent & Pons (1995) report the occurrence of Pleistocene acid sulphate soils, formed during the worldwide sea-level fall, on late glacial terraces in Surinam that were subsequently covered by Holocene sulphidic clay and peat. Most authors suggest that in tidal settings the cumulative nature of these soils implies that through time the build-up of sediments causes drainage to improve, the upper horizons being progressively brought above mean high-tide level or neap- or springtides (Dent 1986; Kraus 1998). Oxidation of sulphidic material is also reported to be due to tectonic uplift (Dent 1978) or to changes in the tributary channels of deltas (Dent & Pons 1995). The result is that

218

Nicosia et al., Micromorphology and Mineralogy of the SEV Core (Santarém, Portugal)…

normally oxidized (sulphuric) materials occur in the top part of the sequence, above waterlogged nonoxidized (sulphidic) materials. A possible explanation for this specific case is that the oxidation of the sulphides occurred upon exposure after sampling. This hypothesis is corroborated by a comparison between pictures taken immediately after the core was extracted and others taken at the moment of sub-sampling, and also by the ‘fresh’, unweathered appearance of gypsum crystals.

4. Conclusions Soil micromorphology in combination with palynological, sedimentological and geochemical data proved to be a useful tool for the reconstruction of past ‘hidden’ landscapes. As already mentioned the SEV core represented the transition between two distinct environments. At the base is a deltaic/estuarine sedimentary system, occasionally receiving saline waters, with very fine-grained and organic-rich sediments characterized by mineral associations of pyrite, jarosite and gypsum. The formation of the jarosite/gypsum mineral association is probably due to oxidation of sulphidic materials caused by postsampling aeration. The upper part of the core represents the transition towards an alluvial plain during the middle and upper Holocene. Notwithstanding the presence, observed macroscopically, of pedofeatures such as carbonate nodules and redoximorphic accumulations/depletions, no indicators of prolonged surface stability with pedogenesis were observed in thin-section. Moreover the samples in this interval present almost no porosity and no features (such as faunal excrements) that could derive from biological activity, and no structural aggregation. The intervention of man as an environmental modifier has been encountered at a depth of 3.5 m where fragments of former topsoil horizons were found that had been eroded and re-deposited within the alluvial sediments. These ‘pedo-relicts’ contained numerous inclusions (charcoal, charred vegetal fragments, phytoliths) that can be ascribed to practices of land management and clearance involving the use of fire (but the occurrence of natural fires cannot be excluded). Such practices were, and in many parts of the world still are, responsible for the increased erosion and severe degradation of soils.

References

and potential for paleoenvironmental reconstruction. Geoderma 76, pp. 221-252. Bullock P., Fedoroff N., Jongerius A., Stoops G. & Tursina T. 1985. Handbook for soil thin section description. Wolverhampton: Waine research. Dent D.L. 1978. Saline gleys and acid sulphate soils, in Rijkse W.C. (ed.), Soil groups of New Zealand: gley soils. N.Z. Soc. Soil Sci. Wellington, N.Z. Govt. Printer, pp. 44-66. Dent D.L. 1986. Acid Sulphate Soils: a baseline for research and development. ILRI Publication 39. Dent D.L. & Pons L.J. 1995. A world perspective on acid sulphate soils. Geoderma 67, pp. 263-276. Kraus M.J. 1998. Development of potential acid sulfate paleosols in Paleocene floodplains, Bighorn Basin, Wyoming, USA. Palaeogeography, Palaeoclimatology, Palaeoecology 144, pp. 203-224. Gebhardt A. 1993. Micromorphological evidence of soil deterioration since the mid-Holocene at archaeological sites in Brittany, France. The Holocene 3, pp. 335-339. Macphail R.I. 1992. Soil micromorphological evidence of ancient soil erosion, in Bell M. & Boardman J. (eds), Past and present soil erosion. Oxford: Oxbow, pp. 197-216. PiPujol M.D. & Buurman P. 1997. Dynamics of iron and calcium carbonate redistribution and palaeohydrology in middle Eocene alluvial paleosols of the southeast Ebro Basin margin (Catalonia, northeast Spain). Palaeogeography, Palaeoclimatology, Palaeoecology 134, pp. 87-107. Poch R.M., Fitzpatrick R.W., Thomas B.P., Merry R., Self P.G. & Raven M.D. 2004. Contemporary and relict processes in a coastal acid sulphate soil sequence: microscopic features, in SuperSoil2004: 3rd Australian New Zealand Soil Conference, 5-9 December 2004, Sidney, Australia. Pons L.J., Van Breemen N. & Driessen P.M. 1982. Physiography of coastal sediments and development of potential soil acidity, in Kittrick J.A., Fanning D.S. & Hossner L.R. (eds), Acid Sulfate Weathering. Madison: Soil Science Society of America, Inc., pp. 1-18. Stoops G. 2003. Guidelines for analysis and description of soil and regolith thin sections. Madison: Soil Science Society of America, Inc. Ramos C., Ramos Pereira A., Azevêdo T.M.M., Nunes E., Freitas C.M., Andrade C., Mozzi P. & Favaretto S. 2007. Middle Tagus paleoenvironments since the Last Glacial. Publicações da Associação Portuguesa de Geomorfólogos, Vol. V, APGeom, Lisboa 2007, pp. 191-199. Wieder M. & Yaalon M.H. 1974. Effect of matrix composition on carbonate nodule crystallization. Geoderma 11(2), pp. 95-121.

Acknowledgements The investigation was carried out within the project ‘Geology and Archaeology of the Tagus River Floodplain over Time’ – Geotarif – POCTI/CTA/39427/2001, financed by the Fundação para a Ciência e Tecnologia (FCT) and co-financed by the European Communitary fund FEDER.

Becze-Deák J., Langohr R. & Verrecchia E.P. 1997. Small scale secondary CaCO3 accumulations in selected sections of the European loess belt. Morphological forms

219

27 Deposits and prehistoric occupation of Ribeira Ponte da Pedra (Central Portugal) Pierluigi Rosina Instituto Politécnico de Tomar, Dep. Território, Arqueologia e Património, Estrada da Serra, 2300-313 Tomar & Centro de Geo – Ciência da Universidade de Coimbra, Portugal – [email protected] Abstract Ribeira Ponte da Pedra is a tributary stream of the Tagus river, located in central Portugal. The homonymous archaeological site is being excavated since 1999. Aiming a geo-archaeology research, this article includes a cultural investigation according to chronostratigraphical characterization. The site sediments belong to two Quaternary fluvial terraces of the Tagus (Q3, middle terrace and Q4a, low terrace) and to colluvial deposits dating to 300,000, 90,000 and 25,000 years BP. Dating allow us to correlate these deposits with the isotopic stages 9, 5, and 2. Archaeological remains (mainly quartzite pebble lithic industries) have been attributed to the Lower Middle Palaeolithic, and an Upper Palaeolithic hearth was also discovered.

1. Introduction Around the world, many human artefacts are found in quaternary fluvial deposits. Fluvial systems are dynamic and diversified, and the same goes for their deposits. The stratigraphical and geo-archaeological understanding of these deposits is complicated by lateral and vertical variations. Fluvial sedimentary environments are characterized by different depositional morphologies: channel, bar, and floodplain. These morphologies, coexisting at the same time, are correlated to different lithologies: from the coarsest channel fills (pebbles, gravels and sands) to the finest floodplain sediments (silts). An additional variability is added according to water energy (that essentially depends on the slope), resulting in two different systems: braided and meandering. Lithology, sedimentary structure and deposit geometry depend on the combination between depositional morphologies and energy. During a fluvial sedimentary cycle, the water energy can change, with repercussions on deposit formation (most of all on lithology, but also on sedimentary structures). Sedimentary cycles are further influenced by climatic oscillations, and in the case of the fluvial deposits this may result in the formation of fluvial terraces. Near the sea, the formation of fluvial terraces is strictly linked to sea level eustatic oscillations; consequently each sedimentary cycle should be related to one such oscillation. But in reality it is difficult to relate a terrace formation with each quaternary eustatic/climatic oscillation. The result of a single sedimentary cycle is a deposit formed by different lithologies, sedimentary structures, and geometries, which change both horizontally and vertically. Archaeological remains found in fluvial deposits therefore frequently require a detailed geological investigation in order to

define their position in the fluvial system: among the channel fills, upon a bar, in the middle of the floodplain, etc. The interpretation of fluvial archaeological sites strongly relies on a geo-archaeological analysis. Furthermore, there is the problem of chronology and chrono-stratigraphy of fluvial terraces. The common lack of elements for absolute or relative dating, typical of this kind of sedimentary environment, has often required the use of a morpho-stratigraphical approach for the chronostratigraphical, and chronological, attribution. The results of morphostratigraphy studies could be influenced by different factors like substrate or local neotectonics (cf. Cunha et al. 2005). In central Portugal — in the Tagus river basin — the oldest human occupations are predominantely associated with fluvial terrace deposits. At present, at least six fluvial terraces have been recognised in this region (cf. Rosina 2002, 2004; Mozzi et al. 1999; Corral Fernandes 1998a). Following the custom of existing geological maps, these are named (from oldest to youngest) Q1, Q2a, Q2b, Q3, Q4a, and Q4b. In the last three terraces (mostly in the more recent ones), several archaeological sites have been identified with abundant lithic industries and very rare fossil remains. In the last 15 years research has aimed at a detailed and updated characterization and cartography of the Tagus Quaternary fluvial terraces and deposits. Besides the geo-chrono-cultural characterization of several open-air archaeological sites found in these deposits, understanding patterns of Palaeolithic settlement, behaviour and exploitation in the river plains has also been a goal of this research. The Ribeira Ponte da Pedra archaeological site excavation was carried out in this research framework, being the third intervention in fluvial terraces made by our team.

221

Hidden Landscapes

and 25,000 years BP. Such dates allow us to correlate these deposits with the isotopic stages 9, 5 and 2. Although the majority of archaeological remains (quartzite lithic industries) have been attributed to the Lower Middle Palaeolithic, an Upper Palaeolithic hearth was also identified. However, the focus of the present discussion will be on the sedimentary sequence and chrono-stratigraphy of the fluvial terraces at this site.

2. Geographical and geomorphological characterisation Figure 1 – Location of study area (higlighted).

Figure 2 – Hydrographical network around Ribeira da Ponte da Pedra.

Ribeira Ponte da Pedra is a tributary of the Tagus river, located in central Portugal, and the homonymous archaeological site has been excavated since 1999. As mentioned before, the aim of the intervention was (and still is) a geo-archaeological study along with the chrono-stratigraphical characterization of the cultural sequence. The site sediments belong to two Tagus Quaternary fluvial terraces (Q3, middle and Q4a, low) and to colluvium deposits dating (Dias et al. 2009) respectively 300,000, 90,000

Ribeira (= ‘stream’) da Ponte da Pedra, also called Ribeira da Atalaia, is a tributary of the Tagus in the region of Ribatejo (= ‘Tagus bank’). The homonymous archaeological site is situated a few kilometres north from the confluence with the Tagus, and administratively belongs to the municipality of Vila Nova da Barquinha. In this region we find three principal geological units: the Cenozoic Tagus sedimentary basin (in the geological map named ‘Tagus and Sado Tertiary Basin’), that stretches mostly along the lower Tagus river valley; to the west and northwest this basin is limited by the essentially Mesozoic Estremenho limestone massif; and in the east and northeast by the PreCambrian and Palaeozoic schist-metamorphic complex (Ancient Massif). Lithologically, the Estremenho Massif consists of limestones and marls (with weak presence of flint). Schist, greywacke, quartzite and granite are the more abundant rocks of the Ancient Massif, and clay, silts, sands, and pebbles make up the detritic drainage basin. The recent alluvial sediments, the wide Pleistocene fluvial terraces, the karstic cave fills (in the limestone massif), and the detritic cover form the Quaternary deposits of the region. Such lithological diversity affects the erosive action of the streams, producing residual elevations. During the Plio-Pleistocene, the hydrographical net deeply embanked itself for 100-150 m with respect to the original surface (now at more than 200 m asl). This net developed according to regional tectonics, which is why the larger Tagus tributaries come from the north. The Ponte da Pedra stream also flows from north to south, its valley being in its entirety excavated in the Tagus sedimentary basin, and its sides therefore formed by Cenozoic (Miocene to modern) detritic sediments. Until the Middle Pleistocene, the valley was longer than it is now, and continued in the Nabão Valley (cf. Mozzi 1998).

222

Rosina, Deposits and prehistoric occupation of Ribeira Ponte da Pedra…

The landscape around the Ribeira Ponte da Pedra site is characterised by the fluvial terraces that cover the sides of the nearby low (less than 140 m) hills, but the large and flat Tagus valley is visible too. Currently, we are able to describe and map six fluvial terrace levels (Rosina 2002 and 2004, op. cit.) in the Ribeira da Ponte da Pedra valley, but there may be more. These levels were named, following the existing geological maps, Q1, Q2a, Q2b (high terraces), Q3 (middle terrace), and Q4a and Q4b (low terraces). The chronology of the Tagus terraces is not yet fully known, however there are some absolute dates for the more recent ones (cfr. Prudencio et al. in press, Cunha et al. in press; Raposo & Santoja 1995; Raposo & Cardoso 1998). At the Ribeira da Ponte da Pedra site, the dating results confirm that the Q3 terrace is formed in the Middle Pleistocene, and the Q4a (and Q4b) in the Upper Pleistocene. Therefore, when correlating with oxygen isotopic stages, Q4b should correspond to i.s. 3, Q4a should be ascribed to i.s. 5, while Q3 bottom should agree with i.s. 9.

As the area, named Fonte da Moita, was going to be destroyed by urban expansion, a rescue excavation was conducted in May and June of 1998 (Grimaldi et al. 1999 and 2000; Lemorini et al. 2001; Jaime 2002). The second site (Santa Cita) was discovered in 1990 by Ferring and was excavated, mostly for its Epipalaeolithic levels, since 1991 (Bicho & Ferring 2001); the final intervention at this site, concentrating on the Q4b terrace Mousterian levels, occurred between October 1999 and March 2000 (Lussu et al. 2001), shortly before its complete destruction. The middle terrace scarp at Ribeira Ponte da Pedra was excavated in a single long trench, in an attempt to observe the relationship between the lower and middle terraces, between the terraces and the substratum, and finally between these deposits and the colluvium. Archaeological remains found in both terraces (only lithic industries, mostly made of quartzite) have been attributed to the Lower Middle Palaeolithic; an Upper Palaeolithic hearth was discovered in colluvial sediments.

3. Archaeological intervention

4. Stratigraphy

A very high quantity of lithic implements, mostly made of quartzite pebbles, were found during surface survey in this region. Over the last 15 years, research in the TEMPOAR — Territory, Mobility and Population — framework aimed to establish the chronological and techno-typological characterization of these macrolithic industries, as well as to obtain a greater understanding of the archaeological sites in relation with the Quaternary deposits of the Middle Tagus and its tributaries (Oosterbeek et al. 2002). To achieve these goals, this research project set out to study surface lithic collections, to excavate both Palaeolithic and Holocene sites, and to generate some thematic maps (elaborated with Geographical Information System) in order to provide a more rigorous articulation of archaeological sites and their respective geological deposits. It is in this framework that Ribeira Ponte da Pedra (Ribeira da Atalaia) archaeological site has been excavated since 1999. It is the third excavation on Pleistocene fluvial deposits carried out by our team in this region, and the only one still ongoing since the others were salvage interventions. Our first excavation in a fluvial terrace dates back to 1998 and was carried out on the archaeological site of Fonte da Moita. In the same year, lithic implements (attributed to the Lower Palaeolithic) were found eroding out of a natural section in the Q3 (middle) terrace near Vila Nova da Barquinha.

The morpho-stratigraphical description of the fluvial terraces in this area (Rosina et al. 2005; Corral Fernandes 1998b) places the Q3 deposits between approximately 40 and 54 m asl, the top of the Q4a sediments at 35 m, and that of the Q4b terrace at 25 m (there are no sections showing the bottom of the lower terraces). The Q3 and Q4a terraces are separated by at least 5 m of substrate and colluvia, whereas the Q4a and Q4b terraces are in contact. Our trench, with more than 8m of elevation difference (from 33-34 m to 41-42 m asl), cuts various stratigraphical and geological units. At the moment we are able to distinguish four geological units (from the oldest to the youngest):

Figure 3 – Overview of the excavation.

223

Hidden Landscapes

1. 2. 3. 4.

Miocene substrate bottom of Q3 fluvial terrace top of Q4a fluvial terrace a cover of colluvia Firstly, two strata represent the Miocene: one represented by a reddish clay and the other by a very well sorted white sand that covers it.

Figure 4 – Miocene strata and overlying colluvia.

Secondly, the Q3 terrace bottom is formed by at least four different depositional morphologies: • a bar (formed by reddish coarse sand and pebbles); • a channel (filled with large pebbles, until 35 cm, and reddish coarse sand) that cuts the bar; • flood plain silt deposits (grey to yellow); • transverse channels (filled with reddish sand and pebbles, until 10 cm) that have a very erosive contact with the silts. Thirdly, the top Q4a terrace sands are rather difficult to describe due to the contact with colluvia, but are represented by well-sorted medium sands. Finally, the colluvia are divided in three parts (including the disturbed surface level). They consist of poorly sorted coarse sands and pebbles, orange and reddish (oldest level) to brown. The dating results (Prudencio et al., in press) indicated an age of 300,000 years BP for the Q3 silts, 90,000 years BP for the Q4a top sediments, and 25,000 years BP for the bottom of the colluvia, in which the hearth was discovered. The Q3 bottom formation could therefore be associated with isotopic stage 9, the Q4a top matches up with isotopic stage 5, and the oldest colluvia could be related with isotopic stage 2.

5. Discussion

Figure 5 – Q3 channel and flood plain deposits (detail).

Figure 6 – Q4 Miocene deposits.

There are now two main issues to consider: one regards the position and distribution of archaeological materials in the fluvial sediments, the other the chrono-stratigraphy of the terraces and their prehistoric occupation. Artefacts recovered from fluvial deposits have different meanings: in fact, though the archaeological material found in channel sediments should be considered transported, the objects collected in floodplain sediments or in the top of a bar could be considered in situ. It is therefore necessary to define the correct position of archaeological remains in order to delineate the Palaeolithic settlement and exploitation patterns of these river plains. To better understand the changes occurring between the late Lower Palaeolithic and the beginning of the Middle Palaeolithic, it is very important to agree on the chronology of the terraces. According to the absolute datings, the chronology of the terraces shows an apparent gap: no deposits corresponding to isotopic stage 7 have been identified yet. There are three potential explanations: 1. The Q3 middle terrace was actually formed during two cycles, and for that reason it is thicker 224

Rosina, Deposits and prehistoric occupation of Ribeira Ponte da Pedra…

than the others. If this hypothesis were true, then the top of the terrace should have an age of about

200,000 years BP; it would also comply with other observations of lithic industries in the Tagus

Figure 7 – Stratigraphy of the Ribeira da Ponte da Pedra excavation.

Cultural stage EpiPalaeolithic

Upper Palaeolithic

Middle Palaeolithic

Lower Palaeolithic

Tagus Archaeological terrace sites

Layer description Years BP

–

Error

Aeolian

–

R. Ponte da Pedra

Hearth

24,878

± 2096

PA1 Colluvial (ITN-Lum-35) burned sand

B – OSL

–

R. Ponte da Pedra

Hearth

24,897

± 2194

PA5 Burned (ITN-Lum-35) quartzite

TL

–

R. Ponte da Pedra

Hearth

24,810

± 2184

PA6 Burned (ITN-Lum-35) quartzite

TL

–

R. Ponte da Pedra

Colluvia

25,374

± 1173

PA3 Colluvial (ITN-Lum-35) Sands

B – OSL

Q4b

Santa Cita

Top terrace

Q4a

R. Ponte da Pedra

Top terrace

90,204

± 17,776

–

–

Dating technique

Santa Cita

–

–

Sample description

–

–

–

Sample code

–

PA4 Fluvial sands (ITN-Lum-38)

–

– B – OSL

Q4a

–

?

–

–

–

–

–

Q4a

–

Bottom terrace

–

–

–

–

–

Q3

–

Top terrace

–

–

–

–

–

–

–

–

Q3

Fonte da Moita

Bottom terrace

–

–

Q3

R. Ponte da Pedra

Bottom terrace

302,085

± 12,635

Table 1 – Summary of dating information.

225

PA2 Fluvial silt (ITN-Lum-36)

B – OSL

Hidden Landscapes

basin, such as those of Vale do Forno (Alpiarça), approximately 50 km from Ribeira Ponte da Pedra (cf. Raposo et al. 1985). 2. The Q4a terrace deposits include two cycles. If this is true, the bottom sediments of this terrace must be older than those above them. 3. The sediments of isotopic stage 7 were not conserved. With the available data, we were able to elaborate a new model of the relationships between cultural stages and fluvial deposits for the last 300,000 years (table 1; cf. Rosina 2002 and 2004; Rosina et al. 2005).

References Bicho N. & Ferring C. 2001. O sítio arqueológico de Santa Cita, Tomar: as intervenções arqueológicas de 1990 a 1997, in Cruz A.R. & Oosterbeek L. (eds), Arkeos 11, Tomar, CEIPHAR, pp. 71-88. Corral Fernandez I. 1998a. Depositos Cuaternarios en el Área de Constância-Barquinha-Entroncamento y la Riba del Bezelga, in Cruz A.R., Oosterbeek L. & Pena dos Reis R. (eds), Quaternário e Pré-História do Alto Ribatejo (Portugal), Arkeos 4, Tomar, CEIPHAR, pp. 59-144. Corral Fernandez I. 1998b. Secciones com material arqueologico en estrato en las proximidades de Atalaia, in Cruz A.R., Oosterbeek L. & Pena dos Reis R. (eds), Quaternário e Pré-História do Alto Ribatejo (Portugal), Arkeos 4, pp. 197220. CEIPHAR-Tomar. Cunha P.P., Martins A.A., Daveau S. & Friend P.F. 2005. Tectonic control of the Tejo river fluvial incision during the late Cenozoic, in Ródão – central Portugal (Atlantic Iberian border). Geomorphology 64, pp. 271-298. Cunha P.P., Martins A.A., Huot S., Murray A. & Raposo L., in press. Dating the Tejo River lower terraces (Ródão, Portugal) to assess the role of tectonics and uplift, in Silva P.G., Audemard F.A. & Mather A.E. (eds), Impact of Active Tectonics and Uplift on Fluvial Landscapes and River Valley Development. Geomorphology, Special Issue. Dias M.I., Prudêncio M.I., Franco D., Rosina P., Oosterbeek L., Cura S. & Grimaldi S. 2009. Luminescence dating of a fluvial deposit sequence: Ribeira da Ponte da Pedra – Middle Tagus Valley, Portugal, in Prudêncio M.I. & Dias M.I. (eds), Proceedings of the XVI UISPP Congress – Volume “Archaeometry”, BAR International Series 2045, pp. 103-113. Grimaldi S., Rosina P. & Boton F. 2000. Um sitio ao ar livre do pleistoceno médio no Alto Ribatejo (Portugal): Fonte da Moita, in Paleolítico da Península Ibérica, Acta do 3º Congresso de Arqueologia Peninsular, vol. 2, Porto, ADECAP, pp. 123-136. Grimaldi S., Rosina P. & Boton F. 1999. A behavioral perspective on “archaic” lithic morphologies in Portugal. The case of Fonte da Moita open-air site. Journal of Iberian Archaeology 1, Porto, ADECAP, pp. 33-57. Grimaldi S., Rosina P., Cruz A.R. & Oosterbeek L. 1999. A geoarcheological interpretation of some “Languedocian” lithic collections of the Alto Ribatejo (Central Portugal), in Cruz A.R., Milliken S., Oosterbeek L. & Peretto C. (eds), Human

Population Origins in the Circum-Mediterranean Area: Adaptation of the Hunter-Gatherer groups to environmental Modification, Arkeos 5, Tomar, CEIPHAR, pp. 231-243. Jaime A. 2002. La industrie lithique de Fonte da Moita, in Cruz A.R. & Oosterbeek L. (eds), Arkeos 13. Tomar, CEIPHAR. Lemorini C., Grimaldi S. & Rosina P. 2001. Observações funcionais e tecnológicas num sítio ao ar livre no Portugal Central: Fonte da Moita, in Cruz A.R. & Oosterbeek L. (eds), Arkeos 11, Tomar, CEIPHAR, pp. 117-140. Lussu T., Rosina P., Oosterbeek L. & Costa F. 2001. O Musteriense de Santa Cita (Tomar, Alto Ribatejo, Portugal): investigação e conservação, in Cruz A.R. & Oosterbeek L. (eds), Arkeos 11, Tomar, CEIPHAR, pp. 13-70. Mozzi P., Raposo L., Cruz A.R., Oosterbeek L. & Pena Dos Reis R. 1999. Morpho-stratigraphy of Quaternary deposits and the archaeological record: the case of the Tejo and Nabão valleys (Ribatejo, Portugal), in Cruz A.R., Milliken S., Oosterbeek L. & Peretto C. (eds), Human Population Origins in the CircumMediterranean Area, Arkeos 5, Tomar, CEIPHAR, pp. 63-84. Mozzi P. 1998. Evoluzione Geomorfologica della bassa valle del fiume Nabão, in Cruz A.R., Oosterbeek L. & Pena dos Reis R. (eds), Quaternário e Pré-História do Alto Ribatejo (Portugal), Arkeos 4, Tomar, CEIPHAR, pp. 37-58. Oosterbeek L., Cruz A.R., Cura S., Rosina P., Grimaldi S. & Gomes J. 2004. Ribeira da Ponte da Pedra – Relatório da campanha de escavação de 2003, Techne 9, Tomar, Arqueojovem, pp. 21-54. Oosterbeek L., Cruz A.R., Rosina P., Figuereido A., Grimaldi S. 2002. TEMPOAR – Territórios, Mobilidade e Povoamento do Alto Ribatejo (Portugal) – 1998-2001 (síntese global dos trabalhos realizados), in Cruz A.R. & Oosterbeek L. (eds), Arkeos 12, Tomar, CEIPHAR, pp. 261-322. Raposo L. 1993. O Paleolítico Médio, in O Quaternário em Portugal – balanço e Perspectivas, APEQ e Colibri, Lisboa, pp. 147-161. Raposo L., Carreira J.R. & Salvador M. 1985, A estaçao acheulense final de Milharos, Vale do Forno, Alpiarça. Actas da I reuniao do Quaternario Ibèrico 2, pp. 41-60. Lisboa. Raposo L. & Cardoso J.L. 1998. O sítio paleolítico da Conceição. CEMA – Alcochete. Raposo L. & Santonja M. 1995, The earliest occupation of Europe: the Iberian peninsula, in Roebroeks & Kolfschoten (eds), The earliest occupation of Europe, Leiden. Rosina P. 2004. I depositi Quaternari della media valle del Tago e le industrie litiche associata. Unpublished PhD thesis, University of Ferrara. Rosina P., Oosterbeek L., Jaime A. & Cura S. 2005. Archaelogical sites associated with Tagus middle valley deposits (Alto Ribatejo – Portugal, in Santoja, Pérez-Gonzáles, Machado (eds), Geoarqueologia y Patrimonio en la Península Ibérica y el Entorno Mediterranéo, ADEMA. Soria, pp. 273-282. Rosina P. 2002. Stratigraphie et Géomorphologie des terrasses fluviatiles de la Moyenne Vallée du Tage (Haut Ribatejo – Portugal), in Cruz A.R. & Oosterbeek L. (eds), Arkeos 13, Tomar, CEIPHAR, pp. 11-53. Serviços Geologicos De Portugal 1977. Carta Geologica de Portugal, esc. 1:50.000 foglio 27-D Abrantes e nota esplicativa. Direcçao-Geral de Minas e Serviços Geológicos, Lisboa.

226

Rosina, Deposits and prehistoric occupation of Ribeira Ponte da Pedra…

Acknowledgements I would like to name some colleagues and friends who collaborated in the excavation and in the project — Sara Cura, Stefano Grimaldi, Luiz Oosterbeek and José Gomes — for explicit mention and particular thanks. I am also grateful to Susana Silvério for revising this text.

227

28 The rediscovery of a 5000-year-old landscape: a copper-age village on the shores of ancient lake Maccarese (Fiumicino, Rome) Alessandra ManfrediniI, Cecilia Conati BarbaroI, Giovanni CarboniI, Carlo GiraudiII I Dipartimento di Scienze dell’Antichità, ‘Sapienza’ Università di Roma, Italy – [email protected]; [email protected]; [email protected] II ENEA, Italy The village of Le Cerquete Fianello is located in the Bonifica di Maccarese at 2-3 m above sea level. The site was discovered in the late 1980s after new land reclamations, when many Neolithic, Copper and Bronze Age sites were recognised during field surveys. The site of Le Cerquete was selected for further excavation campaigns (fig. 1). Extensive excavations were conducted from 1992 to 2002, exploring a to-

tal area of 1000 sq.m, which corresponds to roughly 1/10th of the estimated area of the village. At the same time, palaeoenvironmental and geomorphological research began in an attempt to reconstruct the landscape evolution of the area, which was characterized by a coastal lake, as well as of the human settlement dated between the 4th and the 2nd millennium BC. Deep cores were taken in order to reconstruct the shoreline of the ancient lake and to collect geological as well as palynological data (fig. 2). Geological research allowed us to identify a series of sediments and erosion surfaces which document the palaeoclimatic and palaeoenvironmental evolution of the area during the last 5000 years. Facies variation of sediments due to changes in water level and to pedogenetic phases were recognised: higher water levels are associated with humid and cool phases while lower levels correspond to arid phases. During the Copper Age, the settlement of Le Cerquete was located on a shallow sandy peninsula, almost at the level of the lake (fig. 3). Five huts were discovered, four of them oval (8 × 5 m), one circular,

Figure 1 – Site distribution in the ancient lake Maccarese area. d, n: Final Neolithic; e, g, h, i, j, k, l: Copper Age; f, m: Middle Bronze Age.

Figure 2 – The ancient Maccarese lakeshore and the study area.

Figure 3 – Aerial view of the Copper Age village of Le Cerquete – Fianello.

229

Hidden Landscapes

se). Risultati preliminari, Archeologia Laziale XII(2), pp. 349-360. Manfredini A., Carboni G. & Conati Barbaro C. 2000. Nuovi spunti nello studio dell’Eneolitico dell’Italia centrale: una comunità di villaggio a nord del Tevere, in Silvestrini M. (ed.), Recenti acquisizioni, problemi e prospettive della ricerca sull’Eneolitico dell’Italia Centrale, Atti dell’Incontro di Studio, Ancona, pp. 203-214. Manfredini A. (ed.) 2002. Le dune, il lago, il mare. Una comunità di villaggio dell’età del Rame a Maccarese, Origines, Firenze.

Figure 4 – Settlement and abandonment phases of the Maccarese area from the Neolithic to the Medieval period.

together with hearths, storage pits and waste pits. At the eastern edge of the village, close to the lakeshore, a pit contained the burial of a horse with two dog puppies. Two human burials were also identified in the southern edge of the village. On the basis of a set of C14 dates the settlement was dated to 3380-2700 cal. BC (2Ƴ). As soon as the water level of the lake began to increase, the village was flooded and abandoned. The area has a long history of palaeoenvironmental change, as can be inferred from the archaeological evidence: during the Final Neolithic, part of the Copper Age, the Middle Bronze Age and the medieval period there were phases of lower water levels and a substantial reduction of the size of the lake; by contrast, the lake was much larger between the Copper Age and the Middle Bronze Age and during the early modern Little Ice Age (fig. 4). The fluctuations of the water level of Lake Maccarese correspond to those of many other lakes in Central Italy. Given the various origins of the lakes, these variations were probably caused by general climatic changes.

References Manfredini A., Carboni G. & Conati Barbaro C. 1995. Scavi nel sito eneolitico di “Le Cerquete-Fianello” (Maccare-

230

29 A hidden perilacustrine settlement: a village and its fields during the Middle Bronze Age Micaela AngleI, Eva SacchiII, Annalisa ZarattiniI I Soprintendenza per i Beni Archeologici del Lazio, via Pompeo Magno 2, 00192 Roma, Italy – [email protected]; [email protected] II Dipartimento di Scienze della Terra, ‘La Sapienza’, Università di Roma, P.le Aldo Moro 5, 00185 Roma, Italy – [email protected] Abstract A synthesis of the data resulting from two years of excavation of a site named ‘Il Villaggio delle Macine’, set along the shores of the Albano crater lake, is presented. The data allow us to describe a perilacustrine settlement ascribed to the period spanning from the 20th to the 16th centuries BC and to reconstruct, quite completely, the economic activity of the inhabitants of that Bronze Age village. Evidence for growing of cereals, kitchen-gardening, animal breeding as well as hunting and fruit gathering has been found. This suggests a permanent settlement, enabled by the continuous exploitation of all the different resources. Sedentarization also led to a significant metallurgic production and to a consistently strong overexploitation of the surrounding woodland. An over intensive land use was probably the main reason for a dramatic deforestation event and for the resulting low water productivity and biological changes within the lake. A stable population of around 100 to 150 inhabitants has been estimated on the basis of the extent of the settlement, of the land available for cultivation and forests, and on the assumption of high land fertility.

1. Introduction A ‘hidden landscape’ literally emerged from the studies carried out on a submerged archaeological site, the so called ‘Villaggio delle Macine’ (Village of the Millstones), which was discovered just below the water-level along the northern shore of the Albano crater lake (Castelgandolfo, Rome, central Italy). Most of the currently available data resulted from the authors’ studies conducted during and after the archaeological excavations. Specialists were asked to undertake the study of various types of recovered materials. The results of these studies represent a much-needed increase of our knowledge and ultimately improve our understanding of the site. However, to obtain a comprehensive picture of the site and of the life of its community, a synthesis of the studies was needed as well as a synergic interpretation of the data. An updated synthesis of the now available information on the site has therefore been prepared, encompassing the data obtained either directly by the authors (Angelini et al. 2006; Angle & Guidi 2007; Angle et al. 2008 with references) or by collaboration with different specialists (Carra et al. 2007). The results of previous studies were also taken into consideration, such as the basic works by Guilizzoni et al. (2002) and Magri & Follieri (1992).

the end of the Albano-Artemisio volcanic activity (fig. 1). The site was excavated for two years by a team directed by the Soprintendenza per i Beni Archeologici del Lazio. Archeological as well as palaeobotanic, archeozoologic, geological and limnological integrated studies allow the recognition of a well-supported outline of the structured human activity that developed at the site. The establishment of a complex human activity such as a sedentary perilacustrine settlement was always tied to the available natural resources; in our case, the nature of soils, mainly composed of organic rich silts originating from the weathering of the outcropping volcanic sediments and the crater morphology, which provides a natural richness in water and a sheltered micro-climate, offer prime conditions for a Middle Bronze Age

2. The ‘Villaggio delle Macine’ settlement The studied site is located along the shores and mostly under the water-level, of the Albano Lake. The lake, about 10 km south of Rome, originated approximately 45 ka ago (Karner et al. 2001), after

Figure 1 – Geographical setting of the Villaggio delle Macine along the shore of the Albano crater lake.

231

Hidden Landscapes

Figure 2 – A portion of the excavation showing some of the piles still in the original position.

and 2003 allowed the uncovering of a quite large settlement characterized by the presence of wooden structures associated with pile dwellings (Angelini et al. 2006), and ascribed to the period from the 20th to the 16th century BC; a chronological attribution that has been confirmed by isotopic dating (unpublished 14 C data from G. Calderoni). Today, the extent of the village is estimated to be more than 1 ha. Within the excavated portion some 5000 pile-heads were recognized, as well as some superimposed sub-horizontal levels each of which corresponds to a different settlement phase. Throughout the lifetime of the village, at least four different dwelling phases occurred, both on piles and, after a phase in which water-levels fell, on the dried-out and reclaimed lakeshore. 3.1 The metallurgic handicraft

Figure 3 – Millstone made out of grey leucithic lava.

Figure 4 – Bronze axe with original wooden handle.

settlement. The abandonment of the village and the subsequent sedimentation of a relatively thin cover of nearly anoxic lacustrine silts due to a high-stand phase of the water-level, ensured a very good preservation both of the artifacts and of ecofacts (fig. 2).

3. Archaeology The site was named the ‘Villaggio delle Macine’ because of the frequent (more than 100) finds of millstones made out of the grey scoriaceous leucithic lavas outcropping in the surrounding area (fig. 3). The archaeological excavations conducted between 2001

Both from the surveyed and excavated levels and from previous findings at the site, evidence of a widespread metallurgic production was recovered, related to Middle Bronze Age 1-2 and represented by daggers, axes (fig. 4) and other tools (Angle & Guidi 2007). The production of metal seems to have been specialized, at least to a certain degree, because the shapes of the objects show meaningful similarities, and there is an apparent ability to control the composition of the alloy. Indeed, the plasma spectrometer (ICP) analyses show quite uniform results. On the basis of the tin content, the main alloying element, two groups of axes are present: one with a tin content around 4% and another with the tin content varying from 7% to 9%. No obvious correspondences emerged between dating and tin content, or between tin content and technological level. On the other hand, because typology and purpose of the production were probably the same, the observed differences could be ascribed either to poor control over the casting process, or to a conscious choice to employ a lower tin content in the alloys in order to economise on this relatively rare metal. 3.2 Ceramic handicraft

Pottery and bronze objects resulting from survey, from scattered finds and – mostly – from the excavation, permit a precise framing of the exploitation of the site within the early Middle Bronze Age (phase 1-2); in fact most of the pottery can be related to the Grotta Nuova facies, though elements referable to the contemporaneous protoapenninic phases are also present.

232

Angle – Sacchi – Zarattini, A hidden perilacustrine settlement: a village and its fields during the Middle Bronze Age

Nevertheless, within the lowest levels of the site, and therefore in a stratigraphically controlled framework, numerous fragments and whole pots were found that could be dated to the end of the Early Bronze Age (fig. 5). These are comparable to materials from Mezzano 1 (Petitti & Mitchell 1993, table VIII/1; Franco 1982, table XV, M1-22), but also to pottery from other southern Latial sites, such as the Grotta del Peschio Tornera (Guidi & Pascucci 1996, fig. 2/16); see also type 20 of the Latial group and type 17 of the groups common to the Siena region and Latium, according to Cocchi Genick’s 1998 typology. Short of cross-checking all the contexts of Latium (to the south of the river Tiber) of Early and early Middle Bronze Age, these types must be considered the result of the formal persistence of some characteristics, mostly in pots with narrow mouths of medium and large dimensions. The Grotta Nuova facies is well attested by the numerous band-handles, bowls with upright neck and rim, carinated cups and carinated bowls with rim flattened on the inside. Indeed, some shapes occur widely at the site, such as the deep carinated bowls with a rim flattened on the inside and a flat segmented perforated handle, that have been found in almost all levels. This last is one of the most common pot types, and it has parallels mainly in groups from northern Italy (Candalla, Farneto/Monte Castellaccio and Belverde, in Cocchi Genick et al. 1995, fig. 191 A/185 and B/53A, fig. 195 E/47B). At the same time, protoapenninic influences are suggested by some cups with a surmounted bandhandle or with a handle on the rim (fig. 6), and confirmed by the presence of a sopraelevazione con apici lobati that is comparable only to specimens from the Canterno lake (Angle & Gianni 1986).

animals deer are absolutely dominant, revealing a high incidence of deer hunting (which in turn reflects the presence of large forested areas near the village). The other small wild species, with the exclusion of foxes, were probably just sporadic preys. Among the domestic animals remains of ovicaprines and swine are widespread while dogs and cows occur less frequently. So, probably goats, sheep and pigs were the animals typically bred for meat production. The swine remains pertain to very young individuals and young adults, and to only one adult. This kind of selection points to a limited attention to the amount of meat to be produced. Although cow remains are particularly scarce, their ages suggest the use of these animals both as meat producer and for working power. Butchering traces are clearly visible on the trunk and leg bones of several mammals and on some ribs. Sometimes the vertebrae are split lengthwise, which indicates that the carcass was split in two early in the butchering process.

Figure 5 – Jar with two handles, datable to the end of the Early Bronze Age.

3.3 The primary resources: hunting and breeding

Archaeozoological analysis of more than 750 bone remains, mainly fragmentary and of difficult attribution (indeed, only 35% of the material was determined to a taxon), by A. Tagliacozzo and B. Pino Uría (Angelini et al. 2006; Angle et al. 2008) revealed the presence of several species, both wild and bred. The importance of hunting in this community is evident from the number and variety of the recognized non-domestic species: Cervus elaphus, Capreolus capreolus, Sus scropha, Vulpes vulpes, Meles meles, Lepus europaeus, Felis silvestris. The presence of Emys orbicularis has also been attested, even if the small sample size does not allow us to assume a regular collecting activity of this kind of prey. Among the remains ascribed to wild

Figure 6 – Carinated bowl with sopraelevata band-handle, close to ProtoApenninic types.

233

Hidden Landscapes

On the basis of these archaeozoological data some inferences about the economy can be made. The zoo-economy appears to be based mainly on breeding, accompanied by deer hunting as fundamental economic resources for the village inhabitants during the Middle Bronze Age. Bones were also used for the production of artifacts (fig. 7). The most interesting element is an arrow-head with a squared cross section, obtained from the apex of a deer antler. At the base of one of the facets an engraved decoration is present. The arrow-head shows similarities with some tools recovered from the terramare of the Emilia region (Provenzano 1997, fig. 298/6-9). 3.4 The primary resources: gathering and cultivation

Both paleobotanic and ethnobotanic analyses clarify some aspects of the economy of this Bronze Age community (Carra et al. 2007). Thanks to the preservation by silty deposits, information could be collected on the types of growing, harvesting and gathering. The abundance and the diversity of the recognized species, both of wild and cultivated

plants, provide a realistic insight into the activities carried out within the territory: • cereal growing was based on varieties with dissimilar productivity (Triticum spelta, T. dicoccum, T. monococcum, T. aestivum/durum, Panicum miliaceum, Setaria italica), probably cultivated together with other types of corn that require more moisture (e.g. Hordeum vulgare). The joint growing of different varieties could be related to a strategy of diversification, adopted to limit the risk of losing the whole harvest; • in harvested cereals, the presence of annual infesting plants, recognized both at the ‘Villaggio delle Macine’ and at the similar site of Canterno (Angle & Gianni 1986) lets hypothesize cultivation methods with weeding repeated throughout the season; • horticulture is represented by plants with spring and summer harvesting moments (Vicia faba, Linum usitatissimum, Papaver somniferum), probably appreciated for food as well as medicine (fig. 8), fibres and psychotropic effects (Nencini 2004); • the gathering of wild (Cornus, Corilus, Rubus, Quercus) and semi-domestic (Ficus, Malus, Vitis) fruits emphasizes widespread knowledge and continuous exploitation of all types of available fruits, both for immediate use and for processing and preserving.

4. Pollen analysis

Figure 7 – Bone arrow-head.

Figure 8 – Papaver somniferum seed head found in the site (magnification × 50).

The study of two cores, sampled in the framework of the European project PALICLAS (Palaeoenvironmental Analysis of Italian Crater Lake and Adriatic Sea Sediments), allowed to recognize a strong decrease of organic-matter content within the sediment. This denotes an oxidizing environment with low water productivity during the 4th millennium BP (Chondrogianni et al. 1996). Under the same project the substitution, among Cladocerans (Arthropoda), of the genus Daphnia with Bosmina (Guilizzoni et al. 2002) has been recognized (fig. 9). This biological event, also recorded in other perilacustrine areas, correlates well with a deforestation event (see also Martinelli in Angelini et al. 2006). In fact the analysis showed an increasing rate of sedimentation and a prevalence of non-arboreal pollen in the examined interval, which were recognized as the results of one of the major deforestation events ascribed by Guilizzoni et al. (2002) to agricultural activity during the Bronze Age. Both the slope erosion, with the resulting increase in the rate of sedimentation, and the recognized bio-

234

Angle – Sacchi – Zarattini, A hidden perilacustrine settlement: a village and its fields during the Middle Bronze Age

logical events are consistent with a dramatic reduction of forest cover, probably resulting in part from an over-exploitation of the woodlands (fig. 10). A similar situation was recognized for the Mezzano lake settlement, for which the analysis of pollen confirms large-scale deforestation, both to obtain combustibles (needed in large amounts in metallurgy) and to clear for farming (Sadori et al. 2004).

lakeshore resources. In fact the evidence for animal breeding, hunting, fruit gathering and cereal cultivation speaks for a community in which fishing was a limited and secondary activity. Thus, the inhabitants were not ‘water-people’ but, more probably, adapt-

5. Data synthesis Today the collected data permit us to reconstruct a no longer hidden landscape and the living conditions of a small population living along a lakeshore during the Bronze Age. All the data suggest that a small number of individuals permanently settled near the lake for some hundreds of years. That permanence was enabled by the continuous exploitation of the fertile landscape and of all the other available resources. The fertile fields close to the village had an extent of nearly 100 ha. According to the studies by Ampolo (1980) of the material conditions of production during Latial protohistory, the data suggest that (on the basis of assumed land fertility and also taking into consideration the need of the land to rest between crops) cereal growing and horticulture on such a large area could have supported a group of 150 inhabitants. Ampolo also stated that such calculations must be regarded as indicative, and that, in addition to the land set aside for growing and left fallow, further areas would have been needed for the grazing and forestry essential to the needs of the community. Similar conclusions are also reached using the method of land potential evaluation (FAO 1976) suggested by Cremaschi (1990) to estimate the soil productivity. We must underline that while Cremaschi limited its evaluation to the production of cereals, in our case we must consider the nourishment obtained from animal breeding, fishing and hunting and from fruit gathering. Anyway, its results evoke the ratio between the area occupied by the wooden structures of the settlement (more than 1 ha), the cultivated lands (nearly 100 ha), and the probable woodland surface on the flanks of the caldera (nearly 100 ha). Such areas could well have sustained, to the minimal sustenance level based only on cereals, the community of more than 150 people calculated above (Angle 2008).

Figure 9 – Summary of fossil Cladocera from the Albano lake core (redrawn and simplified after Guilizzoni et al. 2002).

6. Conclusions In view of all the examined data we can be confident that the ‘Villaggio delle Macine’ did not spring up for defensive reasons or for the exploitation of

Figure 10 – The deforestation event in the Alban hills compared to the larger climatic framework (redrawn after Guilizzoni et al. 2002, courtesy of B. Giaccio).

235

Hidden Landscapes

ed themselves to live in an area subject to uncontrollable water-level changes in a way that today can still be observed near the town of Ouedo-Gbadji on the Nokoué Lake, in the Benin (Pétrequin 1984, fig. 2). There, dwellings were built on low piles, on humid soil subject to inundation. The settlement location, which subsequently allowed a quite long sedentarization, was apparently chosen for the obviously resource-rich environment, characterized by the presence of large fertile land surfaces and of the relatively huge forested area available at least in an early phase of the colonization. In conclusion, all the data converge in a consistent scenario that provides an almost complete image of the settlement type, of the use of resources in the area (depicting an intensive exploitation of the full resource spectrum) and of the ancient landscape during the long life of that village.

References Ampolo C. 1980. Le condizioni materiali della produzione. Agricoltura e paesaggio agrario, Dialoghi di Archeologia, n.s., 1, pp. 15-46. Angelini I., Angle M., Artioli G., Bellintani P., Lugli F., Martinelli N., Polla A., Tagliacozzo A. & Zarattini A. 2006. Il Villaggio delle Macine (Castelgandolfo, Roma), Lazio & Sabina 3, pp. 157-168. Angle M. 2008. “Salute a coloro che hanno sete”. Un paesaggio reale sul lago Albano durante il bronzo medio, in Negroni Catacchio N. (ed.), Atti VIII Convegno Preistoria e Protostoria in Etruria, Firenze, Octavo Franco Cantini. Angle M. & Gianni A. 1986. Testimonianze dell’età del bronzo sul lago di Canterno, Quaderni di Protostoria 1, pp. 253-265. Angle M. & Guidi A. 2007. L’antica e media età del bronzo nel Lazio meridionale, in Atti della XL Riunione Scientifica dell’Istituto Italiano di Preistoria e Protostoria, Firenze, Istituto Italiano Preistoria Protostoria, pp. 147-178. Angle M., Cattani L., Franco A.R., Lugli F., Micarelli F., Molinaro A., Rosa C., Savelli A., Tagliacozzo A. & Zarattini A. 2008. Castelgandolfo. Villaggio delle Macine, in M. Angle & C. Belardelli (eds), Repertorio dei siti protostorici, Firenze, Octavo Franco Cantini. Carra M.L., Cattani L. & Rizzi M. 2007. Sussistenza e paleoambiente a Villaggio delle Macine (Lago di Albano, Castel Gandolfo – Roma). Studio carpologico. Atti della XL Riunione Scientifica dell’Istituto Italiano di Preistoria e Protostoria, Firenze, Istituto Italiano Preistoria Protostoria, pp. 775-786. Chondrogianni C., Aritzegui D., Niessen F., Ohlendorf C. & Lister G.S. 1996. Late Pleistocene and Holocene sedimentation in Lake Albano and Lake Nemi (central Italy), in Guilizzoni P. & Oldfield F. (eds), Palaeoenvironmental Analysis of Italian Crater Lake and Adriatic Sediments. Memorie dell’Istituto Italiano di Idrobiologia 55, pp. 23-38. Cocchi Genick D. 1998. L’antica età del bronzo nell’Italia centrale, Firenze, Octavo Franco Cantini.

Cocchi Genick D., Damiani I., Macchiarola I., Peroni R. & Poggiani Keller R. 1995. Aspetti culturali della media età del bronzo nell’Italia centro-meridionale, Firenze, Octavo Franco Cantini. Cremaschi M. 1990. Pedogenesi medio olocenica ed uso dei suoli durante il Neolitico in Italia settentrionale, in Biagi P. (ed.), The Neolithisation of the Alpine Region, Natura Bresciana 13, pp. 71-89. FAO 1976. A framework for Land Evaluation, Soil Bulletin 32. Franco M.C. 1982. L’insediamento preistorico del lago di Mezzano, Roma, Edigraf. Guidi A. & Pascucci P. 1996. Nuovi dati sull’antica età del bronzo nell’area medio-tirrenica, in L’antica età del bronzo in Italia, Firenze, Octavo Franco Cantini, pp. 459-474. Guilizzoni P., Lami A., Marchetto A., Jones V., Manca M. & Bettinetti R. 2002. Palaeoproductivity and environmental changes during the Holocene in central Italy as recorded in two crater lakes (Albano and Nemi). Quaternary International 88, pp. 57-68. Karner D.B., Marra F. & Renne P.R. 2001. The history of Monti Sabatini and Albano Hills volcanoes: groundwork for assessing volcanic-tectonic hazards for Rome, Journal of Volcanology and Geothermal Research 107, pp. 185-219. Magri D. & Follieri M. 1992. Caratteri della biostratigrafia pollinica dell’Olocene in Italia centrale, Memorie della Società Geologica Italiana 42 (1989), pp. 147-153. Nencini P. 2004. Il fiore degli inferi. Papavero da oppio e mondo antico, Roma: Muzzio. Petitti P. & Mitchell E. 1993. Dati preliminari sulla topografia dell’abitato sommerso del Lago di Mezzano, in Vulcano a Mezzano, Comune di Valentano, pp. 17-31. Pétrequin, P., 1984. Gens de l’eau, gens de la terre. Ethnoarchéologie des communautés lacustres, Paris, Hachette. Provenzano N. 1997. Produzione in osso e corno delle terramare emiliane, in M. Bernabò Brea, Cardarelli A. & Cremaschi M. (eds), Le Terramare. La più antica civiltà padana, Milano, Electa, pp. 524-544. Sadori L., Giraudi C., Petitti P. & Ramrath A. 2004. Human impact at Lago di Mezzano (central Italy) during the Bronze Age: a multidisciplinary approach, Quaternary International 113, pp. 5-17.

236

30 A reconstruction of geomorphological and environmental transformations and the peopling of the territory south-east of Rome during recent prehistory Anna Paola AnzideiI, Giovanni CarboniII, Alessandra CelantIII, Renato FunicielloIV, Guido GiordanoIV I Soprintendenza speciale per i Beni Archeologici di Roma, Piazza delle Finanze 1, 00185 Roma, Italy – [email protected] II Dipartimento di Scienze dell’Antichità, Sapienza Università di Roma, Piazzale A. Moro 5, 00185 Roma, Italy – [email protected] III Dipartimento di Biologia Ambientale, Sapienza Università di Roma, Piazzale A. Moro 5, 00185 Roma, Italy – [email protected] IV Dipartimento di Scienze Geologiche, Università di Roma Tre, Largo S. Leonardo Murialdo 1, 00146 Roma, Italy – [email protected], [email protected] Abstract The eastern part of the suburbs of Rome up to the Alban Hills was the focus of extensive archaeological investigations that demonstrated the presence of recent/final Neolithic and Eneolithic settlements as well as of numerous Eneolithic necropolises. Geological studies led to a revision of our understanding of Holocene primary volcanic deposits and lahar flows coming from the crater of the Alban Lake. The lack of early Neolithic settlements where lahar deposits occur in the investigated area agrees with the dates taken immediately below the most impressive lahar deposits, which are referable to a period corresponding to the recent/final Neolithic. Geological and palaeoenvironmental investigations are currently being carried out with the aim of identifying the extent of the lahar flows as well as the timing and the modes of the occupation of the territory immediately above these. These studies are linked to a study of the variations in the vegetation during the Neolithic and the Eneolithic.

The territory of the suburbs of Rome between the Via Prenestina and Via Appia up to the slopes of the Alban Hills was the focus of extensive archaeological investigations carried out by the Soprintendenza ai Beni Archeologici di Roma in particular during the period 1999-2007. These studies demonstrated the presence of recent/final Neolithic and Eneolithic settlements as well as Eneolithic necropolises. At the same time the Department of Geology of the Università di Roma Tre carried out a territorial survey that, thanks to the stratigraphic sections made available by the archaeological investigations, led to a revision of our understanding of Holocene primary volcanic deposits and lahar flows coming from the crater of the Albano Lake. The Alban Hills (fig. 1) are a quiescent volcano. The area of the plain extending from Ciampino to the Caffarella valley represents a geomorphological unit where the fluvial network, eroded during the last glacial period, has been completely obliterated by overlying volcanic and lahar deposits (Funiciello et al. 2002; 2003) that occurred between the end of the Neolithic and the Eneolithic. The Holocene sequence filling the Würmian palaeo-network is called the Tavolato Formation. Stratigraphic markers indicate a depositional interval between the base at 23±7 ky (carbonate level dated with U/Th method by Soligo et al. 2003) and

the top at 5.8±0.1 cal. ky (palaeosol dated with 14C method, in Funiciello et al. 2002). The presence of primary volcanic and lahar deposits indicates the persistence, until very recent times, of phreatic and phreatomagmatic eruptive activity in the Albano crater. The lahar and fluvial deposits may be related to overflows of the Albano Lake with water coming out of the lowest threshold of the crater rim facing the Ciampino plain. In this regard, the description reported by several ancient historians (from Plutarch to Livy) about the sudden overflow of the lake in the 4th century BC suggests that water over-spilling from

Figure 1 – The Alban hills, with the Tavolata formation and the buried site of Quadrato di Torre Spaccata (QTS).

237

Hidden Landscapes

the lake may have occurred and may therefore still occur, possibly originating in a sudden surge of deep fluids rich in CO2. The investigated area so far does not yield any settlement referable to the earliest phases of the Neolithic, with the exception of a few ceramic fragments with impressed decoration of the middle Tyrrhenian type and a handle of the Serra d’Alto type. These were recovered in secondary position in levels attributable to the Final Neolithic in the Quadrato di Torre Spaccata settlement, and were dated to 5280±50 and 5270±50 BP (cal. 1V 4170-4040 and 4170-4090 BC; Anzidei & Carboni 1995). Typologically similar material has been recovered in situ in the locality of Tenuta di Torrenova (Cazzella & Moscoloni 1984), at the top of a small hill that because of its high position was spared by the lahar flows. Archaeological evidence referable to the Middle Neolithic (linear ceramic) is completely absent from the surveyed territory, with the exception of the Casale del Pescatore l site located close to the

ancient lacustrine basin of Castiglione along the Via Prenestina in an area unaffected by lahar flows (Carboni 1992-1993). The dates so far obtained in the G.R.A./Lucrezia Romana area from the levels immediately below one of the most impressive lahar flows (5150±70 BP, 5100±100 BP, 5090±100 BP, cal. 1V 4050-3930, 3990-3770, 3990-3760 BC) correspond to the recent/final Neolithic. The field surveys document in any case the existence of several lahar flows that periodically modified the morphology of the territory, but at the same time created wide open and fertile spaces that allowed the development of intensive agricultural and farming activities, as evidenced by the intensity of human occupation in this territory (Anzidei & Carboni 1995; Anzidei et al. 2007). The vegetation of the Campagna Romana during the last several thousand years was reconstructed from the pollen analyses of Valle di Castiglione and anthracological analyses from the Neolithic settlement of Quadrato di Torre Spaccata (Celant 2000),

Figure 2 – Vegetation reconstructions, 5000 BC and 3000 BC.

238

Anzidei et alii, A reconstruction of geomorphological and environmental transformations…

located above the lahar deposits. It is made up of mixed forest with oak, both deciduous and evergreen, elm, hazel, box, ash, poplar and other subMediterranean elements, and more or less broad grassland clearings. Between 5000 and 3000 years BC there is a significant change in the plant landscape, characterized by a progressive increase in both forest cover, creating closed communities, and floral richness as indicated by the introduction of numerous new taxa (fig. 2). This was an unexpected trend, because it occurred at the same time as the intensification of human settlement and of agricultural and farming activities, and it indicates that both climatic and edaphic conditions were favourable to the development of the forest. The research area is currently being investigated from a geological and palaeoenvironmental point of view with the aim of identifying the extent of the lahar flows whose presence was demonstrated by the archaeological excavations. These investigations have allowed us to identify several levels of lahar deposits interspersed with (sometimes very thick) palaeosols, and to identify the type of flora present in the area when the catastrophic event occurred from several plant impressions at the base of the lahar levels. A study of the timing and the manner of the resumption of human activities in the levels immediately above these sediments is also in progress. This study is linked to a program of pollen analyses in the deposits of the Castiglione river-bed, which aims to clarify the development of the vegetation during the Neolithic and the Eneolithic in the area between the Alban Hills and the Tiber and Aniene rivers.

Romana. in La neolitizzazione tra Oriente e Occidente, Convegno di Studi, Udine 1999, pp. 355-364. Funiciello R., Giordano G., De Rita D., Carapezza M.L. & Barberi F. 2002. L’attività recente del cratere del Lago Albano di Castelgandolfo, Rend. Fis. Acc. Lincei, s. 9, v. 13, pp. 113-143. Funiciello R., Giordano G. & De Rita D. 2003, The Albano maar lake (Colli Albani volcano, Italy): recent volcanic activity and evidence of pre-Roman Age catastrophic lahar events, J. Volcanol. Geotherm. Res. 123, pp. 43-61. Soligo M., Tuccimei P., Giordano G., Funiciello R. & De Rita D. 2003. New U-series dating of a carbonate level underlying the Peperino Albano phreatomagmatic ignimbrite (Colli Albani, Italy), Italian Journal of Quaternary Sciences, spec. Vol INQUA, 16(1bis), pp. 115-120.

References Anzidei A.P. & Carboni G. (eds) 1995. L’insediamento preistorico di Quadrato di Torre Spaccata (Roma) e osservazioni su alcuni aspetti tardo neolitici ed eneolitici dell’Italia centrale, Origini XIX, pp. 55-325. Anzidei A.P., Carboni G., Castagna A., Celant A., Cianca M., Egidi R., Favorito S., Funicello R., Giordano G. & Malvone M. 2007. L’abitato eneolitico di Osteria del Curato-via Cinquefrondi: nuovi dati sulle facies archeologiche di Laterza e Ortucchio nel territorio di Roma, Atti XL Riunione Scientifica IIPP I, pp. 477-508. Carboni G. 1992-93. Giacimento neolitico con ceramiche della facies del Sasso a Casale del Pescatore (Montecompatri – Roma). Documenta Albana, II serie, 14-15, pp. 13-33. Cazzella A. & Moscoloni M. 1984. Testimonianze del Paleolitico superiore e del Neolitico iniziale a Tor Vergata, in Bietti Sestieri A.M. (ed.), Preistoria e Protostoria nel territorio di Roma, L.S.A. 3, De Luca editore, Roma, pp. 105-116. Celant A. 2000. Nuovi dati archeobotanici su ambiente e agricoltura nel Neolitico del Lazio: un esempio dalla Campagna

239

31 Palaeo-environmental observations on the ancient Holocene of the Florentine area Marialetizia CarraI, Paul MazzaII, Sara PerusinIII, Lucia SartiIII I ArcheoLaBio, Centro di Ricerche di Bioarcheologia, Alma Mater Studiorum, Università di Bologna, via S. Vitale 28‐30, 48100 Ravenna, Italy – [email protected] II Università degli studi di Firenze, Dipartimento di Scienze della Terra, via La Pira 4, 50121 Firenze, Italy – [email protected] III Università degli Studi di Siena, Dipartimento di Archeologia e Storia delle Arti, via Roma 56, 53100 Siena, Italy – [email protected]; [email protected] Abstract This paper is a report of preliminary archaeobotanical and archaeozoological analyses in prehistoric and protohistoric settlements in Sesto Fiorentino. The study is aimed at reconstructing the palaeoenvironmental evolution as well as the influences of Holocene human activities in this area.

1. Introduction The present paper is a preliminary contribution to the increasing body of knowledge on the archaeozoology and archaeobotany of Sesto Fiorentino. We analysed remains from various rescue excavations in the Florentine area, spanning a period ranging from the pre-Bell Beaker period to the Roman period: Olmicino (pre-Bell Beaker, 3550-2900 BC cal. 2σ); Olmi 1, Ambrosetti, Querciola, and Semitella (Bell Beaker, 2460-2190 BC cal. 2σ); Lastruccia 3 – strati 8/6 and Frilli Ovest area P1-P3 (post-Bell Beaker, circa 2000-1975 BC cal. 2σ); Campo del Sorgo, S. Antonio – area A, A1, A2 (Bronze Age); and Frilli Ovest – area C (Roman period). These sites are all located in the alluvial plain of Florence, Prato and Pistoia. Specific geomorphological circumstances affected the preservation of bones and botanical remains in this area. Moreover, rescue excavations were being carried out under great pressure from the valley’s constant urbanization, with negative effects on both recovery and accuracy of data recording. Despite this, the accumulated palaeoenvironmental information adds to the existing body of geomorphological and palaeogeographical data, creating a rich database for future reconstruction of past landscapes and of human induced modifications of the natural environment. L.S.

2. Archaeobotanical observations 2.1 Materials and methods

Both micro- and macro-botanical remains have been examined in detail. Standard chemical treatment was used to prepare the samples for palynological analysis: flotation in a heavy fluid was used to concentrate the spores and pollen (the palynological analyses were carried out by Dr. Silvia Ricciardi and Dr. Camilla Calò). The samples of macroremains were collected on the basis of homogeneity

and capillarity flotated, manually sieved with finemesh screens (2.5 mm, 1 mm, 0.5 mm), and finally identified with the aid of stereoscopic microscopy. The plant remains were determined using specialized atlases, but also by comparing them with fossilized and modern specimens. The Sesto Fiorentino sample contains abundant carbonized macroremains of plants. In settlements, plant material is normally found carbonized regardless of the climatic (dry-hot or dry-cold climates) or sedimentary (anoxic perilacustrine sediments) conditions it was preserved in. Most of the Sesto Fiorentino remains have become charred at high temperatures in fireplaces as it happens during the cooking of cereals or the baking of bread. 2.2 Results

Pollen are scarce, possibly because of a clay matrix surplus. Only resistant granules (Tilia, Pinus, Compositae, etc.), and NPP structures (non-pollen palynomorphs) survived the highly selective degradation processes. The analysis revealed low percentages of arboreal pollen and the presence of deciduous Quercus, Corylus, and Fraxinus, as well as a high frequency of anthropogenic plants (Graminaceae, Graminaceae cfr. Cerealia, Chenopodiaceae, Urticaceae, Compositae Cruciferae, Plantaginaceae), which points towards a dense concentration of settlements with intense human activity. Also, an abundance of Concentrycistis (related to the green algae Zygnemataceae) indicates shallow waters and temperate to subtropical conditions. With regard to botanical macroremains, large amounts of cereals (caryopsis and parts of the spike) and of weeds and ruderal species indicate that the cultivated areas were close to the settlements. In various samples of macroremains, cultivated and wild

241

Hidden Landscapes

species are associated (table. 1; fig. 1). The ratios suggest that agriculture was largely preferred over gathering, although distortion due to selective carbonization can not be ruled out.

Only cereals (wheat and barley) were found; legumes are absent. In general only a few wild species were used (fig. 2), such as acorn, hazel, grape, cornel (fig. 3), and some Rosaceae. The presence of acorns, Bell Beaker Olmi 1 Area B

Transitional Frilli Ovest Area P1

Bronze Age S. Antonio A, A1, A2

Roman Age Frilli Ovest Area C

GRAMINACEAE

Hordeum vulgare L.

3.00%

 

Hordeum/Triticum

2.00%

 

Triticum dicoccum Schrank

2.00%

 

Triticum dicoccum/spelta

2.00%

 

Triticum monococcum/dicoccum

0.50%

 

Triticum spelta L.

0.50%

0.50%

 

Triticum sp. L.

8.00%

1.50%

 

Cerealia

70.00%

44.00%

86.00%

43.00%

Crops

 

88.00%

44.00%

91.00%

43.00%

FAGACEAE

Quercus sp. L.

JUGLANDACEAE

Juglans regia L.

Wood

 

1.00%

CORNACEAE

Cornus mas L.

2.00%

1.00%

CORYLACEAE

Corylus avellana L.

0.25%

0.50%

ROSACEAE

Prunus sp. L.

 

Rubus gr. fruticosus L.

 

Rosaceae indet.

2.00%

VITACEAE

Vitis vinifera L.

0.50%

Edgewood

 

5.00%

CHENOPODIACEAE

Chenopodium gr. album

EUPHORBIACEAE

Euphorbia helioscopia L.

GRAMINACEAE

Melica sp. L.

0.05%

 

Graminaceae indet.

0.16%

LABIATAE

Lamium sp. L.

0.05%

LEGUMINOSAE

Vicia sp. L.

0.20%

 

Leguminosae indet.

0.18%

POLYGONACEAE

Fallopia convolvulus (L.) Holub.

0.05%

 

Polygonum aviculare L.

 

Polygonaceae indet.

0.16%

PORTULACACEAE

Portulaca oleracea L.

0.05%

RUBIACEAE

Galium sp. L.

0.05%

UMBELLIFERAE

cfr. Torilis sp. Adanson

0.05%

VIOLACEAE

Viola sp. L.

Herbs

 

1.00%

11.00%

3.00%

15.00%

Indeterminate

 

5.00%

34.00%

4.00%

17.00%

2.00% 1.00%

1.00% 3.00% 3.00%

11.00%

0.50%

0.25% 22.00% 11.00%

2.00%

22.00%

1.20%

14.00%

9.00% 1.00%

1.60% 1.00%

0.20%

1.00%

Table 1 – Percentages of the botanical macroremains.

242

Carra – Mazza – Perusin – Sarti, Palaeo-environmental observations on the ancient Holocene…

100%

Roman Age

90% 80%

Bronze Age 70% 60%

Transitional 50% 40%

Bell Beaker 30% 20%

0%

20%

40%

60%

80%

100%

10% 0% Bell Beaker

Transitional

Bronze Age

Crops

Edgewood

Wood

Herbs

Roman Age

Cornus mas L.

Corylus avellana L.

ROSACEAE

Vitis vinifera L.

Figure 2 – Relative frequencies of the different forest-edge species by period.

Indeterminate

Figure 1 – Relative frequencies of cultivated and wild species by period.

Roman Age

Bronze Age

Transitional

Bell Beaker 0%

Figure 3 – Seed of Cornus mas L. (Cornel).

20%

40%

60%

80%

100%

Chenopodium gr. album

LEGUMINOSAE

Lamium sp. L.

Galium sp. L.

Portulaca oleracea L.

GRAMINACEAE

Viola sp. L.

POLYGONACEAE

Euphorbia helioscopia L.

cfr. Torilis sp. Adanson

Figure 4 – Relative frequencies of the herbaceous vegetation by period.

100%

100%

90%

90%

80%

80%

70%

70%

60%

60%

50%

50% 40%

40%

30%

30%

20%

20%

10%

10%

0% Pre-Bell Beaker

0% Pre-Bell Beaker

Bell Beaker

Domestics taxa

Post-Bell Beaker

Early Bronze Wild taxa

Figure 5 – Relative frequencies of domestic versus wild taxa by period (based on NISP counts).

Bell Beaker

Post-Bell Beaker

Early Bronze

Ubiquitarian

Forest dwellers

Open habitat dwellers

Humid habitat dwellers

Figure 6 – Relative frequencies of wild taxa of the different habitats by period (based on NISP counts).

243

Hidden Landscapes

used as food both by humans and by animals, prove the occurrence of nearby oak trees. Remains of herbaceous plants are rare (fig. 4); they are dominated by arable weeds (Vicia sp. L. and Chenopodium gr. album, which are sometimes also used as food; Galium sp. L. and Portulaca oleracea L.) or by ruderal plants common in cultivated areas (Lamium sp. L, Fallopia convolvulus (L.) Holub.). They are additional proof of human impact on the area. M.C.

3. Archaeozoological observations

biases in the determination of the relative frequencies of small vertebrates. The NISP counts (Number of Identified Specimens) of the different samples were used to approximate the relative frequencies of the taxa. Determining the relative frequencies of domestic versus wild taxa was a means to understand the degree of human impact on the environment. However, it was the wild species that provided most of the palaeoenvironmental information for our reconstruction of the past landscape of the Sesto Fiorentino area.

3.1 Material and methods

The archaeofauna analyzed here derives from the sites of Olmicino (Corridi, pers. comm.); Olmi 1; Ambrosetti; Querciola; Semitella; Lastruccia 3 – strati 8/6; Frilli Ovest area P1-P3 and Campo del Sorgo (Corridi 1997; 2000; 2001; Perusin 2003; Perusin et al. 2008). Unfortunately, the impossibility of sieving due to the rush of the rescue excavations inevitably caused

ubiquitarian Cervus elaphus Pre-Bell Beaker

73.30%

Bell Beaker

57.00%

Post-Bell Beaker

75.00%

Early Bronze

66.60%

Ursus arctos

Vulpes vulpes

Testudo hermanni

3.40%

1.40%

0.70%

8.30% 9.50%

14.30%

Table 2 – Percentages of wild ubiquitous taxa: red deer (Cervus elaphus); brown bear (Ursus arctos); fox (Vulpes vulpes); turtle (Testudo hermanni).

forest

open habitat

Capreolus capreolus

Lepus sp.

Pre-Bell Beaker

20.00%

6.60%

Bell Beaker

15.50%

1.40%

Post-Bell Beaker

12.50% 4.80%

Early Bronze

4.80%

Table 3 – Percentages of wild taxa of forest and open habitats.

humid habitat Arvicola terrestris Bufo bufo Emys orbicularis

3.2 Results

Samples of animal remains included large numbers of large mammals but far fewer small mammals, reptiles and amphibians. The relative frequencies of the domestic (Bos primigenius taurus, Ovis vel Capra, Sus sp.) (total NISP counts = 5299) versus wild taxa (total NISP counts = 210) suggest that hunting was a subsidiary practice for gathering animal food (fig. 5). The wild fauna is dominated by taxa that exhibit broad ecological preferences and tolerances (tables 2, 3 and 4; fig. 6). Capreolus capreolus, the only typical forest dweller in the assemblage, is comparatively well represented, which indicates that woodlands existed near the settlements. Open and especially humid environments are also attested to, although rarely, by Lepus sp. and Arvicola terrestris, Bufo bufo and Emys orbicularis, respectively. S.P., P.M.

4. Conclusions Although still incomplete, the information gathered on the early Holocene flora and fauna at Sesto Fiorentino shows the effects of the impact of early settlement on the local landscape. Humans were already basing their subsistence largely on agriculture, which was one of the principal causes of the rapid deforestation of the environment. The presence of wild faunal and flora in our samples, however, proves that woodlands still persisted at the margins of the settlements. Archaeological research in alluvial plains will greatly benefit from the use of taphonomic analysis, which is very effective in identifying the many different agents involved in site formation. M.C., S.P., P.M.

Pre-Bell Beaker Bell Beaker

0.70%

9.50%

Post-Bell Beaker Early Bronze Table 4 – Percentages of wild taxa of humid habitats.

10.80% 4.20%

References Barfield L.H., Cremaschi M. & Castelletti L. 1975. Stanziamento del Vaso Campaniforme a S. Ilario d’Enza (RE), in Preistoria Alpina 11, Museo Tridentino di Scienze Naturali, Trento.

244

Carra – Mazza – Perusin – Sarti, Palaeo-environmental observations on the ancient Holocene…

Cappers R.T.J. 1995. A palaeoecological model for the interpretation of wild plant species, Review of Vegetation History and Archaeobotany 4, Springer-Verlag, Berlin Heidelberg. Carra M., in press. Plant macroremains from the cave of Sesto Fiorentino “Olmi 1 – Area B”. A preliminary study. Corridi C. 1997. Archeozoologia, in Sarti L. (ed.), Querciola. Insediamento campaniforme a Sesto Fiorentino, Garlatti e Razzai editori, Firenze, pp. 21-32. Corridi C. 2000. Lastruccia 3: i dati archeozoologici nel quadro biocronologico di Sesto Fiorentino, in Sarti L. & Martini F. (eds), Insediamenti e artigianati dell’età del Bronzo in area fiorentina. Le ricerche archeologiche nei cantieri Consiag (1996-1998), Firenze, pp. 45-47. Corridi C. 2001. Archeozoologia del Campaniforme in area fiorentina, in Nicolis F. (ed.), Bell Beakers Today. Pottery, people, culture, symbols in prehistoric Europe, Atti del convegno internazionale di Riva del Garda (TN), 11-16 maggio 1998, vol. 2, pp. 659-661. Costantini L. & Stancanelli M. 1994. La preistoria agricola dell’Italia centro-meridionale: il contributo delle indagini archeobotaniche, in Origini, preistoria e protostoria delle civiltà antiche. Bonsignori editore, Roma. Guidi A. & Piperno M. (eds) 1995. Italia preistorica. Laterza, Roma-Bari. Jacquat C. 1988. Hauterive Champréveyres: les plantes de l’âge du Bronze, contribution à l’histoire de l’environnement et de l’alimentation, catalogue des fruits et graines. Archéologie Neuchâteloise 7-8. Ruau, Saint-Blaise. Kroll H. 1992. Einkorn from Feudvar, Vojvodina, II. What is difference between emmer-like two-seeded einkorn and emmer?, Review of Palaeobotany and Palinology 73, Elsevier, Amsterdam. Perusin S. 2003-2004. L’insediamento della prima età del Bronzo di Campo del Sorgo (Sesto Fiorentino – Firenze): archeozoologia, regimi economici e ambiente, Università degli Studi di Firenze, Facoltà di Lettere e Filosofia, unpublished PhD thesis. Perusin S., Di Giuseppe Z., Corridi C. & Mazza P. 2008. The Sesto Fiorentino fauna – Subsistence strategies from the late third millennium to the early second millennium B.C.: preliminary data, in Bell Beaker in every day life. Proceedings of the 10th Meeting “Archéologie et Gobelets” (Florence – Siena – Villanuova sul Clisi, May 12-15, 2006), Firenze. Pignatti S. 1982. Flora d’Italia. Edagricole, Bologna. Schoch W.H., Pawlik B. & Schweingruber F.H. 1988. Botanical macro-remains. Paul Haupt, Berne.

245

Hidden Landscapes:

DISTRIBUTION PATTERNS

32 Fossil Landscapes in the Fiora Valley. Settlement development of prehistoric and late prehistoric communities read through GIS and Remote Sensing Nuccia Negroni CatacchioI, Martina Rusconi ClericiII, Alberto TagliabueII I Università degli Studi, Milan, Italy – [email protected] II Centro studi di Preistoria e Archeologia, Milan, Italy – [email protected]; www.preistoria.it Abstract After a brief introduction on the concept of ‘fossil landscape’ we examine the basin of the Fiora River, on the border of Tuscany and Latium between the Amiata mountain and the Tyrrhenian sea. This basin, a complex area characterized by volcanic and sedimentary soils, has been studied for many decades, and much geological, ecological and archaeological information is already available. This information was re-examined for the purposes of this study with the aid of GIS and Remote Sensing tools, to analyze the factors that influenced settlement location choices. It is found that in each cultural period the inhabitants had different preferences regarding the geomorphic locations in which to build their villages and their cemeteries: along riverbanks, on top of tuff cliffs with steep slopes, or on hillsides. These preferences appear to depend on ‘mental landscapes’ rather than on any physical changes to the landscape, and are therefore influenced by the ideological and cultural models of the ancient communities.

1. Introduction This paper describes a GIS research project intended to study the settlement dynamics of prehistoric and late prehistoric communities in the valleys of the Fiora and Albegna rivers. Data were collected into a database which currently holds 555 sites in the research area, from the Palaeolithic to the Iron Age. We will briefly introduce the territory of the Fiora and Albegna valleys, before describing the trajectory which led to the completion of the georeferenced database. Next, we comment on some of the cartographic layers we produced and explain the results of the research.

2. The territory The research group based at the Università degli Studi and at the Centro Studi di Preistoria e Archeologia di Milano has been occupied for quite some time in casting the information derived from the subsoil in the form of a structured combination of fossil landscapes. By ‘fossil landscape’ we mean structured clusters of all the archaeological sources for specific ancient communities which succeeded each other in the territory and transformed it. We have studied the basins of the Albegna and Fiora rivers between the Amiata mountain and the Tyrrhenian sea – the rivers run almost parallel to each other, the first still in the territory of Tuscany, a few kilometers north of the border with Latium, and the second just south of that border, for many decades. Much geological, ecological and archaeological data is available for this complex area, characterized by volcanic and sedimentary soils. Most recently, during a workshop in the summer of 2006 on ‘Paesaggi reali e paesaggi mentali’ (real

and mental landscapes), we have tried to assess the importance of the physical characteristics of the territory for the settlement choices of the ancient communities, as opposed to their mentality and their cultural models. We found that there is a correlation between similar geographical and geomorphic landscape types (which didn’t change between Neolithic and Bronze Age), and the observed preferences for specific settlement and burial locations. Such preferences were based on the ascription of a specific meaning to space and territories, whether sacred or functional, and this interaction between man and territory created ‘mental landscapes’ that were influenced by the ideological and cultural models of the ancient communities. As the physical characteristics of the landscape were basically the same during the whole of the Holocene, they do not seem to have been important in the locations chosen for villages, cemeteries, sacred places, etc. These in fact changed according to the different cultural facies: villages went up the steep hills and necropolises came down to the narrow river plains. This means that it is the culture and mentality of ancient communities that guided such choices and created new human landscapes, even as the physical landscape remained the same (these concepts are analyzed in depth in Negroni Catacchio 2008). For the purpose of this paper we will only look at the physical landscape data, which will be interpreted using two different techniques: Remote Sensing and GIS. Our main purpose is not to create a model to identify new sites, but rather to analyze the reasons which influenced the choice of settlement location. Our studies using Remote Sensing started over

249

Hidden Landscapes

a decade ago, in collaboration with Neda Parmegiani and Maurizio Poscolieri of CNR (National Research Institute), and have been widely published. They enabled us to plot the sites known from surveys and excavations onto multispectral Landsat images, so that for each site parameters such as Altitude, Brightness, Greenness and Wetness could be specified. Thus we were able to increase the volume of GIS information in the database and to analyse for each period the physical characteristics which guided the choice of site locations. N.N.C.

3. The geodatabase With regard to the chosen methodology the project had two goals: firstly, the creation of an open architecture that could also easily incorporate new and

Relational alphanumerical DB

GIS station

GIS platform

Media DB

Figure 1 – Digital workflow of the project.

DATA CUBE

different types of data; and secondly, the description of the quality of the information. Regarding the former, using standard systems that are readily available in the market avoids the problem of their becoming obsolete; it also makes it easier to upgrade the system. Our information system is based on a cartographic database, enhanced with an Access routine that contains spatial and chronological information on the sites and that is organized by Site Form and by Archaeological Unit Form. The analyzed archaeological records can thus be combined and plotted onto modern digital cartography, in order to arrive at a hypothetical reconstruction of social, economic and cultural ancient communities. The construction of the geodatabase consisted of the following steps: 1. information coding; 2. structuring of the forms in the database; 3. creation of a control vocabulary to ease the dataentry; 4. design of a user interface in order to simplify the data-retrieval operations and the data visualization; 5. data entry; 6. production of the digital and georelevant cartography; 7. integration of the digital and georelevant data with the database. Step 2, the database design, is related to the analysis of the system structure. The GIS platform has a ‘triangular’ design, whose vertices set up three different IT solutions, each one correlated to the others (figure 1). We also achieved three other important goals: the creation of a data cube (as already mentioned, the open structure makes data entry, grouping and sorting easier); the creation of handy consultation tools, including quick and easy data management during data entry, updating, and analysis; and real- time and customized output of models for data analyses. A.T.

4. Results

ANALYSIS

Figure 2 – Feedback cycle of the project.

RESULTS

In order to register, analyze and visualize our information according to the above-mentioned model, we built a GIS platform aimed at integrated management of archaeological data. Archaeological records were superimposed onto digital cartography in order to reconstruct the social, economic and cultural organization of prehistoric and late-prehistoric communities around the basins of the Albegna and Fiora rivers. The GIS base consists of the program Arcview 8.1 of ESRI, which manages both geodata 250

Negroni Catacchio – Rusconi Clerici – Tagliabue, Fossil Landscapes in the Fiora Valley…

and a numbered archive, and which also retrieves the different graphs formatted with Access. All known sites were georeferenced as points on digital cartography, which had been previously imported in Arcview, but not differentiated by type (dwelling, necropolis, isolated finding etc.) or period. Using the Join tool, each layer was joined to a database so as to have access to all the data at each single GIS site. Furthermore, the site data were sorted both chronologically and typologically into groups of layers through the command Select by Attribute. For each of four periods (Copper Age, Ancient Bronze Age, Middle Bronze Age and Late Bronze Age) we have identified: a. the sub-group of settlement contexts, which includes dwellings, archaeological materials and anthropogenic layers; b. the sub-group of burial site contexts, which includes burial sites, isolated tombs and caves; c. the sub-group of sporadic finds, which includes isolated finds of ceramics and lithics. We then turned to the geomorphological data with the support of the georeferenced geological map sheets 135 and 136 at scale 1:100,000. After combining geological and morphological data we found seventeen areas with similar characteristics: Area Vulcanica Laziale Settentrionale, Caldera di Latera, Selva del Lamone, Area dei Travertini di Ponte San Pietro, Area Vulcanica Laziale Meridionale, Piana di Vulci, Area dei Travertini di Montemerano e Saturnia, Colline di Manciano Nord, Colline di Manciano Sud, Monti di Castro, Colline di Magliano, Colline di Marsiliana, Colline di Capalbio, Area Costiera Meridionale, Colline di Scansano, Bassa Valle dell’Albegna, and Colline di Orbetello. For each of these we created polygon shapes, in order to be able to draw the outline in the GIS and to use these polygons for further analysis. Having structured the data we applied various spatial analytical procedures in order to describe in formal and objective language the spatial structure of our archaeological data, and to determine whether this is random or demonstrates types of spatial organization that might allow us to trace the processes which caused them. Using different colour gradients, this allowed us to visualize site densities in the 17 geomorphologic areas (fig. 3). By distinguishing, for example, between the Copper Age sites from the Rinaldone Culture, the Bell Beaker Culture and the Sassi Neri – Tuscan Group, and then superimposing them on one map in order to show their distribution in the study area, we can observe what has been

Figure 3 – Density of finds in the different parts of the territory.

described as the ‘core area’ of the Rinaldone culture. This extended into the area characterized by tuff plateaus that originate from the lava flows of the Volsini mountains, in particular the area we have called the Northern and Southern Latium volcanic areas. Two interesting ‘lobes’ also appeared, which seem to have encroached upon the upper valley of the Albegna, and to have moved down the river Fiora towards the sea. These lobes could be taken to indicate an important communication channel between southern and northern Etruria, later known as the so-called ‘route of the hoards’. The rest of the study area seems to have been sparsely but uniformly populated, except for one unpopulated area corresponding to the ridge separating the Albegna and Ombrone valleys. Further visualisation of the different cultural aspects of the area enabled us to see that the distribution of Bell Beaker Culture settlements displays a greater uniformity than that of the Rinaldone culture, and that the Sassi Neri – Tuscan Group was limited to the caves of the coastal area. Although the data quality is not entirely satisfactory, we do notice that these population dynamics persisted from the Ancient Bronze Age into the Recent Bronze, and became even more pronounced in the Final Bronze Age, when the attractiveness of the coastal areas decreased while the trend towards favouring the tuff plateaus between Latium and Tuscany (the northern and southern Latium volcanic areas) remained constant. There is far more information on Copper Age burial sites than there is on settlements, and we can conclude that most burial sites contained only a few burials while some, such as La Selvicciola, Ponte San Pietro and Le Calle, had more than 15. We must, however, allow for biases in the sample because many finds were made in the course of rescue excavations. Of all burials, 88% were facing either south (69%) or south-east (19%) and only 12% were facing west

251

Hidden Landscapes

Figure 4 – Locations of cemeteries with numbers of known tombs.

or east. Copper Age burial sites were mainly situated at high elevations or on the edges of plateaus leading to the lower gullies, never in the plain or on valley bottoms. Although our data on settlements are clearly incomplete, these probably occupied the flat areas on top of the higher hills, while the burial sites probably occupied the slopes of those hills. For all successive periods there is more information on settlements than there is on burial sites. During the Ancient Bronze Age, settlements were sometimes situated in dominant positions, but more often along rivers and in valleys between steep slopes. The number of finds declines significantly. Middle Bronze Age settlements were also situated in a variety of locations, but there was a noticeable increase in the number of sites at higher altitudes. There is, however, no evidence yet for the typical defensive hilltop settlements that were to characterise the Final Bronze Age. The Recent Bronze Age shows similar settlement dynamics to the previous period, but there is a further noticeable decrease in the number of finds, which already began with the last phase of the Middle Bronze Age. By contrast, the number of finds increases again significantly during the Final Bronze Age, and settlements are

now located in naturally defensible positions such as steep-sided cliffs alongside watercourses; the corresponding burial sites, if linked to a specific settlement, are usually located on nearby hills. Our data do not allow us — at least for the periods and the areas we looked at — to find a direct link between the geomorphologic characteristics of the landscape and the preferred settlement locations through time. This means that, for example, if in the Ancient Bronze Age people preferred to settle along rivers and on the riverbanks (though we can still find villages in higher positions), we cannot infer that this settlement choice depended on the physical landscape. Rather, the inhabitants, once they decided to settle in the territory, looked for the most suitable location: either the top of a hill, or the hillsides, or a valley or a riverbank. M.R.C.

References Negroni Catacchio N. (ed.) 1981. Sorgenti della Nova – Una comunità protostorica e il suo territorio nell’Etruria meridionale (catalogo della mostra), Milano. Negroni Catacchio N. 2002. Paesaggi d’acque dell’Etruria protostorica, in Negroni Catacchio N. (ed.), Preistoria e Protostoria in Etruria. Atti del Quinto Incontro di Studi, “Paesaggi

252

Negroni Catacchio – Rusconi Clerici – Tagliabue, Fossil Landscapes in the Fiora Valley…

d’acque” Sorano – Farnese 2002, Centro Studi di Preistoria e Archeologia Milano, pp. 3-10. Negroni Catacchio N. 2008. Paesaggi reali e Paesaggi mentali dell’Etruria protostorica, in Negroni Catacchio N. (ed.), Preistoria e Protostoria in Etruria. Atti dell’Ottavo Incontro di Studi, “Paesaggi reali e paesaggi mentali” Pitigliano – Valentano 2006, Centro Studi di Preistoria e Archeologia Milano, pp. 21-40. Parmegiani N. & Poscolieri M. 2008. Il contributo del telerilevamento alla ricostruzione dei paesaggi dell’Etruria meridionale, in Negroni Catacchio N. (ed.), Preistoria e Protostoria in Etruria. Atti dell’Ottavo Incontro di Studi, “Paesaggi reali e paesaggi mentali” Pitigliano – Valentano 2006, Centro Studi di Preistoria e Archeologia Milano, pp. 97-112. Rittatore Vonwiller F., Falchetti F. & Negroni Catacchio N. 1977. Preistoria e Protostoria della valle del fiume Fiora, in La civiltà arcaica di Vulci e la sua espansione, Atti del X Convegno di Studi Etruschi e Italici, Florence, pp. 99-165. Tagliabue A., Negroni Catacchio N. & Cardosa M. 2010. Lands of the Middle Fiora Valley in Prehistory and late Prehistory. From survey to GIS, in F. Niccolucci & F. Hermon (eds), Beyond the Artifact. Digital Interpretaton of the Past (proceedings of CAA2004, Prato 13-17 April 2004), Budapest, Archeolingua, pp. 162-165. Rusconi-Clerici M. & Tagliabue A. 2006. Paesaggi tra Fiora e Albegna, dal survey al Gis: l’Eneolitico, in N. Negroni Catacchio (ed.), Preistoria e Protostoria in Etruria. Atti del Settimo Incontro di Studi, “Pastori e guerrieri nell’Etruria del IV e III millennio a.C. La civiltà di Rinaldone a 100 anni dalle prime scoperte” Viterbo – Pitigliano – Valentano 20034, Centro Studi di Preistoria e Archeologia Milano, pp. 304-314.

253

33 A hidden urban landscape: Populonia between survey and excavations Franco CambiI, Valeria AcconciaII Department of Archaeology and History of Arts, University of Siena, via Roma 56, 53100 Siena, Italy – [email protected] II Department of Antiquity, University of Rome ‘La Sapienza’, piazza A. Moro 5, 00185 Roma, Italy – [email protected] I

Abstract The paper aims to describe the preliminary results of a series of archaeological investigations undertaken in the site of the ancient city of Populonia (Etruria) in the years 2000-2006. Some new archaeological prospections (magnetometry, resistivity, and hand-augering) can be added to the traditional practices of investigation (excavation and survey). These studies are showing that the origins of the ancient town date back to the 8th century BC. The new research conducted in the Gulf of Baratti now contributes to build new hypotheses in relation both to the organization of the port of the ancient city and to the production of iron between the Etruscan and Roman period.

1. Hidden urban landscapes: Populonia and its territory The ancient town of Populonia is a useful case-study which underlines how the theme of hidden landscapes can be focused on urban contexts and their surroundings. Populonia can be considered a sort

of ‘elusive’ example of town. Textual sources, travel reports of the 15th century onward, and research carried out during the late 19th and 20th centuries confirmed its correct location (never really lost) as the only Etruscan town that arose directly on the coast (see Fedeli 1983, pp. 15-44; Strabo V, 2, 6; Plinius

Figure 1 – Populonia and the Gulf of Baratti. a. Poggio del Telegrafo; b. Poggio del Castello; c. Poggio della Guardiola; d. la Porcareccia; e. the San Cerbone-Casone funerary area; f. the Piano and Poggio delle Granate funerary area.

255

Hidden Landscapes

N.H. III, 5, 50-51), at the western edge of the Gulf of Baratti, in front of Elba Island, on the CastelloTelegrafo and Guardiola hills (fig. 1, a-c). Even if its long-period frequentation is well known, from the Early Iron age to romanization, and then through to the Early Medieval Period, its archaeological (and historical) framework is still based mostly on assumptions rather than on material evidence. For instance, our knowledge about the truly Etruscan period of the town (from the 9th to the 3rd century BC, i.e. before the beginning of the romanization process) mainly results from excavations of the funerary areas along the Gulf of Baratti (fig. 1, d-f) (Minto 1922 and 1943; Fedeli 1983; Fedeli et al. 1993, pp. 16-17). For the inhabited area, until the end of last century, substantial archaeological evidence was available only for the two ‘upper’ and ‘lower’ town wall circuits (De Agostino 1962) and for the metal-working area on the Poggio della Porcareccia (Cristofani-Martelli 1979, pp. 75-76; Martelli 1981; Cristofani-Cristofani Martelli 1985, pp. 84-86). The inhabited area within the upper walls was not extensively excavated for the Etruscan period until the 1990s, nor was the absolute chronology of the walls established by accurate stratigraphical excavations: their attribution to the 6th and 4th centuries BC was based only on the manner of their construction and on references by historical sources (for the upper walls, Minto 1943, pp. 18-19; Fedeli 1983, pp. 125-126; Romualdi in Fedeli et al. 1993, p. 108; also Colonna 1981, pp. 443-444). It is only from 1999 that the Universities of Siena and Pisa have been carrying out a research project focusing on the late-Republican phase of the ‘acropolis’ area (located in the depression between the Poggio del Castello and the Poggio del Telegrafo), bringing to light monumental buildings such as temples and a large public complex. Most of our questions about Populonia and its historical topography therefore remain without clear answers or conclusive evidence, although we do have some starting points from previous research. We can therefore specify the main research themes relevant to the urban development of Populonia: • the beginnings of the centre as one of the most important Etruscan towns, traditionally founded by foreign people or by other Etruscan cities (from Corsica or Volterra) with a proper urban status (Servius, ad Aen. X, 172). • the character of the urban organization: the distinction between upper and lower towns, the former being the political and ritual core of the city

and the latter being mainly functional the area for port, commerce and craft activities. • the metallurgy practised at Populonia, its development from the Iron age, and the relationship between Populonia and the iron ores available on Elba Island. F.C, V.A.

2. Between town and countryside The complexity and the scale of the territory of Populonia have been analysed from 1999 with a survey project carried out by the Chair of Landscape Archaeology of the University of Siena. This has shed light on some of the themes presented above, and has highlighted areas of interest such as the top of the Poggio del Telegrafo for the early phases of the town, the ancient routes, the city walls, the ports and the whole of the countryside up to the Campiglia Mountains and the southern edge of the coastal plain. A survey, combined with geophysical analyses, was carried out on the hilltop of Poggio del Telegrafo in 1999-2001. These preliminary studies enabled the research staff to plan a short program of random excavation tests. These were not simple ‘shovel tests’ but a systematic large-scale assessment of monuments and other material traces on the terrain, meant to generate new research questions. The topography of the site after the Orientalising period was investigated with special attention to the city walls and the port location. As mentioned above, the upper and lower wall circuits were dated by their different architectural characters to the late 6th and the 4th century BC respectively. The upper circuit should therefore be the first, built to resist the expansion of Syracusan interests in the northern Tyrrhenian; the lower circuit should be dated in the early Hellenistic period, presumably again to face Syracusan expansion but now also Roman expansion. The first excavations of the lower circuit in the early 20th century did not yield any useful dating evidence, but recent research carried out by the Soprintendenza ai Beni Archeologici della Toscana uses stratigraphical evidence to attribute it to the Archaic period. A survey of the wall circuits carried out with GPS from 2001 to 2006 helped to position its fleeting traces amongst the mediterranean scrub which characterizes the area of the ancient town nowadays. Literary sources suggest that Populonia, like other coastal centres on the Tyrrhenian Sea, had an internal docking basin that may have corresponded to a lagoon now filled in. The application of resistivity survey and a hand-augering test, carried out in 2006 in the area of

256

Cambi – Acconcia, A hidden urban landscape: Populonia between survey and excavations

the Baratti shoreline, south of the Casone buildings, revealed the probable presence of a channel and a natural basin, both filled in, which could be interpreted as traces of this ancient port. F.C.

3. The beginnings of the town: new evidence from the acropolis Surveys carried out on the top of the Poggio del Telegrafo in 2000 revealed the presence, among a dense pattern of Republican and Early Imperial remains, of two scatters of sherds dated between the 9th and the start of the 6th century BC (Aprosio 2002). This suggested that an inhabited area was located on a part of the hill that corresponds to the akron recorded by historical sources (Ptolemaeus Geogr. I, 1, 324, distinguished an upper Poplonion akron and a lower Poplonion polis; Strabo V, 2, 6, distinguished between an upper polichnion-akron and a lower epineion). This akron, composed of the Poggio del Castello, Poggio del Telegrafo and a part of Poggio della Guardiola, was therefore traditionally accepted by scholars as one of the nuclei of early pre-urban Populonia. Other evidence of such an early presence on the Populonia acropolis includes four chamber tombs dated to the 9th-8th century BC, found in 1971-1972 on the western slopes of the Poggio del Castello (Fedeli 1983, pp. 335-336; 1985; 2000; Romualdi 1994, p. 71), and some sporadic finds such as a mould for bronze pins (8th century BC?) found on the western slopes of Poggio del Telegrafo (fig. 2, c) (Fedeli 1982-1983; Fedeli et al. 1993, p. 85). In 1974 further sporadic materials dated to the Iron Age and the Archaic period were picked up from the depression between the Castello and Telegrafo hills (Fedeli 1983, p. 338, site nr. 202, fig. 305); at San Cerbone Vecchio on the Poggio del Castello, in 1914 A. Minto found remains of walls made out of rough and squared blocks with Archaic and Republican pottery (Minto 1914, pp. 412-416; 1943, p. 22; Fedeli 1983, p. 126, pp. 340-341, sites nrs. 205-207, see also fig. 5); and during excavations in 1926 of the lower wall circuit on Falda della Guardiola, under one of the bastions, a bronze hoard came to light that was probably deposited in the 8th century BC but was composed of earlier materials such as a ‘Sa Idda’ type sword, a small bronze nuragic ship, five axes and a fibula (fig. 2, d) (Minto 1926, pp. 347-375; 1943, p. 53; Fedeli 1983, p. 93 and p. 347, nr. 218; Fedeli et al. 1993, p. 86; Romualdi 1994, p. 172). Research carried out by the Soprintendenza per i Beni Archeologici della Toscana and the University

Figure 2 – Populonia, the acropolis. a. the south-eastern slopes excavation area; b. the north-eastern slopes excavation area; c. the Poggio del Telegrafo funerary area; d. the Poggio della Guardiola hoard; e. depression between the Castello and Telegrafo hills; f. Poggio del Telegrafo southern slopes; g. the Punta delle Tonnarelle area; h. the San Quirico area.

of Pisa and Siena since 1984 and in the period 20002007 found traces of pre-Roman presence in the depression between the Castello and Telegrafo hills. Remains of late Archaic structures came to light, associated with residual sherds dated between the 9th and 5th centuries BC (fig. 2, e)(Fedeli et al. 1993, pp. 77, 93 and 110; Romualdi 2002, p. 12; Costantini 2003; Ansaldi 2004; Bigalli 2004). Surveys on the southern slopes of Poggio del Telegrafo revealed the presence of mid-Republican buildings (Dallai 2003, p. 247), as well as an area on the Punta delle Tonnarelle that was frequented for a long time, with materials dated from the Iron Age to the Republican period (fig. 2, f-g) (Shepherd & Dallai 2001, p. 201; Acconcia et al. 2004-2005a, p. 59). Finally, on the Poggio della Guardiola, 4th-3rd century BC structures came to light during the 2005-2006 excavations of the medieval San Quirico monastery (fig. 2, h).1 Briefly, it can be concluded that the first phase of the city, which could be called ‘proto-urban’ (9th8th century BC) is attested over the entire surface of the acropolis. These findings should be added to the bronze hoard of Falda della Guardiola. The period of greatest vitality of the lower city, between the 6th and the 4th century BC, is surprisingly absent or scarcely documented on the two hills of Poggio del Telegrafo and Poggio del Castello, and is yet to be satisfactorily investigated in the low area between the two tops. The Acropolis comes into use again,

257

Hidden Landscapes

Figure 3 – The Poggio del Telegrafo north-eastern slopes excavation area (from Acconcia et al. 2006).

and with great intensity, from the early 3rd century BC onwards. It is conceivable that the reoccupation of an area so important, especially from the ideological point of view, was a consequence of the Roman conquest of the area. The framework which emerges from these observations is obviously incomplete, with occasional suggestions of the use of the acropolis before romanization, and does not provide much information about the settlement pattern or its development for the early phases of the town. The most widely accepted theories about this period were constructed by connecting these occasional traces of inhabited sites to knowledge derived from funerary areas. A continuity in the settlement pattern from the Late Bronze Age to the Iron Age is traditionally maintained for Populonia, during which the centre is supposed to correspond to the ‘sparse-settlement site’ model. In this model, the main Iron Age necropolis along the Baratti Gulf is linked not to Populonia but to other settled areas (cfr Bartoloni 1991, p. 5; Fedeli et al. 1993, p. 102; Pacciarelli 2001, p. 135; Fedeli 1983, pp. 79-90; Bartoloni 1991, p. 10), the evidence for which was found in scattered Late Bronze Age

finds from the Pineta del Casone (middle part of the Gulf), from the slopes of Poggio delle Granate (northern part), and from the Poggio del Molino, the latter connected to the small necropolis of Villa del Barone (Fedeli 1980-1981; Fedeli et al. 1993, pp. 7072 and 78-79; for a bronze fibula (late Bronze age?) from the Baratti shore, see Bartoloni 1991, p. 5). Assuming a continuity of use of the Bronze Age sites, the San Cerbone–Porcareccia–Casone funerary area has been related to a hypothetical village located in the inner part of the gulf, and the necropolis of Piano and Poggio delle Granate to another village which continued the site on Poggio del Molino. The small Iron Age chamber tomb nucleus on the Poggio del Telegrafo, would then be the only one related to a site on the acropolis. It is only during the 7th century BC that this settlement pattern is supposed to have been changed by a centralization of the earlier sparsely inhabited nuclei, a process that in the 6th century BC resulted in the definition of an urban status for Populonia, as a polis (Fedeli 1983, pp. 93-94; Romualdi 1994, pp. 171 and 180). This model for Populonia, however, is not supported by material evidence. The Late Bronze Age sites along the gulf do not show a long-term frequentation during the Iron Age, and the relation between the necropolis of Poggio delle Granate and the site at Poggio del Molino appears to be different: the necropolis was used from the early of Iron Age onwards, precisely when Poggio del Molino (and its funerary area) was abandoned. More generally, the ‘sparse-settlement sites’ model was applied also to other Etruscan cities but was brought into question by most recent scholars, who now recognize a single formation period after the end of the Bronze Age in places such as Tarquinia, Veii, Vulci and also the northern town of Volterra (Pacciarelli 2001; Zanini 1997, pp. 160-163). As previously said, the 2003-2006 research on Poggio del Telegrafo revealed the presence of an extended area used from the Iron Age onwards, which can considered a new starting point to interpret the beginnings of the town. The University of Rome ‘La Sapienza’, in collaboration with Siena University, investigated two areas on the Poggio del Telegrafo: one on the southern slopes of the upper terrace of the hill, the other on the north-eastern side, above the Republican complex known as ‘Le Logge’ (Acconcia et al. 2004-2005a-b; Acconcia et al. 2006; Acconcia & Bartoloni 2007) (fig. 2, a-b). Both areas seem to be inhabited from the 9th century BC, and to have been abandoned in the early 6th century. Their use began

258

Cambi – Acconcia, A hidden urban landscape: Populonia between survey and excavations

— as in other Etruscan towns — with the Iron Age (even if for Tarquinia, Vulci and Volterra we know of a Late Bronze Age phase) and continues in the 8th and 7th centuries BC without significant interruptions. In the north-eastern slopes area, traces of the late th 8 – early 7th century are preserved. Here, the superimposition of Republican buildings destroyed structures related to the second half of the 7th century, but the phase is attested by residual materials (bucchero and etrusco-corinthian pottery) in later archaeological levels (fig. 3). In the southern slopes area, traces of a 9th century oval hut came to light, succeeded by a square one dated to the beginnings of the 7th century and ritually destroyed with the deposition of a significant number of kyathoi in one of the postholes (fig. 4). In the same site another similar hut was then built during the first half of the 7th century, and a series of structures followed each other until the late 7th – early 6th century BC. Both excavation areas are located only 80 m apart on the top of the Populonia acropolis and, although excavated separately, they must be part of the same inhabited site. We may assume the site was located in a relevant position, well defended, and overlooking the territory, the port and the sea with the islands as well.

The material character of the site suggests the presence of a community with evident status markers. The pottery found in both areas includes types and shapes usually found in the Iron Age and Orientalising-period tombs of the Baratti necropolis. For the Iron Age, with the usual incised pottery, there are a small number of Sardinian bottles — the first time that this shape occurs in inhabited areas. For the following period, although the presence of imported pottery is still very sporadic, the fine impasto and bucchero shapes agree with those typical of funerary assemblages at Populonia in the same period (see Pitzalis & Milletti 2004-2005; De Castro et al. 2006). The location, the evidence for ritual practices connected to the use and abandonment of the hut structure, and the character of the materials found for this period, all suggest that the top of the Poggio del Telegrafo could already from the First Iron Age have been the site of the Populonia aristocratic elites. The establishment of the proto-urban centre would then fit the settlement pattern known from other Etruscan towns, with a clear concentration process occurring in the 9th-8th centuries, connected with the abandonment of sparse Late Bronze age

Figure 4 – The Poggio del Telegrafo south-eastern slopes excavation area: the rectangular hut and the ritual pit of the kyathoi (from Acconcia & Bartoloni 2006).

259

Hidden Landscapes

sites (cfr Zanini 1997). The presence of the funerary nucleus of Poggio del Telegrafo is clearly problematic in this regard, but its brief use could be connected with the Falda della Guardiola bronze hoard mentioned before; recently G. Bartoloni suggested that its location might be connected to the urban boundaries marked at a later stage by the lower wall circuit (Bartoloni 2004, p. 247). The ancient town must have from its beginnings included the Telegrafo-Castello and Guardiola hilltops, even if this became evident only when the monumental walls were built. We cannot exclude the possibility of a ritual definition of the town limits, by features such as the Falda della Guardiola hoard and other sacred areas, as recently highlighted by A. Zifferero (Zifferero 2005). Nor can we exclude that the lower wall circuit may have replaced an older one, perhaps built in another technique, such as excavations have recently revealed at Veii and Tarquinia (see Boitani 2008; Baratti et al. 2008). The Falda della Guardiola hoard, marking the extent of the town area, could also mark the organization of a proper centralized community during the 8th century BC, when the use of the necropolis of Poggio del Telegrafo came to an end. The abandonment of the Poggio del Telegrafo upper terrace at the start of the 6th century BC

— which seems not include the lower depression (Romualdi 1993, p. 110) — may be connected with a change in the social and economic order. For this period we do not see a decline in funerary areas or any sign of decline for the rest of the town. On the contrary, the graves seem to be wealthy, although the monumental type of the tumulus tomb is replaced by more regular square-plan tombs. In this period, intensive metallurgy begins in the lower town (Fedeli 1983, p. 119; Romualdi 1993, p. 108; Martelli 1981). This suggests a radical change in social order, which must adapt aristocratic assets to the needs of a dynamic system. The installation of craftsman quarters on the slopes of Poggio della Guardiola and Poggio della Porcareccia seems to be contemporary with the reduction of the site on the acropolis. Although we do not know about the development of the rest of the hypothetical urban area (the slopes of the Poggi, the port, etc.), we may assume that events there would be connected. From this period onward, the top of the Poggio del Telegrafo could have been used as the ritual core of the town, as is shown by the presence of later monumental temples and buildings both on the upper terrace and in the saddle depression. A new political and social order began, directed towards the intensive exploitation of iron ores and the production of metal. V.A.

Figure 5 – The Gulf of Baratti. 1. the Poggio della Porcareccia buildings; 2. the Baratti shore excavation area; 3. the Campo Sei area (from Acconcia & Giuffré 2009).

260

Cambi – Acconcia, A hidden urban landscape: Populonia between survey and excavations

4. Excavations on the Baratti shoreline (2002-2006) The connections between landscape archaeology and intensive excavation shed new light also on the theme of ancient metallurgy and its relations with the urban development at Populonia. The natural bent of the town and its territory towards the exploitation of the ‘mixed sulphides’ and iron ores of the Campigliese district and of Elba Island was traditionally known from literary sources (Corretti 2004; Colonna 1981; Zifferero 2002). During the last century, first A. Minto and then M. Cristofani and M. Martelli excavated buildings in the Poggio della Porcareccia area outside the lower wall circuit. This was used from the 6th to the 3rd century BC as a living and productive quarter, with traces of intensive iron metallurgy (fig. 5, 1). The excavation has recently been started again by the University of Pisa (Bonamici 2004-2005; 2008). Outside the walls, in the so-called Campo Sei, another extensive metallurgical and residential area was recently excavated but not yet widely published (Facchetti & Romualdi 2000, pp. 181-182 )(fig. 5, 3). Recently, surveys combined with geophysical analysis brought to light a third site, just in front of the shoreline at Baratti, north of the Fontanile di San Cerbone, which is characterised by a long-time frequentation. This site was excavated from 2002 by the University of Siena, in collaboration with the Universities of Florence and Rome (fig. 5, 2). It, too, was situated outside the lower wall circuit and not far from the Poggio della Porcareccia and Campo Sei; we can therefore assume the presence of a large metalworking quarter already from the 6th century BC. The intensity of craft production is revealed by the huge iron slag-heaps spread out into the gulf and excavated in the first half of the 20th century in order to rescue metals useful in supporting the Italian economy during World Wars I and II. The site on the Baratti shoreline was first identified by prof. M. Benvenuti and L. Chiarantini from Florence University and by S. Guideri, director of the Archaeological Park of Populonia, on a cliff strongly eroded by the sea. The traces of intensive metallurgy visible on this cliff can be related to ancient craft structures, such as furnaces or forges. Prof. Benvenuti and Chiarantini then arranged the sampling of slags and charcoals, some of which, associated with copper slags taken from the bottom of the cliff, were dated to the middle of the 8th century BC and therefore suggest an early stage of metallurgy, probably based on the exploitation of

Figure 6 – The Baratti shore area: charcoal and slag layers.

copper ores quarried in the Campigliese hinterland (Chiarantini et al. 2004-2005; 2006). This early date for the beginning of metallurgy at Populonia was confirmed by the already mentioned surface find of an 8th century BC mould for bronze pins on the Poggio del Telegrafo, and subsequently by a similar one found in the same area in 2004. The excavations on the Baratti shoreline carried out during 2002-2003 and 2005-2006 (Cambi et al. 2007; Acconcia 2008; Acconcia & Giuffré 2009) brought to light a complex stratigraphical sequence. The area was heavily disturbed by a large pit made in 20th century to find and extract iron slag. This pit, irregular in shape and in depth, destroyed a large part of the ancient archaeological levels, exposing at various stages different phases and their related evidence. At the moment, the earliest deposits investigated are assigned to a 3rd to 1st century BC metalworking phase; remains of craftsmen buildings following one another in quick succession have been documented, in association with forges and working surfaces made from layers of charcoal, burnt clay, roasted haematite, slag, etc. (fig. 6). The latter appear to have been mainly composed of discard materials, in order to raise rapidly the working level. The area probably did not have a residential use in this period, only an industrial one. Iron metallurgy seems to be exhausted during the 1st century BC, and the site seems to be occupied again from the middle of the 1st century AD for residential purposes. The excavations of 2002-2003 and 2005-2006 therefore

261

Hidden Landscapes

probably brought to light the final phase of intensive craftsmanship at Populonia, which was mostly directed to iron metallurgy. The radiocarbon dates cited above suggest that the beginnings of this process can now be dated to the Late Iron Age, perhaps with the production of copper and bronze, changing into intensive iron metallurgy certainly before the 6th century BC. The end of the craftsmanship use of the area occurs in the late Republican period, and could be caused by a sort of relocation of the iron manufacture. Alternatively, one might think that the end of the metallurgical activities in the Gulf of Baratti is to be connected with a decree of the Senate called ‘ancient’ by Pliny the Elder.2 The decree would have involved, according to the writer, the end of mining and metallurgical activities on the Italian peninsula. Although the date and character of the senatusconsultum are still disputed, and although it is possible that the closing of the mines did not include those on Elba Island, there are no later traces of intensive metallurgy in the Baratti area. In the Imperial period the area was occupied by two residential buildings separated by a small road. The heaps of iron slags are covered now, and invisible. In the second half of the 2nd century AD the building becomes a cetaria for salting fish. A few years later a man and a woman of high social status (as shown by the wealth of funerary depositions, in particular the gold ornaments worn by the woman) are buried not far from the building. From then on until late antiquity, an entirely different story began, very far from the noise and pollution of previous centuries. F.C., V.A.

References3 Acconcia V. 2008. Lo scavo della Spiaggia di Baratti: la campagna 2006, in Acconcia V. & Rizzitelli C. (eds), Materiali per Populonia 7, Pisa, pp. 227-242. Acconcia V. & Bartoloni G. 2007. La casa del re, in Botarelli L., Coccoluto M. & Mileti M.C. (eds), Materiali per Populonia 6, Pisa, pp. 11-29. Acconcia V. & Giuffré E.M. 2009. Lo scavo della Spiaggia di Baratti: campagne 2007-2008, in Ghizzani Marcìa F. & Megale C. (eds), Materiali per Populonia 8, Pisa, pp. 127-160. Acconcia V., Milletti M. & Pitzalis F. 2004-2005. Populonia, Poggio del Telegrafo: le ricerche nell’abitato degli anni 2003-2004, Scienze dell’Antichità 12, pp. 57-69. Acconcia V., Di Napoli A., Galante G., Milletti M. & Pitzalis F. 2004-2005b. Poggio del Telegrafo (Piombino, LI): saggi di scavo sull’acropoli di Populonia (PdT 2003), Rassegna di Archeologia 21B, pp. 9-44. Acconcia V., Carraro C.C., De Castro F.R., Gabbrielli L., Galluzzi G., Milletti M., Mottolese C., Nomi F., Palone V., Picucci S.,

Re V. & Taloni M. 2006. Scavi sulla sommità del Poggio del Telegrafo o del Molino, in Aprosio M. & Mascione C. (eds), Materiali per Populonia 5, Pisa, pp. 13-78. Ansaldi C. 2004. Brevi note su alcuni reperti d’epoca romana, in Gualandi M.L. & Mascione C. (eds), Materiali per Populonia 3, Firenze, pp. 129-132. Aprosio M. 2002. Le ricognizioni sull’acropoli di Populonia, in Cambi F. & Manacorda D. (eds), Materiali per Populonia 1, Firenze, pp. 43-50. Baratti G., Castaldi M. & Mordeglia L. 2008. La cinta fortificata di Tarquinia alla luce della nuova documentazione, in La città murata in Etruria, Atti del XXV Convegno di Studi Etruschi e Italici (Chianciano Terme – Sarteano – Chiusi 2005), Pisa–Roma, pp. 155-170. Bartoloni G. 1991. “Populonium etruscorum quodam hoc tantum in litore”. Aspetti e carattere di una comunità costiera nella prima età del Ferro, ArchClass XLIII, pp. 1-36. Bartoloni G. 2004. Populonia: l’insediamento della prima età del Ferro, in Gualandi M.L. & Mascione C. (eds), Materiali per Populonia 3, Firenze, pp. 237-249. Bigalli C. 2004. Un frammento di ceramica attica a figure nere: una proposta di lettura iconografica, in Gualandi M.L. & Mascione C. (eds), Materiali per Populonia 3, Firenze. Boitani, F. 2008. Nuove indagini sulle mira di Veio nei pressi di porta Nord-Ovest, in La città murata in Etruria, Atti del XXV Convegno di Studi Etruschi e Italici (Cianciano Terme – Sartiano – Chiusi 2005), Pisa–Roma, pp. 135-154. Bonamici M. 2004-2005. Recenti indagini nel quartiere industriale di Populonia, in Scienze dell’Antichità 12, pp. 89-103. Bonamici M. 2008. Nuove ricerche nel quartiere industriale di Populonia, in Etruschi, Greci, Fenici e Cartaginesi nel Mediterraneo centrale, Atti del Convegno di Orvieto 2007 (Annali della Fondazione per il Museo “Claudio Faina” XIV), Roma, pp. 431-453. Cambi F., Acconcia V., Camusso G. & Quaglia L. 2007. Lo scavo della Spiaggia di Baratti (Populonia), in Botarelli L., Coccoluto M. & Miletti M.C. (eds), Materiali per Populonia 6, Pisa, pp. 303-334. Chiarantini L., Benvenuti M. & Guideri S. 2004-2005. Recenti ricerche sui processi di riduzione del ferro nel Parco di Baratti e Populonia nel I Millennio a.C., Rassegna di Archeologia 21B, pp. 171-182. Chiarantini L., Benvenuti M. & Guideri S. 2006. La produzione di rame, ferro e bronzo a Populonia in epoca etrusca: nuove acquisizioni, in Cavallini M. & Gigante G.E. (eds), De Re Metallica. Dalla produzione antica alla copia moderna, Roma, pp. 171-182. Colonna G. 1981. Presenza greca ed etrusco-meridionale nell’Etruria Mineraria, in L’Etruria Mineraria, Atti del XII Congresso di Studi Etruschi e Italici (Firenze, Populonia, Piombino 1979), Firenze, pp. 443-452. Corretti A. 2004. Per un riesame delle fonti greche e latine sull’Isola d’Elba nell’antichità, in Gualandi M.L. & Mascione C. (eds), Materiali per Populonia 3, Firenze, pp. 269-289. Costantini A. 2003. La ceramica corsa, in Mascione C. & Patera A. (eds), Materiali per Populonia 2, Firenze, pp. 55-60. Cristofani M. & Martelli M. 1979, Ricerche archeologiche nella zona “industriale” di Populonia, in Prospettiva 16, pp. 75-76.

262

Cambi – Acconcia, A hidden urban landscape: Populonia between survey and excavations

Botarelli L. & Dallai L. 2003. La ricognizione archeologica nel Golfo di Baratti. Rapporto preliminare, in Mascione C. & Patera A. (eds), Materiali per Populonia 2, Firenze, pp. 233250. De Agostino A. 1962. La cinta fortificata di Populonia, in Studi Etruschi 30, pp. 1-52. Facchetti F. & Romualdi A. 2000. Etruschi, Firenze. Fedeli F. 1982-1983. Forma per fusione da Populonia, Rassegna di Archeologia 3, pp. 157-167. Fedeli F. 1983. Populonia. Storia e territorio, Firenze. Fedeli F., Galiberti A. & Romualdi A. 1993, Populonia ed il suo territorio. Profilo storico-archeologico, Firenze. Martelli M. 1981. Populonia: cultura locale e contatti con il mondo greco, in L’Etruria mineraria, Atti del XII Convegno di Studi Etruschi e Italici (Piombino–Portoferraio 1979), Firenze, pp. 399-427. Minto A. 1914. Populonia. Relazione preliminare intorno agli scavi governativi nella necropoli eseguiti nell’anno 1914, NotSc, pp. 411-418, pp. 443-463. Minto A. 1922. Populonia. La necropoli arcaica, Firenze. Minto A. 1926. Populonia. Relazione sugli scavi archeologici durante il 1925-26, NotSc, pp. 346-378. Minto A. 1943. Populonia, Firenze. Pacciarelli M. 2001. Dal villaggio alla città. La svolta protourbana del 1000 a.C. nell’Italia tirrenica, Firenze. Romualdi A. 1994. Populonia tra la fine dell’VIII e l’inizio del VII sec. a.C.: materiali e problemi dell’orientalizzante antico, in La presenza etrusca nella campagna meridionale, Atti del Convegno (Salerno – Pontecagnano 1990), Firenze, pp. 171-180. Romualdi A. (ed.) 2002. Populonia: ricerche sull’Acropoli, Pontedera, pp. 143-177. Shepherd E.J. & Dallai L. 2002. Attività di pesca al promontorio di Piombino (I sec. a.C. – XI sec. d.C.), in Atti del II Convegno nazionale di Archeologia Subacquea (Castiglioncello 2001), Bari 2002. Zanini A. (ed.) 1997. Dal Bronzo al Ferro. Il II millennio a.C. nella Toscana centro-occidentale, Catalogo della mostra (Livorno 1997), Pisa. Zifferero A. 2002. Attività estrattive e metallurgiche nell’area tirrenica. Alcune osservazioni sui rapporti tra Etruria e Sardegna, in Etruria e Sardegna centro-settentrionale tra l’età del Bronzo finale e l’Arcaismo, Atti del XXI Convegno di Studi Etruschi ed Italici (Sassari, Alghero, Oristano, Torralba 1998), Pisa – Roma, pp. 179-211. Zifferero A. 2006. Confini e luoghi di culto nel suburbio e nell’agro populoniese: un contributo alla ricerca, in Aprosio M. & Mascione C. (eds), Materiali per Populonia 5, Pisa, pp. 391-427.

that Elba was exempted from the so-called interdictum issued by the Roman Senate. 3

Since 2007, the scene of archaeological research in Populonia (both the Acropolis and the shoreline of Baratti) has been enriched with the following important publications: Botarelli L., Coccoluto M. & Mileti M.C. (eds) 2007, Materiali per Populonia 6, Pisa, ETS. Acconcia V. & Rizzitelli C. (eds) 2008. Materiali per Populonia 7, Pisa, ETS. Ghizzani Marcìa F. & Megale C. (eds) 2009. Materiali per Populonia 8, Pisa, ETS. Cambi F., Cavari F. & Mascione C. (eds) 2009. Materiali da costruzione e produzione del ferro. Studi sull’economia populoniese fra periodo etrusco e romanizzazione, Bari, Edipuglia.

Notes 1

We thank prof. G. Bianchi for showing us the Hellenistic pottery from the S. Quirico excavation. 2

Plinius, N.H. 3, 138. Haec est Italia diis sacra … metallorum omnium fertilitate nullis cedit terris; sed interdictum id vetere consulto patrum Italiae parci iubentium. N.H. 33, 78: Italiae parci vetere interdicto patrum diximus; alioqui nulla fecundior metallorum quoque erat tellus. In N.H. 34, 41, however, Pliny maintains

263

34 ‘Background noise’ and landscape exploitation in the Late Iron Age Andalusian countryside Victorino Mayoral HerreraI, Antonio UriarteII, Teresa ChapaIII, Juan VicentIV, Ana CabreraV I Instituto de Arqueología-Mérida (CSIC – Junta de Extremadura – Consorcio de Merida), Spain – [email protected] II Departamento de Prehistoria, Instituto de Historia (CSIC), Spain – [email protected] III Universidad Complutense de Madrid, Departamento de Prehistoria. Facultad de Geografía e Historia, Ciudad Universitaria s/n, 280040 Madrid, Spain – [email protected] IV Departamento de Prehistoria, Instituto de Historia (CSIC), Spain – [email protected] V Universidad Complutense de Madrid, Departamento de Prehistoria, Spain – [email protected] Abstract In this paper we present the results from a survey carried out from 2000 to 2003 in the Guadiana Menor valley (Jaen province, Spain). The sampling strategy was designed to explore the presence of archaeological finds in very different landscape units, from the floodplains to the uncultivated slopes of the Cazorla massif. Field survey revealed very high contrasts in find densities, suggesting different patterns of land use. On the valley floor we recorded a weak carpet-like pattern of sherds dating from the Iron Age to Roman times. Its distribution suggests intensive agricultural activities such as manuring in a context of increasing pressure on resources during the Romanization process. In the highland area, the much more scarce evidence seems to be connected with pastoral activity, intensive farming of small valleys, and control over strategic mountain passes. We will here present methodological issues relating to the field survey and reliability problems raised by erosion and land use history of the study area.

1. Introduction From the early 1980s onward the Complutense University of Madrid, in collaboration with CSIC (the Spanish Council for Higher Research) and the University of Jaen, has conducted a regional project on the landscape evolution of the upper Guadalquivir valley (eastern Andalusia) from Prehistory to Roman times. The study area focuses on the Guadiana Menor valley, an important route that connects Andalusia and south-eastern Spain (fig. 1). This region had a very strong previous research tradition on the Iberian Culture (6th to 1st centuries BC). In the first half of the 20th century several great funerary areas were excavated in Tugia (Peal de Becerro, Jaen; Pereira Sieso 1979, Madrigal Belinchón 1997, Blánquez Pérez 1999), Castellones de Céal (Hinojares, Jaén; Chapa Brunet et al. 1998), Basti (Baza, Granada; Presedo Velo 1982) and Tutugi (Galera; Pereira Sieso et al. 2004). Their work and the reviewing of old research has been the main focus of the project until the end of the 1990s. During this stage, archaeological surveys were directed toward the selective examination of certain location types, extrapolating settlement patterns from several historic periods. In 1998 a new interdisciplinary approach was begun, with the aim of incorporating new spatial technologies in the study of landscape evolution (Chapa et al. 1998). Our team carried out new survey campaigns, dealing with both micro-scale analysis of sites and extensive prospection of wide areas. The

mountainous and semi-desert nature of the region challenged traditional survey methods and demanded continuous adaptation. Intense erosive activity generates very complex patterns of sherd distribution. As has happened in other Mediterranean regions, we have documented widely dispersed artifacts that cast doubts on the traditional definition of archaeological sites. With regard to the two more extensive survey campaigns we conducted in 2001 and 2003, the first was focused on the middle course of the Guadiana Menor river,

Figure 1 – Location of the study area in the Iberian Peninsula. Distribution of main archaeological sites of the Late Iron Age cited in the text.

265

Hidden Landscapes

where it passes through a narrow valley flanked by the Magina and Cazorla mountains. Altitude and geological differences create strong contrasts in land use and water availability. High mountain areas provide good pasture in summer for herds coming from the Guadalquivir valley floor. Middling heights are mainly calcareous terrain with dense conifer woods and many springs. Closer to the river, gypsum is the predominant substrate, giving a poor support for agriculture. Here the sparse vegetation consists of shrubs, some of great economic significance like the esparto grass. In contrast, the river floodplain provides clayey soils very suitable for intensive agriculture by irrigation. The campaign of 2003 was conducted along the lower course of the Guadiana Menor, close to its confluence with the Guadalquivir. This is an area with a much softer relief, mainly of calcareous origin. The valley is very open, with an increasingly wide river bed. The predominant land use in traditional agriculture was dry cultivation of wheat and barley combined with olive trees, but in recent times there has been a great expansion of the latter.

2. Objectives and methodology The aim of our survey strategy was twofold. On the one hand, we wanted to develop methods to obtain a high resolution picture of the distribution of archaeological remains over large areas. This required the design of reliable, flexible and not very expensive protocols for the recording of surface finds. But obtaining more abundant and higher-quality data also put more demands on the need for developing information systems for their management and analysis.

On the other hand we were keenly aware that it is the landscape as a whole, not just the archaeological remains in themselves, that provides the information necessary for the reconstruction of its historical evolution. We must therefore incorporate every available source of evidence, from palaeobotany to ethnography. We must identify longe-dureé land use patterns, labor processes and changing territorial structures. Since we were aware of the factors affecting the representativity of surface collections, we chose an experimental, non-reconstructive approach. Survey methods were designed with an analytical rather than a descriptive purpose in mind. One may think that the desideratum of any landscape study should be to obtain a rigorous picture of ‘how it really was’ in the past but we think it may be more profitable to focus on the understanding of the distribution of preserved and observed evidence. Our survey strategy therefore took the individual find as the most elementary unit. At first, the analytical work concentrates on assessing differences in find density and other variables (size, weight, roundness, shape, function, chronology…) over the study area. After that, all these data are correlated with contextual information such as surface visibility, topography (prominence, slope, drainage capacity), land use and any other factors related to the alteration of archaeological deposits and its interpretation. The basic criterion for the delimitation of survey areas was to provide a representative sample of the diversity of landscape units in the region outlined above. We therefore used a stratified random sampling design, each sampling area being a square with

Figure 2 – Views of the different survey areas in each landscape unit.

266

Mayoral Herrera et alii, ‘Background noise’ and landscape exploitation in the Late Iron Age…

Figure 3 – Setting up the survey sampling design (left) and field grid (right).

Figure 4 – Layout of survey sample areas with density of Iberian finds in the lower Guadiana Menor valley. A, B, G: open countryside; C, D: alluvial plain; E, F: steep uncultivated slopes.

a surface area of 26 ha (we used squares of 510 × 510 in order to include 17 × 17 sampling units of 30 m diameter: this value was selected because we were using a Landsat TM image with this pixel resolution for land use classification). Within each of these squares, a grid defines a population of 289 sampling

units of 30 by 30 m, and a random sample of 30 units is prospected. Data collection was accomplished within a small circular area, which allows for fast recording in hilly terrain. The procedure for each unit used the following steps:

267

Hidden Landscapes

• Location using GPS navigation. We used a single-frequency receiver (Trimble Geoexplorer 3) with an accuracy of around 5 m. Originally we had to load each pair of coordinates, previously obtained with a GIS software, individually. Once found, the new location was recorded in order to calculate the difference between the expected and observed coordinates. Differential correction was made by post processing. • Setting up of the sampling unit. Once its center had been determined, its limits were marked by four stakes placed 15 m away from it. • The inspection of the area outlined in this manner was conducted with a fixed number of prospectors (4) during a fixed time (10 minutes). The aim was to keep observation conditions as standardized as possible. All the finds were collected. • During the unit examination, a form was filled in to record factors that can cause bias with regard to the surface archaeological record. • Documentation of the unit was completed by photographs and sketches. Finds were processed in the laboratory. We tried to provide as detailed as possible a description of the pottery. Quantitative analysis was structured in several stages of increasing complexity, from basic descriptive statistics to significance tests and numerical taxonomy.

each survey area, establishing meaningful links between surface visibility and terrain characteristics. The highest off-site find densities were recorded in the open countryside close to the Guadalquivir river. Quantifying sherd roundness and fragmentation, we were able to obtain a highly detailed picture of the low-density distributions around sites such as the late Iberian settlement of Cortijo de los Castellones. This is a small site close to the oppidum of Tugia, characterised by a walled enclosure made from large stones on the top of a little hill. A very high density of sherds is observed on the southern side, suggesting the existence of a little habitation area. The life of the site span from the beginning of the 1st century BC to the first half of the 1st century AD. A sampling area was deployed around it, and the resulting trends suggest a continuous waste accumulation around the stream bed that crosses the survey

% with finds

% not surveyable

. . .

.

.

. .

.

3. Results

.

.

Analysis of the survey results showed strong contrasts in background noise from Protohistoric to Roman times throughout the study region (fig. 5). We statistically explored the finds probability in

Fragments by unit

Figure 5 – Boxplot of find densities in survey sample areas. D zone overlaps with part of the Tugia oppidum.

Figure 6 – Cortijo de los Castellones. A: view of the survey area from NW; the arrow indicates the location of the archaeological site. B: average sherd roundness value in the survey area.

268

Mayoral Herrera et alii, ‘Background noise’ and landscape exploitation in the Late Iron Age…

area. Here density was very low but homogeneous. Slopes and low ridges showed higher find densities and lower sherd erosion values. A limited activity area was tentatively identified in the NW corner of the survey area; perhaps a temporary dwelling for agricultural labor. No obtrusive sherd dispersions were identified in the nearby hills. This site provides a good example of how the ‘background noise’ is much more clustered in the small stretches of alluvial soils bordering tributary streams in the middle course of the Guadiana Menor river. We have also found evidence of activity in late Iberian times around some salt springs and iron mines. Lastly, Iberian sherds are almost completely absent from areas like the foot of the Cazorla massif and the Granada highlands; these were agriculturally colonized in much more recent times.

4. Concluding remarks Our survey strategy has allowed us to document abundant surface finds not related to ‘archaeological sites’. We think that useful tools have been developed for the application of a ‘distributional’ (non-site oriented) approach, but some aspects can still be improved. Firstly, while well-suited to find the more widely distributed materials, our sampling strategy runs the risk of missing more discrete sherd concentrations. Common sense therefore suggest that spaces between sampling units should be fieldwalked as well (as was done for example in Cortijo de los Castellones). Secondly, we must make use of the exponential improvement of spatial technologies in recent years. Integration of GPS and GIS is now much easier, so it is possible to create, upload and record samples very quickly and with greater accuracy (see, for example, Campana 2006). High levels of ‘background noise’ may indicate the existence of an intensively exploited agrarian landscape mainly between the late 3rd and late 1st centuries BC. It is commonly accepted, on the basis of survey data, that during the 5th and 4th centuries BC the fortified village (oppidum) remained the almost exclusive settlement type in the upper Guadalquivir area, but after the Roman conquest scattered rural settlement began in areas like the Guadiana Menor, Guadalimar, Jandulilla and Guadalbullón valleys. This last case study has been very well recorded by open-area excavations, and show an intensive agriculture system developed around small stream beds. Quadrangular fields were crossed by ponds and irrigation channels, and dotted with small rural huts and more permanent households (Ruiz Rodriguez et

Figure 7 – Castellones de Céal, a late Iberian settlement in the middle course of the Guadiana Menor valley.

al. 2007). In this context, low-density sherd distributions could be at least partly the result of manuring for horticulture, a practice with strong implications for the organization of labor in peasant communities. Nevertheless, there are great contrasts in land use throughout the Guadiana Menor valley during the Iron Age, which could be related to the emergence of first order settlements in the 5th century BC on both sides of the Cazorla and Magina massifs. These oppida had a role on the regional scale due to their strategic location in ethnic and political frontier zones (Chapa & Mayoral 1998). Settlements on the limits of their territories combined control over trade routes with the exploitation of the scarcer and more concentrated resources. A good example of this is Castellones de Ceal, located above a ford through the Guadiana Menor river, on the edge of its cultivable meadows. This site was established at the start of the 4th century BC, and abandoned in the first quarter of the 1st century BC (Mayoral 1996).

References Blánquez Pérez J. 1999. La necrópolis del Cerro de la Horca y la Cámara de Toya, in Pérez J.B. & Gómez L.R. (eds), La cultura ibérica a través de la fotografía de principios de siglo. Un homenaje a la memoria, vol. 1. Madrid. Campana S. 2006. DGPS e mobile GIS per l’archeologia dei paesaggi, in Campana S. & Francovich R. (eds), Laser scanner e GPS: paesaggi archeologici e tecnologie digitali 1, atti del workshop, 3 marzo 2005 Grosseto. Firenze, Edizioni all’Insegna del Giglio. Chapa Brunet T., Pereira Sieso J., Madrigal Belinchón A. & Mayoral Herrera V. 1998. La necropolis iberica de los castellones de Ceal (Hinojares, Jaen), Sevilla, Junta de Andalucia, Consejeria de Cultura. Empresa Publica de Gestion de programas Culturales. Universidad de Jaen. Chapa T. & Mayoral V. 1998. Explotación económica y fronteras políticas: diferencias entre el modelo ibérico y el roma-

269

Hidden Landscapes

no en el límite entre la Alta Andalucía y el Sureste. Archivo Español de Arqueología 71, pp. 63-77. Chapa T., Vicent J., Rodríguez A. & Uriarte A. 1998. Métodos y técnicas para un enfoque regional integrado en Arqueología: el proyecto sobre el poblamiento ibérico en el área del Guadiana Menor (Jaén). Arqueología Espacial 1920, pp. 105-121. Madrigal Belinchón A. 1997. El ajuar de la cámara funeraria ibérica de Toya (Peal de Becerro, Jaén). Trabajos de prehistoria 54, pp. 167-184. Mayoral V. 1996. El hábitat ibérico tardío de Castellones de Ceal. Organización del espacio y estructura socio-económica. Complutum 7, pp. 225-246. Pereira Sieso J. 1979. La cerámica ibérica procedente de Toya (Peal de becerro, Jaén), en el M.A.N. Trabajos de Prehistoria 36. Pereira Sieso J., Chapa Brunet T., Madrigal Belinchón A., Uriarte González A. & Mayoral Herrera V. 2004. La necrópolis ibérica de Galera (Granada). La colección del Museo Arqueológico Nacional, Madrid, Ministerio de Cultura. Presedo Velo F. 1982. La Necrópolis de Baza. Ministerio de Cultura, Excavaciones Arqueológicas en España 119. Ruiz Rodriguez A., Serrano Peña J.L., Molinos Molinos M. & Rodriguez Ariza M.-O. 2007. La tierra y los iberos en el Alto Guadalquivir, in Rodríguez Díaz A. & Pavón Soldevilla I. (eds), Arqueología de la tierra. VI Cursos de Verano Internacionales de la Universidad de Extremadura. Cáceres, Universidad de Extremadura.

Acknowledgements The work presented in this paper was developed in the framework of the research project ‘El Poblamiento Ibérico en el valle del Guadiana Menor. Una perspectiva desde la Arqueología del Paisaje’ (DGES-PB98/0775), led by Teresa Chapa Brunet.

270

35 The region of Marghine-Planargia in Sardinia (Italy): digital terrain modelling and spatial analysis in archaeology using GIS software. Hidden landscapes from the Late Neolithic to the Early Iron Age Gianmarco CattariI, Carlo TozziI, Marina BissonII I Department of Archaeological Science, Section of Human Palaeontology and Ethnology, University of Pisa, Via Santa Maria 53, 56100 Pisa, Italy – [email protected]; [email protected] II Istituto Nazionale di Geofisica e Vulcanologia, sezione di Pisa, via della Faggiola 32, 56100 Pisa, Italy – [email protected] Abstract Geographical Information Systems (GIS) are becoming widespread in archaeological studies ranging from inventories to spatial analyses. This study shows that these applications have been very helpful in the understanding of hidden landscapes of MarghinePlanargia, in the past one of the most populated areas of Sardinia. The region consists of mountains and broad uplands. It was periodically surveyed by different teams over 15 years and the results, published in 2000, show a very high concentration of monuments (681 identified and classified) ranging in date from the Neolithic to the Roman period. We created a digital terrain model of the area using a GIS database derived from archaeological and geographical data, enabling the production of thematic maps for each period and monument class. Distribution analyses revealed an extensive void in the central upland area during the Neolithic and Early Bronze Age, whilst the maps for Medium and Final Bronze Age/Early Iron Age showed a dense concentration of sites, suggesting a change in the exploitation of the plateau area over time. Two different groups of Neolithic tombs, associated respectively with the mountains and uplands of the eastern portion of the region, have been also identified. Preliminary conclusions show that the mountains and uplands of the Marghine-Planargia region each played a different role during the later Stone Age.

1. The project The aim of the project is to study the pre- and protohistoric landscape of the plateau/mountainous Marghine-Planargia area in Sardinia (Italy). In preparation, we have created a digital terrain model (DTM) of approximately 1800 km2 (60 by 30 km) of the region, and transferred monument data including information such as Site Name, Counties, Monument Typology, Period, Distance from rivers from the survey publication Ricerche Archeologiche nel Marghine-Planargia (Moravetti 2000) to a GIS database. At present approximately 20% of the potential 3D applications (spatial and visibility analyses) have been carried out.

2. The study area The region of Marghine-Planargia (fig. 1) is located in the mid-western area of Sardinia, covering an area of 806 km2. The western portion consists of low mountains up to approximately 200 m above the sea level, the central portion is characterised by wide continuous volcanic uplands covering an are of 13 by 10 km, with heights ranging from 300 to 700 m. Finally the eastern part consists of a mountain chain running NE-SW for a distance 19 km, rising to a height of over 1000 m. A very high concentration of monuments shows that Marghine-Planargia was one of the most pop-

ulated areas in Sardinia. A total of 681 sites have been recognised and published (Moravetti 2000): 57 subterranean tombs (domus de janas) and 34 portal tombs (dolmen) mainly belonging to the Late Neolithic (3800-2900 cal. BC)/Copper Age (2900-2100 cal. BC); 523 sites such as stone towers (Nuraghi), stone hut villages and stone tombs belonging to the Nuragic civilization (1800-580 cal. BC); 19 megalithic monuments and 17 sites that are generally considered to be Roman.

3. Archaeological analyses GIS software is an important instrument to manage and analyze spatial data for different purposes; in our case GIS tools offered by ArcView 9.0 are used to understand hidden landscapes better. Thematic maps related to site distributions are created through database querying; they represent the first

Figure 1 – The region of Marghine-Planargia in Sardinia.

271

Hidden Landscapes

Figure 2 – Distribution of Late Neolithic tombs (domus de janas).

Figure 3 – Distribution of Middle Bronze Age stone towers (protonuraghi).

Figure 6 – 3D Images of the modern landscape of Marghine-Planargia in Sardinia (source: www.sardinia3d.it. License used on 04/05/2007).

Figure 4 – Distribution of Middle-Final Bronze Age/Iron Age sites (Nuragic civilisation).

Figure 5 – Distribution of Late Neolithic portal tomb (Dolmen).

step of spatial analysis, highlighting settlement patterns in relation to periods or monument typology. Next, the maps are related to the DTM (Digital Terrain Model): the figures 2, 3, 4, 5 show the GIS archaeological database overlaid on a Digital Elevation Model with 10 m vertical resolution. Comparisons between period maps show two different settlement patterns that can be related to the various geographic conditions in the region of Marghine-Planargia. Analysis of the Late Neolithic and Middle Bronze Age maps (figs. 2-3) clearly shows a very similar settlement pattern: two main site groups, located in the western and eastern areas of the region, with a void in the central area and in both the northernmost and southernmost uplands. In contrast, during the Middle-Final Bronze Age/Iron Age (fig. 4) these areas were for over a thousand years fully occupied by a complex soci-

272

Cattari – Tozzi – Bisson, The region of Marghine-Planargia in Sardinia (Italy)…

ety, the Nuragic Civilization, which left over 400 monuments. On the map of Late Neolithic tombs (domus de janas) (fig. 2), another difference is evident in the site distribution: while the western group shows a random distribution of settlements, in the eastern group two different settlement patterns can be observed, one associated with the mountains (northern clusters) and the other with the uplands (southern clusters). The observed architectonic similarities of the subterranean Late Neolithic tombs are consistent with two alternative hypothetical interpretations for the observed site distribution: either two contemporary groups of people but each with a different economy, or a single group with changes in settlement. A different settlement pattern is detectable in the dolmen map (fig. 5), where the northernmost uplands, with their ancient paths along the natural river courses running in a NW-SE direction through the main mountain chain, seem sparsely occupied. Probably the change in settlement patterns is due to economic factors and this will be investigated in the next stage of the research.

4. Future work Studying 3D images in order to study the relationship between geomorphology-geology and monuments will be the next stage of this research project. The recent publication of a digital terrain model of Sardinia (fig. 6) by the Regione Autonoma della Sardegna, using up-to-date colour orthophotos (from 2006 onwards) and its 3D free-view access at www.sardinia3d. it, will be an important instrument in understanding the Marghine-Planargia archaeological landscape. This will allow a comparison to be made of communication routes from ancient to modern times in Sardinia and it will also improve viewshed analysis.

References Contu E. 1997. Storia della Sardegna antica e moderna. Sassari, Arti Grafiche Editoriali «Chiarella». Fisher P., Farrelly C., Maddocks A. & Ruggles C. 1997. Spatial Analysis of Visible Areas from the Bronze Age Cairns of Mull. Journal of Archaeological Science 24, London, pp. 581-592. Guidi A. & Piperno M. 1992. Italia preistorica, Bari. Kvamme K.L. 1991. GIS and archaeology, in Lock G. & Moffett J. (eds), Computer and Quantitative Methods in Archaeology 1991, BAR International Series S577, Oxford, BAR Publishing. Lilliu G. 1963. La civiltà dei Sardi dal Neolitico all’età dei Nuraghi. Torino, Eri Eds. Moravetti A. 2000. Ricerche Archeologiche nel MarghinePlanargia,Vol. I-II. Sassari, Carlo Delfino Eds.

273