Charcoals From the Past: Cultural and Palaeoenvironmental Implications: Proceedings of the Third International Meeting of Anthracology, Cavallino - Lecce (Italy), June 28th - July 1st 2004 9781407302942, 9781407333014

Citation preview

BAR S1807 2008 FIORENTINO & MAGRI (Eds) CHARCOALS FROM THE PAST

B A R

Charcoals from the Past: Cultural and Palaeoenvironmental Implications Proceedings of the Third International Meeting of Anthracology, Cavallino - Lecce (Italy), June 28th - July 1st 2004

Edited by

Girolamo Fiorentino Donatella Magri

BAR International Series 1807 2008

Charcoals from the Past: Cultural and Palaeoenvironmental Implications Proceedings of the Third International Meeting of Anthracology, Cavallino - Lecce (Italy), June 28th - July 1st 2004

Edited by

Girolamo Fiorentino Donatella Magri

BAR International Series 1807 2008

ISBN 9781407302942 paperback ISBN 9781407333014 e-format DOI https://doi.org/10.30861/9781407302942 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

Contents

Introduction G. Fiorentino, D. Magri

IV

Preface F. D’Andria

IX

Special techniques for the anatomical study of charcoal G. Angeles, F. Ortega-Escalona, C. Madero Vega Archaeo-environmental studies of cultivation terraces in the Enveig Mountain (Cerdagne) in Eastern Pyrenees (France). Use of pedo-anthracology M.C. Bal, R. Harfouche, P. Poupet, P. Campmajo, C. Rendu Did calcareous grasslands exist in prehistoric times? An archaeobotanical research on the surroundings of the prehistoric settlement above Kallmünz (Bavaria, Germany) A. Baumann, P. Poschlod

1

11

25

Charcoals in context: anthracological analysis at Muro Tenente, south-eastern Italy G.J. Burgers, D. Lentjes

39

Wood in arid zones’ prehistoric architecture I. Caneva

47

A fuoco lento: strutture di combustione nell’abitato dell’età del Bronzo di Coppa Nevigata (Manfredonia - FG) A. Cazzella, G. Recchia

53

An approach to Holocene vegetation history in the middle Rhone valley (France): anthracological data from the « TGV-Méditerranée » excavations C. Delhon, S. Thiébault

63

Environment and ritual in a late Iron Age context: an example from Raffin Fort, Co. Meath, Ireland M. Dillon, C. Newman, K. Molloy, M. O’Connell

75

Charred organic matter and phosphorus in black soils in the Lower Rhine Basin (Northwest Germany) indicate prehistoric agricultural burning E. Eckmeier, R. Gerlach, U. Tegtmeier, M.W.I. Schmidt

93

Charcoal as environmental and ethnological evidence from medieval archaeological sites in NW-Italy I. Ferrari Fontana, B.I. Menozzi, C. Montanari

105

I

Pollen and micro-charcoal evidence of vegetation dynamics and human impact along the southern Bulgarian Black sea coast M. Filipova – Marinova, H. Angelova

111

Rapid and accurate estimates of microcharcoal content in pollen slides W. Finsinger, W. Tinner, F.S. Hu

121

Arts du feu et du forgeron en Mauritanie C. Fortier

125

Metallurgy in ancient Lecce: new evidence from the excavations of Piazzetta Epulione and Piazzetta Castromediano C. Giardino, A. Quercia

129

La capanna rituale di Serra Cicora (Nardò-LE) E. Ingravallo

139

Recenti ricerche sulla produzione di utensili lignei a Karatepe-Aslantas, Turchia m.r. Iovino, c. Altinbilek

141

Experimental charcoal-burning with special regard to anthracological wood diameter analysis T. Ludemann

147

Gli accampamenti invernali e primaverili dei nomadi dell’Arkhangaï e dell’Ovorkhangaï settentrionale: i ricoveri per gli animali (Mongolia centro-occidentale) F. Lugli

159

Two long micro-charcoal records from central Italy D. Magri

167

The “fires” of Aeolian villages at the end of Middle Bronze Age: the case of Portella site in the Salina island (ME - Italy) M.C. Martinelli, G. Fiorentino

177

Combining charcoal and pollen analysis: Holocene vegetation dynamics, tree species composition and woodland use in the Bavarian Forest O. Nelle

183

Environment and agriculture in the early Neolithic of the Arene Candide (Liguria) R. Nisbet A contribution to the forest history of the Markstein area in the southern Vosges (France) W. Nölken Il controllo delle alte temperature e l’inizio della metallurgia nel Vicino Oriente A. Palmieri II

193

123

207

Food Processing in the Levant during the Middle Bronze Age. Fire installations cooking pots and grinding tools at Tell Mardikh-Ebla (Syria)-Two Case Studies L. Peyronel, G. Spreafico High resolution AMS radiocarbon dating of archaeological charcoals G. Quarta, M. D’Elia, L.Calcagnile

213

225

Environmental history in the Mediterranean basin: microcharcoal as a tool to disentangle human impact and climate change L. Sadori, M. Giardini

229

Collapsed beams and wooden remains from a 3200 BC temple and palace at Arslantepe (Malatya, Turkey) L. Sadori, F. Susanna, F. Balossi Restelli

237

The Use of Wood: Traditional Building Techniques in the Swat Valley (Pakistan). An Ethno-Archaeological Research I.E. Scerrato

251

Environmental and cultural history of South American temperate forests: an interdisciplinary approach M.E. Solari

259

A critical assessment and experimental comparison of microscopic charcoal extraction methods R. Turner, A. Kelly, N. Roberts

265

Domestic fires and vegetation cover among Neanderthalians and Anatomically Modern Human Groups (>53-30 kyr BP) in the Cantabrian Region (Cantabria, Northern Spain) P. Uzquiano

273

Fuel Supplies for Pompeii. Pre-Roman and Roman charcoals of the Casa delle Vestali R. Veal, G. Thompson

287

Anthracology and Mediterranean landscape, classical and new approaches J.L. Vernet

299

Solar influence on Holocene fire history K.J. Willis, K.D. Bennett, S.G. Haberle

307

III

Introduction The Third International Meeting of Anthracology, entitled “Charcoals from the Past. Cultural and palaeoenvironmental implications”, was organized at Cavallino (Lecce) from 28th June to 1st July 2004, four years after the previous meeting of anthracologists, held in Paris in September 2004 and chaired by S. Thiébault and J.-M. Pernaud. This tradition was started by initiative of J.-L.Vernet, who organized the first meeting in Montpellier in 1991.

the last glacial-interglacial cycle, in relation to abrupt events and to the succession of different vegetation types 3. Holocene records, focussing on the role of human activity in determining the frequency and intensity of fires, and on the possible importance of solar forcing. A special session, and a very successful one, was devoted to the identification and counting of microcharcoal particles, through the projection of sample slides under a light microscope. This open practical exercise confirmed that, although counting methods and interpretation of the data may vary among the researches, the identification criteria are very similar.

During the four days at Cavallino, over 50 oral communications and as many posters were presented by researchers from several European countries, as well as from Algeria, Australia, Canada, Chile, Mexico and New Zealand. On the whole, over 200 participants contributed in the scientific discussions in the following research fields:

Another positive aspect of the workshop was the integration of palynologists and anthracologists, who generally form two distinct scientific communities, although they share many interests in palaeoenvironmental research. Knowledge of the respective methods and interests is certainly a way to promote fruitful cooperation and effective research.

• Anthropological implications of the interaction between Man and fire • Hearths and fuel structures. Pyrotechnology and management of fire • Ethno-archaeological suggestions for wood exploitation • Charcoal from wood: new methodologies in charcoal analysis • Anthracological analysis: different regional and chronological patterns • Pedoanthracology: taphonomical and anatomical implications

The final day of the Meeting was devoted to an excursion to the Salentina Peninsula, where everybody expressed enthusiastic comments for the beauty of the landscape, the boat trip along the coast, the important archaeological sites, the swimming stop, the visit to the art town of Otranto and to the pollen site at the Alimini lakes. On the whole, all of us had the opportunity to present new data, see new results, meet new people, collect new ideas, visit new sites, and we hope to enjoy a friendly and fruitful meeting.

A subject peculiar to this meeting was the connections between archaeology and fire, through the analysis of fuel structures, in relation to both domestic and handicraft activities. Besides, special attention has been paid to the anthropological implications of Man-Fire relationships, through an ethnographical approach and ethnoarchaeological evidence. Anthracological research focused on a variety of chronological and geographical contexts, especially concerning the Mediterranean regions.

Acknowledgements We would like to thank all those involved with the organization of the Third International Meeting of Anthracology and conference sponsors, especially Università del Salento, Facoltà di Beni Culturali, Scuola di Specializzazione in Archeologia Classica e Medioevale “Dinu Adamesteanu”, Dipartimento di Beni Culturali, Regione Puglia, Provincia di Lecce, Città di Lecce, Comune di Cavallino, Ministero per i Beni e le Attività Culturali, Dottorato di Ricerca in Geomorfologia e Dinamica Ambientale dell’Università di Bari, Istituto Italiano di Preistoria e Protostoria, INQUA – Commission of Palaeoecology and Human evolution, CNR-Istituto IBAM, Collége de France – Ecole des Hautes Etudes et Science Sociales – Laboratoire d’Anthropologie, Società Botanica Italiana – Gruppi di Paleobotanica e di Palinologia, Mario Lippolis and Nikon Instruments, Quarta Caffè company, without whom the Meeting would not have been possible. The Meeting has benefited in the organizing phase from the friendly help by Rino D’Andria (Secretary of Scuola

A special workshop on “Micro-charcoal as evidence of Mediterranean fire histories” was organized within the Meeting of Anthracology”, based on the lack of any previous attempt either to assess critically or harmonise laboratory methods, or to provide any kind of regional synthesis for changes in fire frequency/intensity in the Mediterranean regions, although fire is well known to be an important element in the ecology of many Mediterranean landscapes. Participants from Italy, Switzerland, Great Britain, France, Bulgaria and New Zealand shared their research experience on three main topics of micro-charcoal and pollen research in the Mediterranean countries: 1. methodological approaches, including counting and preparation methods, and comparison with modern pollen-microcharcoal relationships 2. long microcharcoal records, aimed at the comparison between climate changes and fire incidence during IV

di Specializzazione in Archeologia Classica e Medioevale “Dinu Adamesteanu”) and Giampiero Colaianni, Milena Primavera and Francesco Solinas and all others student of the Archaeobotanical and Palaeoecological Laboratory of Lecce. We thank all the Authors who contributed in this volume. They are responsible for the English language of their texts. Each paper was reviewed by two independent referees, and we are grateful to all those that helped us in this task. We thank Valentina Caracuta and Alessandra Celant for reading through the manuscripts and all the reviewers, Giovanni Pietro Marinò for the graphic project and format of the volume. Girolamo Fiorentino Laboratorio di Archeobotanica e Paleoecologia Dipartimento di Beni Culturali Università del Salento Via D. Birago, 64 73100 Lecce, Italy Donatella Magri Dipartimento di Biologia Vegetale Sapienza Università di Roma P.le Aldo Moro, 5, 00185 Roma, Italy

V

The scientific excursion to the Salentina Peninsula

Geomorphological stop along the coast

The palaeolithic site Grotta della Zinzulusa

Visit to the art town of Otranto

VI

The organizers of III International Meeting of Anthracology Meeting (photo by N. Roberts)

J.-L. Vernet, G. Fiorentino, F. Damblon

VII

Preface The publication of this volume dedicated to Anthracology is the end point of a journey which began in the rooms and cloisters of the 17th century Dominican monastery in the town of Cavallino (near Lecce in southern Italy). Over the last few years, in this prestigious location, Salento University’s School of Specialisation in Archaeology has promoted a series of initiatives aimed at providing training for young people in the cutting-edge sectors of modern archaeology, which are characterised by their interdisciplinary nature. These initiatives include the meeting on “Human beings, plants and animals from the religious viewpoint” (June 2002), the Summer School on Greek ceramics in archaeological contexts (2001), and the International Conference on Archaeometallurgy (2006). These initiatives were rendered even more special by the presence in Cavallino of the “Museo Diffuso”, an extensive open-air workshop in an archaeological site which was once the location of an ancient city, built in the 6th century BCE by the Messapians, who inhabited the Salento Peninsula at that time. In this large area, teachers and students can participate in learning activities such as those that took place during the Meeting on Anthracology, when experiments were conducted on equipment and techniques used in ancient pyrotechnology. These top-level training courses also benefit from the support of the Bioarchaeology unit in the Dominican monastery, which includes the three Laboratories of Archaeobotany, Archaeozoology and Physical Anthropology. The Acts of the Conference reflect the dynamism of the initiatives sponsored by the University of the Salento in the archaeological sector: there have been numerous projects concerning Lecce and the surrounding area, but also studies in various other areas of the Mediterranean. In the field of AMS-radicarbon dating of organic finds, a substantial contribution has been made by the University’s CEDAD Laboratory headed by Lucio Calcagnile. The papers included in this volume reflect a broad range of interests, and concern not only anthracological analyses in archaeological contexts but also in natural environments, with reference to the many issues linked to climate change and processes of environmental transformation. In this sense, if studied with appropriate methods, charred remains can become a powerful tool for acquiring knowledge of ecosystems, and, in the long term can help to forecast future events. The actions of human beings contribute to decisive changes in the landscape, as in the case of the South of France where studies have been able to trace the transformation of the tree cover back to the Neolithic period. In this framework the landscape becomes, in the words of J.L. Vernet, “a memory complex and obstinate of human influence”. The use of fire broadens into themes of cultural anthropology, taken up by many papers in this volume, and helps us to understand social structures and forms of settlement, as Claude Lévi-Strauss taught; we should remember

however that the binary structure of “the raw and the cooked” was part of a more complex system of oppositions that included the sacred and the profane, celibacy and marriage, male and female, central and peripheral. In conclusion, I would like to extend my thanks to Girolamo Fiorentino for his help in making this initiative a reality, through his work in the Laboratory of Archaeobotany, in collaboration with the School of Specialisation in Archaeology. Thanks also to Rino D’Andria, the School’s secretary, who at all stages of the project’s organisation provided valuable support in meeting the needs of each participant. Francesco D’Andria Head of the School of Specialisation in Archaeology “Dinu Adamesteanu” University of Salento, Lecce, Italy

IX

Special techniques for the anatomical study of charcoal G. Angeles, F. Ortega-Escalona, C. Madero Vega Unidad de Recursos Forestales. Instituto de Ecología, A. C. Km. 2,5 Carretera Antigua a Coatepec No. 351. Xalapa, Veracruz 91070. Mexico E-mail: [email protected]

Abstract A series of techniques for the study of charcoal anatomy were tested in artificially charcoalified wood from different sources, as well as in a sample of archaeological charcoal. Our results proved to be effective in distinguishing detailed anatomical features, such as resin canals, simple and bordered pits, growth rings, axial and ray parenchyma, etc. The main advantages of the techniques presented here are their low cost, since no special equipment is needed to carry them on.

Introduction Charcoal anatomy has been studied mostly using fresh fractured surfaces illuminated with incident light. Since charcoal is of homogeneous color and it reflects light intensively, its surfaces are not easily observed in this way, and it is very difficult to obtain good photomicrographs from them. Because of its brittle structure, thin sections of charcoal are practically impossible to obtain with a microtome. Relatively more or less thin sections can be obtained by polishing pieces of charcoal which had previously been embedded in resin, in the same way archaeological samples are processed. However, this is a very tedious, time consuming process, which produces poor quality sections. Recently, a new technique using hidroxyethyl metacrylate as an embedding medium has been reported (Igersheima, Cichockib 1996), producing thin sections with a microtome. The whole process is said to take from 24 to 36 hours. In our laboratory we have tried several techniques to study charcoal anatomy; they are easy to reproduce and do not require specialized equipment or chemical reagents. We were able to obtain good quality transverse sections of charcoalified Pinus wood and of some archaeological samples. Radial and tangential indirect replicas can be obtained with practically any kind of material. Microcasting of vessels, tracheids and fibers of charcoalified plant material were obtained with polysiloxane mixtures. Here, we present our results obtained with charcoalified wood of Pinus taeda (Pinaceae), bamboo (Guadua aculeata, Poacecae), and some archaeological samples. We choose Pinus taeda wood for this study, because its anatomy presents very interesting features that are difficult to observe even in histological wood preparations. Thus, we took the challenge of demonstrating the benefits of our technique with this difficult sample. In our xylarium, we had several wood samples as well as microscopic slides from this species. We also tried different species of bamboo and some archaeological charcoal samples obtained from Teotihuacan, Mexico, a very important archaeological site from the fifth century. We describe the techniques used for the charcoali-

fied pine in detail; for the others samples, we describe only those steps which differed from the former process.

Material and Methods To compare the anatomy of characoalified and non charcoalified wood from the same sample, we used a specimen of Pinus taeda from our xylarium. We removed blocks of 1.5 by side from this specimen and oriented it perfectly in the transverse, radial and tangential faces. Two blocks were prepared for sectioning in a sliding microtome, following standard procedures (Berlyn, Micksche 1976; Ruzin 2000). Sections made in the three planes were stained with safranin and fast-green and mounted with synthetic resin (Ruzin 2000). Using a digital camera, we took images from each plane under a Nikon Eclipse 600 microscope with bright field or polarized light. To charcoalify wood samples, three 1.5 X 1.5 X 4 cm blocks were wrapped individually in heavy- duty aluminum foil, and placed together in a porcelain mortar. We covered the aluminum foil wraps with silica sand, and placed them in a furnace at 200 ºC for two hours. After cooling them at room temperature for 30 minutes, we removed the sand and opened the aluminum foil, to recover the charcoalified samples.

Replicas We obtained radial and tangential surfaces from charcoalified samples, by fracturing the charcoal with a sharp razor blade, aided with a hammer. To obtain direct replicas from these surfaces, equal amounts of light body and activator components of a polysiloxane paste (Speedex, A.G. Altstätten, Switzerland) were mixed up with a metal spatula on a clean glass surface. When thoroughly mixed, the paste was applied on the freshly obtained surfaces of charcoalified wood, with the spatula in order to obtain a direct impression (Fig. 1A). When the silicone paste was perfectly solidified, it was removed and examined under the dissecting microscope, with incident light, to examine the quality of the impression. [These di-

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

G. Angeles, F. Ortega-Escalona, C. Madero Vega rect replicas can also be shadow-casted with gold-palladium (or another similar conductive metal) to be observed under the SEM (Angeles 2001), if desired.] To obtain indirect replicas from the direct ones, a thin layer of transparent nail polish was applied to the surface of the silicon replica and allowed it to dry at room temperature for two to five minutes. When the nail polish was completely dry, the indirect replica was removed from the direct one very carefully, deposited on a clean microscope slide, taking care to place the replica side up (Fig. 1A), and mounted dry with a cover slip. The cover slip could be fastened to the microscope slide with a drop of nail polish, with a clear, transparent cement, or with thin strips of adhesive tape, to avoid losing the indirect impression. After this, the indirect replicas can be examined under a compound microscope, either with bright or dark field. For optimal results, double interference contrast (Nomarski optics) is suggested.

had solidified. The matrix around the sample provided an excellent support for microtoming the sample with a sliding microtome, allowing us to obtain sections as thin as 20 µm. In the case of bamboo, it was not possible to obtain thin sections in this way, but for pine charcoal it worked quite well.

Microcasting It was possible to obtain moulds of vessels, tracheids, and fibers lumina, by microcasting. A piece of charcoal, ovendried at 70 °C overnight, was embedded in a mixture of catalysed base solution A of Batson’s #17 (Polyscience, Inc. Warrington, PA., USA), consisting of a catalyser plus monomer. This solution was forced inside the empty charcoal spaces under vacuum, for 15 minutes. Then, the samples were transferred to a solution containing equal volumes of catalysed solution A, prepared as described above, and solution B, which was prepared by adding 24 drops of the promoter to 50 ml of monomer (solution A). This mixture was infiltrated into the empty spaces of charcoal under vacuum, for 30 minutes. Samples were left under vacuum over night. The next day, the block of resin, containing the charcoal at its core, was broken with a chisel and hammer, to expose the charcoal. These pieces of resin-containing-charcoal were placed in an aqueous solution of 3% KMnO4 for 24 hours, at 70 ° C in the oven, to release the moulds. The released moulds were then washed in distilled water in an ultrasonic bath, changing the water several times, to get rid of the KMnO4 and the charcoal debris. The moulds were picked up with a Pasteur pipette and deposited on a clean glass slide. The slides were dried in an oven at 70 °C for 24 hours, and sputter-coated with Au-Pd, to observe them under the SEM, at 20 Kev. In this way, it was possible to measure tracheid lumina length.

In transverse surfaces, it was difficult to obtain good indirect replicas of polysiloxane, because the paste penetrates deep into the lumina of tracheids or vessels, making it hard to detach from the charcoal surface. For transverse surfaces, two different procedures were followed. In one, long pieces of charcoalified wood are broken simply by pressing both ends with both hands and pushing up with the thumbs (Fig. 1B). The fractured segment produces a very fine transverse section, which can be observed under the dissecting microscope under incident light. However, the contrast is inadequate for photomicrography. Contrast was increased by filling the tracheary elements lumina with white type-writing corrector fluid. This was achieved by painting the freshly exposed transverse surface with the fluid, using the brush provided with it. After the fluid had dried up, another layer of fluid was applied, allowing it to dry again. The process was repeated two or more times until the lumina were totally filled up. Then the excess fluid was removed with a sharp razor blade. Examination of this surface under the dissecting microscope showed the lumina in white, while cell walls remained black (see examples in the Results section). Another technique that we tried consisted of infiltrating the charcoalified wood with tile adhesive caulk, which is composed mainly of calcium carbonate and phthalate ester (DAP Inc., Baltimore, MD, USA). We introduced individual charcoal blocks in a copper tube, perfectly sealed at the bottom (Fig. 2A). We covered the sample completely with the caulk adhesive paste, to the rim of the tube. Then, we introduced a cylindrical piece of wood, that fitted tightly on the tube; it acted as an embolus, to force the caulk paste into the tracheid lumina. In order to obtain an even pressure, we built the device shown in Fig. 2B. By pressing the nuts at the same time, an even pressure was exerted on the caulk, forcing it to enter in the lumina. Some paste was expelled through the tube rim (Fig. 2B), but the rest of it created a solid matrix around the sample (Fig. 2C) and a cast inside the tracheids. After 24 hours, the caulk paste

Results The procedures for obtaining direct and indirect replicas of charcoal surfaces are illustrated in Fig. 1. Tangential and radial surfaces are ideal for producing these kind of replicas, as illustrated in Figs 3 to 6. It is more tricky to obtain indirect replicas from transverse sections, because the polysiloxane paste tends to penetrate deeply into the lumina of tracheal elements or fibers (see Fig. 2B). If care is taken not to press very hard on the transverse surface when applying the polysiloxane, then it is possible to obtain direct polysiloxane replicas for cases in which indirect nail polish replicas could be obtained. Fracturing charcoal samples as illustrated in Fig. 1B yields perfect transverse surfaces of charcoalified wood regardless of the source. Even if the whole surface is not on the same plane, it is always possible to find a flat area wher to obtain a flat area. Painting these surfaces with liquid corrector paper, increases the contrast. All the lumina are filled with the white ink, providing a stark contrast with the black cell walls (Fig. 3A, 4C and D, 6A). Annual rings (Fig. 6A) and resin canals (Fig. 4C and D), for example, are 2

Special techniques for the anatomical study of charcoal References

clearly differentiated in this way. Tile adhesive caulk worked very well with charcoalified pine wood. Figs 5A and C are examples of the kind of sections which we obtained with this technique. However, neither charcoalified bamboo nor charcoalified oakwood was as suitable for sectioning in this way. We will try this technique in other sources of charcoal. Polysiloxane paste provided excellent direct replicas of the longitudinal surfaces of any kind of charcoalified wood. Coating them with gold-palladium was enough to observe them under the scanning electron microscope (Figs 3B and G). Another advantage of polysiloxane direct replicas is the facility with which indirect replicas can be obtained from them with nail polish (Figs 1A, 3D, 5B, E, F and 6B-F). When observed under Double Interference Contrast (Nomarski optics), indirect replicas made with nail polish provide excellent three dimensional views of longitudinal surfaces (Figs 6B-F). Microcasting of tracheal elements, fibers, and parenchyma of charcoalified wood provided high fidelity images of the lumina of these elements (Figs 7A-C, F-H). In fact, the fidelity of the microcasts was so good, that even the margo of a pit membrane was observed (Figs 7C). Micromoulds of fibers provide very good detail of the most intricate corners of fiber lumina, as illustrated in Fig. 7G. The complicated sculpture of the microcast reveals many indentations in the inner cell walls of charcoalified bamboo fibers of bamboo (Fig. 7G inset).

angeles, g., 2001 - New techniques for the anatomical study of charcoalified wood. IAWA Journal 22, 245—254.

berlyn, g. p., miksche, j. p., 1976 - Botanical Micro tech-

nique and Cytochemistry. The Iowa State. Iowa: University Press. Ames. igersheima, a., cichockib o., 1996 - A simple method for microtome sectioning of prehistoric charcoal specimens, embedded in 2- hydroxyethyl methacrylate (HEMA). Review of Palaeobotany and Palynology 92 (3-4), 389-393. ruzin, se., 1999 - Plant Microtechnique and Microscopy. New York: Oxford University Press.

Discussion We found that not every sample of charcoal can be handled in the same way. Charcoal characteristics depend on the species from which they are obtained. In our experience, charcoal from pine wood is very easy to section when it is embedded in a solid matrix, like tile adhesive caulk; however, we found charcoalified bamboo to be unsectionable, even though it was embedded in the tile adhesive. Direct or indirect replicas of tangential and radial surfaces are very easy to obtain from any charcoalified sample regardless of its source. It must be kept in mind that nail polish tends to fracture over time, so it is not possible to keep samples for too long. When stored in hermetic boxes, in a dry place, they can last several months. However, direct polysiloxane replicas can be stored indefinitely, so it is always possible to come back to the original direct replica to obtain new indirect replicas, whenever it is necessary. Fresh indirect replicas are excellent for photomicrography. Microcasting of tracheal elements and fibers of charcoalified wood provide the mean for taxonomical studies of charcoalified plant remnants. To our knowledge, this has never been tried before, as far as we know. 3

G. Angeles, F. Ortega-Escalona, C. Madero Vega

Fig. 1. Illustration of the replica technique. A. Direct and indirect replicas obtained with polysiloxane and nail polish. B. Free sample fracturing.

4

Special techniques for the anatomical study of charcoal

Fig. 2. Embedding charcoal samples with tile adhesive caulk paste. A. Two copper cylinders (lower arrow) containing charcoal samples like the one to the left (vertical arrow) and the adhesive paste. Upper horizontal arrow: wood embolus used to exert pressure on the adhesive paste. B. Two cylinders inside the wooden frame used to exert uniform pressure on the cylinders. Arrow points to the excessive adhesive paste being ejected from the copper cylinders. C. Charcoal sample embedded in a matrix of adhesive paste (arrow).

5

G. Angeles, F. Ortega-Escalona, C. Madero Vega

Fig. 3. Charcoalified bamboo (Guadua angustifolia). A. Fragment of charcoalified bamboo, lightly painted with typewriting correction fluid. Notice how the vessel walls are outlined in white (arrow). Scale = 16 mm. B. Mould of the inner part of a charcoal sample, obtained with polysiloxane paste, as seen under the Scanning Electron Microscope (SEM). Arrow points to a couple of metaxylem vessels. The very hard parenchyma cell walls of bamboo permitted their individual cells to be cast. Scale = 500 μm. C. Cross section of a bamboo culm, obtained close to the epidermis. The large bundle of sheaths are made of metaxylem vessels (arrows), protoxylem vessels, and phloem. They are embedded in parenchyma cells (arrow head). Scale = 200 μm. D. Indirect replica of bamboo, obtained from the same region as C. The profile of metaxylem vessels (arrows) and parenchyma cells (arrowhead) are very clear. Scale = 200 μm. E. Tangential view of a young culm under the SEM, showing a long metaxylem vessel (arrow), with annular thickenings of its cell wall. Scale = 200μm. F. Close up of E, showing the annular thickenings in detail. Scale = 300 μm. G. Direct replica in polysiloxane of a longitudinal surface of charcoalified bamboo, showing a metaxylem vessel. Notice the abundance of pits (arrowheads) and the union of two vessels (arrow). The vessel is in contact with fibers to the left and with parenchyma cells to the right. Scale = 100 μm.

6

Special techniques for the anatomical study of charcoal

Fig. 4. Charcoalified pine (Pinus taeda). A. Cross section of earlywood. To the left there is a large resin canal, surrounded by subsidiary cells (arrow). Scale = 50 μm. B. Indirect replica of the same area as A, showing the outlines of tracheids and resin canal. A large cavity indicates the position of the subsidiary cells (arrow), which collapsed due to the charcoalifying procedure. Scale = 25μm. C. Smooth transverse surface of charcoalified pine, previously embedded in caulking paste. The small white dots are tracheid lumina (arrows). The square delimits a resin canal. Scale = 130 μm. D. Detail of the area within the square in C. Most of the tracheids are filled with caulking paste (arrowheads). The cavity of the subsidiary cells is completely full of the paste (arrow), but not the resin canal itself. Scale = 50 μm.

7

G. Angeles, F. Ortega-Escalona, C. Madero Vega

Fig. 5. Charcoalified pine (Pinus taeda). Longitudinal views. A. Radial section of charcoalified wood, approximately 30 µm thick. Arrow points to two bordered pits. Scale = 25 μm. B. Indirect replica of a radial face of charcoalified wood, showing two alternate pits (arrow). Scale = 10 μm. C. Detail of A. Arrow points to a Sanio bar; arrowhead points to a bordered pit. Scale = 20 μm. D. Cross—field pitting characteristic of pine wood (arrow). Scale = 6 μm. E. Indirect replica of a charcoalified wood sample, in radial section. Arrow points to cross-field pitting. Scale = 10 μm. F. Indirect replica of the same area shown in D. Arrow indicates the pinoid type pitting. Scale = 10 μm.

8

Special techniques for the anatomical study of charcoal

Fig. 6. Archaeological charcoal. Mixture of species. A. Cross section of a charcoalified gymnosperm wood. Annual rings are clearly delimited (arrows). Scale = 15 mm. B. Indirect replica of a tangential surface of charcoalified angiosperm wood. Vessels (arrows) and ray parenchyma (asterisk) are clearly visible under the compound microscope. Scale = 30 μm. C. Indirect replica of a radial surface of charcoalified angiosperm wood. The body of a ray is shown at the center (arrow). Scale = 45 μm. D. Indirect replica of a tangential surface of charcoalified angiosperm wood. Vessel pitting (arrow) and inclined perforation plate (arrowhead) are observed with the compound microscope. Scale = 25 μm. E. Radial view of charcoalified angiosperm wood, observed with double interference contrast (Nomarski optics). Vessels (arrow), axial (arrowhead), and ray parenchyma (asterisk) are seen in high- relief. Scale = 50 μm. F. Another radial view, observed with Nomarski optics. A cross-field is seen at the center. Arrow points to a perforation plate. Scale = 40 μm.

9

G. Angeles, F. Ortega-Escalona, C. Madero Vega

Fig. 7. Micromoulds. A. Micromould of charcoalified pine wood. Arrow points to a pit aperture. Scale = 10 µm. B. Same as A. Upper arrow shows an open pit aperture; lower arrow points to a membrane protruding through the pit aperture. Scale = 20 µm. C. Micromould of a pit. Fine details of torus (arrowhead) and margo (arrow) can be seen. Scale = 5 µm. D. Dissociated material of bamboo culm. Relative sizes of metaxylem vessels (V), fibres (arrow), and parenchyma (p) can be observed. Vessel cell wall is covered with numerous pits. Scale = 15 µm. E. Single fiber from a bamboo culm. Notice the fiber tip (encircled). Scale = 15 µm. F. Micromould of a metaxylem vessel from a charcoalified bamboo culm. Arrow shows the numerous pits in the cell wall. Scale = 20 µm. G. Micromould of a fiber from a charcoalified bamboo culm. Notice the extent of the bifurcation. of a metaxylem vessel from a charcoalified bamboo culm. Inset: Detail of the tip of the micromould of a metaxylem vessel from a charcoalified bamboo culm. Scale = 500 µm; inset scale = 100 µm. H. Micromould of parenchyma cells from a charcoalified bamboo culm. Arrow points to a single micromould of parenchyma cell. Scale = 10 µm.

10

Archaeo-environmental studies of cultivation terraces in the Enveig Mountain (Cerdagne) in Eastern Pyrenees (France). Use of pedo-anthracology M.C. Bal1, R. Harfouche2, P. Poupet3, P. Campmajo4, C. Rendu5

CNRS UMR 5602, GEODE géographie de l’environnement – Université Toulouse le Mirail, Maison de la Recherche, 5 all A. Machado, 31058 Toulouse cedex 1, France. E-mail: [email protected] 2 CNRS UMR 5608, UTAH - Université Toulouse le Mirail, Maison de la Recherche, 5 all A. Machado, 31058 Toulouse cedex 1, France. 3 CNRS UMR 5140 - Archéologie des sociétés méditerranéennes : milieux, territoires, civilisations. Toulouse cedex 1, France. 4 Chercheur associé, Centre Anthropologie UMR 8555 – Allées J. Guesde, 31000 Toulouse, France. 1

CNRS UMR 5136, FRAMESPA - Université Toulouse le Mirail, Maison de la Recherche, 5 all A. Machado, 31058 Toulouse cedex 1, France.

5

Abstract since twenty years the Enveig mountains (Cerdagne, eastern Pyrenees, France) are subject of interdisciplinary research on past people communities and on the summering of cattle on mountain pasture as archaeological remains attest settlements since the Neolithic period. Here we present archaeo-environmental data obtained by pedo-anthracology, archaeology, soil science and palynology at two sites. Cultivation terraces dating to the Bronze Age were found at altitudes between 1700 and 2000 m a.s.l, and charcoal sampling followed archaeological and pedological structures. Processes explaining the production and preservation of charcoal in organic horizons of palaeosoils were described, thus including the evaluation of the probable functioning and evolution of cultivation terraces since the Bronze Age.

The Enveig Mountain: A workshop area about pastoral archaeology and the palaeoenvironmental reconstitution. East of the Pyrenees, Cerdagne (Fig. 1) is a large altitude plateau about equivalent to the high Sègre Valley, which runs from France towards Spain, to Seu d’Urgell. It’s surrounded with the Carlit Range (2921m) in the Northwest, and the Puigmal, in the southwest, which reaches 2910m at its highest point. The subject under study is the Enveig Mountain (Fig.2) which is located at the southern extremity of the Carlit. A Joint Researches Programme has been based upon the choice of workshop areas and sites chosen according to their possibilities to give answers to the summering of cattle on moutain pastures issue and social structuration of the cerdan moutain area (Programme Ministère de la Culture et CNRS, coordination C. Rendu (Rendu 2003) since 2003). This Mountain has the distinctive feature of being composed with projecting ledges broken with steep slopes. Old lands of the village followed by hamlets settled with ashtrees (Fraxinus excelsior), birches (Betula verrucosa) and hazel trees (Corylus avellana) stretch out as high as 1400m. Beyond 1700m, on the left shore of the Brangoli, the intermediate area provides a fine grouping of quite faded cultivation terraces which actually soften the strong verticalness of the mountainside. Some are also surrounded with a high wall: they are called Devesas del Cavaller. This easter mountainside is currently covered with a type of vegetation that is distinguished by

colonizing species in open environments, as known as Cytisus purgans, Juniperus communis and J. nana, Ilex aquifolium, Rosa canina, Salix cf pirenaica, Pinus uncinata and Pinus sylvestris, Rhododendron ferrugineum. At 1900m, at the place known as l’Orri d’en Corbill, on the opposite mountainside, the first pastoral projecting ledge with terraces, enclosures, and pastoral structures that have been the subject of archaeological excavations can be found. Still nowadays, the cattle comes for pasturing on this area, which is distinguished by an open environment composed with Pinus uncinata, Juniperus communis and J. nana, Cytisus purgans and Calluna vulgaris. Beyond 2100m, a large pastoral projecting ledge can be reached, as known as le Pla de l’Orri. Its center is settled with a peat bog, which was the subject of a palynological analysis (Galop 1998). This open environment features a few Juniperus communis and J. nana, Pinus uncinata and Cytisus purgans. A lot of cabins and pastoral enclosures can be found there, some being currently excavated. The third pastoral projecting ledge, Maurà, found at 2300m, introduces a vegetation that is characterized isolated trees (Pinus essentially). The grass covering each one of the projecting ledges are mostly composed with Festuca eskia and Nardus stricta. The whole ‘workshop area’ therefore introduces a low-density vegetation, composed with lone trees between big erroded granite blocks. The pedo-anthracological approach set up on the site started in september 2002, with the realisation

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

M.C. Bal, R. Harfouche, P. Poupet, P. Campmajo, C. Rendu

Fig. 1. Location of the site of Cerdagne.

of trenches carried out the terraces retaining structures. This first archaeo-environmental approach brought to light a carbonaceous lentil of Abies alba dated from 4120+/-30 BP (2764 (2701) 2579 CalBC), at the surface of an organic horizon located 80 cm deep in the so-called Cerd5 terrace (cultivation terrace next to Cerd6). Furthermore, takings from this excavation revealed the important abundance as regards charcoals from this organic horizon located deep in a differentiated soil. Facing these structures, it became essential to develop a charcoal takings method taking account of the archaeo-pedological organisation of the soils. A specific excavation process devoted to terraces systems studies of the Enveig Mountain was set up from may 2003. During the excavations in each of the cultivation terraces, an archaeo-pedological description allows to work out the pedo-anthracological takings protocol. During the first excavations, the choice of the trenches was made according to the localisation of the sites which were already subject to studies (archaeological sites (Rendu 2003; Rendu et al. 1996), anthracological sites (Davasse et al. 1997), peat bogs (Galop 1998)). A first prospecting session took place in september 2002 according to an altitude transect, at the Deveses del Cavaller (1700m ), then at Orri d’en Corbil (1900m). The first pedoanthracological takings were realised the same year in the pastures of Maurà, at 2300m. The following year, the first ‘terraces yard’ enabled the excavations of two terraces at Devesas del Cavaller (1700m, Cerd5bis et Cerd6), one terrace and an enclosure at l’Orri d’en Corbil (1940m, Cerd3bis et Cerd9), then one enclosure at the Pla de l’Orri (2100m, Cerd7), and one trench into a pasture area, next to the peat bog (Cerd8). During the campaign in May 2004, the mechanical shovel opened up a terrace (Cerd13) at Deveses del Cavaller, another one at l’Orri d’en Corbil (Cerd11). The last trench corresponds to the extraction of the upstream from a wall, pointed out

as an enclosure at Orri d’en Corbil (Cerd12). The only pedo-anthracological results from the Cerd6 and Cerd3bis trenches shall be introduced. Charcoals from the other trenches are currently under identification process.

Method  Pedo-anthracology has developed much since 1992 (Thinon 1992; Carcaillet and Thinon 1996; Carcaillet 1996; Carcaillet and Talon 1996; Carcaillet and Brun 2000). It consists, from the identification of the charcoals taken in the soils, in regenerating palaeo-vegetations; together with the carbon-14 dating, it makes it possible to precise when taxons were present, on scale of the mountainside. The remaining charcoals assemblings left on the soil after fire come from burnt vegetation on scale of the Valleyside (Thinon 1992). A fine mesh sifting process (0.4mm) of the sediments enables to retrieve small charcoal samples corresponding to small woody remains such as Ericaceae or Cytisus, but also to obtain valuable information about the shrub which may, depending on the species, show the low-density of the forest cover. The taking protocol The pedo-anthracological taking protocol was set up by M. Thinon (1992) and detailed by Carcaillet and Thinon (1996) as regards the altitude areas. This protocol consists in a 6 to 10kg sediment per level taking in a trench reaching the mother rock, upon a the slope of a projecting ledge. The different taking levels are determined after the description of the trench. The alteration horizons of the mother rock correspond to one or two taking levels as after sifting, they reveal charcoals (Thinon 1992; 12

Archaeo-environmental studies of cultivation terraces

Fig. 2. Enveig mountainside.

13

M.C. Bal, R. Harfouche, P. Poupet, P. Campmajo, C. Rendu

Fig. 3. Cerd6, Devesas del cavaller, 1700m, section of terrace, later trench / qualitative and quantitative anthracological results.

Talon 1997; Carcaillet 2001). This kind of taking is mostly used with mountain soils, out of the archaeological context, that is to say soils which are supposed free from a human intervention (Thinon 1992), for instance soils in the altitude of the Alps, having been under studies regarding the evolution of supra-forest areas (Talon 1997). The hereby work foregrounds the interests yielded to charcoals coming from the cultivation terraces soils, in a geo-archaeological context. The main difficulty as regards sampling is due to the fact that these structures are part of a complex soil system, which was subject to a human intervention during the setting up of the cultivation terrace (Poupet 2000). Generally speaking, the first step, as regards the construction process, was the erection of a substrate wall; then went the filling of the space that was left empty between the mountainside and the wall (Alcaraz 1999; Harfouche 2003). The soils modifications lead all the more to modifications as regards the charcoals in these soils. Therefore the matter is to determine the taking levels upon pedo-archaeological descriptions of these cultivation terraces. The open terraces revealed an ancient terrace featuring its function soil (palaeosoil) buried under the soil of the terrace currently shown on the surface. The taking protocol consists in a rigorous distinction of the different levels regarding the palaeosoils and those of the recent terrace. The takings are made from the base to the surface of the trench in order to avoid interlevel contamination as much as possible. Most generally, one to two levels are defined within the paleosoils. They correspond to the alteration of the granite mother rock, as regards the depth levels; then to a level in the organic horizon of the palaeosoil, or even two levels if this horizon is thick (Cerd6). The soil of the terrace showing on the surface, which has

been submitted to a new pedogenesis since the filling of the ancient terrace (Harfouche et al. in press), is also under study, with two to three taking levels. The boundaries for these different levels will be explained taking account of every excavated structure in the following part of presentation.

Cerd6(Fig. 3) The Cerd6 trench, located at Deveses del Cavaller (1700m) and excavated thanks to a mechanic shovel, is 2m deep, and reveals a first terrace, previously built before the terrace currently showing on the surface. This superposition shows that every single terrace has its own function soil: the buried terrace consists in a 1m thick palaeosoil (Harfouche et al., in press). Seven taking levels had to be defined in the whole trench. Level I (5 to 30cm) corresponds to a surface organo-mineral horizon introducing important roots. From 35 to 60cm, Level II is, as well as the surface horizon, distinguished by a biological activity notably testified by the existence of earthworms. Level III is delimited by a browner horizon involving angular aggregates. Levels IV and V were subject to takings in the palaeosoil organic horizon. The latter being relatively thick, it was divided into two taking levels. Levels VI and VII correspond to the granite alteration. A charcoal lentil (Abies alba) located on surface of the palaeosoil was subject to takings. It aims at the quantitative and qualitative comparison of the charcoals coming from alteration horizons, with those included in the palaeosoils organic horizons, in order to understand the terms in the way charcoals shift. 14

Archaeo-environmental studies of cultivation terraces

Fig. 4. Cerd3bis, Orri d’en Corbill, 1930m, section of terrace / qualitative and quantitative anthracological results.

as regards some pedological horizons. These observations are supported by a micro-morphological analysis led by D. Sordoillet in some pastoral structures at Pla de l’Orri (Sordoillet 2004). Some tunnels made by micro-mammals could sometimes also be noticed. The biological activity may explain some charcoal shiftings in the soils (Thinon 1992; Carcaillet and Talon, 1996; Carcaillet 2001).

Cerd3bis (Fig. 4) At l’Orri d’en Corbill (1940m height), the taking (Cerd3bis), upstream from the wall of the terrace showing on the surface, could also reveal a surimposition terraces system, each having its own function soil (Harfouche et al. in press). The previously described pedo-archaeological scheme, meaning a stacked terraces system, is reproduced in this case, with the exception that there is no carbonaceous lentil. The first taking level (level I) corresponds to the surface organo-mineral horizon and features a number of roots and micro-roots. Level II’s boundaries were set between 25 and 45cm of depth. The palaeosoil organic horizon accounts for the third taking level (level III). A last level IV was subject to takings in alteration horizons of the mother rock, located in the deepest part of the trench. The whole observed soil profiles (as well as the ones which will not be described in this article) feature a biological activity, revealed by the earthworm activities sometimes reaching a depth of 50cm. This bioturbation is equally confirmed by the granular structures of the organo-mineral horizons, or else the presence of a number of aggregates

Studies in laboratory The process consists in several steps (Thinon 1992). The water sieving is executed on takings, dried at fresh air beforehand in order to increase the impact resistance of the charcoals. The cement mixer causes a mixing movement of the sediment in the water, which enables to crumble the lumps and most of the agregates. The lightest roots and charcoals get back to the surface. These remains floating on the surface are retrieved with a 0.4 diameter sieve. The base of the steeper is poured on a column composed with four sieve (5 mm, 2 mm, 0.8 mm and 0.4 mm). The content of these sieve is cleaned with care thanks to a limp brush, in a stream. The process is to be repeated until the disappearance of the very last aggregates. All of the sieving remains are swilled, then dried before they are sorted 15

M.C. Bal, R. Harfouche, P. Poupet, P. Campmajo, C. Rendu out. The cleaning process consists in cleaning off the clayey layer covering the charcoals thanks to a ultrasonic sounds device. The charcoals are cleaned one by one in a water tank. The samples are then dried to be identified. The identification process requires a preliminary work consisting in their selection according to three linear sections (transversal, tangential longitudinal and radial longitudinal) thanks a thin-blade scalpel under the binocular microscope. They are thereafter set up in a cup containing either poppy seeds or Fontainebleau sand in order to angle them more easily. The observation is processed thanks to a reflected light microscope (100x, 200x, 500x). Identifying the charcoals leans on the referring to works dealing with the anatomy of the wood (Schweingruber 1990; Jacquiot et al. 1973; Jacquiot 1955). It also requires a comparison with the burnt wood samples of the reference collection (laboratory Pyrenean reference collection, GEODE).

or not. These lattest reflect the burnt vegetation, but they also depend on a number of breaking up and transitingrelated processes, but also to bioturbation which contributes to the splintering of the charcoals; and finally their incorporation in the soil. The assemblage realized out of charcoals coming from the palaeosoil organic horizon reflects the burnt vegetation which correspounds to the buried terrace functionning process. The anthracological pattern corresponding to the surface horizon gives useful information about how the terrace showing on the surface works. In the particular case of the Enveig Mountain, it becomes possible to assume that these patterns are submitted to a stratigraphy hierarchy related to the terraces superposition, eventhough some charcoal shiftings disrupt this stratigraphy. The anthracological results obtained on the terraces and enclosures in Cerdagne led to think taphonomically speaking. The production and preservation processes of the charcoals within these whole complex structures will be described in the following part, in order to understand the way these terraces work. Those terraces are related directly to agro-sylvo-pastoral activities in Pyrenean mountain communities and societies. It is therefore essential to take account, one the one hand, of the charcoal shiftings related to anthropogenic interventions on the soils; and on the other hand to find tangible explanations about the charcoal abundance in these palaeosoil organic horizons of the buried terraces.

The choice for datings AMS datings, as regards the charcoals, have been realised at the VERA laboratory (Austria). Generally speaking, each one of the samples sent corresponds to a single charcoal bit which has been selected, rather than a charcoal mixture, which provides an average dating and leads to a loss of information about the presence of a taxon at a given period. The state of some charcoals together with their low size doesn’t allow the choice of a single charcoal fragment. That’s the case for the Cerd6 terrace, for which a set of Abies alba charcoal fragment extracted from the carbonaceous lentil located in the palaeosoil organic horizon, was sent. These charcoal fragments were highly breakable and would reduce to powder, in contact with water. In the case of the Cerd3bis terrace, two charcoal fragments coming from the same depth level, the palaeosoil organic horizon (level III) were independantly dated (one bit of Abies alba and one bit of Pinus sylvestris/uncinata). A third bit could be dated, a fragment of Pinus sylvestris/uncinata coming from the taking level located above the palaeosoil organic horizon (level II).

Production (Fig. 5) Natural or anthropogenic fires are responsible for charcoals production. From then those may be submitted quite quickly to the action of various factors. The wind erosion and the hydric erosion are the first causes to a fast scattering of the charcoals to the ground surface after a fire (Thinon 1992; Clark 1988; Trabaud 1989; Lynch 2004). The fire action weakens the organic horizons on the soil surface which therefore become vulnerable to erosion (Choromanska 2001; Ketterings et al. 2000; Eliot and Robichaud 2001; Whelan 1995). Its effects decrease as soon as herbaceaous and lineous organisms settle on the burnt area (Thinon 1992). Bioturbation phenomena are essentially due to earthworm, enchytreids and micro-vertebrate organisms activity (Darwin 1881; Thinon 1992) which cause charcoal shiftings in soils. Those activities contibute to the burying and splintering of the charcoals (Thinon 1992; Carcaillet 1996; Carcaillet and Talon 1996; Talon 1997). The alteration horizons of the mother rock enclose charcoals which reveals how important is the burying. The roots take part in this splintering by creeping inside the charcoals anatomic structures (Thinon 1992; Carcaillet 1996). The vertical movement, essentially realized due to the soil mesofauna, contribute to the incorporation of the smallest charcoals in the round surface horizons.

Studies on the charcoals taphonomy, their growing and preservation within the terraces Pedo-anthracology aims at a reconstruction process of the composition of the burnt ligneous vegetation on a scale of the moutainside from a qualitative study of the identified taxons, but also the assemblages formed by these taxons. These assemblages depend on the ecological possibility for the various taxons to stand for vegetal groups, related to current ground knowledge and observations. The disruptions regarding the charcoal stratigraphy have been featured in a number of works (Thinon 1992; Carcaillet and Talon 1996; Carcaillet 2001) and lead to think about whether these assemblages are valid 16

Archaeo-environmental studies of cultivation terraces Transport

(Métailié 1981; Sigault 1975). This eventuality can be corroborated by the taxonomic composition, particularly the actual presence of shrubs in comparison with the low proportion in trees. Several kinds of functionning systems as regards the terrace soils, are thereafter conceivable and moreover related to permanent agriculture: a no land fallowing culture process involving manuring; practice of a long-lasting land fallowing; or else short-lasting land fallowing.

Processing a charcoal transit by human intervention is considered as fortuitous, as regards altitude soils in a supra-forest boundary (Talon 1997). This anthropogenic lack of intervention regarding the soils merge with an anthropogenic lack of charcoals shifting within the soils (due to a farming-related activity, for instance). As far as the areas highly pastured on are concerned, but within which the trenches don’t feature any anthropogenic intervention, as seen in Maurà, the surface organic horizons disruptions are only due to the cattle trempling them down. It’s going quite in another way regarding altitude terraces soils, given that the main interventions are in this case realised by men at the time the terrace is being processed. The retaining structure is directly built right on the granite mother rock, after a soil evacuation. The coals found in small amounts, currently in the alteration horizons of the mother rock, are most likely to come from fires, previous to elaboration of the buried terrace, but also probably from deforestation fires related to this elaboration. Generally speaking, the soil encloses marks of an ancient environment (Gebhart 2000), such as charcoals. The palaeosoils organic horizons of the buried terrace stands for the surface horizon which is directly related to the way this terrace works. In the current work, this is therefore this horizon which holds the palaeo-environmental informations related to the system of these terraces. In each of the excavated terraces, > 2mm and > 0.8mm sized charcoals can be found profusely in the former surface horizon, they are generally bony-shaped and do not feature any recent section made by the mixing during the sifting process. A comparison with charcoals coming from human interventions-free soils and located in the same geographical context is essential. At a height of 2400m, Maurà, the two pedo-anthracological trenches dug on a pasture soil revealed a very small amount of < 0,8mm sized coals, their dispersion being spread out within the whole profile. The size, shape, and quantity of the charcoals found in the terraces palaeosoil organic horizon, thus, may give indications about the systems of these soils, which means in the very case, the use of swidden cultivation. Indeed, the low-splintered charcoals in the soil is suggestive of this agricultural practice (Gebhardt 2000) the lattest correspounding to a culture based upon swidden cultivation, which therefore only requires a minor soilwork before sowing the seed, with no additional ploughing needed (in spanish : roza y quema. To know more about this topic and the debate concerning tropical agriculture, see also: Grenand 1996; Barrau 1996). The functionning system of these terraces may also be more complex, and is likely to involve farming practices from other times such as slashing and burning or permanent agriculture. Slashing and burning was practiced in heath and grassy grounds and required a great work as regards the superficial layer peeling of the heath and grassy soils thanks to a heavy hoe, the [ecobue]; divots were piled up in stoves to be dried out, then burnt off, the ashes being scattered on the soil

Charcoals preservation It should be notified that the charcoals extracted from the palaeosoil organic horizon benefit from an outstanding preservation due to a post-deposit fossilization of the palaeosoils. This was enabled by a colluvium process or thanks to an anthropogenic sedimentary supply in order to build the most recent terrace. That step was quickly enabled by a preservation process involving the charcoals and the carbonaceous lentil (Cerd6). The protection level can then reach its maximum point, thus preventing the charcoals from transiting and splintering in the palaeosoil. In the case of a delayed protection, the elements on the ground surface may have been submitted to the various processes described at the beginning of the subsection, meaning scattering and bioturbation, before a post-deposit fossilization. The digging of the second terrace was realized the same way as it had been done for the buried terrace. Its function soil features a differentiated pedogenesis (Harfouche et al. in press; Harfouche and Poupet, 2003). The taking level, corresponding to the surface organic horizon, reveals 0.8mm to 2mm-sized coals. They come from fires previous to the functionning of the buried terrace. This study about taphonomy, production and preservation of the charcoals enables to understand the anthracological results. It actually gives information on the buried terraces working system, and provides estimation about the functionning period of theses terraces, and finally they give valuable palaeo-environmental informations as regards the reconstruction of the burnt vegetation related to the agrosylvo-pastoral activities.

Results  The pedo-anthracological results are classified in a chart introducing the taxons and their respective anthracomass, with a descriptive scheme introducing the taking levels regarding Cerd6 and Cerd3bis trenches. This representation process enables the possibility to obtain a taxonomic presentation, and a quantitative presentation within the same scheme. The qualitative respects enable the possiblity the rebuild palaeo-vegetations out of the identified taxons and their current ecological characteristics (Thinon 1992). Identifying the taxons makes it possible to differentiate trees (e.g. Pinus) from shrubs (Cytisus purgans). The quantitative respects won’t allow a quanti17

M.C. Bal, R. Harfouche, P. Poupet, P. Campmajo, C. Rendu

Fig. 5. Studies on taphonomy, production and preservation of charcoals coming from the terraces soils

18

Archaeo-environmental studies of cultivation terraces tative interpretation of the obtained results, but reveals information about the organization of the charcoals in the soils (Carcaillet and Talon 1996). The anthracomass should be used in this case. It stands for the weight (in mg) of the extracted charcoals (diameter>0.4mm), and the whole weight (in kg) of the dried soil (fraction < 5mm). There is an actual anthracomass standing for each taxon per level (AST), and/or an anthracomass for the actual charcoals within one level (ASN), and/or an anthracomass of the actual charcoals on the whole profile (ASG). The one which can’t be determined are wrongly considered as one single taxon. This process aims at simplifying the rendering of the results.

of the buried terrace, which correspound to levels IV and V, involving 318.38 mg kg -1 and 3444.16 mg kg -1 respective ASTs. At Orri d’en Corbill (Cerd3bis; altitude: 1900m), Pinus sylvestris/uncinata and Abies alba are present on the whole trench. Ericaceae and Papilionaceae (mostly small branches) could be identified in the palaeosoil organic horizons and the surface levels. Juniperus (communis and nana) appear in level II. Corylus avellana and Salix (cf pirenaica) are found exclusively in level III. A high proportion of unidentified elements come from level III. Vitrified elements can only be found in small amounts in the deepest level. This altitude is distinguished by the fact that are a majority of Pinus sylvestris/uncinata in comparison with Abies alba on the whole trench, and most particularly in the palaeosoil organic horizon, with a 786.45 mg kg-1AST.

Quantitative profiles All of the taking levels include charcoals, from the granite mother rock to the surface horizons. A generic identification is possible if the charcoals are ≥ 0.4mm sized. The most frequently indentified taxons are two conifers: Abies alba and Pinus sylvestris/uncinata (found in 9 levels out of 9 levels subject to takings). Ericaceae (Calluna vulgaris) and Papilionaceae (type Cytisus purgans) only appear in the Cerd3bis levels at 1900m (respectively found in 2 and 3 levels out of 9). Only two taxons appear in one single level (the one included in the palaeosoil organic horizon in the Cerd3bis): Corylus avellana and Salix (cf pyrenaica). Juniperus (communis and nana) only appear in level II of the Cerd3bis trench. Charcoals that couldn’t be identified are present in the taking levels corresponding to palaeosoils organic horizons of the two trenches. These charcoals can’t be identified due to their altered structures. The vitrified ones are separated from the ones which couldn’t be identified in order to obtain an estimation of their proportion in the different levels. All of the identified taxons are currently present on the site under study, with the exception of Abies alba. The two Cerd6 and Cerd3bis trenches feature highly-rated (3926.26 mg kg-1 as for Cerd6 and 1328.79 mg kg-1 as for Cerd3bis) global anthracomasses (ASG). They also feature a very high charcoal proportion in the palaeosoils organic horizons in the two terraces. Indeed, these horizons provide over 95% of the global anthracomass. The anthracomasses per level analysis, regarding the two trenches, are not homogeneously spread due to the abundance in charcoals within the palaeosoils organic horizons. At Deveses del Cavaller (Cerd6; altitude: 1700m), Abies alba is present on the whole trench. Pinus sylvestris/uncinata is only missing in level V. The elements that can’t be identified are found in level IV. Vitrified elements are located in the deepest level, in the base of the palaeosoil organic horizon, and in level III. As far as this pedo-anthracological profile is concerned, Abies alba is the element that is the most represented, especially in the former function horizon

Composition of the assemblages In this very instance, the composition of the anthracological assemblages is processed in connection with the building and functionning mode of the terraces. Each of the open terraces features a buried terrace, and each of these terraces has its own function soil. Charcoals are extremely abundant in the palaeosoils organic horizons. Most of them are also 5mm and 2mm (and over) sized; a few of them reach a two centimeter-diameter, or more. No sign of splintering due to extraction or water-sieving (these processes may bring on cracks which can be easily spotted, as they are not covered with clay) is noticed. The sieving remains (0.8mm sifter) also include a high charcoals rate. The quantitative analysis, the size of the charcoals, their aspect and also their origin enable the consideration of anthracological assemblages in connection with agrosylvo-pastoral activities. In this instance, the assemblages are not processed according to ecological respects of the identified taxons and their ability to give birth to potential vegetal formations. Therefore should the anthropogenic dimension be included into their composition. This is the reason why making researches on taphonomy (as far as charcoals are concerned in this context) was an essential step before the constitution of assemblages in order to bring forward an interpretation. Three patterns can thus be put forward as for each one of the excavated structures. The first one can be composed with the charcoals extracted from the granite sand, the second assemblage includes the taxons which were identified in the palaeosoils organic horizons, and the third one is composed with the charcoals coming from the upper horizons of the soil as regards the current terrace shown on the surface. Therefore, the pedo-anthracological Cerd6 trench features a first pattern composed with a small proportion of Abies alba and Pinus sylvestris/uncinata identified in depth 19

M.C. Bal, R. Harfouche, P. Poupet, P. Campmajo, C. Rendu levels. The second assemblage, composed with levels IV and V, is distinguished from the other levels by its high rate of Abies alba and Pinus sylvestris/uncinata charcoals which only appear in level IV. The surface levels charcoals (III, II and I) are gathered to constitute the third assemblage which includes the same taxons as in depth levels. The first assemblage in the Cerd3bis trench is composed with the two most abundant taxons: Pinus sylvestris/uncinata and Abies alba coming from depth levels. The first assemblage in trench Cerd3bis is composed with the two most frequently identified taxons, namely Pinus sylvestris/uncinata and Abies alba coming from depth levels. The second pattern is distinguished by the high rate in charcoals extracted from the palaeosoil organic horizon, but also by the apparition of heliophilous taxons such as Corylus avellana, Salix cf pirenaica, Cytisus and Calluna vulgaris in this level. The last assemblage features Pinus sylvestris/uncinata and Abies alba coals coming from surface levels.

Interpretation  This interdisciplinary work gathering together archaeologists, pedologists, and charcoals specialists first made it possible to render the aspect of the burnt vegetation located on the mountainsides of Enveig spatialized more accurately; besides it also enabled to locate this palaeovegetation on a timeline in relation with the mechanism of the terraces, and finally to provide information about the organisation modes of these farming structures. This study thus foregrounds new items, regarding palaeo-environmental respects, but also archaeo-environmental respects about the cultivation terraces of the cerdan mountain. The anthracological results obtained as far as trenches Cerd6 and Cerdbis are concerned show that the taxons dispersion in altitude is preserved on the whole profiles. At a height of 1700m, Abies alba is dominant in comparison with Pinus sylvestris/uncinata; from 1900m on, Pinus sylvestris/uncinata becomes dominant in comparison with Abies alba, although the lattest is still highly present at this altitude. At an altitude of 2100m, at Pla de l’Orri, the fir-tree is not part of the identified taxons, for this altitude has to do more with a Pinus uncinata distribution (results of a trench/not released yet). These results render a coherent representation of the taxons organisation on the moutainsides of Enveig. Anthracological analysis of the pastoral structures hearths together with palynological results, reveal the presence of Abies alba in Enveig (Davasse et al. 1997). Pedo-anthracology supports these results and provides information more accurate, spatially speaking, about the distribution in altitude of Abies alba from the Final Neolithic until the Ancient Bronze age. Abies alba in not part of the present taxons on the studied moutainsides. As far as trench Cerd3bis is concerned, the heliophilous (Corylus avellana, Salix cf pirenaica, Cytisus and Calluna vulgaris) appear in the level correspounding to the palaeosoil organic horizon. This conveys a nonnegligible opening of the environment, for it enables the development of these taxons at Orri d’en Corbill. It is also in those horizons that unidentified and vitrified elements are prominent, as far as the two trenches are concerned.

Carbon-14 datings (Table 1) The Abies alba lentil coming from the palaeosoil surface of Cerd6 trench was dated 3590+/-25 BP (1984 (1943) 1887 CalBC). It provides an estimation of the functionning period as far as the buried terrace is concerned, at the Ancient Bronze age. The three obtained datings concerning Cerd3bis trench (see table below) enable to locate the presence of Abies alba during the Chalcolithic and Acient Bronze age eraes. The fragment of Pinus sylvestris/uncinata is dated from the Final Bronze age. Most of the obtained datings as regards the two trenches described in this article are therefore linked to the Bronze age. Considering this timescale, the charcoal lentil coming from Cerd5 trench (not described in this article) is not far from the Cerd6 lentil dating.

20

Archaeo-environmental studies of cultivation terraces The pastoral structures, dated from the Ancient Bronze age (Rendu 2000; Rendu et al. 2004) together with the numerous forest clearing phases recorded by the fire alarm demonstrate the existence of agro-sylvo-pastoral activities which have turned the forest environment into a more open environment due to successions of processes such as holding, abandonment, and re-taking up of the mountainside. During this period, Abies alba was subject to clearings at an average altitude (Galop 2000; Vannière 2001). These practices contributed on long-range to global Abies alba clearing on the mountainsides of Enveig, thus modifying the specific composition of cerdan forests. On the mountainside correspounding to Deveses del Cavaller, anthracological analysis of the palaeosoil organic horizon enable the possibility to render a vegetal landscape, mainly composed with Abies alba also including a low proportion of Pinus sylvestris/uncinata. The analysis supports the presence of an environment planted with trees, an Abies alba wood, as no ligneous charcoal could be identified. Charcoals coming from the surface levels reflects the same representation in relatively lower proportions. At Orri d’en Corbill, the analysis of the charcoals coming from level III of trench Cerd3bis reveals the existence of a pine-wood with Abies alba. This pine-wood seems to be scattered due to the presence of arbustive taxons such as Calluna vulgaris and Cytisus purgans. This vegetation profile is fairly close to the one obtained thanks to an anthracological analysis of the pastoral cabins at l’Orri d’en Corbill as for the end the Iron age, and maybe until the Antique era (Davasse et al. 1997). Between altitudes of 1600m and 1900m, the cultivation terraces present on the East mountainside of the Brangoli and at Orri d’en Corbill, are signs of former exploitations which, through their construction and functionning modes reshaped the mountainsides. The two excavated terraces revealed the same scheme: an interlocked terrace system. The abundance in charcoals of the palaeosoils organic horizons testifies the existence of a burnt palaeo-vegetation that can be temporally located in respect to the building of the terraces. In the two instances (Cerd6 and Cerd3bis), the burnt vegetation rendered out of charcoals coming from the palaeosoil, is prior to the construction of the terrace currently shown on the surface. The charcoals coming from the most superficial level constitute a palaeo-vegetation settled subsequently to the construction of the oldest terrace. Using pedo-anthracology in an archaelogical context, with the support of archaeologists and pedologists enables, in this very instance, to render vegetation settled before the construction of the terraces shown on surface. The anthracological assemblages in the levels corresponding to palaeosoils organic horizons mainly provide a precious archaeological information, for they bring forward rendering models of the agro-sylvo-pastoral activities related to the functionning mode of the terraces. The anthracological analysis of each structure has shown the abundance in charcoals in the palaeosoils organic horizons. In this geo-archaeological context, the charcoals give clues on

the functionning modes of the buried terraces. The studies led on the production and preserving of the charcoals over 5 and 2mm enable the consideration of this mechanism in connection with a complex cyclic system organised in respect with the swidden cultivation during the Bronze age. The Abies alba carbonaceous lentil (3590+/-25 BP (1984 (1943) 1887 CalBC) found at the Cerd6 palaeosoil surface positions the functionning of the buried terrace at the begining of the Ancient Bronze age. As far as the terraces at Orri d’en Corbil (Cerd3bis) are concerned, the two datings obtained out of two isolated charcoal fragments in the palaeosoil organic horizon (Abies alba in level III: 3730+/20 BP (2118 (2133) 2200 CalBC Ancient Bronze; (Pinus in level III  : 2745+/-30 BP (940 (881) 819 CalBC Final Bronze)), won’t help to make an accurate temporal location of the functionning of the buried terrace. The dating correspounding to an Abies alba fragment (4135+/-25 BP 2794 (2729) 2620 Cal BC) coming from the taking level located above the palaeosoil, can be explained by the apparition of charcoals thanks to colluvium. Willing to handle the interpretation with care, due to the important gap between these two datings, it is essential to attribute datings to several charcoal fragments coming from this horizon in order to determine a probability function making it possible to locate the functionning of the ancient terrace temporally. It is possible to render a part of the mountainsides of Enveig as cultivation terraces, whose functionning can be at least dated from the Bronze age (Cerd3bis can be more detailed), out of anthracological analysis made on Cerd3bis and Cerd6 terraces. The landscape present at this period has nothing to do with the one present nowadays. The fir-wood in Deveses del Cavaller and the pine-wood in Orri d’en Corbill are most likely to have been subject to one or more clearing phases in order to set up the first terraces. Then, the used farming techniques such as swidden cultivation, in connection with slashing and burning, and permanent agriculture, kept this environment open. The identified shrubs support this hypothesis. Sometimes abandonment phases favour the development of heliophilous which can be found in the anthracological stream, as it can be notified in Orri d’en Corbill. The datings attributed to charcoals fragments isolated in the palaeosoil organic horizon the Cerd3bis buried terrace involves a long-lasting functionning, during the Bronze age era. The distinction of the ground constituted above the palaeosoil and standing for the most recent terrace, in the two instances (Cerd6 and Cerd3bis), also demonstrates the oldness of the buried terraces. The charcoals coming from the superficial horizon of this soil renders a palaeovegetation composed with the same taxons as in depth levels: Abies alba and Pinus sylvestris/uncinata in Cerd6 ; Pinus sylvestris/uncinata, Abies alba, Cytisus purgans and Calluna vulgaris in Cerd3bis. Datings and several Abies alba fragments coming from this horizon are pending in order to know the time boundaries of this taxon on the moutainsides of Enveig. The same farming techniques (as 21

M.C. Bal, R. Harfouche, P. Poupet, P. Campmajo, C. Rendu previously described) could be used regarding the terraces shown on surface. The charcoals enclosed in this surface horizon were not ‘protected’ by some post-deposit fossilization, which may therefore explain their quantitative lack contrasting with the anthracological abundance of the palaeosoil organic horizon. A second hypothesis would tend towards a more open environment which would produce less charcoals after being set on fire.

55-58. Carcaillet, C., Thinon, M., 1996. Pedoanthracological contribution to the study of the evolution of the upper treeline in the Maurienne Valley (North French Alps): methodology and preliminary data. Review of Palaeobotany and Palynology, 91, 399 – 416. Carcaillet, C., 1996. Evolution de l’organisation spatiale des communautés végétales d’altitude depuis 7000BP dans la vallée de la Maurienne (Alpes de Savoie, France): une analyse pédoanthracologique. Thèse de doctorat en sciences. Aix Marseille III. Carcaillet, C., 2001. Soil particles reworking evidences by AMS 14 C dating of charcoal. Earth and Planetary science, 332, 21-28. Carcaillet, C., Brun, J.J., 2000. Changes in landscape structure in the northwestern Alps over the last 7000 years : lessons from soil charcoal. Journal of vegetation science, 11, 705-714. Carcaillet, C., Talon, B., 1996. Aspects taphonomiques de la stratigraphie et de la datation de charbons de bois dans les sols: exemple de quelques sols des Alpes. Géographie physique et quaternaire, 50 (2), 233-244. Choromanska, U., Deluca, T.H., 2001. Prescribed fire alters the impact of wildfire on soil biochemical properties in a ponderosa pine forest. Soil Science Society, 65, 232-238. Clark, J.S., 1988. Particle motion and the theory of charcoal analysis: source area, transport, deposition, and sampling. Quaternary Research, 30, 67 – 80. Davasse, B., Galop, D., Rendu, C., 1997. Paysages du Néolithique à nos jours dans les Pyrénées de l’Est d’après l’écologie historique et l’archéologie pastorale. In J. Burnouf, J.P. Bravard., G. Chouquer, eds. La dynamique des paysages protohistoriques, antiques, médiévaux et modernes, XVIIe Rencontre Internationale d’Archéologie et d’Histoire d’Antibes, Sophia-Antipolis : APDCA, 577-599. Darwin, C., 1881. The formation of vegetable mould through the action of worms with observations on their habits. London: John Murray and Co. Eliot, W.J., Robichaud, P.R., 2001. Comparing erosion risks from forest operations to wildfire. In P. Schiess, F. Krogstad, eds. Proceedings of the international mountain logging and 11th pacific northwest skyline symposium: 2001 – A forest engineering Odyssey. Seattle, WA: College of Forest Resources, University of Washington and IUFRO,78-89. Galop, D., 1998. La forêt, l’homme et le troupeau dans les Pyrénées, 6000 ans d’histoire de l’environnement entre Garonne et Méditerranée. Touluse: GEODE, Laboratoire d’Ecologie Terrestre, FRAMESPA. Galop, D., 2000. Les apports de la palynologie à l’histoire rurale, la longue durée des activités agro-pastorales pyrénéennes. Etudes rurales, 153-154, 127-138. Gebhart, A., 2000. Le rôle de la micro-morphologie des sols dans la formation des paysages. Etudes rurales, 153-154, 139-149. Grenand, F., 1996. L’abattis contre l’essart. Journale d’Agriculture Traditionelle et de Botanique Appliqee, XXXVIII (1), 19-53. Guilaine, J., 2000. Changeons d’échelles: pour la très longue durée, pour de larges espaces. Etudes rurales, 153-154, 9-21. Harfouche, R., Poupet, P., 2003. Construction et évolution des paysages montagnards: sols, paléosols et aménagements agricoles, sondages dans les terrasses de la montagne

Conclusion Cerdagne appears to be an ideal place to lead studies about the history of the fire. This interdisciplinary work gathering together archaeologists, pedologists, and charcoals specialists first made it possible to reveal interlocked terraces systems between altitudes of 1700 and 1900m ; then to develop a taking pedoanthracological in connection to these complex structures protocol. Eventhough it is impossible to say wether the fires have anthropogenic or natural origins - this study however reveals, out of charcoals, a farming terrace-based system between altitudes of 1600 and 1900m, during the Bronze age. In this accurate context, pedo-anthracology brings forward informations about the altitude terracing of Abies alba and of Pinus uncinata/sylvestris taking account of the agro-sylvo-pastoral practices consequences on the repartition of taxons. It also renders palaeo-landscapes on the scale of the mountainside, and provides more accurate indications, spatially speaking, in respect with palynology. The anthracological results reveal the using of swidden cultivation involving a more complex system including slashing and burning, and permanent agriculture from the beginning of the Bronze age. Studies led on taphonomy, the production and preservation processes of the charcoals made the interpretations attributed to the rendering of the palaeo-landscape and the mechanism of terraces possible and more accurate. This work is essentially connected to the construction and the mechanism of the buried terraces. The current aim is to narrow these pedo-anthracological interpretations out of other excavations realised on the same site, and to conduct the researches towards the mechanism of the terraces shown on surface. The main difficulty comes from the fact that the soil correspounding to the most recent keeps on evolving, for it is not subject to a post-deposit fossilization. That is why it is critical to bring datings to a maximum number of charcoal fragments coming from the organo-mineral horizon. .

References

Alcaraz, F., 1999. Les terrasses méditerranéennes, entre terroirs et paysages (nord-ouest du bassin méditerranéen). Thesis. Université du Mirail, Toulouse, 2 vols. Barrau, J., 1996. Vous avez dit « essart » ? Journal d’Agriculture Traditionnelle et de Botanique Appliquée, XXXVIII (1),

22

Archaeo-environmental studies of cultivation terraces d’Enveig. In C. Rendu et al., eds. PCR Cerdagne, Rapport intermédiaire 2003. Estivage et Structuration sociale d’un espace montagnard. Harfouche, R., 2003. Histoire des paysages méditerranéens au cours de la protohistoire et de l’antiquité: aménagements et agriculture. Thesis. (Doctorat). Université Aix Marseille I. Harfouche R., Poupet P., Ruas Mp., Campmajo P., Rendu C., Bal, Mc., in press. Aux marges de l’ager : forêt, pâturages et… agriculture dans la montagne pyrénéenne. VII Colloque AGER, Renne, octobre 2004. Ketterings, Q.M., Bigham, J.M., 2000. Soil color as an indicator of slash-and-burn fire severity and soil fertility in Sumatra, Indonesia. Soil Science Society, 64, 1826-1833. Jacquiot C., 1955 – Atlas d’anatomie des bois des conifères. Centre technique du bois, Paris. 2 vol. 175 p. Jacquiot, C., Trenard, Y., Dirol, D., 1973. Atlas d’anatomie des bois des Angiospermes (essences feuillues). Paris: Centre technique du bois. Lynch, J.A., Clark J.S., Stocks, B.J., 2004. Charcoal production, dispersal, and deposition from the fort providence experimental fire: interpreting fire regimes from charcoal records in boreal forests. Canadian Journal of Forest Research, 34, 1642-1656. Métailié, j.p., 1981. Le feu pastoral dans les Pyrénées Centrales (Barousse, Oueil, Larhoust). Toulouse : C.N.R.S. Poupet, P., 2000. Science du sol et archéologie, à propos d’un exemple délien. Etudes rurales, 153-154, 91-114. Rendu, C., 2000. Fouiller des cabanes de bergers: pour quoi faire ? Etudes rurales, 153-154 : 151-176. Rendu, C., 2003. La montagne d’Enveig, une estive pyrénéenne dans la longue durée. Canet : Editions Trabucaire. Rendu, C., Campmajo, P., Crabol, D., 2004. Bilan de la campagne de fouilles 2004 sur l’occupation pastorale de la montagne d’Enveig. In: Rendu C. et al., eds, Estivage et structuration sociale d’un espace montagnard: la Cerdagne. Rapport PCR. Rendu, C., Campmajo, P., Davasse, B., Galop, D., Grabol, D., 1996. Premières traces d’occupation pastorale sur la montagne d’Envei. Travaux de Préhistoire catalane, 8, 35-44. Centre d’Etudes Préhistoriques Catalanes, Université de Perpignan. Schweingruber, F.H. 1990. Anatomie europaïscher Hölzer – Anatomy of european woods. Bern, Stuttgart: Haupt. Sigaut, F., 1975. L’agriculture et le feu, rôle et place du feu dans les techniques de préparation du champ de l’ancienne agriculture européenne. Paris, La Haye : EHESS, Mouton & Co. Sordoillet, D., 2004. Approche micro-morphologique de quelques structures archéologiques. In  : C. Rendu et al., eds. Estivage et structuration sociale d’un espace montagnard : la Cerdagne. Rapport PCR. Talon, B., 1997. Evolution des zones supra-forestières des Alpes sud-occidentales françaises au cours de l’Holocène. Analyse pédoanthracologique. Thesis (3e cycle). Université d’Aix-Marseille III. Thinon, M., 1992. L’analyse pédoanthracologique, aspects méthodologiques et applications. Thesis. Université d’Aix Marseille III. Trabaud, L., 1989. Les feux de forêts – mécanismes, comportement et environnement. France-sélection. Vanniere, B., Galop, D., Rendu, C., Davasse, B., 2001. Feu et pratiques agro-pastorales dans les pyrénées orientales:

le cas de la montagne d’Enveitg (Cerdagne, Pyrénées orientales, France). Sud-ouest Européen, 11, 29-42. Whelan, R.J., 1995. The ecology of fire. Cambridge: Cambrigde University Press.

23

Did calcareous grasslands exist in prehistoric times? An archaeobotanical research on the surroundings of the prehistoric settlement above Kallmünz (Bavaria, Germany) A. Baumann, P. Poschlod

University of Regensburg, Institute for Botany, Universitätsstraße 31, D-93040 Regensburg, Germany E-mail: [email protected] Abstract A combination of a charcoal and a pollen analysis is the basis for the reconstruction of the prehistoric vegetation history of the surroundings of the prehistoric settlement of Kallmünz (Bavaria, Germany), which is dominated today by calcareous grasslands. The analysis of charcoals out of one of the greatest prehistoric settlements in Southern Germany, revealed the development of land-use from the Younger Bronze Age until the Late Iron Age (Latène period). Quercus (oak) charcoals dominate all spectra due to human preference. High values of Quercus in pollen samples from prehistoric periods are a sign for the existence of wide-spread oak forests. The existence of Juniperus, Plantago lanceolata etc. in the pollen record indicates pastoralism and habitats rich in calcareous grasslands. In the surroundings of the settlement arable fields existed. The values of Pinus (pine) are dominant in the charcoal samples as well as in the pollen record. The existence of light pine-forests on abandoned grasslands or in grazed forests is discussed. Fagus (beech) charcoals are absent in the samples of the Younger Bronze Age, but Abies (silver fir), a typical species of present beech forests, was verified. The values of beech increase in the samples from the Middle Bronze Age (3 %) to those of the Latène Age (16 %). The expansion of beech-forests as a consequence of a reduction of human impact in oak dominated forests will be debated. The increase of Fagus in the Latène Age is discussed as a result of an abandonment of settlements in the urn-pit period (Late Bronze Age). During the Bronze and Iron Age oak and pine forests in the surroundings of the prehistoric hillfort were opened by the prehistoric settlers and converted to arable fields and calcareous grasslands.

Introduction In Central Europe, calcareous grasslands naturally occur only in small-size habitats with shallow slopes, e.g. rocks or crevices (Ellenberg 1996). Nearly all calcareous grasslands have originated and developed under human landuse practices like burning, grazing by cattle and sheep and hay-making (Pott 1995; Bredenkamp et al. 2002; Poschlod and WallisDeVries 2002). However, the date of the first appearance of man-made dry grasslands is mostly unknown, the intensity of prehistoric pasturalism and interactions between grasslands and arable fields are still poorly understood. Archaeobotanical indication or reconstruction of calcareous grasslands is difficult due to bad preservation conditions for non-carbonized macroremains in dry soil conditions and the rarity of peatlands and other pollen assemblages in karst regions. Despite of these facts there are some ideas on the history of calcareous grasslands. Since Neolithic Age forest-grazing was common (Kreuz 1990; Lüning 2000; Segerström and Emanuelsson 2002), as well as tree pollarding and leaf-foddering (Haas and Rasmussen 1993; Akeret et al. 1999; Gardner 2002; Kienlin and Valde-Nowak 2002), species of grasslands could enter the opened forests (Weisel 1971; Pott 1995, 1996). Kerney et al. (1964) suggest that, according to an interpretation of a molluscan analysis, forests of the chalk escarpment near Brook (England) were cleared for Neolithic pasturalism. In the Middle Bronze Age woodland clearances, arable fields and pastures increased, but dry grasslands were restricted to more or less small scaled sites (Kerney et al. 1964; Shimwell 1976;

Thorley 1981; Dimbleby 1984; Evans 1993; Preece and Bridgland 1999; Waller and Hamilton 2000; Bouby and Billaud 2001; Lang et al. 2003; Wilkinson 2003). An analysis of preserved sodes of grassland of the cairn “Magdalenenberg” (Southern Germany) dating from the Hallstatt Age (Early Iron Age), provides a remarkable insight into plant communities of prehistoric calcareous grasslands (Fritz 1977; Körber-Grohne and Wilmanns 1977; Fritz 1979; Fritz and Wilmanns 1982; Wilmanns 1997). However, from the centre of the distribution of calcareous grasslands in Central Europe, which are the Jurassic mountains Swabian and Franconian Jura (Royer 1991; Dierschke 1997), no study has been carried out on the origin of dry grasslands so far. The study has also another background. Calcareous grasslands belong to the most species-rich habitats in Central Europe. The factors causing species richness are still poorly understood. Land-use history and the age of an ecosystem can have significant effects on its species richness (Rackham 1980; Peterken and Game 1984; Zoller and Wagner 1986; Rackham 1995; Austrheim et al. 1999; Bruun et al. 2000; Cousins and Eriksson 2002). Grasslands with different history can differ in their floristic composition (Hard 1964), contrasting Pärtel et al. (1999). From forests we know, that ancient forests are more species rich than recent forests and exhibit a certain set of species which does not occur in recent forests (Graae et al. 2004). The grasslands around the study area are extremely species-rich (Gauckler 1938; Zielonkowski

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

A. Baumann, P. Poschlod

Fig. 1. Survey area. The survey area is situated in the Franconian Alb around the confluence of the rivers Vils and Naab. 26

Did calcareous grasslands exist in prehistoric times? 1973; Sendtko 1993; Baumann et al. 2005). Therefore, one hypothesis is that in this case the origin of calcareous grasslands may date back to prehistoric times.

For the following periods, the Middle Bronze Age and the Late Bronze Age (urn-pid period) a continuous settlement could be proved (Sandner and Schauer 2003). Since there are not any remains of the Early Iron Age (Hallstatt period), it seems, that the hillfort then was abandoned during this period. Human activities in the Later Iron Age (Early Latène period) are reflected in ceramic remains (Harnest and Schauer 2000, 2002; Sandner and Schauer 2003). A system of an inner and an outer rampart constituted a man-made protection of the plateau, steep slopes towards the rivers were used as a natural protection (Sandner and Schauer 2003). Several archaeological excavations were undertaken on this prominent settlement (Reinecke 1956; Stroh 1975; Harnest and Schauer 2000, 2002; Sandner and Schauer 2003).

The area was not only chosen, due to its actual diversityrich grasslands and more or less continuous human activities in the prehistoric periods (Sandner and Schauer 2003) but also due to the availability of suitable and combinable archives for the reconstruction of the history of grasslands. In the research training group “palaeoecosystems and history” of the German Scientific Foundation (DFG) the Institutes for Prehistoric Archaeology, Physical Geography and Botany of the University of Regensburg (Germany) cooperate in the reconstruction of the prehistoric and historic land-use in the region of Regensburg. The surroundings of Kallmünz constitute a main study area of archaeological, pedological-sedimentological and archaeobotanical researches. Since 1998 two archaeological excavations took place (Harnest and Schauer 2000, 2002; Sandner and Schauer 2003). The impacts of historic and prehistoric land-use on soils were analysed by Nelle and Schmidgall (2003) and Schmidgall (2003).

Methods In the anthracological research, archaeological samples were analysed from archaeological excavation sites, which are directly situated above actual and potential prehistoric grasslands. The site, where the drilling core was taken, was located in a peatland below these grasslands in the floodplain of the river Naab. Both methods were the fundament of the reconstruction of the archaeological and palaeoecological conditions. Prehistoric human activities are reflected in the remains of archaeological sites. Charcoals are typical archaeological remains, as they derive from campfires, cookers or burned wooden constructions etc. Despite human use and preferences of several tree species, charcoal spectra might reflect palaeoecological conditions (Figueiral and Mosbrugger 2000). The standard method to reconstruct vegetation history is the pollen analysis (Fægri and Iversen 1964; Erdtman 1969).

Survey area The survey area is situated in the Jurassic Mountains of the Franconian Alb (Bavaria, southern Germany). The climate is continental. Climatic data were obtained from weather stations in Parsberg (542 a.s.l.) and Schwandorf (372 a.s.l.). Mean annual temperatures are 7.4˚C/ 7.8˚C. Annual precipitation is 649 mm (Deutscher Wetterdienst). The Jurassic bedrock with its three hills Schloßberg (433.6 a.s.l.), Hirmesberg (453.1 a.s.l.) and Kirchenberg (429.8 a.s.l.) mainly consists of Malm ε and Malm δ. Müller (1961) and Meyer and Schmidt-Kaler (1995) provide information about the geology and Figge (2002) and Schmidgall (2003) about the pedology of the survey area.

Generally, 20 – 30l soil from apparently homogenous contents of the inner and outer rampart was taken for anthracological researches. While greater charcoals were selected to avoid fragmentation, smaller charcoals were extracted by a careful wet sieving of the soil samples. From the two samples, dating in the Late Early Bronze Age (1/ 4/ 2 and 1/ 4/ 3), only 2 - 3 l soil material were analysed. The stratigraphy of the sample sites is according to Harnest and Schauer (2002) and Sandner and Schauer (2003). Charcoals were analysed with an episcopic microscope, keys from Grosser (1977) and Schweingruber (1990) and a reference collection.

The potentially natural vegetation (Tüxen 1956) of the calcareous soils are beech-forests; a Lathyro-Fagetum on fresher and a Carici-Fagetum with xerophilous and stresstolerating species on drier sites (Seibert 1968; Künne 1969; Hohenester 1989; Rosskopf 1989). Gauckler (1938) predicts Pyrolo-Pinetum or Cytiso-Pinetum communities for rocky habitats. Throughout the study region calcareous grassland are scattered on steeper slopes (Gauckler 1938; Zielonkowski 1973; Hohenester 1989; Sendtko 1993). First indicators of human activity in the area date in the Neolithic times (Sandner and Schauer 2003). A prehistoric settlement near Kallmünz, the prehistoric hillfort, situated above the confluence of the two rivers Naab and Vils on the plateau of Jurassic mountains, is one of the most important Bavarian archaeological sites (Stroh 1975). The foundation of the settlement in the Late Early Bronze Age is concluded from the dates of the archaeological remains.

Holocene sediments in the floodplain of the river Naab were investigated using a hand auger and a soil column gauge auger, which was hammered into the soil, applying a gasoline mechanical hammer. The drilling core Ka27

A. Baumann, P. Poschlod

Fig. 2. View on the prehistoric settlement and the palynological sample site. View from south-east on the village Kallmünz, the prehistoric settlement on the plateau of the Schloßberg, Kirchenberg and Hirmesberg (Bavarian Office for the Preservation of Ancient Moments, Aerial Archaeology, from 14.10.1982, author O. Braasch, archiv-nr. 6936/001, 2529-22, modified). The palynological sample site is marked in the peatland.

ence to Beug (1961); Moore et al. (1991); Reille (1995, 1998, 1999); Fægri et al. (2000); Beug (2004). The determination of Poaceae pollen followed Fægri et al. (2000). If possible, spores were determined after Moore et al.. (1991); Boros et al. (1993); Fægri et al. (2000). Macroremains were determined according to Landwehr (1966); Grosse-Brauckmann (1972, 1974); Schweingruber (1990); Frahm and Frey (1992); Hather (1993). For the identification of pollen and macro remains a reference collection was used. Micro-charcoals were recorded in the categories 10 - 20 µm and >20 µm. The pollen diagrams have been made with Tilia 2.0 and TiliaView 2.0.2. Percentages of tree pollen types were calculated on the sum of tree pollen. Trees of the local vegetation, like Alnus, Salix and Populus were excluded. The pollen sum for NAP-pollen (herbs) was calculated on tree pollen and pollen of non-aquatic species-groups. The identification of pollen assemblage zones was performed after Moore et al. (1991) and tested with numerical methods after Grimm (1987). Eight sediment samples of the drilling core and eleven charcoals from the archaeological excavations were sub-

llmünz6 for the pollen analysis was obtained from top soil (0 cm) to the sandy sediments at 3.00 m, using a Russian corer (Jowsey 1966) and stored at 4°C prior subsampling. The drilling core Kallmünz6 was embedded in an associated sedimentological and pedological research project. 1 cm3 sediment was taken every cm, dried and stored at 4° C for further investigations. The pollen preparation and concentration was obtained, using standard methods (Fægri et al. 2000) with the addition of Lycopodium tablets to enable the calculation of the concentration of pollen (Stockmarr 1971). The prepared pollen samples were embedded in glycerine jelly as mounting medium. Pollen analysis was undertaken using a Leica DM/ LM phase-contrast microscope with a magnification of 400 x. For critical identifications a magnification of 1000 x oil-immersion was used. For each sample, saturation curves were calculated for Abies, Alnus, Fagus, Juniperus, Picea, Pinus, Quercus and Salix in steps of 50 counted pollen grains. The total number of counts differed between the samples due to the saturation of the dominant tree pollen types. Generally, a minimum of 500 land pollen grains, including Cyperaceae, were counted. Pollen identification were made with refer28

Did calcareous grasslands exist in prehistoric times? Table 1 – Radiocarbon dates In total 11 pieces of charcoals from the inner and the outer rampart have been submitted to the Physical Institute from the University Erlangen (Germany) for radiocarbon dating by accelerator mass spectrometry (AMS). All radiocarbon dates are calibrated after Stuiver et al. (1998). Lab.-Nr.

part

ExcaLayer vation-Nr.

14 C-Age BP

1

4

2

Erl-5961

3474 ± 49 3752 ± 117 (95,0 %) Late Early Bronze Age

1

4

11

Erl-5967

2318 ± 48 2384 ± 86

(63,3%)

1

8

2

Erl-5965

2753 ± 51 2856 ± 95

(95,4 %) Late Bronze Age

1

12

7

Erl-5964

2636 ± 51 2789 ± 79

(90,1%)

1

13

5

Erl-5962

1981 ± 46 2044 ± 33

(92,9 %) Late Iron Age

1

17

3

Erl-5963

2293 ± 48 2215 ± 62

(49,6 %) Late Iron Age

1

19/ 21

2

Erl-5966

2237 ± 48 2248 ± 97

(93,8 %) Late Iron Age

2

23

38

Erl-5957

3061 ± 47 3292 ± 88

(82,4 %) Middle Bronze Age

2

23

32

Erl-5958

3147 ± 47 3395 ± 76

(80,2 %) Middle Bronze Age

2

23

14

Erl-5959

3165 ± 48 3397 ± 77

(87,2 %) Middle Bronze Age

2

23

33

Erl-5960

3171 ± 48 3398 ± 77

(88,7 %) Middle Bronze Age

Ram-

calib. Age BP (2ı)

Archaeological periods

Late Iron Age

Late Bronze Age

Table 1. Radiocarbon dates.

terized by an absence of Fagus and a presence of Alnus. There are no samples from the inner rampart dating from the Middle Bronze Age (B2), all contemporary samples of this period derive from the outer rampart. Half of all the samples contain charcoals only from Quercus. In samples of the outer rampart Fagus, Abies and Taxus were detected.

mitted to the Physical Institute of the University of Erlangen for radiocarbon dating by accelerator mass spectrometry (AMS). All radiocarbon dates are calibrated after Stuiver et al. (1998).

Results Charcoal analysis

Pollen analysis

Several stratigraphic layers from the archaeological excavation pits of the inner and outer rampart surrounding the prehistoric settlement above Kallmünz are rich in charcoals. The oldest charcoals are dating from the Late Early Bronze Age, the youngest from the Late Iron Age (Latène Age). The radiocarbon-dates of charcoals correlate with the dates of archaeological remains. In total, 358 charcoal pieces were analysed, 44 (12 %) mostly very small pieces could not be determined. Charcoal from 13 species or species-groups according to Schweingruber (1990), were found. In all samples, Quercus is the dominant species group. Pinus is frequent in most samples throughout all archaeologically proven periods. In both ramparts charcoals of Abies, Picea and Malus / Crataegus were present. Only in the inner rampart, charcoals from Salix, Corylus and Alnus cf. glutinosa were recorded. The anthracological analysis of all samples is reported in Fig. 3, the changes of the spectra correlating to time in Fig. 4. Oldest layers of the inner rampart (B1) are charac-

The subfossil pollen record of Kallmünz 6 is shown in Fig. 6 and fig. 7. The drilling core contains well preserved pollen throughout the profile from 280 cm to the top. There is no selective pollen destruction. A sedimentological analysis from K. Heemskerk from the Institute for Physical Geography at the University of Regensburg showed that the C-concentration in the sediment is less than 15%. Two sandy pollen-free layers interrupt the pollen record in a depth of 264 – 266 cm and 279 – 269 cm. The pollen profile reflects mainly historical periods, only the pollen zones KA 1a and KA 1b derive from prehistoric times. Six local pollen assemblage zones have been identified (KA 1- KA 6). KA 1 is defined by high values of Quercus. Pollen zone 1 is divided in subzones 1a and 1b. Only the subzone 1a is definitely prehistoric and due to this fact, only this pollen zone is discussed in this article. The values of Pinus increase in KA 2; Quercus decrease. Fagus is infrequently present. KA 3 is characterized by a decrease of Pinus, a continuous presence of Fagus and an increase of Cereal-type, Secale-type, 29

A. Baumann, P. Poschlod

Fig. 3. Charcoal diagram of the prehistoric settlement Kallmünz. The spectra of all anthracological samples are: Pre – Prehistoric remains; Iron – Remains from the Iron Age; MA – Medieval Times.

Fig. 4. Prehistoric vegetation history of the Schloßberg-Plateau of Kallmünz.

30

Did calcareous grasslands exist in prehistoric times? was covered with light oak and pine forests. Light, pine rich oak forests can be assumed on sites with a use of wood, field-grass-rotation system or grazing of cattle. Ložek (1986) describes a fundamental change of East European karst landscapes, the destruction of forests and the appearance of steppe-ecosystems, in the Early and Middle Bronze Age due to climatic and increasing human impacts. Due to the indirect and hypothetical identification of dry grasslands in the surroundings of inner rampart of the prehistoric settlement, no quantification is possible. However, dry grasslands covered the deforested and eroded slopes, which were used as pastures in the surroundings prehistoric hillfort of Kallmünz.

2000

1500

calibrated Age

1000

y = -6,58x + 1767

500

0 0

50

100

150

200

250

300

-500

-1000

depth (cm)

Middle Bronze Age

Fig. 5. Age-depth relationship. There is a significant regression between the depth of the sediment and its age (R2 = 0.94; p < 0.001). The pollen containing sediment began to grow in the Late Iron Age (calibrated Age -301 BC ± 89).

All radiocarbon dates from charcoals of different layers of the excavation pit throughout the outer rampart on the Kirchen- and Hirmesberg indicate its origin in the Middle Bronze Age (Sandner and Schauer 2003). Charcoals of Quercus are very frequent in the outer rampart: this might be due to wide-spread oak dominated forests and a man-caused overrepresentation. Five samples only contained oak charcoal. Sandner and Schauer (2003) interpret some charcoal rich layers of the rampart as a burned wooden wall structure. The heat of the burning event burned neighbouring clay, the construction material for the rampart. However, not only these layers contained charcoals, distant layers were even rich in charcoals. Fagus, Abies and Taxus, which are typical for natural Beech forests, are present in the charcoal samples of the outer rampart. The proportion of Pinus is very small. In the Middle Bronze Age the Kirchenberg and Hirmesberg were covered with light beech-rich oak forests with yew and fir on areas, less affected by man. In a sedimentological research, Schmidgall (2003) detected that the area around the outer rampart was not used for intensive agriculture. If grasslands existed around the outer rampart at the Middle Bronze Age, they were not so dominant like those in the surrounding of the inner rampart. In some European studies, dry grasslands appeared or increased in the Middle Bronze Age (Kerney et al. 1964; Thorley 1981; Dimbleby 1984; Evans 1993; Waller and Hamilton 2000; Wilkinson 2003).

Cannabis/Humulus-type and Juniperus. Values of Pinus and Juniperus increase in zone KA 4, Fagus and wild grass pollen type decline. Pinus stays at high values, Fagus is infrequently present at low percentages and all tree pollen increase in KA 5. The pollen zone KA 6 is defined by an increase of Fagus, Quercus, cereals, Secale and wild grasses.

Discussion Late Younger Bronze Age (Early Middle Bronze Age) Only the samples from the base of the inner rampart date from the Younger Bronze Age (B 1). The reconstruction of the inner rampart, however, must be dated back to the Younger Bronze Age. The spectrum of this archaeological period is characterized by an absence of Fagus-charcoals and a presence of Alnus-charcoals. Alnus was a dominant species of the prehistoric vegetation of the floodplain of the river Naab (Fig. 8). However, Fagus was present in the vegetation of Kallmünz during the Younger Bronze Age (Nelle 2002; Nelle and Schmidgall 2003). Different hypotheses can explain this fact: i) The spectrum could be regarded as non-representative, because of a too small number of analysed charcoals (N=43). ii) Large beech forests did not exist or they have been rare on the upper slopes and on the top of the plateau of the Schloßberg. Due to several prehistoric charcoal spectra of the Schloßberg, Nelle and Schmidgall (2003) discuss a prehistoric distribution of Quercus on upper and on drier slopes. They propose that Fagus and Abies grew only on fresher habitats. iii) Fagus probably was not as dominant in prehistoric forest vegetation as it is today. In several Central European pollen records, Fagus sylvatica was recorded infrequently over several centuries. This species became dominant, especially during or past prehistoric impacts and clearings (Speier and Pott 1995; Pott 2000). In this period, the Schloßberg

Late Bronze Age (Urn-pid period) The samples of the Middle and Late Urn Pid period (Late Bronze Age) (B 3 of Fig. 3) are spectra, which derive from two layers from different excavation pits of the inner rampart. The diversity of these spectra is rich because eight species (-groups) were recorded. Compared with the charcoal spectra B1 and B2, the value of Fagus increased again. Ceramic remains and the structure of the inner rampart seem to suggest that the origin of this 31

Su m

N AP

(S em

i-) Aq ua tic

s 1525 AD ± 89

Li th o AP logy

0

de st C ro ha ye rc d Tr oa ee ls R s eg Ly ion co al s po lle Zo n ne

A. Baumann, P. Poschlod

20 40

1330 AD ± 56

60 80

1300 AD ± 75 965 AD ± 64

717 AD ± 53

120

Depth (cm)

calibratred Dates (1 Sigma)

100

140 160 180 200

333 AD ± 198

220 240

340 AD ± 79 301 BC ± 89

260 280

176 391 156

266 666 382

396 1020 730

135

375

547

180

519

321

156

372

1047

412

567

940

224

333

500

273

529

609

204 277

273 386

491 320

195

293

275

226

412

261

297

453

339

203

375

321

198

445

324

139

357

314

175

463

681

97

291

397

122

305

373

127

343

281

155

321

575

191

499

538

81

205

330

124

344

476

171

477

559

178

498

673

189

473

468

312

437

450

214

402

425

160 150

430 285

660 896

217

292

360

132

343

372

140 132

380 213

326 149

147 141 92

332 315 177

256 381 415

CONISS

KA 6

KA 5

KA 4

KA 3

KA 2 KA 1b KA 1a

300 20

40

60

80 100

20

40

60

20

20

2

4

6

Total sum of squares

Glyceria-type rich sediment

Carex-type rich sediment

Sand

Fig. 6. Summary diagram of the pollen record Kallmünz 6.

rampart dates back to the Late Bronze Age. It is most likely that the area around the rampart was intensively used and treeless (Sandner and Schauer 2003). Light pine-rich forests, which were connected to pastures with dry grasslands, covered probably the surroundings of the inner rampart. The Roman-Germanic museum (Römisch-Germanisches Zentralmuseum) in Frankfurt, Germany detected in archaeobotanical samples, gained from the archaeological excavation 1956/ 57 undertaken by Armin Stroh, remains of agriculture (Triticum dicoccum, Hordeum, Lens, Panicum milleaceum), arable weeds (Galium cf. spurium and Polygonum spec.) and two charcoals of oak (Archive of the Bavarian Office for the Preservation of Ancient Moments, Regensburg). The carbonized seeds were interpreted as been related to archaeological remains of the Late Bronze Age: their origin is unknown. Unfortunately, sediment traps do not contain sediments from the Late Bronze Age (Schmidgall 2003).

Late Iron Age (Latène Age) Compared with the Bronze-Age samples out of the inner rampart, the values for Pinus decreased. This might be affected by a fallow of the settlement during the following Early Iron Age (Hallstatt period). Archaeological remains of this period are missing (Frisch 1998; Harnest and Schauer 2000; Sandner and Schauer 2003) and in sediment traps less eroded soil material was recordable (Schmidgall 2003). Abandoned agricultural areas, like pastures and arable fields, probably developed to forests, additional light forests were closed. These vegetation changes are reflected in a continuous rise of beech since its first appearance in the spectra of the Middle Bronze Age. In the charcoal spectra of the Late Iron Age, Younger and Middle Latène Age, the fraction of Fagus increased to 16%. Due to a radiocarbon date of the lowest peat layers (Erl7651), the pollen containing sediment in the floodplain 32

Fig. 7. The percentage pollen diagram Kallmünz 6.

33

300

280

260

240

220

200

180

160

140

120

100

80

60

40

20

0

Pi

20

s nu

40

60

Pi

20

a ce

40

20

40

20

20

20

20

20

20

20

40

20

Poaceae

Cultural indicators

KA 1b KA 1a

KA 2

KA 3

KA 4

KA 5

KA 6

Herbaceous species . r . e iff o iff yp s und aj are a t t nd . . s m ul f is l u e a u f l u / i e ff . o a ic n iff nd ops umucea e eae typ ndi e e ya eae ce edi av pe d u p n y a c e c l e n y c u e p -t H la e e yp s la m m u -t a i e cea -Ga b is/ hyl io id cula thu ae ca- t s e l-typ -ty ass e ia ure p od go go nu x s us a s u r e s p a e u i p i n o r e a e r e n s e l in s a o m ta ta ty a lu e la c ra t a o o n a c n a rc tu nip ilia o ry a rp bie agu a ph cer ea- e re eca ildg ac rtem e nt h en aliu lan lan olyg u m pia ste allo a nn a ry ich a nu h in o sa ero r tic ne ue T C C A F D A Z C S W A A B C C C R R R V U Q Be J u Po A C C G P P P R Zo

Trees

Did calcareous grasslands exist in prehistoric times?

Depth (cm)

A. Baumann, P. Poschlod of the river Naab began to grow in the Late Iron Age (Latène Age). The short and interrupted (hiatus) section in the core does not allow a detailed interpretation of the prehistoric section of the pollen record. The high fraction of Quercus and Pinus in the pollen record correlates with the high values of both species-groups in the charcoal samples. The abundance of Pinus and the presence of typical species of dry pastures or grazed light forests, like Juniperus, Artemisia, Plantago lanceolata, indicate the existence of dry grasslands since Iron Age. High values of Cereals and indicators of arable fields, like Centaurea cyanus, Chenopodiaceae, show the existence of arable fields (Behre 1981). The local tree vegetation of the floodplain of the river Naab was dominated by Salix and Alnus. Fritz (1977); Körber-Grohne and Wilmanns (1977); Fritz (1979); Fritz and Wilmanns (1982) showed the existence of dry grasslands during the Early Iron Age (Hallstatt Age) in Villingen-Schwenningen at the border of the Black forest.

operation was excellent. Prof. Dr. J. Völkel, Institute for Physical Geography at the University of Regensburg, was the coordinator of the research training group 462/ 2 “Palaeoecosystems and history” of the German Science Foundation (DFG). He supported this research.

References Akeret, Ö., Haas, J.N., Leuzinger, U., Jacomet, S., 1999. Plant macrofossils and pollen in goat/sheep faeces from the Neolithic lake-shore settlement Arbon Bleiche 3, Switzerland. The Holocene, 9, 175-182. Austrheim, G., Olsson, E.G.A., Grønvedt, E., 1999. Landuse impact on plant communities in semi-natural grasslands of Budalen, central Norway. Biological Conservation, 87, 369-379. Baumann, A., Blattner, S., Poschlod, P., 2005. Die neuzeitliche Geschichte der Kalkmagerrasen in der Umgebung von Kallmünz, Mittlere Frankenalb, Bayern. Hoppea, Denkschrift der Regensburger Botanischen Gesellschaft, 66, 469-488. Behre, K.E., 1981. The interpretation of anthropogenic indicators in pollen diagrams. Pollen et spores, 23, 225-245. Beug, H.J., 1961. Leitfaden der Pollenbestimmung für Mitteleuropa und angrenzende Gebiete. Stuttgart: Gustav Fischer. Beug, H.J., 2004. Leitfaden der Pollenbestimmung für Mitteleuropa und angrenzende Gebiete. München: Verlag Dr. Pfeil. Boros, Á., Járai-Komlódi, M., Tóth, Z., Nilsson, S., 1993. An Atlas of recent European Bryophyte Spores. Budapest: Svcientia Publishing Budapest. Bouby, L., Billaud Y., 2001. Late Bronze age agrarian economy in the border of the Bourget Lake (Savoy, France). Comptes Rendus de l’Académie des Sciences - Series IIA - Earth and Planetary Science, 333, 749-756. Bredenkamp, G.J., Spada, F., Kazmierczak, E., 2002. On the origin of northern and southern hemisphere grasslands. Plant Ecology, 163, 209-229. Bruun, H.H., Fritzboger, B., Rindel, P.O., Hansen, U.L., 2001. Plant species richness in grasslands: the relative importance of contemporary environment and land-use history since the Iron Age. Ecography, 24, 569-578. Cousins, S.A.O., Eriksson, O., 2002. The influence of management history and habitat on plant species richness in a rural hemiboreal landscape, Sweden. Landscape Ecology, 17, 517-529. Dierschke, H., 1997 Pflanzensoziologisch-synchorologische Stellung des Xerothermgraslandes (Festuco-Brometea) in Mitteleuropa. Phytocoenologia, 27, 127-140. Dimbleby, G.W., 1984. Anthropogenic changes from Neolithic through Medieval times. New Phytologist, 98, 57-72. Ellenberg, H., 1996. Die Vegetation Mitteleuropas mit den Alpen in ökologischer, dynamischer und historischer Sicht. Stuttgart: Ulmer. Erdtman, G., 1969. Handbook of Palynology - An Introduction to the Study of Pollen Grains and Spores. Copenhagen: Munksgaard. Evans, J.G., 1993. The influence of human communities on the English chalklands from the Mesolithic to the Iron Age: the molluscan evidence. In: F. M. Chambers, ed. Climate Change and Human Impact on the landscape. London:

Conclusions Palaeoecological data from one of the greatest prehistoric hillfort in Southern Germany, situated in the Franconian Alb, which were gained in archaeological excavations and a neighbouring pollen record, show the existence of dry grasslands from the Later Younger Bronze Age to the Late Iron Age and prehistoric forest developments. First archaeological remains and charcoal spectra derive from the Later Younger Bronze Age. This first spectrum is characterized by high values of Pinus and Quercus and the absence of Fagus. Light oak and pine forests covered the surroundings of the prehistoric settlement. The values of Pinus und Quercus stay high in all spectra. Pine forests are discussed as connected to dry pastures, covered with calcareous grasslands. Fagus increases continuously since the first appearance in the Middle Bronze Age. Several possibilities for this development of beech are discussed. Possibly, oak forests were the natural forest vegetation of the calcareous sites and were shifted to beech forests after prehistoric human impacts. The oldest part of a pollen record date in the Late Iron Age. This pollen subzone (1a) is characterized by high values of Pinus and Quercus. These results are similar to the results of the anthracological research. Cereals, Centaurea cyanus and Chenopodiaceae indicate the presence of arable fields and Juniperus, Galium and Plantago lanceolata derive from dry grassland rich habitats. Dry grasslands were present in prehistoric times, during Bronze and Iron Age, like in other European regions. The results are similar to other fundamental landscape-change in Karst regions during these prehistoric periods.

Acknowledgement

R. Sander, Institute for Prehistoric Archaeology at the University of Regensburg, executed the archaeological excavations. The co-

34

Did calcareous grasslands exist in prehistoric times? und Laubfutterwirtschaft in der Schweiz - eine alte Landwirtschaftspraxis kurz vor dem Aussterben. Dissertationes Botanicae, 196, 469-489. Hard, G., 1964. Kalktriften zwischen Westrich und Trockenund Halbtrockenrasen, Trockenwäldern und Trockengebüschen. Annales Universitatis Saraviensis. Reihe: Philosophische Fakultät A. Langen. Heidelberg, Carl Winter Universitätsverlag. Harnest, M., Schauer, P., 2000. Der Schloßberg oberhalb von Kallmünz, Oberpfalz - Vorbericht über die archäologischen und bodenkundlichen Untersuchungsergebnisse des Jahres 1999. Archäologisches Korrespondenzblatt, 30, 513-525. Harnest, M., Schauer, P., 2002. Der Schlossberg oberhalb Kallmünz, Oberpfalz. Vorbericht über die archäologischen Untersuchungsergebnisse von 1999 2000. Archäologisches Korrespondenzblatt 32, 401-407. Hather, J.G., 1993. An Archaeobotanical Guide to Root and Tuber Identification. Volume 1. Europe and South West Asia. Oxbow Monograph 28. Oxford: Oxbow Books. Hohenester, A., 1989. Zur Flora und Vegetation der Fränkischen Alb. In: F. Tichy, R. Gömmel, eds. Fränkische Alb. Neustadt a.d. A., Schriften des Zentralinstituts für fränkische Landeskunde und allgemeine Regionalforschung an der Universität Erlangen-Nürnberg. Germany, 77-94. Jowsey, P.C., 1966. An improved peat sampler. The New Phytologist, 65, 245-249. Kerney, M.P., Brown, E.H., Chandler, T.J., 1964. The lateglacial and post-glacial history of the chalk escarpment near brook, Kent. Philosophical Transactions of the Royal Society of London, Series B, 248, 135-204. Kienlin, T.L., Valde-Nowak, P., 2002. Neolithic transhumance in the Black Forest Mountains, SW Germany. Journal of Field Archaeology, 29, 29-44. Körber-Grohne, U., Wilmanns O., 1977. Eine Vegetation aus dem hallstattzeitlichen Fürstengrabhügel Magdalenenberg bei Villingen - Folgerungen aus dem pflanzlichen Befund. In: K. Spindler, ed. Magdalenenberg V. VillingenSchwenninge: Neckar- Verlag, 51-68. Kreuz, A., 1990. Die ersten Bauern Mitteleuropas. Eine archäobotanische Untersuchung zu Umwelt und Landwirtschaft der ältesten Bandkeramik. Analecta Praehist. Leidensia, 13, 1-256. Künne, H., 1969. Laubwaldgesell-schaften der Frankenalb. Dissertationes Botanicae, 2, 1-177. Landwehr, J., 1966. Atlas van de Nederlandse Bladmossen. Koniklijke: Nederlandse Natuurhistorische Vereniging. Lang, A., Niller, H.P., Rind, M.M., 2003. Land Degradation in Bronze Age Germany: Archaeological, Pedological and Chromometrical Evidence from a Hilltop Settlement on the Frauenberg. Niederbayern. Geoarchaeology, 18, 757778. Ložek, V., 1986. Anthropogene Umweltveränderungen während der Urnenfelder-Bronzezeit, dargestellt am Beispiel mitteleuropäischer Karstlandschaften. Veröffentlichungen des Museums für Ur- und Frühgeschichte Potsdam, 20, 133-136. Lüning, J., 2000. Steinzeitliche Bauern in Deutschland - die Landwirtschaft im Neolithikum. Seminar für Vor- und Frühgeschichte der Universität Frankfurt, Band 58. Bonn: Rudolf Habelt Verlag Meyer, R.K.F., Schmidt-Kaler, H., 1995. Rund um Regensburg. Wanderungen in die Erdgeschichte, 7. München: Verlag

Chapman & Hall, 147-156. Fægri, K., Iversen, J., 1964. Textbook of Pollen Analysis. Copenhagen: Munksgaard. Fægri, K., Iversen, J., K., Kaland, P.E., Krzywinski, K., (eds.) 2000. Textbook of Pollen Analysis. (reprint of IV edn.) New York: Wiley and Sons. Figge, B., 2002. Geomorphologische Reliefanalyse und sedimentologisch-pedologische Untersuchungen zu anthropogenen Veränderungen der Landschaft auf dem Schloßberg bei Kallmünz/ Oberpfalz. Institute for Geography, University of Regensburg. unpubl. MasterThesis. Figueiral, I., Mosbrugger, V., 2000. A review of charcoal analysis as a tool for assessing Quaternary and Tertiary environments: achievements and limits. Palaeogeography Palaeoclimatology Palaeoecology, 164, 397-407. Frahm, J.P., Frey, W., 1992. Moosflora. Stuttgart: Ulmer. Frisch, A., 1998. Der Schloßberg bei Kallmünz - eine vorgeschichtliche Höhenbefestigung der Bronze- bis Latènezeit. Beiträge zur Achäologie der Oberpfalz, 2, 285-310. Fritz, W., 1977. Zur ökologischen Interpretation einer Moosvegetation aus dem Grabhügel Magdalenen-berg. In: K. Spindler, ed. Magdalenenberg V. VillingenSchwenninge: Neckar- Verlag, 21-34. Fritz, W., 1979. Ein hallstattzeitlicher Festuca-Brometea-Rasen. In: O. Wilmanns, R. Tüxen, eds. Werden und Vergehen von Pflanzengesellschaften. Berichte der Internationalen Vereinigung für Vegetationskunde, Rinteln 1978. Vaduz, 165-180. Fritz, W., Wilmanns, O., 1982. Die Aussagekraft subfossiler Moos-Synusien bei der Rekonstruktion eines keltischen Lebensraumes - Das Beispiel des Fürstengrabhügels Magdalenenberg bei Villingen. Berichte der Deutschen Botanischen Gesellschaft, 95, 1-18. Gardner, A.R., 2002. Neolithic to Copper Age woodland impacts in northeast Hungary? Evidence from the pollen and sediment chemistry records. The Holocene, 12, 541-553. Gauckler, K., 1938. Steppenheide und Steppenheidewald der Fränkischen Alb in pflanzensoziologischer, ökologischer und geographischer Sicht. Berichte der Bayerischen Botanische Gesellschaft zur Erforschung der heimischen Flora, 23, 1-134. Graae, B.J., Okland, R.H.P., Petersen, P.M., Jensen, K., Fritzboger, B., 2004. Influence of historical, geographical understorey composition and and environmental variables on richness in Danish forests. Journal of Vegetation Science, 15, 465-474. Grimm, E.C., 1987. CONISS: A Fortran 77 program for stratigraphically constrained cluster analysis by the method of incremental sum of squares. Computery & Geosciences, 13, 13-35. Grosse-Brauckmann, G., 1972. Über pflanzliche Makrofossilien mitteleuropäischer Torfe. I. Gewebereste krautiger Pflanzen und ihre Bestimmung. Telma, 2, 19-56. Grosse-Brauckmann, G., 1974. Über pflanzliche Makrofossilien mitteleuropäischer Torfe. II. Weitere Reste (Früchte, Samen, Moose u.a. und ihre Bestimmungsmöglichkeiten). Telma, 4, 51-118. Grosser, D., 1977. Die Hölzer Mitteleuropas - ein mikrophotographischer Lehratlas. Berlin, New York, Heidelberg, Tokyo: Springer. Haas, J.N., Rasmussen, P., 1993. Zur Geschichte der Schneitel-

35

A. Baumann P. Poschlod Kallmünz - Vorbericht über die archäologischen Ergebnisse 2002 - 2003. Archäologisches Korrespondenzblatt, 33, 505-524. Schmidgall, J., 2003. Bodenkundlich-sedimentologische Untersuchungen zum anthropogen induzierten Landschaftswandel von Karstgebieten am Beispiel des Schloßbergs bei Kallmünz (südöstliche Fränkische Alb). Thesis (PhD) Institute for Geography, University of Regensburg. Schweingruber, F.H., 1990. Microscopic Wood Anatomy. Birmensdorf: Eidgenössische Forschungsanstalt für Wald, Schnee und Landschaft. Segerström, U., Emanuelsson, M., 2002. Extensive forest grazing and hay-making on mires - vegetation changes in south-central Sweden due to land use since Medieval times. Vegetation History and Archaeobotany, 11, 181190. Seibert, P., 1968. Vegetation und Landschaft in Bayern. Erläuterungen zur Übersichtskarte der natürlichen Vegetationsgebiete von Bayern. Erdkunde, 22, 294-313. Sendtko, A., 1993. Die Flora und Vegetation der Kalkmagerrasen am Schloßberg und Hutberg bei Kallmünz (Landkreis Regensburg). Hoppea, Denkschrift der Regensburger Botanischen Gesellschaft, 54, 393-454. Shimwell, D., 1976. Festuco-Brometea Br.-Bl. and R. Tx 1943 in the British isles: the phytogeography and phytosociology of limestone grasslands. I. General introduction; Xerobromion in England. Vegetatio, 23, 1-28. Speier, M., Pott, R., 1995. Paläobotanische Untersuchungen zur Entwicklung prähistorischer und historischer Waldfeldbausysteme im Lahn-Dill-Bergland. In: B. Pinsker, ed. Eisenland – Zu den Wurzeln der nassauischen Eisenindustrie. (Museums Katalog). Wiesbaden, 235256. Stockmarr, J., 1971. Tablets with spores used in absolute pollen analysis. Pollen et spores, 13, 615-621. Stroh, A., 1975. Die Vor- und Frühgeschichtlichen Geländedenkmäler der Oberpfalz. In: K. Schwarz, ed. Materialhefte zur Bayerischen Vorgeschichte. Reihe B Inventare Geländedenkmäler 3. Kallmünz. 260-261. Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S., Hughen, K.A., Kromer, B., McCormac, G., Van der Pflicht, J., Spurk, M., 1998. INTCAL98 radiocarbon age calibration, 24.000-0 cal BP. Radiocarbon, 40, 10411083. Thorley, A.J., 1981. Pollen analytical evidence relating to the vegetational history of the chalk. Journal of Biogeography, 8, 93-106. Tüxen, R., 1956. Die heutige potentiell natürliche Vegetation als Gegenstand der Vegetationskartierung. Angewandte Pflanzensoziologie, 13, 5-42. Waller, M.P., Hamilton, S., 2000. Vegetation history of the English chalklands: a mid-Holocene pollen sequence from the Caburn, East Sussex. Journal of Quaternary Science, 15, 253-272. Weisel, H., 1971. Die Bewaldung der nördlichen Frankenalb Ihre Veränderungen seit der Mitte des 19. Jahrhunderts. Erlanger Geographische Arbeiten, 28, 1- 72. Wilkinson, K.N., 2003. Colluvial deposits in dry valleys of southern England as proxy indicators of palaeoenvironmental and land-use change. Geoarchaeology, 18, 725-755. Wilmanns, O., 1997. Zur Geschichte der mitteleuropäischen Trockenrasen seit dem Spätglazial – Methoden, Tatsachen,

Dr. Pfeil. Moore, D.P., Webb, J.A., Collinson, M.E., 1991. Pollen Analysis. Oxford: Blackwell Scientific Publications. Müller, M., 1961. Zur Entwicklung von Malm und Kreide im Raum Parsberg-Kallmünz. Erlanger Geologische Abhandlungen , 40, 1-48. Nelle, O., 2002. Zur holozänen Vegetations- und Waldnutzungsgeschichte des Vorderen Bayerischen Waldes anhand von Pollen- und Holzkohleanalysen. Hoppea, Denkschrift der Regensburger Botanischen Gesellschaft, 63,1-361. Nelle, O., Schmidgall, J., 2003. Der Beitrag der Paläobotanik zur Landschaftsgeschichte von Karstgebieten am Beispiel der vorgeschichtlichen Höhensiedlung auf dem Schloßberg bei Kallmünz (Südöstliche Frankenalb). Eiszeitalter und Gegenwart, 53, 55-73. Pärtel, M., Mändla, R., Zobel M., 1999. Landscape history on a calcareous (alvar) grassland in Hanila, western Estonia, during the last three hundred years. Landscape Ecology, 14, 187-196. Peterken, G.F., Game, M., 1984. Historical factors affecting the number and distribution of vascular plant species in the woodlands of central Lincolnshire. Journal of Ecology, 72, 155-182. Poschlod, P., WallisDeVries, M.F., 2002. The historical and socioeconomic perspective of calcareous grasslands - lessons from the distant and recent past. Biological Conservation,104, 361-376. Pott, R., 1995. The origin of grassland plant species and grassland communities in Central Europe. Fitosociologia, 29, 7-32. Pott, R., 1996. Die Entwicklungsgeschichte und Verbreitung xerothermer Vegetationseinheiten in Mitteleuropa unter dem Einfluß des Menschen. Tuexenia, 16, 337-369. Pott, R., 2000. Die Entwicklung der europäischen Buchenwälder in der Nacheiszeit. Rundgespräche der Kommission für Ökologie, 18, 49-75. Preece, R.C., Bridgland, D.R., 1999. Holywell Coombe, Folkestone: A 13,000 year history of an English Chalkland Valley. Quaternary Science Reviews, 18, 1075-1125. Rackham, O., 1980. Ancient woodlands - its history, vegetation und uses in England. London: Edward Arnold. Rackham, O., 1995. The history of the countryside. London, UK : Dent. Reille, M., 1995. Pollen et Spores d´Europe et d´Afrique du nord. Supplément 1. Marseille: Laboratoire de Botanique Historique et Palynologie. Reille, M., 1998. Pollen et Spores d´Europe et d´Afrique du nord. Supplément 2. Marseille: Laboratoire de Botanique Historique et Palynologie. Reille, M., 1999. Pollen et Spores d´Europe et d´Afrique du nord. vol. I. Marseille: Laboratoire de Botanique Historique et Palynologie. Reinecke, P., 1956. Der Ringwall von Kallmünz. Die Oberpfalz , 44, 231-247. Rosskopf, M., 1989. Vegetationskundliche Untersuchungen an Laubwaldgesellschaften im Naabtal nördlich von Pielenhofen. Hoppea, Denkschrift der Regensburger Botanischen Gesellschaft, 47, 55-90. Royer, J.M., 1991. Synthèse eurosibérienne, phytosociologique et phytogéographique de la classe des Festuco-Brometea. Dissertationes Botanicae , 178, 1-296. Sandner, R., Schauer, P., 2003. Der Schlossberg oberhalb

36

Did calcareous grasslands exist in prehistoric times? Hypothesen. Phytocoenologia, 27, 213-233. Zielonkowski, W., 1973. Wildgrasfluren der Umgebung Regensburg. Hoppea, Denkschrift der Regensburger Botanischen Gesellschaft, 31, 5-181. Zoller, H., Wagner, C., 1986. Rückgang und Gefährdung und Rückgang von Mesobromium-Arten im Schweizer Jura. Veröffentlichungen des Geobotanischen Instituts der Eidgenössischen Technischen Hochschule, Stiftung Rübel, 87, 239-259.

37

Charcoals in context: anthracological analysis at Muro Tenente, south-eastern

Italy

G.J. Burgers, D. Lentjes

Archaeological Centre, Vrije Universiteit, Amsterdam E-mail: [email protected] Abstract Urbanization has long been a major theme in the archaeology of ancient Italy. In Southern Italy, however, the scholarly attention was clearly dominated by the Greek colonisation in the seventh and sixth century BC. Their neighbouring indigenous were considered of less importance. This trend led to a lack of well-documented settlement contexts in most of the regions concerned. It was not until recently that an interest for non-Greek settlement research emerged, revealing that urbanizing trends also occurred in the indigenous world. The upsurge of settlement research has been particularly intense on the Salento peninsula. One of the fortified ancient towns investigated there is known as ‘Muro Tenente’. The site is located some 18 km south-west of Brindisi and was probably inhabited from the eighth century BC until the Roman Imperial Age. From 1992 until 2002 it constituted the research object of the Archaeological Institute of the Free University of Amsterdam (AIVU). With the excavations at Muro Tenente we have wished to contribute to the theme of urbanization. By including also archaeobotanical research, we wanted to investigate also the relationship between urbanisation and agriculture, landscape and the environment in general. To that aim, archaeobotanical data were collected, in close collaboration with prof. Girolamo Fiorentino of Lecce University. Between 1998 and 2002, some 200 samples were taken from contexts all over the site. 64 of them contained charcoal. In this presentation, we wish to show the results of the determination of this charcoal and present some preliminary conclusions.

Introduction Urbanization has long been a major theme in the archaeology of ancient Italy. And yet, not long ago a survey of ancient urban sites in Sout­hern Italy demon­strated a striking lack of well-documented sett­lement contexts in most of the regions concerned. Ma­ny expla­nations can be given for this negative outco­me. Promi­nent among them is the traditional fascination of Clas­sical archaeology for a parti­ cular type of urban site, namely the Greek one. In Southern Italy ancient Greek towns are repre­sented only in very small numbers. Notwithstanding, until recently they total­ly domi­nated scholarly attention, to the neglect of investigations into the neighbouring, non-Greek indigenous regions of Southern Italy. Fortunately, this focus is gradually being coun­terba­lan­ced by an upsurge of non-Greek settlement re­search, revealing that urbanizing trends also occurred in the rest of the South. The upsurge of settlement research has been particularly intense on the Salento peninsula thanks above all to the investigations of Lecce University. The undula­ting land­scape of Salento is literally dotted with forti­fied an­cient towns, most of which are shown on Fig. 1. A series of these sites was investigated in the context of the so-called Brindisino project of the Archae­ologi­cal Centre of the Free University of Amster­dam, the ACVU (formerly AIVU); during the 1980’s and 1990’s various project teams carried out intensive field surveys in a number of samples throughout the Brindisi region, in the northwest of Salento. The field walking ranges from total coverage surveys of catchment areas of major ancient towns, like Oria and Valesio to urban surveys like those at Muro Tenente, Muro Maurizio and Li Castelli di

San Pancrazio Salentino and transect surveys crosscutting various physiographic units, as at Ostuni in the Murge area (Yntema 1993a; 1993b; Burgers 1998a; Burgers et al 2003). Moreover, in the context of the regional project large-scale excavations were carried out at a number of large fortified sites where surveys had previously been conducted, notably at the site of Muro Tenente (Burgers 1998b; Burgers and Yntema 1999; Alberda et al. 1999). The latter site was investigated from 1992 until 2002. It is located some 18 km south-west of Brin­di­si. Nowadays, its most conspicuous aspect are its fortifications which are clearly visible on aerial photographs (Fig. 2) as well as in the field; they enclose an area of 52 hectares. Considering that the site is situa­ted in open coun­tryside and that it is for­mal­ly pro­tec­ted from destruc­tive agri­cultu­ral acti­vi­ties, it is admira­bly suited for large-scale excava­ti­on aiming at gaining insight into its chrono­logical, functional and spatial charac­te­ristics. With these excavations we have wished to contribute to the theme of urbanization that has been discussed above. By including also paleo-ecological research we wished to investigate also the relationship between urbanisation and agriculture, landscape and the environment in general. To that aim, archaeobotanical data were collected and investigated. The present article offers a preliminary discussion of the charcoal samples taken from the Muro Tenente digs, relating them to spatial and economic contexts highlighted by other fieldwork. Excava­tions at Muro Tenente were preceded by intensive field sur­veys of both the intra-mural area and its extramural perip­hery. These surveys allowed us to build hypotheses on structural transforma­tions in local

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

G.J. Burgers, D. Lentjes

Fig. 1. Salento, major archaeological sites.

landscape and settlement organiza­tion. A series of small trenches was subse­quently laid out in order to test the surface data. Both surveys and test trenches indicate that the site was continuously and intensively occu­pied from the 8th century BC until within the early Imperial age. It reached its maximum expan­sion in the early Hel-

lenistic period, in the late 4th or e­arly 3rd century BC. In this same phase a fortifi­cation wall was built, clearly with the aim of enclosing the entire resident population. This population was concentra­ted in large builtup nuclei separated by open areas. Two of these nuclei have been selected for large scale excavati­ons, one at the centre of the site and one at its nor­thern periphery. In the periphery of the site a continuously built-up domestic quarter was excavated, with parallel streets and lined-up houses of more or less equal dimensions. This quarter clearly belongs to the phase of maximum expansion of the site. In contrast, in the central settlement area, adjacent to domestic structures dating to the same phase, a necropolis was excavated that started as early as the 6th century BC.

The archaeobotanical research at Muro Tenente The archaeobotanical investigations at Muro Tenente started only in 1998. As yet, of the total amount of 72 charcoal samples that were collected, 52% or 72% have been analysed. It is important to mention that these samples don’t represent all excavated contexts; they mainly concern the early Hellenistic domestic layers of the central excavation area. Fig. 3 shows the distribution of the analysed samples

Fig. 2. Muro Tenente. Aerial photograph of the site.

40

Charcoals in context: anthracological analysis at Muro Tenente

Fig. 3. Muro Tenente. Spatial distribution of the analysed charcoal samples in trench no. 18 in the centre of the fortified area.

Fig. 4. Total number of charcoal fragments from the Muro Tenente digs that could (partially) be analysed.

41

G.J. Burgers, D. Lentjes

Fig. 5. Taxonomic graph of the analysed charcoal pieces found at Muro Tenente. The fragments that could not be determined (from which 2297 dicotyledons, 97 monocotyledons and 386 undeterminable) are left out.

in the central trench. Most of them were collected within the domestic areas in the north and northwest, relatively few were found in graves. Unfortunately, large part of the charcoal pieces studied here proved to be very hard to determine in detail (Fig. 4). In fact, most of the Muro Tenente charcoal samples were heavily damaged, in some cases even glazed. This is the case, for instance, with regard to the few samples that were taken from graves, almost all of which were either pulverized or glazed, and sometimes both. As yet, it is not clear which formation processes must be accounted for to explain this phenomenon, although it can be assumed that the samples were burnt at high temperatures; it could not be determined whether the glazed charcoal fragments were related to pyrotechnical craft structures, as has been positively established for the Archaic Salento site of Cavallino (Fiorentino and Colaianni 2005, p.98). More in general, it must be concluded that due to these biases, drawing conclusions from the charcoal samples from Muro Tenente is highly complicated.

indicated in the graph came from 9 different samples. In contrast, the 28 pieces that were found of Quercus cfr. ilex (oak, probably the evergreen holm-type) came from two different samples. All the other species that were found in the Muro Tenente samples only appear once: 13 pieces that could be identified as Mormon tea (cfr. Ephedra sp., 13 pieces, found in the periphery trench); six pieces of Myrtle (Myrthus communis: these were found in a foundation trench in the central excavation area; nearby, in a Hellenistic floor level, another 27 probable Myrthus communis-specimen were found); 29 specimen of Stone pine (Pinus pinea / halepensis: found in a grave in the central trench; these fragments could derive either from a Stone or an Aleppo pine, but the former is much more common in this region); 29 pieces of the Pistacia-family (found in an ash dump in the central trench); 31 pieces of Poplar or Willow (Populus/Salix, from the periphery trench); 8 small fragments identified as pear or apple wood (cfr. Pyrus/Malus, from a domestic context in the central trench); 18 specimen of Buckthorn or Narrowleaved vetch (Rhamnus/Phillyrea: from a domestic context in the central trench); 14 fragments of pomegranate wood (cfr. Punica granatum, from the periphery trench); some 200 large pieces that were identified as oak wood (cfr. Quercus sp., from postholes in the central trench) and finally, 10 tiny fragments of Sage (Salvia sp., from the periphery trench).

Fig. 5 shows the total number of fragments of particular species found at Muro Tenente. In 15 samples, a total number of 778 pieces was found of different types of Erica (heath tree), some of which were identified as belonging to the arborea and the multiflora type. Olive (Olea europaea) also appeared in more than one sample: the 97 fragments 42

Charcoals in context: anthracological analysis at Muro Tenente was built in the same period. That the settlement expanded according to traditional principles of spatial organization is suggested by the degree of dispersion of the new early Hellenistic domestic quarters. In no way can the site be considered as a truly agglomerate settlement. As in the past, relatively large open spaces can be assumed in those zones between the domestic clusters where we have documented a very low density of artefacts of any kind. Drillings and stratigraphical trenches have proved these zones to be lacking early Hellenistic archaeological structures in the subsoil. They are likely to have been maintai­ned deliberately in order to separate the various settlement nuclei and to accommodate for small scale agriculture and the stocking of herds. To find more or less ‘wild’ vegetation in these contexts, like the large amount of maquis-taxa, can be held as a confirmation of this hypothesis.

Fig. 6. Muro Tenente. Fragment of olive wood (Olea europaea), found in trench no. 18.

Of particular interest to our research objectives at Muro Tenente are the specimen of olive wood found amongst the charcoal samples (Fig. 6). As already mentioned, fragments of Olea europaea were found in nine different samples, that could all be dated in the Hellenistic period (around 280-230 BC). One of these samples was collected in the periphery trench, the others come from the excavations in the centre of Muro Tenente: five from domestic contexts and three from graves. The significance of these finds relates to the theme of urbanization, central to our research in Salento; one of the major aims of the archaeobotanical research at Muro Tenente is to investigate to what degree the ongoing process of urbanization was accompanied by transformations in the agricultural regime of the site. We hypothesize that, apart from an expansion of agriculture, specialization in arboriculture was among the major assets of this transformation. Unfortunately, as yet it is too early to prove this hypothesis on the basis of the charcoal samples collected at Muro Tenente, or even whether olive trees were actually cultivated in or near the site, or that they were merely part of high maquis-vegetation. There is, however, other evidence, textual and archaeological, suggesting that during the early Hellenistic period arboreal cultivation became firmly established besides traditional cultivation of cereals. With regard to the ancient texts, in themselves they do not tell us much on the item of Salento agriculture. If the region is mentioned at all, then mainly in connection to the Roman period. This also applies to arboricul­ture. Both Salento viney­ards and olive groves are frequently encounte­red from the 2nd century BC onw­ards1. They clearly had more than regional fame and were ranked among the better known orchards of Italy. From these references, which suggest substan­tial experien­ce with viticulture and the cultiva­tion of olives in the 2nd century BC, one could even deduce the retrospective hypo­thesis that arboriculture was

Interpreting these data, a series of observations can be made. First one may point to such rare specimen as Pinus cfr. pinea and Populus/Salix. With regard to the former, one must emphasize the context in which the fragments were found, i.e. in a grave (no. 805). This leads to investigate further the hypothesis that Pinus could have had a special significance connected to local burial rites. Similarly rare is the occurrence of Populus / Salix, of which a fragment was found in a stratigraphical layer covering Hellenistic domestic structures in the periphery trench. These structures were abandoned at the end of the 3rd century BC, and shortly afterwards covered with alluvial soils. This could explain the presence of Populus/Salix, because it favours wet soils. Unfortunately, the Populus-specimen was in such a bad state that a more precise analysis proved to be impossible. As for the other samples analysed, the general picture is formed by low maquis taxa such as Erica, Myrthus, Salvia and Ephedra, and (middle-) high maquis species like Quercus, Rhamnus/Phillyrea and Pistacia. One must conclude that in particular Erica-vegetation was a common phenomenon in Hellenistic Muro Tenente. Fifteen samples, which is to say almost 30%, throughout the two main trenches contained this particular small, tough plant. Erica is pyrofile: it grows quite well on recently burnt grounds. Its over-representation could be interpreted as a sign that the site was burnt down at some point, but archaeological evidence for this is absent. Another possibility to be considered is that Erica was omnipresent in the vicinity of the habitation quarters. To test this hypothesis, spatial information on domestic structures must be evaluated. In this context we are fortunate to be able to point to the data proffered by the systematic on-site field surveys carried out at Muro Tenente in 1993. These surveys have demonstrated that in the early Hellenistic period a centrifugal expansion of habitation took place from the central area outwards in all directions, halted only by the fortification enceinte that 43

G.J. Burgers, D. Lentjes (Per­kins and Walker 1990, p. 58 for Etruscan Doganella). These conclusions can be compared with the results of the surveys on the Adriatic coast near Valesio (see above). At Valesio concentrations of tiles, amphoras and dolia are also found outside the defences. Interestingly, seven of such sites are located near the defences, to a maximum distance of 1 km. The concentration of these storage sites close to the settlement is significant. It can be explained by drawing a parallel with the characteristic model of land use zonation as documented by historical geograp­hers for Italian towns of more recent times2. The spatial orga­nization of this typical agricultural landscape is determined mainly by transport and labour demands. In many cases this led to a pattern of concentric zones around the town, with the more labour-intensive crops planted closer to the town. Thus, a first zone, nearest to the town, is generally given over to irrigated vegetable gardens, vineyards and olive groves. This zone is surrounded by arable lands, beyond which a third zone is left uncultivated, for permanent grazing. If, indeed, the agricultural landscape around the early Hellenistic Brindisino sites was comparable to some degree to such a zonation, it is not surpri­sing to find sheds specific for the storage of the products of intensive agriculture concentrated near the surveyed sites, lying either within the periphery of the walled area or immediately outside it. One could suggest that these sheds were standing in the midst of orchards making up a first zone of cultivation surrounding the ancient sites.

Fig. 7. Muro Tenente. Grape pip (Vitis vinifera) from grave 27 in trench no. 18 (figure 3; north).

introduced in an earlier phase. Thus, as early as the first half of the 2nd century BC Cato menti­ons olives of the Salento type, which need to be planted in heavy soils: “in agro crasso et caldo oleam ... sallentinam ... eam maxime serito” (Cato, De Agricultura 6.1; Lombardo 1992, 38-39, no. 65). This characterization is significant in the present context. Considering that it takes 10 to15 years for olive trees to bear fruit and that it would undoubtedly have taken much longer to acquire a reputation, one could suggest that the cultivating olives was already firmly established in the 3rd century BC. Although the written evidence for this hypothesis is admit­ tedly weak, it can be more firmly proposed on the basis of archaeology. Particularly revealing in this context is the discovery of large amounts of probably 3rd century BC amphora fragments within the Brindi­sino survey areas, discussed in the introduction. The great majority of these fragments can be classified as Graeco-Italic amphoras. They are generally considered to be containers for the storage and transportation of wine and/or olive oil. Dense concentrations of these amphoras were found for instance at Muro Tenente and Valesio, immediately along the inner side of the defences. Here, the distribution and density of amphora fragments is such that the existence of sheds for storing amphoras and dolia can be hypothesized at several spots within the walls of the early Hellenistic towns (Yntema 1993b, pp. 61-63). If this interpretation is accepted, the location of these sheds can be called significant. It can be argued that they were located deliberately in peripheral zones within easy reach of the entran­ce ways of the settlements to facilitate storage of olive oil and wine from vineyards and olive groves in the surrounding country­side

The above field survey results clearly add much information to the unfortunately still meagre archaeobotanical data in Salento. However, in order to test these results, more problem oriented archaeobotanical field work is definitely needed. For this reason we hope that further investigation of the Muro Tenente botanical samples might throw new light on the issue of arboriculture; some 200 samples, containing botanical macrorests, are still waiting to be investigated. In the present context it is noteworthy to mention that preliminary analysis has confirmed that at least six of the samples from the central excavation trench and two from the periphery trench contain grape pips (Fig. 7). In our view, the issue to what degree fruit-bearing species like grapes (Vitis vinifera) and olives (Olea europaea) were cultivated should be one of prime importance in the agenda of the pre-Roman archaeology of the Salento region. In the same line of reasoning, we hold that with future problem-oriented research, archaeobotany has much to offer to the issue of ancient urbanization in general.

Acknowledgements

The archaeological and botanical research at Muro Tenente would not have been possible without the generous support of the Scuola di Specializzazione in Archeologia Classica e Medievale ‘Dinu Adamesteanu’ of the University of Lecce. In the present context we especially wish to thank its director prof. Francesco D’Andria, and also prof. Girolamo Fiorentino for offering the facilities that have enabled this study. Moreover, we are due to

44

Charcoals in context: anthracological analysis at Muro Tenente dott. Francesco Solinas for his help with the botanical analysis. We would also like to thank Bert Brouwenstijn, Jaap Fokkema and Harry Burgers (Vrije Universiteit of Amsterdam) for preparing the figures. The English text was edited by the Vrije Universiteit’s Language Centre.

Italy. Final Report on the Field Survey in the Town Area. Bulletin Antieke Beschaving, 68, 49‑70.

Endnotes Cato, De Agricultura 6.1 (Lombardo 1992, 38-39, no. 65); Varro, Res Rusticae 1.8.2/1.24.1/2.6.5 (Lombardo 1992, 50-52, nos. 81, 82 and 84); Dionysius Halicarnassus, Antiquitates Romanae 1, 37, 2 (Lombardo 1992, 78, no. 133); Strabo, Geographi­ca 6.3.5 (Lombardo 1992, 97-98, no. 183); Columella, Res Rustica 12.50.3 (Lombardo 1992, 119-120, no. 212); Plinius, Naturalis Historia 15.4.20/17.35.166 (Lombardo 1992, 131-132, nos. 234/236­); Macrobius, Saturnalia 3.20.6 (Lombardo 1992, 192, no. 359); Hesychius, Lexicon B 629 (Lombardo 1992, 199, no. 382). 2 See especially Chisholm 1968, pp.43-67 and Delano Smith 1979, pp.172-176 with extensive bibliography. 1

References Alberda van, K., Burgers, G.J., Burgers, H., Karel, D., Yntema, d., 1999. In: A. Nitti, ed. Muro Tenente. Centro messapico nel territorio di Mesagne. Manduria. Boersma, J.S., Burgers, G.-J, Yntema, D.G. 1991. The Valesio Project; final interim report, in: Babesch 66, 115‑131. Burgers, G.J., 1998a. Constructing Messapian Landscapes. Settlement Dynamics, Social Organization and Culture Contact. Dutch Monographs on Ancient History and Archaeology, XVIII. Burgers, G.J., 1998b. Muro Tenente: indagini archeologi­che della missione olandese. In: M. Lombardo, C. Marangio, eds. Il territorio brindisino dall’età messapica all’età romana. Atti del IV Convegno di Studi sulla Puglia Romana. Galatina: Congedo, 137-150. Burgers, G.J, Yntema, D.G. 1999. The Settlement of Muro Tenen­ te. Third Interim report. Bulletin Antieke Beschaving, 74, 111-132. Burgers, G.J., Attema, P.A.J., van Leusen, M., 2003. Walking the Murge. Preliminary report on The Ostuni Field Survey’. Studi di Antichità, 11, 1-26. Chisholm, M., 1968. Rural Settlement and Land Use. London: Hutchinson University Library. Delano Smith, C., 1979. Western Mediterranean Europe. A Historical Geography of Italy, Spain, and Southern France since the Neolithic. London: Elsevier Science & Technology Books. Fiorentino, G., Colaianni, G., 2005. L’analisi archeobotanica, in: F. D’Andria, ed. Cavallino. Pietre, case e città della Messapia arcaica. Ceglie Messapica. 96-99. Lombardo, m., 1992. I Messapi e la Messapia nelle fonti letterarie greche e latine. Galatina. Lyding Will, E. 1982: Greco-Italic Amphoras, Hesperia 51, 338‑351. Perkins, P., Walker, L., 1990. The Etruscan city at Doganella. Papers of the British School at Rome, 58, 1-143. Yntema, D.G., 1993a. In Search of an Ancient Countryside. The Free University Field Survey at Oria, Province of Brindisi, South Italy (1981‑1983). Amsterdam: Thesis Publishers Amsterdam. Yntema, D.G., 1993b. The Settlement of Valesio, Southern

45

Wood in arid zones’ prehistoric architecture I. Caneva

Dipartimento di Beni Culturali, Università del Salento, Via D. Birago, 64, 73100 Lecce (Italy). E-mail: [email protected]

Abstract A short overview on the prehistoric architecture of the Near East and the Upper Nile valley leads to contradict the idea of environmental deterioration or technological improvements at the origin of the progressive contraction in the use of wooden architecture. For different reasons, and in different cultural contexts, wood continued to be used and valued, while more durable materials were reserved for buildings intended to last for a long time, either fortifications, monuments, or funerary structures, with this variety being linked to the specific ideology of the groups.

Trees can grow in virtually all the environments in which human life is possible and thus constitute one of the most persistent component of human landscapes. Besides fruits and leaves, they offer wood for a variety of potential practical use: fuel, structures, containers, furniture, perfumes, handles or parts of tools, art or symbolic objects. Wood from prehistoric excavations informs, therefore, not only on a number of climatic and ecological factors, but also on different technological aspects of human life, including cultivation, trading and, above all, a wide range of artefact manufacture. Among the technological uses of wood, are various aspects of architecture, such as whole structures, or planking, roofing, flooring, and others. Wood was most probably the material used to build the primeval hut in the world, and till today has never ceased to be used in one or another construction phase. Unfortunately, wood is among the most perishable materials and its finding in archaeological excavations is limited to carbonised remains. In many cases, the use of wood can only be inferred from indirect evidence, such as post holes or wood imprints on mud plaster, or, even more indirectly, from the microscopic observation of distinctive wood working wear traces on the edges of lithic tools. Wood and vegetables were used in architecture in various forms, as planks and posts for the structure, as branches and reeds for filling the spaces between the poles, and, finally, as chaff mixed with mud for roof covering and floor and wall plastering. The simple presence of mud plaster, even without wood or reed imprints, coupled with the absence of other building materials, is sometimes enough for the archaeologists to recognise the presence of a wattle and daub architecture, which implies the use of wooden posts and frameworks. The type and intensity of the use of wood as a building material defines specific kinds of architecture. In archaeology, this definition is usually translated into a social one, and considered to shed light on the kind of society which produced that kind of buildings: architecture, more than any other human artefact, can in fact be considered as the culture-specific material expression of human groups, with their technology and internal organisation and interrelations (Hall 1972). Wood is doubtless a more fragile, perishable and short

living material than stone or mud, and probably wooden structures reflected this “short run” perspective in their construction and spatial organisation. The reading of the organised social contacts through the layout of the internal or external spaces in a prehistoric village is, therefore, much more difficult when wooden structures were used, while it is easier and more significant when consitent and durable stone or brick buildings were found in the archaeological contexts. The question is a relevant one, since, as said, architecture is usually equated to the corresponding society. Archaeological studies have been heavy conditioned by this view, and have focused on the cultures characterised by consistent architectural remains, attaching much less importance to the others, which were considered as marginal and historically irrelevant. This was particularly the case with the analysis of the beginning of the dynastic civilisations in the Near East and in the Nile valley. The gradual emergence of complex societies in these areas has been defined by the steps in the development of “complex” architecture, considered as their material image, with almost no attention being paid to the less consistent forms of architecture co-existing in the region, although it should be kept in mind that the related groups must have influenced, in their turn, the urbanisation process of the region. The role of wood in architecture has indirectly became, in this way, a critical element for the reconstruction of histori-

Fig. 1. Pre-pottery Neolithic A (PPNA) house at Beidha (post holes are visible on the floor).

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

I. Caneva cal processes. This role, however, is probably conditioned by a variety of factors, which can not be simply summarised in the traditional assumptions that wood is an easy, primitive and non sophisticated building material, and that its use depends on its availability in the territory, and is therefore only a land-specific element (Aurenche 1981). As many authors have said, not only typological features, but also a variety of technological (Schirmer 1990), social (Flannery 1972) and symbolic (Forest 1987, Hodder 1990) aspects should be considered when dealing with the forms, size and materials of ancient architecture. Based on these considerations, this paper attempts to a preliminary recognition on the conditionments which might have played a role in the use of wood in the Levant and the Upper Nile valley before urbanisation.

Wood in the Near East The beginning of house building in the Levant dates back to the Natufian and contemporary Zagros and Taurus epipalaeolithic cultures, in the mid XIII millennium BC, or even slightly before. Most of these early villages were established on sloping areas, and the houses built on artificial terraces, with the rear wall being built against the face of the terrace. The first structures were tents or shelters made of organic materials, erected against a stone built terrace wall. The first huts were partially sunk into the soil, at least on one side, and wood constituted the entire wall and roof structure. The wooden poles were thus the first prefabricated building material ever produced by man. The poles were supported by stone foundations and socles, and were plastered with straw tempered mud. The structures were circular in shape and had a limited size, ranging from 2-3 m to 4-5 m in diameter (Bar Yosef 1992). In the following Khiamian and Pre-Pottery NeolithicA cultures, the same technology continued, with many settlements, such as Jerf el Ahmar, Nahal Oren, Netiv Hagdud, el Khiam, and others, being again terraced on sloping areas. In most of the sites, however, the huts were no longer sunk into the soil, but entirely built. The houses had stone foundations, wooden post structures, and branch and reed walls. In many villages, plano-convex mud bricks, best known from Jericho, were already used. Jericho shows also the first consistent public architecture documented in the world, consisting in a massive stone wall and a huge tower. Even in these early village phases, different materials were used in the same site, wood, stone and mud bricks, suggesting that function, more than technology or raw material availability, conditioned the choice. Different types of wood were used in the Neolithic Near East. Poplar tree, which could reach 6 m in length, was commonly used at Mureybet, in Syria, in all Neolithic phases, although the carbonised wood fragments found in the deposits could not always be associated to architectural functions rather than to fire wood or artefact making. Oak tree was certainly used for the central pole in one of the

Fig. 2. Pre-pottery Neolithic B (PPNB) ligneous subdivisions in the house at Mureybet.

PPNA round house in this site, indicating that a specific wood was selected according to its function in the structure. Oak was used also in much arid regions, such as southern Jordan and Palestine. At Beidha, it occurred together with pine, pistachio and juniper trees (Aurenche 1981, p.81). In many sites wood was obviously imported from other areas, often from far away, suggesting that the specific properties of each kind of wood were well known and quite rationally used. Probably less valuable than wood tree were shrubs and reeds, which were used to fill the spaces between the poles or to cover the roof. In southern Iraq, where wood has always been extremely rare, timber could have been replaced by long reeds assembled in bundles, as is documented in present day huts: the bundles,

Fig. 3. Impressions of reed matting at Jericho (PPNB levels).

48

Wood in arid zones’ prehistoric architecture Kent Flannery (1972) associated the change from round to rectangular in the shape of the dwelling structures to a change in the social organisation. The round shape reflected a non producing or a proto agricultural society, with each hut hosting one function or one member of the extended family which constituted the village group, and with communal food storing occurring in the open space between the huts. The multi-roomed rectangular house, instead, was associated to a nuclear family, in a food-producing society, with each room hosting separate activities and food storing, and with several separate families inhabiting the village. While Flannery emphasised the economic and social significance of the change in building shape and materials, other authors referred to symbolic aspects, such as the consideration of the house as a sacred microcosm opposed to the wild outside world, and includes life and death, the latter being ideologically “domesticated” by assuming the dead inside the “domus” (Hodder, 1990). The need to better separate the external world, from the internal, domestic one, and to protect the latter from contamination and danger, would have pushed people to build more solid houses. Probably due to a combination of these factors, more durable building materials, such as dressed stones or mud bricks or pisé, started to be widely used in the Pre-Pottery NeolithicB phase, in all the areas considered. Wattle and daub walls were the first to be replaced, but also the standing wooden posts were soon replaced by mud brick or stone pillars, as in the site of Nemrik (Kozlowski 1990).

Fig. 4. Reed structure for floor lathwork (PPNB levels at Jericho).

which are often more than 9 m long and 35 to 70 cm in diameter, are flexed to form at the same time walls and roof of big huts (Aurenche 1981, p.79). In the Neolithic Levant, the tree trunks were variously modified and prepared, with the bark and branches being removed, often leaving a biforcation to facilitate the assembling. The base was either pointed, or squared. The posts were sunk into the soil, in deep holes, sometimes lined with stones. Post holes are often the only architectural element preserved. Their size and depth, and the distance between them, indicate the type of structure and its function, either roofed hut or shelter, as well as the technique of wall framing and lining. In some cases, the poles were close to one another, in others, they were more than one meter apart, and reeds and branches were arranged between them and plastered with mud. In the sunken huts, the posts had the double function of vertical and lateral support, and the intervals between the poles were often reinforced with stone containing walls, as can be observed at Jerf el Ahmar, in Syria (Stordeur 2000), at Hallan Cemi (Rosemberg, Davis 1992), in south-eastern Turkey, and at Beidha, in Jordan (Kirkbride 1966). The growing practice of agriculture, at the end of the PrePottery NeolithicA, introduced in the settlements a number of changes, related both to the practical need for distinct spaces for cereal storing and manipulating, and to the “nuclearisation” of distinct families within the group. The new structures were bigger and had internal wooden or mud walls to separate different zones. In the Pre-Pottery NeolithicB phase the circular shape was no longer used, replaced by rectangular multi-roomed houses. The change was gradual, with transitional phases with multi-roomed round houses being attested, as in the well known examples from Mureybet (Cauvin 1994) and Beidha (Kirkbride 1966). Though wooden posts were still used, stone and cigar-shaped mud-bricks were apparently more effective building materials for these larger houses. Wood was probably limited to the roof, while other vegetables, such as branches or reeds, were used, as attested at Jericho, for floor preparation and revetment (Kenyon 1957). In a famous comparative ethnoarchaeological analysis,

Fig. 5. Modern buildings made of fastened reeds in southern Iraq.

49

I. Caneva The round shape was however not completely abandoned. At Jerf el Ahmar, both shapes were present in the same level, with the round huts being built with the same early technique, sunken into the soil, with a wooden superstructure. The fact that these houses were clearly reserved to ceremonial functions suggests that wooden structures, far from being out of fashion, were taken in good consideration, though no longer practical for daily life. Other examples of a similar differentiation from rectangular stone built dwelling houses and roundish partially sunken public structures can be found in eastern Anatolia, at Nevali Cori (Hauptmann 1988), where in the public building stone monoliths replace the wooden posts. During the following development of the late neolithic and chalcolithic cultures of the Levant, architecture showed a variety of dwellings, due to the new variety of specialised activities, agriculture, hunting, and herding flocks (Bar Yosef 1992; Porath 1992). Dwellings vary from natural shelters, to subterranean pit houses, to wooden huts, and to a solid stone architecture, in which wood was often eliminated even from the roof, with stone vaulted or domed structures being adopted. The resumption of the round plan wooden hut characterised for several millennia the pastoral cultures of southern Levant, similarly to those of the Nile valley.

Fig. 6. Acacia wood huts with the entrance daubed with mud in Sudan.

to leave any trace on the soil. Since burials were frequently found in the settlements, below the occupation level, it has been assumed that they were organised according to the dwelling structures, either outside the door, or below the floor, but these dwelling elements are at present indistinguishable (Arioti, Caneva 2004). These cultures had in the hinterland a complementary supply area, organised in a complex settlement pattern which covered a vast territory. The system comprised permanent sites along the river and long-term camp-sites in the hinterland, used alternatively in the dry and in the rainy seasons. Both settlement areas were complemented with a number of hunting lo-

Wood in the Upper Nile valley, Sudan Unlike what can be observed in the Near East, where dwelling models were formal and quite standardised elements of cultural definition, in the north-African regions, prehistoric settlements were rare and poorly preserved and architectural features hardly recognisable and classifiable in definite categories through time. As in southern Iraq, wood is rare in these landscapes, unlike clay and sand. In spite of this, wood appears to have been the only building material ever used in prehistory. In the hypothetical reconstruction of the Holocene vegetation cover of the upper Nile valley, thorny trees and shrubs of different Acacia species predominated in the drier the savanna landscapes, away from the Nile (Wickens 1982). The Nile flows for about 1500 km across the desert regions of Sudan and Egypt, making life possible on its edges only, though human life was certainly possible in a larger area 9000 years ago, when the latest post-palaeolithic hunterfishers inhabited the central Sudan. The Mesolithic settlements occupied an almost uninterrupted strip along the river, with thick concentrations of stone and bone implements, and food residues. One of the peculiarities of the late hunter-gatherers of the Sudan is their sedentariness, attested first of all by the considerable thickness of the deposits, in spite of the marked deflation these territories underwent. The quantities of large, heavy pots and grinding stones, coupled with the good state of preservation of food residues, are additional indicators of settlement duration and stability. Their huts, however, were probably too light

Fig. 7. Wattle and daub techniques in a modern Turkish house.

50

Wood in arid zones’ prehistoric architecture cations, on top of the rocky outcrops (Caneva, Santucci 2006). In spite of the great stability of this system through time, the dwelling perspective of these people was obviously a short term occupation. The present day nomads use round structures made intertwining branches of Acacia trees, which may suggest that similar constructions were made in the past. The twisted branches provide a solid self standing structure which does not need to be sustained with sunken posts. When the group moves, the structure can easily be dismantled and taken away, without leaving any trace on the soil.

Conclusions Although this short overview is far from being exhaustive, and no conclusive assumption can be made until a more comprehensive analysis is made, it enlightens the inconsistency of some traditional assumptions. First of all, it contradicts the idea of a progressive contraction in the use of wooden architecture depending on environmental deterioration or technological improvements: as for all human artefacts, changes occur in conformity to specific cultural needs, in a mixture of technical, practical and ideological elements. In the case of wooden architecture, environmental availability was a secondary problem: in the Near East, the change in architectural techniques regarded the whole area, from the arid zones of southern Jordan, to the forested ones of eastern Anatolia; in both the Levant and the Nile valley, the widest and most constant use of wood was testified in the most arid zones. Technical knowledge was also a minor element, since mixed techniques were in many cases used in the same construction, or different types of construction co-existed in the same village, always displaying different functions. An interesting recurrent aspect in this overview, is that wooden structures were not disdained when other building techniques were in fashion, but they were instead highly valued and reserved to ceremonial functions, as shown for instance by the Pre-Pottery Neolithic B level at Jerf al Ahmar, in Syria, and by the 3rd millennium city of Kerma, in Upper Nubia, as well as again by the Pre-Pottery Neolithic B phase at Nevali Cori, in south-eastern Anatolia, where wooden posts were significantly replaced by stone pillars (Hauptmann 1988). In fact, it is the practical aspect, particularly that of duration, which appears to have been more restrictive and prevented people from using wood for buildings intended to last for a long time. It therefore depended on the long or short run perspective with which the buildings were made that one or the other material was selected: in all cultures, both in prehistoric Levant and in the Nile Valley, the most durable materials were reserved for fortifications, monuments, or funerary structures, with this variety being linked to the

Fig. 8. Acacia tree in Sudan.

In the following phase, in the prehistoric cultural sequence of these regions, the introduction of domestic sheep and goat produced probably a growth of the animal populations, both wild and domestic, which impoverished the vegetation cover, leading to a quick advance of the desert. Very soon, at least at the end of the fifth millennium BC, a fully nomadic life style was necessary to feed the animals, and the previous village system was abandoned. The new nomadic cultures are since then best known through their graves. A few settlements of this period recently excavated in Nubia testify for wooden round plan structures, with the village being protected by wooden palisades, still used during the 3rd mill. BC (Honegger 2003, 2005). In the city of Kerma itself, levels with mud brick buildings alternated with levels with wooden huts, and the central, public hall of the classical city was a round hut with wooden posts (Bonnet 1990, p.34-35). In the same periods, more consistent stone structures were reserved for graves. The excavation of several stone tumuli in different sites has recently provided a cultural and chronological definition of this widely used funerary form, revealing that although they appear to have a similar shape, they belong to different cultures and lasted for several millennia, from the late Neolithic to Islam. These groups were highly mobile, their lifestyle being organised in a wide nomadic cycle and the cemetery appears to have been their main reference point as regards both the geographical and social organisation.

Fig. 9. Nomadic camp of shepherds in Sudan.

51

I. Caneva specific ideology of the groups. In the Near East, prehistory was characterised by a trend towards settlement stability. With the growing dependence on their own cultivated fields, people became sedentary and changed the earlier seasonal, short run settlement perspective into a long run one. First the internal structures of the hut, such as mortars, storage pits and ovens, became permanent. Than the hut itself changed into a house and started to include all aspects of life and death. Quite significantly, this change was not applied to the public buildings. In the upper Nile Valley, prehistory was characterised by an inverse trend, towards settlement mobility, from a seasonal shifting between two residential areas, to a wider nomadic cycle. The first huts were light, almost portable structures, and probably never changed in the following millennia as a common dwelling structure. In the cultural history of this area, the other building materials were not unknown, but were reserved to permanent structures, such as temples or burial tumuli, as is best testified by the monuments of the dynastic periods in both Sudan and Egypt. It should be stressed, however, that these materials were never particularly valued per se, compared to wood, and not only at Kerma, but also at Hierakonpolis, in Upper Egypt, the proto-dynastic public buildings consisted of wooden structures (Adams 1974). In the same perspective, present day semi-nomadic groups build mud brick houses as storerooms or animal stables, but prefer to live in the more costly wooden huts or tents (Casciarri 1999).

d’un code symbolique. Anatolia Antiqua, II, 1-42. Hall, A., 1972. People and building. Hauptmann, H., 1988. Nevali Çori, Architektur. Anatolica, 15, 99-110. Hodder, I., 1990. The domestication of Europe. London: Blackwell. Honegger, M., 2003. Peuplement préhistorique dans la region de Kerma. Genava LI, 281-90. Honegger, M., 2003. Kerma et le début du Néolithique Africain. Genava, LIII, 239-49. Kenyon, K.M., 1957. Digging up Jericho. London: Benn. Kirkbride, D., 1966. Five seasons at the Pre-Pottery Neolithic Village of Beidha in Jordan. Palestine Exploration Quaterly, 98/1, 8-61. Kozlowski, S.K., 1990. Nemrik 9. Pre-Pottery Neolithic site in Iraq. Warsaw: Wydawnictawa Universit Warsawskiego. Porath, Y., 1992. Domestic architecture of the Chalcolithic period. In: A. Kemoinsky, R. Reich, eds. The Architecture of Ancient Israel. Jerusalem: Israel Explotation Society, 40-48. Rosemberg, M., Davis, M.K., 1992. Hallan emi tepesi, an early aceramic Neolithic site in eastern Anatolia. Anatolica, 18, 1-18. Schirmer, W., 1990. Some aspects of building at the “aceramic neolithic” settlement of Çayönü Tepesi. World Archaeology, 21, 363-87. Stordeur, D., 2000. Jerf el Ahmar et l’émergence du Néolithique au Proche Orient. In: J. Guilaine, ed. Premiers paysans du monde, Paris: Errance. Wickens, G.E., 1982. Palaeobotanical speculations and Quaternary environments in the Sudan. In: M.A.J. Williams, D.A. Adamson eds., A land between two Niles. Rotterdam: Balkema, 23-50.

References Adams, B., 1974. Ancient Hierakonpolis. Warminster. Arioti, M., Caneva, I., 2004. Sepolture di pastori nell’Africa nord-orientale. In: M. e Barogi, F. Lugli, eds. Atti del Convegno Nazionale di Etnoarcheologia. Rimini: Raffaelli, 194-201. Aurenche, O., 1981. La maison orientale, Paris: P. Geuthner. Bar Yosef, O. 1992. Building activities in the prehistoric periods until the end of the Neolithic. In: A. Kemoinsky, R. Reich, eds. The Architecture of Ancient Israel. Jerusalem: Israel Exploration Society, 31-39. Bonnet, Ch., éd., 1990. Kerma, Royaume de Nubie. Université de Genève. Caneva, I., Santucci, E., 2006. Late hunter-gatherers of Central Sudan: land use and settlement pattern. In: A. Roccati, I. Caneva, eds. Nubica, Proceedings of the X International Conference for Nubian Studies, Roma 2002. Casciarri, B., 1999. Essere ‘arab tra gli Ahamda del Sudan centrale. La ricerca folklorica, 40: Numero speciale: Società pastorali d’Africa e d’Asia. M. Arioti e B. Casciarri eds., 117-134. Cauvin, J. 1994. Naissance des divinités, naissance de l’agriculture. Paris: CNRS éditions. Flannery, K., 1972. The evolution of the village as a settlement type. In: P. Ucko, R. Tringham, G. W. Dimbleby, eds. Man, Settlement and Urbanism. London  : Duckworth, 23-53. Forest, J.D., 1987. Çatal höyük et son décor: pour le déchiffrement

52

A fuoco lento: strutture di combustione nell’abitato dell’età del Bronzo di Coppa Nevigata (Manfredonia - FG) A. Cazzella1, G. Recchia2

Università di Roma “La Sapienza”, Dipartimento di Scienze Storiche, Archeologiche, Antropologiche dell’Antichità. Via Palestro, 63 – 00185 ROMA. E-mail: [email protected]. 2 Università di Foggia, Dipartimento di Scienze Umane. Via Arpi, 155 – 71100 FOGGIA. 1

Abstract The recognition of “fireplaces”, in a broad sense, in an archaeological context is generally considered one of the easiest interpretive cases, as regards the connection between archaeological traces and human activities that produced them. On the contrary, the specific function of the various elements linked to the use of fire is not so understandable. On the basis of the rich evidence of firing structures from Coppa Nevigata (XVIII – XII BC) the authors deal with the criteria to classify them, considering their technical feature, location inside the settlement, relations with other structures and way of using them. The analysis leads us to suggest hypotheses as regards some behaviours linked to the use at Coppa Nevigata of firing structures and the areas where they were found, and their transformations in time. The principal area of excavation at Coppa Nevigata is a marginal part of the Bronze Age settlement, near the walls, at least till to the Subapennine phase (XIII – XII BC). The concentration of firing structures in that area between XVI and XIV BC is likely related to its marginal location to avoid dangerous effects. We cannot rule out they were linked to food preparing and cooking in a collective way (kin groups?). During the Early Protoapennine (XVIII – XVII BC) and the Subapennine the firing structures, as for example ovens, seem to be related to single dwellings, likely of nuclear families.

Il riconoscimento di punti in cui è stato utilizzato il fuoco in un contesto archeologico è considerato, in genere, uno dei procedimenti più semplici di interpretazione del deposito in relazione al rapporto tra tracce archeologicamente visibili ed azioni umane che le hanno prodotte. Tuttavia la funzione specifica dei diversi elementi connessi con l’uso del fuoco non è di altrettanto immediata comprensione. I parametri da considerare sono molteplici e vanno anche al di là di quelli strettamente legati alla conformazione delle varie strutture di combustione stesse, come, ad esempio: la posizione all’interno o all’esterno di ambienti coperti e il rapporto funzionale complessivo con il contesto di rinvenimento, le esigenze di illuminazione e riscaldamento, le abitudini alimentari che condizionavano le modalità di cottura, le caratteristiche degli eventuali recipienti impiegati per la cottura (come il tipo di materia prima in cui sono realizzati, la forma generale etc..), le attività di altro tipo svolte per mezzo del fuoco, i tipi di combustibile utilizzati. Su quest’ultimo aspetto le analisi archeobotaniche in generale e, nello specifico, quelle in corso sui resti antracologici di Coppa Nevigata da parte di G. Fiorentino per il riconoscimento dei diversi tipi di legno utilizzati, anche in relazione ai vari “punti di fuoco” individuati nell’insediamento, potranno fornire dati specifici. Il tipo di combustibile impiegato, infatti, può aiutare a comprendere le diverse “qualità di calore” ottenuto (ad es. fuoco “vivo” o braci), che, come si vedrà, possono essere collegate non solo alle modalità di cottura ed alle attività svolte, ma anche alla scelta ubicativa stessa del punto di fuoco, nonchè al tipo di struttura di combustione adottata. Le proposte di analisi, interpretazione e classificazione funzionale delle evidenze legate all’uso del fuoco, almeno in ambito italiano, sono spesso legate a situazioni specifiche o a singole strutture: se da un lato è chiaro come solo lo studio

contestuale possa fornire i dati necessari alla formulazione di ipotesi concrete e compiute, dall’altro le analisi comparate condotte su campioni più ampi, l’esplicitazione di criteri generali, l’individuazione degli indicatori archeologici utili e gli studi di tipo etnoarcheologico, senza voler creare una tassonomia generale di riferimento, possono aiutare nella comprensione dei singoli casi. In un recente lavoro, condotto in particolare sulle strutture di combustione relative all’età del Bronzo di Broglio di Trebisacce, C. Moffa (2002) ha giustamente ribadito la distinzione tra diversi tipi di strutture funzionali e, ad esempio nell’ambito di quelle fisse, tra focolari, forni e fornaci; l’autore pone inoltre l’accento sull’importanza delle caratteristiche tecniche e delle analisi archeometriche per il riconoscimento dei diversi tipi. Tuttavia anche all’interno delle diverse categorie proposte si possono avere numerose differenze, oltre quelle relative alle caratteristiche tecniche della base. Ad esempio per i focolari, tra le principali, accanto alla forma stessa (quadrangolare/circolare), vi sono quelle legate alle dimensioni, alla conformazione complessiva, alla collocazione nello spazio, all’eventuale rapporto seriale tra strutture simili o con altri tipi di sistemazioni. L’insediamento dell’età del Bronzo di Coppa Nevigata (una sintesi dei risultati degli scavi Puglisi in Puglisi, Coccolini 1982; Cazzella, Moscoloni 1987 e Cazzella et al. 2004; sugli scavi in corso: Cazzella, Moscoloni 1999a, con aggiornamenti in Cazzella et al. 2001, 2003, 2006) dove si sono messe in luce un buon numero di strutture di combustione per ciascuno dei suoi diversi livelli di vita (dal Protoappenninico -XVIII sec. a.C.- al Subappenninico avanzato –XII sec. a.C.) può costituire una buona situazione campione per prendere in esame una serie di strutture diversificate che appaiono comunque connesse

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

A. Cazzella, G. Recchia con l’uso del fuoco, benchè rimanga difficile stabilire la precisa funzione correlata con ciascun “tipo”. Per quel che riguarda l’analisi delle attrezzature di combustione si è cercato di individuare da un lato alcuni elementi di differenziazione (come la stabilità e la posizione rispetto ad ambienti coperti o strutture diverse), dall’altro caratteristiche strutturali generali, in relazione alle modalità di dispersione del calore, legate sia alla copertura che alla preparazione del sottofondo. Tali elementi sono stati ritenuti significativi in relazione al tipo di uso generale che poteva essere fatto delle strutture; per ciascuno di essi si è cercato di esplicitare i relativi indicatori archeologici (Tab.1). In sintesi una prima distinzione può essere fatta tra strutture fisse ed elementi mobili; tanto le prime che i secondi possono essere poi, in prima istanza, separati tra “aperti” e “chiusi”. Un parametro ritenuto tra quelli maggiormente significativi, inoltre, è la posizione dei punti di fuoco nell’abitato e la loro relazione con strutture o aree aperte. Di seguito si cercherà dunque di illustrare la varietà delle attrezzature messe in luce a Coppa Nevigata sulla base dei criteri esplicitati, a partire da quelle “fisse” ed “aperte” poste all’interno di strutture abitative e non, a quelle dalle medesime caratteristiche poste però in spazi aperti, per poi passare a quelle “fisse” e “chiuse”, quindi a quelle mobili, sempre tenendo conto della loro localizzazione al momento dello scavo.

cronologiche, ed il carattere pluristratificato del sito fanno sì che non siano al momento numerosi gli ambienti, presumibilmente coperti, messi in luce o che comunque si siano potuti scavare in modo integrale. Tale situazione può aver contribuito all’evidenza che le strutture di combustione poste all’interno di ambienti di questo tipo siano piuttosto rare nell’abitato; esse appaiono comunque diversificate tra loro per il tipo di sottofondo, nella forma, nelle dimensioni, etc. Tra l’altro, proprio l’ambiente nel quale sono stati individuati due piani in argilla di piccole dimensioni, con sottofondo in cocciame, esposti a fonti di calore (Fig. 2, n.8; Cazzella, Moscoloni 1999b,Figg. 6 e 7), non è necessariamente interpretabile come abitazione (si tratta di una struttura esterna rispetto alle prime mura, riferibile a una fase finale del Protoappenninico: Recchia 2001; Cazzella et al. 2002). L’evidenza di tracce di esposizione al calore in un punto concentrato, di forma sub-circolare, che insiste su un ampio piano preparato rappresenta invece un caso diverso: tale testimonianza è forse interpretabile come un “focolare” (Fig. 1, n.1), con caratteristiche non definibili, ed è posto all’interno di una struttura abitativa di forma probabilmente rettangolare, riferibile alla fase del primo abitato protoappenninico circondato da mura, dunque leggermente più antico rispetto al primo caso ricordato. Per l’Appenninico iniziale non si conoscono ancora ambienti coperti o ben delimitati, mentre alcune indicazioni si hanno per l’Appenninico Recente. In tale periodo l’area della porta protoappenninica di accesso all’abitato risulta tamponata: si determinò quindi uno spazio delimitato dall’alzato delle torri e dal tamponamento stesso, non necessariamente coperto, che non sembra direttamente interpretabile come interno di un’abitazione (Cazzella et al. 2005, Fig. 2). Qui era presente un piano di cottura in argilla con sottofondo in cocciame (Fig. 4, n.20; Cazzella et al. 2003, Fig. 5); un possibile secondo piano di cottura adiacente, molto mal conservato, sembra non essere contemporaneo in senso stretto. La circostanza che in seguito siano stati accesi fuochi nel medesimo spazio, ancora nel corso dell’Appenninico Recente, sia pure in assenza di focolari strutturati (Fig. 4, n.21; Cazzella, Moscoloni 1993, Fig. 4), fa ipotizzare che ci si trovi di fronte ad attività condotte con qualche variazione nel tempo, ma comunque ripetute in quest’area, che in questo caso potrebbero eventualmente esulare dalla cottura dei cibi. Almeno due piani di cottura in argilla subcircolari, adiacenti, probabilmente con episodi di rifacimento nella medesima posizione, sono stati inoltre individuati all’interno di una struttura ellissoidale ancora ascrivibile all’Appenninico Recente e posta poco a sud-est dell’area della porta protoappenninica (Fig. 4, n.23; Cazzella et al. 2006, Fig.3). Tale struttura, delimitata da pietre a secco, potrebbe costituire semplicemente un recinzione, priva di copertura, di un’area dedicata nello specifico ad attività connesse con l’uso del fuoco, forse in questo caso legate al trattamento di beni alimentari.

Fig. 1. Coppa Nevigata, livelli del Protoappenninico (XVII-XVI sec. a.C.).

Strutture di combustione fisse, aperte, poste all’interno di ambienti coperti o meno La dislocazione dell’area principale di scavo a Coppa Nevigata, marginale rispetto a quella che doveva essere la parte centrale dell’abitato durante le diverse fasi 54

A fuoco lento: strutture di combustione nell’abitato dell’età del Bronzo di Coppa Nevigata

Table 1. Schema sinottico dei parametri ritenuti significativi nell’analisi delle strutture di combustione e dei relativi indicatori archeologici.

55

A. Cazzella, G. Recchia Com’è noto sia nell’ambito degli scavi Puglisi sia in quelli attualmente in corso sono state poste in luce diverse strutture, riferibili ai livelli subappenninici, generalmente a pianta rettangolare, verosimilmente interpretabili come abitazioni. All’interno di nessuna di queste, però, erano conservati piani di cottura o strutture di combustione, attestati invece, almeno in un caso, subito all’esterno; l’unica eccezione potrebbe sarebbe costituita dall’evidenza di una concentrazione di carboni di legno, forse impiegati come combustibile, in una porzione specifica del vano meridionale della struttura rinvenuta nei settori E4 degli scavi Puglisi in estensione (Fig. 5, n.28), che potrebbe testimoniare la presenza di un focolare non delimitato (Moscoloni et al. 2002).

base, attestati a Coppa Nevigata in particolare durante l’Appenninico Recente nell’area della porta tamponata (nei livelli formatisi per accumulo volontario al di sopra del piano di cottura già ricordato; Fig. 4, n.21); un altro esempio è rappresentato da un’area caratterizzata da carboni, tendenzialmente semicircolare, non delimitata, che utilizzava come sottofondo il piano della strada consolidata con pietrisco che nel Protoappenninico attraversava la porta di accesso all’abitato (Fig. 2, n.4; Cazzella et al. 2003, Fig.4). Nell’ambito dell’insediamento, durante le sue varie fasi di vita, tali soluzioni risultano comunque piuttosto rare. Una situazione in parte differente è rappresentata da una struttura circolare di piccole dimensioni, costituita da una leggera depressione con ciottoli alla base, senza un piano di copertura dei ciottoli stessi, protetta da un muretto e rinvenuta presso uno dei due ambienti sub-rettangolari ascrivibili al Subappenninico Recente Fig. 5, n.27; Cazzella et al. 2004, Fig. 1). Una struttura ugualmente circolare, circondata da pietre (diametro interno 80 cm ca.), ma senza una vera e propria base, venne messa in luce nel settore F degli scavi Puglisi in profondità, ed è ascrivibile all’Appenninico Iniziale (Fig. 3, n.14; Cazzella, Moscoloni 1987, Fig. 64). Si potrebbe ipotizzare che si tratti in questi casi di normali focolari destinati al riscaldamento della zona circostante e/o alla cottura diretta di cibi con fuoco vivo (ad esempio carni allo spiedo) o per mezzo di contenitori ceramici. Può risultare in contrasto con questa interpretazione la rarità di tale elemento, che si immagina dovesse essere ricorrente, ma che forse era effettivamente poco utilizzato all’interno dell’abitato per limitare i rischi di incendio. Più frequenti, non solo a Coppa Nevigata ma anche in altri siti coevi della Puglia (ad esempio Monopoli, Punta Le Terrare, Porto Perone1, Scalo di Furno, solo per citarne alcuni; Cinquapalmi 1995 e 1998; Lo Porto 1963 e 1986; Recchia, Radina 1998; Radina et al. 2002), appaiono i casi in cui fu realizzata una sistemazione del sottofondo con cocci e/o piccole pietre, rivestita con uno strato di fango (il termine “malta” è forse più adatto a situazioni in cui un preparato è usato come legante tra materiali costruttivi) più o meno accuratamente lisciato in superficie e cotto per l’esposizione al calore, definibili come piastre di cottura fisse. E’ probabile che tali struttture non fossero veri e propri focolari per un fuoco vivo, ma fossero destinate solo ad accogliere letti di braci; la cottura di cibi (sia sfarinati che porzioni carnee o pesce o vegetali) poteva poi avvenire anche a diretto contatto con il piano stesso, precedentemente surriscaldato. La presenza di frammenti ceramici nella preparazione della base, oltre che essere legata a una funzione di isolamento dall’umidità del terreno, potrebbe aver contribuito soprattutto alla restituzione del calore immagazzinato; la lisciatura del piano poteva favorire invece l’eliminazione accurata di braci e cenere, per consentire la cottura senza ulteriori contenitori. Non si può escludere dunque che un elemento di questo tipo, sia pure differenziato perché

Fig. 2. Coppa Nevigata, livelli del Protoappenninico Recente (XVI sec. a.C.).

Strutture di combustione fisse, aperte, esterne Le attrezzature di combustione poste all’esterno di strutture, coperte e non, sono molto più frequenti nel sito (anche per i motivi sopra accennati) e, probabilmente, da correlare con attività diverse, tra cui quelle stesse di cottura dei cibi potrebbero essere state svolte in più modi. Le testimonianze, benchè numerose, una volta suddivise tra le diverse fasi dell’insediamento sono comunque insufficienti per collegare con certezza alcuni “tipi” direttamente con momenti cronologici: nella trattazione la pertinenza al periodo sarà quindi esplicitata, ma questo avverrà solo nell’ambito delle varie categorie individuate. Una prima distinzione può essere fatta in base al tipo di sottofondo: si possono citare i semplici punti di fuoco, in cui questo venne acceso senza che si avessero chiari elementi di delimitazione né preparazioni consistenti della 56

A fuoco lento: strutture di combustione nell’abitato dell’età del Bronzo di Coppa Nevigata fisso, svolgesse una funzione simile a quella delle piastre di cottura mobili, non molto frequenti a Coppa Nevigata. A Vivara, Punta d’Alaca, è presente una soluzione intermedia, con più piastre accostate, poste a formare una struttura di fatto fissa (Tusa 1991, pp. 85-86 e Fig. 63). Gli stessi contenitori ceramici di forma bassa e larga (altrimenti definiti “teglie”)2, inoltre, potrebbero attestare un sistema di cottura non diverso rispetto alla piastra/ piccolo piano fisso, che viene fortemente riscaldato e poi destinato ad accogliere l’alimento da cuocere; naturalmente le “teglie”, in quanto elementi mobili, potevano in alternativa essere state esposte al calore all’interno di un forno o soltanto al di sopra di un letto di brace o, più difficilmente, di un piano di cottura. (A.C.)

ogni caso quest’area come destinata ad attività legate alla cottura. Forse confrontabili e correlabili, data la distanza non eccessiva, potrebbero essere inoltre due piani, rinvenuti durante gli scavi Puglisi, ascrivibili alla stessa fase, uno alloggiato all’estremità sud-orientale della postierla messa in luce nel 1955 (Fig. 3, n. 16) e l’altro addossato alla faccia interna del residuo delle mura protoappenniniche, nel settore H (Fig. 3, n. 17). A suo tempo questi vennero interpretati da chi eseguì lo scavo come parti basali di “forni a cupola”, ma potrebbero invece rientrare nell’ampia categoria dei piani di cottura. Delimitazioni su uno o più lati, presenza in un’area relativamente concentrata e forma poco attestata in altri casi a Coppa Nevigata (dove i piani di cottura sono tendenzialmente subcircolari) costituiscono i loro caratteri distintivi: è difficile stabilire se si tratti di strutture destinate alla cottura dei cibi, presumibilmente svolta in modo collettivo, o a qualche altra attività di interpretazione problematica, condotta comunque contemporaneamente da più persone. Per quanto riguarda i piani concotti connessi con l’uso del fuoco, privi di delimitazione in elevato, la differenza fondamentale è costituita dall’ampiezza, cui si accompagna il contesto generale in cui si pongono. Strutture di dimensioni medio-piccole, di forma subcircolare3, si ritrovano in posizioni tra loro relativamente differenziate. Un piano di cottura subellittico di dimensioni medie (1,5x0,80 m), relativo al Protoappenninico tardo, fu ricavato riutilizzando come base un forno a cupola defunzionalizzato posto nell’area esterna alle prime mura di fortificazione (Fig. 2, n.5; Cazzella, Moscoloni 1999b, Fig. 5). Parte di una struttura analoga, di dimensioni mediopiccole (diametro ca. 80 cm), riferibile all’Appennninico Iniziale, è stata individuata nel settore F degli scavi Puglisi (Fig. 3, n. 13; Cazzella, Moscoloni 1987, Fig. 64), non lontano dalla piastra subrettangolare delimitata da pietre, sopra ricordata, e da quella che sembra essere la base di un vero e proprio “forno a cupola” (Fig. 3, n.12). Un piano di cottura, riferibile all’Appenninico Recente (diametro ca. 60 cm), è stato rinvenuto in anni recenti nello spazio residuo tra le mura difensive dello stesso periodo e il fossato4 (Fig. 4, n.19), facendo pensare che non fosse destinato alla normale cottura di cibi, ma a qualche attività (forse ugualmente nell’ambito delle preparazioni alimentari), eventualmente temporanea, realizzata in uno spazio fortemente marginale dell’abitato. Un piano di dimensioni piuttosto piccole (dim. massima conservata ca. 30 cm), risalente alla stessa fase, è stato messo in luce all’interno dell’abitato (Fig. 4, n.22), nell’ambito di un’area ancora in corso di scavo che appare interessata, anche se con lievi scarti temporali, dalla presenza di diverse strutture di combustione, come le piastre inserite nella struttura ellissoidale sopra ricordata, relative ad un livello sottostante il piano dove era alloggiata la piccola piastra in questione. Queste attrezzature hanno dimensioni piuttosto variabili, anche se complessivamente non grandi, e sono dissimili tra loro anche per il contesto in cui sono collocate: di volta in volta possono dunque essere connesse con attività diverse, a differenza delle strutture

Fig. 3. Coppa Nevigata, livelli dell’Appenninico Antico (XV sec. a.C.).

Gli elementi principali di distinzione tra le piastre di cottura stabili possono essere rappresentati dalla forma, dalle dimensioni e dalla presenza o meno di elementi fissi di delimitazione in elevato. Questi ultimi sono rari a Coppa Nevigata: il caso meglio documentato, riferibile all’Appenninico Iniziale, è costituito da un piano quadrangolare (di circa 50 cm di lato), definito da pietre poste di taglio su due lati e con il terzo lato appoggiato al residuo dell’alzato delle mura protoappenniniche ancora visibile (Fig. 3, n.11; Cazzella, Moscoloni 1997, Fig. 4). Esso trova, tra l’altro, un preciso confronto con un piano rettangolare simile nella forma e nella delimitazione in elevato da Punta le Terrare (Recchia, Radina 1998, Figg. 516 e 518). A Coppa Nevigata accanto a tale piano ne sono stati rinvenuti altri, purtroppo fortemente danneggiati, di cui uno in posizione simile e quindi con almeno un lato rettilineo: la loro concentrazione lascia pensare che si trattasse di un insieme di strutture, alcune delle quali forse in uso contemporaneamente, che caratterizzano in 57

A. Cazzella, G. Recchia di maggiori dimensioni che, invece, tendono ad essere più ripetitive. Le strutture più grandi (diametro 1,5-2 m) pongono a loro volta problemi di diverso tipo. Documentate in particolare per la fase finale del Protoappenninico nell’ambito degli scavi in corso (Fig. 2, nn.7 e 9; Cazzella et al. 2001, Fig. 6), nelle adiacenze di una coppia di forni a cupola, e per quella iniziale dell’Appenninico dagli scavi Puglisi (Fig. 3, n. 15; Cazzella, Moscoloni 1987, Fig. 64). queste, rinvenute anche a gruppi, condividono con quelle più piccole il piano ben lisciato, realizzato in argilla o in terra battuta, e appiano in taluni casi delimitate da una canaletta più o meno profonda. Per la presenza di tracce di fuoco più ampiamente diffuse nell’area dove esse si collocano non si può essere del tutto certi che si tratti di strutture di combustione; qualora quest’ultima ipotesi fosse confermata, però, le dimensioni ampie e la presenza di resti di cereali carbonizzati potrebbero far ipotizzare che fossero utilizzate, riscaldando la superficie, per tostare i cereali stessi, al fine di consentirne una migliore conservazione nel tempo. Ai livelli dell’Appenninico Recente esplorati durante gli scavi Puglisi sono riferibili due strutture sub-circolari, confrontabili con le ultime citate per dimensioni (diametro intorno ai 2 m, anche se una delle due sembra essere di forma ovaleggiante) e presenza di un piano battuto esposto al calore, ma che si differenziano da esse per una delimitazione in elevato, almeno parziale, realizzata con una fila di pietre e in parte utilizzando la faccia interna delle mura contemporanee (Fig. 4, nn. 25 e 26; Cazzella, Moscoloni 1987, Fig. 66). Lembi di piani di cottura riferibili al al Protoappenninico (Fig. 2, n.10: in questo caso la posizione subito al di fuori

delle prime mura protoappenniniche ricorda quella dei piani di ampie dimensioni sopra citati), all’Appenninico Iniziale (Fig. 3, n.18) e al Subappenninico Iniziale (Fig. 5, n.30), sono stati rinvenuti nel corso degli scavi Puglisi 1972-75 nell’area sud-occidentale del sito. Lo stato di conservazione piuttosto limitato non ne rende possibile definire forma e dimensioni e proporne dunque un inquadramento specifico nelle diverse categorie proposte. Un aspetto a sé stante, apparentemente di più facile interpretazione, è costituito dai forni a cupola. Questi si ritrovano in tutte le fasi con caratteristiche qualitative e dimensionali simili e sono stati rinvenuti sempre in aree “esterne”, benchè la loro collocazione in specifiche zone dell’abitato o in relazione a singole strutture, quando identificabile, sembra essere piuttosto diversificata. Un esemplare (Fig. 1, n.2; Cazzella, Moscoloni 1999b, Fig. 2), attribuibile alla fase di occupazione protoappenninica del sito dopo la costruzione delle prime mura di fortificazione, si trova all’interno dell’insediamento presso la probabile struttura abitativa a pianta quadrangolare (sopra ricordata per le tracce di un possibile “focolare” interno; Fig. 1, n.1). Almeno dietro la metà posteriore di questo forno si conservano tracce di un elemento che lo racchiude, forse la base di un muretto, di cui non è del tutto chiara la funzione: consentire una copertura dell’area di uso del forno per le stagioni meno clementi? definire la pertinenza alla struttura abitativa? A una fase finale del Protoappenninico, nell’area al di fuori delle prime mura sopra citata per altre strutture di combustione, sono stati rinvenuti due forni ravvicinati (Fig. 2, nn.5 e 6, Cazzella et al. 2001, Figg. 3 e 4). La posizione periferica e la duplicazione della struttura potrebbero far pensare che in questo caso non si tratti della normale cottura del cibo per una famiglia nucleare, ma per un gruppo più ampio o di qualche attività di trasformazione dei prodotti, sia pure alimentari. A suo tempo Dennell (1972), per contesti neolitici della penisola balcanica, propose che forni simili servissero per tostare i cereali. La stessa funzione qui è stata invece suggerita, anche se in maniera molto dubitativa, per probabili strutture di combustione circolari di dimensioni maggiori rinvenute nelle immediate vicinanze dei forni accoppiati stessi, per cui è verosimile che almeno in uno dei due casi l’ipotesi debba essere rifiutata, a meno che le diverse tipologie strutturali non fossero funzionali a fasi differenti del medesimo procedimento. Va notato che in un momento successivo, come accennato, uno dei due forni in disuso divenne la base per realizzare un piano di cottura di forma ellissoidale. Un altro forno è stato messo in luce in relazione con i livelli dell’Appenninico Antico in posizione simile a quella della coppia appena descritta, subito al di fuori delle prime mura, ma più a sud-ovest, nel settore F degli scavi Puglisi (Fig. 3, n.12; Cazzella, Moscoloni 1987, Fig. 64), in posizione non lontana anche in questo caso dalla coeva struttura circolare di grandi dimensioni con tracce di bruciato, delimitata da

Fig. 4. Coppa Nevigata, livelli dell’Appenninico Recente (XIV sec. a.C.).

58

A fuoco lento: strutture di combustione nell’abitato dell’età del Bronzo di Coppa Nevigata una canaletta sopra ricordata. L’unico forno per ora individuato nei livelli dell’Appenninico Recente, sempre nell’ambito degli scavi Puglisi, si trova invece appoggiato alla faccia interna delle mura riferibili a questo periodo, in una situazione comunque marginale all’abitato e che poteva facilmente consentire una copertura dell’area ad esso adiacente (Fig. 4, n.24; Cazzella Moscoloni 1987, Fig. 66). Il forno meglio documentato è quello riferibile alla fase terminale del Subappennico Recente (Fig. 5, n.29 e Fig. 6), situato in un’area aperta, per la quale è stata a suo tempo realizzata la raccolta dei reperti per metro quadrato. La distribuzione e la selezione dei resti di fauna, in particolare, hanno fatto ipotizzare che il forno potesse servire anche per la cottura di pietanze a base di carne precedentemente trattata (Moscoloni et al. 2002). L’area in cui si trova, al di là della possibile copertura, è aperta, ma adiacente alle tracce di un piano battuto quadrangolare probabilmente da ricollegare con la presenza di una struttura abitativa. In una lettura sintetica dei dati fin’ora disponibili sulla localizzazione spaziale dei forni si possono distinguere due situazioni tendenziali: la presenza di un singolo forno presso specifiche strutture (forse di carattere abitativo) e la dislocazione di uno o due forni ravvicinati nell’ambito di aree, marginali all’abitato, caratterizzate anche dalla presenza di altre strutture di combustione di tipologia differente.

Fig. 5. Coppa Nevigata, livelli del Subappenninico (XIII – XII sec. a.C.).

fosse destinato soltanto ad attività di tipo artigianale, come la metallurgia, ma avesse uno spettro di utilizzi molto più ampio. Se questa impressione non dipende solo dalla marginalità prevalente delle aree scavate, l’uso intensivo del carbone di legna potrebbe rientrare tra gli espedienti posti in atto per un uso controllato del fuoco all’interno di un insediamento con un tessuto presumibilmente denso, dove il fuoco vivo costituiva un elemento di pericolo.

Strutture di combustione mobili Un rapido accenno si può infine fare alle attrezzature per l’uso del fuoco mobili e in particolare ai fornelli. A Coppa Nevigata sembrano essere attestati a partire dall’Appenninico, forse già da un momento iniziale, anche se si tratta di piccoli frammenti per i quali non è certa l’attribuzione a tali elementi, per i quali indicazioni più evidenti si hanno dalle fasi successive, in particolare nel Subappenninico. Sembrano mancare a Coppa Nevigata i forni mobili a due camere, proposti per Torre Mordillo e S. Maria di Ripalta (Moffa 2002), la cui interpretazione lascia tuttavia perplessi, dal momento che non è chiaro come fosse possibile accedere alla camera in cui erano posti gli elementi da cuocere. Nel caso dell’esemplare di Torre del Mordillo si potrebbe alternativamente pensare (capovolgendo la ricostruzione) a un grande contenitore sub-troncoconico per filtrare liquidi, appoggiandovi sopra un elemento in mateiale deperibile a maglie piccole, con uno scolo in basso. Anche nel caso dei fornelli è possibile che l’alimentazione avvenisse per mezzo di carbone di legna e non con fuoco vivo. Questa ipotesi, unita a quella che si utilizzasse il medesimo tipo di combustibile per i piani di cottura sopra ricordati, nettamente più frequenti rispetto ai focolari veri e propri e ai punti in cui vi sono tracce di accensione del fuoco, potrebbe far pensare che a Coppa Nevigata durante l’età del Bronzo la produzione del carbone di legna non

Conclusioni Il contesto più generale e il rapporto spaziale ravvicinato tra più strutture di combustione, in alcuni casi già citato, possono fornire indizi sull’uso complessivo di intere aree in cui queste si collocano. In particolare va tenuto presente, come evidenziato all’inizio, lo scavo ha interessato soprattutto zone marginali dell’abitato, vicine alle mura, almeno fino al Subappenninico, quando il sistema di fortificazioni è meno definibile, al di là del persistere del fossato realizzato nell’Appenninico Recente; per quest’ultima fase, nell’area oggetto di scavo, sono state invece rinvenute diverse strutture di cui alcune probabilmente di carattere abitativo. Proprio perché si tratta di zone marginali nei momenti precedenti, è possibile che strutture di combustione di diverso tipo, un elemento di pericolo all’interno dell’abitato, fossero qui concentrate. Fenomeni simili sembrano riscontrarsi in altri insediamenti pugliesi coevi, come ad esempio Masseria Chiancudda e Punta Le Terrare (Cinquepalmi et al. 2004; Radina et al. 2002). 59

A. Cazzella, G. Recchia Note Anche se ormai è opinione comune, vale la pena ricordare come le strutture dello strato “e” definite “capanne” con sottofondo in “potsherds pavements” dall’autore dello scavo possano essere riferite a strutture di combustione, sia per le dimensioni, troppo ridotte per essere ambienti, sia per le caratteristiche di realizzazione. 2 Tali elementi sono piuttosto diffusi a Coppa Nevigata (in particolare a partire dai livelli subappenninici) così come in altri insediamenti coevi dell’Italia meridionale (per un esempio ed una più ampia descrizione funzionale: Recchia 1997, Fig.23). 3 Sebbene la forma spesso non sia regolare, a causa dell’assenza di delimitazioni, sembra comunque difficile che tali elementi derivino sempre da piani in origine quadrangolari, degradatisi. 4 Tale spazio risulta dunque esterno all’abitato vero e proprio ma ancora “protetto” dal fossato, e veniva evidentemente utilizzato per lo svolgimento di alcune attività. 1

Fig. 6. Coppa Nevigata, forno a cupola del Subappenninico Recente dagli scavi Puglisi in estensione.

In sintesi, riguardo le diverse strutture concentrate in aree marginali, si possono ricordare in ordine cronologico: per un momento finale del Protoappenninico, i due “forni a cupola” (di cui uno trasformato in piano di cottura), la struttura con due piani di cottura all’interno e, forse, i piani circolari più ampi, qualora siano effettivamente connessi con l’uso del fuoco (Fig. 2). Per l’Appenninico Iniziale l’insieme dei piani di cottura di forma almeno in parte quadrangolare, il “forno a cupola”, il “focolare”, il piano di cottura circolare e forse il grande piano circolare delimitato da canaletta degli scavi Puglisi (Fig. 3). Per l’Appenninico Recente il “forno a cupola” e forse i piani delimitati da pietre degli scavi Puglisi, il piano di cottura e in seguito i punti di fuoco nell’area della porta protoappenninica tamponata e forse alcune concavità con tracce sporadiche di fuoco, ad essa adiacenti, le due piastre inserite nella struttura ellittica e la piccola piastra di cottura rinvenuta nella stessa area posta poco più all’interno dell’abitato (Fig. 4). Anche se gli indizi sono estremamente limitati, qualora tutte queste strutture fossero connesse con la cottura dei cibi o comunque con preparazioni alimentari, si potrebbe pensare che, come sembra avvenire per altri aspetti (ad esempio la conservazione dei cereali), tali attività avessero un carattere più collettivo, probabilmente a livello di famiglia estesa, nel Protoappenninico finale e nei momenti immediatamente successivi, e viceversa più ristretto, in via d’ipotesi a livello di famiglia nucleare, sia in una fase precedente del Protoappenninico stesso, che in un momento avanzato del Subappenninico. Per questi ultimi periodi, infatti, le attrezzature di combustione appaiono essere tendenzialmente più collegate a singole strutture, forse di carattere abitativo: alle prime fasi del Protoappenninico, ad esempio è relativo il forno a cupola adiacente la struttura rettangolare all’interno dell’insediamento (Fig. 1), mentre per il Subappenninico (Fig. 5) si possono ricordare il focolare presso una struttura rettangolare dagli scavi in corso, le evidenze della struttura bicellulare in E4 e il forno da pane in D5 nell’ambito degli scavi Puglisi in estensione. (G.R.)

Bibliografia Cazzella A., Minniti C., Moscoloni M., Recchia G., 2005. L’insediamento dell’età del Bronzo di Coppa Nevigata (Foggia) e la più antica attestazione della produzione della porpora in Italia, Preistoria Alpina, suppl. 1, v. 40 (2004), Trento, pp. 177-182. Cazzella A., Moscoloni M., 1987. Età del Bronzo. La ricerca archeologica, in Cassano S.M., Cazzella A., Manfredini A., Moscoloni M. (a cura di), Coppa Nevigata e il suo territorio, Quasar, Roma, pp. 109-190. Cazzella A., Moscoloni M., 1993. Nuovi dati sui livelli dell’età del Bronzo di Coppa Nevigata, in Gravina A. (a cura di), Atti del 13° Convegno Nazionale sulla Preistoria, Protostoria, Storia della Daunia, San Severo, pp.55-65. Cazzella A., Moscoloni M., 1997. Gli scavi nell‘insediamento dell‘età del bronzo di Coppa Nevigata: nuovi risultati, in Gravina A. (a cura di), Atti del 15° Convegno Nazionale sulla Preistoria, Protostoria, Storia della Daunia, San Severo, pp. 103-116. Cazzella A., Moscoloni M., 1999a. Coppa Nevigata: i risultati degli scavi in estensione 1983-97, in Tunzi A.M. (a cura di), Gli ipogei del Tavoliere e la preistoria in Daunia, Grenzi, Foggia-Napoli, pp. 102-107. Cazzella A., Moscoloni M., 1999b. Coppa Nevigata: campagna di scavo 1998, in Gravina A. (a cura di), Atti del 19° Convegno Nazionale sulla Preistoria, Protostoria, Storia della Daunia, San Severo, pp. 143-156. Cazzella A., Moscoloni M., Recchia G., 2001. Coppa Nevigata: campagna di scavo 1999 e 2000, in Gravina A. (a cura di), Atti del 21° Convegno Nazionale sulla Preistoria, Protostoria, Storia della Daunia, San Severo, pp. 153-170. Cazzella A., Moscoloni M., Recchia G., 2003. Coppa Nevigata: campagne di scavo 2001 e 2002, Gravina A. (a cura di), in Atti del 23° Convegno Nazionale sulla Preistoria, Protostoria, Storia della Daunia, San Severo, pp. 201-213. Cazzella A., Moscoloni M., Recchia G., 2004. Coppa Nevigata, in Cocchi Genick D. (a cura di), L’età del Bronzo Recente in Italia, Octavo, Firenze. Cazzella A., Moscoloni M., Recchia G., 2006. Gli scavi 20032005 nell’insediamento fortificato dell’età del Bronzo di

60

A fuoco lento: strutture di combustione nell’abitato dell’età del Bronzo di Coppa Nevigata dell’età del Bronzo. Ricerche lungo il versante adriatico pugliese, Schena, Fasano, pp. 185-195. Tusa S., 1991. Scavi alla Punta d’Alaca, in Cazzella A., Damiani I., di Gennaro F., Marazzi M., Moscoloni M., Pacciarelli M., Rosi M., Sbrana A., Tusa S., Vezzoli L. (a cura di), Vivara I, Bagatto Libri, Roma, pp.75-93.

Coppa Nevigata (Manfredonia), in Gravina A. (a cura di), Atti del 26° Convegno Nazionale sulla Preistoria, Protostoria, Storia della Daunia, San Severo, pp. 113-124. Cazzella A., Recchia G., Baroni I., Minniti C., 2002. Coppa Nevigata: analisi dell’uso dello spazio in una struttura protoappenninica, in Peretto C. (a cura di), Analisi informatizzata dei dati delle strutture di abitato di età preistorica e protostorica in Italia, Origines, Istituto Italiano di Preistoria e Protostoria, Firenze, pp. 427-442. Cinquepalmi A., 1995. L’insediamento protostorico di Monopoli: ricerche in Piazza Palmieri e via Papacenere, Taras, XV, pp. 313-333. Cinquepalmi A., 1998. Monopoli centro storico - le ricerche in Piazza Palmieri: i livelli inferiore e medio, in Cinquepalmi A., Radina F. (a cura di), Documenti dell’età del Bronzo. Ricerche lungo il versante adriatico pugliese, Schena, Fasano, pp. 109-124. Cinquepalmi A., Fiorentino G. Recchia G. 2004. Masseria Chiancudda (Cisternino, Prov. di Brindisi), Rivista di Scienze Preistoriche, LIV, pp. 631-632. Dennel R.W., 1972. The interpretation of plant remains: Bulgaria, in Higgs E.S. (ed.), Papers in Economic Prehistory, Cambridge University Press, Cambridge, pp. 149-159. Lo Porto F.G., 1963. Leporano (Taranto). La stazione protostorica di Porto Perone, Notizie degli Scavi di Antichità, 88, pp. 280-380. Lo Porto F.G., 1986. Le importazioni micenee in Puglia. Bilancio di un decennio di scavi, in Marazzi M., Tusa S., Vagnetti L. (a cura di), Traffici micenei nel Mediterraneo, Palermo, pp. 13-20. Moffa C., 2002. L’organizzazione dello spazio sull’acropoli di Broglio di Trebisacce. Dallo studio delle strutture e dei manufatti in impasto di fango all’analisi della distribuzione dei reperti, Grandi contesti e problemi della Protostoria italiana, 6, All’insegna del Giglio, Firenze. Moscoloni M. Recchia G. Baroni I. Minniti C., 2002. Coppa Nevigata: analisi funzionale delle strutture subappenniniche dei settori E4 e D5 (scavi PuglisiPalmieri in estensione), in Peretto C. (a cura di), Analisi informatizzata dei dati delle strutture di abitato di età preistorica e protostorica in Italia, Origines, Istituto Italiano di Preistoria e Protostoria, Firenze, pp. 443-465. Puglisi S.M., Coccolini G.B.L., 1982. Coppa Nevigata (Manfredonia - Foggia), in Vagnetti L. (a cura di), Magna Grecia e mondo miceneo. Nuovi documenti, Taranto, pp. 45-52. Radina F. Baroni I., Minniti C., Recchia G. 2002. L’uso dello spazio nell’ambito di un’area aperta: il caso dei livelli appenninici di Punta le Terrare (BR), scavi 1966, 1969 e 1981, in Peretto C. (a cura di) “Analisi informatizzata e trattameno dati delle strutture di abitato di età preistorica e protostorica in Italia”, Origines, Istituto Italiano di Preistoria e Protostoria, Firenze, pp. 411-426. Recchia G., 2001. Archeologia della vita: funzione dei vasi ed aree interne all’abitato. Un esempio da Coppa Nevigata, in Gravina A. (a cura di), Atti del 21° Convegno Nazionale sulla Preistoria, Protostoria, Storia della Daunia, San Severo, pp. 245-252. Recchia G., 1997. L’analisi degli aspetti funzionali dei contenitori ceramici: un’ipotesi di percorso applicata all’età del Bronzo dell’Italia meridionale, Origini, XXI, pp. 207-306. Recchia G., Radina F., 1998. Punta le Terrare: il saggio A 1969, in Cinquepalmi A., Radina F. (a cura di), Documenti

61

An approach to Holocene vegetation history in the middle Rhone valley (France): anthracological data from the « TGV-Méditerranée » excavations C. Delhon, S. Thiébault

UMR 7041 ArScAn, MAE, 21 allée de l’Université, 92023 Nanterre, France, E-mail: [email protected] Abstract The construction of the « TGV-Méditerranée » railway, connecting Paris to the Mediterranean, has allowed large-scale rescue excavations to take place throughout the middle Rhone valley. These provided wood charcoal fragments not only from archaeological settlements but also from off-site sequences. Their analysis has allowed us to propose a high resolution reconstruction of the vegetation history, that has proved to be quite different from the pattern previously admitted for the karstic peripheral hills. An early colonisation of the area by deciduous oak is attested during the Late Glacial interstades. The Preboreal and Boreal are marked by high proportions of pine before deciduous oak formations became definitively dominant. If the Middle Neolithic experiences a specialised management of the landscape by agro-pastoral populations, the vegetation that settles at the transition with the Bronze Age seems to be more influenced by a climatic change towards more humid conditions. The early Bronze Age is marked by a fall-off in agricultural activities together with a probable climate cooling. On the contrary, the final Bronze Age and even more the Iron Age are periods of vegetation diversity due to the increasing human impact on the landscape, while the climatic context gets dryer. The Classical period is a time of strong anthropisation of the landscapes favouring vegetation biodiversity. During the early Middle Ages, deciduous oak forests partly regenerate thanks to a short decrease of agricultural activities. During the late Middle Ages, cooling tracers tend to decrease while anthropogenic markers tend to increase. The modern period seems to be marked by a loss of biodiversity, following a fall-off in agricultural activities. Most of the lost taxa reappear during the sub-contemporary period, maybe because of a climatic improvement and of a reorganisation of agricultural lands towards vine monoculture, beneficial to “garrigue” species.

Introduction From 1994 to 1997, the construction of the « TGV-Méditerranée » railway, connecting Paris to the Mediterranean, allowed large-scale rescue excavations through out the middle Rhone valley. These have provided palaeoenvironmentalists an exceptional south-north transect crossing every stage of vegetation, from the lower mesomediterranean vegetation stage to the mountain vegetation stage. Thanks to the coordinators of the project, who expressed the will that palaeoenvironnents should be documented as well as archaeological settlements, palaeobotanical remains have been largely sampled, not only on the archaeological sites but also in numerous “off-site” pedosedimentary sequences. Such a sampling strategy has allowed us to determine more than 40 000 charcoal fragments, distributed among 1000 assemblages from several dozens of different sites and offsite sequences and to propose a first approach to the vegetation dynamics during the last 15 millennia. Biogeographical context The studied area is located between the towns of Orange (south) and Valence (north), along the river Rhone (Fig. 1). It mainly concerns the alluvial plains of the Rhone and its tributaries, but also involves hills of the pre-alpine piedmonts. The valley is delimited to the west by the slopes of the Cévennes hills (southern Massif Central) and to the

east by the pre-Alps of the Diois and the Baronnies. The Middle Rhone valley is located in an ecotone situation, straddling the northern limit of the Mediterranean climatic zone, at the boundaries of the Medio-European and Alpine climatic influences. This location makes the vegetation of the middle Rhone valley highly reactive to both climatic changes and anthropogenic pressure. This area belongs to a humid sub-Mediterranean bioclimate, marked by highly irregular precipitation and temperatures, associated mainly with the supra-Mediterranean vegetation stage (Ozenda 1975). The Mediterranean dryness (two months of water deficit in summer) is still marked, but the annual pluviometric average appears rather abundant (between 700 and 1000 mm). Rain concentrates over two seasons (autumn and spring). Because of the various climatic influences, of microclimatic conditions related to various topographies (plains, slopes, hills and terraces), and of strong and ancient human pressure, the landscape of the middle Rhone valley presents a characteristic vegetation mosaic. Nowadays, the forests are mainly constituted by oak formations, dominated by Quercus cf. pubescens and pine forests (Pinus sylvestris). The transitional aspect of this area is well recorded in the vegetation which shows a clear substitution of Pinus halepensis by Pinus sylvestris and Quercus ilex by Quercus pubescens at the north of the Donzère defile (Fig. 1). That substitution corresponds to the passage from a mesomediterranean vegetation stage to a supramediter-

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

C. Delhon, S. Thiébault

GEOSYSTEMS

Romans Bourg de Péage

L 'I s è r e

Alixan L a B a rbe r o tte

St-Marcel-les-Valence

L e G uima

n

et

d

Montèlier

L e Mia la n

e V é or

Chabeuil

La

Valence

Plain of Valence

Châteauneuf-sur-Isère

Rh ône

Montvendre

Le

Montmeyran

u yrie

Upie

x

TGV-Méditerranée

Vaunaveysla-Rochette

Eurre

Main sampled sites or sequences

La

èz e uv

Loriol

Crest

Chabrillan

Gr

Divajeu

r R if N oi

e e tt en

O

L a D rô me

Lower Drome valley

L 'E

La Roche sur Grane

Marsanne

Roynac

Le R hô

n oubio Le R

Bonlieu

ub ion

an

so

n

La Laupie Sauzet

Valdaine

ne

L ' A nc e lle

M

Le

L e J a bron

Le

Ro

Montélimar

Montboucher-sur-Jabron

Le Teil

Espeluche La

C ite ll e

Allan L a R ia ille

Donzère defile

Malataverne La

Be

rre

Roussas

Donzère

Les Granges-Gontardes La Garde Adhèmar Pierrelatte

Saint-PaulTrois-Châteaux

Le

C an

al M

Bourg-Saint-Andéol

on-D ra g o nd

Bolléne

Lau zon

L ' A rd è ch e

re onzè

Lapalud

L'Hrin

Lez Le

Tricastin

L e L ez

ôn Rh

e

Lamotte du Rhône Mondragon

Mornas Vénéjan

es ygu L'A

Orange L e R hône

La Cze

800 m 600 m 400 m 200 m

0

5km

G. Macabéo

Fig.1. Localisation of the studied area in the middle Rhone valley.

64

Caderousse

Plain of Orange

Piolenc

Saint-Etienne des Sorts

An approach to Holocene vegetation history in the middle Rhone valley

Occurences:

presence of a taxon in a sample = 1 absence of a taxon in a sample = 0 nb. of fragments Taxon A Sample 1 Sample 2

10 0

Sample 3

12

Sample 4

0

nb. of occurences of each taxon for the period % of occurences of each taxon for the period*

1 0 1 0

Taxon B

Taxon C

15

0

1

1

2 20%

1 1

50

number of occurences per sample

nb. of occurences

1 1

0 8 0

4 40%

0 0 1 0

1 10%

(= number of taxa per sample)

Taxon D

1 3

1

3

1

2

1

4

0

1

1 0

3 30%

10

Total number of occurences

(= base sum for the calculation of %)

* the % calculated on less than 30 occurences are indicated in grey on the diagram (figure 4)

Fig. 2. Principle of the calculation based on occurrences.

ranean vegetation stage. In the northern part of the studied area and sometimes on the peripheral heights, mountain vegetation characterised by beech forests replaces oak formations. The middle Rhone valley can be divided into several smaller geosystems (figure 1). The plain of Orange is the single one that currently belongs partly to the lower mesomediterranaean stage. Holm oak (Quercus ilex) is commonly found in that area. The Tricastin is included in the upper mesomediterranean vegetation stage; Holm oak is still present, but Mediterranean pubescent oak (Quercus pubescens) formations dominate. In the Valdaine area, the vegetation clearly belongs to the supramediterranean stage, except the highest parts of the hills where mountain vegetation formations tend to prevail. The mountain influences increase with the latitude, but even in the plain of Valence the vegetation stays mainly of supramediterranean type.

or dry sifting, most often using a 2 mm mesh, according to usual procedures (Chabal et al. 1999). In the case of offsite sampling, charcoal has been extracted from sediment by sifting through a 500 µm mesh. Every stratigraphic unit has been sampled, even if no macroscopic charcoal was noticed at first sight (Delhon 2005). Identification Even though the fragments are in average far smaller than those recovered from archaeological sites, “off-site” charcoal (also called pedo-charcoal) is identified following the same protocol. The fragments are manually broken and the characteristic planes are observed under a reflected-light microscope in order to compare the anatomical features with those described in wood anatomy atlases (Schweingruber 1990) and/or with those of charcoal from the reference collection (laboratoire d’Archéobotanique et de Paléo-écologie, MAE, Nanterre). In spite of the small size of pedo-charcoal fragments, it is usually possible to determine them at least to the family level, but most of the time to the genus level (sometimes species can even been identified). In the study presented here charcoal from drainage ditches dated from Antiquity were partly identified by L. Rousseau (Rousseau 1996); charcoal from most of medieval levels were identified by L. Fabre (Delhon et al., forthcoming).

Method Sampling Charcoal fragments have been recovered from archaeological settlements or from off-site sequences. In the first case, they were sampled in the archaeological structures by wet 65

66

15000

10000 11000

9000

6700

6000

5000

3900

3600

3200 3300

2600

2400

2000

1600

1000

400

200

Subatlantic

late

Subboreal

early Neo.

middle Neo.

recent/final Neo.

Bell Beakers

early Bronze Age

lack of data

final Bronze Age

first Iron Age

second Iron Age

Classical period

early M. A.

early/late M.A.

late M.A.

modern period

4800

3500

2500

2200

1400 1700

800

500

120

AD

400

1000

cc An Mil

1500

1700

cal BC

0

20

30

40

deciduous Quercus

10

50

60

70 20

Quercus sp.

10

30

10

20

10

20

10

10

20

10

10

Salix Ilex

Cor

Prun

ac Ros

Fig.3. Anthracological diagram of the TGV-Méditerranée based on the occurrences of taxa.

�P�enner P3�3� � ���$>�� '3P3� � �77�3 � -%��$��� � *��)($������� , J.E., �2004. Improving global estimates )��*���������)�)�*��������!��0��%�������"��&�)��$���5*���0������!���1�

/#�����%����� 123�����*������ � !��� � ����0�� � !��0 � $%��$��� � ��$��)� � �� � ������ � !������3� !��%�����������%����$�����($������!����������!����%����� �� -���)����P�(������!��������������$%�������9��/�9+93 �%� � 0��� � �!!�$���� � 2� � �� � *��$��) � !�� � 0�$���$�*�$� D�$.����)���3D3����������-3"3�3���F��$F��P3D3��D������A3�3���88�3��%�� $%��$�������� ��������%��$�(�������!�$%��$����!���0������ ��$�����($������!�'������������������������ �!��0���>����)�0�����/��� -�0*��������! �-%��$�����"���������)�0��������$������) ����$%�0�$��� 2��%�(���� ��(*��!�(�(�������0���(��0������������0����3� �)�5��3�R(������� ��$���$����&��2����7��+�/���?��%�(�)���� @)����)� � '3O3� � �88�3 � �%� � !��� � %����� � �! � .����% � %���%���) � ����� � ��� $�(���)��� ������3���2�&�����!��%������������:(��������&�� ��!��$��) � � � *����� � ��) � $%����) � *����$��� � �� � ��>� � ��)�0����3 � �%�� ����0�����!�������$%��$����$������������)�0����� �)�*������� ����$����������,/��93

W. Finsinger, W.Tinner, F.S. Hu of atmospheric emissions from biomass burning. Journal of Geophysical Research-Atmospheres, 109, D14. Lynch, J.A., Clark, J.S., Stocks, B.J., 2004. Charcoal productions, dispersal, and deposition from the Fort Providence experimental fire: interpreting fire regimes from charcoal records in boreal forests. Canadian Journal of Forest Research, 34, 1642-1656. MacDonald, G.M., Larsen, C.P.S., Szeicz, J.M., Moser, K.A., 1991. The Reconstruction of Boreal Forest Fire History from Lake Sediments - a Comparison of Charcoal, Pollen, Sedimentological, and Geochemical Indexes. Quaternary Science Reviews, 10, 53-71. Odgaard, B.V., 1992. The fire history of Danish heathland areas as reflected by pollen and charred particles in lake sediments. The Holocene, 2, 218-226. Patterson, W.A., Edwards, K.J., Maguire, D.J., 1987. Microscopic charcoal as a fossil indicator of fire. Quaternary Science Reviews, 6, 3-23. Sarmaja-Korjonen, K., 1998. Latitudinal differences in the influx of microscopic charred particles to lake sediments in Finland. The Holocene, 8, 589-597. Schröter, D., Cramer, W., Leemans, R., Prentice, I.C., Araujo, M.B., Arnell, N.W., Bondeau, A., Bugmann, H., Carter, T.R., Gracia, C.A., de la Vega-Leinert, A.C., Erhard, M., Ewert, F., Glendining, M., House, J.I., Kankaanpaa, S., Klein, R.J.T., Lavorel, S., Lindner, M., Metzger, M.J., Meyer, J., Mitchell, T.D., Reginster, I., Rounsevell, M., Sabate, S., Sitch, S., Smith, B., Smith, J., Smith, P., Sykes, M.T., Thonicke, K., Thuiller, W., Tuck, G., Zaehle, S., Zierl, B., 2005. Ecosystem Service Supply and Vulnerability to Global Change in Europe. Science, 310, 1333-1337. Swain, A.M., 1973. A history of fire and vegetation in northeastern Minnesota as recorded in lake sediment. Quaternary Research, 3, 383-396. Thevenon, F., Williamson, D., Vincens, A., Taieb, M., Merdaci, O., Decobert, M., Buchet, G., 2003. A late-Holocene charcoal record from Lake Masoko, SW Tanzania: climatic and anthropologic implications. The Holocene, 13, 785-792. Tinner, W., Conedera, M., 1995. Indagini paleobotaniche sulla storia della vegetazione e degli incendi forestali durante l’Olocene al Lago di Origlio (Ticino meridionale). Bollettino della Società Ticinese di Scienze Naturali, 83, 91-106. Tinner, W., Conedera, M., Ammann, B., Gäggeler, H.W., Gedye, S., Jones, R., Sägesser, B., 1998. Pollen and charcoal in lake sediments compared with historically documented forest fires in southern Switzerland since AD 1920. The Holocene, 8, 31-42. Tinner, W., Conedera, M., Gobet, E., Hubschmid, P., Wehrli, M., Ammann, B., 2000. A palaeoecological attempt to classify fire sensitivity of trees in the southern Alps. The Holocene, 10, 565-574. Tinner, W., Hu, F.S., 2003. Size parameters, size-class distribution and area-number relationship of microscopic charcoal: relevance for fire reconstruction. The Holocene, 13, 499505. Tinner, W., Hubschmid, P., Wehrli, M., Ammann, B., Conedera, M., 1999. Long-term forest fire ecology and dynamics in southern Switzerland. Journal of Ecology, 87, 273-289. Waddington, J.C.B., 1969. A stratigraphic record of the pollen influx to a lake in the Big Woods of Minnesota. Geological

Society of America, Special Paper, 123, 263-283. Whitlock, C., Larsen, C., 2001. Charcoal as a fire proxy. In: J.P. Smol, H.J.B. Birks, W.M. Last, eds. Terrestrial, Algal, and Siliceous Indicators. Tracking Environmental Change Using Lake Sediments. Dordrecht, The Netherlands: Kluwer Academic Publishers, 75-97. Wright, R.F., 1976. The impact of forest fire on the nutrient influxes to small lakes in northeastern Minnesota. Ecology, 57, 649-663.

124

Arts du feu et du forgeron en Mauritanie C. Fortier

CNRS, Laboratoire d’Anthropologie Sociale (CNRS-EHESS-Collège de France) E-mail: [email protected]

Abstract In the Moorish society of Mauritania, the blacksmiths constitute a scorned caste. Although the blacksmith has a protective magic knowledge, he, nevertheless, is associated with the demon because he handles the fire, matter constitutive of the démoniaques creatures which are the djinns, and because he, in addition, has a transformer activity of the matter similar to the woman, considered herself, in the Moorish society, like close relation of the devil. There is, in addition, a positive relation between iron and fire, since iron is worked by the blacksmith thanks to fire. Thus, one can put forth the assumption more general than iron, under the terms of its capacity to resist fire, became a quasi universal symbol of protection. And the handling by the blacksmith of the antagonistic forces that are iron and fire, makes of him a powerful and dangerous character, attributes explaining why he is maintained, in many societies, in a state of inferiorisation.

Introduction Les Maures de Mauritanie sont des bédouins ayant majoritairement abandonné le nomadisme. Ils sont musulmans et parlent un dialecte arabe nommé hassâniyya1. La société maure est extrêmement hiérarchisée; au sommet de cette hiérarchie, on trouve les marabouts et les guerriers, et à sa base, les tributaires, les anciens esclaves ainsi que deux castes constituées par les forgerons et les griots. Il s’agit bien de castes, même si le mot n’est pas employé localement, dans la mesure où ces groupes se distinguent par leur activité et leur endogamie, et sont par ailleurs méprisés. Un tel mépris concerne plus profondément les forgerons que les griots, ce qui tient en grande partie à leur métier qui les amène à manipuler le feu. Activité artisanale La division sexuelle du travail chez les forgerons est relative aux matériaux transformés par chaque sexe; alors que la femme est spécialisée dans le travail du cuir, l’homme est affecté à celui du bois et du métal. Le fait que les matières dures comme les métaux soient travaillées exclusivement par les hommes et les matières souples par les femmes, se retrouve dans de nombreuses sociétés, en raison, entre autres, de la différence de force physique attribuée aux deux sexes (Fortier 2002). La forge, alimentée par du charbon, consiste en une cavité creusée dans la terre, parfois surmontée d’un canari (gadra). Les forgerons distinguent différentes qualités de fer, depuis le fer blanc, qualifié de « pur » (hadid harr), jusqu’au « fer faible » (hadid lmar), en passant par le fer allié à un autre métal, dit « mélangé  » (hadid hartânî), terme qui renvoie également à la manière dont on désigne les anciens esclaves. L’outillage du forgeron consiste en un soufflet (hanût), une enclume (zabra), un socle d’enclume (ashagdal), un petit marteau (sammâra), un marteau (matrarga), une

masse (ma‘wan), un maillet (marzubba), un étau (zayyâr), un chalumeau (jabat an-nafakh), une cisaille (mgas), un compas ou une équerre (khattat), un rabot (makharta), un rabot plane (‘ablawna), une drille (‘âbla), une scie (gatta‘ ou majrad), une lime plate (mabrad), une lime triangulaire (munzalfa), un poinçon (ashubbâd), un poinçon pour le cuivre ou le cuir (marsham), une herminette (amâyk ou amâyyag), un fer à souder (kawya), un tournevis (balûalâb), une vis (lûlab), une pointe (musmâr), une hachette (gâdûm), une planche (jalfa), un burin (magta), une pelle (khubbi), une tenaille (kallâba), une pince (nattâr), des pincettes (khaylib), un crochet (mâlaga). Il y a encore une cinquantaine d’années, lorsque les Maures pratiquaient le nomadisme, les forgerons jouaient un rôle très important dans la vie nomade, produisant la plupart des objets usuels. Parmi les objets domestiques en fer, forgés par eux, on trouve la théière (barâd) ou la bouilloire (maghraj). Par ailleurs, avec les métaux (étain, cuivre), les forgerons confectionnent le harnachement (lakhzama et ars) des bêtes de monte, chameaux ou chevaux, ainsi que des cadenas finement ciselés qui permettent de fermer les sacs de voyage masculins (tasûfra) ou féminins (tîziyâtan). Le fer est aussi utilisé avec le bois de talh (Acacia raddiana) pour réaliser des outils, y compris ceux employés par le forgeron. Ces mêmes matériaux servaient autrefois à fabriquer différents types d’armes, du couteau jusqu’au fusil. Activité magique Le terme de ma‘lam (pl, ma‘lmin) qui désigne le forgeron dans la société maure et qui signifie littéralement «  celui qui sait » en arabe, semble faire référence à un savoir magique qui lui permet de fabriquer des objets ayant non seulement une fonction usuelle mais également prophylactique (Fortier 2006).

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

C. Fortier Ainsi, les motifs tracés par les forgeronnes sur les coussins en cuir des femmes, ou ceux gravés par les forgerons sur les bijoux féminins, n’ont pas seulement une signification « esthétique » mais également symbolique. Ils comportent en effet des symboles liés à la fécondité féminine tel le motif stylisé du triangle auquel s’ajoutent les lettres arabes waw et dal2. Il est caractéristique que ces signes bénéfiques figurent surtout sur des objets destinés aux femmes, visant ainsi à garantir leur fécondité qui est essentielle au renouvellement du groupe. Étant donné que la mystique musulmane accorde un statut particulier aux lettres de l’alphabet arabe selon une science connue sous le nom de «  science du secret des lettres  » (‘ilm al-huruf), se reporter aux textes mystiques et magiques musulmans les plus classiques permet de savoir si les lettre arabes qui sont en l’occurrence utilisées par les forgerons maures, le waw et le dal, ont une signification magique. Or, il apparaît que la mystique musulmane confère à ces lettres une valeur toute particulière; la lettre waw figure parmi les plus nobles selon Ibn ‘Ârabî, grand mystique du début du XIIIe siècle (1988, p. 425). Tandis que son contemporain, al-Bûni, célèbre auteur de traités de magie3, souligne la puissance magique de la lettre dal; elle fait en effet partie des sept lettres (suâqit al-fâtiha) qui n’apparaissent pas dans les sept versets de la première sourate du Coran et dont l’absence est l’objet de spéculations métaphysiques (Doutté 1984, p. 159). Par ailleurs, les forgerons ornaient autrefois d’un motif étoilé à valeur talismanique les planchettes des élèves coraniques ayant terminé la mémorisation d’une partie du Coran. Un tel motif est nommé «  khâtam de Salomon  » (khâtam Sulaymân). Cette expression arabe a ici plusieurs sens4; elle désigne l’achèvement et évoque en l’occurrence la fin de l’apprentissage, ou aussi le sceau, faisant alors référence au motif dessiné, ou encore la bague, renvoyant ainsi à l’anneau miraculeux de Salomon. Selon le Coran, Salomon, fils de David, avait la particularité de commander les djinns (XXXIV, 12), pouvoir qu’il tenait d’un anneau (khâtam) où figuraient des inscriptions ésotériques (Walker-Fenton 1998, p. 857). Celles-ci auraient inspiré une figure talismanique de forme hexagonale et étoilée — qui porte elle-même le nom de khâtam — qu’on retrouve dans les traités de magie les plus célèbres, notamment ceux d’al-Bûni. Cet auteur en donne par ailleurs la signification ésotérique, expliquant que les signes gravés sur la bague de Salomon renfermaient le « grand nom » de Dieu, comme ceux inscrits sur le cœur d’Adam (Doutté 1984, p. 157). De cette analyse, il ne faudrait pas conclure que les forgerons maures se sont directement inspirés de ces textes de magie, mais plutôt qu’ils ont repris dans leur art certains symboles qui puissent leur signification dans la mystique musulmane.

le fait que si les forgerons ont la capacité de produire des objets prophylactiques, ils sont eux-mêmes censés porter malheur, considérés comme plus proches de Satan (Iblîs) que de Dieu (Allâh). Certains marabouts croient, en particulier, que croiser un forgeron le matin est un mauvais présage ; aussi, évitent-ils de lui serrer la main et rentrentils chez eux pour éviter l’infortune risquant d’advenir. Car, le fait de forger et d’avoir une activité démiurgique dans la société maure apparentent moins les forgerons au divin qu’au démon. Dans la mesure où le forgeron utilise le feu dans son activité artisanale, il est associé au monde obscur des djinns qui constitue un univers parallèle à celui des hommes. Ces créatures démoniques et démoniaques sont particulièrement craintes par les hommes parce qu’elles peuvent les posséder et les rendre fous, d’où le terme qui désigne la folie en arabe, majnûn, et qui renvoie au fait d’être habité par des jnûn, autre pluriel qui permet de nommer les djinns. L’épilepsie est également causée par ces démons ainsi que le montre le terme arabe qui la désigne, « la maladie des djinns ». C’est aussi toute forme de paralysie, telle l’hémiplégie, qui, dans la société maure, est attribuée à l’attaque d’un djinn. En toute hypothèse, les forgerons sont présumés capables d’entrer en contact avec les djinns dans la mesure où ils manipulent la matière même dont sont faites ces créatures démoniques, le feu. En effet, selon les croyances locales inspirées du Coran (LV, 14), les djinns sont faits de feu tandis que les hommes sont faits d’argile : « Il a créé l’homme d’argile, comme la poterie; Il a créé les djinns d’un feu pur » (trad. Masson 1967, p. 664 t. 2). Les djinns sont considérés comme d’autant plus puissants et dangeureux pour les hommes qu’ils préexistaient à la création du premier humain, Adam, ainsi que le montre un verset coranique (XV, 26-27) : « Nous avons créé l’homme d’une argile, extraite d’une boue malléable. Quant aux Djinns, nous les avions créés auparavant, du feu de la fournaise ardente » (trad. Masson 1967, p.317 t. 1 ). Activité transformatrice D’autre part, le rapport que le forgeron entretient avec le feu permet de rapprocher l’activité de production artisanale de la reproduction humaine. Or, dans ce processus de fabrication, le forgeron tient non pas le rôle masculin, mais le rôle féminin. Si l’on se réfère au verset coranique (LVII, 25), « Nous avons fait descendre le Fer », la matière première du forgeron provient de Dieu, par essence masculin. Et le forgeron transforme cette matière, en objet, par la technique de la forge. De ce point de vue, le rôle du forgeron est analogue à ceui de la femme dans le processus de génération tel qu’il est conçu dans la société maure et en islam (Fortier 2001). Au cours ce processus, la femme n’apporte aucune substance dans la conception d’un enfant, mais son ventre contient une source de chaleur qui transforme la matière première

Activité démoniaque Les forgerons sont donc possesseurs d’un savoir secret apparenté à la magie5. Mais un des paradoxes consiste dans 126

Arts du feu et du forgeron en Mauritanie qu’est le sperme masculin en un enfant, selon un processus analogue à celui de la cuisson. Dans le texte coranique, la reproduction physiologique est l’image réduite de la création divine ; toutes deux ont en effet la même matière première, le sperme. Le nutfâ, qui désigne le sperme dans le Coran (XVI, 4), apparaît à la fois comme la substance primordiale de la création : « Il a crée l’homme d’une goutte de sperme (nutfâ)  » (trad. Masson, 1967, p. 323 t.1), et comme le premier stade du développement embryonnaire (XXII, 5). Le forgeron est donc associé au démon parce qu’il manipule le feu, matière constitutive des créatures démoniaques que sont les djinns, et parce qu’il a par ailleurs une activité transformatrice de la matière qui l’assimile à la femme, considérée elle-même, dans la société maure, comme proche du diable.

des harpes (ârdîn) des griottes. 3 Ahmad ibn ‘Ali al-Bûni, originaire de Bône (‘Annâba), serait mort en 622/1225. Son œuvre est connue en Mauritanie ainsi que dans d’autres pays musulmans. 4 Selon E. Doutté (1984, p. 83), le terme de khâtam, qui se retrouve dans d’autres langues sémitiques, désigne divers dessins auxquels on attribue un pouvoir surnaturel. 5 Le terme de sunnâ‘ qui désigne l’artisan au sens générique en arabe classique, est rarement utilisé en Mauritanie.

Références Coran, 1967. Trad. de D. Masson, 2 tomes. Paris: Gallimard. Doutté, E., 1984. Magie et religion dans l’Afrique du Nord. Paris: J. Maisonneuve et P. Geuthner. Fortier, C., 2001. « Le lait, le sperme, le dos. Et le sang ? ». Représentations physiologiques de la filiation et de la parenté de lait en islam malékite et dans la société maure de Mauritanie. Les Cahiers d’Études Africaines, XL(1), 161, 97-138. Fortier, C., 2002. De la forge à l’écriture. De l’indépendance à l’aliénation. Le statut ambivalent du forgeron dans la société maure. La Forge et le forgeron. Tome I. Pratiques et croyances. Paris: L’Harmattan (Eurasie), 125-153. Fortier, C., 2006. Intelligence pratique du berger et art magique du forgeron dans la société maure de Mauritanie. In : S. d’Onofrio et F. Joulian, éds. Dire le savoir-faire. Cahiers d’Anthropologie Sociale (1), Paris : L’Herne, 55-65. Ibn ‘Ârabî, 1988. Les Illuminations de la Mecque. Textes choisis. Trad. de M. Chodkiewicz. Paris : Sindbad. Walker, J., Fenton, P., 1998. Sulaymân b. Dâwûd  . Encyclopédie de l’islam 9. Leiden: E.J. Brill et Paris, G.P. Maisonneuve et Larose, 857-85.

Activité paradoxale Le caractère magique et démoniaque du forgeron tient par ailleurs au fait qu’il forge le fer, matériau dont le Coran (LVII, 25) parle en termes énigmatiques, affirmant à la fois son caractère dangereux et utile pour les hommes: « Nous avons fait descendre le Fer qui contient pour les hommes un mal terrible et des avantages » (trad. Masson 1967, p. 677 t. 2). En outre, la sourate où figure ce verset porte ellemême le nom de ce métal, le fer. Or, le forgeron a le pouvoir de transformer les pouvoirs néfastes du fer en pouvoirs bénéfiques puisque les objets en fer dans le monde musulman ont un caractère prophylactique, et ont en particulier la propriété de protéger les hommes des démons. Ainsi, dans de nombreuses sociétés musulmanes, dépose-t-on un objet en fer près du nouveauné pendant quarante jours pour éloigner les djinns qui pourraient l’échanger contre un de leur enfant. Notons que l’utilisation prophylactique du fer est par ailleurs bien connue d’autres sociétés, comme par exemple la nôtre, ainsi que l’illustre l’usage protecteur du fer à cheval. Cette croyance partagée semble renvoyer à la relation positive entre le feu et le fer, puisque ce métal est travaillé par le forgeron au moyen du feu. Par conséquent, on peut émettre l’hypothèse plus générale que le fer, en vertu de sa propriété de résister au feu, est devenu un symbole quasi universel de protection. Et le maniement par le forgeron des forces antagonistes que sont le fer et le feu, en fait un personnage potentiellement puissant et dangereux, attributs expliquant qu’il est maintenu, en de nombreuses sociétés, dans un état d’infériorisation.

Endnotes Dans la translittération des mots arabes, le souligné a été substitué au point de rigueur pour le h aspiré et les lettres emphatiques. 2 La lettre waw se retrouve sur les vibrateurs (harba) métalliques 1

127

Metallurgy in ancient Lecce: new evidence from the excavations of Piazzetta Epulione and Piazzetta Castromediano C. Giardino1, A. Quercia2

University Suor Orsola Benincasa, Naples E-mail: [email protected] 2 University of Lecce, Dipartimento di Beni Culturali, Via D. Birago, 64 73100 Lecce (Italy) 1

Abstract Recent excavations made by University of Lecce in Piazzetta Epulione and Piazza Castromediano (Lecce) have produced structures and artifacts related to metal-working. Most of the evidence dates back to the Roman period; traces of metallurgical activity can also be traced to the periods preceding and following the Roman period. Iron slags and semi-finished objects have been found in Piazzetta Epulione. Of particular interest are the vetrified slags related to the tuyeres, which were probably connected to smithing works. Archaeometallurgical analyses (microscopy and XRD) of objects found in the excavations support the hypothesis that the metallurgical workshops of Lecce were involved in the production of iron.

Piazzetta Lucio Epulione: the archaeological evidence Piazzetta Epulione lies in the south-eastern part of the historical centre of Lecce, just south-east of the Roman theatre; the area is located in a part of the modern town where archaeological evidence was already discovered, between the end of the 19th and the first half of the 20th century1. Preventive archaeological investigations took place in 2001, during works of maintenance of the city sewer system; they led to the discovery of archaeological deposits in section. Archaeological excavations started in 20022 in an area of 215 squares metres; they brought to light a stratigraphy that underlined the continuity of occupation of the piazzetta from the 8th century B.C. up to the contemporary age and they allowed to acquire an interesting page of the history of Lecce3 (fig. 1). The earliest metallurgical evidence dates back to the Messapian phase (4th-3rd centuries B.C.). It deals with finds coming from layers identified in the eastern part of the excavation, where it was possible to deepen the dig up to Messapian levels. They are, in most of the cases, layers of accumulation (units 246, 270), that obliterate the remains of a Japigian hut datable between the end of the 8th and the 7th century B.C. (fig. 1, n. 1); the aforesaid layers can be attributed to the period between the second half of the 4th and the 3rd century B.C., according to the finds. Despite the few evidences, it is possible to hypothesize that metalworking took place in the area of Piazzetta Epulione already between the 4th and the 3rd century B.C. (A.Q.) The excavation of the forges Piazzetta Epulione area was interested by the presence of a metallurgic workshop, active in iron manufacturing during the Roman period (fig. 1, n. 2). In an area placed near a road, endowed with a wall of delimitation (unit 7),

with a pedestrian sidewalk (units 38 and 144) and with a roadway for the vehicles (unit 122) - datable between the second and the first half of the 3rd century AD4 - a series of evidence like a sequence of plans of burnt clay, layers of ash and remains of hearths, overlapped between them and characterized by abundant iron slag were recovered to the west. The layers were delimited to north by an east-west wall built with blocks. The presence of adjustments and alteration of medieval and post-medieval period makes rather difficult a general interpretation of the evidences related to this workshop. (A.Q.) Starting from the upper layers, the remains of a hearth of semicircular shape was discovered (unit 253); the were covered by an accumulation layer dating back to the late antiquity (Fig. 2). This hearth is laid down on a reddish and very compact plan, whose characteristics are related to the high temperatures reached (unit 254); most of the slag and semi-finished iron products came from this level. This layer could be interpreted as a working plan for metallurgical activity, to which the forging hearth was functionally connected. It was flanked by two pits of quadrangular shape, whose walls were filled with stones (units 257, 259), in which a pair of anvil bases were probably inserted; a small pit filled with discarded material was found next to them (unit 255). (AQ, C.G.) The removal of this level brought to light a layer of local crushed limestone (“tufina”) in the central part of the area (unit 272). It covered a hearth for all its extension; this layer presented thermal alterations producted by the contact with the underlying structure. This hearth, probably a second forge (fig. 3), had a rough rectangular form slightly lengthened to the west, and it was composed by two different layers, divided by an alignment of small fragments of stones and “tufina”. To the west the unit 282 was composed by a very soft, dark grey soil (more compact to the centre),

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

C. Giardino, A. Quercia

Fig. 1. Lecce, Piazzetta Lucio Epulione. General map showing the position of the contexts with metallurgical activity.

mixed up with fragments of calcarenite, thermally alterated, and anthracological finds; to the east the unit 283 had lengthened form, reddish and very compact. Probably the metal was heated to high temperatures in the western side of the forge and it was subsequently cooled in the eastern part. The upper layer of tufina (unit 272) sealed the forge when it was not in use any more, even if it cannot be excluded that this layer had a role in the metalworking processes5. Other possible bases of forges were found under this level; they had different dimensions and they were made by burnt clay and pressed calcarenite (units 289, 291, 292). The largest one (unit 289) had the same dimension as the forge unit 282-283, that was just above; it was charac-

terized by chromatic gradations (from dark-grey to red), according to the distance from the fire. The bases were built on a compact floor (unit 273) that was probably the occupation surface of the productive area; the same floor was detected to the west (units 139 and 194), for about 7 m long. The data allow to affirm that the area was used for iron processing; the hearths and the bases in burnt clay can be interpreted as forges. The productive area underwent small transformations in the spatial organization during the time. In fact the lower burnt clay layers (units 282-283, unit 289) did not change their position, while the upper hearth (unit 253) was shifted slightly to the south. 130

Metallurgy in ancient Lecce

Fig. 2. Lecce, Piazzetta Lucio Epulione. The productive area; the forging hearth unit 253, the plan unit 254 and the pits units 257 and 259.

The productive area was delimited by walls made by square limestone blocks different in dimensions6, badly preserved because of medieval and modern interventions; therefore the general interpretation is rather complex. The best preserved wall was identified to the north (unit 58); it presents two rows of blocks. The floors and the plans of the productive area leant to it: to the south the wall unit 172 and to the west (orthogonal to the last one) the unit 171. Another portion of wall (unit 173) was identified to the west, perhaps related to the same structure7. According to the alignments of the walls hypothetically it is possible to reconstruct a rather large building, of about 55 square metres (10 m long, 5.5 m wide). Probably it was divided into two rooms: the eastern room for productive activities8, while the western one – smaller

and with a tile covering - had not industrial structures and could have different functions (commercial and/or residential ?)9. The few pottery found in the levels of use10 fixed the chronology of the smithery activity between the late second and the first half of the 3rd century A.D.; it was, therefore, contemporary to the adjoining road, that was probably functional to the productive structure. Some slag were found in the preparation layer of the most recent pavement of the road (unit 37), dated to the middle of the 3rd and the beginning of the 4th century A.D. An intense metallurgical activity held in the area of Piazzetta Epulione during the Late Antiquity. A certain number of metallic slag were found in the accumulation 131

C. Giardino, A. Quercia

Fig. 3. Lecce, Piazzetta Lucio Epulione. The productive area: the forge units 282- 283.

layers and in the discards (units 40, 53, 79, 128), in close association with a great amount of pottery. This pottery was discarded to level the Roman road and the surrounding area. The pottery11 fix the chronology of this intervention between the 5th and the 7th century A.D. (A.Q.)

Forges were used in two different processes, that took place after iron ore smelting. The first operation was to refine the bloom in to a semi-finished bar; the smelting furnace produced, in fact, a spongy mass, the bloom, that was composed by iron mixed with charcoal remains and slag. A second, later process transformed the bar into finished implements. Both processes do not need the reduced atmosphere, that is necessary in the smelting activity; in fact, the blacksmith have only to heat the iron, in order to allow the hammering (Giardino 1999, pp. 205-207; Giardino et al. 2003, p. 643). According to Diodorus of Sicily, sometimes the bloom was splitted into smaller parts, which were then sold (Diodorus, History 5.13.1-2).

The forges of Piazzetta Epulione: the archaeometallurgical evidence As a rule, the forges have a very simple shape; often they were made mostly by clay, because they were not permanent structures. Therefore their identification can be rather difficult in archaeological excavations. The reuse of the area limits generally their evidence to more or less regular spots of burnt surfaces that are surrounded and filled by ashes, charcoals and slag. 132

Metallurgy in ancient Lecce represent the blacksmith’s workshop. A marble slab from the Domitilla’s Catacombs in Rome represents two artisans; it is dated to the 1st century A.D. The first craftsman works at the anvil, where he is hammering a bar kept with the tongs; the second blacksmith stay at the forge, operating at the bellows. Unlike the evidences from Lecce, in this relief the fire is placed on a raised base and it has a kind of wall (perhaps made by bricks) on the back, to protect bellows and blacksmith from the overheating. The forges found in Rome, on the Capitoline Hill, were rather similar to the structures from Lecce, Piazzetta Epulione, but they are rather more ancient: they belong to levels contemporary to the construction of the temple of Jupiter (Giardino, Lugli 2001, p. 327). (C.G.).

The archaeological identification of smithig hearts can be difficult; key finds are slag. The Roman forges have generally an elongated elliptical plan, usually 1-2 m long; they were occasionally constructed of clay and stone and they have occasionally an internal clay lining (Schrüfer-Kolb 2004, pp. 31-32, Fig. 17). The two forges of Piazzetta Epulione did not worked at the same time; nevertheless they both can be date period between the end of the second and the first half of the 3rd century A.D. These forges represent a rather rare archaeological evidence in the Italian Peninsula, because the paucity of similar remains in Italy. The most recent forge (unit 253) appeared partly damaged, but it can be reconstructed as a hearth of coarse elliptic shape, measuring nearly 100 x 50 cm. Unfortunately, nothing of the upper structure was preserved, a part from few scorified clay fragments; they are an indication that the forge walls were made (and probably lined) by clay. Using the different traces of combustion in the soil as a guide, it is probably possible to distinguish two different and separate areas inside the forge. In one side the tuyere was placed; therefore the highest temperatures were reached in this area. The other side was located at a certain distance from the air outlet: this was the area where the manufactured objects had the opportunity to cool slowly. A similar space organization can be reconstructed also in the oldest and largest forge, measuring 130 x 90 cm (unit 282-283); it was preserved better than the previous one. Evidence of the walls were not preserved, except few fragments of vitrified burnt clay. An irregular dark spot was found very close to the first forge; a concentration of anthracological remains distinguished this unit (unit 254): it could be the storage place for the charcoal used as fuel. Two pits (unit 257 and 259) were discovered near this forge. Most likely the anvil was located in one of them (probably the unit 259). It is also possible that the two pits housed a couple of anvils, typologically different, used for specific purposes in metal production. The blacksmith’s stele from the oppidum of Le Bure near Saint Dié (Vosgi) offers a good iconographical example of the co-presence of more anvils in the same workshop: three different anvils are represented on it (Cima 1991, p. 184). Anvils have to be very close to the forge, because the iron must be hammered when it is incandescent. If the interpretation of the pit unit 259 is really, as supposed, an anvil base,, the anvil stayed less than a meter from the smithing hearth. The small pit unit 255 found near the forge could be used to discard the residues. Agricola’s pictures from De Re Metallica (Agricola 1556, book IX) show late Medieval and Renaissance blacksmith’s workshops where pits for discarded materials close to the forges were in use. Many Roman relieves, generally of funerary character,

Functional interpretation of the slag A large amount of slag was found in the excavation of the blacksmith’s workshop of Piazzetta Epulione. They represent, together with scorified clay and tuyeres, important evidences to detect the processes involved in iron making processes. A correct interpretation of these evidences needs to com-

Fig. 4. Slag related with tuyeres, found in Messapian levels (4th3rd century B.C.) (unit 246).

Fig. 5. Whetstones from Piazzetta Epulione, Messapian levels.

133

C. Giardino, A. Quercia Two large concavo-convex slag are particularly meaningful; they both came from unit 254 (Fig. 6). These slag are different in dimensions and weight (g 181 and g 378 respectively); both magnetic, the smallest one shows a lower magnetism. They are the remains of smithing hearth bottoms. It can outlined that the lower surface is rather smooth in both slag; this is an indication that the blacksmith kept the hearth constantly clean. The blacksmith therefore acted in order to maintain the job under control, because he wanted to produce objects of a certain quality: systematic cleaning of the hearth improves the visual control of the temperature and reduced the risk to contaminate the iron with other materials (Fluzin 1999, p. 75). A small iron ingot (weight g 125) was found in unit 37, which can be dated between the second half of the 3rd and the beginning of the 4th century B.C. This find is probably one of the forges products, a semi-finished bar ready to be exported to other workshops, or transformed in the same place into a finished implement (Fig. 7). Fragments of small semi-finished bars were recovered in the unit 128 also, belonging to the Late Antiquity levels (5th – 7th century A.D.). Numerous residual shapeless slag (unit 40), together with a concavo-convex slag (unit 39), were found also in contemporary levels; they suggest the continuation of the metallurgical activity for some centuries in the area of Piazzetta Epulione (C.G.).

bine archaeometrical data, obtained by the structural analysis of the finds, with the information coming from the careful study of the archaeological contexts. Sometimes the results of archaeometrical investigations can be ambiguous: smelting slag process can physically and chemically resemble the slag produced during the smithing operations, especially if the analysed specimens are small and broken, as it happens with the finds from the blacksmith’s workshop of Lecce (Giardino et al. 2003, p. 641). The first evidence of metallurgical activity in Piazzetta Epulione come from the Messapian levels (4th – 3rd century B.C.). Two vitrified clay fragments were found in unit 246; one of them has a rough cylindrical shape and can be interpreted as the imprint left by the tuyere, or else as a slagged fragment of the tuyere (fig. 4). Two whetstones came from the same unit; they were used for finishing the metal objects (fig. 5). Because of the few remains, it was not possible to establish if iron or copper alloys (or both) were manufactured on the site in the Messapian time. Clear evidence of iron metallurgy come from the Roman layers (second and first half of the 3rd century A.D.); light, glassy and not magnetic slag were found; they are related with the sintering process of the clay from the forges induced by high temperatures.

The metallurgical evidence from Piazzetta Castromediano Other evidence, concerning metallurgical activity in Lecce, come from an excavation made in Piazzetta Castromediano, a site located in the eastern part of the historical centre of Lecce, just north of the Roman amphitheatre. In 2002, works for the maintenance of the city sewer system brought to light archaeological levels and structures. Therefore the digging area was enlarged; a series of archaeological excavations took place12 in an area of about 3600 square metres between 2002 and 2004. The place interested by the research was so large that it was not possible to complete the investigations up to the bedrock. A complex stratification was found attesting the occupation from the Iron Age up to nowadays13. (Fig. 8) The earliest metallurgical evidence can be dated to the Late Roman phase (3rd-4th centuries A.D.): slag coming from a level of discard (unit 819) that obliterated the remains of the earthen floor of a north-south road, dated to the 2nd3rd century A.D., according to the associated pottery. The road was paved with small stones and was flanked by a sidewalk (unit 828) delimited by limestone blocks. Other slag come from a level of discard (unit 717) belonging to the same phase (end of the 3rd – beginning of the 4th century A.D.). This layer was cut by a small circular cooking hearth built with small shapeless stones (unit 728). It can be dated to the medieval period. Other indications of metallurgical activity were found in

Fig. 6. Concavo-convex slag from the levels of the second-first half of the 3rd century A.D. (unit 254).

Fig. 7. Semi-finished iron bar from Piazzetta Epulione (unit 37).

134

Metallurgy in ancient Lecce the 16th century filling14 of a lacus olearius of republican age (2nd - 1st century B.C.); they can be referred to a later occupation phase of the area. (C.G., A.Q.) During the excavation, some slag were found; they are connected with a late Roman forging activity like at Piazzetta Epulione. A peculiar interest had a concavo-convex slag found in unit 819 (Fig. 9). Unlike the 2nd-3rd century A.D. samples recovered at Piazzetta Epulione, this one shows an irregular lower surface. It means that here the blacksmith did not pay much attention to the cleaning of the forging hearth; besides, the presence of green lumps still attached to the surface is an indication that the forge was used, perhaps desultorily, not only for iron, but also to melt copper alloys too. (C.G.). The ore sources Where the iron worked in Lecce was smelted is still a matter of discussion; this problem is connected with the location of the mining areas. Puglia is commonly regarded as a region lacking in iron ore deposits that could be used in antiquity. It is possible to hypothesize that the raw material was imported from other regions, like Calabria, Sicily, Etruria, the Eastern Alpine area or also, on the other side of the Adriatic Sea, Albania, Macedonia, Bosnia or Serbia (Gliozzo, Arthur in press, p. 8). Iron ore deposits exist in Apulia and especially the Salento area, in form of bauxites ore outcrops; a very large one is located south-east of Lecce, and it covers a vast area between S. Donato and Santa Maria di Leuca (Cotecchia, Dell’Anna 1959, p. 5; Crescenti, Vighi 1964, pp. 288307). Recently the bauxite was exploited, starting from the middle of the nineteenth century, as mineral to extract the aluminium. Nevertheless bauxites contain in fact high percentages of iron, that can also reach 30-40% (Boni 1972,

Fig. 8. Lecce, Piazzetta Castromediano. General plan of the excavation with the location of the contexts where evidence from the workmanship of the iron has been found: 1) levels of Late Roman period (units 819 and 717) 2) discards of the 16th century A.D. located inside the lacus olearius of republican age.

Fig. 10. Lecce, Piazzetta Epulione, bauxite nodule from the levels of 4th-3rd century B.C. (unit 270).

Fig. 9. Concavo-convex slag from Piazzetta Castromediano (unit 819).

135

C. Giardino, A. Quercia

Fig.11. Lecce, Piazzetta Epulione, bauxite nodule from the levels of IV-III century B.C.: XRD spectrum.

p. 487). Laterite is a kind of rock very similar to the bauxites found in Puglia and this ore was usually smelted to obtain iron in Central Africa; iron metallurgy existed in that country at least from the first millennium B.C. (Fluzin et al. 1995, pp. 56-58; Quéchon, Person 2000, pp. 9-11; Bocoum 2000, p. 16; Serneels 2005, p. 1). It is therefore possible that, as happened for the African laterites, the Salentine bauxite too was exploited in antiquity as iron ore. Until Nineteenth century, bauxite was considered a useful ore for aluminium and iron, according with its content (D’Achiardi 1883, p. 399). As a matter of fact, iron was smelted from bauxites in modern times, by the Bourbons’ government, just before

the unification of Italy (L. Baldacci, letter of 19/05/1909, in Mosso 1909, p. 316; Costa 1864, p. 369). A bauxite nodule was found at Piazzetta Epulione in layers dating back to the 4th-3rd century B.C. (unit 270) (Fig. 10). It was analyzed by X-Ray Diffraction and observed with the metallographic microscope, and it is very similar to those recovered in the Salentine ore deposits (Figs. 11, 12). The few data available indicate that Roman smithing sites were located in Puglia in the cities of regional importance, where probably iron production was one of the industrial activities carried out on the site. In these cities, like in Lecce, professional blacksmiths’ workshops were established. If the hypothesis of ancient iron smelting starting from the local bauxite ore is correct, the Roman iron production in Apulia has some similarities with models observed in other parts of Europe as well. In Roman Britain – where Roman iron production is relatively well known - smelting sites are correlated with the major ore outcrops, because smelting was undertaken close to the ore deposits. Professional smithing sites are placed in small towns or in the major cities, with important population concentrations (Schrüfer-Kolb 2004, pp. 44-48, pp. 131-132). According to this model, we could perhaps have to expect to find the smelting place in the fields of Salento, not too far away from the bauxite ore deposits. More detailed field work, strictly linked to archaeometallurgical analyses, is needed to identify and to study the metallurgical sites in Puglia region. The examination of the evidence from Lecce is a first, basic step to the comprehension of the complex network of Roman productive activities.

Fig.12. Lecce, Piazzetta Epulione, bauxite nodule from the levels of 4th-3rd century B.C.: microphotograph, 20 x.

136

Metallurgy in ancient Lecce (C.G.).

African D Sigillata (types Hayes 87A and 104B) and Eastern Late Roman C (type Hayes 3) are documented, together with Eastern (Late Roman 2 and 4) and Africans (spatheia, Keay LXII) transport amphorae, plain and painted pottery, and kitchen ware. 12 The excavations were directed by prof. Francesco D’Andria. 13 A short synthesis of the living phases of the area in D’Andria 2004, pp. 54-55. 14 Analysed slag come from the upper part of the filling (unit 256). 11

Acknowledgments

Our first debt is to prof. Francesco D’Andria, who gave to us the possibility to present this preliminary study. The excavations at Piazzetta Epulione, directed by prof. F. D’Andria, within the Project of Urban Archaeology “Lecce Sotterranea”, and financed by the Municipality of Lecce, are developed in collaboration with the Archaeological Superintendence of Puglia (dott. Luigi Tondo, Operational Office of Lecce); they were conducted by a team of archaeologists from the Department of Cultural Heritage of the Lecce University, coordinated by Alessandro Quercia. The graphic documentation of the excavation was elaborated by Fabrizio Ghio. The excavation at Piazzetta Castromediano were also directed by Prof D’Andria. A team of archaeologists from the Department of Cultural Heritage of the University of Lecce took part to the digging, coordinated by Caterina Polito, in cooperation with the Archaeological Superintendence of Puglia and financed by the Municipality of Lecce within the Project of Urban Archaeology “Lecce Sotterranea.”. We are grateful to Caterina Polito, who supported the excavation documentation of Piazzetta Castromediano. The aerophotograph of Piazzetta Castromediano was realized by Corrado Notario. We should also like sincerely to thank Umberto Savarese and Lorena Scarpato for inking the drawing Fig. 18. Our thanks are also due to Francesca Silvestrelli for the English translation of the text.

References Agricola, G., 1556. De Re Metallica libri XII. Basilea. Boni, M., 1972. Bauxiti dell’Italia centrale, meridionale e della Sardegna (Bibliografia ragionata). L’industria mineraria, XXIII (oct.), 487-519. Cima, M., 1991. Archeologia del Ferro. Torino: Grafo. Cotecchia, V., Dell’Anna, L., 1959. Contributo alla conoscenza delle bauxiti e terre rosse del Salento. Memorie e note dell’Istituto di Geologia Applicata di Napoli, VII, 3-20. Crescenti, U., Vighi, L., 1964. Caratteristiche, genesi e stratigrafia dei depositi bauxitici cretacici del Gargano e delle Murge; cenni sulle argille con pisoliti bauxitiche del Salento (Puglie). Bollettino della Società Geologica Italiana LXXXIII, 1, 285-338. D’Achiardi A. 1883. I metalli. Loro minerali e miniere II. Hoepli: Milano. D’Andria, F., 2004. Il sottosuolo come risorsa di conoscenza e sviluppo. In: M. De Stefano, ed. Lecce. Riqualificazione e valorizzazione ambientale, architettonica ed archeologica del centro storico. Roma : Ibam, 46-67. Fluzin, P., 1999. Il processo siderurgico: evoluzione storica e indizi archeometrici. In: C. Cucini Tizzoni, M. Tizzoni, eds. La miniera perduta. Cinque anni di ricerche archeologiche nel territorio di Bienno. Comune di Bienno, 61-92. Fluzin, P., Benoit, P., Kienon, H.T., Kiethega, J.B., El Kedim, O., 1995. Apports de l’archéométrie à la restitution de la chaîne opératoire des procédés sidérurgiques directs à partir des vestiges archéologiques; interêts des comparaison ethnoparcéologiques. In  : G. Magnusson, eds. The Importance of Ironmaking. Technical Innovation and Social Change. Papers presented at the Norberg Conference on May 8-13, 1995. Stockholm, 56-64. Giardino, C., 1999. I metalli nel mondo antico. Introduzione all’archeometallurgia. Roma – Bari: Laterza. Giardino, C., Guida, G., Migliarelli, A., Vidale, M., 2003. La documentazione archeologica dell’antica lavorazione del ferro. In: C. D’Amico, ed. Atti del Secondo Congresso Nazionale di Archeometria. Bologna 2002. Bologna, 639-656. Giardino, C., Lugli, F., 2001. L’attività siderurgica nel Giardino Romano. Bullettino della Commissione Archeologica Comunale di Roma, CII, 327-328. Giardino, L., 1994. Per una definizione delle trasformazioni urbanistiche di un centro antico attraverso le necropoli: il caso di Lupiae. Studi di Antichità, 7, 137-203. Ghiozzo, E., Arthur, P., in press. Archaeometallurgic study of Bizantine and Medieval slags from the Salento (Southern Apulia, Italy): new evidence and open problems. Mosso, A., 1909. Stazione preistorica di Coppa Nevigata presso Manfredonia. Monumenti Antichi dell’Accademia dei Lincei XIX, coll. 305-386.

Endnotes Some Messapian graves were found in the surrounding areas (particularly in the streets F. Di Aragona and Conti di Lecce and the church of St. Matteo); they are dated to the 4th-3rd century B.C.: Giardino 1994, pp. 199-200, nn. 210-217. 2 The excavations were directed by prof. F. D’Andria, within the Project of Urban Archaeology “Lecce Sotterranea”. 3 A short synthesis about the phases of life of the piazzetta appeared in D’Andria 2004, p. 54. Preliminary notes about the excavations are in Quercia 2003 (for the post-medieval phase), Quercia 2005 and Quercia in print. 4 The level of the road was raised again later, adding a new pavement (unit 57) that enlarged the roadway, covering the sidewalk. This intervention can be dated between the middle of the 3rd and the beginning of the 4th century A.D., according with the pottery. 5 Currently, in Salento, some blacksmiths smooth the metals using cakes of tufina. We thank Antonio Mangia for the information. 6 The largest blocks are 100 cm long and 30 cm wide; the smallest are between 70 to 30 cm. 7 The west closing wall was not found; it cannot be excluded that it was never built and that the productive area leaned out on the road; the wall unit 58 is tied up to the wall that delimitated the road (unit 7). 8 Another plan made by tufina was recovered behind the angle among the walls units 171 and 172; it had with evident traces of combustion (unit 239), cut by a circular pit (unit 240) filled with the remains of a discarded hearth, as attested by the reddish colour of the filling (unit 241). The scarce diagnostic material into the pit dates generically the activity of the hearth in middleimperial age (2nd-3rd century A.D.). 9 The floor of this room (unit 194), was covered by the tile roof collapse (unit 212); the collapse chronology cannot be determined because of the small amount of pottery found. 10 African A/D and C Sigillata and African kitchen pottery are documented. 1

137

C. Giardino, A. Quercia Quéchon, G., Person, A., 2000. The iron age. In: eds. The Iron Roads in Africa – Les Routes du Fer en Afrique. Paris: UNESCO, 9-14. Quercia, A., 2003. Lecce, piazza Lucio Epulione, 2001-2002. Archeologia Postmedievale, 7, 301-302. Quercia, A., 2005. Lecce, Piazzetta Lucio Epulone. Taras, 22 (1-2) 156-159. Quercia, A., 2007. Lecce, Piazzetta Lucio Epulone. Taras, 23, 187-190. Schrüfer-Kolb, I., 2004. Roman Iron Production in Britain. Tecnological and socio-economic landscape development around the Jurassic Ridg. BAR British Series 380, Oxford: BAR Publishing. Serneels, V., 2005. An Ongoing Research Project on Iron Production and Use in the Dogon Country, Mali. HMS News, Historical Metallurgy Society, 60, Summer 2005, 1-3.

138

La capanna rituale di Serra Cicora (Nardò-LE) E. Ingravallo

Dipartimento di Beni Culturali, Università degli Studi di Lecce, via Birago 64, 7300 Lecce, Italy E-mail: [email protected]

Abstract In the site of Serra Cicora two moments of the Neolithic occupation are documented: the first in the early Neolithic with impressed pottery and the second during the Serra d’Alto-Diana period. To this second moment the funeral monuments, two of them with a small Serra d’Alto pot as equipment for the buried, can be ascribed. The radiocarbon analyses carried out on some of the remains date the complex to about the V millennium BC, a phase of profound transformations of the Neolithic societies. At the centre of the plateau, surrounded by funerary circles, there is an oval-planned hut with a floor of beaten earth and a hole for a pole in the interior. The whole site presents signs of fire, probably due to specific moments of the ritual.

Serra Cicora è un sito neolitico ubicato su un pianoro lungo la costa ionica nei pressi di Nardò (Ingravallo 2004). Tre datazioni, di cui due da carbone e una da osso umano, fanno risalire alla I metà del VI millennio BC l’inizio delle frequentazioni che si prolungano nel V millennio. Alla facies culturale della ceramica impressa si sovrappone quella di Serra d’Alto-Diana, durante la quale il pianoro viene usato come sede di una necropoli. La parte centrale del pianoro, in particolare, viene occupata da circoli funerari con deposizioni singole o multiple datate alla metà del V millennio BC. In posizione dominante rispetto a questi e conservatasi grazie all’incendio, è la base ovale (140x130 cm) di una capanna in miniatura, di cui resta la parte inferiore delle pareti di intonaco, alte in alcuni punti 40 cm dal piano di campagna, e una buca da palo all’interno. Sulla probabile soglia, deducibile dall’interruzione della parete di intonaco, era un individuo deposto secondo una modalità inconsueta rispetto alle altre deposizioni: in decubito laterale destro ma in direzione nord-sud, al posto dell’orientamento sud-nord che caratterizza il resto degli inumati. Dalla giacitura, lo si sarebbe detto un guardiano dei morti. Al pari degli altri individui, non era stato minimamente toccato dall’azione del fuoco. Buche da palo, per quanto rare, si segnalano anche nella zona dei circoli, suggerendo la presenza di strutture forse di copertura, a protezione dei monumenti funerari. L’incendio subito dalla capanna ha prodotto una grande quantità di intonaco con relativa dispersione su tutta l’area. Interessanti sono soprattutto le impronte straminee che si osservano su molti campioni che rivelano un complesso lavoro di carpenteria. L’insieme dei resti bruciati è in fase di studio, oggetto di una tesi concordata con il Prof. Fiorentino, docente di paleobotanica, con il quale è stata elaborata una scheda in grado di estrarre il maggior numero possibile di informazioni da ogni singolo pezzo. Resta il problema di risalire alle cause degli incendi, i cui resti si trovano negli insediamenti neolitici e che, nella letteratura, vengono di solito attribuiti a fattori imprevedibili che seguono la fase dell’abbandono. Studi sperimentali sull’argomento hanno dimostrato che molto difficilmente un incendio ha la forza di radere al

suolo un villaggio o anche una sola capanna: per farlo, ha bisogno di essere continuamente attizzato e alimentato. Sulla base di tali osservazioni M. Stevanovic (2002) ha analizzato i resti bruciati di alcuni siti neolitici dell’Europa sud-orientale, giungendo alla conclusione che gli incendi delle case – lungi dall’essere spontanei – vanno attribuiti a un’azione sociale deliberata e dall’alto contenuto simbolico: con la distruzione delle abitazioni si intendeva probabilmente suggellarne il tempo di vita e interdirne l’uso futuro. Paradossalmente, le case sarebbero state bruciate per assicurare visibilità e continuità alla coesione sociale del gruppo che abitava in un dato insediamento. Costruzione, uso, distruzione delle case erano le tappe di un processo continuo con cui rimarcare la proprietà della terra, il radicamento dei suoi abitanti, il legame tra mondo dell’al di qua e mondo dell’aldilà. Numerosi sono gli esempi etnografici relativi alla distruzione e riedificazione di case per ragioni strutturali o sociali, come per esempio presso gli Athapaskans (sud-ovest USA), che a volte abbandonano le loro dimore per la paura di spiriti ancestrali, o presso i Navajo dove si può bruciare o rendere inabitabile una casa in cui è morto qualcuno. Emblematico, d’altro canto, è il caso dei tell dell’Europa sud-orientale e del Vicino Oriente, là dove numerose generazioni hanno edificato le loro abitazioni su precedenti livelli di occupazione per sottolineare – nel ciclo ininterrotto di costruzione, uso, distruzione delle case – la presenza e la continuità del proprio lignaggio. Non si tratta, naturalmente, di fare indebiti confronti tra realtà differenziate come l’Europa sud-orientale e l’Italia meridionale, essendo nota la varietà dei “neolitici” europei. E, tuttavia, è probabile che la capanna di Serra Cicora – tenuto conto del luogo in cui sorge e del ruolo cultuale che avrà avuto – sia stata deliberatamente bruciata a compimento di cerimonie particolari. L’uso controllato del fuoco a scopo rituale è, d’altra parte, ben documentato a Serra Cicora in due delle strutture funerarie: in una (T9) l’individuo era adagiato su una superficie regolare di pietre, a loro volta sistemate sul livello basale fatto con blocchi di argilla cotta. Poco distante, una concavità naturale della roccia recava

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

E. Ingravallo evidenti segni di una prolungata azione del fuoco visibile dal suo accentuato stato di calcinazione. L’altra struttura funeraria (T12) conteneva quattro individui affiancati in coppia, di cui una a nord e l’altra a sud: presso uno di loro era una scapola di bue all’altezza della mandibola. Tutti giacevano, molto compressi data l’esiguità dello spazio, su un letto di terra al di sotto del quale era uno spesso pacco di argilla cotta. I riti funerari rappresentano da sempre il tentativo di addomesticare la morte, per antonomasia la forza indomabile dell’universo, che con il fuoco condivide l’indole imprevedibile e selvaggia. Questo non autorizza, comunque, a generalizzare l’equazione fuoco uguale morte e, quindi, ad attribuire un significato esclusivamente metaforico all’uso del fuoco, trascurandone l’apporto strutturale o, talvolta, l’aspetto enigmatico come nel caso delle sepolture citate, il cui trattamento le differenzia dalle altre presenti a Serra Cicora. Quanto alla capanna, invece, si possono avanzare alcune ipotesi di spiegazione facendo ricorso alla sfera simbolica: distruggerla con il fuoco poteva essere il modo di costringere la natura (la morte, il fuoco) entro gli spazi dominabili della cultura (la casa, lo spazio chiuso). Ma la capanna incendiata poteva anche essere il “sema” che dava potere fondativo alla proprietà della terra in cui erano sepolti i propri morti. Le comunità del V millennio, non a caso, introducono l’uso generalizzato delle necropoli che rimanda al bisogno di radicamento nel territorio e alla necessità di tradizioni condivise con cui alimentare la coesione interna. In un periodo in cui si allarga la rete degli scambi e dei rapporti con la circolazione dei beni di lusso su larga scala, la competizione tra comunità esigeva un forte senso di appartenenza e di autostima. La cultura di Serra d’Alto è nota, d’altro canto, per il cerimonialismo e la ritualizzazione delle occasioni di scambio: la Grotta dei Cervi di Porto Badisco diventa il “santuario” presso il quale formalizzare momenti rituali e rinsaldare la rete di relazioni sociali presente nel territorio. Gli abitanti di Serra Cicora, in altre parole, dando degna sepoltura ai defunti con l’atto finale del fuoco, hanno voluto celebrare se stessi grazie alla forza evocativa di un rito di cui restasse memoria nel tempo.

References ingravallo, e.,

2004. Il sito neolitico di Serra Cicoria. (NardòLe): note preliminari. Origini, XXVI, 87-119.

stevanovic,

m., 2002. Burned Houses in the Neolithic of Southeast Europe. In: D. Georghiu, ed. Fire in Archaeology. BAR International Series 1089. Oxford: BAR Publishing, 55- 62.

140

Recenti ricerche sulla produzione di utensili lignei a Karatepe-Aslantas, Turchia M. R. Iovino1, C. Altinbilek2

Centro Internazionale di Sperimentazione, di Documentazione e di Studio per la Preistoria e l’Etnografia dei Popoli Primitivi. Via S. Zosimo 10, 96100 Siracusa (Italia). E-mail: [email protected] 2 Ýstanbul Üniversitesi Edebiyat Fakültesi Arkeoloji Bölümü Prehistorya Anabilim Dalı – Istanbul, Turkey. 1

Abstract At Karatepe-Aslantas (Turkey), nearby the Mediterranean coast, about 28 km to north of Osmaniye, to east of Adana and to north respect Antakya, the traditional technology of spoon wood production used from the elderly of the village has been investigated. Wood represents still today a matter of use of leading importance. Beside the obvious employment in the construction of the huts, palisades, boats and other, wood can be employed in the preparation of big part of the household tools, tools for hunting and for defense. In the field of the agricultural and pastoral cultures the wood was and is still today the predominant material for the accomplishment of a lot of tools and containers, sticks, knots for the animals, spatulas, ladles of different type, spoons for the skimming of the milk. The spoon, particularly, acquires also a social meaning. The making of spoon converges towards the cycle of preparation and of food consumption. Farming and pastorals life-style introduced new foods, new methods of food processing and new feeding style. The usual ethnographical description of wood working shows details on the utilization of metal tools but in the case-study here presented glass tools are also attested. This feature contributed to raise research hypothesis for the functional study of obsidian prehistoric tools. The production technologies observed at Karatepe-Aslantas follow rules of ancient cultural tradition but small variations due to the craftsman and to a belonging to a particular ethnic group are here testified.

Introduzione La Turchia è una grande penisola con una superficie di circa 780 600 km2, situata tra l’Asia sud occidentale e l’Europa. Essa confina a nord con la Grecia e la Bulgaria, ad est con la Georgia, l’Armenia e l’Iran, a sud con l’Iraq e la Siria. È circondata da quattro mari: il Mar Nero al nord, il Mare mediterraneo al sud, il Mare Egeo ad ovest, ed il Mare di Marmara a nord-ovest. Situata alla congiunzione tra l’area mediterranea, l’Irano-Turaniana, e l’area Euro-siberiana, sebbene l’altopiano dell’Anatolia Centrale e le parti orientali montuose siano sottoposte

ad un clima continentale, la maggior parte del territorio turco si trova sotto l’influenza del clima mediterraneo. Le specie vegetali indigene si sono largamente conservate nelle regioni dell’area Mediterranea ed Irano-Turaniana (Akman, 1982). La ricerca qui presentata è stata svolta in Cilicia (Fig. 1), nell’area rurale attorno alla riserva naturale di KaratepeAslantaş (provincia di Adana, coordinate geografiche 37°16’59”N 36°14’20”E) situata nella vallata del fiume Ceyhan, poco distante dal sito Ittita di Domuztepe-Aslantaş

Fig. 1. Carta geografica della Turchia. area di Karatepe- Aslantaş (Adana) indicata da triangolo, cartina gentilmente concessa da University of Texas Libraries, The University of Texas at Austin. g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

M. R. Iovino, C. Altinbilek (Çambel 1948, 1996). Qui è presente una foresta dominata da Pinus brutia Ten. e da Pinus halepensis, e macchia mediterranea con Quercus ilex, Quercus aegilops e Quercus infectoria, Olea europaea, Cornus mas, Osyris alba, Buxus sempervirens, Phillyrea angustifolia and Pistacia lentiscus. La vegetazione fluviale include Tamarix spp. e Nerium oleander. Il legno, in questo contesto ecologico, rappresenta una risorsa economica di facile reperibilità e la popolazione è interessata al suo sfruttamento con una produzione- a conduzione domestica e non industriale- di manufatti di uso quotidiano che, soprattutto in estate, trovano una distribuzione anche tra i turisti, incrementando l’economia del villaggio.

a lavorare il legno, i cucchiai in legno che venivano usati dagli abitanti di Karatepe provenivano da questa città. Gli altri artigiani della zona, che hanno imparato a produrre utensili in legno da Hasan, non producono i cucchiaini ma sono cucchiai e contenitori. Nella produzione dei cucchiai e cucchiaini di legno Hasan ha sviluppato

Metodologia Questo contributo è stato strutturato in modo da raccogliere dati su: origine dell’attività a Karatepe, modalità di apprendimento, conoscenza e ricerca della materia prima, tecnologia disponibile e scelte umane, organizzazione sociale dell’attività, risvolti dell’attività nel contesto sociale di appartenenza. A Karatepe-Aslantas è stato possibile incontrare il signor Hasan Karaca che ci ha aiutati a comprendere le tecniche, gli arnesi, le scelte umane e i i gesti necessari per alcuni processi di trasformazione del legno. Con H. Karaca abbiamo dialogato sulla necessità della produzione dei manufatti in legno, tra cui i cucchiai e i cucchiaini, e sulle valenze passate e attuali di questa produzione. I dati qui presentati si basano principalmente sull’intervista ad H. Karaca e sulle osservazioni raccolte durante una sessione di lavoro pratico durata in totale due giorni. Le informazioni sono state raccolte in forma scritta e i vari momenti dell’attività, per quanto possibile, sono stati documentati attraverso la ripresa fotografica. E’ stato possibile campionare alcune delle entità vegetali (bosso, platano e sandalo) utilizzate durante l’attività pratica.

Fig.2. Manico bifido di cucchiaio con estremità superiori scolpite a testa di uccello.

un suo modello molto originale che presenta il manico scolpito a forma di testa di uccello (Fig. 2). Per questo tipo di realizzazione Hasan nel passato è stato molto contestato sia dalla sua comunità sociale sia all’interno della sua famiglia perché non si era attenuto alla norma aniconica tradizione spirituale sufica che è relativa al divieto di riprodurre qualsiasi essere vivente. Dopo un tormentato periodo interiore, in cui per lungo tempo non ha prodotto cucchiaini in legno, Hasan ha trovato la soluzione. Ha ripreso a scolpire il manico a forma di uccello ma privando la rappresentazione o degli occhi (Fig. 2) o del becco. Una produzione così difettosa non può essere più interpretata come una sfida verso il creatore. I cucchiai in legno variano in grandezza e forma e hanno nomi diversi in relazione al loro uso e cibo, e secondo la letteratura etnografica reperibile in Turchia (Eren 1984, p. 9) per la loro manifattura vengono utilizzati nove utensili: testere (sega), keser (accetta), bıçak (punteruolo), eğdi e İç sıyırgı (due diversi tipi di sgorbia), yassı eğe (raspa piatta), sistire (pialla), törpü (lima semplice) e üçgen eğe (raspa semitonda). La raspa, a differenza della lima, presenta sulla superficie della piastra di metallo che la compone molti denti grandi e sporgenti; il suo impiego serve a togliere il legno in eccesso, per arrotondare gli

Risultati La produzione tradizionale di cucchiai in legno, in turco Kaşık, è ancora diffusa in Turchia in diverse località tra cui Akseki, Kaş, Gediz, Geyve, Tarakli, Bolu, Kastamonu, Eskişehir, Konya, Kayseri, Karatepe. La produzione dei manufatti in legno a Karatepe Arslantas è una attività domestica eseguita principalmente da artigiani di sesso maschile, negli ultimi anni si è un pò intensificata visto l’interesse e la richiesta dei turisti verso cucchiai, cucchiaini e contenitori in legno. Hasan è l’unico nel villaggio che produce anche i cucchiani, poiché lui ha imparato questa tradizione direttamente da suo nonno e sin dalla sua infanzia ha prodotto per suo uso personale ma anche per la sua comunità manufatti in legno. Il nonno di Hasan proveniva da Kahramanmaras, una città un pò più ad est rispetto Adana, ma non molto distante. A Kahramanmaras la tradizione di produrre cucchiai in legno è antica, e nel passato, prima che Hasan imparasse 142

Recenti ricerche sulla produzione di utensili lignei a Karatepe-Aslantas, Turchia spigoli o per spianare una superficie irregolare. Poiché lascia una superficie molto ruvida dopo è necessario utilizzare la lima. I tipi di legno tradizionalmente utilizzati nell’area di Aslantas e nelle altre aree di produzione dei cucchiai sono i seguenti: bosso - Buxus sempervirens (nome comune in turco Sim sir), sandalo- Arbutus andrachne (Sandal è il nome comune ricordato da Hasan), pero - Pirus communis (nome comune in turco Armut), ginepro - Juniperus sp. (nome comune in turco Ardıç), quercia - Quercus sp. robur (nome comune in turco Meşe), carpino - Carpinus sp. L. (nome comune in turco Gűrgen), platano - Platanus orientalis L. (nome comune in turco Çinar) e olmo - Ulmus carpinifolia (nome comune in turco Karaağaç). Il legno di bosso, di colore giallo, ha una grana molto fine, con caratteristiche di durabilità e indeformabilità. A differenza delle foglie e della corteccia, che contengono lo alcaloide bussina, l’uso del legno di bosso non presenta rischi di tossicità. La scelta del legno usato da Hasan è comunque determinata dai seguenti fattori: 1) reperibilità in zona, 2) tipo di fibra, 3) densità del legno. Hasan preferisce utilizzare legni compatti, facili da intagliare, senza nodi e venature. Generalmente preferisce lavorare il legno di bosso, di quercia ma anche l’olmo ed il legno di sandalo. Il taglio del legno avviene principal-

mente in inverno perché il lavoro di produzione dei cucchiai viene svolto preferibilmente nelle giornate fredde e piovose, quando le attività dei campi richiedono minor impegno. Hasan ha un suo luogo di lavoro speciale, poco distante dalla sua casa, per il taglio e la sbozzatura del legno (Fig. 3). Nel momento in cui si è svolta questa attività era possibile riconoscerlo come un luogo di lavorazione del legno per la grande quantità di trucioli accumulata tutt’attorno. Hasan ci ha però informate che l’accumulo di trucioli, quando diventa secco, viene periodicamente rimosso per alimentare i focolari domestici, rendendo quindi molto evanescenti i residui che potrebbero essere

Fig. 4. Fase inizale del lavoro di produzione.

indicatori di questa attività. Hasan posiziona il ramo di legno su un piano orizzontale, anch’esso in legno, e lo divide longitudinalmente utilizzando una lunga lama in ferro e un percussore in legno (Fig. 4). E’ molto impor-

Fig 3. Area di lavoro per la produzione dei cucchiai e cucchiaini in legno.

Fig. 5. Sagomatura del cucchiaio mediante accetta.

143

M. R. Iovino, C. Altinbilek

Fig. 6. Sagomatura del cucchiaio mediante accetta.

Fig. 8. Scavo della bocca del cucchiaio mediante ascia a margine concavo.

tante nello svolgimento di questa operazione non segare ma spaccare il ramo in due. Si rimuove la corteccia e si inizia a dare la forma al cucchiaio (Fig. 5, 6). In questo caso Hasan procede sempre con l’accetta, che presenta un margine di forma diritta, ma ci informa che è possibile eseguire queste fasi di lavoro anche usando un coltello. Pertanto, la sbozzatura del manico viene eseguita con l’ausilio di una lama in ossidiana ( Fig.7). Lo scavo della

bocca del cucchiaio è eseguito con l’ascia a margine concavo (Fig. 8), ma Hasan - talmente abile a proseguire nel lavoro solo con l’uso dell’ascia - dà anche dimostrazione dello scavo tramite l’utilizzo di una sgorbia (Fig.9). Ultimata la realizzazione della forma del cucchiaio è necessaria una fase di stagionatura prima di passare alle due operazioni conclusive: la levigatura e la protezione della superficie con olio. La levigatura è una operazione

Fig. 9. Scavo della bocca del cucchiaio tramite l’utilizzo di una sgorbia.

Fig. 10. Levigatura della superficie tramite un frammento di vetro.

Fig. 7. Sagomatura del manico del cucchiaio mediante lama in ossidiana.

144

Recenti ricerche sulla produzione di utensili lignei a Karatepe-Aslantas, Turchia arboree che ha già subito dei cambi e che è nella fase di subirne degli altri. La scelta del tipo di legno da lavorare dipende dal contesto ambientale ma alcune specie sono preferite rispetto ad altre perché facilitano le operazioni di lavoro. L’attività non è di tipo continuativo ma sfrutta tempi di stasi dalle altre attività svolte dagli individui di sesso maschile, durante il periodo invernale. Hasan non utilizza come riportato in bibliografia tutti gli utensili tradizionalmente preposti alla produzione dei cucchiai e/o cucchiaini, sia perché sceglie legni a grana compatta sia perchè egli ha oramai acquisito una tale manualità da effettuare quasi tutte le operazioni principali con l’accetta e l’aiuto di un coltello. La trasformazione della tecnologia di produzione iniziale è già in atto. L’utilizzo anche se ormai teorico- di raspe e lime e la levigatura finale con il vetro- anche se in procinto di essere sostituito dalla più pratica carta vetrata- si è rivelato di particolare interesse per formulare delle nuove ipotesi di utilizzo per la sperimentazione relativa alla comprensione delle potenziali funzioni di utensili in vetro vulcanico (ossidiana), in particolare quelli provenienti da Cayonu. Gli studi archeobotanici della Turchia orientale (Bottema, van Zeist, 1981) hanno contribuito a tracciare un quadro sulla vegetazione del passato. Le risorse arboree hanno giocato un ruolo molto importante nell’ambito della sussitenza umana, ma da un punto di vista archeologico l’interpretazione circa le tecnologie e la produzione delle risorse lignee rimane spesso molto ipotetica. L’evidenza etnografica della lavorazione del legno a Karatepe- Aslantas e l’attestazione di un utilizzo specifico di materiale in vetro nelle fasi finali ha dato il via a un vasto programma di sperimentazione, ancora in corso, diretto alle specie arboree mediterranee potenzialmente utilizzate dai gruppi umani preistorici per la preparazione di manufatti e utensili. Infine, e comunque a nostro avviso molto importante, è da evidenzire l’approccio ideologigo al lavoro da parte di Hasan. Trasformando il legno Hasan non vuole creare solo un utensile funzionale ma desidera trasmettere alla materia una forma che sia espressione della sua cultura e dell’ambiente ricco di stimoli naturali in cui vive. L’ innovazione di Hasan, la raffigurazione del manico a forma di testa di uccello, è diventato un bisogno incontrollato che gli ha creato disagi sociali nel contesto religioso di appartenenza, ed è stato grazie alla sua capacità di approfondire il processo simbolico che Hasan ha trovato la giusta mediazione tra i concetti del suo territorio e i concetti della comunità di appartenenza.

necessaria per evitare che possano staccarsi frammenti di legno, soprattutto per i cucchiai che si usano per portare il cibo alla bocca, ma anche per far durare a lungo il cucchiaio. Hasan ci dimostra le ultime due operazioni su dei cucchiaini già stagionati, prodotti l’inverno precedente. Egli in genere utilizza un frammento di vetro (Fig. 10) per levigarne la superficie, il vetro a suo ricordo era il materiale utilizzato tradizionalmente per la levigatura delle superfici di oggetti in legno. Questo è un sistema che sta cadendo in disuso perché, ci informa Hasan, i giovani artigiani, quando ve ne è la disponibilità, preferiscono utilizzare della carta vetrata. L’attestazione dell’utilizzo del vetro nella fase finale di lavorazione di manufatti in legno è stata riscontrata anche in Grecia, Tunisia e Sicilia (Uccello, 1992). La lucidatura della superficie dei cucchiaini con l’olio (Fig. 11) ha la funzione di proteggere il legno da una eccessiva disidratazione che ne causerebbe la spaccatura. E’ infatti buon uso delle donne del villaggio provvedere alla manutenzione periodica con l’olio dei manufatti in legno di uso quotidiano.

Discussione Questa esperienza ci ha permesso di conoscere una attività tradizionale di sfruttamento delle risorse naturali

Ringraziamenti

Hasan Curaça, Halet Çambel, Murat Akman e Martina SickerAkman hanno costituito le fondamenta per l’inizio di questa ricerca. Grazie alla loro disponibilità, alla loro capacità e desiderio di trasmissione della loro conoscenza. La çorba di menta e yougurth del cuoco della missione di Karatepe, così come l’ospitalità ricevuta, resta indimenticabile. Grazie ad Isabella Caneva per il supporto di sempre.

Fig. 11. Lucidatura della superficie dei cucchiaini con olio.

145

M. R. Iovino, C. Altinbilek Bibliografia Akman, Y., 1982. Climats et bioclimats méditerranéens en Turquie. In: Quezel ed. Définition et localisation des écosystémes méditerranéens terrestres, Colloque de Saint-Maximim (France), 16.-20/11/981. Ecologia Mediterranea (Marseille), Tome VIII, Fasc. 1/2: 73-87. Bottema, S., Van Zeist, W., 1981. Palynological evidence for the climatic history of the Near East, 50000-6000 BP. Philadelphia: University of Pennsylvania Press Çambel, H., 1948. Karatepe: an archaeological introduction to a recently discovered Hittite site in Southern Anatolia. Oriens 1, 147-162. Çambel, H., 1996. Archaeology in Turkey: A cospectus of recent evidence on the prehistoric and early historic periods. Archaeometry 29. (The Proceedings of the 29th International Symposium on Archaeometry, Ankara, 9-14 Mayıs 1994): 339- 350. Eren, N., 1984.   Kaşik ve Kaşikcilik. Arkeoloji ve Sanat Yayınları, El Sanatları, Folklor ve Etnografya Dizisi, 2, 3-32. Türkçe. Uccello, A., 1992. La civiltà del legno in Sicilia,. Palermo: Cavallotto editore.

146

Experimental charcoal-burning with special regard to anthracological wood diameter analysis T. Ludemann

University of Freiburg, Institute of Biology II,Dept. of Geobotany,Schaenzlestrasse 1, D-79104 Freiburg, Germany [email protected]

Abstract A standardized anthracological analysis was developed and has been applied for several years in many studies, including a specific wood diameter determination. Thereby the individual charcoal fragments were sized by the curvature of the tree rings and by the angles of the rays to each other. Five diameter size-classes were distinguished, and a size-class distribution was obtained for each sample and each taxon. Based on this distribution, a single mean value was calculated, in order to facilitate the comparison of samples. Specific distributions of the charcoal fragments to the size-classes were found and could be used to characterize and classify the historic charcoal samples anthracologically. Information on the dimensions of the wood exploited in the past was deduced. In order to verify the method and the interpretation, samples from experimental charcoal-burning done in the traditional way were analysed. Hereby, the wood used as fuel in the kiln as well as that used for the kiln construction was measured in detail, and its theoretical distribution in diameter size-classes was calculated. After finishing the burning and removing the charcoal produced from the site of the experimental kiln, charcoal samples were taken in the same way as at the ancient sites. The results of fuel and construction wood measurement and volume calculation were compared with the anthracological results. Comprehensive information about the wood of the experimental kiln was recorded by charcoal analysis. The quantitative differences in proportions and dimensions of the tree taxa used could usually be clearly deduced from the anthracological results, with a few exceptions. Experimental charcoal-burning indicates that the method applied in the analyses of historic samples is a sensitive and specific tool to establish detailed quantitative information about the wood taxa combinations and dimensions used in the past.

Introduction For several years remains of historic mining and charcoalburning in the Black Forest have been analysed systematically, especially charcoal from hundreds of kiln sites. For this, a standardized anthracological method was developed, with special regard to the diameters of wood exploited in the past. Diameter size-class distributions and mean diameter values were determined for most of the charcoal samples and the taxa found. This method was applied in many studies and for large numbers of charcoal remains (Ludemann 1995, 1996, 1999, 2001; Ludemann and Britsch 1997; Ludemann and Nelle 2002; Nelle 2002a, 2002b; Noelken 2003, 2004). From anthracological analyses we attempt to deduce information on the taxa composition and dimensions of the exploited wood. But, in fact, we do not know the real compositions and dimensions of the wood from which the charcoal remains come. Up to now the calculated mean diameter values were only relative values. The calculation was done in order to simplify the comparison of samples and to describe the relation of the wood dimension of different taxa and samples within an order of magnitude. First indications of the quantitative relation of the diameter of the used wood and the anthracologically determined diameter were given by Ludemann (1996), Ludemann and Nelle (2002) and Nelle (2002a; 2002b). In the last few years remains of recent and experimental charcoal-burning were studied, whereby the used wood was known, in order to verify the interpretations of the historic material. The main questions were:

- Which diameter size-class distribution of the charcoal remains corresponds to a specific dimension of the exploited wood? - Which wood diameters were exploited in the past? We want to verify the quantitative relation between anthracologically determined size-class distributions and mean diameter values on the one hand and the actually used wood diameters on the other. In this study an experimental kiln was burned and analysed by the same standardized method, which was applied at the historic sites. Before burning, the wood used as fuel in the kiln as well as that used for the kiln construction was measured in detail and its theoretical distribution to diameter size-classes was calculated. Based on this information, the size-class distributions and mean diameter values were calculated for the charcoal theoretically arising from the used wood. These calculations were compared with the actual anthracological results of the experimental kiln.

Material and method Experimental charcoal-burning, study site Experimental charcoal-burning was conducted by the communal forest administration of the city of Freiburg, Southwest-Germany. The experimental kiln site was located in a small valley neighbouring the city at the edge of the Black Forest. Before charcoal-burning, there were no charcoal remains in the soil and at the ground surface

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

T. Ludemann

Fig. 1. Base and centre construction of the experimental kiln. Kiln centre with three small Abies trunks (vertical) and kiln base with large Pseudotsuga boards and some medium-size trunk sections of Fraxinus and Alnus beneath. 29 April 2002.

Fig. 2. The experimental kiln. Fuel wood of Fagus, Fraxinus, Quercus, Abies and Prunus trunks. 29 April 2002.

1 7

6

5

2

Fig. 3. The experimental kiln. Green cover of Picea twigs and branches. 29 April 2002.

3

4

Fig. 4. The construction of the experimental kiln. 1 Fuel wood (Fagus, Fraxinus, Quercus, Abies, Prunus). 2-4 Kiln base: 2 Fraxinus and Alnus stem sections. 3 Pseudotsuga boards. 4 Abies branches. 5 Kiln centre, Abies trunks. 6 Green cover, Picea twigs. 7 Soil cover (Drawing by Nelle 1998, from Ludemann, Nelle 2002, modified).

of the site. The experimental charcoal-burning was done in the traditional historic manner, as transmitted by the last professional charcoal burners and especially from Mr. S. Riesterer (Muenstertal, Black Forest). Figs 1 to 4 show the experimental kiln in construction (figs 1-3) and schematically (fig. 4).

Cone volume: Vol_cone = Vol_cyl/3 = π(d/2)(d/2)l/3 Cone part volume: Vol_pcone = πl(r1*r1+r1*r2+r2*r2)/3

The wood used The wood used was measured in detail before charcoalburning and while building-up the experimental kiln together with the charcoal burner. Moreover, the charcoal-burning process was observed regularly (personal observations and experience during kiln site visits).

with d, diameter; l, length of cylinder, cone or cone part; r1 and r2, minimum and maximum radius of cone part. The distinguished diameter size-classes correspond to those distinguished in anthracological wood diameter determination (cf., section on anthracological analysis). The calculated volume distributions for wood cylinders and cones of specific diameters are given in Table 1, Table 2 and Fig. 5. Based on these regular volume distributions, the same calculations were done for the experimental kiln, namely for both its wood and - in a second step - its (theoretically) charred wood. In the latter calculations, the reduction of diameter and volume caused by the radial, tangential and transversal shrinking of wood during carbonisationeach about 20 % (Brockhaus 1931; 2001; Herder 1954; Schlaepfer and Brown 1948) - was taken into account-

Diameter size-class distribution of wood and charred wood In order to link anthracological results and wood use, the theoretical volume distribution of the used wood to diameter size-classes was calculated, based on the following calculation and mathematical formulas for cylinders, cones and parts of cones: Cylinder volume: Vol_cyl = π(d/2)(d/2)l 148

Experimental charcoal-burning with special regard to anthracological wood

100 90 80 70 60

%

50 40 30 20 10 0

r ne C o l ind e r ne y m C C o l ind e 5c ne r y m C c C o l ind e 0 ne er y 1 cm C C o l ind 0 ne r y 2 cm C C o l ind e 0 ne r 3 cm C y C o l ind e 40 cm C y 50

I

II

III

IV

V

Fig. 5. Diameter size-class distributions of wood cylinders (grey columns) and cones (black columns) with a diameter of 5, 10, 20, 30, 40 and 50 cm. Diameter size-class I: 0-2 cm; II: >2-3 cm; III: >3-5 cm; IV: >5-10 cm; V: >10 cm.

Sampling of charcoal remains A sampling strategy most similar to that used at the historic kiln sites was applied at the experimental kiln site after charcoal-burning. Charcoal fragments of sizes larger than 0.25 cm3 (corresponding to a mesh width of 5 mm) were collected by hand from all over the charcoal layer at the soil surface. Both charcoal remains and charcoal intended for sale were analysed.

V 2

I II

1

IV III

4

3

- 2 cm 2 - 3 cm 3 - 5 cm

Anthracological analysis

5 - 10 cm

The determination of wood taxa of the charcoal fragments follows Schweingruber (1990), using a stereoscope (Leica MZ 12) and an incident-light microscope (Zeiss Universal M III C) as well as a reference collection of charred known wood. The quantity of an established taxon was determined by the number of charcoal pieces found. Altogether, 2000 charcoal fragments with a total weight of 9547 g were analysed. For diameter determination the charcoal fragments were

> 10 cm

Fig. 6. Diameter template, diameter size-classes (I-V) and charcoal fragments to be sized (1-4, examples). Result of sizing (examples): Fragment 1, class II; fragment 2, class III; fragment 3, class IV; fragment 4, class V.

149

T. Ludemann

Diameter cm 1.0 2.0 2.5 3.0 4.0 5.0 7.5 10.0 12.5 15.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0

Volume-% in Diameter Classes I II III IV 100.0 100.0 64.0 36.0 44.4 55.6 25.0 31.3 43.8 16.0 20.0 64.0 7.1 8.9 28.4 55.6 4.0 5.0 16.0 75.0 2.6 3.2 10.2 48.0 1.8 2.2 7.1 33.3 1.0 1.3 4.0 18.8 0.4 0.6 1.8 8.3 0.3 0.3 1.0 4.7 0.2 0.2 0.6 3.0 0.1 0.1 0.4 2.1 0.1 0.1 0.3 1.5 0.1 0.1 0.3 1.2 0.0 0.1 0.2 0.9 0.0 0.1 0.2 0.8

SizeV

36.0 55.6 75.0 88.9 93.8 96.0 97.2 98.0 98.4 98.8 99.0

Volume* cm³ 79 314 491 707 1257 1963 4418 7854 12272 17671 31416 70686 125664 196350 282743 384845 502655 636173 785398

sized by the curvature of the annual growth rings and by the angles of the rays to each other, using a diameter template. Five wood diameter size-classes (I to V) were distinguished; class I: 0-2 cm; II: >2-3 cm; III: >3-5 cm; IV: >5-10 cm; V: >10 cm. In this way a size-class distribution of each taxon was obtained, based on the number of pieces in the individual classes. Based on the size-class distribution, a single value was calculated, termed the mean diameter mD = (nI+nII*2.5+nIII* 4+nIV*7.5+nV*15)/N; where nI to nV is the number of charcoal fragments in the respective size-class; and N, the total number of fragments per sample. The mean diameter can theoretically vary between 1 and 15 cm, if all fragments are in the first class or all in the fifth class. Indeed, successful determination requires charcoal samples with a sufficient number of fragments, which are large enough and suited wood-anatomically for this kind of sizing. 1911 charcoal fragments (of 2000 analysed from the experimental kiln) have been sized in this way; for 89 charcoal pieces the determination of a diameter size-class was not possible.

mD cm 1.0 1.0 1.5 1.8 2.8 3.2 5.6 6.4 9.5 11.2 12.9 14.0 14.5 14.7 14.8 14.8 14.9 14.9 14.9

Table 1. Volume distribution of wood to diameter size-classes I to V (class I: 0-2 cm; II: >2-3 cm; III: >3-5 cm; IV: >5-10 cm; V: >10 cm). Calculated for wood cylinders of a specific diameter. mD mean diameter (cf., section on anthracological analysis). *Cylinder length 100 cm.

Diameter cm 1.0 2.0 2.5 3.0 4.0 5.0 7.5 10.0 12.5 15.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0

Volume-% in Diameter Classes I II III IV 100.0 100.0 89.6 10.4 74.1 25.9 50.0 34.4 15.6 35.2 29.6 35.2 17.5 17.7 38.9 25.9 10.4 11.2 28.4 50.0 6.9 7.7 20.7 54.4 4.9 5.5 15.5 48.1 2.8 3.3 9.6 34.4 1.3 1.5 4.6 18.5 0.7 0.9 2.7 11.3 0.5 0.6 1.8 7.6 0.3 0.4 1.2 5.4 0.2 0.3 0.9 4.1 0.2 0.2 0.7 3.2 0.1 0.2 0.6 2.5 0.1 0.1 0.5 2.1

SizeV

10.4 25.9 50.0 74.1 84.4 89.6 92.6 94.5 95.7 96.6 97.2

Volume* cm³ 26 105 164 236 419 654 1473 2618 4091 5890 10472 23562 41888 65450 94248 128282 167552 212058 261799

Results Fuel and construction wood used (Table 3, Fig. 4, Appendix 1) (1) Trunk wood of Fagus sylvatica (37 %) and Fraxinus excelsior (23 %) was mainly used as fuel wood, in addition that of Quercus petraea and Abies alba, with 13 % each, and of Prunus avium (8 %; Table 3). The diameter of the used trunks was larger than 22 cm, with the exception of Prunus with diameters between 8 and 22 cm. The largest wood came from Abies and Fagus, with trunk diameters mainly larger than 70 cm. Details given in the Appendix 1. Logs of Abies have been used preferentially (a) to fill the centre of the kiln when starting the burning process and (b) to refuel the kiln while burning if necessary. The construction wood consisted of (2) 14 trunk sections of Fraxinus excelsior and 2 of Alnus glutinosa with diameters between 8 and 12 cm and a length of 130 cm, which lay underneath the Pseudotsuga boards, in order to guarantee a good air circulation while charcoal-burning they have been charred only partly, approximately 20 % of them, (3) Pseudotsuga boards of 3 cm thickness, from the periphery of large trunks (in German “Schwartenbretter”; sawing waste), that built an 8 m² wood layer onto which the fuel wood was piled up - the boards overlapped over half of the area (total volume: 0.03 m * 12 m² = 0.36 m³), (4) 12 m branches of Abies with a diameter of 5 cm at the average (kiln base periphery), (5) three small trunks of Abies with a diameter of 4, 5 and 6 cm and a length of 2.20 m (kiln centre) and one of Picea with a diameter of 5 cm and a length of 3 m (kiln base) and (6) the green cover of the kiln with 120 m branches and

mD cm 1.0 1.0 1.2 1.4 2.0 2.5 4.1 5.3 6.7 8.3 10.6 12.7 13.6 14.1 14.4 14.5 14.6 14.7 14.8

Table 2. Volume distribution of wood to diameter size-classes I to V (class I: 0-2 cm; II: >2-3 cm; III: >3-5 cm; IV: >5-10 cm; V: >10 cm). Calculated for wood cones of a specific diameter. mD mean diameter (cf., section on anthracological analysis). *Cone length 100 cm.

150

Experimental charcoal-burning with special regard to anthracological wood 40 W ood Charcoal

30

% 20

10

s nu Al

ce a Pi

us Pr un

Ps eu

do

ts ug

a

s ie Ab

rc us Q

ue

us Fr ax in

Fa

gu

s

0

Fig. 7. Tree taxa proportions of used wood (grey columns; cf., Table 3) and analysed charcoal (black columns; cf., Table 5) of the experimental kiln.

35 30 25 20

% 15 10 5 0 Fagus

Fraxinus

Abies

Quercus

Pseudotsuga

Prunus

Picea

Alnus

V III IV II I

Fig. 8. Distribution of used wood volume (charred; grey columns) and analysed charcoal fragments (black columns) to taxa and diameter size-classes I to V. Diameter size-class I: 0-2 cm; II: >2-3 cm; III: >3-5 cm; IV: >5-10 cm; V: >10 cm. Based on Tables 4 and 5.

151

T. Ludemann twigs of Picea with a maximum diameter of 5 cm and an average diameter of 2.5 cm (main branches). The proportion of Picea wood of the green cover was estimated in two ways, (a) calculating a hemispherical area of 25 m² (radius 2 m) and a layer thickness of 40 cm with a twig density of 1 %, (b) 40 branches of 3 m length and 5 cm maximum diameter, both resulting a total wood volume of about 0.08 m³. The finer branching of the twigs was not taken into account in this calculation. Numbers (1) to (6) correspond to those in Fig. 4.

Abies centre wood (not present in class IV after charring; cf., Tables 3 and 4). Consequently, the effects established for mD-values of large wood were quite moderate; the larger the wood and the mD-values were, the smaller was the theoretical effect of carbonisation on the mD-value (smallest relative differences). The opposite was the case for small diameters and small mD-values. In addition, effects of the size-class limits, which were determined subjectively, were visible, for example for construction wood of Abies and Picea (reduction of mD-values from both 3.2 and 3.7 cm to 2.8 cm) and of Fraxinus and Alnus (reduction of mD-value from 8.1 to 5.9 cm). The calculated total volume reduction of nearly 50 %, from 7.5 m³ to 3.8 m³, corresponds quite well with the general values given in the literature (Brockhaus 1931, 2001; Schlaepfer and Brown 1948).

Diameter size-class distribution and mean diameter of the used wood The calculated volume distributions to wood diameter size-classes and the corresponding mD-values are given in Table 3. In the case of Picea cover material, the calculation was done for wood cones; for all others for wood cylinders. Pseudotsuga and all fuel wood taxa reach mD-values of more than 14 cm, with the exception of Prunus (mD 11.7 cm). Corresponding to the different size of the used wood (cf., Appendix 1), the mD-values increased from Prunus, Quercus and Fraxinus to Fagus, Abies and Pseudotsuga. Wood of Pseudotsuga was exclusively of the largest sizeclass, resulting in the maximum mD-value of 15 cm. The construction wood of Picea and Abies mainly reached mD-values around three cm. Calculation for 8 to 12 cm Fraxinus and Alnus trunk parts resulted in mD-values of 8 cm. The dimensions of Pseudotsuga construction wood have been already mentioned.

Charcoal and anthracological analysis (Table 5) In the charcoal material analysed (2000 fragments) we found 9 wood taxa. These were, with decreasing frequency, Fagus, Fraxinus, Quercus, Abies, Pseudotsuga, Prunus, Picea, Alnus and Salix. More than 60 % of the material belonged to Fagus and Fraxinus the most frequently used fuel wood. Third place is taken by Quercus, Abies and Pseudotsuga, with similar proportions, around 10 %. The smallest proportion of a fuel wood taxon was established for Prunus with 6 %. Few charcoal pieces were found of Picea and Alnus. In addition one piece of Salix was established. Most of the wood taxa had most of their fragments in the largest diameter size-class. Considering the size-class distributions and mD-values, Pseudotsuga and Fagus obtained the highest values, about 15 cm, with the largest proportions of pieces in the largest class. Second diameter position is taken together by Quercus, Fraxinus and Abies, followed by Prunus (Table 6). Alnus had a similar proportion in the fourth and fifth diameter class - but only a few pieces -, resulting in a mD-value of 8.5 cm. The smallest material found was of Picea (mD-value 1.7 cm).

Diameter size-class distribution and mean diameter of charred wood and charcoal Calculation for charred wood (Table 4) The (theoretical) size-class distributions and mD-values calculated for wood cylinders of the specific diameters of the experimentally used (charred!) wood are given in Table 4, assuming 20 % shrinking in the three spatial dimensions (radial, tangential, transversal) and thereby taking in to account the actual diameter and volume reduction. Most of the fuel wood of Prunus, diameters 8 to 22 cm (minimum size of fuel wood), after carbonisation theoretically belongs into the second largest and the largest diameter size-classes. The other (larger) fuel wood diameters calculated (> 22 cm) have a distinct volume maximum further in the largest size-class (Fagus, Fraxinus, Abies, Quercus; grey columns; Fig. 8). In contrast, distinct shifts in size-class distribution after charring were found for construction wood of smaller diameters, especially at size-class limits. The latter is the case for example for Alnus and Fraxinus construction wood (not present in class V after charring) as well as

Discussion Wood use and anthracological results General features Comprehensive detailed information about the fuel and construction wood of the experimental kiln has been recorded by the charcoal remains and their analyses. Especially the important quantitative differences in proportion and dimension of the tree taxa used could be clearly deduced from the anthracological results of the charcoal samples (Figs 7 and 8, Table 6); 152

Experimental charcoal-burning with special regard to anthracological wood Taxon

Use

Volume total % dm³ Fagus Fuel wood 36.6 2735 Fraxinus Fuel wood 22.8 1702 Quercus Fuel wood 13.1 976 Abies Fuel wood 13.0 970 Prunus Fuel wood 7.6 571 Total Fuel wood 93.1 6953 Pseudotsuga Kiln base 4.8 360 Picea Kiln cover 1.1 79 Fraxinus Kiln base 0.4 30 Abies Kiln base 0.3 24 Abies Kiln centre 0.2 13 Picea Kiln base 0.1 6 Alnus Kiln base 0.1 4 Total Kiln construction wood 6.9 515 Picea Total 1.1 84 Abies Total 13.5 1007 Fraxinus Total 23.2 1732 Total Kiln wood 100.0 7468

Volume (dm³) in Diameter size-classes I II II IV V 2.1 2.7 8.6 40.2 2680.9 3.8 4.7 15.1 70.7 1607.6 2.8 3.5 11.3 53.0 904.9 0.5 0.6 1.9 8.8 958.0 9.1 11.4 36.4 165.2 348.8 18.3 22.9 73.3 337.9 6500.1 0.0 0.0 0.0 0.0 360.0 27.6 23.2 27.6 0.0 0.0 1.1 1.4 4.6 16.3 6.3 3.8 4.7 15.1 0.0 0.0 2.1 2.6 6.7 1.9 0.0 0.9 1.2 3.8 0.0 0.0 0.2 0.2 0.7 2.3 0.9 35.7 33.4 58.5 20.5 367.2 28.6 24.4 31.4 0.0 0.0 6.3 7.9 23.7 10.7 958.0 4.9 6.1 19.7 87.0 1613.9 54.1 56.3 131.8 358.5 6867.3

mD cm 14.8 14.5 14.4 14.9 11.7 14.4 15.0 2.5 8.1 3.2 3.7 3.2 8.1 11.7 2.6 14.5 14.4 14.3

Table 3. Volume distribution of wood to taxa and diameter size-classes I to V. Calculated for fuel and construction wood of the experimental kiln. Diameter size-class I: 0-2 cm; II: >2-3 cm; III: >3-5 cm; IV: >5-10 cm; V: >10 cm. mD mean diameter value (cf. section on anthracological analysis).

Taxon

Use

Fagus Fuel wood Fraxinus Fuel wood Quercus Fuel wood Abies Fuel wood Prunus Fuel wood Total Fuel wood Pseudotsuga Kiln base Picea Kiln cover Fraxinus Kiln base Abies Kiln base Abies Kiln centre Picea Kiln base Alnus Kiln base Total Kiln construction wood Picea Total Abies Total Fraxinus Total Total Kiln wood

Volume total % dm³ 36.6 1400 22.8 871 13.1 499 13.0 497 7.6 292 93.1 3560 4.8 184 1.1 40 0.4 15 0.3 12 0.2 7 0.1 3 0.1 2 6.9 264 1.1 43 13.5 515 23.2 887 100.0 3824

Volume (dm³) in Diameter size-classes I II II IV V 1.7 2.1 6.9 32.2 1357.2 3.0 3.8 12.1 56.5 796.0 2.3 2.8 9.0 42.4 442.9 0.4 0.5 1.5 7.1 487.1 7.3 9.1 29.2 125.4 121.3 14.7 18.3 58.6 263.6 3204.5 0.0 0.0 0.0 0.0 184.3 20.1 13.8 6.3 0.0 0.0 0.9 1.1 3.7 9.5 0.0 3.0 3.8 5.3 0.0 0.0 1.7 2.1 3.1 0.0 0.0 0.8 0.9 1.3 0.0 0.0 0.1 0.2 0.5 1.4 0.0 26.6 21.9 20.1 10.9 184.3 20.9 14.8 7.6 0.0 0.0 5.1 6.3 9.9 7.1 487.1 3.9 4.9 15.7 66.1 796.0 41.2 40.2 78.8 274.5 3388.8

mD cm 14.7 14.3 14.0 14.8 9.9 14.1 15.0 2.0 5.9 2.8 2.8 2.8 5.9 11.4 2.0 14.4 14.1 14.0

Table 4. Volume distribution of charred wood/charcoal to taxa and diameter size-classes I to V. Calculated for the wood of Table 3 after carbonisation (radial, tangential and transversal shrinking, each 20 %). Diameter size-class I: 0-2 cm; II: >2-3 cm; III: >3-5 cm; IV: >5-10 cm; V: >10 cm. mD mean diameter value (cf. section on anthracological analysis).

153

T. Ludemann Taxon

Number of charcoal pieces Total in Diameter size-classes % n I II III IV Fagus 34.0 680 0 0 0 17 Fraxinus 27.1 542 0 0 11 70 Quercus 11.5 229 0 0 3 19 Abies 10.4 207 3 7 7 4 Pseudotsuga 9.2 184 0 2 0 0 Prunus 5.7 114 0 0 3 28 Picea 1.5 29 18 8 3 0 Alnus 0.7 14 1 0 3 5 Salix 0.1 1 0 0 1 0 Total 100.0 2000 22 17 31 143

V 642 450 167 183 178 74 0 4 0 1698

? 21 11 40 3 4 9 0 1 0 89

mD (cm) 14.8 13.8 14.1 13.8 14.9 12.7 1.7 8.5 4.0 14.0

Table 5. Frequency distribution of charcoal fragments to taxa and diameter size-classes I to V. Diameter size-class I: 0-2 cm; II: >2-3 cm; III: >3-5 cm; IV: >5-10 cm; V: >10 cm. mD mean diameter value (cf., section on anthracological analysis).

Taxon

Pseudotsuga Fagus Abies Fraxinus Quercus Prunus Alnus Picea Total Reference

Average wood diameter Construction Fuel wood wood cm cm (>50) 71.4 5.0 90.7 10.0 42.3 36.1 15.4 10.0 2.6 3.1* 32.6 Appendix 1

mD mean diameter value Wood Charred total wood cm cm 15.0 15.0 14.8 14.7 14.5 14.4 14.4 14.1 14.4 14.0 11.7 9.9 8.1 5.9 2.6 2.0 14.3 14.0 Table 3 Table 4

Charcoal cm 14.9 14.8 13.8 13.8 14.1 12.7 8.5 1.7 14.0 Table 5

Rank

1 2 3 4 5

Table 6. Measured diameters of used wood and calculated mD-values of wood, charred wood and charcoal. mD mean diameter value (cf., section on anthracological analysis). *without Pseudotsuga.

- mainly Fagus wood with large diameters was used and recorded anthracologically, as well as - a smaller proportion (second position) and smaller diameters of Fraxinus; - third position in an order of proportion was taken by Abies and Quercus together, with mean diameter values similar to that of Fraxinus; considering Abies, it was due to its having a mixture of smaller and larger wood diameters (Table 6); - than the exception of Pseudotsuga with a considerable difference in wood and charcoal proportions (Fig. 7); while indeed only very large wood was used and correspondingly recorded in the charcoal remains (Table 6); - next position was taken by Prunus, considering both its frequency as well as its mD-value;

- smallest proportions and diameters were used and established anthracologically for Picea and Alnus, especially the very small diameters of Picea twigs (kiln cover material). Further evidence of the high sensitivity of the method is given by a specific detail of the diameter distribution of Abies charcoal. Its distribution showed a second poor accumulation of charcoal pieces in size-classes II and III (Table 5; Fig. 8), caused by the additional use of small Abies wood for kiln construction, besides the main use of large trunk wood as fuel (Tables 3 and 4). In contrast, construction wood of small diameters was not used for the other fuel wood taxa (Fagus, Quercus, Fraxinus, Prunus). Consequently, we found fewer or no fragments of them in size-classes I, II and III. 154

Experimental charcoal-burning with special regard to anthracological wood Exceptions and methodological aspects

very small total volume of small wood was used and (2) often the sized charcoal fragments ranged over more than one diameter size-class, of which only the largest was registered (cf. Fig. 6). Recording every involved size-class in such cases could not solve the problem, because always different volume proportions of a charcoal fragment belong to the different size-classes. This general methodological problem cannot be solved without considerable additional effort. This results in higher mD-values of charcoal samples and must taken into account in the interpretation. The larger the analysed fragments, the larger the effect could be. The larger the diameter of the used wood, the smaller this effect is; considering large wood (diameter > 30 cm; mD > 13 cm) the difference is very small (Fagus, Quercus; Table 6). The effect is especially strong at the limits of the size-classes and, in addition, for wood diameters between 10 and 15 cm, because pieces of the fifth class were calculated with a mean diameter of 15 cm (determined subjectively), also if their diameters were between 10 and 15 cm. Examples of this effect are given by Prunus and Alnus mD-values (Table 6). Picea diameters: The calculated mD-value of charred Picea wood was larger than the one determined anthracologically, because finer branching was not taken into account in the calculation. The finer differentiation of (trunk) wood much larger than 10 cm is not limited in principle by the minimum diameter of the largest size-class of only 10 cm. Indeed, it is possible to deduce quantitative information about larger wood dimensions, because wood of a distinct diameter has a specific volume distribution into diameter size-classes (Tables 1 and 2). The quantitative resolution depends only on the number and size of fragments available of a charcoal sample and fitted wood-anatomically for the determination. Of course the number of fragments must be very high (hundreds) to deduce such information. Considering the similar mD-values of Abies, Fraxinus and Quercus, it became evident that different combinations of wood diameters can result in the same mD-value.

Besides the general features pointed out above some exceptions and methodological aspects have to be discussed. Tree taxa proportions The largest difference between used wood and anthracologically recorded charcoal proportions was established for Pseudotsuga. Its proportion in charcoal is much larger than in kiln wood (Fig. 7). Charcoal production requires much energy, which must be provided by the fuel wood itself. This energy and wood volume burn and are thus lost, resulting in a volume and energy reduction of 50 % of the total fuel wood (Brockhaus 1931, 2001; Herder 1954). We suppose that construction wood at the kiln base was not involved in this process with the same intensity like the fuel wood, so that a smaller proportion of the base construction wood was lost than that of the central parts of the kiln. Moreover, the spatial position at the kiln base may have led to an accumulation of charcoal of such construction wood underneath the kiln at the soil surface, where the charcoal samples were taken (Fabre and Auffray 2002). Of course the proportions of all base construction taxa (Pseudotsuga, Alnus, Fraxinus) were larger in charcoal samples than in wood used. Inversely, such effects at the kiln base would explain the fact, that the fuel wood taxa generally were less frequent in charcoal than in used wood. Independently for Abies fuel wood we expected a larger part to be burned completely and be lost, because it was mainly used to start the burning process and to refuel the kiln. Picea proportions: Charcoal samples of historic kiln sites often showed an accumulation of coniferous charcoal in the smallest size-class (Ludemann 1995, 1996, 2001; Ludemann, Britsch 1997; Ludemann, Nelle 2002; Nelle 2002b), interpreted as accumulated remains of charred cover material of the kilns. The same effect was established by the anthracological analysis of a recent kiln site, on which the kilns were covered traditionally by Picea and Abies twigs (Ludemann in press). Consequently, a moderate accumulation of remains of the cover material of the experimental kiln was expected and indeed recorded anthracologically. Salix was intentionally not used in the experimental kiln. The single Salix charcoal piece must be a contamination, originating from the close surroundings of the experimental kiln site, where some Salix shrubs and small trees grow.

Conclusion Experimental charcoal-burning and its anthracological analysis indicate that the applied method is a sensitive tool to establish comprehensive detailed information on the wood used. Not only qualitative information on the wood taxa used could be deduced from anthracological analyses but also quantitative data about the taxa compositions and dimensions used. Valuable additional information can be provided especially by anthracological diameter analysis. At the same time limits of the method could be recognized by experimental charcoal-burning. Further investigations should be undertaken on recent or experimental charcoal-burning of medium wood sizes (diameter 6 to 30 cm, mD-value 4 to 13 cm).

Diameter analyses In contrast to the theoretical calculation for charred wood, we found no charcoal fragments of Fagus, Fraxinus, Quercus and Prunus in size-classes I and II (Table 4 and 5, Fig. 8). This effect is caused by the fact that (1) only a 155

T. Ludemann Acknowledgements

Meilerplaetzen in den Suedvogesen. ZAM Zeitschrift fuer Archaeologie des Mittelalters 31 (2003). Bonn: Habelt. Schlaepfer, P., Brown, R., 1948. Ueber die Struktur der Holzkohle. EMPA-Bericht (Eidg. Materialpruefungsanstalt f. Industrie, Bauwesen u. Gewerbe) 153. Zuerich. Schweingruber, F.H., 1990. Microscopic wood anatomy. Structural variability of stems and twigs in recent and subfossil woods from Central Europe. Birmensdorf, Schweiz: Swiss Federal Institute for Forest, Snow and Landscape Research, 3 edn.

We are grateful to the communal forest administration of the city of Freiburg, Southwest-Germany, and especially to forester Philipp Schell, the young “experimental charcoal burner”, for conducting the experimental charcoal-burning and managing the construction of the kiln and the charcoal-burning process accurately. Moreover, we thank Randy Cassada (University of Freiburg) for linguistic editing.

References Brockhaus, 1931. Der Große Brockhaus. Handbuch des Wissens in 20 Baenden. Leipzig: Brockhaus, 15 edn. Brockhaus, 2001. Brockhaus. Die Enzyklopaedie in 24 Baenden. Leipzig: Brockhaus, 20 edn. Fabre, L., Auffray, J.C., 2002. An anthracological method for the study of charcoal kilns in relation to historical forestry management. BAR International series, 1063. Oxford: BAR Publishing, 193-199. Herder, 1954. Der Große Herder. Nachschlagwerk fuer Wissen und Leben. Freiburg: Herder, 5 edn. Ludemann, T., 1995. Zwei Kohlplaetze im Mittleren Schwarzwald. Mitteilungen badischer Landesverein Naturkunde Naturschutz N.F. 16, 319-334. Ludemann, T., 1996. Die Waelder im Sulzbachtal (SuedwestSchwarzwald) und ihre Nutzung durch Bergbau und Koehlerei. Mitteilungen des Vereins für forstiliche Standortskunde Forstpflanzenzuechtung. 38, 87-118. Ludemann, T., 1999. Zur Brennstoffversorgung einer roemischen Siedlung im Schwarzwald. Internationalia Archäeologie,9. Studia honoraria, 165-172. Rahden, Westfalen: Leidorf. Ludemann, T., 2001. Das Waldbild des Hohen Schwarzwaldes im Mittelalter. Ergebnisse neuer holzkohleanalytischer und vegetationskundlicher Untersuchungen. Alemannisches Jahrbuch 1999/2000, 43-64. Ludemann, T., Britsch, T., 1997. Wald und Koehlerei im noerdlichen Feldberggebiet, Suedschwarzwald. Mitteilungen badischer Landesverein Naturkunde Naturschutz, 487-526. Ludemann, T., Nelle, O., 2002. Die Waelder am Schauinsland und ihre Nutzung durch Bergbau und Koehlerei. Freiburger Forstliche Forschun, 15, 139. Ludemann, T., in press. Anthracological analysis of recent charcoal-burning in the Black Forest, SW Germany. BAR International Series, 1063. Oxford: BAR Publishing. Nelle, O., 1998. Waldstandorte und Koehlerei am Schauinsland (Suedschwarzwald). Thesis. Dept. Geobotanik. Institut für Biologie II, Universitaet Freiburg. Nelle, O., 2002a. Charcoal burning remains and forest stand structure - Examples from the Black Forest (south-west Germany) and the Bavarian Forest (south-east Germany). BAR International Series, 1063. Oxford: BAR Publishing. 201-207. Nelle, O., 2002b. Zur holozaenen Vegetations- und Waldnutzungsgeschichte des Vorderen Bayerischen Waldes anhand von Pollen- und Holzkohleanalysen. Hoppea, Denkschrift der Regensburger Botanischen Gesellschaft, 63, 161-361. Noelken, W., 2003. Holzkohleanalytische Untersuchungen zur Waldgeschichte der Vogesen im Tal von Miellin. Freiburger Universitaetsblaetter 160 (2), 111-118. Noelken, W., 2004. Holzkohleanalytische Untersuchungen an

156

Experimental charcoal-burning with special regard to anthracological wood Taxon Volume (dm³)

Picea

Alnus

Prunus Quercus Fraxinus Fagus Abies

Pseudotsuga

Construction wood

84

4

0

0

30

0

37

360

Fuel wood

0

0

571

976

1702

2735

970

0

Total kiln wood

84

4

571

976

1732

2735

1007

360

15.4

36.1

10.0 42.3

71.4

5.0 90.7

(> 50)

Diameter average (cm) Construction wood 2.6 10.0 Fuel wood Dimension and form Diameter Form Length (m) (cm) Construction wood

Volume (dm³)

4 5 5 6

Cylinder Cylinder 3.00 Cone 120.00 Cylinder

8 12 >50 Fuel wood

Cylinder Cylinder Board

8

Cylinder

2.00

11 12 13 14 15 16

Cylinder Cylinder Cylinder Cylinder Cylinder Cylinder

2.00 2.00 2.00 4.00 2.00 4.00

17 19 20

Cylinder Cylinder Cylinder

5.00 1.00 2.00

21

Cylinder

1.00

22

Cylinder

2.00

23 26 29 34 35

Cylinder Cylinder Cylinder Cylinder Cylinder

36 37 39 40 42

Cylinder Cylinder Cylinder Cylinder Cylinder

42

Cylinder

2.00

43 45 47 48 50 52 69 72 73 74

Cylinder Cylinder Cylinder Cylinder Cylinder Cylinder Cylinder Cylinder Cylinder Cylinder

1.00 2.00

74

Cylinder

87 90 95

Cylinder Cylinder Cylinder

2.20 14.20 2.20 0.26 0.26

1.82 1.82

360

1.00 1.00 1.00 1.00 1.00 1.00 2.00 1.00 2.00

1.00

1.00 1.00 1.00 1.00 3.00 1.00 1.00 1.00 0.83 0.50 0.50 0.50

Appendix 1. Construction and fuel wood of the experimental Kiln.

157

Gli accampamenti invernali e primaverili dei nomadi dell’Arkhangaï e dell’Ovorkhangaï settentrionale: i ricoveri per gli animali (Mongolia centrooccidentale) F. Lugli

Associazione Italiana di Etnoarcheologia E-mail: [email protected] Abstract A large portion of the Mongolian population is still nomadic. This ancient way of life, however, is being severely threatened by encroaching “modernism” and thus, there is an impelling need to study and document these nomadic traditions and models before they undergo drastic change or are tragically lost altogether. Thus, an ethnoarcheological mission headed by the Author has been underway in Mongolia since 2002 (sponsored by the Italian Ethno archeology Association and co-funded by the Foreign Affairs Ministry), whose aim is to document the various models of nomadism still extant in the country and the traces it has left behind. The northern regions of Arkhangaï and Ovorkhangaï, with an average altitude of 2,414 m, are covered by forests and vast expanses of steppe, rivers and lakes. Rich in pastures, it is considered excellent grazing land for livestock, prevalently yak, horses, sheep and goats. In these regions where winter temperatures can sometimes drop to as low as –50°C, the nomads travel, together with their livestock, an average of four times a year with the change of seasons. For transport, they use solid wooden carts put together by carving and fitting the joints, using a bare minimum of metal elements. In the winter and spring, the nomads camp on higher ground, while in the summer and autumn, they descend to the valleys to camp near river streams. In addition to the ‘ger’, the well-known circular tents made of heavy felt and used by the nomads all year round, some open-air pens are put up for the animals in the camps which are used in the warmer months, but no other mobile or semi-fixed structures are built. On the other hand, in the winter and spring camps which are used for several years, there are, in addition to the ‘ger’, some quite sturdy animal shelters, used above all for sheep and goats, consisting of partly covered solid wooden structures, as well as a small wooden pantry. While the tents leave very scant and hard to interpret traces, the animal shelters, once they are abandoned, can provide important clues for archeological records useful to the nomadic archeology of the steppe. Wood is extremely precious and often rare, even in the mountainous regions of MongoIia and is not to be wasted. Thus, the substantial use of wood and the techniques used in building ‘ger’ and animal shelters provide food for thought.

Premessa La Mongolia (Fig.1), dove il nomadismo si affermò a partire dall’età del bronzo, tutt’oggi rappresenta il modello di adattamento al territorio più diffuso all’esterno delle aree urbane. In questo paese, infatti, grande approssimativamente cinque volte la Francia, dove si stima che vivano attualmente circa 2.800.000 di persone con una densità media di 1,8 abitanti per kmq, circa la metà della popolazione vive di pastorizia spostandosi più volte sul territorio secondo il succedersi delle stagioni, smontando e rimontando ogni volta la tradizionale tenda di feltro ‘ger’, nonché le strutture per gli animali1 per le quali, contrariamente alle abitazioni descritte più volte da storici e viaggiatori (a partire da Marco Polo e Guglielmo di Rubruc), non si dispone di un ampio materiale a riguardo. Il sistema di vita nomadico, estremamente radicato nella cultura mongola, è oggi fortemente minacciato dalla ‘modernità che incalza, spesso con concetti totalmente estranei e in antitesi ai presupposti di una cultura non stanziale. E’ urgente, quindi, studiare e documentare il nomadismo mongolo, erede dei grandi imperi, prima che tradizioni e modelli insediamentali vengano drasticamente sovvertiti o tragicamente annientati. Dal 20022 è, quindi, in corso una missione etnaorcheologica in Mongolia promossa

dall’A.I.E. e diretta dalla scrivente che si propone di documentare i diversi modelli di accampamenti dei nomadi nelle diverse regioni. Gli accampamenti dell’Arkhangaï e dell’Ovorkhangaï settentrionale L’Arkhangaï e la parte settentrionale dell’Ovorkhangaï, regione famosa per il legname utilizzato per lo scheletro delle ‘ger’ dove sorgeva l’antica Karakorum3. sono situati sul versante settentrionale della catena montuosa Khangaï Nuruu della Mongolia centro-occidentale ad un’altitudine media di 2414 m. Queste regioni montuose, caratterizzate da immense distese di steppa e fiumi, ricche di pascoli, di laghi e anche foreste, sono considerate ottimali per il bestiame, costituito essenzialmente da yak, cavalli, pecore e capre. Qui nel 2000 erano presenti, infatti, circa 80.000 nomadi su 110.000 abitanti, in maggior parte dell’etnia dei Khalka che rappresentano all’incirca il 78,8% della popolazione della Mongolia. In queste regioni le temperature invernali possono scendere talvolta anche fino a –50°C mentre nel periodo più caldo dell’anno possono raggiungere anche più di + 30°C.

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

F. Lugli

Fig. 1. La Mongolia divisa per Aimak. L’Arkhangaï e la parte settentrionale dell’Ovorkhangaï, sono situati sul versante settentrionale della catena montuosa Khangaï Nuruu della Mongolia centro-occidentale ad un’altitudine media di 2414 m.

Dopo il 2000, si sono succeduti tre inverni estremamente duri nonché l’estate del 2002 che è stata particolarmente secca. A seguito di ciò i nomadi di queste regioni hanno subito gravi perdite, perso numerosi capi di bestiame e in alcuni casi sono stati costretti ad abbandonare il nomadismo e a cercare un’occupazione nei villaggi4. La tenda utilizzata in queste regioni è, come del resto in gran parte della Mongolia, la tradizionale ‘ger’ (Fig. 2), tipica dei popoli delle steppe5, utilizzata spesso anche all’interno delle aree urbane. Nel Khovsgöl settentrionale, invece, i pastori di renne, gli Tsaatan, utilizzano una tenda conica (urts) di pelle di renna e feltro che è simile al tipi americano (Bruno 2003), realizzata direttamente dai nomadi. Storicamente la diversa forma e struttura delle tende è stata da tempo posta in relazione a diverse categorie di utilizzo del territorio (Grousset 1939, p. 249-251). Di fatto il tipi è funzionale ad una maggiore mobilità6, spesso con una non frequente interazione con i gruppi stabili. La ‘ger’, invece, la cui struttura è piuttosto complessa e molto pesante, è realizzata di norma da un artigiano stabile che risiede in un villaggio o in una cittadina; questa produzione costituisce, quindi, un importante settore della lavorazione del legno in Mongolia. Ciò può essere un importante aspetto del rapporto di interazione tra i nomadi e i nuclei stabili della Mongolia che, insieme alla cultura materiale in genere, costituiscono un interessante spunto di analisi e discussione (Koryakova 2000, p. 13-17). La tenda, sempre a pianta circolare, ha un diametro variabile dai 6 agli 8 metri circa, ed un’altezza massima di

circa m. 2,50. Lo scheletro ligneo, è costituito da moduli pieghevoli (hana) che fungono da pareti perimetrali e da numerosi pali (uni) inclinati posti a raggiera che, poggiando tutt’intorno sui moduli, convergono in alto verso un cerchio ligneo centrale (toono) (attraverso il quale passa lo sfiato della stufa), in cui sono incastrati. Solo due o più pali verticali su cui è posizionato il cerchio, non infissi nel terreno, fungono da supporto al ‘tetto’ oltre all’appoggio delle pareti. La porta (haalga), montata in una intelaiatura con cardini, è sempre rivolta verso meridione. L’inclinazione della luce che entra dall’apertura centrale indica l’ora, in conformità al ciclo di 12 anni del calendario lunare7. Poiché il vento è spesso molto forte, le tende vengono rese più stabili da zavorre, solitamente costituite da possenti pietre, tenute da corde passanti da parte a parte al di sopra della tenda. Talvolta, in caso di vento eccessivo, le ‘ger’ possono venire legate le une alle altre (Tsultem 1988). La struttura lignea della tenda è rivestita da diversi strati di feltro, più o meno spessi e numerosi a seconda delle temperature. L’isolamento dal terreno è garantito da tappeti, da feltro o da pavimentazioni lignee lievemente sopraelevate tramite traversine. La zona interna antistante la porta d’entrata, è, invece, di norma, priva di copertura. Solitamente, nell’area di una ‘ger’ appena smontata si pongono alcune pietre per “non lasciare vuoto lo spazio utilizzato”. La mancanza di pali portanti, di elementi fissi e la copertura del terreno fa sì che rimangano poche tracce sul terreno dopo lo smontaggio e, quindi, un record archeologico di non facile lettura. Lo studio delle tracce lasciate dalle ‘ger’ 160

Gli accampamenti invernali e primaverili dei nomadi in Mongolia centro - occidentale una volta rimosse, è stato uno degli obiettivi della campagna svolta nel 20028. I Nomadi delle regioni centro-occidentali della Mongolia si spostano di media quattro volte l’anno con il cambiamento delle stagioni per mezzo di carri di legno pieno assemblato ad incastro, con uso limitatissimo di elementi metallici (Fig. 3). In inverno e in primavera si accampano in zone più elevate (tra i 1650 e i 1800 m slm) e discendono verso valle (intorno ai 1500 m slm) e, quindi, in prossimità del corso dei fiumi, in estate e in autunno (che è di fatto la stagione più secca dell’anno) (Fig. 4, 5). Questo fatto è facilmente spiegabile, come ben descrisse Guglielmo di Rubruc (2002, p. 13), poiché nei mesi estivi il fiume garantisce pascoli e acqua mentre nei mesi invernali la piana alluvionale diventa prima a rischio di inondazioni e poi un’unica lastra di ghiaccio o una distesa di neve assai profonda non atta agli animali9. A settembre, spesso, senza smontare gli accampamenti, alcuni componenti delle famiglie si trasferiscono con il loro bestiame in cerca di pascolo, fino a qualche decina di chilometri dal campo base, spostandosi a cavallo e dormendo in tende leggere da spostamento veloce. Gli accampamenti sono, di norma composti da un minimo di un nucleo familiare ad un massimo di sette/otto e ogni famiglia può possedere da un minimo di poche decine di animali ad un massimo di diverse centinaia. Negli accampamenti delle stagioni calde, oltre alle ‘ger’, vengono montati solo alcuni recinti per gli animali, privi di copertura, e non sono presenti altro genere di strutture, sia mobili che semi-fisse. Anche la ‘dispensa’ è localizzata in una ‘ger’, sebbene di dimensioni ridotte rispetto a quelle abitate. Senza ripercorrere una certa tendenza degli studi che considera, di fatto, ‘fragile’ la cultura materiale

Fig. 2. La ‘ger’ è costituita da uno scheletro ligneo coperto da feltro e tela. Le pareti sono costituite da moduli pieghevoli ed il tetto da lunghi pali inclinati posti a raggiera convergenti verso l’apertura centrale. Queste ‘ger’ non posseggono pali portanti infissi nel terreno.

Fig. 3. Gli spostamenti sono effettuati dai nomadi dell’Arkhangaï e dell’Ovorkhangaï utilizzando dei carretti lignei trainati dagli yak. I carri sono presentano due ruote di legno pieno protette da una fascia metallica collegate da un asse su cui poggia il piano effettivo del carretto. Scarsi sono gli elementi metallici utilizzati poiché, essenzialmente, le diverse parti che compongono la struttura sono assemblate ad incastro.

161

F. Lugli Il legno e i ricoveri lignei per gli animali invernali e primaverili: l’esempio del sito 3 (Huliin Shilin – Havarjaa). Negli accampamenti invernali e in quelli primaverili sostanzialmente, la presenza e la disposizione delle strutture sono essenzialmente simili10. Esistono però alcune importanti differenze a livello macroscopico: le postazioni primaverili sono ad una quota più bassa e sono quasi privi dell’isolamento di sterco e paglia posto tra i legni che costituiscono le pareti e le coperture delle strutture; i siti invernali, inoltre, più che quelli primaverili, per isolarli dal freddo, sorgono su uno spesso strato di sterco spesso ammucchiato anche per una certa altezza intorno alle stalle, che può costituire un importante indicatore nel record archeologico. Il sito 3 (Huliin Shilin – Havarjaa) (Fig. 6) posto a 1650 metri slm nell’Arkhangaï centro settentrionale, costituisce un esempio rappresentativo degli accampamenti delle stagioni fredde. L’accampamento, come di norma orientato a sud, era stato utilizzato durante la primavera del 2002 da una famiglia11,

Fig. 4. Negli accampamenti estivi e in quelli autunnali non sono presenti stalle per gli animali ma solo dei recinti. Anche per la piccola dispensa, che negli accampamenti dei mesi freddi è una solida struttura lignea, qui viene utilizzata una ‘ger’ di dimensioni ridotte.

dei nomadi (Cribb 1991, p. 65), si può dire, però, che le tracce che gli elementi strutturali di questi insediamenti lasciano sul terreno una volta smontati, non sono molto consistenti. La situazione cambia sostanzialmente negli accampamenti dei mesi freddi, ovvero invernali e primaverili, dove vengono utilizzate solide stalle lignee ben riparate dal vento e dalla neve, costruite dai nomadi stessi piantando possenti pali nel terreno e dotate di poderose coperture. Con la medesima tecnica vengono anche erette delle piccole costruzioni che fungono da dispensa. Queste strutture possono costituire certamente un significativo indicatore a livello archeologico poiché lasciano ingenti tracce sul terreno dopo l’abbandono, tra cui, numerose, le buche di palo, estremamente preziose se si pensa alle complesse problematiche dell’archeologia del nomadismo delle steppe. In queste postazioni, inoltre, le ‘ger’, coperte di ulteriori strati di feltro, spesso presentano un robusto impiantito ligneo, per un maggiore isolamento dal terreno, che lascia alcune interfacce negative di non impossibile lettura.

Fig. 6. Rilievo del sito 3 (Huliin Shilin – Havarjaa), posto a 1650 metri slm nell’Arkhangaï centro settentrionale. Questo accampamento era stato utilizzato nella primavera del 2002 da una famiglia composta dai due anziani genitori e dal figlio con la moglie e i tre bambini che possedevano in tutto circa 450 animali di cui: 60 yak, 60 cavalli e 330 ovini.

Fig. 5. Negli accampamenti invernali e primaverili sono presenti delle solide costruzioni lignee per proteggere gli animali dal freddo (sito 3 Huliin Shilin – Havarjaa).

162

Gli accampamenti invernali e primaverili dei nomadi in Mongolia centro - occidentale

Fig. 7. Pianta e prospetto Sud della stalla per gli ovini con dispensa e tracce della ‘ger’ dei ‘nonni’.

composta dai due anziani genitori e dal figlio con la moglie e i tre bambini che possedevano in tutto circa 450 animali di cui: 60 yak, 60 cavalli e 330 ovini. L’accampamento è stato realizzato dai due uomini della famiglia con legni già utilizzati in un altro accampamento, e lo spostamento degli animali dall’accampamento invernale era stato effettuato

dal nonno e dal nipotino maggiore di età in meno di una settimana12. I nomadi intervistati hanno riferito che i prossimi inverno e primavera non avrebbero rioccupato medesimi accampamenti, tra cui, quindi, il sito 3, ma che non li avrebbero smontati poiché la zona scelta per ‘svernare’ presentava 163

F. Lugli

Fig. 8. Pianta e prospetto Sud della stalla e recinto per gli yak e per i cavalli.

delle strutture già erette e abbandonate da poter utilizzare. E’ interessante ricordare, infatti, che le strutture per gli animali degli accampamenti invernali e primaverili non vengono sempre smontati per essere allestiti altrove. Spesso i nomadi utilizzano strutture già erette e abbandonate da altri gruppi. I legni per le stalle vengono smontati e riutilizzati soprattutto nel caso in cui i nomadi si spostino in una zona che ne è priva. Il luogo in cui spostarsi nei mesi

invernali è frutto di una scelta fatta in base a delle ricognizioni preventive effettuate dagli uomini della famiglia che, di conseguenza, decidono se occupare strutture già abbandonate oppure smontare e rimontare le ultime utilizzate. Nel sito 3 erano state montate due ‘ger’ di cui erano ben visibili le impronte. Entrambe le due ‘ger’, infatti, presentavano evidenti le tracce delle interfacce negative del pavimento ligneo che era stato utilizzato, nonché del ‘battuto’ 164

Gli accampamenti invernali e primaverili dei nomadi in Mongolia centro - occidentale per il quale il loro ricovero è in realtà un ampio recinto scoperto con annesso un piccolo ambiente coperto (circa m 4,60 x 7,50). Le recinzioni delle due strutture (per ovicaprini e per yak e cavalli) presentano una differenza: le prime hanno i tronchi orizzontali che poggiano per terra per impedire l’uscita degli animali mentre nel secondo caso, queste possono non arrivare fino a terra. Le pareti sono realizzate mediante la sovrapposizione di più pali orizzontali incastrati, in due o più punti, tra due pali ben infissi nel terreno. I pali orizzontali vengono bloccati tramite l’inserimento all’estremità di un cuneo ligneo piantato col martello che ne aumenta il diametro (Fig. 9). Nei punti di intersezione tra le pareti, i tronchi sono ‘maschiettati’ e bloccati l’un l’altro mediante degli incavi scavati alle estremità. I pali portanti, così come anche i pali verticali che contengono le pareti perimetrali, sono solidamente conficcati nel terreno e spesso, sono affiancati da due pali di dimensioni più ridotte che rendono più solido l’appoggio (Fig. 10). Il tetto è a un solo spiovente ed è costituito, anch’esso da pali affiancati che poggiano su pali verticali e da rompi tratta.

Fig. 9. Particolare della tecnica di costruzione mediante inserimento di pali orizzontali all’interno di due verticali, allargati all’estremità mediante l’inserimento di un cuneo.

Il sito 3 presentava a livello macroscopico alcune suppellettili residue, soprattutto elementi di carro non più funzionali. sostanzialmente di legno (poveri di guarnizioni e elementi strutturali metallici) e, soprattutto diversi processi formativi individuabili all’interno e all’esterno delle diverse strutture. A livello archeologico, quindi, un accampamento invernale o primaverile lascia sul terreno tracce consistenti della propria esistenza. Fig. 10. Particolare della tecnica di costruzione con palo portante supportato da una coppia di pali di diametro minore.

Conclusioni

antistante la soglia. E’ interessante osservare che i nomadi di queste regioni non utilizzano quasi mai chiodi per le proprie costruzioni lignee salvo, talvolta, come in questo caso, per il pavimento in legno della ‘ger’. La tenda utilizzata dai ‘nonni’ era stranamente ubicata a ridosso delle stalle perché in questo modo era più protetta dai venti che soffiano spesso da nord. Erano presenti, inoltre, una dispensa (circa m. 4,70 x 4), alle spalle della tenda dei due anziani, e due ricoveri per gli animali. Uno per gli ovini ed un’altra più distante, a circa 70 metri, per i bovini e i cavalli. La stalla per gli ovini, come si può vedere dall’esempio del sito 3 (Fig. 7), presenta un corpo centrale coperto (circa m 12,50 x 7,40) e dei recinti annessi privi, invece, di copertura. Gli ovini devono essere particolarmente protetti dal freddo e dalla neve, motivo per cui le loro stalle sono sempre ottimamente isolate. Per quanto riguarda i cavalli e gli yak (Fig. 8) invece, non debbono essere tenuti al riparo allo stesso modo, motivo

L’uso del legname13 per la costruzione delle ‘ger’ e dei ricoveri per gli animali, nonché come combustibile per la preparazione dei cibi, il riscaldamento e le attività artigianali quali la metallurgia, deve avere contribuito nel tempo in modo massiccio sulla perdita di manto forestale in gran parte del paese. Oggi la Mongolia appartiene, secondo la FAO ai paesi più poveri di risorse forestali. La deforestazione, infatti, ha costituito in passato e costituisce oggi un problema grave le cui principali cause sono da collegare principalmente agli incendi estivi e agli insetti, e poi all’uso del legname come combustibile e per le costruzioni14. Nel 2000 la superficie coperta da foreste era valutata circa il 8,1 % dell’intero paese con una perdita di manto forestale negli ultimi venti anni di circa 60.000 ettari all’anno. Il legno, quindi, è in Mongolia, anche nelle aree di montagna un bene estremamente prezioso e spesso raro da non sprecare. Se nelle aree rurali come combustibile da cucina viene facilmente sostituito dallo sterco bovino, per le ‘ger’ e per i ricoveri degli animali rimane l’unica materia uti165

F. Lugli lizzata. Negli accampamenti invernali e primaverili, in cui gli animali devono essere efficacemente protetti dal freddo l’uso del legname è consistente. Vengono utilizzati, infatti, tronchi sbozzati ad ascia, di notevoli dimensioni assemblati mediante la tecnica dell’incastro, aiutati talvolta solo da cordame di pelo animale e assai raramente da chiodi. Per isolare ulteriormente le strutture, nei campi invernali tra legno e legno viene posta una sorta di ‘stuccatura’ composta da sterco e paglia. I nomadi dell’Arkhangaï e dell’Ovorkhangaï si spostano di media quattro volte l’anno in base al susseguirsi delle stagioni. Le strutture abitative sono sempre costituite dalle tradizionali tende di feltro ‘ger’ mentre per gli animali vengono erette consistenti strutture negli accampamenti dei mesi freddi e solo delle recinzioni per quelli dei mesi caldi. Se le tende sul terreno lasciano tracce non consistenti e di non semplice lettura, i ricoveri per gli animali una volta smontati possono costituire un importante indicatore del record archeologico, utile ad un’archeologia del nomadismo delle steppe. Il consistente utilizzo del legno, inoltre, sia per le ‘ger’ che per i ricoveri per gli animali, rappresenta un interessante spunto di riflessione per una migliore comprensione di questo genere di vita e per il rapporto tra economie stabili e nomadi, nonché per una corretta ricostruzione della storia della deforestazione in Mongolia.

grafie o ritratti dei componenti della famiglia. La parte sinistra è riservata alle donne, mentre quella di destra agli uomini. Gli uomini, infatti, all’interno si dovrebbero spostare verso ovest sotto la protezione del cielo, mentre le donne verso est, ovvero sotto la protezione del sole. A destra, quindi, vengono riposti selle, attrezzature per i cavalli, per la caccia e cordami, mentre a sinistra gli utensili per la casa e per ottenere il formaggio e l’airag. Lungo le pareti vengono posti i letti e gli armadi. 8 I risultati inerenti a questa problematica sono stati presentati al 3° Convegno Nazionale di Etnoarcheologia (Mondaino 2004) e sono in corso di stampa. 9 La possibilità di documentare gli spostamenti sul territorio costituisce un altro importante aspetto della ricerca etnoarcheologica. 10 La disposizione delle ‘ger’ e dei diversi elementi che si trovano negli accampamenti sono un importante oggetto di studio ed analisi. 11 La famiglia di Sterennadmid Nastag. 12 Stando ai racconti dei protagonisti. 13 Mentre nelle regioni meridionali della Mongolia è diffuso il saxaul, nelle regioni settentrionali, invece, vi sono essenzialmente il larice siberiano, i pini silvestri e i pini siberiani. 14 Attualmente tale fenomeno è accelerato, inoltre, da un’ intensa esportazione verso la Cina (resa possibile a seguito della liberalizzazione del commercio estero nel 1996), che ha costretto lo stato, di fatto possessore dei beni forestali, a varare in merito leggi severissime difficilmente rispettate. L’esportazione illegale di legname, infatti, è un serio problema del paese le cui vaste e non popolose frontiere sono quasi impossibili da controllare.

Note

Bruno, I., 2003. Yourtes et tipi. Paris: Hoёbeke. Cribb, R., 1991. Nomads in Archaeology. Cambridge: Cambridge University Press. di Rubruc, G., 2002. Viaggio nell’impero dei Mongoli. (1820). Genova: Marietti. Grousset, R., 1939. L’empire des steppes. Attila, Gengis-khan, Tamerlan. Paris: Payot. http://encarta.msn.com/encyclopedia 761565003_3/Mongolia (country). html http://www.fao.org//docrep/003/x6900e/6900e0k.htm http://www.mongolie.mn/index3.php?page=./bcentre/ humanitaire/humanitaire.html Koryakova, L., 2000. Some Notes about the Material Culture of Eurasian Culture. In J. Davis-Kimball et alii, eds. Kurgans, Ritual Sites, and Settlements in Eurasian Bronze and Iron Age. BAR International Series, 890. Oxford: BAR Publishing, 13-17. Tsultem, N., 1988. Mongolian architecture. Ulan-Bator: State Publishing House.

Bibliografia

Secondo le stime governative per l’estate 2005 (http://www. fao.org//docrep/003/x6900e/6900e0k.htm) e (http://encarta.msn. com/encyclopedia_761565003_3/Mongolia_(country).html. ) 2 Nel 2002 si è svolta la prima missione etnoarcheologica promossa dalla’Associazione Italiana di Etnoarcheologia e coofinanziata dal MAE, in collaborazione con l’Accademia delle Scienze e l’Università di Ulaan Baatar, diretta da Francesca Lugli. Alla missione hanno partecipato, inoltre, l’arch. Graziano Capitini (A.U.A. srl) per il rilievo degli accampamenti e la prof.ssa Tserenkhand, docente di etnologia dell’Università di Ulaan Baatar, per la documentazione etnologica. 3 Centro urbano e commerciale, fatta costruire dal figlio e successore di Gengis Khan, Ögedeï. 4 http://www.mongolie.mn/index3.php?page=./bcentre/humanitaire/humanitaire.html 5 I mongoli delle steppe utilizzano per gli spostamenti veloci anche una tenda a due spioventi simile alla ‘canadese’ a due posti. 6 Il tipi, infatti, è una struttura agile e leggera che pesa relativamente e che può essere montata e smontata facilmente e in breve tempo. 7 L’interno della ‘ger’ è estremamente confortevole e lo spazio vi è utilizzato secondo antichi schemi che regolano la disposizione delle suppellettili. L’apertura è sempre rivolta a meridione. Al centro, in corrispondenza dell’apertura nel tetto si trovano: la stufa, talvolta sopraelevata, che funge sia per il riscaldamento che per la preparazione dei cibi, un tavolinetto e degli sgabelli o delle sedie. Opposto alla porta, si trova il luogo d’onore per ospiti o persone anziane e una sorta di altare con immagini sacre e foto1

166

Two long micro-charcoal records from central Italy D. Magri

Dipartimento di Biologia Vegetale, Università “La Sapienza”, P. le Aldo Moro, 5, 00185 Roma, Italy E-mail: [email protected] Abstract Two upper Pleistocene and Holocene micro-charcoal records from crater lakes in central Italy (Valle di Castiglione near Rome and Lagaccione near Lago di Bolsena), are compared with the respective pollen diagrams, with the aim of evaluating the relationship between vegetation and fire history in a Mediterranean country since the last interglacial. The two pollen diagrams show remarkable changes in the vegetation composition, structure and density. Very diverse landscapes followed each other, with alternate dominance of mixed deciduous oaks forest, beech forest, Mediterranean evergreen vegetation, conifer woodland and steppe-grassland formations. The micro-charcoal records do not highlight univocal relationship with the vegetation structure and density. Although micro-charcoal is generally more abundant during the forested periods, there are several exceptions, with only moderate increases of fire frequencies in correspondence with remarkable forest expansions. The time-periods with well-developed evergreen Mediterranean vegetation generally show high fire frequencies, while no clear correlation is found between pine-dominated vegetation and micro-charcoal concentrations. The fire frequencies of the postglacial do not appear significantly different from those recorded during the Eemian interglacial, when the human impact on the vegetation was negligible or absent. On the whole, the micro-charcoal records show similar trends in the two sites, indicating that they may be considered generally representative of the regional fire history.

Introduction Microscopic charcoal records from lake sediments are widely used to reconstruct local and regional fire histories, extending the fire chronology back several millennia. Micro-charcoal analysis is often associated with pollen analysis, as a simple and useful complement that helps understanding long-term impacts of fire on vegetation, particularly with respect to prehistoric human activity (Carrión, van Geel 1999; Galop et al. 2002, Sadori et al. 2004; Filipova-Marinova, Angelova, 2008). Other studies focus on the relationship between fire incidence and climate conditions, by comparing the micro-charcoal record with different vegetation types (e.g. Scott 2002; Carrión et al. 2003, Hallett, Hills 2006), or with climate changes induced by solar cycle oscillations (Willis et al. 2008). In most cases, the studied records deal with postglacial times, when human activity has become increasingly important, so that distinguishing anthropogenic and naturally induced fires may be a very difficult task. Only few records have been published that go back in time into the last glacial (e.g. Carrión, van Geel, 1999) or into previous interglacials (Kershaw 1986; Wang et al. 1999; Teed 2000).

tive of regional fire history? The two records of Valle di Castiglione and Lagaccione appear especially suitable to discuss these questions because they include very important vegetation changes, with some periods characterized by dense forests and other periods with open vegetation types. In addition, the two records show a wide range of intermediate situations. Pollen concentrations, available at both sites, show changes of up to four orders of magnitude, suggesting considerable changes of plant biomass. The composition of vegetation varies largely from time to time within the same records, showing alternate dominance of deciduous and evergreen forests, conifers and steppe-grassland formations. Besides, the two records include various forest periods preceding the postglacial, when the human impact on the landscape was negligible. Finally, the two sites are located at only 100 km from each other, a distance that, in spite of local differences of vegetation composition, allows the reconstruction of the regional features of the past vegetation and so may help in deciphering the local and regional inputs of micro-charcoal.

In this study, two long micro-charcoal records from central Italy, compared with the respective pollen records, offer the opportunity to reconstruct the fire history in a Mediterranean country since the last interglacial. The long micro-charcoal records from Valle di Castiglione and Lagaccione in central Italy, spanning 130,000 yr and 100,000 yr respectively, are presented with the aim of discussing the following questions: is the concentration of microcharcoal in sediments related to vegetation structure? to plant biomass? to vegetation composition? are postglacial micro-charcoal records significantly different from previous interglacials? is the micro-charcoal record representa-

Study sites Valle di Castiglione is an artificially dried out crater lake, located approximately 20 km east of Rome (lat. 41°43’30”N, long. 12°45’35”E), in a large plain extending from the Tyrrheanian Sea to the Alban Hills volcanic complex and the foothills of the Apennines. The mean annual temperature is 15°C, and the mean annual rainfall is about 800 mm. The lake bed, approx. 1 km in diameter, is situated at an elevation of 44 m a.s.l. The sediments of Valle di Castiglione have been drilled to a depth of 88 m. Multidisciplinary studies have been carried out on the lacustrine sediments,

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

D. Magri

Fig. 1: Synthetic pollen and micro-charcoal diagram from Valle di Castiglione.

Fig. 2: Synthetic pollen and micro-charcoal diagram from Lagaccione. 152 168

Two long micro-charcoal records from central Italy including lithostratigraphical and tephrochronological investigations (Narcisi et al. 1992), pollen analysis (Follieri et al. 1988; Magri 1989) and micro-charcoal analysis (Magri, Ciuffarella 1991). The chronology of the core is based on a series of 21 radiocarbon dates, counts of annual laminations (Magri, Narcisi 1994) and correlations with the precessional term (Magri 1989; Magri, Tzedakis 2000) leading to an estimated linear accumulation rate of about 0.3 mm yr-1. Although the last three millennia are not recorded, the core from Valle di Castiglione, spanning a time-period of over 250,000 yr, has provided one of the longest continuous palaeoenvironmental records in Europe (Tzedakis et al. 1997, 2001). Micro-charcoal particles were extracted and quantified in 105 samples. The samples were processed with a standard pollen preparation procedure, including treatment with HCl, HF and NaOH. They were neither sieved nor acetolysed. A known number of Lycopodium spores was added to each sample in order to determine the pollen and micro-charcoal concentrations. In two depth intervals, between 26.1 m and 22.3 m and between 47.5 m and 43.0 m, pollen is so scarce that it is not possible to present the results in a percentage diagram. In these intervals microcharcoal was however present. The concentration of microcharcoal in each sample was determined using the Clark (1982) point count method in 50 fields of view (Magri, Ciuffarella 1991).

Lagaccione is an artificially dried out maar-lake, located approximately 100 km northwest of Valle di Castiglione (lat. 42°34’N, long. 11°51’E). It occupied an explosion crater formed during the final period of activity of the Vulsini volcanic district. The present mean annual temperature is around 13.5°C, and the mean annual rainfall is 1000 mm. The lake bed, approx. 800 m in diameter, is situated at an elevation of 355 m a.s.l. The sediments of Lagaccione were drilled down to the volcanic basement, that was reached at 46 m depth. The palynological study of the Lagaccione core has provided a palaeoenvironmental record for the last 100,000 years (Magri 1999). The chronology of the sequence is based on a series of 16 radiocarbon dates, on a tephra layer and on palynostratigraphical correlations. The pollen samples were treated with HCl, HF and NaOH and a known number of Lycopodium spores was added to each sample. The concentration of micro-charcoal in 130 pollen samples was determined using the Clark (1982) point count method in 100 fields of view. Three size classes of micro-charcoal were originally counted (5-25 μm, 25-50 μm and >50 μm), but as they do not show significantly different trends, they are represented here in a single total curve. The pollen record from Lagaccione (Fig. 2), starting at the base of the St Germain I forest phase and ending about 3000 years ago, shows that remarkable changes occurred in the floristic composition, in the vegetation structure and in the plant biomass, including also long-lasting periods of open woodlands during the last pleniglacial (Magri 1999). The records of Lagaccione and Valle di Castiglione can be easily correlated on the basis of their radiocarbon dates and vegetational characters. Compared to the Valle di Castiglione record, at Lagaccione the vegetation of the last 100,000 years was generally more mesophilous, as can be expected considering that even the modern climate is wetter and cooler. Lagaccione shows higher percentages of AP than Valle di Castiglione during the forest phases, lower values of Mediterranean taxa, and higher proportions of montane elements, reaching values of over 80% during the St Germain I forest phase (Figs 1 and 2). Despite these differences the two records are very similar, showing vegetation changes valid at least at a regional scale.

A summary pollen diagram from Valle di Castiglione is represented in Fig.1. It shows that very different landscapes have followed each other in the Roman plain during the last glacial-interglacial cycle: the concentration of arboreal pollen (AP) indicates that dramatic changes of forest density have occurred, from situations with very few trees (e.g. some phases of the last glacial) to very lush vegetation during the interglacials. The AP percentages, by and large indicating the structure of vegetation, show four main forest phases, corresponding – from bottom to top - to Eemian, St Germain I, St Germain II and Holocene. Some other periods have very open vegetation, for example the last glacial maximum. A number of intermediate AP percentages indicating sparse trees in the landscape, are recorded during the pleniglacial interstadials. The vegetation composition of the non-forested periods is rather homogeneous, being dominated by Artemisia, Poaceae and Chenopodiaceae. On the contrary, the forest composition is appreciably different from period to period: the Mediterranean vegetation (including evergreen oaks, Olea, Cistus, and Phillyrea) appears to have been especially important during the Eemian and the second part of the Holocene; the mixed oak forest (deciduous oaks, Ulmus, Zelkova, Tilia, Acer, Carpinus, Ostrya, and Corylus) is well represented in all the forest phases, particularly during the Holocene, while the so-called montane forest (including taxa that mostly live at present over 800-1000 m a.s.l. in Italy: Fagus, Abies and Picea) was the dominant vegetation type during St Germain I.

Results In Fig. 3 the AP percentages and the micro-charcoal concentrations of Valle di Castiglione and Lagaccione are compared. The two micro-charcoal records show generally similar trends, with high values during the Holocene, very low concentrations in the last glacial maximum (around 14 C age 20,000 BP), fluctuating but generally significant concentrations during the middle pleniglacial interstadials, moderate values during the St Germain II, and somewhat higher values during St Germain I. The Eemian interglacial, recorded only at Valle di Castiglione, shows remark169

D. Magri

Fig. 3: Correlation of the AP percentages (Pinus and Juniperus excluded) and micro-charcoal concentrations from Valle di Castiglione and Lagaccione.

Fig. 4: Correlation of the AP and micro-charcoal concentrations from Valle di Castiglione and Lagaccione. 154 170

Two long micro-charcoal records from central Italy

Fig. 5: Correlation of the Pinus and Juniperus percentages and micro-charcoal records from Valle di Castiglione and Lagaccione.

Fig. 6: Correlation of Mediterranean vegetation and micro-charcoal records from Valle di Castiglione and Lagaccione. 156 171

D. Magri able peaks, comparable to those of the Holocene. When the micro-charcoal record is compared with the AP percentages, no biunique relationship can be found: although micro-charcoal is generally more abundant during the forested periods, at both sites there are intervals of the Holocene and St Germain forest expansions with very low charcoal concentrations. On the contrary, during the moderate increase of arboreal pollen vegetation of the pleniglacial interstadials significant increases of charcoal are often observed. However, when the AP percentages are low (e.g. during the last glacial maximum and in the stadial between St Germain I and St Germain II) charcoal values are generally also low. The AP concentrations of Valle di Castiglione and Lagaccione are represented in Fig. 4. They compare to each other quite well, showing very high values during the forested periods and very reduced values during the glacial phases with mainly grasslands and steppe vegetation. In both cases the pleniglacial interstadials fluctuate around moderate values. The main discrepancy between the two records is found in the St Germain II, when low concentration values at Valle di Castiglione are probably due to a disturbed sediment accumulation. The comparison of AP and micro-charcoal concentrations (Fig. 4) indicates that in some cases very high pollen concentrations correspond to high micro-charcoal values (e.g. some peaks of the Holocene and of the St Germain I). In other cases, when pollen concentration is very high, charcoal is low (e.g. St Germain II and first half of the Holocene at Lagaccione, upper part of the Holocene at Valle di Castiglione). When pollen concentration is low, the microcharcoal presents a range of intermediate values. Pollen concentration is generally low during the last glacial, when the prevailing arboreal AP type is Pinus, often accompanied by abundant Juniperus (Fig. 5). High frequencies of pine and juniper may be biased by percentage calculations in a time-period when other AP are very scarce, but it cannot be excluded that at least in some periods (e.g. at the end of St Germain II) pine was really abundant in the landscape. Although pine and juniper are commonly considered to be very flammable, and therefore suited to live in frequently burning environments, no clear relationship is evident with the micro-charcoal concentration (Fig. 5). The highest micro-charcoal peaks (Holocene, Eemian) do not correspond to high values of pine, but a number of increases of pine and juniper in the pleniglacial interstadials match moderate increases of charcoal. The high percentages of pine of the last glacial maximum correspond to micro-charcoal minima. This could indicate either that there were few pines on the landscape, or that the environment was unsuitable for burning. The Mediterranean vegetation is particularly abundant in the plain site of Valle di Castiglione, especially during the Eemian, an interglacial forest phase that is missing at Lagaccione (Fig. 6). A generally good correspondence is found between micro-charcoal concentrations and Mediterranean taxa: the highest micro-charcoal peaks occur 172

when evergreen vegetation is dominant, e.g. during the Eemian and the second half of the Holocene. There are however many cases with high micro-charcoal values and sparce Mediterranean vegetation, for example during the lateglacial, in some oscillations of the pleniglacial and in the stadial between Eemian and St Germain I (at Valle di Castiglione). Mixed oak forest is the most common forest type both at Valle di Castiglione and at Lagaccione (Fig. 7). Deciduous trees are always present in both records, although with very reduced frequencies during the glacial phases. The highest percentages of mixed oak forest are found in the postglacial, with maximum values of over 60% at Valle di Castiglione and 90% at Lagaccione. The comparison of micro-charcoal concentrations with the mixed oak forest record provides the same results of the comparison with the total AP percentage curve: charcoal is generally very abundant during the forested periods, particularly in the postglacial, but there are a number of exceptions. For example, the remarkable expansion of mixed oak forest of the St Germain II corresponds to only moderate peaks of charcoal. Montane vegetation is especially abundant during the St Germain I forest phase at both sites, with values of up to 80% at Lagaccione and 45% at Valle di Castiglione (Fig. 8). Correlation with micro-charcoal is generally not good, as during the beech forest development, micro-charcoal values appear moderate to low, especially at Lagaccione. The highest peaks of charcoal at Valle di Castiglione (Holocene and Eemian) correspond to periods with sparse montane taxa.

Discussion The comparison of the micro-charcoal and pollen records from Valle di Castiglione and Lagaccione suggests some answers to the addressed questions: The comparison with AP percentages, showing a wide range of situations over the last glacial-interglacial cycle indicates that the incidence of fires is not always directly correlated with the degree of forestation and the vegetation structure. Although this is true in some cases, exceptions are so numerous that a relationship between forested landscapes and fire frequencies cannot be firmly established. When plant biomass, as indicated by AP concentrations, is very high, fires may be abundant; this relation is however not always true. Although large amounts of plant material are necessary to produce large fires, it appears that there is no clear correlation between biomass and incidence of fires. Concerning the relationship between fires and vegetation composition, a flexible answer may be suggested. Time periods with well developed Mediterranean vegetation generally have also high micro-charcoal values. On the contrary, diffused montane vegetation does not favour

Two long micro-charcoal records from central Italy

Fig. 7: Correlation of mixed oak forest and micro-charcoal records from Valle di Castiglione and Lagaccione.

Fig. 8: Correlation of montane vegetation and micro-charcoal records from Valle di Castiglione and Lagaccione. 158 173

D. Magri the occurrence of fires. Mixed oak forest shows a more ambiguous behaviour. This result suggests that a marked Mediterranean climate, with dry and hot summers, may have favoured the incidence of fires, whereas wet and cool summers, which are needed for the development of mountain forests, may have prevented vegetation from burning. In both cases, it appears that summer climate may be crucial for fire ignition. Unexpectedly, no clear relationship is visible between pine-dominated vegetation and fire frequency. Although some mediterranean pine species are well adapted to fire (e.g. Pinus halepensis and P. pinea) and other pine species are particularly well suited to dry climates (e.g. Pinus sylvestris), it appears that micro-charcoal values were not especially high when pine and juniper were the dominant trees. This result suggests that although percentages of Pinus were high during some glacial phases, the actual number of trees in the landscape could be rather low. Both at Valle di Castiglione and at Lagaccione the last three millennia are missing in the records, as the cores were drilled in dried out lakes transformed in arable land. For this reason it is not possible to examine the time-period when human activity has been most important. However, considering the record of Mesolithic, Neolithic and Bronze age times at Valle di Castiglione, the incidence of fires is comparable to the Eemian record, suggesting the possibility that the prehistorical human impact did not significantly increase the occurrence of fires with respect to the natural incidence. At Lagaccione, where the Eemian interglacial is not recorded, no other forest period is found with climate and vegetation features similar to the Holocene, so it is impossible to determine whether the marked increase of fires of the postglacial is due to human activity or to a natural trend related to the increase of Mediterranean evergreen vegetation. On the whole, the two micro-charcoal records match quite well with each other. They show very low values during the most severe phases of the last glacial, e.g. in the last glacial maximum, in the open vegetation phase following the St Germain II (corresponding to cold stage 4 of the marine isotope stratigraphy), and between the St Germain I and St Germain II forest phases (corresponding to stadial 5b of the marine stratigraphy). The highest peaks of microcharcoal concentrations are found in both records during the Holocene. In the rest of the records there are fluctuating concentrations of micro-charcoal, with intermediate values during the full glacial interstadials and the two St Germain forest periods. Interestingly enough, also the long micro-charcoal record from Navarres 3, near Valencia in Spain (Carrión and van Geel 1999), spanning the last 30,000 years, shows very low values during the last glacial maximum, and increasing concentrations during the postglacial, with very high values after 6000 BP, while the interstadial oscillations correspond to intermediate values of charcoal. When the two Italian records are compared in more detail, no clear relationship is visible in the timing and duration of the single micro-charcoal oscillations. In

particular, these discrepancies may be examined in the upper part of the records, where the chronology is supported by a number of radiocarbon dates. On the other hand, the composition of the vegetation is also somewhat different at the two sites, as often happens in the mosaic-like vegetation typical of the Mediterranean environments. On the whole, it seems that while the details of the records are different, the degree of similarity of the two charcoal records at the scale of 103-104 years is comparable to the similarity between the two pollen records. This observation induces to put forward the hypothesis that a broad correlation may exist between climate change and fire incidence at a regional scale. However, to validate this hypothesis, an objective method of comparison between records, making use of statistical tests as suggested by Willis et al. (this volume) should be applied, and new data from long records spanning at least the last glacial should be collected.

References Carrión, J.S., van Geel, B., 1999. Fine-resolution Upper Weichselian and Holocene palynological record from Navarrés (Valencia, Spain) and a discussion about factors of Mediterranean forest succession. Review of Palaeobotany and Palynology 106, 209-236. Carrión, J.S., Sánchez-Gómez, P., Mota, J.F., Yll, R., Chaín, C., 2003. Holocene vegetation dynamics, fire and grazing in the Sierra de Gádor, southern Spain. The Holocene 13, 839-849. Clark, R.L., 1982. Point count estimation of charcoal in pollen preparations and thin sections of sediments. Pollen et Spores 24, 523-535. Filipova-Marinova, M., Angelova, H., 2008. Pollen and microcharcoal evidence of vegetation dynamics and human impact along the southern bulgarian Black sea coast, (this volume). Follieri, M., Magri, D., Sadori, L., 1988. 250,000-year pollen record from Valle di Castiglione (Roma). Pollen et Spores 30, 329-356. Galop, D, Vannière, B, Fontugne, M., 2002 Human activities and fire history since 4500 BC on the northern slope of the Pyrenees: a record from Cuguron (central Pyrenees, France). In: Thiébault S., ed. Charcoal Analysis. Methodological approaches, palaeoecological Results and wood uses. BAR International Series 1063. Oxford: BAR Publishing, 43-51. Hallett, D.J., Hills, L.V., 2006. Holocene vegetation dynamics, fire history, lake level and climate change in the Kootenay Valley, southeastern British Columbia, Canada. Journal of Paleolimnology 35, 351-371. Kershaw, A.P., 1986. Climatic change and Aborigenal burning in the north-east Australia during the last two glacial/ interglacial cycles. Nature 322, 47-49. Magri, D., 1989. Interpreting long-term exponential growth of plant populations in a 250,000-year pollen record from Valle di Castiglione (Roma). New Phytologist 112, 123-128. Magri, D., 1999. Late-Quaternary vegetation history at Lagaccione near Lago di Bolsena (central Italy). Review of Palaeobotany and Palynology 106, 171-208. Magri, D., Ciuffarella, L., 1991. Incendi e vegetazione nella Campagna Romana durante il Quaternario superiore.

174

Two long micro-charcoal records from central Italy Giornale Botanico Italiano 125 (3), 288. Magri, D., Narcisi, B., 1994. Annually laminated sediments at Valle di Castiglione (Rome, Italy). PACT (Journal of the European Study Group on Physical, Chemical, Mathematical and Biological Techniques Applied to Archaeology), 41, 9198, Rixensart, Belgium. Magri, D., Tzedakis, P.C., 2000. Orbital signatures and longterm vegetation patterns in the Mediterranean. Quaternary International 73/74, 69-78. Narcisi, B., Anselmi, B., Catalano, F., Dai Pra, G., Magri, G., 1992. Lithostratigraphy of the 250,000 year record of lacustrine sediments from the Valle di Castiglione crater, Roma. Quaternary Science Reviews 11, 353-362. 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, 5-17. Scott, L., 2002. Microscopic charcoal in sediments: Quaternary fire history of the grassland and savanna regions in South Africa. Journal of Quaternary Science 17, 77-86. Teed, R., 2000. A. 130,000-Year-Long Pollen Record from Pittsburg Basin, Illinois. Quaternary Research 54, 264274. Tzedakis, P.C., Andrieu, V., Beaulieu, J.-L. de, Birks, H.J.B., Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N.J., Wijmstra, T.A., 2001. Establishing a terrestrial chronological framework as a basis for biostratigraphical comparisons. Quaternary Science Reviews 20, 1583-1592. Tzedakis, P.C., Andrieu, V., Beaulieu, J.-L. de, Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N.J., Wijmstra, T.A., 1997. Comparison of terrestrial and marine records of changing climate of the last 500,000 years. Earth and Planetary Science Letters 150, 171-176. Wang, X., van der Kaars, S., Kershaw, P., Bird, M., Jansen, F., 1999. A record of fire, vegetation and climate through the last three glacial cycles from Lombok Ridge core G6-4, eastern Indian Ocean, Indonesia. Palaeogeography, Palaeoclimatology, Palaeoecology 147, 241–256. Willis, K.J., Bennett, K.D., Haberle, S.G., 2008. Solar influence on Holocene fire history, (this volume).

175

The “fires” of Aeolian villages at the end of Middle Bronze Age: the case of Portella site in the Salina island (ME - Italy)

M. C. Martinelli1, G. Fiorentino2

Soprintendenza BCA Messina (Sicily, Italy) Servizio per i Beni Archeologici Laboratorio di Archaeobotanica e Paleoecologia, University of Lecce - Italy Email: [email protected]

1 2

Abstract Archaeological investigations in the Middle Bronze age village of Portella in the Salina island reveal the presence of fire layers in the huts. Anthracological analysis and ethnographic comparison with the typical house of the shepherds in the Peloritani mountains, shed new light on carpentry elements and roof construction in relation to vegetal composition.

The prehistoric village of Portella has been dated at Middle Bronze age (XV-XIII B.C.) when in the Aeolian island (Fig. 1) the Milazzese culture was widespread. In this period in the Aeolian island the people built their villages on high rocks to defend the territory around. The settlements on Milazzese cape in Panarea, Montagnola in Filicudi, Acropoli in Lipari and Portella in Salina are well known. The prehistoric village of Portella was discovered in 1954 when a road connecting Santa Marina to Malfa was built. Archaeological remains were found digging through a ridge. Archaeological excavations were performed in 1955 by L. Bernabò Brea and M. Cavalier (1968, pp. 144-180) and later in 1999-2000 by Martinelli (2002; 2003; 2004). The village is spread along the entire Portella ridge, a 300 m slope of pyroclastic sediments, mostly incoherent. The ridge is cut by an erosional valley at the top, and is bordered by deep erosional valleys on both sides. Erosions and collapses are still visible at present. In 1955 ten buildings (A-I and F1) were discovered and seven more buildings (L-R) during the excavations of 1999-2000. Buildings are oval, sometimes nearly circular (their maximum diameter is 4 m). Each building is built inside a platform excavated in incoherent volcanic rocks. Several variations are observed in the structure: huts Q and R are surrounded by a stone wall (Fig. 2), M and P have no stone wall, the rear vertical side of the N and O platforms is

Fig. 2. Huts Q and R.

Fig. 1. Eolian Islands ( Sicily-Italy).

covered with stones, the rear vertical side of the L platform is covered with clay. The walls are made of big stones collected from the beach or in the village area. The threshold is a flat stone plate, the entrance is lower than the outside ground level. Each building includes a single occupation layer very rich in archaeological remains. Floors are made of clay and stone plates. The interdisciplinary research on the site has allowed the reconstruction of a model for Bronze age buildings1, and the possibility to understand their functional use. The archaeological documentation for the Middle Bronze

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

M. C. Martinelli, G. Fiorentino

Fig. 3. Portella site in Salina Island: Map of the village. The grey circles indicate the buildings maintaining clear traces of fire.

age in the Aeolian islands is very problematic: villages were abandoned and destroyed during this period. The Ausonians, a new population from the Italian peninsula, arrived in Lipari and lodged only in the Acropoli village. The other islands were uninhabited until the Greeks arrived in 580 BC . The village of Portella is a good witness of this period, because:

2. in some buildings, the floor is covered by a black charcoal layer caused by one or more fires that destroyed the vegetal roof. In the O hut big fragments of charcoal were preserved. In Fig. 3 the map of the village is shown. In the buildings marked with a grey circle traces of fire were found. The 14C dating (Fig. 9) shows two group of ages separated by about one century. The fire has damaged some of the buildings of the village. It is possible that the occurrence

1. in the huts (structures) all the objects used at the time when the huts were deserted are preserved 178

The “fires” of Aeolian villages at the end of Middle Bronze Age

Absolute anthracological diagram Villaggio di Portella (Salina - ME) n° 368 0

10

20

30

40

50

60

70

80

90

Prunoideae

Pomoideae

Olea

Pistacia

Phillyrea/Rhamnus

taxa

Erica sp.

cfr. Arbutus

Myrtaceae

Leguminosae ind.

cfr. Genista

Cistus sp.

Indeterminable

Fig. 4. General anthracological diagram.

dynamics of the deposit: first the collapse of the roof and then the re-building of it occurred. In the anthracological record there are numerous fragments of the roof. These fragments were deposited during the fire and preserved by the desertion of the huts. The fragments are represented by branches of little or medium diameter (max 4 cm) of Leguminosae, that in detail belong to Genista sp. (Fig. 6) and some Pomoideae. We present two cases of collapse of the roof: 1) Hut O (US29; US25): collapse of the roof in the centre of the hut (Fig. 5) 2) Hut L (US28): collapse of the roof on spatial distribution. Fig. 5. Hut O: spatial distribution of charcoals.

of natural fires was very common in this Mediterranean habitat and that the leaves roof of the buildings fed the fire itself. The anthracological analysis (Fig. 4) established that the palaeovegetation was composed by Mediterranean taxa, thermophilous and heliophilous. The heather and the Cistus testified an anthropic landscape where fires and grazing frequently occured. In this landscape the presence of broom and wild olive was important. The traces of fire are unmistakable in same structures. During the excavation a thick stratum was found, black in colour and rich of charcoals. This stratum covered the floor of the hut and the objects (pottery and lithic tools) which were inside. The macroscopic charcoals were studied in their original position in the stratum. So it was possible to reconstruct the

Fig. 6. Transversal section of Genista sp.

179

M. C. Martinelli, G. Fiorentino in Sicily (fig. 7a-b) (Li Vigni 1997). In this part of Sicily the typical house of the shepherds was “U pagghiaru”: a circular building with stone walls and conic roof, build with Genista branches (fig. 8a-b). In these buildings the branches are embedded in stone wall, to support the conic roof. The branches were of different sizes (lengths and thicknesses). The structure radiating under the roof was

Fig. 7a. Peloritani mountains (Messina, Sicily): this hut was inhabited until modern time.

Fig. 7b. Peloritani mountains (Messina, Sicily). Structure of the roof: there are no central poles as in the Portella huts.

Hut L (diameter max 3,20 m.; height 1,80 m.) The structure has a circular plant but part of it was destroyed during the construction of a new provincial road (1954). The coal-black stratum (US28) was distributed in two sectors. this is evidence of a great fire. It is 40 cm thick. The plaster coating on the internal surface of the hut was very cooked. Hut O (diameter max 3.00 m; height 1.80 m) The structure has a circular plan, but it is not complete because part of it collapsed. The coal-black stratum, 50 cm thick, covers the floor. The charcoals were grouped in semicircles. Large fragments of charcoal are well preserved. We suppose that the roof of these structures was conic and was build with branches of Genista sp. The conic roof rests on tools-wall or on a rock. In the huts of the Portella site there were no poles to support the roof. The reconstruction of the roof was made easier by the ethnographic comparison seen on the Peloritani Mountains

Fig. 8a-b. The reconstructed hut and its section.

180

The “fires” of Aeolian villages at the end of Middle Bronze Age

Fig. 9. 14C dating.

irregular. The roof covering was built with small bands of fronds of young broom on many layers. Some branches were crossed to support the structure. The bigger branches of broom were tied with branches of willow (today the branches of willow are used to weave baskets) and the little branches were tied with leaves of broom. A layer of fronds of broom was changed every year because the weather damaged it. The interdisciplinary study together with the ethnographic research has allowed to fully understand the building techniques and the topography of the village during the Bronze Age. This archaeobotanical analysis was very helpful to the archaeological and environmental research, as it will help to a better understanding of the other Bronze Age villages of the Aeolian islands.

Martinelli M.C., 2004. Le isole Eolie nella preistoria: Breve sintesi sulle culture e la recente ricerca scientifica, Seminario “ Il fare e il suo senso. Tecniche saperi e culture dai cacciatori paleolitici agli agricoltori neolitici”, Dipartimento Beni Culturali, Università di Lecce, Congedo ed. Martinelli M.C.-Baroni I.-Lopes L.-Minniti C.-Recchia G., 2002. La Portella, analisi funzionale delle strutture L e P, Atti Riunione Scientifica “Analisi informatizzata e trattamento dati delle strutture di abitato di età preistorica e protostorica in Italia”, Ferrara 26-27 novembre 2001, pp. 477-488. Martinelli M.C.-Lopes L.-Calabria M.T.-Calderoni G.Fiorentino G.-Levi S.T., 2003. Nuove ricerche sul sito dell’età del Bronzo di Portella nell’isola di Salina, Atti XXXV Riunione Scientifica dell’IIPP “Le comunità della Preistoria italiana: studi e ricerche sul Neolitico e le età dei metalli” in memoria di Luigi Bernabò Brea, Lipari 2000, pp. 883-888.

Endnotes The reconstruction was made with the collaboration of the technical designer Leandro Lopes, during the archaeological excavation.

1

Bibliography Bernabò Brea L., Cavalier M. 1968. Meligunis Lipara III, Flaccovio Editore, Palermo. Li Vigni V. 1997. Sopravvivenze architettoniche. Il Pagliaro, 1997 in Tusa S. (a cura di), Prima Sicilia, Ediprint, Palermo, pp. 609-615

181

Combining charcoal and pollen analysis: Holocene vegetation dynamics, tree species composition and woodland use in the Bavarian Forest O. Nelle

Ökologie-Zentrum der Christian-Albrechts-Universität Kiel, Olshausenstr. 75, 24118 Kiel, Germany E-mail: [email protected]

Abstract Pollen and charcoal analysis were carried out in the Western part of the Bavarian Forest (Vorderer Bayerischer Wald, south-east Germany) and adjacent areas to reconstruct postglacial vegetation dynamics and anthropogenic impact on the forest. Data from two pollen profiles in submontane and montane altitudes were combined with anthracological data from 66 historic charcoal production sites (medieval and postmedieval/modern times) and samples from archaeological excavations (Bronze Age fortification at Bogenberg near Straubing; Latène-period „Viereckschanze“ at Plattling-Pankofen, and others).

Introduction Most research into the Holocene vegetation history of Europe is either palynological or anthracological (in this context: species identification through charcoal analysis). There are few studies which combine the potentials of both approaches to overcome the shortcomings of each. Charcoal production sites have proved to be a precious palaeo-archive for the investigation of woodland history (Müller 1939/40; Krause 1972; Hillebrecht 1982; Bonhote, Vernet 1988; Ludemann 1994; 1996; 2003; Ludemann, Nelle 2002; Nelle 2002b, 2003; Nelle, Kwasniowski 2001; Fabre 1996; Davasse 2000; Bonhote et al. 2002; Montanari et al. 2002). The combination of site-related, fine-scaled information of woodland composition from charcoal production sites with pollen data has been practiced very rarely so far (Smettan 1995; Hildebrandt et al. 2001). In the framework of an interdisciplinary research training group (Graduiertenkolleg) at the University of Regensburg, anthracological and palynological studies were carried out in the Western Bavarian Forest (Vorderer Bayerischer Wald), Germany. The research training group (involved disciplines: geography, archaeology, botany) aimed to work interdisciplinary at the same sites and/or in the same study area, in order to reconstruct landscape change under climatic and human impact, in the surroundings of Regensburg in southEastern Bavaria. Here summarised results of the combined anthracological and palynological research (see Nelle 2002a for the detailed publication) are presented. The study adressed the following questions: 1. How did vegetation develop and change since the last glaciation in the Western Bavarian Forest? For the Eastern Bavarian Forest (Hinterer or Innerer Bayerischer Wald) and the Bohemian Forest, the general Holocene vegetation dynamics are known thanks to the studies of Stalling (1987), Kral (1979), Traut-

mann (1952), Langer (1962), Svobodová, Soukupová (2000), Svobodová et al. (2001), Brande (1995). However, the Western Bavarian Forest (Vorderer Bayerischer Wald) lacked palaeoecological investigations so far. 2. When and how did human activities impact the vegetation? What were the ecological conditions the first settlers faced? Published studies so far suggest that the mountainous range of the Bavarian Forest was settled during medieval times. 3. How was the natural forest composition before human impact? To establish a landscape and forest management which favours a near-natural species composition, palaeoecological studies access the past to provide information on the forest composition before man altered it. In the region under study, it is of special importance to get a detailed picture of the migration, distribution and role of spruce (Picea abies) before the onset of forest management. 4. Do we gain more insight into woodland dynamics with the combination of charcoal and pollen analysis? Is it possible to reconstruct regional and local vegetation dynamics when combining the two data sets? Pollen analysis reveals vegetation change on a local and regional level over continuous time spans, whereas charcoal analysis provides site-related information on species occurence and forest composition, especially when sediments containing pollen are lacking. Methodological research on diameter measurements of kiln charcoal samples was carried out and confirms the methods‘ potential for reconstructing forest stand structure and improves interpretation of historic charcoal (for details see Nelle 2002a).

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

O. Nelle

Fig. 1. Study area and sample sites in the Western Bavarian Forest, adjoining Danube valley and “Passauer Abteiland” (kiln sites at Ringelai, no. 8).

not until medieval times, that people founded settlements in the mountainous regions and cut down the forest. It is assumed that the “cultural landscape” of today’s Bavarian Forest was formed since the High Middle Ages.

Study area The “Vordere Bayerische Wald” is the Western part of the central European mountain range of the Bavarian/ Bohemian Forest, the whole range shared by three nations (Austria, Czech Republic, Germany). The altitudinal range of the study area (Fig. 1) spans from 300 m a.s.l. in the Danube valley to 1100 m a.s.l. at the tops of the Vordere Bayerische Wald. Landscape today is a mixture of forested areas, pastureland, meadows and few arable fields. Forests are generally dominated by Picea abies, which was used in reforestation from the 19th century onwards. Climate is suboceanic to subcontinental. Zonal potential vegetation of the Western Bavarian Forest is assumed to be mainly beech forest (communities of the Luzulo-Fagenion) with varying proportions of Abies alba and Picea abies, and Acer pseudoplatanus in upper montane forest stands (Janssen and Seibert 1991; Rüther 2003). In submontane elevations, especially on south-facing slopes towards the Danube valley, Pinus sylvestris and Quercus robur/Q. petraea play a considerable role in the forest communities. The plain of the Danube was settled in prehistoric times. However, it was

Material and Methods Pollen profiles, pollen analysis Two mires – submontane Filzmoos (635 a.s.l.) and montane Kugelstattmoos (870 a.s.l.) – were cored with a Russian peat sampler (0.5 m, 7 cm in diameter). Pollen preparation followed standard methods, treating the samples with 10 % HCl, 10 % KOH, HF, acetolysis and mouting in glycerine (Faegri and Iversen 1989). Generally, an arboreal pollen sum of min. 600 grains was aimed at. Pollen type identification followed Moore et al. (1991), Beug (1961), Faegri, Iversen (1993), and a reference collection. Basis for the %-calculation of the arboreal pollen types was the arboreal pollen sum. Nonarboreal pollen types were calculated using the overall land pollen sum. 184

Combining charcoal and pollen analysis: the Bavarian Forest

Fig. 2. Relative simplified pollen diagram from “Filzmoos” near Wiesenfelden.

Charcoal analysis

site were sampled, with rare exceptions of smaller sample size. Diameter of fragments was measured using a stencil, and pieces were allocated thereafter to 5 arbitrary chosen diameter classes (-2 cm, 2-3 cm, 3-5 cm, 5-10 cm, >10 cm) (Ludemann 1996; Ludemann, Nelle 2002; Nelle 2002a, 2002b). A histogram (Fig. 3, 4) gives the proportions of the five diameter classes for a sample and illustrates the distribution of the charcoal pieces among the diameter classes (size class distribution). A mean diameter was calculated for each sample using the formula:

Charcoal samples derived from archaeological excavations (Bronze Age, Bogenberg: charcoal fragments n = 335; Iron Age, Plattling-Pankofen: n = 227), but mainly from charcoal production sites in the study area. Overall 66 kiln sites were sampled, with a total number of 7826 charcoal pieces. Species identification followed Schweingruber (1990a, 1990b), and a reference collection was used. A minimum of 100 charcoal fragments at each kiln 185

O. Nelle

Fig. 3: Charcoal production pit “K 66” in the „Forstmühler Forst“ S Altenthann, results of the archaeological excavation. a, f: Pit form in the field. b: Planum. c: W-profile. d: N-profile. f: Results of charcoal analysis. 1-2: C-horizon, 3-9: stratigraphically differentiated layers. black: charcoal-rich, dark grey: with some charcoals, light grey: charcoal-free. TB: animal hole. (from Nelle et al. 2003, altered)

5 to 10 % and did not play a considerable role in the natural woodland, opposed to the Eastern Bavarian Forest (Hinterer Bayerischer Wald with National Park Bavarian Forest), where Picea pollen makes up to 30 % or more in the diagrams (Stalling 1987). Human impact is weak until medieval times: some Plantago lanceolata and Rumex type pollen might indicate local disturbance by humans, but can also be explained by distant transport from the Danube valley, where human impact was intense since the Neolithic. Vegetation change due to human settlements took place in medieval times. The Fagus-Abies woodlands were partly cleared, opportunistic pioneer species such as Pinus, Betula, and Picea increased. Cereal pollen grains show the presence of arable fields in the vicinity.

mean diameter (mD) [cm] = [n(0-2cm) *1cm + n(3-5cm) *2.5cm + n(3-5cm) * 4cm + n(5-10cm) *7.5cm + n(>10cm) *15cm] / ntotal Mean diameter can theoretically be min. 1 cm and max. 15 cm. Methodological studies on charcoal samples from known wood sizes have shown that fire places have mDvalues from 1,5-6,2 cm, kiln sites where wood from coppiced stands were used have 7-9 cm, and kilns where wood from trees with stem diameters 15-20(-30) cm were charred showed mD-values 10-11 cm (Nelle 2002a). Beech dominated woodlands since the Subboreal: the Filzmoos pollen profile The Filzmoos mire started to grow in the Late Glacial, the basis was AMS-dated to 12266 ± 101 conv. a BP. Between 9300 cal a BP and 5600 cal a BP, a hiatus can be stated by a 3 cm mineral layer and marked changes in the peat stratigraphy and the pollen diagram. Fagus-Abies-forests established around 5000 a BP. Picea pollen was between

Charcoal production sites at Forstmühler Forst In a small valley in the Forstmühler Forst, 10 km north of the Danube plain, we found over a dozen charcoal production pits and close-by sites where upright circular kilns were operated. It is said that charcoal producing technol186

Combining charcoal and pollen analysis: the Bavarian Forest

Fig. 4. Location of mire “Kugelstattmoos“ and kiln sites (solid circles) at Hirschenstein (Vorderer Bayerischer Wald). Results of charcoal analysis of 10 kiln sites, histograms showing proportions of species and diameter size classes. Cursive: mean diameter values. (n = data basis). “pioneers”: Betula, Corylus. Other (K 51): Alnus.

ogy from pits to circular upright kilns on even plates took place in Late Middle Ages/early modern times. One pit was excavated (Fig. 3), and several layers rich in charcoal, separated by charcoal-free layers, could be distinguished (Nelle et al. 2003): the deepest and middle layers contained only Quercus charcoal, whereas the upper layer contained only Fagus and Abies. Other charcoal pits showed a spectrum rich in Fagus and Quercus, but without

conifers. The diameter histograms show that wood from large stems was used. Two pits were dated by AMS radiocarbon dating (Physikalisches Institut, Universität Erlangen): K 21, Erl-4075, 1408 ± 44 a BP, K 28 Erl-4076, 1399 ± 41 a BP. These are the earliest datings of anthropogenic structures so far known in this mountainous area. The sites where upright circular kilns were operated showed Abies-rich spectra, with an admixture of Fagus and Picea. 187

O. Nelle

Fig. 5. Relative simplified pollen diagram from Kugelstattmoos near Hirschenstein.

One site was dated dendrochronologically by Franz Herzig, Dendroarchaeological Laboratory, Thierhaupten: AD 1692-1734.

but there is a considerable proportion of Abies, and some Picea. Around the top, spectra are nearly 100 %-Fagus, but with a constant appearance of Acer (presumably Acer pseudoplatanus) in the samples. Wood mean diameter has medium values at the mire-neighboured sites, whereas hilltop sites show higher values. In the pollen diagram (Fig. 5), a marked fall of Abies pollen curve in 24 cm depth and Fagus pollen curve in 20 cm depth might be partly due to cutting down trees for the kilns. This coincides with a peak of charcoal particles in

The combination of charcoal and pollen data: Hirschenstein In upper altitudes we found a unique ensemble of a mire and very close-by charcoal kiln sites (Fig. 4). The charcoal spectra at the margin of the mire were dominated by Fagus, 188

Combining charcoal and pollen analysis: the Bavarian Forest

Fig. 6. Summarised results along a „transect“. Bronze Age hillfort Bogenberg with Bronze Age and mediaval/modern times spectra; pollen diagrams “Filzmoos” and “Kugelstattmoos” in a simplified, summarised form for three time periods: Bronze Age to medieval, medieval to modern times, modern times. Sites of a kiln site ensemble are summarised in the vertical columns, with given no. of sites per column and no. of charcoal pieces. bz: Bronze Age 3000-800 BC; ma: Middle Ages 8th-15th century; mo: Modern Times 16th-20th century.

the upper layer, which is possibly connected to charcoal production. The charcoal analysis reveals a species spectra pattern which cannot be derived from the pollen analysis: around the mire “Kugelstattmoos”, the stands had higher proportions of Abies, whereas around the top of “Hirschenstein”, more or less pure Fagus stands were growing at time of charcoal production, which took place presumably during early modern times. The stands the charcoal burners used might be already altered by woodland pasture; the mean diameter values of the mire-near sites point at the usage of small and medium sized wood, typical for coppiced or grazed woodlands. Interpreting solely the pollen diagram, we would reconstruct Abies-rich Fagus woodland with a considerable admixture of Picea abies. But spruce was mainly growing at the margins of the mire, on wet sites, which can be told from the rarity of the conifer in the charcoal spectra.

Synthesis and conclusions Combined results of charcoals from archaeological sites, all examined kiln site ensembles and the two mire profiles (Fig. 6) will now be discussed focusing on the time span since the Bronze Age. Charcoal from archaeological sites in the lowland of the Danube is mainly Quercus and some pine (Pinus sylvestris). Fortifications of the Bronze Age Bogenberg settlement were built with oak and beech wood, hornbeam (Carpinus betulus) was presumably used as fire wood. In the Danube valley, which is poor in mires, the anthracological proof of Carpinus is of special interest, since this tree was the last to reappear in Middle Europe since the last glaciation, and its migration history along the Danube is still poorly understood. During the Subboreal period beech-(Fagus sylvatica)-fir (Abies alba)-forests developed in the Vordere Bayerische Wald. The forest at submontane and montane altitudes was 189

O. Nelle dominated by Fagus with a high proportion of Abies, as revealed by pollen analysis. Today, they are assumed to be the potential natural vegetation in an area with man-made high proportions of spruce (Picea abies). Human impact altered the vegetation since the Middle Ages. Close to the Danube, charcoal production sites show shortage of Fagus due to intense use and thus the dominance of conifers – Pinus, Picea, Abies – in the charcoal spectra. East of Regensburg in the „Forstmühler Forst“ we found numerous pit kilns dating back to early medieval times. Charcoal makers used oak and beech, whereas on 18th century circular kiln sites fir trees were carbonised. Interestingly, in early medieval times oak stands were present at “Forstmühler Forst”, considerably far away from the thermophile southern slopes of the Bavarian Forest. Due to plant sociological studies (Rüther 2003), this is surprising, since dominance of Fagus would not allow growth of Quercus at this altitude and under the site conditions given. Possibly, the forest was already altered by man (woodland pasture) which favoured the growth of oak. Another explanation could be that some oaks still strived as relics from a past age within the beech-forest, before Fagus completely dominated the mesophile mountainous sites. Due to the usage of beech, but also due to favourable site conditions, the proportions of Abies were enriched especially on the lower slopes of the little valley, and were then cut down for charcoal making in the 18th century. In montane elevation at Hirschenstein charcoal kiln sites were investigated in close proximity to the pollen profile of the „Kugelstattmoos“. Combined data suggest as natural vegetation a beech dominated forest around the mountain peaks, with a few fir and spruce, and regular appearance of sycamore (Acer pseudoplatanus). High values of carbonised particles in the upper level of the peat profile indicate charcoal production. In the woods near Ringelai (Freyung-Grafenau), the charcoal spectra of 13 kiln sites, including four which were dendrochronologically dated, show a beech-fir-woodland with sparse spruce in the early 19th century, conspicuously similar to the actual stand composition. It seems that usage of the woodland for charcoal making did not alter the vegetation too much. Spruce (Picea abies) played a minor role in the natural forests. With the anthropogenic opening of the forests, it was favoured in reforestation processes which followed on fallow lands, presumably also favoured by the little ice age. The examples show that anthracological and palynological data complement one another to create a detailed picture of woodland history and landscape change. Given the presence of suitable palaeo-archives, palaeoecological studies on woodland dynamics should combine charcoal and pollen analysis to increase the resolution of the constructed historic scenarios.

Foundation). Thanks are due to Franz Herzig for dating charcoals dendrochronologically, and to Esther Guggenbichler for comments on the manuscript draft.

References Beug, H.J., 1961. Leitfaden der Pollenbestimmung für Mitteleuropa und angrenzende Gebiete. Stuttgart: Lieferung 1. G. Fischer. Bonhote, J., Vernet J.L., 1988. La memoire des charbonnieres. Essai de reconstitution des milieux forestiers dans une vallee marquee par la metallurgie (Aston, Haute-Ariege). Revue Forestiere Francaise, 40 (3), 197-212. Bonhote, J., Davasse, B., Dubois, C., Izard, V., 2002. Charcoal kilns and environmental history in the Eastern Pyrenees (France). A methodological approach. In: S. Thiébault, ed. Charcoal analysis. Methodological approaches, palaeoecological results and wood uses. Proceedings 2nd Int. Meeting of Anthracology, Paris, September 2000, BAR Int. Series 1063. Oxford: BAR Publishing 219-228. Brande, A., 1995. Pollenanalyse zur Bestandesgeschichte der Hochlagenwälder am Plöckenstein (Böhmerwald). Centralblatt für das gesamte Forstwesen, 112 (1), 1-17. Davasse, B., 2000. Forêts charbonniers et paysans dans les Pyrénées de l’est du moyen âge à nos jours. Une approche géographique de l’histoire de l’environment. Toulouse : GEODE (Géographie de l’environment). Fabre, L., 1996. Le charbonnage historique de la chênaie à Quercus ilex L. (Languedoc, France): conséquences écologiques. PhD-Thesis, University of Montpellier II. Faegri, K., Iversen, J., 1989. Textbook of pollen analysis. Chichester, u. a.: John Wiley & Sons. Faegri, K., Iversen, J., 1993. Bestimmungsschlüssel für die nordwesteuropäische Pollenflora. Stuttgart - New York : G. Fischer. Hildebrandt, H., Heuser-Hildebrandt, B., Stumböck, M., 2001. Bestandsgeschichtliche und kulturlandschaftsgenetische Untersuchungen im Naturwaldreservat Stelzenbach, Forstamt Nassau, Revier Winden. Pollenanalyse aus Geländemulden und Auswertung von Holzkohlespektren historischer Meilerplätze. Mainzer naturwissenschaftliches Archiv Beihef, 25. Hillebrecht, M.L., 1982. Die Relikte der Holzkohlewirtschaft als Indikatoren für Waldnutzung und Waldentwicklung. Untersuchungen an Beispielen aus Südniedersachsen. Göttinger Geographische Abhandlungen, 79. Göttingen: Verl. Erich Goltze. Janssen, A., Seibert, P., 1991. Potentielle natürliche Vegetation in Bayern. Anmerkungen zur Arbeitsmethode der Transektkartierung und Auswertung der Ergebnisse. Hoppea, Denkschriften der Regensburgischen Botanischen Gesellschaft, 50, 151-188. Kral, F., 1979. Pollenanalytische Untersuchungen zur Waldgeschichte des Kubany-Urwaldreservates „Boubínský prales“ (Böhmerwald, CSSR). Forstwissenschaftliches Centralblatt, 98, 91-110. Krause, A., 1972. Bestimmung von Meilerkohlen aus dem Hunsrück und ihre vegetationskundliche Aussage. Decheniana, 125, 249-253. Langer, H., 1962. Beiträge zur Kenntnis der Waldgeschichte und Waldgesellschaften Süddeutschlands. Bericht der Naturforschenden Gesellschaft Augsburg, 14 (73)..

Acknowledgements Work was done with the financial support of a research grant from the Deutsche Forschungsgemeinschaft (German Research

190

Combining charcoal and pollen analysis: the Bavarian Forest Ludemann, T., 1994. Vegetations- und Landschaftswandel im Schwarzwald unter anthropogenem Einfluss. Berichte der Reinhold-Tüxen-Gesellschaft, 6, 7-39. Ludemann, T., 1996. Die Wälder im Sulzbachtal (SüdwestSchwarzwald) und ihre Nutzung durch Bergbau und Köhlerei. Mitteilungen Verein für Forstliche Standortskunde und Forstpflanzenzüchtung, 38, 87-118. Ludemann, T., 2003. Large-scale reconstruction of ancient forest vegetation by anthracology - a contribution from the Black Forest. – Phytocoenologia, 33, (4), 645-666. Ludemann, T., Nelle, O., 2002. Die Wälder am Schauinsland und ihre Nutzung durch Bergbau und Köhlerei. Freiburger Forstliche Forschung, 15. Montanari, P., Scipioni, S., Calderoni, G., Leonardi G., moreno, D., 2002. Linking anthracology and historical ecology: suggestions from a post-medieval site in the Ligurian Apennines (north-west Italy). In: S. Thiébault, ed. Charcoal analysis. Methodological approaches, palaeoecological results and wood uses. Proceedings 2nd Int. Meeting of Anthracology, Paris, September 2000. BAR Int. Series 1063. Oxford: BAR Publishing, 235-241. Moore, P.D., Webb, J.A., Collinson, M.E., 1991. Pollen analysis. Oxford: Blackwell Scientific Publications. Müller, K., 1939/40. Das Waldbild am Feldberg jetzt und einst. Dargestellt auf Grund neuer Untersuchungen. Mitteilungen Badischer Landesverein Naturkunde und Naturschutz. Neue Folge, 4, (3-4), 143-156. Nelle, O., 2002a. Zur holozänen Vegetations- und Waldnutzungsgeschichte des Vorderen Bayerischen Waldes anhand von Pollen- und Holzkohleanalysen. Hoppea, Denkschriften der Regensburgischen Botanischen Gesellschaft, 63, 161-361. Nelle, O., 2002b. Charcoal burning remains and forest stand structure - examples from the Black Forest (south-west Germany) and the Bavarian Forest (south-east Germany). In: S. Thiébault, ed. Charcoal analysis. Methodological approaches, palaeoecological results and wood uses. Proceedings 2nd Int. Meeting of Anthracology, Paris, September 2000. BAR Int. Series 1063. Oxford: BAR Publishing, 201-207. Nelle, O., 2003.Woodland history of the last 500 years revealed by anthracological studies of charcoal kiln sites in the Bavarian Forest, Germany. Phytocoenologia, 33, (4), 667-682. Nelle, O., Kwasniowski, J., 2001. Untersuchungen an Kohlenmeilerplätzen im NSG Eldena (Vorpommern) - Ein Beitrag zur Erforschung der jüngeren Nutzungsgeschichte. Greifswalder Geographische Arbeiten, 23, 209-225. Nelle, O., Guggenbichler, E., Putz, U., Schmidgall, J., 2003. Eine mittelalterliche Kohlenmeilergrube im Vorderen Bayerischen Wald. Ergebnisse archäologischer, anthrakologischer und bodenkundlicher Untersuchungen. Archäologisches Korrespondenzblatt 33, (3), 457-467. Rüther, C., 2003. Die Waldgesellschaften des Vorderen Bayerischen Waldes, mit einem Beitrag zur jüngeren Waldgeschichte. Hoppea, Denkschriften der Regensburgischen Botanischen Gesellschaft, 64, 475876. Schweingruber, F.H., 1990a. Anatomie europäischer Hölzer. Bern, Stuttgart: Haupt. Schweingruber, F.H., 1990b. Mikroskopische Holzanatomie. Formenspektren mitteleuropäischer Stamm- und Zweighölzer zur Bestimmung von rezentem und

subfossilem Material. Birmensdorf/Schweiz: Eidg. Anstalt für das forstl. Versuchswesen . Smettan, H.W., 1995. Archäoökologische Untersuchungen auf dem Albuch. In: M. Böhm A. Hauptmann, Kempa M., eds. Beiträge zur Eisenverhüttung auf der Schwäbischen Alb. Forschungen und Berichte zur Vor- und Frühgeschichte in Baden-Württemberg 55, 37-136. Stalling, H., 1987. Untersuchungen zur spät- und postglazialen Vegetationsgeschichte im Bayerischen Wald. Dissertationes Botanicae, 115. Berlin: Stuttgart. Svobodová, H., Soukupová, L., 2000. Mires of the Šumava Mountains: 13,000 years of their development and present-day biodiversity. GeoLines, 11, 108-111. Svobodová, H., Reille, M., Goeury, C. 2001. Past vegetation dynamics of Vltavský luh, upper Vltava river valley in the Šumava mountains, Czech Republic. Vegetation History and Archaeobotany, 10, 185-199. Trautmann, W., 1952. Pollenanalytische Untersuchungen über die Fichtenwälder des Bayerischen Waldes.

191

Environment and agriculture in the early Neolithic of the Arene Candide (Liguria) R. Nisbet

Viale Rimembranza 7, 10066 Torre Pellice, Italy E-mail: [email protected]

Abstract

In summer 1997, the Soprintendenza per i Beni Archeologici della Liguria carried out a limited stratigraphic sampling in the inner part of the cave of Arene Candide, in between the previous excavations by Cardini and Tinè. The new excavation, involving an area of approximately 3 m2 was aimed to make straight the residual section of the 1972-1977 excavation, so that it could be accessed by the public. On this occasion the lower levels (us 9A, us 9B, us 9Bbase, us 10 nuovo scavo), coinciding with the levels 25 and 28 of the Cardini-Bernabò Brea excavation related to the Ceramica Impressa Ligure and Ceramica Cardiale of the Early Neolithic phases, were checked. In these levels, thanks to standardized flotation techniques, the most ancient traces of cereal cultivation in Liguria were found. A barley caryopsis (us 10) was radiocarbon (AMS) dated 6830±40 BP (lab. Beta 110542). The archaeobotanical remains include caryopses of Triticum monococcum, T. dicoccum, T. aestivum/compactum/durum and Hordeum vulgare, in addition to some Lens and Vicia seeds.

Introduction Since 1940 Bernabò Brea’s excavations at the cave of Arene Candide provided one of the early and most impressive stratigraphy in the Middle Holocene of the Mediterranean coast. Followed by other researches in recent years in its Upper Paleolithic (1970) and Neolithic layers (197277; 1997), this site gave the opportunity to study its organic context in several fields of work. The present paper deals particularly with the latest charcoal data, focusing on some palaeo-environmental problems referring to the woodland cover, the use of the timber and some aspects of palaeo-agriculture.

Definition and open problems The present study refers to the problematic levels of the Early Neolithic of the Arene Candide. A discussion on these levels is made possible thanks to the by now ample collection of materials, done during a period of over half century of systematic sampling of charcoals, and in a lesser measure, from the publication of pollen sequences of the cave (Branch 1997; Arobba 1982; Arobba et al. 1999). The stratigraphy of the Arene Candide, characterized by a regular sequence of layers of ashes and charcoals over a large part of the deposit (Maggi 1997a) shows a hiatus of some millennia in the early Holocene. However, the oldest Holocene levels known to now in the cave, are dated to the beginning of the eighth millennium BP and they correspond to a local aspect of the Early Neolithic (Impressed “ligurian” Ware) (Maggi 1997b). For this reason nothing is known in this area of the earliest phases of the wood colonization that took place in the early Flandrian and the setting of the forests of the Atlantic period, as they are recognized in the analyses of pollen and charcoal in the deposit of the cave till now.

This gap is not restricted to the Arene Candide. All the prehistoric caves of Liguria show a similar situation. The same, for example, happens in the not distant Arma dell’Aquila and in the Arma dello Stefanin, in the val Pennavaira. Again in Val Pennavaira, the Arma di Nasino has an anthracological diagram very similar to that of the Arma dell’Aquila, because of the absence of the Mesolithic levels thus resulting in a sudden passage from the Late Glacial, dominated by the pines, to the Early Neolithic with dominant oak (Vernet 1974). The study of the environment of the Early Neolithic is of crucial importance to document the way through which the first phases of the productive economy developed, how man gradually originated some changes in the woodland environment during the passage from a form of nomadism to sedentarism. A further element of interest concerns the forms in which the forest of the Atlantic period have been able to spread, assuming specific characteristics in the diverse surroundings, depending on the climate and soil. This problem has been debated after the stimulating hypotheses recently proposed by Vera (2000). A further difficulty, besides that of finding points of contact between the different disciplines, arises when one considers changes which occurred in a period that covers over a millennium and that seldom offers a stratigraphic and cultural articulation in a same site. Previous data In a recent synthesis of the anthracology related to the Neolithic in some coastal sites between Savona and Marseille, S. Thiébault (2001) has shown that the Early Neolithic is dominated by the deciduous oak woodland. Later on, during the Middle Neolithic, this association was replaced by a more open forest with evergreen sclerophyllous species, that will eventually develop into the mediterranean

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

R. Nisbet “macchia”. At the Arene Candide, till now data have been obtained in regard with the whole VIII millennium BP. Both Fancelli Galletti (1972) and Nisbet (1997), studying the materials of the excavations of Bernabò Brea, have been able to examine the layers 25, 26, 25-26 and 27 in a chronological succession referable to the Early Neolithic, but they could not find any relationship with the post-Glacial woodland environment. However, Castelletti, Castiglioni (1999) could gather an articulation in the anthracological diagram, noticing that in the passage from an earlier holocenic level (layer 16 of Tinè’s excavation, preceding the Neolithic presence but of an unclear age) to the Early Neolithic (15 and 14), no variations are shown in the charcoal remains. They refer therefore the layers 16-14 fully within the Atlantic period. As a consequence the observable variations in the passage from the layer 15 to 14 should be related to an intensified pressure when the human presence in the cave become more systematic. Nisbet (1997) notices that for the whole Early Neolithic vegetational changes are not recorded in the composition of the charcoal samples, perhaps because the human presence was neither temporally continuous nor so numerous as to require quantities of wood for heating and building or other pastoral work. He has observed that the represented species are numerous, but only some (Quercus sp., Pistacia lentiscus, Phillyrea sp, Arbutus unedo) systematically appear in the samples. The recent excavations on circumscribed sectors of the Arene Candide and a rigorous stratigraphic control, allows

us now to individualize precisely two cultural moments within the Early Neolithic. It has also been helpful that several researchers have worked on the site independently and in different times. In all we dispose now of 1235 charred fragments for a reconstruction of the wood environment of the Early Neolithic. This fact allows us to frame with better precision the anthracological evolution on the whole VIII millennium BP. The new anthracological data (Excavations 1997) As shown in Fig. 1 and in Table 1, in the us 10, the earliest Neolithic layer so far studied in the cave (Impressed “ligurian” Ware), 9 taxa are documented on a sample of 120 charcoals. Of these, four are more frequent and, among these, the deciduous oak is largely dominant. To a lesser extent Arbutus unedo, Phillyrea sp. and Pistacia terebinthus were found. There is little evidence for thermophilous and xerophilous species. The layer 9b, about 500 radiocarbon years later, shows a potentially uniform forest structure, very similar to the previous one. The dominance of the deciduous oak continues, with a good presence of Phillyrea and Arbutus. The Turpentine tree has disappeared, but this should be considered as a consequence of the reduced sample dimensions rather than a true vegetational change, since the environmental context appears uniform and stable during the whole period. The occurrence of Ash is of interest, and remarkable is the presence of Pinus halepensis. The number of taxa slightly increases, in com-

Arene Candide - Early Neolithic 70

60

50

40

%

Neolitico cardiale Impressa ligure

30

20

10

0 u erc Qu

sc

s o s a re us us sis sp. sp. sp. uc . s p. ilex gu ed ga ore lmu en ad tisc inth us ae un us cia rus rea vul arb cf U halep len rat xin reb pe illy sta erc a i m tus i a C e a h c u r n u u i t i P f r P F Q c Ju Er ia us tac Arb ust tac Pin Pis Lig Pis

Taxa Fig. 1.

194

Environment and agriculture in the early neolithic of the arene candide stratigraphy us 9A sez. S 1-5

Neolitico Antico a ceramica cardiale

us 9A

us 9B

us 9B sez. S 1-5

Neolitico Antico a Impressa ligure

stratigrafia us 10

us 10 sez. S 1-5

Taxon Quercus caduc. Arbutus unedo Phillyrea sp. Pistacia sp. Pistacia lentiscus Juniperus sp. Quercus caduc. Phillyrea sp. Arbutus unedo Pistacia sp. cf Crataegus Quercus caduc. Arbutus unedo Phillyrea sp. cf Crataegus Pinus halepensis Juniperus sp. Erica arborea Quercus ilex Pistacia sp. Phillyrea sp. Quercus caduc. Arbutus unedo Erica arborea Pistacia terebinthus Ligustrum vulgare cf Crataegus Fraxinus sp. Quercus ilex

carried out anthracological analyses in these regions. In the oak woodland several arboreal forms are present, though clearly subordinate (like ash, elm, strawberry tree, different species of Pomoideae, hornbeam, Laburnum elder, maple). This variety of forms, even though in reduced quantity, seems to suggest the presence of deep and mature soils, inherited from the early Holocene. During the following centuries there appeared a vegetation formed by species that will become dominant some millennia later in the mediterranean “macchia”. First of all we find Quercus ilex or, less probably, another evergreen oak. Other elements arise, forming the shrub layer of the woodland, such as Rhamnus, Erica cf arborea, Pistacia lentiscus and P. terebinthus, Phillyrea sp. These plants played a limited role in the oak woodland, but occasionally they could prevail where the latter was reduced for occasional reasons (fire, erosion of the soil etc.). With regard to this aspect the presence of Pinus halepensis, already documented by the analyses of S. Thiébault on the French coast, is of particular interest. Its rarity at Arene Candide (us 9b) does not allow us to establish whether it constituted indeed the consequence of a forest fire clearing, or it simply reflected an opening of the oak woodland due to exposition, morphology etc. reasons. The first hypothesis can be advanced, if we take into account the pressure that the economy of the first settlers had played on the environment some centuries before (sheep, agriculture etc.). A similar consideration can be suggested for the presence of junipers (Juniperus oxycedrus or J. phoenicea), that are generally favored by woodland fires more than by grazing. Again in this case an environment favorable to the spread of these bushes could be represented by the rocky slopes surrounding the cave, or even by the surface of the dune in front of the cave.

n. fragms. volume 35 mL 21 11 11 1 1 1 235 mL 45 16 7 1 1 380 mL 68 9 17 4 1 2 1 2 4 125 mL 4 21 4 1 1 1 2 1 4

Taxon

n. frr.

Quercus caduc. Phillyrea sp. Arbutus unedo cf Ulmus Pistacia terebinthus cf Crataegus Quercus ilex Erica arborea

43 16 24 1 4 3 1 3

Arbutus unedo Quercus caduc. Erica arborea Pistacia terebinthus Quercus cf ilex Juniperus sp.

2 10 3 7 2 1

volume 270 mL

60 mL

cf T. dicoccum 7,7%

Hordeum 13,8%

T. dicoccum 9,2%

cf Hordeum 3,1%

T. monococcum 3,1%

Table 1.

cf T. monococcum 6,2%

parison with the us 10. That could be simply due to the larger number of analyzed charcoals (147). Finally, in the upper layer (us 9a, 116 charred fragments), with a possible date around the first or second century of the seventh millennium BP (therefore following the 9b layer by about one or two centuries) we find a small number of taxa, but again suggesting a similar stable woodland environment: predominance of the broad-leaved deciduous oak, good representation of Phillyrea and Arbutus; to a lesser extent Pistacia lentiscus, Pomoideae and Juniperus. From the above data the following observations can be made. The association of the deciduous oak in the coastal regions of the Liguria and the Provence is once more confirmed, as already observed by all the authors that have

cf Triticum sp 7,7% Triticum sp 1,5%

Cerealia 38,5% cf T. "aestivum" 1,5%

Arene Candide - Early Neolithic

Fig. 2.

195

T. "aestivum" 7,7%

R. Nisbet Some shrubby plants such as Arbutus unedo and Erica arborea (and in general, all the evergreen shrubs such as different species of Cistus), find a favorable development in the open spaces of the oak woodland. Early crops The first evidence of pre-Middle Neolithic agriculture from Arene Candide has been documented during the rescue excavation of 1997 in the Impressed “ligurian” Ware layers of the cave (Fig. 2). With the analogous materials from S. Sebastiano di Perti (Arobba, Vicino 1997-99), these are the only remains of crops of Early Neolithic in the whole coastal areas of Liguria so far recognized. A radiocarbon (AMS) date performed on a single caryopsis (us 10) gave the date Beta 110542: 6830±40 BP (Binder, Maggi 2001). Given the limited surface of the 1997 dig it was decided to float the whole volume of sediment for level, i.e. the us 9a (250 l), 9b (300 l) and 10 (250 l). The samples from the flotation occupy a volume of 620 ml (9b), 470 ml (9a) and 410 ml (10). In total, 42 fragments of palaeo-agriculture have been recognized, plus 23 measurable caryopses. In the deepest layer (us 10) 5 caryopses of Triticum aestivum/durum have been determined, 1 of T. monococcum, 6 of T. dicoccum, 7 of Hordeum vulgare; 34 caryopses of cereals remain of uncertain determination. In us 9b there are present 1 caryopses of Triticum monococcum and 1 of Hordeum vulgare, plus 9 of uncertain determination. In the us 9a only a caryopsis of Hordeum vulgare is present. All the caryopses of Hordeum vulgare appear twisted, with an asymmetrical axial furrow. For this reason we would attribute them to H. vulgare rather than to the one with a symmetrical furrow, H. distichon. In us 10 there are also present one cotyledonal fragment of lentil, Lens culinaris, and a seed of Chenopodium album, a weed frequent on trampled soils, in gardens, in loose and nitrogenous soils. In us 9a there are present two spherical seeds of Vicia sp. In a comprehensive view, the palaeo-agriculture of the Arene Candide does not differ from what already known elsewhere in Northern Italy. Along with the S. Sebastiano findings, it nevertheless comes to fill what appeared to be an incomprehensible gap among Ligurian coasts and the Mediterranean France. It also settles the problem of the palaeo-economic relationships with the nearby geo-ecological and cultural areas, both coastal and continental. At the same time it confirmed also for the Italian Western side the diffusion of agriculture as early as, or even a few centuries earlier than those Neolithic sites already known in North-Eastern Italy (Improta, Pessina 1998), since the first half of the sixth millennium BC. The data from Mediterranean France is larger. Here we find numerous Cardial (Early Neolithic) sites with evidence of wheat (above all Triticum aestivo-compactum, T. monococcum and T. dicoccum) and of barley, for instance at La-Font-des-Pigeons (Châteauneuf-les-Martigues) (Cour196

tin et al. 1976), at Grotte de l’Aigle (Méjannes-le-Cap) (Erroux 1979; Courtin, Erroux 1974) at Fontbrégoua (Courtin, Erroux 1974); and also of spontaneous pulses since the Mesolithic at Balma Abeurador (Vaquer et al. 1986) and again at Fontbrégoua. Previous analyses by J Renfrew (Evett, Renfrew 1971) at Arene Candide have shown the presence of impressions of spiklets of Triticum dicoccum in layer 22 (excavations Bernabò Brea-Cardini), and therefore referable to the culture SMPI (Middle Neolithic). Still to the culture SMP belong some caryopses of Triticum compactum Host. (Nisbet 1985). A conspicuous contingent of seeds and caryopses of the Middle Neolithic layers from Arene Candide is under study. However the existence of some agricultural activity in the cave has already been suggested from the presence of elements related to the prehistoric agrarian technology. The microscopic study of some worn areas of possible grindstones and the microwear analysis of chipped stone tools from the layers 25-27 have allowed to point to the presence of few elements of “sickle” or “mill” (Starnini, Voytek 1997a; Starnini and Voytek 1997b).

Interpretation of the data To interpret the wood/agriculture general data of the Early Neolithic, some ecological elements of primary importance should be appraised: 1. The direct or indirect action of both man and wild and domestic grazers); 2. The previous Early Holocene forest history; 3. The location of the coastline in the seventh millennium and its earlier variations; the morphology of the slope. 1. The analyses of Rowley-Conwy (1997) show that the oldest Neolithic level documented in the cave contains bones of sheep and pig. However, grazing animals like red and roe-deers were certainly present before the arrival of the Neolithic settlers. In particular, these wild ruminants have different alimentary customs. According to Vera, the roe deer feeds preferably on leaves and twigs of trees and shrubs, while the red deer eats also large quantities of grasses. As a result both species could have well contributed to the formation of gaps in the canopy and of small parkland areas in the dominant deciduous oak forest, in this way favoring the development of the early pastoral economy based on sheep. 2. Unfortunately the archaeological deposits with evidence of a Mesolithic presence in the coastal Liguria are not known; neither are investigations on charcoal exist for that period in the region. 3. The rapid post-glacial eustatic rise of the sea-level played a non secondary role in the forest regeneration. This would have caused the disappearance of some kilometers of coastal lowland and the consequent disappearance or the compression of elements of the coast flora on the lower slopes.

Environment and agriculture in the early neolithic of the arene candide 4. Even confining ourselves to the study of the present morphology, we find that, in a range of one km from the cave, in the territory currently emerged, the 35% of the ground has a slope of 75% or more; a more restricted part was in antiquity occupied by the dune, and only a 30% is gently sloped, therefore permitting soil formation. In such a complex morphology it is not surprising that a mosaic of vegetation was represented in the charred wood samples. This could indeed represent an excellent argument for the choice of the cave as a quasi-permanent prehistoric setting. Admitting that man appeared in the region in correspondence with the us 10 of the 1997 dig, then the corresponding anthracological spectrum should represent with a certain degree of probability the forest structure of the area as it developed in the early millennia of the Holocene. It is observed that in a forest dominated by the broad-leaved oak, the evergreen sclerophyllous trees and bushes are already present. They probably grew in the areas less favorable for the diffusion of the oak, such as the steep or the rocky slopes, the unfavorable soils and exposure, or previous episodes of spontaneous occasional fires. The analyses carried out by several authors, on this and on other coastal sites of Liguria and Provence, show that practically all the species that will characterize, in the following millennia, the formations of the “macchia” (garrigue) are already present in the forest environment of the Atlantic period, though probably reduced or limited to zones peripheral to that of the deciduous oak woodland. The mosaic structure of the local vegetation is confirmed by other analyses. The study of the land molluscs (Girod 1997) assigns, for the Early Neolithic, a notable importance to associations distinctive of stony soils and rocky environments (Solatopupa similis, Helicigona cingulata), excluding any type of arboreal vegetation. This was still present around the cave, as proved by other species of molluscs of shaded and humid environment. Finally the mesophile snail association is present as a non secondary component, that could largely correspond to a tall scrub environment. Also the pollen analyses (Branch 1997) show the presence of warm associations, where the grassy component is important, forming about the 40% of the whole pollen spectrum for the Early Neolithic. This clearly shows that the local canopy forest was not completely closed. This is confirmed, on the other hand, by the direct presence of cereals consumed in the cave, and certainly cultivated in the surroundings . Even without attributing a crucial role to the presence of wild herbivores in the local wood formation, and accepting the criticism expressed to Vera’s model (e.g. Moore 2005) it seems however that the vegetation of the slope knew a notable variety of species before the arrival of man, a diversity that we would attribute to a variety of ecological factors operating during the ancient Holocene. The local economy will work as an ecological factor, in a first moment, with the increase of pastoral activities (forage, construction of

enclosures, firewood) and of the agriculture. Nevertheless it seems that, for the whole Early Neolithic and the Middle Neolithic this action was not able to produce consistent changes on the structure of the ancient wood cover. The most important modification was the strong contraction of the broad-leaved oak and the expansion of the evergreen oaks, contemporary to the definitive expansion of the macchia with arboreal and shrubby species like Arbutus unedo, Erica arborea, Rhamnus, Phillyrea and Pistacia. All these species, however, were already present, though secondarily, in marginal areas since the ancient Holocene. From the ancient phase of the Early Neolithic to the recent phase of the Early Neolithic apparent changes are not evident, although agriculture is already appearent in the layer 10. On the other hand it has been underlined that the ancient cereals did not require particular techniques of preparation of the ground for sowing, consisting namely in the preliminary eradication of the grasses and in the fragmentation of the soil. It is therefore unlikely that these changes could have had a strong impact on the local vegetation, at least until the agriculture was definitely practiced on reduced surfaces by semi-permanent human groups.

References Arobba, D., 1982. Étude palynologique de la Grotte des Arene Candide (Ligurie italienne): premiers résultats sur quelques niveaux du Mésolithique et du Néolithique inférieur et moyen. Ier Congrès International de paléontologie humaine, Nice 16-21 Octobre (résumés des communications). Nice: Louis-Jean, 171-172. Arobba, D., Vicino, G., 1997-99. Segnalazione di macroresti botanici nel sito neolitico di S. Sebastiano di Perti (SV). Bollettino dei Musei Civici Genovesi, XIX-XXI, 29-35. Arobba, D., Giacobini, G., Castelletti, L., Gardini, G., Meriggi, A., Ottoboni, F., 1999. Analisi di un coprolite rinvenuto nei livelli del Neolitico Medio. In: S. Tinè, ed. Il Neolitico nella Caverna delle Arene Candide (scavi 1972-1977). Coll. Monografie Preistoriche Archeologiche, X. Bordighera: Istituto Internazionale di Studi Liguri, 2535. Binder, D., Maggi, R., 2001. Le Néolithique ancien de l’arc liguro-provençal. Bulletin de la Société Préhistorique Française, 98 (3), 411-422. Branch, N.P., 1997. Palynological study of the early and middle Neolithic cave deposits of Arene Candide: preliminary results. In: R. Maggi, ed. Arene Candide: a functional and environmental assessment of the Holocene sequence (excavations Bernabò Brea-Cardini 1940-50). Memorie dell’Istituto Italiano di Paleontologia Umana,V, 89-102. Castelletti, L., Castiglioni, E., 1999. Resti antracologici. In S. Tiné (a cura), Il Neolitico nella Caverna delle Arene Candide (scavi 1972-1977). In: S. Tinè, ed. Il Neolitico nella Caverna delle Arene Candide (scavi 1972-1977). Coll. Monografie Preistoriche Archeologiche, X. Bordighera: Istituto Internazionale di Studi Liguri, 1824. Courtin, J., Erroux, J., 1974. Aperçu sur l’agriculture préhistorique dans le Sud-Est de la France. Bulletin de la Société préhistorique française, 71 (1), 321-334.

197

R. Nisbet Courtin, J., Erroux, J., Thommeret, J., 1976. Les céréales du Néolithique ancien de Châteauneuf-les-Martigues (Bouches-du-Rhône). Bulletin Musée Histoire Naturelle Marseille, 36, 11-15. Erroux, J., 1979. Détermination de graines carbonisés. In: J.L. Roudil et al., eds. La grotte de l’Aigle à Méjannes-le-Clap (Gard) et le Néolithique Ancien du Languedic-Oriental. Mémoires Société Languedocienne Préhistorique (1). Evett, D., Renfrew, J., 1971. L’agricoltura neolitica in Italia: una nota sui cereali. Rivista di Scienze Preistoriche, 26 (2), 403-409. Fancelli Galletti, M.L., 1972. I carboni delle Arene Candide e l’evoluzione forestale in Liguria dopo l’ultima glaciazione. Atti della Società Toscana di Scienze Naturali, Serie A, Pisa, 79, 206-212. Girod, A., 1997. Arene Candide: Holocene land-snails. In: R. Maggi, ed. Arene Candide: a functional and environmental assessment of the Holocene sequence (excavations Bernabò Brea-Cardini 1940-50). Memorie dell’Istituto Italiano di Paleontologia Umana, V, 125-135. Improta, S., Pessina, A., 1998. La neolitizzazione dell’Italia settentrionale. Il nuovo quadro cronologico. In: A. Pessina, G. Muscio, eds. Settemila anni fa il primo pane. Ambienti e culture delle società neolitiche. Udine: Museo Friulano di Storia Naturale, 107-115. Maggi, R., 1997a. The excavation by Luigi Bernabò Brea and Luigi Cardini of the Cave of Arene Candide within the historical context of the study of prehistory in Italy. In: R. Maggi, ed. Arene Candide: a functional and environmental assessment of the Holocene sequence (excavations Bernabò Brea-Cardini 1940-50). Memorie dell’Istituto Italiano di Paleontologia Umana, V, 11-30. Maggi, R., 1997b. The radiocarbon chronology. In: R. Maggi, ed. Arene Candide: a functional and environmental assessment of the Holocene sequence (excavations Bernabò Brea-Cardini 1940-50). Memorie dell’Istituto Italiano di Paleontologia Umana, V, 31-52. Moore, P.D., 2005. Down to the woods yesterday. Nature, 433, 588-89. Nisbet, R., 1985. Aspetti cronologici dell’agricoltura neolitica dell’Italia settentrionale. In : A. Moroni, A. Anelli, O. Ravera, eds. Atti Secondo Congresso Nazionale della Società Italiana di Ecologia, Padova, giugno 1984. Padova: Società Italiana di Ecologia, 943-945. Nisbet, R., 1997. Arene Candide: charcoal remains and prehistoric woodland use. In: R. Maggi, ed. Arene Candide: a functional and environmental assessment of the Holocene sequence (excavations Bernabò BreaCardini 1940-50). Memorie dell’Istituto Italiano di Paleontologia Umana, V, 103-112. Rowley-Conwy, P., 1997. The animal bones from Arene Candide. Final report. In: R. Maggi, ed. Arene Candide: a functional and environmental assessment of the Holocene sequence (excavations Bernabò Brea-Cardini 1940-50). Memorie dell’Istituto Italiano di Paleontologia Umana, V, 153277. Starnini, E., Voytek, B., 1997a. The Neolithic chipped stone artefacts from the Bernabò Brea-Cardini excavations. In: R. Maggi, ed. Arene Candide: a functional and environmental assessment of the Holocene sequence (excavations Bernabò Brea-Cardini 1940-50). Memorie dell’Istituto Italiano di Paleontologia Umana, V, 349426.

Starnini, E., Voytek, B., 1997b. New lights on old stones: the ground stone assemblage from the Bernabò Brea excavation at Arene Candide. In: R. Maggi, ed. Arene Candide: a functional and environmental assessment of the Holocene sequence (excavations Bernabò BreaCardini 1940-50). Memorie dell’Istituto Italiano di Paleontologia Umana, V, 427-511. Thiébault, S., 2001. Anthracoanalyse des établissements néolithiques de la région liguro-provençale. Bulletin de la Société Préhistorique Française, 98 (3), 399-409. Vaquer, J., Geddes, D., Barbaza, M., Erroux, J. , 1986. Mesolithic plant exploitation at the Balma Abeurador (France). Oxford Journal of Archaeology, 5 (1), 1-18. Vera, F.W.M., 2000. Grazing ecology and forest history. Wallingford, UK: CABI. Vernet, J.L., 1974. Précisions sur l’évolution de la végétation depuis le Tardiglaciaire dans la région méditerranéenne d’après les charbons de bois de l’Arma de Nasino (Savone-Italie). Bulletin AFEQ, 39, 65-72.

198

A contribution to the forest history of the Markstein area in the southern Vosges (France) W. Nölken

Albrecht-Ludwigs-Universität, Institut für Biologie II / Geobotanik, Schänzlestr. 1, 79104 Freiburg (Germany) E-mail: [email protected]

Abstract Former charcoal burning places are used by Anthracology as an archive, in which information about the tree combination at the time of use is conserved. The aims of the researches in the Markstein area are to find charcoal places, determine the wood species used and therefore get a picture of the former forest stands close to the burning spot. To find similarities or differences in the charcoal assemblages, statistical methods are used. As an explanation of the different distributions of the tree species found in the charcoals, the elevation above sea level could be worked out as the mayor ecological factor. The often suspected firewood selection by the charcoal burner was not confirmed by the investigations. Because spruce was not existing in the samples, the general introduction of this species in the Vosges could be used as a timemark for the charcoal producing period.

Introduction How the tree species have naturally been distributed in a given area is always a difficult question to answer. Different methods can be used to achieve that goal, depending on the spatial resolution. If one wants an overview on a regional scale, than the pollen analysis is the method of choice. To get a higher spatial resolution, other methods must be taken into account. One source of information with an “on spot” description of a species community can be written records. But those information are scarce, especially before the introduction of the forest laws (Code Forestier 1827). Another source of information about the species composition at the local resolution scale are charcoal places. Because the charcoal burners depended on the natural supply of wood in the close surroundings, this bioarchive can be used to reconstruct the tree species compositions at the burning sites. An advantage of this method is the remarkably high distribution of such places in the mountainous areas of middle Europe (Wullschleger 1979; Hillebrecht 1982; Metailie 1993; Montanari et al. 2000; Nelle 2003; Ludemann 2003). The burning spots are easy to find and usually contain a lot of determinable charcoal pieces in a black layer. The Markstein area in the Vosges mountains was chosen as an example for the fine-scale resolution of the charcoal places analyses. The main aim of the research is to find out the natural tree species composition at the different sites. Another aspect of the research is to find out, if the species found in the charcoal can be connected to the altitudinal belts of the Vosges. Combined with this task is the question at which time the forest in the Markstein area was used for charcoal making. After the determinations of the charcoals the results are compared to today’s tree compositions at the different sites.

Fig 1. The Vosges mountains in eastern France.

Study area The investigation area is situated at 690 m above sea level on the east side, rises up to 1230 m and than goes down to 860 m on the west side (Fig. 1). Situated in the arc with the highest peaks in the Vosges, Greywacke builds up the Markstein massif and the established soils are rankers, podzols and brown earth (Walter 1992; Boudot et al. 1981). The average annual rainfall ranges from 1800 mm in the lower regions to 2160 mm at the top (Reklip 1995). The average temperature in the peak region is 5° to 6° Celsius, in the lower regions it ranges from 6° to 8° (Reklip 1995).

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

W. Nölken

Fig. 2. The study area “Markstein” and analysed charcoal places.

Material and Methods

Results

The material of the study are the remains of the former charcoal burning processes. Today the flat basis of charcoal kilns can still be found in the field. The flat basis is easy to find in the mountainous terrain and has an average size of 10 x 10 m. Usually a black and charcoal enriched horizon can be found at those sites. At all charcoal places one sample set was taken out of the top charcoal layer. The set consists of five sub samples with at least 20 to 25 charcoal fragments each, making sure that 100 pieces per place can be determined. The size of the charcoal fragments depended on the available material. It was always tried to get the same amount of small (0.25 cm3) and big (up to 5 cm3) pieces. This sampling design lowers the chances of collecting only pieces of a certain taxa due to a local enrichment, which can not be excluded. (Davasse 2000). At all places just one charcoal layer could be distinguished. In a 50 m semicircle above the charcoal place, the actual forest communities and tree composition were ascertained. The determination of the charcoals follows Schweingruber (1982), using a stereoscope (Leica MZ 12) and an incidentlight microscope (Zeiss Universal M III C) as well as a reference collection. In addition, the diameter of every charcoal piece was measured. The fragment was fit into a stencil of different diameter sizes, according to the bending of the growth ring. The five diameter classes are up to 2 cm, 2 to 3 cm, 3 to 5 cm, 5 to 10 cm and bigger than 10 cm.

Altogether 25 charcoal places were sampled in the Markstein area (Fig. 2), with 3056 pieces. The collected material weighs 2.7 kg and consists of wood from 11 different taxa (tab. 1). On average, 122 pieces of charcoal could be determined per sample. Today 12 different tree taxa can be found in the close vicinity of the charcoal places. Betula, Populus and Prunus are present only in the charcoal material; Picea, Pseudotsuga, Sambucus and Aesculus only in the recent forest composition. All other taxa are present in both categories. The today’s trees species are beech (Fagus sylvatica), silver fir (Abies alba), maple (Acer pseudoplatanus), ash (Fraxinus excelsior), elm (Ulmus glabra), hazel (Corylus avellana), rowan (Sorbus aucuparia and S. aria), willow (Salix caprea and S. aurita), spruce (Picea abies), douglas fir (Pseudotsuga menziesii), elder (Sambucus nigra) and horse chestnut (Aesculus hippocastanum). The anatomic determination of wood is only possible down to genus level, but some genera have only one species in the study area. The charcoals of these genera can therefore be assigned to a single species (Fagus, Abies, Fraxinus and Corylus). All other genera have more species in the closer study area. But, knowing the present day tree population and the ecological demand of the species, it is possible to come to a realistic assessment of the other charcoals species. The genus Acer, Ulmus, Sorbus and Salix are most likely the same species as in today’s vegetation (see 200

Prunus

Abies

Acer

Fraxinus

Ulmus

Corylus

Sorbus

Salix

Picea

Pseudotsuga

Sambucus

Aesculus

2 1 1 3 0 0 0 0 0 0 1 0 1 3 1 0 3 3 3 3 3 3 3 3

2 0 3 2 0 1 1 0 0 0 3 1 0 1 1 1 3 1 1 1 2 2 2 2

3 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 2 2 1 1 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0

1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1

0 0 0 0 1 0 0 0 0 0 3 1 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0

1 0 0 2 0 2 2 3 3 1 0 2 1 0 3 3 0 0 3 3 2 4 4 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 3 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0

0 0 22 1

2 0 12 1

3

3

2

0

0

1

0

0

1

0

1

0

6

4

25 17 19 8

2

8

3

2

18 2

4

2

0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 0 1 0 0 0 0 0

0 0 0 1 0 3 3 5 0 0 0 1 0 0 0 0 8 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0

130 48 46 5 0 3056 2177 689 138 4

0 2

0 6

0 4

2

2

3

25

17 18 2 45 1 1 1 3 4 0 0 1 1 28 28 40 15 43 29 39 27 74 38 29

23

Acer

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0 2 0

Abies

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0

Fagus

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0

79 74 88 38 99 93 95 89 95 100 94 98 99 70 68 56 36 55 57 60 67 21 59 69

4 8 10 17 0 4 1 3 1 0 6 0 0 0 4 5 27 2 8 1 5 4 1 2

20

1

1

1

Fagus

Salix

4 4 3 3 3 4 4 3 3 4 3 3 4 3 3 2 3 3 1 1 3 2 2 3

Populus

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0

Sorbus

Ma32 690 Ȉ (pieces) Ȉ of places with Taxon

0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 8 0 1 0 0 0 0 0

Corylus

110 119 102 121 119 107 107 110 115 117 130 105 119 111 117 240 105 110 112 132 150 119 133 116

Betula

860 950 970 1010 1230 1200 1200 1190 1190 1180 1160 1130 1120 1000 945 915 905 850 790 780 710 710 710 690

Ulmus

n charcoals

Mat10 Mat11 Mat12 Mat13 Ma2 Ma21 Ma14 Ma13 Ma12 Ma1 Ma39 Ma23 Ma10 Ma11 Ma38 Ma37 Ma3 Ma9 Ma22 Ma5 Ma36 Ma35 Ma34 Ma33

Treecover in the fuel catchments semicircle*

Fraxinus

Elevation (m)

Charcoal: Genus [%]

Burning n = 25

place

A contribution to the forest history of the Markstein area in the southern Vosges

Table 1. Anthracological results and the present tree vegetation in the close vicinity of the charcoal places in the Markstein area (* cover: 4 = 51-100%, 3 = 26-50%, 2 = 25-6%, 1 = 1-5%).

above). Betula (birch) can most probably be assigned to B. pendula, Populus (poplar) to P. tremula and Prunus (cherry tree) to P. avium. Beech is the dominant species in the analysed material, as well as in the present day vegetation, at every study site. It is followed by silver fir and maple. In the present state spruce is playing a mayor role in the forest community. This species was not found in the charcoal material. Silver fir occurs in 17 study sites in the present vegetation, but charcoal fragments of that tree could be found in 23. The plots above 1100 m a.s.l. contribute most to the loss.

some pieces. In the region between 690 m and 1010 m a.s.l. the Fagus – Abies ratio evens out and the percentages do not differ that much. Testing the correlation between Fagus and elevation and Abies and elevation with Spearmann-Rho came to the result, that they are both significant (Fagus – elevation: 0.752, Abies – elevation: –0.772, significant on the 0.01 error level, two – tailed). Two charcoal places are dominated by silver fir remains. These are the results from the burning place Mat13 and Ma35. At the first spot, a good deal of maple could also be found (17 %) and the percentage of Abies is only slightly over that of Fagus (45 : 38 %). The second burning place is clearly dominated by Abies remains with a 74 % share. The place Ma3 is also remarkably. Eight genera could be found in the material of this burning site, maple has its highest share of all analysed charcoal places with

Taking a closer look at the elevation of the study sites and the abundance of Fagus and Abies charcoals a connection can be established. Beech is the dominant species above 1100 m a.s.l., whereas silver fir is only represented by 201

W. Nölken The second group consists of five places with a Fagus share of 48 to 60 % and an Abies share of 39 to 46 %. Seven sites are building up the last unit, with higher Fagus shares of 58 to 79 % and lesser Abies shares of 18 to 30 %, compared to the second unit. These groups are not clearly separated by their elevation.

27 %. Also birch (4 %), hazel (2 %), rowan (8 %) and willow (8 %) have their maximum shares at this site. For the cherry tree it is the only appearance in the places analysed. The materials from Ma22 consist of seven genera, with the only proof of ash and poplar for the whole Markstein area. It is also the place, where the highest number of today’s tree species (9) could be found. In the next step of the analyses it was tried to find distinguishable groups in the charcoal material. For this purpose a hierarchical cluster analyses was used (Fig. 3).

To find out whether these three units are statistically distinguishable or not, a discriminate analyses was used. Two discriminate factors explain 100% of the variance in the data matrix, the first alone 99.1%. The classification of each charcoal place into its group

Fig. 3. Dendrogram of the Markstein charcoal places using Average Linkage (Between Groups). The fundamental basis is the charcoal matrix (table 1).

The dendrogram shows three main units and three charcoal places, which could not be grouped on an high similarity level. These three study sites are the ones with a clear Abies dominance (Ma35) or high Acer values (Mat13 / Ma3). The first group consists of ten places with high to singular beech charcoal content. Taking a look at the elevation of the sites, it can be seen that all but one (Mat12) lay above 1100 m a.s.l. Maple is less frequent in this group than in the other two (Tab. 1).

is significant on the 0.01 error level. Only three study sites could not be assigned to one of these groups (Fig. 4). Because of the high explanation value of the first discriminate function, the next step is to find out what ecological factor hides behind it. Signs point strongly in the direction of the elevation a.s.l., keeping in mind the highly positive correlation of Fagus and the highly negative correlation of Abies, as well as the fact that all study sites of group one lay above 1100 m a.s.l. 202

A contribution to the forest history of the Markstein area in the southern Vosges The diameter values also reflect the potential of the trees to grow at different elevations. In the figure 6 the results of the diameter measurements of silver fir and beech pieces are represented. Comparing the results of the diameters of the charcoal fragments collected above 1100 m a.s.l. and below 1000 m a.s.l. it is obvious, that the pieces of the lesser elevation tend towards the larger categories. The diameters of Fagus and Abies below 1000 m a.s.l. have their maximum in the category > 10 cm. The percentage share for each following diameter class is descending, except for the silver fir share in the up to 2 cm category. Above 1100 m a.s.l., Abies has his maximum in this category with an almost 70 % share. The values of Fagus don’t have a clearly distinguishable maximum in a certain class. The three biggest

Fig. 4. Canonical discriminate analyses. Basis is the charcoal matrix (tab. 1). Three groups of charcoal assemblages were tested and confirmed.

In fig. 5 the values of the first discriminate function of each charcoal place are plotted against their elevation. The scatter plot shows a strong linear correlation between this two elements (r = 0.78).

Fig. 6. Compilation of the diameter values of silver fir and beech charcoals at different elevations.

categories have an almost even share of 23 %. The dating of the kilns was done with the radiocarbon method in the Beta laboratory in Florida. The exact time of charcoal burning in the Markstein area is difficult to define. Seven 14C analysis date the charcoals to the modern time (table 2). In world war one the last verifiable charcoal burning was done in the upper Vosges (Balmier and Roess 2002).

Fig. 5. Scatter plot of the elevation a.s.l of the charcoal places and the values of the 1. discriminate function.

Charcoal Beta No. Conventional 14 places C age (BP) MA1 179259 140 ± 50

2Sigma calibration (95%p) AD 1660 - 1950

MA2

179248

180 ± 50

AD 1640 - 1950

MA3

179243

140 ± 50

AD 1660 - 1950

MA10

179244

120 ± 60

AD 1660 - 1950

MA12

179246

30 ± 50

-

MA22

179247

150 ± 50

AD 1650 - 1950

MA35

179241

120 ± 50

AD 1660 - 1950

Table 2. Results of radiocarbon dating of charcoals from 7 kiln sites in the Markstein area, Vosges (France).

203

W. Nölken Discussion

window of 160 years (1660 – 1820 AD) can be given for the charcoal burning period. Compared to other charcoal place studies in the Vosges (Nölken 2003) the share of pioneer tree species is much less and the diameters of the charcoals are much higher. The forest in that study was partly devastated and in an early stage of succession at the time of charcoal burning. This can be ruled out for the forests analysed at the Markstein massif.

The species found in the anthracological analyses all belong to the natural mountainous forest vegetation of the Vosges (Issler 1942). Eleven different taxa build up the sample material. This can be interpreted as a usage of most of the forest stand for the charcoal production process. The distribution of the wood taxa in the charcoal assemblages can be explained quite well by the differences in the ecological conditions at the study sites, as shown in earlier investigations (e.g. Ludemann et al. 2005). Indications of wood selection could not be found. If, in the time of the charcoal burning, timber had been of higher value than charcoal, than the share of silver fir in the samples would be lower. Whether the burners used the different wood in a single kiln or separate from each other could not be verified. The statistical analyses revealed three main groups of charcoal assemblages in the Markstein region; they also revealed an elevation gradient in the data. The elevation above sea level has an influences on several important ecological factors. The factors enclosed in elevation are: a precipitation gradient, a temperature gradient and the duration of the vegetation period. Also important are the usually less developed and shallower soils at the mountain tops and the strong winds (Carbiener 1966). The exposition of the study sites has no verifiable influence on the species found in the charcoal. The diameter values also indicate an elevation gradient in the data. The analyses of the samples collected above 1100 m a.s.l. shows lower diameters than the material below 1000 m a.s.l. This is due to the growth conditions in case of the beech. The ecological conditions force the beech to grow less (Issler 1942). The use of the mountain tops as pastures for cattle since approximately thousand years (Eggers 1964) also influences the growth of trees by grazing. In case of the silver fir, it must be assumed, that the trees are in a fringe of their growing potential (Issler 1942). It might even be possible, that it didn’t grow in the vicinity of the charcoal places, but was transported by man. This can be supported by the fact, that nine out of the thirteen charcoal pieces found only have a diameter up to 2 cm. The material of silver fir in this category is used mainly as a cover of the kiln, before the earth roof is put on. So it is needed in the process and it is easy to transport. The big share in this category in the samples collected below 1000 m a.s.l. can also be explained in that way. By comparing the results of the charcoal analyses and the present vegetation it is noticeable, that spruce is not existing in the samples. The use of the forest as fuel wood for the charcoal burning process must therefore have taken place before the introduction of the species to the Markstein area. This was during the general enforcement of the forest law in the Vosges (Code Forestier), which passed at 1827 AD (Garnier 2004). So the period of time the charcoal burning took place lies most probably before 1827 AD. Taking into account the radiocarbon dating, a time

Summary Looking at the results of the charcoal analyses it can be deduced, that a mixed deciduous coniferous forest with beech and silver fir domination was covering the elevations between 700 m and 1000 m a.s.l. of the Markstein massif. Maple was a constant companion of the two dominating species. Above 1100 m a.s.l. silver fir was much less frequent, and maybe didn’t even grew there. Sorbus was more abundant and the beech forest consists of smaller trees than in the lower regions. The results of the study correspond with the expectations of vegetation science and forest site ecology for the natural distribution of trees species in the Vosges. The distribution of the charcoal assemblages could directly be linked to the elevation above sea level, a mayor ecological factor in mountainous areas.

Acknowledgments

The research was funded through a scholarship of the German Research Foundation (DFG). The author is grateful to Thomas Ludemann for his support during the whole study and to Maik Berendes for his helpful comments on the manuscript.

References Balmier, É., Roess, D., 2002. Scènes de tranchées dans les Vosges. Strasbourg : Editions du Rhin. Boudot, J.P., Bruckert, S., Souchier, B., 1981. Végétation et sols climax sur le Grauwackes de la série du Markstein (Hautes-Vosges). Annales des Sciences Forestières, 38, 87-106. Carbiener, R., 1966. La végétation des Hautes Vosges dans ses rapports avec les climats locaux, les sols et la géomorphologie. Faculté des Sciences de Paris. Davasse, B., 2000. Forêts Charbonniers et Paysans dans les Pyrénées de l´Est du moyen Âge à nos Jours. Toulouse: Geode. Eggers, H., 1964. Schwarzwald und Vogesen. Ein vergleichender Überblick. Braunschweig: Georg Westermann Verlag. Garnier, E., 2004. Terre de Conquêtes. La forêt vosgienne sous l´Ancien Régime. Fayard : Librairie Arthème. Hillebrecht, M.L., 1982. Die Relikte der Holzkohlewirtschaft als Indikatoren für Wald-benutzung und Waldentwicklung. Untersuchungen an Beispielen aus Südniedersachsen. Göttingen: Göttinger Geographische Abhandlungen 79. Issler, E., 1942. Vegetationskunde der Vogesen. Jena: Gustav

204

A contribution to the forest history of the Markstein area in the southern Vosges Fischer Verlag. Ludemann, T., 2003. Large-scale reconstruction of ancient forest vegetation by anthracology - a contribution from the Black Forest. Phytocoenologia, 33, 645-666. Ludemann, T., Michiels, H.G., Nölken, W., 2005. Spatial patterns of past wood exploitation, natural wood supply and growth conditions - indications of the natural tree species distribution by anthracological studies of charcoal burning remains. European Journal of Forest ResearchDOI 10.1007/s10342-004-0049-z. Metailie, J.P., 1993. La Forêt Charbonnée. Histoire des forêts et impact de la métallurgie dans le Pyrénées ariégeoiises au cours des deux derniers millénaires. In: C. Beck, R. Delort, eds. Pour une Histoire de l´Environnement. Paris : CNRS, 251-258. Montanari, C., Prono, P., Scipioni, S., 2000. The Study of Charcoal-burning Sites in the Apennine Mountains of Liguria (NW Italy) as a Tool for Forest History. In: M. Agnoletti, S. Anderson, eds. Methods and approaches in forest history. IUFRO research series, 3, 79-91. Nelle, O., 2003. Woodland history of the last 500 years revealed by anthracological studies of charcoal kiln sites in the Bavarian Forest, Germany. Phytocoenologia, 33, 667682. Nölken, W., 2003. Holzkohleanalytische Untersuchungen zur Waldgeschichte der Vogesen im Tal von Miellin. Freiburger Universitätsblätter, 160 (2), 111-118. Reklip / Trinationale Arbeitsgemeinschaft Regio-Klima-Projekt, 1995. Klima Atlas Oberrhein Mitte-Süd / Atlas Climatique du Fossé Rhénan Méridional. Atlas und Textband. Zürich: Hochschulverlag. Schweingruber, F.H., 1982. Mikroskopische Holzanatomie. Formenspektrum mitteleuropäische Stamm- und Zweighölzern zur Bestimmung von rezentem und subfossilen Material. Birmensdorf: Eidgenössische Anstalt für das forstliche Versuchswesen. Walter, R., 1992. Geologie von Mitteleuropa. Stuttgart: E. Schweizerbart´sche Verlagsbuchhandlung,. Wullschleger, E., 1979. Über frühere Waldnutzungen. Ein Beitrag zur aargauischen Forstgeschichte. Dargestellt am Beispiel des Gemeindewaldes Schafisheim und Teilen des Gemeindewaldes Gränichen. Birmensdorf: Eidgenössische Anstalt für das forstliche Versuchswesen.

205

Il controllo delle alte temperature e l’inizio della metallurgia nel Vicino Oriente A. Palmieri

CNR - Istituto per le tecnologie applicate ai Beni Culturali Abstract Through the development of new technologies enhanced by fire control, the first artificial materials, such as pottery and metals, were created by means of furnaces and specialized implements. The metal smelting process is here analysed in its chronological progress as regards to minerals, typology of crucibles, and importance of slags. The high temperature control was a difficult conquest, acquired by forcing air into the crucibles or furnaces, by means of leather blowing apparatuses. The archaeological evidence of early metallurgy is summarized in a short overview of the eastern Anatolian and Near Eastern Chalcolithic sites and a description of the main ore sources in the area.

Introduzione Il fuoco ha giocato un ruolo fondamentale nello sviluppo delle civiltà, non solo nella sua funzione di riscaldamento, illuminazione e cottura di cibi, ma soprattutto permettendo la trasformazione di materiali naturali in prodotti complessi, scoperti, ideati e realizzati dall’uomo, come la ceramica e i metalli. Infatti, l’invenzione della ceramica (alla fine dell’VIII millennio nel Vicino Oriente) e, più tardi, l’estrazione e la fusione del metallo per formare oggetti (alla fine del VI millennio, sempre nel Vicino Oriente), hanno segnato l’acquisizione di tecnologie complesse che hanno richiesto lunghe fasi di sperimentazione. Specie per l’estrazione del metallo dai minerali, non deve essere stato né immediato né banale comprendere i vari passaggi di un processo di fusione. Dominio del fuoco, identificazione dei minerali, scelta dei combustibili, fusioni in crogioli o forni, uso di mantici e tuyères, invenzione delle diverse tipologie di forme di fusione, sono alcuni tra i più evidenti aspetti che articolano il complesso fenomeno della scoperta del metallo e che, nella documentazione archeologica, hanno indicato i gradi di acquisizioni tecnologiche fondamentali che hanno portato al completo dominio della metallurgia. Partendo da un fuoco di legna che raggiunge mediamente la temperatura di 750 °C, si è potuta gradualmente raggiungere e superare la temperatura di fusione del rame, di 1083 °C. E’ soprattutto il controllo delle alte temperature che ha richiesto la formazione di officine specializzate, che operavano per ridurre i minerali di rame. Dal Calcolitico al BAII, queste officine si trovavano all’interno dell’insediamento. Successivamente, le tracce materiali legate alla fusione del minerale non sono più presenti nei siti archeologici, dove invece sono presenti quelle della trasformazione in oggetti finiti del rame, venuto probabilmente già in forma di metallo dai centri minerari. Questa differenza deriva probabilmente dalla crescita del fabbisogno di metallo, in relazione alla sempre maggiore richiesta di oggetti di prestigio, armi e ornamenti, che rendeva più vantaggiosa la lavorazione dei minerali in montagna e il trasporto nell’insediamento del solo rame estratto, libero dal peso delle scorie e della ganga minerale.

Nell’ambito del Vicino Oriente, le regioni del sud-est anatolico hanno avuto un ruolo primario per quanto riguarda l’origine della metallurgia, grazie alla concomitanza di particolari condizioni: da un lato, la presenza di numerosi e articolati giacimenti minerari, dall’altro lo sviluppo precoce di società gerarchizzate, simili a quelle, contemporanee, dei primi centri urbani della Mesopotamia meridionale.

Fig. 1. Minerali metalliferi dell’Anatolia.

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

A. Palmieri

Fig. 2. Estrazione del crogiolo da un forno sperimentale.

in epoche antiche. Anche le scorie ritrovate non hanno infatti potuto essere datate. In Anatolia nord-orientale, fino al Caucaso, sono state localizzate molte altre miniere di rame, per esempio a Kure, dove sono presenti calcopirite, pirite, e bornite e a Murgul, nell’importante distretto minerario di Artvin, dove si trovano calcopirite e mineralizzazioni secondarie di bornite, calcocite, e covellite. Anche nella catena del Tauro, che si estende per una parte importante del Sud-Est dell’Anatolia e costituisce una barriera tra la linea di costa Sud anatolica e l’interno, si trovano formazioni metamorfiche di tipo ofiolitico che contengono diversi affioramenti minerari, tra cui recentemente indagati quelli del Bolkardağ (Yener 1993). L’individuazione dei depositi di minerali metalliferi non deve essere stata un problema per i ricognitori preistorici, che meglio di quelli attuali conoscevano le rocce disponibili nel loro territorio, e le loro proprietà ai fini della scheggiatura o della levigatura. Il colore deve essere stato fin dalle origini un elemento importante nella ricerca dei minerali di rame. Questi sono stati infatti utilizzati per lungo tempo come coloranti o come rocce colorate da cui trarre oggetti d’ornamento. La malachite fornisce una polvere colorante verde intenso, ben nota nell’antichità, sin dal Neolitico. La calcopirite ha un colore giallo ottone, la cuprite è rossa, l’azzurrite e la crisocolla hanno vari azzurri e la bornite è iridescente. Tutti colori facilmente riconoscibili anche

Le principali miniere in Anatolia orientale In Anatolia sud-orientale, vicino alle sorgenti del Tigri, troviamo uno dei più ricchi e meglio conosciuti distretti minerari del Vicino Oriente, collegato da rotte commerciali con il centro e il sud dell’altopiano. Si tratta del deposito massivo di Ergani Maden, nella regione di Diyarbakir, il cui nucleo è costituito da solfuri di rame e ferro. Il deposito è stato visitato dall’uomo già dal Neolitico, come dimostra la presenza, in siti di questa età nella regione, come Çayönü, di oggetti e ornamenti fabbricati in malachite, azzurrite, minerali misti, rame nativo, calcopirite e pirite, usati come rocce decorative. Le vene di minerali ossidati e polimetallici ora visibili nei dintorni e in superficie erano quindi facilmente accessibili all’inizio dell’Olocene. Queste categorie di minerali (fahlers), insieme al rame nativo, sono state probabilmente le prime identificate e usate dall’uomo perché situate nella parte superficiale della miniera (gossan e zona di arricchimento). In queste zone superficiali, gli agenti atmosferici hanno ossidato i minerali primari, e le acque di risalita dalle zone profonde hanno trascinato con sé i minerali solubili, convogliandoli verso l’alto e lisciviandoli fuori. Il complesso minerario di Ergani è considerato importante per la sua posizione, al centro di una regione ricca di importanti evidenze preistoriche. Tuttavia, non vi sono stati, sul posto, ritrovamenti di materiali che hanno confermato l’uso di questi giacimenti 208

Il controllo delle alte temperature e l’inizio della metallurgia nel Vicino Oriente

Fig. 3. Fasi di metallurgia sperimentale.

a distanza da un antico prospettore. Anche il sapore deve essere stato un elemento utile nella definizione dei minerali, applicato peraltro ancora dai geologi attuali. Agricola (Agricola 1950) nel XIV secolo afferma che i prospettori dei suoi tempi riconoscevano 6 differenti sapori relativamente ad altrettante categorie di minerali. L’odore agliaceo è, per esempio, notoriamente caratteristico dell’arsenico. Infine, anche la vegetazione può essere legata in modo caratteristico a certi depositi (p.es. il ginepro per il rame, l’Equisetum sp. può essere un indicatore di suoli connessi allo stagno), e la sua distribuzione nel paesaggio può essere stata una guida preliminare nella prospezione.

fase Halaf si trovano per la prima volta manufatti in piombo (de Jesus 1980). Lo sviluppo della metallurgia nel Neolitico e nella prima parte del Calcolitico è molto lento, fino al primo quarto del V millennio. La più antica documentazione di metallo fuso viene dal livello XVI di Mersin-Yumuktepe (4900 BC.), dove sono anche precocemente documentati altri aspetti tecnici di fusione, come l’uso di piombo e argento. Si tratta tuttavia di oggetti non preziosi né nella funzione (asce e spilloni) né nella tecnica. Nel Tardo Calcolitico (4300 B.C.) abbiamo evidenze di una metallurgia che cresce, e a Tell Shemshara (Mortensen 1970) e a Tell es-Sawwan (Al-A’dami 1968), nel corso del V millennio, si ritrovano frammenti di coltelli di rame. La fonte mineraria di questo rame orientale è stata messa in relazione con la via dell’ossidiana e quindi con miniere est-anatoliche. Dal Calcolitico (4300-3500 B.C) comincia una fase cruciale per lo sviluppo della metallurgia in Anatolia, con lo sfruttamento dei grandi depositi massivi di Ergani, Kure, Murgul e degli altri più piccoli, anche nelle aree nord-anatoliche. Per quanto riguarda le fonti del Tauro, nei livelli XI-IX di Mersin (Bronzo Antico II-III: 2800-1800 a.C.) un gran numero di altri oggetti si aggiungono alle asce e spilloni: sigilli, coltelli, pugnali e punte di lancia. I buchi dei rivetti sui manici delle armi suggeriscono che una perfetta immanicatura era necessaria per questi speciali strumenti, mentre la costolatura mediana testimonia una nuova tec-

La documentazione archeologica In Anatolia, la prima testimonianza di trasformazione del rame in oggetti si trova nei siti protoneolitici e neolitici di Aşıklı höyük (Esin 1999) e Çayönü (Muhly 1989), con punteruoli, perle e aghi in rame nativo, lavorati a freddo. A Çatal höyük (6400-5880 B.C.) è stata trovata una “scoria” e numerosi piccoli oggetti decorativi di rame, ma non è certo che vi siano state operazioni di fusione (de Jesus 1972). Anche in Siria, a Tell Ramad I (6500 B.C.), sono documentate perle di rame, e ugualmente in Iran, ad Ali Kosh (Renfrew in Mortensen 1970), e in Iraq, a Tell Sotto (pre- Hassuna), Tell es-Sawwan (livelli inferiori) e Yarim tepe (Hassuna) (Munchaijev, Merpert 1973), dove nella 209

A. Palmieri nica di produzione con una forma bivalve, come anche le teste poliedriche degli spilloni e degli scalpelli. Un sigillo di rame del livello X, decorato con un motivo a spirale, è un incomparabile campione datato in Egitto alla tredicesima dinastia (Garstang 1953, 215). Molti sono i siti di queste fasi in Anatolia Sud-orientale che hanno restituito prove di una metallurgia elaborata come Hassek huyuk, Norsun tepe, Tepecik ed altri. Fra questi, Arslantepe, con la sua sequenza cronologica ininterrotta dal Calcolitico fino ai livelli romani e con i suoi eccezionali ritrovamenti di oggetti in metallo e di strumenti legati alla fusione, ha fornito la possibilità di seguire i vari aspetti della storia della metallurgia (Palmieri et al. 1993). In nord Mesopotamia, troviamo a Tell Brak una grande varietà di oggetti in metallo (Mallowan 1947). Il più antico bronzo conosciuto, uno spillone, viene da Tepe Gaura X (3200 B.C.) (Speiser 1935). Sempre nel IV millennio a Niniveh e ad Arpachiya abbiamo ritrovamenti di oggetti di rame. Il problema per la Mesopotamia è soprattutto le miniere di approvvigionamento visto che la regione ne è sprovvista. Due ipotesi per confronti possibili sono l’Iran e per il sud Mesopotamia le culture del Golfo Arabico. A metà del IV millennio troviamo la fusione del rame a Timna - fornaci e luoghi di preparazione dei minerali, martelli, mortai, strumenti in selce e ceramica del Calcolitico (in Palestina 4000-3300 B.C.) (Rothenberg 1978). Sempre in Palestina Abu Matar ha dato prove di metallurgia verso la metà del 4° millennio con minerali, fornaci, crogioli e oggetti in metallo (Perrot 1957). Nel Levante i più sorprendenti ritrovamenti di metallo vengono dal ripostiglio di Nahal Mishmar (Sud Palestina), con oltre 200 teste di mazza, scettri, corone. Tutti lavorati con eleganti motivi decorativi ad animali. La data ottenuta dalle stuoie che contenevano gli oggetti è intorno alla fine del IV millennio. Questi oggetti rimangono senza provenienza nonostante le varie analisi effettuate. Appare con questi oggetti molto elaborati la testimonianza, forse, della prima fusione a cera perduta del Vicino Oriente (Bar-Adon 1971). Finora questa particolare tecnica era conosciuta a Cipro e nel vicino Oriente nell’età del Bronzo ed è da Cipro che questo procedimento è giunto in Sardegna. Per la stessa natura della lavorazione è estremamente difficile se non impossibile ritrovare delle tracce riconoscibili. Questo procedimento è denominato” a cera persa” in quanto la figura che si voleva realizzare veniva prima modellata in cera in tutti i minimi particolari, compresi i “canali di sfiato” e poi veniva ricoperto di argilla cruda, che veniva poi cotta, eliminando la cera. Quando veniva versato il bronzo, questo scorreva a riempire tutti gli spazi lasciati dalla cera. Si doveva poi attendere il raffreddamento della forma, per romperla ed eliminarla.

sempre di mettersi nelle condizioni che venivano suggerite dai ritrovamenti archeologici. In queste operazioni venivano controllate le temperature mediante termocoppie poste all’interno del crogiolo o in vari punti delle pareti delle fornaci. Venivano inoltre analizzati i fumi per verificare le condizioni ossidanti o riducenti all’interno della camera di combustione. Infine, veniva tenuto esatto conto del peso iniziale del minerale e finale del prodotto, per calcolare la temperatura e i tempi necessari per ottenere un rendimento ottimale. Al di là della strumentazione di controllo, moderna e capace di ridurre i tempi di sperimentazione, gli altri passi delle operazioni rispettavano probabilmente da vicino gli esperimenti condotti, attraverso maggiori difficoltà e con tempi lunghissimi, dagli artigiani calcolatici. L’aspetto cruciale dell’intera operazione è la temperatura, e quindi il combustibile. Nella sperimentazione è stato utilizzato il carbone di legna. Già conosciuto in antico per il suo maggiore potere calorico, e usato per i forni per la ceramica, il carbone potrebbe essere stato un passaggio obbligato per arrivare alla fusione dei minerali metalliferi. Mancano tuttavia studi in proposito per le aree della Transcaucasia e del sud-est anatolico. E’ difficile stabilire, nell’ambito dei frammenti di legno carbonizzato ritrovati negli scavi, quali legni fossero usati come combustibile per le operazioni di fusione. La maggior parte dell’Anatolia era coperta da alberi nei tempi preistorici, costituendo una grande riserva di combustibile; attualmente gli alberi diradati lasciano una traccia nell’altopiano con un visibile disboscamento diffuso a macchia di leopardo. Il processo essenziale che riguarda la produzione di carbone consiste nel bruciare tutte le sostanze resinose del legno e seccarlo senza bruciare i materiali fibrosi. Dai ritrovamenti archeologici è difficile distinguere come si preparava il carbone ma trattandosi di un processo abbastanza semplice e con poche varianti, si può dedurre che non fosse diverso dai periodi più recenti. Il carbone può essere fatto con la maggior parte dei tipi di legno. Ginepro e conifere erano comunemente usati nei climi freddi, mimosa e tamarisco nelle regioni calde. Un albero che si trova ancora abbondantemente sull’altopiano del sud-est anatolico e viene raccolto come combustibile per l’inverno è la quercia. Il legno veniva accatastato in una depressione e coperto completamente con zolle erbose. Prima della copertura completa, veniva acceso e lasciato bruciare lentamente senza fiamma. Un altro possibile tipo di combustibile è rappresentato dagli escrementi di erbivori seccati al sole, tradizione che ancora oggi si riscontra nei piccoli villaggi dell’Est Anatolia e si manifesta con la presenza di torte di letame attaccate per essiccare al muro di casa e poi accatastate in pile nei cortili. Il suo basso costo e il suo alto potere calorico potrebbero averlo reso uno strumento essenziale delle operazioni di fusione, ma la sua combustione non lascia tracce archeologicamente rilevabili. Altro elemento cruciale è la costruzione dei forni. Di fronte alla grande abbondanza di oggetti di rame nell’Anatolia preistorica, poche sono le testimonianze archeologiche di fornaci. Il sito di Kültepe nel

La tecnologia Per comprendere meglio le varie fasi della fusione sono state effettuate prove sperimentali con varianti sui tipi di minerali, usando forni e crogioli diversi, ma cercando 210

Il controllo delle alte temperature e l’inizio della metallurgia nel Vicino Oriente riodo delle Colonie Assire di Kultepe. Forme a due valve sono state trovate a Troia e a Thermi II e IV. Forme per lingotti sono conosciute da Troia IIg e Thermi I, un altro esempio viene da Pulur X (EBII). Le poche forme di lingotti trovate in Anatolia suggeriscono che i lingotti erano fabbricati al di fuori degli insediamenti. I minerali usati per l’estrazione del rame si riscontrano in una precisa sequenza cronologica che si svolge sulla base del fabbisogno del metallo e delle caratteristiche fisico chimiche del minerale. Nei periodi più antichi si usa il rame nativo fino a suo completo esaurimento. Poi si passa agli ossidi e carbonati che sono caratterizzati da una semplice fusione, ad Arslantepe si incontrano dal Calcolitico i minerali polimetallici costituiti da un insieme di minerali (carbonati, ossidi, solfuri e rame nativo) intimamente legati da una ganga silicea che permette una fusione senza aggiunta di materiali fondenti. Infine, quando anche queste fonti si esauriscono e la richiesta del metallo aumenta, si cerca di fondere la calcopirite che rappresenta un nuovo passo tecnologico: è necessario, infatti, un arrostimento preventivo del minerale per eliminare lo solfo, e l’aggiunta di un materiale fondente per separare il ferro dal rame e ottenere una scoria di tipo fayalitico che cattura il ferro e libera il rame, operazione non banale che avrà avuto bisogno di un lungo periodo di sperimentazione. Scorie isolate non costituiscono una sequenza operativa di fusione, perché possono essere state trasportate o essere il risultato di operazioni del fabbro e non di fusione da minerali. Il riconoscimento delle scorie di fusione nei contesti archeologici, invece, è cruciale per gli studi sulla metallurgia di un sito perché permette, a volte, di identificare le operazioni effettuate nel processo di fusione. Ancora vi è molto da fare nell’interpretare la composizione delle scorie. Anche perché gli antichi fonditori, usando metodi empirici avevano problemi a mantenere l’omogeneità durante l’intero processo, ma a volte si possono riconoscere i minerali usati (scorie sinterizzate) e le variazioni composizionali al loro interno. Esistono grandi varietà di scorie, e ciò dipende dai minerali e dalla tecnologia disponibile. Il rame nelle scorie da fornace è basso mentre è alto nelle scorie da crogiolo perché dovuto alla reazione del rame con la ceramica (20-28%). Il rame fonde a 1083°, in condizioni ottimali i minerali di rame possono essere ridotti a rame metallico a circa 800°C. Esperimenti moderni suggeriscono che la fusione dei minerali in antico può essere avvenuta a temperature al di sotto di 1150°C. Le scorie di fayalite del Calcolitico di Timna hanno avuto un punto di fusione fra i 1150-1250°C. I fondenti sono materiali che si aggiungono per ottenere o facilitare le operazioni di fusione. Quando è immesso volontariamente, il tipo di fondente varia secondo il minerale usato. Generalmente i solfuri hanno eccesso di ferro e va aggiunta silice, mentre gli ossidi di rame hanno, al contrario, eccesso di silice e si aggiungono ossidi di ferro. In ogni caso, la scoria liquida che si forma è fayalite (2FeO.SiO2). La mancanza di silice in una fusione potrà causare la combinazione con il rame, formando un materiale pesante, composto essenzialmente

Bronzo Medio è uno dei rari esempi di documentazione, ma si riferisce a periodi relativamente recenti. La fornace è costruita sul pendio di una collina in modo da utilizzare l’aria dei venti dominanti e far scendere le scorie con facilità. Evidenze di fusioni nel Calcolitico e nel Bronzo Antico vengono dal sito di Norşuntepe, dove scorie e minerali sono state ritrovate nell’insediamento, senza tuttavia che vi fossero localizzate fornaci. Fuori dall’Anatolia, fornaci per fusione sono state trovate in vari contesti. Famose sono quelle di Timna, nel Sinai meridionale, e di Tell Iblis (de Jesus 1980). Altri importanti siti connessi alla metallurgia preistorica sono Feinan, in Giordania, che ha restituito fornaci, minerali e materiali archeologici del Bronzo Antico (Hauptmann et al. 1992), e il calcolitico Abu Matar, nel Levante meridionale, con crogiuoli, scorie e oggetti di pregiata manifattura. Tuyères e mantici sono elementi associati alle fornaci per la fusione dei minerali o del metallo. L’uso dei mantici presuppone l’esistenza di tuyères di ceramica. In realtà, è molto più frequente il ritrovamento di tuyères che di mantici, questi ultimi fatti probabilmente con materiale deperibile, come legno e pelle. In siti molto conosciuti per la metallurgia, come a Norşuntepe, a Cipro, a Gezer (Macalister 1912); non sono stati trovati mantici, a Timna, sono state trovate tuyères, in Anatolia a Beycesultan VIII e Poliochni rosso (EBI), probabili frammenti di tuyères vengono da Troia EB III. I più antichi mantici anatolici vengono da Kultepe MBI ma la fusione era un’attività spesso praticata vicino alle miniere lontano dagli insediamenti tradizionali (de Jesus 1980). L’immissione di aria tramite canne con punta in ceramica è documentata in Egitto nel 2800 B.C., non vi sono tracce altrove nel Vicino Oriente. La fusione Ad Arslantepe si notano differenze sostanziali nell’uso dei minerali e nella tipologia dei crogioli. Piccoli e spessi, a forma di ciotola, nei periodi più antichi, evolvono successivamente in forme sub-cilindriche più capienti, per arrivare nel BA III a forme piccole, piatte, con beccuccio e prese ad aletta, utilizzate per la fusione diretta del rame. In questo periodo il rame usato, presumibilmente, era stato estratto vicino alle miniere, evitando così nel trasporto il soprapeso della ganga minerale (Palmieri et al. 2001). Le prime forme di fusione usate in antico potevano essere spaccature in rocce o depressioni in argilla in cui veniva versato il rame fuso. In seguito furono usate forme aperte per asce, scalpelli e altri semplici strumenti. I più antichi strumenti in Anatolia, probabilmente fatti con questo metodo, sono le asce piatte e gli scalpelli tardo calcolitici da Mersin (4900 B.C.). Ben nota è la testa di mazza da Can Hasan (4750 B.C.) fatta di rame quasi puro, sferica con un foro al centro. Le fusioni a cera perduta sono più tarde e vi sono solo evidenze indirette, una forma di ascia viene dall’EBI di Poliochni. Forme di asce con foro sono rare, una viene dal EBAIIIA di Norsuntepe e un’altra dal pe211

A. Palmieri da rame e ferro intimamente legati (metallina). È possibile recuperare gocce di rame in una fornace riducente facendo freddare le scorie e rompendole. Si può ripetere questa operazione diverse volte.

Ceramic in the Society, Proceedings of the 6° European Meeting on Ancient Ceramics, Fribourg, Switzerland. Perrot, J., 1957. Les Fouilles at Tell Abou Matar, près de Beersheba. Syria, 34: 1-38. Rothemberg, B., 1978. Excavations at Timna Site 39 a Chalcolithic copper site and Furnace and its metallurgy. Chalcholitic Copper Smelting, Monograph n. 1. London: Publication of the Institute for Archaeometallurgical Studies, 1-15. Speiser, E.A., 1935. Excavations at Tepe Gawra. I Philadelphia Yener, K.A., Vandiver, P.B., 1993. Tin Processing at Göltepe, an Early Bronze Age Site in Anatolia, American Journal of Archaeology 97: 207-38.

Conclusioni In Anatolia lo storia della metallurgia è stata ed è una ricerca lunga e appassionante dove, di volta in volta, si aggiunge un nuovo tassello che getta luce su apetti tecnologici e culturali ancora non conosciuti. Diventa un data base per confronti e relazioni fra siti, fornendo spunti interpretativi per una visione più completa sincronica e diacronica della storia. In conclusione, abbiamo visto che il dominio delle alte temperature ha portato alla scoperta di nuovi materiali complessi, alla costruzione di crogioli, forni e strumenti idonei alla fusione, che si sono dimostrati elementi indispensabili nella ricostruzione dello sviluppo di una civiltà.

Bibliografia agricola, 1950. De re metallica (tradotto dalla prima edizione latina del 1556. New York: Hoover H.C., Hoover L.H). al - a’Dami, k., 1968. Excavation at Tell es-Sawwan. Sumer, 24: 57-60. bar-aDon, p., 1971. The Cave of the Treasure: The Finds from the Caves in Nahal Mishmar. Jerusalem (in Ebraico). De Jesus, p. s.,1972. Prehistoric Metallurgy - Another Look. Anadolu/Anatolia, 16: 129-140. De Jesus, p. s., 1980. The Development of Prehistoric Mining and Metallurgy in Anatolia. B.A.R. International Series. 74. London: BAR Publishing esin, u., 1999. Copper objects form the pre-pottery Neolithic site of Aşıklı. In: A. Hauptmann, E. Pernicka, T. Rehren, U. Yalcın, eds. Der Anschnitt 9. The Beginnings of Metallurgy. Bochum, 22-30. Garstang, J., 1953. Prehistoric Mersin. Oxford. Hauptmann, A., Begemann, F., Heitkemper, E., Pernicka, E., Schmitt-Strecker, S. 1992. Early copper produced at Feinan, Wadi Araba, Jordan: the composition of ores and copper. Archaeomaterials 6: 1-33. Mallowan, M. ‘Excavations at Brak and Chagar Bazar. Iraq, 9: 1-266. Mortensen, P, 1970. Tell Shimshara. Copenaghen. Muhly, J. D., 1989. Çayönü Tepesi and the Beginnings of Metallurgy in the Ancient World. In: A. Hauptmann, E. Pernicka, G. Wagner eds., Old World Archaeometallurgy, Bochum, 1-11. Munchajev, R.M., Merpert, N.I., 1973. Excavations at Yarim Tepe 1972. Sumer 29: 3-16. Palmieri A.M., Sertok K., Chernykh E., 1993. From Arslantepe Metalwork to Arsenical Copper Thecnology in Eastern Anatolia. In: M. Frangipane, A. Hauptmann, M. Liverani, P. Matthiae, eds. Between the rivers and over the mountains - Archaeologica Anatolica et Mesopotamica Alba Palmieri Dedicata, Mellink. Palmieri, A.M., Morbidelli, P. 2001. Archaeometric study on crucibles from Arslantepe, Turkey (IV-II mill. BC).In:

212

Food Processing in the Levant during the Middle Bronze Age. Fire installations cooking pots and grinding tools at Tell Mardikh-Ebla (Syria)-Two Case Studies L. Peyronel, G. Spreafico Università - IULM - Milano E-mail: [email protected]

Abstract The excavations carried out since 1964 at Ebla-Tell Mardikh (North Inner Syria), have brought to light large sectors of the Old Syrian town dating to the Middle Bronze Age II (c. 1800-1600 BC). A large quarter of houses located in the Area B East, at the foot of the Acropolis in the South-Western sector of the Lower Town , and the Western Fort, built on the top of the rampart defending the town during the Middle Bronze Age, provided a large evidence of fire installations and contexts related to the food processing (cooking pots, grinding tools, and bio-archaeological remains). Within a more wide project of environmental cross-disciplinary researches, including archaeometry, bio-archaeology, palaeogeography, sedimentological and geological analyses, we will focus here the attention on the typology of fireplaces for cooking and/or heating, proposing a comparison between a public defensive complex and a group of private dwellings.

Introduction1 The excavations carried out at Ebla by the Italian Expedition since 1964 have brought to light large sectors of the Old Syrian town dating to the Middle Bronze Age II (c. 1800-1600 BC), allowing to reconstruct the general outline of the outer fortification, the lower town and the acropolis (Matthiae 1989, p. 135-186; 1991; Matthiae et al. 1995, p. 164-178) (Fig. 1). The city was destroyed and sacked at the end of the MB IIB (c. 1650-1600 BC), probably by the Hittites2, and a thick destruction layer with ashes and burnt bricks from the collapsed walls was found in every buildings of this last phase, as a sound proof of the fire which spread during the siege. Therefore, a large amount of materials (pottery vessels, objects, organic materials, etc.) was sealed in situ over the floors and in the layers immediately above them, allowing to investigate the relation between architecture and the original activities performed in. A new project of environmental crossdisciplinary researches (archaeometry, bioarchaeology, palaeogeography, sedimentological and geological analyses) started in 2001 in connection with renewed investigations in Area B East, where a large quarter of well-preserved houses was identified and partially excavated at the beginning of the Seventies (for the first results see Peyronel in press b). At the same time, microstratigraphical test-excavations were conducted at different suitable contexts, such as the large midden on the eastern rampart, dating to the beginning of the Middle Bronze I (c. 2000-1900 BC) (Peyronel in press c) and a complete reanalysis of bio-archaeological data from public buildings previously excavated in connection with a distributive and functional analysis of archaeological materials was planned with the main goals of an identification of activity areas and handicraft production reconstruction3. Within the context of a more wide study of fire installations at Tell Mardikh-Ebla during the Early and Middle Bronze Ages we will present here the evidence from the large Western Fort and from the private houses in Area B, focusing the attention on the typology of fireplaces for

food cooking and/or heating, and stressing similarities or differences between a public defensive complex and the private buildings, both dating at the time of the final destruction of the Old Syrian town. The Domestic Quarter in the Southern Lower Town (Area B East) Different typologies of fire installations were found in the large quarter of private dwellings (Area B East) located in the South-Western sector of the Lower Town at the foot of the Acropolis, between Sanctuary B2 to the West and the Southern Palace to the East (Fig. 2). The domestic quarter is divided in various insulae by perpendicular streets: L.2110 South-North, L.8651 and L.8472 East-West. The Northern Unit, to the North of L.8651, and part of the Central Unit, between L.8651 and L.8472, were brought to light in 70’s. The Central and Southern Units, the latter located to the South of L.8472, was brought to light only in 2001, with the purpose to carry out a detailed study of domestic architecture, activity areas and archaeological materials in a multidisciplinary approach. Each unit includes several dwellings characterized by a very similar arrangement of spaces, but at the same time by a great variety in plan. A typical house consists of a small vestibule, where is often located a staircase leading to the roof, a large central courtyard, where most of the domestic activities took place, and several subsidiary rooms with different functions, such as storerooms, kitchens, and working-rooms4. In particular, each house has one or more rooms devoted to food processing and cooking, as testified by fire installations, cooking pots, and grinding tools found in situ in the destruction level that sealed the floors, firmly dated to the end of the Middle Bronze IIB (c. 1650-1600 BC). Three different types of fire installation can be singled out: the horse-shoe shaped hearth (tannur), the mud-brick or stone platform, and the bread oven (tabun).

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

L. Peyronel, G. Spreafico

Fig. 1. Topographic Plan of Tell Mardikh-Ebla.

employed as fuel for tannur. First branches were probably used to ignite the dung and to raise the temperature, and then hot coals from the dung maintained the heat for a long time. We can suppose the cooking pot was placed over the tannur only after the fire burnt out, so that the coals allowed a slow and uniform cooking. Several vessels of kitchen ware used for cooking food over the tannur were often found near the horse-shoe shaped hearths (Fig. 4). The most common shape is a globular cooking pot with short neck and everted rim, characterized by a thick rim and a round base. The pot’s size agrees with its placement over the hearth, while the thin base should have favoured heat passage as well as the thick rim should have facilitated transport. Cooking happened probably also in the large flat bowls of different shapes recovered at Ebla, as suggested by their size and the traces of fire on the surface of the pottery7. This kind of platters strongly resembles vessels employed by Bedouins for cooking and baking over simple hearths. In Palestine villages an iron bowl (sag), about 34-50 cm in diameter, was used to bake bread over portable hearths (Dalman 1935, p. 3973, Figs. 9, 12.1), while women of the village of al-Hiba

The horse-shoe shaped hearths are composed by a curved terracotta wall partially embedded in the floor. The curved wall encircled the area where fuel was burnt and it was used to avoid dispersion of charcoals and to improve combustion. In some cases the tannur had a second external wall intended to maintain the heat5 (Fig. 3). The hearths are about 30-35 cm in external diameter and 40 cm high, with 3-4 cm thick walls. This kind of fire installation is widely attested in the ancient Near East6 and was employed until recent time by the population of the area, as attested by ethnographic studies (Dalman 1935, p. 21140; Aurenche 1981, p. 241-252). The archaeobotanical analysis carried out by Girolamo Fiorentino (Ascalone et al. in press) on materials collected from burnt layers inside the tannur T.8662, located in the courtyard L.8454 of the central unit, identified charcoals produced mainly by combustion of olive tree branches (Olea europaea) and carporemains characterized by chemical alterations typical of seeds included in dung. As testified by the observations of the modern village life in Palestine, Jordan, Lebanon and Iraq, reported by Dalman (1935, p. 18-21) and Ochsenschlager (1974), dung was usually 214

Food Processing in the Levant during the Middle Bronze Age

Fig. 2. Schematic Plan of Area B East.

215

L. Peyronel, G. Spreafico (southern Iraq) put a disk of sun-dried mud (tabag) on three clay legs over the fire for baking wheat or rice bread (Ochsenschlager 1974, p. 167). The second kind of fire installation consisted of mud-brick or stone platforms (Fig. 5), about 5-10 cm high, on which portable hearths or braziers were placed. The mud-bricks used for these installations are larger (about 53-60 x 35 cm) than those employed to build the walls (usually 38/40 x 38/40 x 15 cm), and their paste is more finished and hardened by the continuous contact with the fire. Evident traces of burning were identified on the top face of platforms, showing that activities related to heating or cooking were performed there. To the last purpose the residents of the quarter probably employed the typical horse-shoe shaped portable hearths attested by the specimens of the Western Fort and Western Palace. Many fragments of these objects were in fact found in the destruction layers that sealed the domestic quarter. However, we can also suppose other fire installations, like those described by Dalman, laid on the platforms. For example the villagers of Palestine used to put turned upside down a simple bowl on three stones over the fire, baking the bread on the back of it (Dalman 1935, p. 39-41, 45-51). Another kind of bread oven was composed by a clay bell-shaped installation with an upper

Fig. 3. Tannur T.8662.

Fig. 4. Kitchen ware from Area B East.

216

Food Processing in the Levant during the Middle Bronze Age is always heated by fire or coals and this should be very expensive for a single family, so the women go to beak their bread in turn, everyone bringing a part of fuel. At the same time the isolated placement prevents the smoke from filling the houses. Only the lower part of the oven T.8760 was found, with the cooking surface blackened by fire and the walls formed by a mixture of clay and rubble. The installation is oval in shape with a diameter of 95-100 cm and an opening to the east 40 cm wide. It probably had a vaulted cover and an upper opening like the bread ovens still widely attested in the Near East. The tabun was built in segments in modern times as well as in the past: the people make the lower segments thicker than the uppers, waiting for the first wall layer to dry before adding the next one (Ochsenslager 1974, p. 169-170). The oven is employed to bake thin, flat disks of dough, who are flattened against the inside walls above the level of the hot coals, when the fire is out. The tabun takes up almost all the room L.8757, with only a little free space in front of the door. This arrangement is useful to maintain the heat and to protect the oven from the wind.

Fig. 5. Mud-brick platform in L.1148.

opening and a clay round cover (Dalman 1935, p.76-79, Figs 12, 13). A similar oven was probably placed on the mud-brick platform of the room L.8756, characterized by a semicircular trace of burning. The brick surface inside the semicircle was glazed by an intense heat, showing that this area was used for cooking. The platforms, as well as the fixed horse-shoe shaped hearths, are spread over all the dwelling quarter, generally located in a specific room (a courtyard or a kitchen-room) where the activities of food preparation and cooking took place. In most cases we found only one tannurhearth or one platform for room, with two exceptions: the long room L.1150 in the Northern Unit, where three horse-shoe shaped hearths were placed one beside the other, and the room L.8756 in the Southern Unit, where we brought to light both a tannur and a brick-platform. In the latter case we can suppose the cooking activities were concentrated in the same room for space reasons in dwelling’s arrangement. On the contrary, L.1150 in a first building phase was probably located outside the house composed by courtyard L.1148 and rooms at the back, forming a common cooking space shared by two or more dwellings of the Northern Unit8. If the horse-shoe shaped hearths and the platforms mirror the daily cooking activity of every single family, the third kind of fire installation, the great bread oven (tabun) T.8760 (Fig. 6), represents a shared activity of more houses. The tabun was placed outside any house, in a small room (L.8757) opening on the southern E-W street L.8472, and it was probably used for making bread by all the people living in the quarter. Once again its placement remembers the customs of modern Palestine and Jordan villages, where several women use to share a single tabun located in an “oven house” (Dalman 1935, p. 76-77; McQuitty 1984, p. 261-264, Fig. 6). This arrangement was found also at the Islamic Jordan rural site of Khirbet Faris, where bread ovens inside the dwelling quarter are placed in separated rooms (McQuitty 1994, p. 68-69, Figs. 12-13). A full oven’s heating it is possible only if the installation

Fig. 6. Tabun T.8760.

217

L. Peyronel, G. Spreafico Several grinding facilities related to the food preparation and storage devices for foodstuffs were brought to light near the horse-shoe shaped hearths and the platforms. Particularly interesting is the evidence from courtyard L.8454 of the central dwelling (Ascalone et al. in press). This is the central poly-functional space of the house, where many domestic activities were performed. In the South-East area of the room, opposite the tannur, we found some basalt grinding tools on the floor: two fragments of millstones, two pestles and a tripod bowl (fig. 7). These objects were probably used to prepare the flour for the bread. First the spikelets were pounded in the basalt bowl, and then the grain, which was sieved several times to eliminate the husk, was ground to coarse flour on a saddle quern with the millstones (Renfrew 1995, p. 95). The corners of the room were employed to store foodstuffs: in the North-West corner we found in situ a large bowl full of charred caryopses of barley, while a concentration of grape fruits was in the South-West corner (see Ascalone et al. in press). In L.8454, as well as in other courtyards of the quarter, was gathered also the water used in everyday life. Rainwater was collected from the roof into vertical clay pipes, and then flowed trough horizontal stone channels into a well. The Western Fort (Fig. 8) The excavations carried out between 1995 and 1998 in Area V brought to light a huge defensive complex located on the top and inner slope of the Western rampart, namely the Western Fort (Matthiae 1997a, p. 10-12; 1998, p.

Fig. 7. Grinding tools from L.8454.

Fig. 8. Schematic Plan of Western Fort.

218

Food Processing in the Levant during the Middle Bronze Age 575-579, Figs. 12-15; 2000, p. 580-584, Figs. 10-11; 2001, p. 44-46, Fig. 8; Peyronel 2000 in press a). It was built at the beginning of the Middle Bronze II or slightly before (c. 1850/1800 BC) to control the external foot of the fortification, and it was definitively destroyed at the end of the MBA (Mardikh IIIB2), when the town was sacked probably by the Hittite under the reign of Mursili I (c. 1650/1600 BC), as revealed by the thick destruction layers with burnt bricks and ashes found over the floors of the fort. The building is 70 m long from S to N and it is composed by several quarters linked by staircases and ramps. A large upper courtyard (L.6315+L.6621) connects the different units: the South-East unit, certainly devoted to food processing, the South-West wing, possibly a sort of ‘treasury’, the western quarter with a residential function, the massive north-western tower (the so-called Fortress V), and the lower eastern sectors, with small rooms (storerooms and barracks) organized on the two sides of the ramp leading to the central upper court. The distributive analysis of the materials collected in situ over the floors and sealed by the destruction layers allowed to identify some rooms of the building devoted to processing and cooking of foodstuffs. The most important is the South-East unit, where three rooms, of which only one is well preserved, were equipped with a large amount of grinding tools for cropprocessing (Fig. 9). In L.6427 five complete basalt tripods and several pestles of different sizes (mainly of the small conical/tronco-conical type, beside two specimens of the elongated type), two querns with fragmentary handstones and large fragments of storage jars were found over the floor9. The predominance of the small pestle perfectly fits with the presence of tripods with shallow basins in the rooms, being used to crushing and pulverizing small quantities of organic materials inside them. The first processing operations were instead made with saddleshaped grinding slabs and handstones, probably also in the nearby rooms, in which some fragments of this kind of basalt equipment were found, despite the very poor state of preservation of the structures. Palaeobotanical analyses showed the presence two-rowed barley and glume wheat, together with lentil, although the organic material was scarce, because the floor lies only 20 cm under the topsoil and the archaeological deposit had partly flowed down; we cannot therefore recognize if the different cereals were stored together or separated in different containers before being crushed in the tripodic mortars and grinding slabs. The location of primary processing activities in peripheral sectors of the Old Syrian public buildings can be observed also in the Northern and Western Palaces in the Lower Town, attesting a common trend during this period. The north-western wing of the Western Palace (probably the crown prince’s residence) is composed basically of rectangular rooms devoted to food preparation in front of court L.3200. The room of the querns (L.3135) at the eastern end of the wing represents the best known and extraordinary evidence for a centralization of food

Fig. 9. L.6427 with grinding facilities in situ.

resources during the MBA in North Inner Syria (Matthiae 1985, tav. 68; 1989, p. 170 tav. 88; Dolce 1990, p. 125126, Figs. 12-13, 15; 1991, p. 330-332): it is a rectangular room furnished with a high mud-brick plastered bench running parallel but separated from the S, N, and W walls. Sixteen saddle-shaped querns and the related grinding stones were still placed on the upper top of this installation. Workers must have been seated or kneeling directly on the

Fig. 10. Pottery vessels and grinding tools from L.6516.

top of the saddle querns, using the back corridor to move around the installations, working the handstones to crush the cereals, which would fall on the cement floor and then be collected in the underground silos through a carefully plastered circular opening10. The peripheral north-eastern wing of the Northern Palace, with small square or rectangular rooms arranged on the two sides of corridor L.4036, was devoted to food processing, equipped with large circular ovens for cooking food (tabun) and andirons11, possibly linked with ceremonial banquets, which is likely that took place in this building, as suggested by a large refuse pit F.5861/F.5701 to the N of the street flanking the northern side of the palace, and by its vicinity to Ishtar’s Sacred area formed by Temple P2, Monument P3 and the Cisterns Square12. A second interesting context of the Western Fort is 219

L. Peyronel, G. Spreafico represented by Fortress V with its vestibule L.6522 and the adjacent room L.6516. The fortress has a standard rectangular plan formed by a southern (vestibule and staircase) and a northern block (six non-communicating rooms, five of them accessible from an upper storey or terrace through ladders), with the entrance located to the South, leading to the broad vestibule (L.6522). From this room it was possible to reach room L.6516 to the North, the upper floor or the terrace through a staircase located in the north-western corner, and the understairs now completely

testified by a group of basalt grinding tools still in situ: a basalt saddle-shaped was found upside down - probably because not in use - in the centre of the installation, with two handstones near it; two other grindstones were found in the corner leaning against the wall in a vertical position, and a second quern was put sidelong against the South wall. This assemblage testifies a use of the working area for crushing small quantities of dry-product, such as cereals, probably by one or two workers. Other food processing activities were performed on the low square platform at the opposite corner, where a fine basalt tripod with two elongated pestles with shaped handles were discovered. Finally, a basalt mortar was found together with the pottery along the Northern wall. The typologically differentiated grinding equipment (see Peyronel 2000, Fig. 9) allows to recognize the different steps of food processing which took place in the fortress, and their spatial allocation. We have here attested almost all the main types of grinding tools of MB Ebla, suggesting the possibility to perform a complete transformation of primary agriculture products in a single room. If we consider materials from the nearby vestibule we can therefore identify the final stage of the process, namely the cooking of the food: a hearth formed by a curved terracotta wall partially embedded in the floor associated with two cooking pots and simple ware (mainly bowls and small jars) is in fact located in the north-eastern corner of the vestibule, immediately at the right of the door connecting with L.6516 (Fig. 12). This type of fire installation is very common at Ebla during the MBA, as previously noted in the analysis of the private quarter in Area B. A large globular cooking pot with square everted rim and a another small vessel of the same morphology, both peculiar of the late MB II horizon, were found in fragments around the tannur. The food processed with grinding tools inside room L.6516 was probably cooked here by personnel related to the control of the building, so that the autonomy of the tower was assured in case of a siege14. The existence of cooking activities in the rooms located at the two sides of the ramp connecting the entrance of the

Fig. 11. L.6516 with the installation for food processing on the left.

missing through a door at the south-western corner. Room L.6516, the only room of the northern wing with a door to the vestibule, was under the control of some official, as revealed by two door-sealing with cylinder seal impressions (Peyronel 2000, Fig. 14). It was a very peculiar storeroom and working place in which a large amount of spatially and functionally diversified materials was found in situ (Fig. 10) (Ascalone – Peyronel 1999, tab. 3, Fig. 7; Peyronel 2000, p. 1359-1364, Figs. 3, 8-9): almost 20 medium or small sized vessels (jugs, small jars, bowls and cooking pots) and a basalt mortar, arranged in two rows along the northern wall; a group of basalt objects (a fine basalt tripod with two elongated pestles with shaped handles) set on a low square bench; a wooden box with stone weights and other small stone and bronze tools in a cluster near the centre of the room, grinding stones and querns related to a working area at the south-eastern corner. Kitchen ware is represented in the pottery equipment of the room by an open bowl with inturned rim, flattened base and circular burnishing, a large open bowl with burnishing inside and outside, a large dish with circular burnishing and incised decoration under the thickened horizontal rim, three hemispherical cooking pots with everted square rims13. It is quite probable that this assemblage was used for cooking food in the vestibule, where a fire installation is located (see below). A low curving clay plastered installation with one end against the East wall isolates a small working area in the SE corner of the room (Peyronel 2000, Fig. 12) (Fig. 11). It was surely devoted to food processing as

Fig. 12. Horse-shoe shaped tannur in L.6522.

220

Food Processing in the Levant during the Middle Bronze Age to study and publish the fire installations from Areas V and B. We also express our gratitude to Prof. Frances Pinnock for the useful suggestions and discussions regarding the pottery production at Ebla. L. Peyronel wrote the introduction and the § on the Western Fort; G. Spreafico wrote the § on the Domestic Quarter. 2 For the chronology of Ebla see Matthiae et al. 1995, p. 86-95; for the historical evaluation of the MB destruction cf. now Matthiae 2002, p. 571-574 and especially Matthiae in press. 3 For the textile manufacture see now Peyronel 2004. A first evaluation of some contexts of food processing and conservation was already published by Dolce (1990; 1991). 4 For a preliminary evaluation of domestic architecture at Ebla during the Middle Bronze Age II see Matthiae 1997b, and Baffi in press. For a general discussion of the houses in area B East see Peyronel in press b. and Ascalone et al. in press for a detailed description of the central dwelling in area B East. 5 This feature has a comparison in the Islamic Jordan site of ‘Aqaba/Ayla (VII-XII sec. AD), where three hearts, built against a house wall, consist of two concentric, cylindrical jars packed with silt, sand and ash (McQuitty 1994, pp.63-65, figs. 7, 9-10). McQuitty (1984, pp. 64-65) underlines that this type of oven does not occur in present, while it is attested in archaeological contexts of various periods, starting form the Bronze Age until the Hellenistic and Umayyad times. 6 For a general discussion of the fire installations in the Near East see Aurenche 1981; for a detailed evaluation of the portable hearths in Anatolia see Diamant and Rutter 1969. 7 This bowls are often characterized by the burnishing of the inner side, sometime spread also on the upper part of the outer side. 8 A similar arrangement with two ranges of tannur-hearts set against the outside walls of dwellings was found at the Jordan site of ‘Aqaba/Ayla in the Abbasid occupation phase (McQuitty 1994, p. 60-65, figs. 7-8). 9 For a preliminary typology of Middle Bronze basalt tools from Ebla see Merluzzi 2000. 10 Similar installations for grinding cereals were found also in the Royal Palace G dating to the EB IVA (c. 2400-2300 BC) (Matthiae 1989, p. 76, tavv. 22-24; Dolce 1991, p. 326-330), in which specific quarters devoted to processing food were under the control of the palatial administration, like the textile ateliers (Peyronel 2004, p. 92-97); for the epigraphic documentation related to agricultural products see Archi 1990; 1991; 1999. 11 Matthiae 1987, p. 157. Thirty large storage jars (probably for olive oil) were fitted on a low mudbrick bench in two narrow storerooms (L.4031and L.4043) to the E of the throneroom: Matthiae 1990, p. 409; Dolce 1990, pp. 126-127, fig. 17a-b; 1991, p. 332-334; Pinnock in press. 12 For the different functions of the Eblaic palatial buildings see Matthiae 1991, p. 327-332; 1997b, p. 129-132; for a first analysis of the pit see De Grossi Mazzorin, Minniti 2000 (animal remains) and Peyronel in press b. 13 For a general description of the main types of MB II kitchen ware at Tell Mardikh-Ebla see Nigro 2002a, p. 324; 2002b, p. 112. 14 The fort certainly went through this kind of situation as is demonstrated by the destruction layers and by two complete burnt male skeletons discovered lying directly on the paving of the triangular court in front of the fortress entrance, sound proof of the battle on the fortification during the siege: Peyronel in press b, fig. 9.

Fig. 13. The terracotta portable andirons from L.7124.

fort with the upper central courtyard, is testified by several fragments of portable clay andirons frequently collected in the destruction layers over the floors. Two of these were part of the equipment of the rectangular room L.7124 (Fig. 13), together with some kitchen ware (a cooking pot, a dish and a large bowl with circular burnishing) and simple ware (a small bowl with flaring rim, a medium-sized jar) vessels; these peculiar portable fire installations are horse-shoe shaped, respectively 40 and 30 cm high with a diameter of 30 and 20 cm, resembling the typology of the tannur, although they were not partially inserted into the floor. Several specimens were found in the Northern and Western palaces, and some of them had incised and applied decorations usually on the front side and knobs on the upper and lateral side to make easier the transport (cf. Matthiae et al. 1995, p. 490 n. 428). Summing up, data collected from this defensive building lead to the following observations: a) The fort was equipped with a specific wing devoted exclusively to food processing (mainly cereals), located at its southern edge: the activities here performed included grinding of grain and storage of foodstuffs inside large jars. b) Diversified activities of cooking were performed in the upper central courtyard and inside the two rows of small rooms symmetrically located at the sides of the ramp, where portable horse-shoe shaped andirons were found; c) The horse-shoe shaped hearth with cooking pots and the grinding installations from Fortress V allowed the identification of the complete sequence of the chaîne opératoire of food transformation and revealed the selfsufficiency of the defensive tower, which, if necessary, enable the resistance of the soldiers in case of external attacks.

Endnotes We would like to sincerely thank Prof. Paolo Matthiae, director of the Italian Archaeological Expedition at Ebla, who allows us

1

221

L. Peyronel, G. Spreafico References

Matthiae, P., 1990. Nouvelles fouilles à Ebla en 1987-1989. Comptes Rendus des Séances de l’Académie des Inscriptions et Belles-Lettres, 1990, 384-431. Matthiae, P., 1991. Architettura e urbanistica di Ebla Paleosiriana. La Parola del Passato, 46, 304-371. Matthiae, P., 1997a. Tell Mardikh, 1977-1996: Vingt Ans de Fouilles et de Découvertes. La Renaissance d’Ebla Amorrheenne. Akkadica, 101, 1-29. Matthiae, P., 1997b. Typologies and Functions in the Palaces and Houses of Middle Bronze II Ebla. In  : C. Castel, M. al-Maqdissi, F. Villeneuve, eds. Les maisons dans la Syrie antique du IIIe millénaire aux débuts de l’Islam. Pratiques et représentations de l’espace domestique. Beyrouth : IFAPO, 125-134. Matthiae, P., 1998. Les fortifications de l’Ebla paléo-syrienne: Fouilles à Tell Mardikh, 1995-1997. Comptes Rendus des Séances de l’Académie des Inscriptions et BellesLettres,1998, 555-586. Matthiae, P., 2000. Nouvelles fouilles à Ébla (1998-1999): Forts et palais de l’enceinte urbaine. Comptes Rendus des Séances de l’Académie des Inscriptions et Belles-Lettres, 2000, 567-610. Matthiae, P., 2001. A Preliminary Note on the MB I-II Fortifications System at Ebla. Damaszener Mitteilungen, 13, 29-51. Matthiae, P., 2002. Fouilles et restaurations à Ebla en 20002001: Le Palais Occidental, la Résidence Occidentale et l’urbanisme de la ville paléo-syrienne. Comptes Rendus des Séances de l’Académie des Inscriptions et BellesLettres 2002, 531-574. Matthiae, P., in press. New Historical Data on the Final Destruction of Ebla at the End of the Middle Bronze II. In: P. Matthiae, L. Nigro, L. Peyronel, F. Pinnock, eds. Proceedings of the International Colloquium ‘From Relative Chronology to Absolute Chronology. The 2nd Millennium BC in Syria-Palestine’, Rome 29th November - 1st December 2001. Roma: Accademia Nazionale dei Lincei. Matthiae, P., Pinnock, F., Scandone Matthiae, G., (eds.), 1995. Ebla. Alle origini della civiltà urbana. Trenta anni di scavi in Siria dell’Università di Roma ‘La Sapienza’. Milano: Electa. McQuitty, A., 1984. An Ethnographic and Archaeological Study of Clay Ovens in Jordan. Annual of the Department of Antiquities of Jordan, 28, 259-267. McQuitty, A., 1994. Ovens in Town and Country. Berytus, 41, 53-76. Nigro, L., 2002a. The MB Pottery Horizon of Tell Mardikh/ Ancient Ebla in a Chronological Perspective. In: M. Bietak, ed. The Middle Bronze Age in the Levant. Proceedings of an International Conference on MB IIA Ceramic Material. Vienna, 24th-26th of January 2001. Wien: Contributions to the Chronology of the eastern Mediterranean III, 297-328. Nigro, L., 2002b. The Middle Bronze Age Pottery Horizon of Northern Inner Syria on the Basis of the Stratified Assemblages of Ebla and Hama. In M. al-Maqdissi, V. Matoian, Ch. Nicolle, eds. Céramique de l’Âge du Bronze en Syrie I. La Syrie du Sud et la Vallée de l’Oronte. Beyrouth: IFAPO, 97-151. Ochsenschlager, E., 1974. Mud Objects from al-Hiba. A Study in Ancient and Modern Technology. Archaeology, 27, 162-174.

Archi, A., 1990. Agricultural Production in the Ebla Region. Annales Archéologiques Arabes Syriennes, 40, 50-55. Archi, A., 1991. Culture de l’olivier et production de l’huile à Ebla. In  : D. Charpin, F. Joannes, eds. Études sur la civilisation mésopotamienne offerts à Paul Garelli. Paris: Editions Recherche sur les Civilisations, 211-222. Archi, A., 1999. Cereals at Ebla. Archív Orientální, 67, 503518. Ascalone, E., Peyronel, L., 1999. The Eblaite Metrology in Middle Bronze Age. Archaeological Context and Distributive Analysys of Wheigths. Proceedings of the XLIV RAI, Venice 7-11 July 1997. Padua: Sargon Ascalone, E., Peyronel, L., Spreafico, G., Caracuta, V., Fiorentino, F., in press. Functional Analysis of a Courtyard in a House of Middle Bronze II at Tell Mardikh-Ebla, Syria. In: H. Kühne et al., eds. Proceedings of the 4th International Congress on the Archaeology of the Ancient Near East, Berlin 28th March – 4th April 2004. Berlin. Aurenche, O., 1981. La maison orientale. L’architecture du Proche Orient Ancien des origines au milieu du quatrième millénaire. Paris: Librairie Orientaliste Paul Geuthner S.p.A. Baffi, F., in press. House Hunting. I am an Archaeologist. In: F. Baffi, R. Dolce, R. Ramazzotti, S. Mazzoni, F. Pinnock, eds. “Ina kibrat erbetti”. Studies in Honor of Paolo Matthiae, Presented by Colleagues and Friends on the Occasion of His 65th Birthday. Roma Dalman, G., 1935. Arbeit und Sitte in Palästina IV. Brot, Öl und Wein. Schriften des Deutschen Palästina Instituts, 7. De Grossi Mazzorin, J., Minniti, C., 2000. The Northern Palace of Tell Mardikh-Ebla (Syria): Archaeozoological Analysis of the Refuse Pit F.5861/F.5701. In: P. Matthiae, A. Enea, L. Peyronel, F. Pinnock, eds. Proceedings of the First International Congress on the Archaeology of the Ancient Near East. Rome, May 18th-23rd 1998. Università di Roma “ La Sapienza”, 311-322. Diamant, S., Rutter, J., 1969. Horned Objects in Anatolia and the Near East and Possible Connexions with the Minoan “Horns of Consecration”. Anatolian Studies, 19, 147177. Dolce, R., 1990. Les magasins et les lieux de traitement des derrés alimentaires à Ebla au III et II millénaires. Annales Archéologiques Arabes Syriennes, 40, 122-145. Dolce, R., 1991. Luoghi e sistemi di accumulo delle derrate. Per una riqualificazione urbanistica degli edifici secolari preclassici in Siria e Mesopotamia. Scienze dell’Antichità, 5, 309-334. Merluzzi, E., 2000. Basalt Tools at Ebla: An Example of “Ground Stone” Industry in a Central Syria Site of the Bronze Age Period. In: P. Matthiae, A. Enea, L. Peyronel and F. Pinnock, eds. Proceedings of the First International Congress on the Archaeology of the Ancient Near East. Rome, May 18th-23rd 1998. Università di Roma “ La Sapienza”, 1061-1078. Matthiae, P., 1985. I tesori di Ebla. Roma, Bari: Laterza. Matthiae, P., 1987. Les dernières découvertes d’Ebla en 19831986. Comptes Rendus des Séances de l’Académie des Inscriptions et Belles-Lettres, 1987, 135-161. Matthiae, P., 1989. Ebla. Un impero ritrovato. Dai primi scavi alle ultime scoperte. Torino: Einaudi.

222

Food Processing in the Levant during the Middle Bronze Age Peyronel, L., 2000. Middle Bronze II Fortress V at Tell MardikhEbla (Syria). Preliminary Analysis of Architectural Contexts and Archaeological Materials. In: P. Matthiae, A. Enea, L. Peyronel, F. Pinnock, eds. Proceedings of the First International Congress on the Archaeology of the Ancient Near East. Rome, May 18th-23rd 1998. Università di Roma “ La Sapienza”, 1353-1377. Peyronel, L., 2004. Gli strumenti di tessitura dall’età del Bronzo all’epoca Persiana. Materiali e Studi Archeologici di Ebla IV. Roma: Missione Archeologica Italiana in Siria. Peyronel, L., in press a. Late Old Syrian Fortifications and Middle Syrian Re-Occupation of the Western Rampart at Tell Mardikh-Ebla. Problems of Relative Chronology and Stratigraphic Sequence, In: M. Bietak, ed. Proceedings of the 2nd Euro-Conference, Vienna. Peyronel, L., in press b. Domestic quarters, refuse pits, and working areas. Reconstructing human landscape and environments at Tell Mardikh-Ebla during the Old Syrian period. In: H. Kühne et al., eds. Proceedings of the 4th International Congress on the Archaeology of the Ancient Near East, Berlin 28th March – 4th April 2004, Berlin. Peyronel, L., in press c. Looking towards the Steppe. The Eastern Rampart at Tell Mardikh-Ebla (Syria): Old Syrian Fortifications in Areas M and EE. In: J.Cl. Margueron, P. de Miroschedji, J.P. Thalmann, eds. Proceedings of the 3rd International Congress on the Archaeology of the Ancient Near East, Paris 15th-19th, 2002. Paris: Eisenbrauns Pinnock, F., in press. La ceramica del Palazzo Settentrionale del Bronzo Medio II. Materiali e Studi Archeologici di Ebla VI, Roma. Renfrew, J.M., 1995. Vegetables in the Ancient Near East Diet. In: J.M. Sasson, ed. Civilizations of the Ancient Near East. New York: Charles Scribner’s Sons, (I), 91-102.

223

High resolution AMS radiocarbon dating of archaeological charcoals G. Quarta, M. D’Elia and L. Calcagnile

CEDAD, Centro di Datazione e Diagnostica, Department of Engineering of Innovation, University of Lecce, Italy E-mail: [email protected] Abstract After 20 years since its development Accelerator Mass Spectrometry radiocarbon dating is now a mature technique, fundamental in many research fields such as archaeology, Earth and environmental sciences, biomedicine. Here we discuss the advantages and the methodological aspects related to the selection and the use of charcoals as samples for 14C AMS dating. A discussion is also reported concerning the measurement the δ13C mass fractionaction term with an accelerator mass spectrometer, referring, in particular, to the achievable precision and accuracy levels.

Introduction The demonstration in the late 70’s of the possibility to use large electrostatic accelerators, initially devoted to nuclear physics studies, as an experimental method alternative to the conventional β-counting approaches (Libby 1955) for radiocarbon dating purposes has represented a real turning point in the history of the radiocarbon dating method and of its application to archaeology (Purser et al. 1977; Muller et al. 1977). The two main advantages of AMS, if compared to conventional counting methods, are the reduction of ed sample material and measuring time of a factor one thousand and one hundred, respectively. Recently the possibility to date by AMS sample of mass as low as 200 μg has been also reported (Hua et al. 2004). In particular the reduction of the required sample mass has resulted, primarily, in the reduction of the destructiveness of the method, which is a crucial factor in the study of object of particular relevance and value or even unique (D’Elia et al. 2004a), but also, in the possibility to use more accurate sample selection strategies both in terms of the samples’ informative content and of sample contamination. Furthermore the possibility to select single portion of the samples even at the molecular level from complex biological matrices (Stafford et al. 1991) such as bones, wood, charcoals and textiles, together with a more close understanding of the 14C uptake mechanisms represents the current frontier in the application of the method. Beside the importance of the method and its potentialities in several research fields such as archaeology, geology, earth and environmental sciences, biomedicine a new generation, high precision and high throughput facility was completely missing in Italy till 2001 when a new dedicated AMS radiocarbon dating facility (CEDAD, Centro di Datazione e Diagnostica) was installed by the University of Lecce (Calcagnile et al. 2004a). Radiocarbon dating of charcoals Charcoal is, by far, one of the sample material more commonly used for radiocarbon analysis. Fig. 1 shows the sample materials analysed in the first year of operation at the

Charcoal 38%

Other 3% Sediments 3% Vegetable Rests 9%

Bones 37%

Wood 6% Marine Carbonates 4%

Fig. 1. Summary of the samples measured at CEDAD in 2004. The charcoals represent 38 % of the total.

AMS facility in Lecce. The charcoals represent the 38 % of the total. Different factors made charcoal particularly suitable for radiocarbon dating: a. It is one of the material most commonly present in archaeological contexts; b. For the vegetables the uptake of radiocarbon take place, almost exclusively, through exchanges with the atmosphere and the so-called radiocarbon reservoir effects are more rare than what observed for other samples such the human bones (van Klinken 1999). Nevertheless possible “reservoir” effects have been reported for particular environmental conditions (G. Quarta et al. 2005). c. The possibility to contemporary use the charcoal samples not only as chronological but also as paleo environmental indicators. One of the main problem related to the use of charcoal as dating element, especially when a high accuracy level is required, is the so called “old wood effect”: the difference between the age of the charcoal itself (which reflect the age of the tree ring or the tree rings it is made of) and the age of the event to be dated (Jung et al. 2004). The identification and the following selection of short lived species or an accurate selection of the tree ring can avoid the occurrence of this problem (Butalag et al. 2005).

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

G. Quarta, M. D’Elia, L. Calcagnile 100 Radiocarbon age measurement uncertainty (570 AMS measurements)

Frequency

80

60

40

20

0

0

20

40

60

80

100

120

140

160

180

200

Radiocarbon Age Uncertainty (years BP) Fig. 2. Frequency diagram showing the uncertainty in a set of 570 radiocarbon measurements. No distinction is done with respect to the age of the samples (ranging from modern to 42 ka BP), their mass (ranging from 0.3 to 2 mg) and their conservation status.

Methodological aspects

of alternate Acid-Alkali-Acid attacks, the combustion of the purified sample material in sealed quartz tubes at 900°C and the final conversion of the extracted CO2 by The radiocarbon dating of charcoal is a complex multi step Figure 2. Frequency diagram showing the of a catalytic reduction reaction into solid graphite. mean process consisting of a series of treatment (both physical uncertainty in a set of 570 radiocarbon The graphite is then pressed into the aluminium sample and chemical) which are intended to remove any possible measurements. No distinction is done with holders of the accelerator mass spectrometer ion source for source of contamination which could fake the radiocarbon respect to the age of the samples (ranging from the measurement of the isotopic ratios. result. The entire sample processing is described in details modern to 42 ka BP), their mass (ranging from The measurement of the 14C/12C and 13C/12C carbon elsewhere (D’Elia et al. 2004b) here we just mention that it 0.3 to 2 mg) and their conservation status. isotopic ratios is carried out by comparing the ion beams consists in the purification of the sample material by mean 160

13

G C measurement uncertainty (570 AMS measurements)

140

-14

100 80 60

G C (‰)

Frequency

120

13

40 20 0

-16

-28 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

13

G C Uncertainty PDB (‰) Figure 3. Frequency diagram showing the uncertainty in a set of 570 measurements of the δ13C term.

226

Figure 3. Frequency diagram showing the uncertainty in a set of 570 measurements of the G13C term.

A A A

-30

High resolution AMS radiocarbon dating of archaeological charcoals

13

G C av = -14.8 ± 0.1 ‰

-16

13

G C (‰)

-14

-28

13

G C av = -29.0 ± 0.1 ‰

Cynodon Dactylon (C4 Plant) Agropyron Repens (C3 Plant) Average for Cynodon Dactylon Average for Agropyron Repens

-30

Sample Figure 4. Measurement of the δ C term for two groups of C3 and C4 plants. The two different photosynthetic pathways are clearly distinguished and identified. 13

The isotopic mass fractionation: the δ13C term

extracted and collected from the unknown samples and those collected from reference materials supplied by the International Atomic Energy Agency (IAEA). A complete description of the entire measuring procedure is reported by Calcagnile et al. (2005). In the first years of operation of the CEDAD the possibility to carry out 14C/12C measurements with a precision level of 0.3-0.4 % on the single determination (corresponding to an uncertainty in the radiocarbon age of ± 30 years) has been demonstrated (Calcagnile et al. 2004b). As a summary Fig. 2 shows the uncertainty obtained in the measurement of a set of 570 samples, with an age ranging from the modern to 42 ka BP, no distinction is made between the samples with respect to their mass (ranging between 0.2 and 2 mg) and their conservation status. Most of the samples are measured with a precision ranging between 20 and 80 years. In the optimal conditions (1 mg of extracted carbon) and for samples younger than 2000 years an uncertainty of ± 30 years is normally achieved for a single measurement. High precision measurement are also possible with an uncertainty of only 20 years. The accuracy is constantly monitored by the measurement of standard materials of different ages and by the participation to international intercomparison programme such as the VIRI, Fifth International Radiocarbon Intercomparison (Scott et al. 2003). A detailed description of the quality assurance protocol developed at CEDAD is given by Calcagnile et al. (2005).

Together with the measurement of the 14C/12C isotopic ratio, used for the age calculation, the accelerator mass spectrometer in operation at CEDAD allows also the on-line measurement of the 13C/12C ratio, with a relative precision better than 0.1 % (Calcagnile et al. 2004b). This gives the important possibility to calculate the δ13C term given, in per mil, by: § 13C · § 13C · ¨¨ 12 ¸¸  ¨¨ 12 ¸¸ © C ¹ sample © C ¹ St G 13C ˜ 1000 § 13C · ¨¨ 12 ¸¸ © C ¹ St 13 § C· ¨¨ 12 ¸¸ Where © C ¹ St refers to the standard internationally

accepted and used, either directly or indirectly, which is a carbonate (Belemnite) from the Peedee formation in South Carolina, USA (PDB: Pee Dee Belemnite). In radiocarbon dating the δ13C term is primarily used to correct the radiocarbon age determination for the carbon isotopes mass fractionation as reported by Stuiver, Polach (1977). Fig. 3 shows, for the same set of 570 samples reported in Fig. 2, the precision in the δ13C measurement: an uncertainty better than 0.3 ‰ is routinely achieved. It is important to underline that the δ13C term measured with the accelerator reflects not only the natural mass 227

G. Quarta, M. D’Elia, L. Calcagnile fractionation occurring in the samples but also all the mass dependent phenomena occurring during the complex sample processing and the following measurement with the accelerator (Quarta et al. 2004). Nevertheless the accuracy level achieved in the measurement of the δ13C term with accelerator mass spectrometers (which are obviously primarily designed for the measurement of rare isotopes such as 14C) is enough to carry out important investigations concerning the stable carbon isotope discrimination of the plants. As an example, Fig. 4 shows the δ13C term measured with the accelerator mass spectrometer in Lecce for six modern plants collected in the same area in Brindisi, Italy. Three of the samples were Cynodon dactylon (with a C4 photosynthetic pathway) and three were Agropyron repens (with a C3 photosynthetic pathway). The measurements have shown that the two different photosynthetic pathways can be clearly distinguished, with an average δ13C value of -14.8 ± 0.1 and -29.0 ± 0.1 for the C4 and C3 plant respectively. Additionally the analysis of the results shows that, within each of the two groups, the data are normally distributed with an overall uncertainty of 0.1 ‰. Furthermore the measurement of international standards such as IAEA C1 (Carrara marble, δ13C = 2.42 ± 0.33 ‰), C3 (Cellulose, δ13C= -29.91± 0.49 ‰), C4 (Subfossil wood, δ13C = -23.96 ± 0.63 ‰), C6 (Sucrose, δ13C=-10.8 ± 0.347 ‰), shows that the overall accuracy in the δ13C measurement is between 1 and 1.5 ‰. It is under investigation the possibility to integrate the age and the δ13C term determinations to obtain paleo environmental information.

Calcagnile, L., Quarta, G., D’Elia, M., Rizzo, A., Gottdang, A., Klein, M., Mous, D.J.W., 2004a. A new accelerator mass spectrometry facility in Lecce, Italy. Nuclear Instruments and Methods in Physics Research B, 223-224, 16-20. Calcagnile, L., Quarta, G., D’Elia, M., Gottdang, A., Klein, M., Mous, D.J.W., 2004b. Radiocarbon precision test at the Lecce AMS facility using a sequential injection system. Nuclear Instruments and Methods in Physics Research B, 215, 561-564. Calcagnile, L., Quarta, G., D’Elia, M., 2005. High resolution accelerator-based mass spectrometry: precision, accuracy and background. Applied radiation and Isotopes, 62, 623629. D’Elia, M., Quarta, G., Sanapo, C., Laudisa, M., Rizzo, A., Calcagnile, L., 2004a. Dating a medieval painting prepared on the new AMS sample preparation lines in Lecce. Proceedings of the 4th Symposium “Radiocarbon and Archaeology” St Cathrine’s College, 9-14th April, 2002. Oxford University School of Archaeology, Monograph 62, 321-326. D’Elia, M., Calcagnile, L., Quarta, G., Sanapo, C., Laudisa, M., Toma, U., Rizzo, A., 2004b. Sample preparation and blank values at the AMS radiocarbon facility of the University of Lecce. Nuclear Instruments and Methods in Physics Research B, 223-224, 278-283. Hua, Q., Zoppi, U., Williams, A.A., Smith, A.M, 2004. Smallmass AMS radiocarbon analysis at ANTARES. Nuclear Instruments and Methods in Physics Research B, 223224, 284-285. Jung, R., Weninger, B., 2004. Kastanas and the chronology of the Aegean late bronze and early iron age. Proceedings of the 4th Symposium “Radiocarbon and Archaeology” St Cathrine’s College, 9-14th April, 2002. Oxford University School of Archaeology, Monograph 62, 209-228. van Klinken G.J., 1999. Bone collagen quality indicators for paleodietary and radiocarbon measurements. Journal of Archaeological science, 26, 687-695. Libby, W.F., 1955. Radiocarbon dating. Chicago: The University of Chicago Press. Muller, R.A., Stephenson E.J., Mast T.S., 1977. Radioisotope dating with an accelerator: a blind measurement. Science 201, 347-348. Purser K.H., Litherland A.E., Gove H.E., 1977. An attempt to detect stable N-ions from a sputter ion source and some implications of the results for the design for ultrasensitive carbon analysis. Review Physics Applications 12, 1487-1492. Quarta, G., D’Elia, M., Calcagnile, L., 2004. The influence of injection parameters on mass fractionation phenomena in radiocarbon analysis. Nuclear Instruments and Methods in Physics Research B, 217, 644-648. Quarta, G., D’Elia, M., Rizzo, G.A., Calcagnile, L., 2005. Radiocarbon dilution effects induced by industrial settlements in Southern Italy. Nuclear Instruments and Methods in Physics Research B., 240, 458-462. Scott, E.M., Bryant, C., Cook, G.T., Naysmith, P., 2003. Is there a Fifth International Radiocarbon Intercomparison (VIRI). Radiocarbon, 45 (3), 493-495. Stafford, T.W. Jr., Brendel, K., Duhamel, R.C., 1991. Accelerator radiocarbon dating at the molecular level. Journal of Archaeological Sciences, 18, 35-72. Stuiver, M., Polach, H.A., 1977. Discussion, reporting of 14C data. Radiocarbon, 19 (3), 355-363.

Conclusions Accelerator mass spectrometry radiocarbon dating, after more than two decades since its introduction, is now a well developed technique, became fundamental in many research field and in particular in archaeology. In the new AMS facility in Lecce, Italy, an uncertainty as low as 30 years is routinely achieved in the radiocarbon measurement of a single sample of mass as low as few milligrams. Furthermore the possibility to measure, together with the 14 C/12C ratio also the δ13C term, with uncertainties of 0.1 ‰ and accuracies of 1-2 ‰, is turning to be very important in many research field such as the paleoenviornment reconstruction and the study of the carbon cycle.

References Butalag, K., Demortier, G., Quarta, G., Muscogiuri, D., Maruccio, L., Calcagnile, L., Pagliara, C., Maggiulli, G., Mazzotta, C., 2005. Checking the homogeinity of gold artefacts of the final Bronze Age found in Roca vecchia, Italy, by Proton Induced X-Ray emission. Nuclear Instruments and Methods in Physics Research B., 240, 565-569.

228

Environmental history in the Mediterranean basin: microcharcoal as a tool to disentangle human impact and climate change L. Sadori, M. Giardini

Dipartimento di Biologia Vegetale, Università “La Sapienza”, Piazzale Aldo Moro, 5, 00185 Roma E-mail: [email protected] Abstract The role of fire in two Mediterranean environments of central (Lago di Mezzano, Latium) and insular (Lago di Pergusa, Sicily) Italy has been detected through microcharcoal analysis. The microcharcoal data have been integrated with pollen percentage and concentration data of lacustrine sediments since mid Holocene, i.e. the last 4500 years. Particular attention was paid to the choice of the sites; both extant lakes are located in lands with an almost continuous history of prehistoric and historic settlements. An attempt was made to integrate archaeological evidence with environmental and human history. The chronology of the two sediment records, based on AMS radiocarbon dates, on bulk sediments or terrestrial plant macrofossils when available, is expressed in cal. years BP. The physiognomy of the vegetation appears to have been quite different during the last millennia in the two regions. The forests of Latium were much denser than those of Sicily all through the investigated period. Regional and local fires appear to have been very strong in Sicily around 4000, slightly before 2100 and around 1200 years BP. In Sicily, local human presence is distinctly recognized in the pollen diagram only in the last two millennia, and is not very clear even in periods in which important settlements in the area were documented, indicating that pollen can only assist in detecting local settlements. On the contrary, fire occurrences indicate that the region underwent a strong regional human impact, not easily detectable by pollen analysis. At Lago di Mezzano, in Latium, important regional and local fires appeared later than in Sicily. An important drop in arboreal pollen concentration found around 3900 years BP, was ascribed, using charcoal analysis, to a climate change towards aridity. Water shortage forced the Bronze Age populations to settle by the lake shores. Local and regional fire dynamics show how humans contributed in shaping the environment, with comparable trends in the period between 3700 and 2900 years BP, when two phases of human presence were detected also by pollen and organic carbon contents. Less important, but continuous traces of burning, were found again around 2400 years BP, when the surrounding land is believed to have been exploited by Etruscans, and between 100 and 600 A.D., during the Roman Empire.

Introduction Disentangling climatic and human impact is an everlasting minefield, in a matter where in fact there is no easy solution. Potentially the best way to detect the cause of landscape change comes from combining continuous records from lake cores obtained as close as possible to archaeological sites, which are the direct link to cultural history. In this paper, pollen and microcharcoal analyses of lacustrine records from the last 4500 years of two extant Italian lakes (Lago di Pergusa, central Sicily and Lago di Mezzano, northern Latium) from different Mediterranean environments, are used. The high potentiality of pollen analysis in palaeonvironmental studies and its use to reconstruct the climate of the past and to evaluate human impact is well known, but the occurrence of natural burnings and the use of fire by prehistoric populations in the forest clearance which resulted in the present-day vegetation state still deserves to be investigated in the Mediterranean Basin. Few and chronologically limited data is in fact available from the area (Caroli et al. 2004; Carrión et al. 2003; Sadori 2003; Sadori and Giardini 2004 a, b; Sadori et al. 2004). There is generally no doubt that major events were happening across a wide geographical range relatively synchronously for which humans cannot have been responsible alone. During the Neolithic and Eneolithic periods, since ca. 8000 years ago, the macrofossil record shows that agriculture was a common practice all over Italy (e.g. Castelletti and Carugati 1994; Costantini and Stancanelli 1995; Follieri 1973, 1977-82; Rottoli 1999)

and both Latium (Bulgarelli et al. 1993; Celant 2000; Coubray 1997; Fugazzola Delpino 1996; Rottoli 1993, 2000-2001) and Sicily (Costantini 1981) experienced the same cultural development as the rest of the country. We should therefore expect that human impact on the landscape increased markedly as populations expanded rapidly under conditions favourable to farming and life. Human impact on Mediterranean landscapes has been hard to detect in pollen diagrams and other contexts before the Bronze Age (Roberts et al. 2004), probably because previously it was local in scale and did not affect regional patterns of vegetation. The fact that even human presence is hardly detectable in Mediterranean regions until the Bronze age, and that in many sites a clear human impact is found only since the Roman periods, is incredible. If prehistoric populations did not produce strong changes on the landscapes (at least on a broad scale), their influence therefore has to be sought in lake archives very close to archaeological sites. It was only during the Bronze Age that the number of perilacustrine settlements in the Italian peninsula increased, probably because in that period water became a less available resource (Sadori et al. 2004). Two arguments are used to explain this lack of evidence and delay in evidence coming from pollen records: 1. The distance between lake and archaeological settlements. If the areas around the lakes were not extensively exploited and frequented by human populations, and if the presence was not local, pollen spectra are unlikely to have recorded changes induced

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

L. Sadori, M. Giardini

Fig. 1. Location of the investigated lakes and main archaeological sites. Latium: 1 – Poggio Evangelista (Bronze Age, Etruscan period); 2 – Lago di Mezzano (Bronze Age, Middle Ages); 3 – Montemerano (Bronze Age); 4 – Monte Senano (Bronze Age); 5 – Monte Bisenzo (Bronze Age, Iron Age); 6 – Sorgenti della Nova (Bronze Age, Middle Ages); 7 – Poggialti Vallelunga (Copper Age); 8 – Scarceta (Bronze Age); 9 – Crostoletto di Lamone (Bronze Age); 10 – Naviglione (Copper Age); 11 – Castelletto di Prato Frabulino (Bronze Age); 12 – La Selvicciola (Copper Age, Etruscan Age, Roman period, Middle Ages); 13 – Ponte S. Pietro (Copper Age); 14 - Grotta Nuova (Bronze Age); 15 – Grotta delle Settecannelle (from Palaeolithic to Iron Age); 16 – Poggio Olivastro (Neolithic Age, Copper Age); 17 – Canino (Middle Ages). Sicily: 1 – Case Bastione (Neolithic Age, Copper Age); 2 – Realmese (Bronze Age, Iron Age); 3 – Malpasso (Iron Age); 4 – Calcarella (Bronze Age, Iron Age); 5 – Enna (Copper Age, Bronze Age, Greek period, Middle Ages); 6 – Cozzo Matrice (Copper Age, Greek period); 7 – Riparo di Contrada S. Tommaso (Bronze Age); 8 - Monte Giulfo (Greek period); 9 – Capodarso (Copper Age, Bronze Age, Greek period); 10 – Rocche (Greek period); 11 – Montagna di Marzo (Greek period); 12 – Rossomanno (Greek period, Middle Ages); 13 – Morgantina (Greek period); 14 – Contrada Gaspa (Roman period); 15 – Contrada Runzi (Roman period); 16 – Canalotto (Middle Ages).

by humans on vegetation. 2. The indigenous origin of many edible plants and difficulty in identification of many anthropic indicators. Most edible plants such as cereals, pulses and fruit trees are native to Mediterranean regions and are found during the whole Holocene and even before in the pollen diagrams. An exemplification can be made quoting the cereal pollen type which includes pollen of both cultivated and wild cereals as well as of other grasses (Andersen 1978). Also Olea europaea L. (olive) and Vitis vinifera L. (grape vine) are native to the Mediterranean regions. Plantago lanceolata L., a synanthropic herb whose finding is attentively taken into account as evidence of human presence in central Europe, has pollen grains which cannot be distinguished from those of other Plantago species indigenous in Italy (Reille 1992b). Another argument used to prove human impact was (Brande 1973; Iversen 1941) and still is (Reille 1992a) forest clearance, easily detectable by a decrease of AP (pollen of Arboreal Plants) percentages and, if used, concentrations (Stockmarr 1971). We know from long south-eastern European sequences such as Valle di Castiglione (Follieri et al. 1988), Tenaghi Philippon (Wijmstra 1969; Wijmstra, Smit 1976), Ioannina (Tzedakis 1994) that similar changes have frequently occurred on long time periods, not only at the interglacial/ glacial transitions, but also during the interglacial phases, in periods in which man cannot be suspected to be the main agent of deforestation. An opening of vegetation in the interglacial woods of the last millennia could therefore also have been caused by climate forcing due to physical 230

causes in regions very sensitive to climatic change. A good tool for disentangling climate change and human impact is the combined interpretation of either pollen percentage and concentration data (the latter being considered as an estimate of the biomass amount) and microcharcoal trends. Dramatic and rapid tree pollen concentration drops, not necessarily matching tree percentage decreases, have been pointed out during forest phases in many pollen diagrams of Latium during the last hundreds of thousands of years (Follieri et al. 1988; Magri 1999; Magri, Sadori 1999) and interpreted as vegetation responses to climate changes. Because similar drops are also found during the Holocene at Lago di Pergusa and at Lago di Mezzano we have to consider the possibility that human communities had caused them. But how could prehistoric populations have induced sudden, enduring, dramatic arboreal phytomass reductions on a large scale only by cutting trees? The necessity to investigate past fires has become clear and carried out using microcharcoals as indicators of forest burnings. This study will try to distinguish lightning and human ignition using other clues recognizable in lacustrine sediments. It should however be considered that human populations used to burn forests not only to obtain open areas for agriculture and cattle breeding, but also for domestic activities, heating, and metallurgy. The investigated sites The common feature of the two lakes chosen for this comparison is the closeness to important archaeological sites dating back to the investigated period (Fig. 1).

Environmental history in the Mediterranean basin Lago di Mezzano (northern Latium, Central Italy, 42°36’N, 11°46’E, 452 m a.s.l., 30 m deep) is a lake of volcanic origin (in an eccentric position in the Caldera of Latera) which experienced in prehistoric and historical times many lake level lowerings (Sadori et al. 2004). Human presence in the caldera is documented since the Palaeolithic (Conti et al. 1993). Three submerged settlement areas of Bronze Age pile-dwellings (M1, M2, M3) with posts, pottery and metal tools have been singled out (Petitti, Mitchell 1993). The archaeological material indicates that the site was inhabited from the early Bronze Age to the first phase of middle Bronze Age. Many bronze axes belong to this period. After a short period of abandonment, a new human presence is documented during the late Bronze Age, but at that time there were probably no settlements by the lake, which was mainly used for cult purposes, as indicated also by the finding of two golden bronze swords (Pellegrini 1993). The archaeological reconstructions was confirmed by geomorphological, geochemical and palynological analyses carried out on lacustrine sediments. During the final part of the Bronze Age, when the lakeside seems to have been completely abandoned, the caldera recorded a new phase of intense occupation, whose traces have been identified both in the southern part of the basin (Monte Starnina and Poggi del Mulino) and in the northern part, where the important site of Poggio Evangelista (Di Gennaro 1986) is located on the top of the caldera rim, not far from the highest point of the lake’s catchment area. Later on, during Etruscan/early Roman times, the site was probably used again by human populations. Some authors identified Lago di Mezzano as the so-called Lacus Statoniensis, which was close to the Roman municipium

of Statonia, later destroyed by Vandals. In the land surrounding Lago di Mezzano remnants of Roman villas are found. During the Middle Ages (900-1348 A.D.) the village of Mezzano, on the crater edge, was a settlement of the Longobard population. Lago di Pergusa (central Sicily, 37°31’ lat. N, 14°18’E long. E, 667 m a.s.l., few decimetres deep at the time of the drilling) is an endangered lake, the only natural inland lake of the island. In the last thirty years decreased precipitations and a bad environmental policy caused the lake level lowering to a few centimetres. A careful program of protection, a control of lake water illicit draining and programmed water supplies from another basin, have been led in the few last years to restore a water depth of more than one metre. As in Latium, also in central Sicily, human activity is documented since the Palaeolithic Age (Tusa 1992). The Eneolithic Age is found not only in the famous necropoles of Malpasso and Realmese (Bernabò Brea 1961) but also in the site of Cozzo Matrice, located on a hill (cozzo in Sicilian) at the edge of the catchment of Lago di Pergusa, and active also during Greek times (Touring Club Italiano 1989). In the surroundings of Lago di Pergusa (Fig. 1) the Bronze and Iron Ages are documented by several sites (Bernabò Brea 1961; Giannitrapani, Pluciennik 1998; Touring Club Italiano 1989; Tusa 1992). Classical writers report that since three thousand years ago the area was settled by Sicans (Diodorus Siculus, Library of History, V, 6, 2-4) and Siculis, then by Greeks, Siracusans and Carthaginians (Touring Club Italiano 1989). Under the Romans, Enna (Castrum Hennae) became a rich and important centre for wheat trade also under the Byzantines and the Arabs.

Fig. 2. Lago di Mezzano and Lago di Pergusa. Regression curves showing the correlation among radiocarbon dates.

231

L. Sadori, M. Giardini

Fig. 3. Lago di Mezzano and Lago di Pergusa. AP percentage, AP and microcharcoal concentration curves.

Materials and methods

In order to estimate the distance from the fire, microcharcoal particles were counted at the microscope, measuring their shortest axe and sorting them in three dimensional classes (between 10 and 50 μm, between 50 and 125 μm, >125 μm). Particles smaller than 50 μm were not taken into account, constituting a background noise, while particles included between 50 and 125 μm are considered as an evidence of regional fires and those bigger than 125 μm are taken as an evidence of local fires (Whitlock 2001). Microcharcoals data are used as concentration (particles/g) values, pollen data are drawn either as percentage or concentration (grains/g).

Microcharcoals were counted in the pollen slides of the sediment records from Lago di Pergusa and Lago di Mezzano, at a magnification of 400 X. The sediment samples processing followed a standard pollen procedure with HCl (37%), HF (40%) and NaOH chemical attacks. This kind of treatment may have implied a charcoal fragmentation and therefore an underrepresentation of the biggest charred particles. For drilling operations, sediment description, methodology, lithostratigraphy, tephra analyses, detailed pollen results of the Holocene lacustrine sediments of Lago di Pergusa, we refer to the published papers (Sadori 2001; Sadori and Narcisi 2001; Narcisi 2002). Regarding Lago di Mezzano, the last 15,000 years were investigated through magnetic susceptibility, lithostratigraphy, geochemistry, tephra, AMS-radiocarbon analyses (Ramrath et al. 2000; Ramrath 1997) of a sediment core spanning the last 30,000 years, while pollen and microcharcoal results were only partially published (Ramrath et al. 2000; Sadori and Giardini 2004a, 2004b; Sadori et al. 2004). The chronological framing of the record was based on radiocarbon AMS dates, both of bulk sediments and macrofossils, and tephra layers whose age was checked with radiocarbon analyses of the associated sediments. The dates have been calibrated with OxCal 3.9 programme (Bronk Ramsey 2003) and the mid point of the interval was chosen to draw an age/depth diagram. To assess the chronological scales, ages were linearly interpolated (Fig. 2).

Results The forests of Latium were much denser than those of Sicily (Fig. 3) throughout the investigated period, i.e., the last 4500 years, either the AP percentages or concentrations were considered. When dealing with absolute pollen curves, two factors should be taken into account when comparing values from the two sites, that are very high at Lago di Mezzano and very low at Lago di Pergusa: first, central Sicily underwent a strong arboreal phytomass drop around 8000 cal. years BP (Sadori, Narcisi 2001) and forests could never completely recover; second, sedimentary and taphonomic processes may have influenced the quantity and quality of pollen found. Also the fire regime shows a diversified pattern. Regional fires (microcharcoals between 50 and 125 μm) appear to have been very strong in Sicily at the beginning of the Bronze Age, soon after 4000 years BP, when also traces 232

Environmental history in the Mediterranean basin diagram from Sicily definitive human presence is detected only in the two last millennia, and is not very clear even in periods in which important settlements in the area were documented (Fig. 1). At Lago di Mezzano, in Latium, important regional and local fires appear later than in Sicily. An important drop in arboreal pollen concentration, possibly due either to climatic or human induced forest clearance, is found around 3900 years BP (Fig. 3). As it seems unlikely that prehistoric populations could produce a reduction of forest biomass at a regional level to less than one/fifth by only cutting trees, the role of fire can help in defining this strong and sudden vegetational change. The microcharcoals data indicates that the forest clearance was not induced by fires as it preceded local and regional fires, which only later influenced AP percentages. This sudden change towards aridification is not exclusive of Lago di Mezzano, but was found in many other lakes of Latium, either in sediment or in low lake levels records (Calderoni et al. 1994; Follieri et al. 1988; Giraudi 2000, 2002; Giraudi, Narcisi 1994; Lowe et al. 1996; Magri 1999; Magri, Sadori 1999; Sadori et al. 2004). The water level of Lago di Mezzano lowered, as it is indicated by pile dwellings found at more than 10 m water depth. During the Early Bronze Age water availability decreased and obliged the populations to stay close to the water, by the lake shores. But one question is not yet answered. What was then the role of man? Was it just the climate that made the difference? Local and regional fire dynamics can help in understanding the human contribution in shaping the environment and show comparable trends in the period between 3700 and 2900 years BP, when two phases of human presence were detected also by pollen (Fig. 5) and organic carbon contents (Sadori et al. 2004). Although less important, continuous traces of burning were found up to 3000 years BP. An isolated peak of the larger charcoal fraction is found at 2700 years BP, matching the first appearance of Juglans pollen. From 2700 up to 1800 years BP (200 A.D.) there is no evidence of strong local land exploitation, either in local fires or in pollen indicators, but a regional human presence cannot be ruled out. Regional fires again took greater importance around 2400 years BP, when the surrounding land is believed to have been exploited by Etruscans, and between 100 and 600 A.D. during the apex and the decline (early Barbarian invasions) of the Roman Empire. In the first part of the second period just mentioned, around 100 and 400 A.D. according to the established chronology, a first phase of land use with increasing local fires is also suggested by intense cultivation (Castanea, Juglans, Olea, Vitis, cereals and legumes). In the following two centuries the landscape changed, no longer being intensely cultivated, but none the less frequented as suggested by the curve of Urticaceae, which shows a parallel behaviour with local fires. The occurrence of local burnings seems to be almost continuous in the last 1200 years, probably linked to local human presence pointed out also by many pollen anthropic indicators.

Fig. 4. Lago di Pergusa. Pollen percentage of selected taxa and microcharcoal concentration diagram.

Fig. 5. Lago di Mezzano. Pollen percentage of selected taxa and microcharcoal concentration diagram.

of local fires were found. As a consequence of fire, trees like deciduous oaks and Olea underwent a decrease, while only slight traces of cultivation can be detected (Fig. 4). Regional fires continued up to the top of the diagram, even if of decreased importance (Fig. 3). Hints of local fires are found again at Pergusa only slightly before 2100 years BP, some centuries after the first appearance of Juglans, dated at 2800 years BP, and around 1200 years BP. In the 233

L. Sadori, M. Giardini Conclusions

populations to settle by the lake. The local living activities of the pile dwellings inhabitants left valuable clues in the lacustrine archive. A similar behaviour is observed also in the diffusion of other anthropic-indicator plants, matching local fires during the first centuries of the Iron Age, until 2500 years BP, during the Roman Empire, and again during Middle and Modern Ages. A foreign presence, but probably widespread in the area, can be found in the Etruscan period. However a number of questions remain open: is the charcoal of the Roman age related to the surrounding Roman villas or to the castrum of Statonia (the city arising along the shores of Lacus Statoniensis, according to some authors identifiable with Lago di Mezzano)? And, if yes, is there any trace of the conflagration which destroyed the town of Statonia by hands of Vandals? The period (9001348) of the Longobard Medieval settlement of Mezzano, on the crater edge of the homonymous lake, is confirmed by the pollen record, as in the corresponding period, plants indicating cultivation and human disturb are found, together with moderate fire use. After the abandonment of the site by Longobards, fires and forest trees slightly increased. As a matter of fact the integrated study of the two sites led to us hypothesize that only local human presence can be easily detected by palynology, while microcharcoal analysis gives a singular contribution to the understanding of regional land occupation dynamics. Only since the Roman time did the human impact became so strong all over the Mediterranean and the land so exploited on a large scale to leave clear and unequivocal traces in the lacustrine sediment records.

The physiognomy of the vegetation appears to have been quite different during the last millennia around the investigated sites from the two regions. This discrepancy is not only due to obvious different environmental features like geomorphology, climate, flora and vegetation at the two investigated sites, but is also due to different land exploitation and to the fact that herbaceous cultivation is often underrepresented in pollen diagrams, because of the pollen grain size (cereals), or the morphology of the flower and pollination strategy (legumes). It is obvious that, just in the areas where the most ancient civilizations originated, the methods of investigation have to be sharpened and many parameters considered and evaluated as a whole. Microcharcoal analyses turned out to be a useful tool and an important complement to palynology in detecting human presence and land use. It deserves to be pointed out once more that at Lago di Pergusa, with very important archaeological settlements in the close surroundings, almost continuous pollen curves with very low values of cultivated and synanthropic pollen taxa are found, but it is not possible to single out distinct and clear land use phases as at Lago di Mezzano. Moreover, it is not clear if the increasing curve of Olea of the Sicilian record indicates local and/or regional cultivation, or only a natural change in vegetation. This uncertainty remains even comparing curves of olive with other taxa like cereals and legumes, whose pollen is not dispersed far away. The fire regime could be of help in understanding the type of land use, as only three spots of local burnings are found at Pergusa, since the beginning of the Bronze Age, and never in concomitance with spreads of anthropic plants. The general impression, using both charcoal and pollen evidence, is that the human settlements were never local, but so widespread and continuous in the territory to leave only traces of cultivation and burning in the lake record. According to this interpretation, early Bronze Age fires (after 4000 years BP) indicate strong human impact in the area, confirmed by the archaeological sites located some kilometres from the lake (Giannitrapani, Pluciennik 1998). The possibility should however considered that the increase of fires could be the expression of the aridification detected in Latium in the same period. After 4100 years BP, there was an increase of fire intensity and number of events matching an increase of sclerophyllous vegetation, as also recorded in south-eastern Spain by Carrión et al. (2003) and possibly linked both to human activity and climate change. At Lago di Pergusa fires generally matched decreases of AP pollen concentration and percentage values, inducing forest clearance. This kind of behaviour is at times found in Latium, but only in the last two millennia. At Lago di Mezzano the early Bronze Age fires were found only after a consistent drop of arboreal phytomass, indicating that a climate-induced forest opening either favoured spontaneous fires or, most likely, pushed human

Acnowledgments

We thank Neil Roberts who kindly reviewed the manuscript, improving it. The map of archaeological sites surrounding Lago di Mezzano was drawn with the crucial help of Patrizia Petitti from Soprintendenza Archeologica all’Etruria Meridionale.

References ANDERSEN, S.T., 1978. Identification of wild grass and cereal

pollen. Danmarks Geologiske Undersogelse Årbog 1978, 69-92. bernabò brea, L., 1961. La Sicilia prima dei Greci. Milano: Il Saggiatore, III Ed. BRANDE, A., 1973. Untersuchungen zur postglazialen Vegetationsgeschichte im Gebiet der Neretva-Niederungen (Dalmatien, Herzegowina). Flora, 162, 1-44. BRONK RAMSEY, C, 2003. OxCal 3.9. Oxford University. BULGARELLI, G.M., COUBRAY, S., PELLEGRINI, E., PETITTI, P., 1993. L’insediamento preistorico di Poggio Olivastro (Canino - VT). Considerazioni e prospettive. Bullettino di Paletnologia Italiana, 84, 435-480. CALDERONI, G., CARRARA, C., FERRELI, L., FOLLIERI, M.,

GLIOZZI, E., MAGRI, D., NARCISI, B., PAROTTO, M., SADORI, L., SERVA, L., 1994. Palaeoenvironmental, palaeoclimatic and chronological interpretations of a Late-Quaternary sediment core from Piana di Rieti

234

Environmental history in the Mediterranean basin In: A. Manfredini, ed. Le Dune, il Lago, il Mare, Una comunità di Villaggio dell’età del Rame a Maccarese, Firenze: Istituto Italiano di Preistoria e Protostoria, 2535. GIRAUDI, C., NARCISI, B., 1994. Ricerche paleoclimatiche in ambiente lacustre. Il contributo della Task Force “Ambiente Globale e Cambiamenti Climatici”. ENEA, RT/AMB/94/31, Roma. IVERSEN, J., 1941. Land Occupation in Denmark’s Stone Age. A Pollen-Analytical Study of the Influence of Farmer Culture on the Vegetational Development. Danmarks Geologiske Undersøgelse, II (66), 1-68.

(central Apennine, Italy). Giornale di Geologia, 56(2), 43-72. CAROLI, I., CALDARA, M., FIORENTINO, G., 2004. Integrating data from micro and macro-charcoal analysis of “Lago Battaglia” core (sud-est Italy). Abstracts of III International Meeting of Anthracology “Charcoals from the past. Cultural and Palaeoenvironmental implications”, June 28 – July 1, 2004, Cavallino (Lecce), Italy, 9-10.

CARRIÓN, J.S., SÁNCHEZ-GÓMEZ, P., MOTA, J.F., YLL, R., CHAÌN, C., 2003. Holocene vegetation dynamics, fire and grazing

in the Sierra de Gádor, southern Spain. The Holocene, 13(6), 839-849. CASTELLETTI, L., CARUGATI, M.G., 1994. I resti vegetali del sito neolitico di Sammardenchia di Pozzuolo del Friuli (Udine), Preistoria e Protostoria del Friuli Venezia Giulia e dell’Istria. Atti della XXIX Riunione Scientifica dell’Istituto Italiano di Preistoria e Protostoria, Trieste, 1990, 167-184. CELANT, A., 2000. Nuovi dati archeobotanici su ambiente e agricoltura nel neolitico del Lazio: un esempio dalla campagna romana. In: A. Pessina, G. Muscio, eds. La neolitizzazione tra Oriente e Occidente. Atti del convegno di Studi, Udine, aprile 1999. Comune di Udine. Edizione del Museo Friulano di Storia Naturale, 355-364. CONTI, A.M., MANCINI, B., PERSIANI, C., PETITTI, P., 1993. Ricerche nel territorio del Lago di Mezzano. In: A.A. V.V., Vulcano a Mezzano. Comune di Valentano, Museo Civico. COSTANTINI, L., 1981. Semi e carboni del mesolitico e neolitico della Grotta dell’Uzzo, Trapani. Quaternaria, 23, 233247. COSTANTINI, L., STANCANELLI, M., 1995. La preistoria agricola dell’Italia centro-meridionale: il contributo delle indagini archeobotaniche. Origini, 18, 149-244. COUBRAY, S., 1997. Analisi preliminare dei macroresti vegetali. In: A. Zarattini, L. Detrassi, eds. Casale del Dolce. Roma: Ministero dei Beni Culturali ed Ambientali, 274-286. DI GENNARO, F., 1986. Forme di insediamento tra Tevere e Fiora dal Bronzo Finale al principio dell’età del Ferro. Firenze: Leo Olschki Editore. FOLLIERI, M., 1973. Cereali del villaggio neolitico di Passo di Corvo (Foggia). Annali di Botanica (Roma), 32, 49-61. FOLLIERI, M., 1977-1982. Le più antiche testimonianze della agricoltura neolitica in Italia meridionale. Origini, 11, 337-344. FOLLIERI, M., MAGRI, M., SADORI, L., 1988. 250,000-year pollen record from Valle di Castiglione (Roma). Pollen et Spores, 30, 329-356. FUGAZZOLA DELPINO, M.A., 1996. Un tuffo nel passato. 8.000 anni fa nel lago di Bracciano, Roma. Roma: BetaGamma editrice. GIANNITRAPANI, E., PLUCIENNIK, M., 1998. La seconda campagna di ricognizione (settembre 1997) del progetto “Archeologia nella valle del Torcicoda”. Sicilia Archeologica, 96, 59-69. GIRAUDI, C., 2000. Le oscillazioni di livello del lago di Vico (Lazio) nell’ambito delle variazioni climatico-ambientali dell’Italia Centrale. Il Quaternario, 13(1/2), 21-30. GIRAUDI, C., 2002. Evoluzione ambientale tardo-olocenica dell’area comprendente il sito eneolitico di Maccarese.

LOWE, J.J., ACCORSI, C.A., BANDINI MAZZANTI, M., BISHOP, A., VAN DER KAARS, S., FORLANI, L., MERCURI, A.M., RIVALENTI, C., TORRI, P., WATSON, C. 1996. Pollen

stratigraphy of sediment sequences from lakes Albano and Nemi (near Rome) and from central Adriatic, spanning the interval from oxygen isotope Stage 2 to the present day. In: P. Guilizzoni, F. Oldfield guest eds. Palaeoenvironmental Analysis of Italian Crater Lake and Adriatic Sediments. Memorie dell’ Istituto italiano di Idrobiologia, 55, 71-98. MAGRI, D., 1999. Late Quaternary vegetation history at Lagaccione near Lago di Bolsena (central Italy). Review of Palaeobotany and Palynology, 106, 171-208. MAGRI, D., SADORI, L., 1999. Late Pleistocene and Holocene pollen stratigraphy at Lago di Vico, central Italy. Vegetation History and Archaeobotany, 8, 247-260. NARCISI, B., 2002. Tephrostratigraphy of the Late Quaternary lacustrine sediments of Lago di Pergusa (central Sicily). Bollettino della Società Geologica Italiana, 121, 211219. PELLEGRINI, E., 1993. Aspetti della metallurgia nel comprensorio del Lago di Mezzano e nella media Valle del Fiora dal Bronzo antico all’XI sec. a.C. In: A.A. V.V., Vulcano a Mezzano. Comune di Talentano, Museo Civico. PETITTI, P., MITCHELL, E., 1993. Dati preliminari sulla topografia dell’abitato sommerso del Lago di Mezzano. In: A.A. V.V., Vulcano a Mezzano. Comune di Valentano, Museo Civico. RAMRATH, A., 1997. Laminierte Sedimenten des Lago di Mezzano (Latium, Italien) – Limnogeologie und Rekonstruktion von Umweltbedingungen der letzten 34.000 Jahre. PhD thesis, Universitàt Potsdam, 115. RAMRATH, A., SADORI, L., NEGENDANK, J.F.W. 2000. Sediments from Lago di Mezzano, central Italy: a record of Late Glacial/Holocene climatic variations and anthropogenic impact. The Holocene, 10(1), 87-95. REILLE, M., 1992a. New-pollen-analytical researches in Corsica: the problem of Quercus ilex L. and Erica arborea L., the origin of Pinus halepensis Miller forests. New Phytologist., 122, 359-378. REILLE, M., 1992b. Pollen et spores d’Europe et d’Afrique du Nord. Laboratoire de botanique historique et palynologie. Marseille : URA CNRS 1152.

ROBERTS, N., STEVENSON, T., DAVIS, B., CHEDDADI, R., BREWSTER, S., ROSEN, A., 2004. Holocene climate, environment and cultural change in the circummediterranean region. In: R.W., Battarbee, F. Gasse and C.E. Stickley, eds. Past climate variability through Europe and Africa,. Dordrecht, The Netherlands: Springer,. 343-

235

L. Sadori, M. Giardini 362.

ROTTOLI, M., 1993. “La Marmotta”, Anguillara Sabazia (RM).

Scavi 1989. Analisi paletnobotaniche: prime risultanze. Appendice I. Bullettino di Paletnologia Italiana (Roma), 84, 305-315. ROTTOLI, M, 1999. I resti vegetali di Sammardenchia-Cûeis (Udine), insediamento del Neolitico antico. In: A. Ferrari, A. Pessina, eds. Sammardenchia-Cûeis. Contributi per la conoscenza di una comunità del primo Neolitico. Pubblicazioni del Museo Friulano di Storia Naturale, 41, 307-326. ROTTOLI, M., 2000-2001. Zafferanone selvatico (Carthamus lanatus) e cardo della Madonna (Silybum marianum), piante raccolte o coltivate nel Neolitico antico a “La Marmotta”? Bullettino di Paletnologia Italiana (Roma), 91-92, 47-61. SADORI, L., 2001. Holocene climatic change in central Sicily (Italy). Terra Nostra, 3, 181-186. SADORI, L. 2003 - Late-Glacial and Holocene fire and vegetation history from Lago di Mezzano, central Italy. XVI INQUA Congress, Reno, Nevada, 23/30 luglio, 193. SADORI, L., GIARDINI, M., 2004a. A new combined approach to detect Holocene environmental changes: the different roles of man and climate. 32nd International Geological Congress, 20–28 August 2004, Florence, Italy, Workshops: program with abstracts, 1499-1500. SADORI, L., GIARDINI, M., 2004b. The role of fire in the Mediterranean landscape: an intriguing question. Abstract of III International Meeting of Anthracology “Charcoals from the past. Cultural and Palaeoenvironmental implications”, June 28 – July 1, 2004, Cavallino (Lecce), Italy, 40-41. 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(1), 5-17. SADORI, L., NARCISI, B., 2001. The postglacial record of environmental history from Lago di Pergusa (Sicily). The Holocene, 11(6), 655-671. STOCKMARR, J., 1971. Tablets with spores used in absolute pollen analysis. Pollen et Spores, 13(4), 615-621. TOURING CLUB ITALIANO, 1989. Guida d’Italia. Sicilia. Milano. TUSA, S., 1992. La Sicilia nella preistoria. Palermo, II Ed. TZEDAKIS, P.C., 1994. Vegetation change through glacialinterglacial cycles: a long pollen perspective. Philosophical Transactions of the Royal Society of London Series B, 345, 403-432. WHITLOCK, C., 2001. Variations in Holocene fire frequency: a view from the western United States. Biology and environment: proceedings of the Royal Irish Academy, 101B(1-2), 65-77. Wijmstra, T.A., 1969. Palynology of the first 30 metres of a 120 m deep section in Northern Greece. Acta Botanica Neerlandica, 18(4),. 511-527. WIJMSTRA, T.A., SMIT, A., 1976. Palynology of the middle part (30-78 metres) of the 120 m deep section in Northern Greece (Macedonia). Acta Botanica Neerlandica, 25(4), 297-312.

236

Collapsed beams and wooden remains from a 3200 BC temple and palace at Arslantepe (Malatya, Turkey) L. Sadori1, F. Susanna1, F. Balossi Restelli2

Dipartimento di Biologia Vegetale, Università di Roma “La Sapienza”. E-mail: [email protected]; 2 Dipartimento di Scienze Archeologiche, Storiche e Antropologiche dell’Antichità, Università di Roma “La Sapienza” 1

Abstract

At Arslantepe in the final centuries of the Late Chalcolithic period (3350-3000 BC) there was evidence of a massive boost given to the system of centralising and redistributing goods. It was in this period, coeval to the Late Uruk phase in Mesopotamia, that a huge monumental public architectural complex was built. Because of its division into sectors with different functions and different architectural features (temples, stores, areas for discarding administrative materials, a courtyard, corridors), it may be considered to be the first known example of a “palace” to have been discovered in the whole of the Near East, with areas set aside for performing the main public religious and secular functions. Two temples were excavated, of which the second one, temple B, consisting of two small entrance rooms, a main big room, and in the adjacent corridor was found, with a large amount of big charred wooden beams, some posts and other wooden structures or objects in situ. More than two hundred samples of charcoal have been recovered in the temple B area and identified as wood of Alnus, Pinus, Juniperus, Populus, Ulmus, Fraxinus, Crataegus, deciduous Quercus. Also some monocotyledonous structures of reeds ascribable to Arundo, which were probably used for the roofs constructions, were found. Neither Alnus, probably preferred for its high trunks, nor Populus, probably locally abundant, are generally considered as a good timber for construction; Pinus is viewed as the best, but we have hypothesised that it was probably less available than the other trees in the area around the site, as it is not very abundant. The fact that oak was used very little for the construction of the palace and temple, is possibly due to the small dimensions of oak at that time, or other properties or meanings that we, today, do not appreciate.

Introduction Arslantepe has been bringing to light extraordinary remains of past prehistoric and protohistoric cultures of Eastern Anatolia. The site, north of the Taurus range, in the Malatya plain, today famous worldwide for its delicious apricots, has been under systematic excavation since 1961, by a team of the University of Rome “La Sapienza”, directed in the last decade by Prof. Marcella Frangipane (Fig. 1). The mound is approximately 15 km west of the banks of the Euphrates and has been occupied for many thousands of years, at least from the V Millennium BC to the end of the Neo-Hittite period (VIII cent. BC) (Frangipane 2004). Its long and complex stratigraphic sequence, that has built up approximately 30 metres of deposit, has been thoroughly investigated in these years, uncovering the continuity of occupation and the alternating cultural

Fig. 1. Location of the site of Arslantepe, Malatya (Turkey).

and political relations of the region over such a long period of time. In this Upper Euphrates region, some of the greatest civilisations of the Ancient Near East have come in contact – those from Caucasia, Mesopotamia and Anatolia - and here testimony has been evidenced of an early development of a centralised form of power, a form of protostate in many respects similar to the early state formations of Lower Mesopotamia (Frangipane 1997, 2002).

The Architectonic Remains Majestic architectonic structures dated to the end of the IV millennium BC (3350-3000 BC, period VIA at Arslantepe, or LC5 in the Santa Fé chronology – Rothman 2001) have come to light in the last two decades (Frangipane 2001) and, through a long and painstaking work of data collecting, material culture analysis and architectural reconstruction, the Italian Archaeological Expedition has been able to reconstruct a complex palatial organisation, most evidently run by a leading class of high status individuals, that not only had political control over the rest of the community, but also evidenced economic centralisation of goods (Frangipane and Palmieri 1983). A large multifunctional public building has been as yet only partially brought to light, over an area of approximately 2000 m², thus suggesting a possibly much larger structure, the most inner and possibly important parts of which might still be unknown (Fig. 2). A long corridor leads right into the building, where temples, storages, administrative areas, rooms of representation, courtyards and other less well characterised areas, indicate the variety of activities that

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

L. Sadori, F. Susanna, F. Balossi Restelli

Fig. 2. Arslantepe, plan of the palatial complex of period VIA.

must have been taking place in the structure. This is probably the most ancient “palace” known today. Thick walls, in mud brick, plastered in white and decorated, have been at times found standing for more than 2 metres. Impressed concentric rhomboidal “stuccos” decorate the long corridor and also the entrance to the inner temple, and red painted stylised human and animal figures are found both at the entrance of the storages and at the far end of the corridor. In the palace, amazing finds have been discovered, that well illustrate the degree of prestige of the leading class and the economic control it had. Copper-arsenic alloy swords and lances have been found, from this period, in which metallurgical production is still at its early stages of development. Enormous quantities of sealings (cretulae) were discovered in the storages and in special dumping places, where they were collected together as probable documentation of economic transactions, which have hinted to the centralised economic system of the palace (Frangipane 2003a). The entrance to the palace led directly into the long corridor, the first part of which (A209) was probably an open air space, moving slightly uphill towards the most inner part of the building (fig. 2). The further away part of the corridor (A796), instead, was probably covered, at least in a second phase, with a ceiling built with wooden beams, straw and mud, as traditional houses in the area still are today. All the wooden beams collapsed on the floor during the conflagration. The walls of this part of the corridor furthermore had been recoated in a second phase of life of the structure, by actually rebuilding a proper facade to the

wall, with wooden posts and mud bricks, which were then covered in mud and plastered. This re-covering of the wall, which also had the effect of narrowing this part of the corridor, covered the red ochre and black paintings depicting a cart pulled by two oxen. Just east of this part of the corridor is Temple B, the entrance to which though, is not from the corridor, but from a courtyard on the opposite side. This temple has a very distinct plan, in all similar to that of the other temple of the palace, temple A. Composed of two rows of rooms, the building has an undivided large central room (A450), where the ritual and ceremonial activities most probably took place, and a small series of access rooms to one side (Frangipane 2003b). The central room is 5.7 m wide and 12 m long, thus a particularly large space. Amongst the lateral rooms the central one (A809) is the actual entrance to the temple and, from this, two windows (A810 and A811) look into the large central room. It is possible that the windows indicate that proper access to the central room was not permitted to all people and these windows might have been used to pass offerings or the like. A great quantity of ceramic was lying on the floor of central room A450, composed of jars of various sizes, a few bowls and some fruitstands. To enter the most sacred part of the temple (where altars, plastered depressions and niches suggest the cultic activity), small room A800 (3 x 3 m) is to be crossed, making a turn to the left for the large room A450. At the entrance of room A800 are some of the impressed rhomboidal figures as those of the corridor, but here these are also decorated in red paint. On the opposite side of en238

Collapsed beams and wooden remains from Arslantepe

Fig. 3. Alnus sp., transverse surface.

Fig. 4. Alnus sp., radial surface, scalariform perforation plates.

Fig. 5. Pinus sylvestris/montana gr., transverse surface.

Fig. 6. Pinus sylvestris/montana gr., radial surface, large fenestriform pits.

trance A809, a flight of stairs probably took up to the roof (A812 and A813). In the temple too, as in corridor A796, the wooden beams of the ceiling and all the other wood contained in the structure had collapsed on the floor on occasion of the devastating fire. Some of these were in good preservation, but others fell apart at touch. The wooden beams are mostly lying directly on the floor, with the mud and earth pack from the ceiling above, and the fallen bricks and other building debris covering all. Only in A800 the beams were found throughout the whole filling suggesting a possibly different construction from that of the other rooms. The sequence of the debris suggests that at least temple room A450 must have been single storey, as only one single stratigraphic “pack” of ceiling material was identified, above the wooden beams (Alvaro 2004). The height of the

temple is difficult to reconstruct, but considering the thickness of its walls and comparing these to ethnographic cases, we could argue that the temple might have been as high as 8 m (Alvaro personal communication). Ethnographic cases indicate a height of 2-3 m in normal dwellings with a wall thickness of about 60 cm (Koyunlu 1982, p. 262); here walls are more than 1 m thick. As for corridor A796, this might have been lower, at least in its second phase, as the vertical wooden posts recovered, that we suppose were holding the roof, do not appear to have reached such a height. 29 radiocarbon dates from the period VIA palace have dated it to the end of the IV millennium BC, but quite enigmatically, there is an evident discrepancy between the dates taken from samples within temple B and the corridor (A796) and the rest of the analysed charcoals (Di Nocera 239

L. Sadori, F. Susanna, F. Balossi Restelli 2000b). In fact, 3 samples from A796 and 5 from the temple, give earlier dates compared to those of the rest of the palace (Di Nocera 2000a; Di Nocera 2000b; Kuniholm 1996). Two alder (Alnus) beams of A450 suggest an age interval between 3506 and 3106 BC calibrated (2σ) (Alessio et al. 1983; Calderoni et al. 1994) (while a piece of charcoal of ash could be slightly younger), whilst two juniper beams of A800 give an age between 3650-3389 BC calibrated (2σ). The poplar beam of A796 indicates an interval of 3906-3543 BC calibrated (2σ) and the two pine beams from the same room date between 3708 and 3534 BC calibrated (2σ), thus partly matching the juniper beams of A800. Dendrochronological analyses (Kuniholm 1996) though indicate an age of 3374 ± 30 years BC (2σ) for the juniper timber of A800, thus slightly moving down the radiocarbon date. Whereas A800 and A450 do have a clear overlap within their dates and wood from those rooms might in fact be contemporary, wood from A796 does indeed appear to be older. This difference finds as yet no justification in the architecture since there are no stratigraphic indications that the corridor might have been built before the temple. Stratigraphic and archaeological evidence furthermore clearly indicate that the chronological development of the rest of the VIA period palace is dated to the interval 3300-2900 BC. Careful stratigraphic investigations are underway to understand whether this latter part of the palace might have been built later. It might also be possible though that wood for these structures, that had greater dimensions and importance than the rest of the building, had been reused and taken from earlier buildings. This is in fact a common practice in the Near East, especially so if the building has some particular symbolic importance. Di Nocera (2000b) has suggested that the timber from each room (A800, A796, A450) came from batches that had perhaps belonged to single pre-existing buildings. Hopefully the scheduled AMS radiocarbon dating of monocotyledon structures (cfr. Arundo) used for the roof construction of A450 together with soil will help precise the date of the structure. The end of the palace’s life is signed by a great fire, that burnt down all the structure; a disaster for the people, but a great help for the archaeologists, who were thus able to discover a nearly intact context, with in situ materials and a lot of charred plant remains, that helped both in reconstructing the environment of that period, in understanding the building techniques used for the palace and in calculating its date (radiocarbon analysis). The collapsed debris furthermore, strongly helped preserve the structure to our discovery.

ity of data. When excavating within a structure, all single objects found are triangulated, in order to be able to reconstruct, three dimensionally, how the building collapsed, as well as its past appearance. In this process do not only undergo whole and fragmentary material finds, but also substantial botanical, faunal or structural remains. The stratigraphic sequence of the collapse within a room is furthermore kept well distinct, by using specific conventional signs, in order to immediately recognise, without having to look at the list of levels at which objects are found, whether they were lying just above the floor or amongst the ceiling collapse, or within the collapse of the walls. In situ materials, those lying directly on the floor, are furthermore mapped and named one by one, with a specific tag, so that these too are immediately recognised. This process was of course also used when excavating the 3000 BC palace and, in this specific case, has permitted to position all botanical remains found in the structure. All were sampled and taken back to Rome, where they were then studied in laboratory. The recovered charcoals, generally well preserved though fragmented, where sorted according to their peculiar anatomical features with the help of a reflected light microscope, at different magnifications and, for detailed wood identification, a Nomarski microscope (phase contrast microscope with differential interference contrast) was used. When necessary, the scanning electron microscope was used as well. In total, more than 200, often very large in size, charcoal remains have been identified, generally analysing all the available charred wood fragments of each sample (at times 10 or more). The fractured beams gathered from the floor ranged in size from 10 cm to 4 m. For charcoal identification various anatomical descriptions and diagnostic surface photographs were employed (Fahn et al. 1986; Greguss 1955, 1959; Schoch et al. 1988; Schweingruber 1978, 1990).

Results 201 samples of charcoals have been recovered in the temple B area and in the covered part of the corridor (A796) and the following woody taxa have been found and identified: Alnus, Pinus, Juniperus, Populus, Ulmus, Fraxinus, Crataegus, Quercus. Their visible and characterising features will be reported of afterwards. Mostly this analysis led to identify timber only at the genus level, but sometimes some species are grouped in the same woody taxon. This is the case of Pinus sylvestris/montana, which includes species of not Mediterranean pines, and also of deciduous Quercus, whose wood can be differentiated from the evergreen one. It has thus not been possible to say which species was present for each identified taxonomic group, but a comparison with the taxa present today in the area might give us an idea of the possibly exploited species. Alnus sp. (alder) charcoals show, along the transverse sur-

Materials and Methods Great attention is given during field investigations to the documentation process, in spite of the fact that this slows enormously the whole excavation. This has permitted, over the years, to collect an exceptional quantity and qual240

Collapsed beams and wooden remains from Arslantepe

Fig. 7. Juniperus sp., transverse surface.

Fig. 8. Juniperus sp., radial surface, cupressoid pits.

Fig. 9. Populus sp., transverse surface.

Fig. 10. Populus sp., radial surface, 2-3 (4) rows of simple large pits.

face, diffuse pores, more or less densely packed in radial files and groups (Fig. 3). Scalariform perforation plates (Fig. 4) even with more than 20 bars are often visible inside the pores, at phase contrast, with differential interference contrast and scanning electron microscopes. On the tangential surface rays appear uniseriate, and are up to 25 cells high. In aggregate rays bi- to three-seriate rays are often present. On the radial surface, rays appear homogeneous and scalariform perforation plates are clearly visible in the pores. Pinus sylvestris/montana gr. (pine) charred woods on the transverse surface always show distinct growth ring boundaries; cells with maximum wall thickness are in general positioned a little before the growth ring boundary, with the early/latewood transition generally abrupt, and abundant resin canals (fig. 5). The rays appear uniseriate and 1-12 cells high on the tangential surface, pluriseriate

and higher when containing resin ducts. On the radial surface, longitudinal tracheids with uniseriate pits and heterocellular rays are generally shown. Parenchyma cells have generally one, rarely two, large fenestriform its per crossfield (fig. 6). In latewood, pits are large. Ray tracheids with dentate walls are present. Juniperus sp. (juniper) charcoals show distinct growth ring boundaries on the transverse surface. Resin canals are absent and there is a very gradual transition from early- to latewood (fig. 7). On the tangential surface rays appear low, with a height generally from 2 to 7 cells. On the radial surface tracheid pits are uniseriate. Rays are homocellular, composed only of parenchyma cells. There are 1 to 2 cupressoid pits per cross-field (fig. 8). Populus sp. (poplar), on the transverse surface (fig. 9), has diffuse- to semi-ring-porous wood. Pores are solitary, in groups or in short radial files of 2-3, rarely clus241

L. Sadori, F. Susanna, F. Balossi Restelli

CHARCOAL A 796

A 450

A 800

A 805

A 809

A 810

Total weight

(g)

(g)

(g)

(g)

(g)

(g)

(g)

Alnus sp.

4492

8704

6

98

93

13393

45.72

Pinis sylvestris/montana gr.

4074

965

255

5294

18.08

5245

17.91

3991

13.62

947

3.23

TAXA

Juniperus sp. Populus sp.

968

55

4841

261

88

2355

158

61

226

939

8

Ulmus sp.

223

%

Fraxinus sp.

257

257

0.87

Crataegus sp.

146

146

0.49

3

18

0.06

6

0.02

29297

100

Arundo sp.

15

deciduous Quercus

5

1

9554

Total

12485

5938

336

668

316

Table 1: Arslantepe, period VIA, distribution of charcoal.

A 450: charcoal

A796: charcoal

TAXA

g

Alnus sp.

8704

69.72

Populus sp.

2355

18.86

Pinus sylvestris/montana gr.

965

7.73

Fraxinus sp.

257

Crataegus sp Juniperus sp.

g

%

Alnus sp.

4492

47.02

Pinus sylvestris/montana gr.

4074

42.64

Populus sp.

968

10.13

2.06

Arundo sp.

15

0.16

146

1.17

deciduous Quercus

5

0.05

55

0.44

Total

9554

100

3

0.02

12485

100

Table 4: Arslantepe, period VIA, charcoal of corridor A796.

Arundo sp. Total

%

TAXA

Table 2: Arslantepe, period VIA, charcoal of A450.

tered. On the tangential surface rays appear uniseriate, from 10 to 15 cells high. On the radial surface, rays are homogeneous; in crossing fields, 2-3 (rarely 4) rows of simple large pits are arranged on each ray cell (fig. 10). The latter are the features which allow to distinguish between Populus (poplar) and Salix (willow). Ulmus sp. (elm) charred woods show a porous ring on the transverse surface. There are groups of 2-4 pores in more or less tangential bands, in the latewood. On the tangential surface multiseriate rays are 4-5 cells wide and less frequently 1-3 cells wide. Ray height is of 30-60 cells. On the radial surface rays appear homogeneous; simple perforation plates and distinct spiral thickenings are visible in the vessels. Fraxinus sp. (ash) has ring porous charcoals visible on the

A800: charcoal g

%

4841

81.5

Ulmus sp.

939

15.8

Populus sp.

158

2.7

5938

100

TAXA Juniperus sp.

Total

Table 3: Arslantepe, period VIA, charcoal of A800.

242

Collapsed beams and wooden remains from Arslantepe tana gr. (8 %) and Fraxinus sp. (2 %) were used as well, both for ceiling beams and for smaller items. The use of Juniperus sp. (0.44 %) was very limited in this room. The longest intact piece of roof beam (4 m) found on the floor is of pine, recovered some metres away from the remnants of a number of superimposed and crossed alder beams, each piece of which rarely exceeding 2 metres. Interesting to note is the identification of Crataegus sp. (1.17 %) charcoal lying on the floor, in a corner of this central room of the temple. In the small, almost square, room A800, mainly Juniperus (81.5 %), with some Ulmus (15.8 %) and Populus (2.7 %) charcoals were found (table 3). Juniper beam remains reach a length of 1 m, but must have been longer (room is 3 x 3 m). In corridor A796, next to the ceiling beams, were also the vertical poles that had been used when the north-eastern wall with the paintings had been covered up and rebuilt, but only a few of the charcoal remains from within the holes left by the posts have been as yet analysed and it is difficult to tell which of the posts lying on the floor were the vertical posts and which the horizontal ceiling beams. Only three of the analysed samples surely belong to such posts and have been identified as Populus and Quercus. The fallen beams found on the floor of A796, the 12 m long and 3 m large corridor, resulted of Alnus (47.02 %), Pinus sylvestris/montana gr. (42.64 %), Populus (10.13 %) and Quercus (0.05 %) trees (table 4).

transverse surface. There are scattered, solitary or radially paired, pores in the latewood. On the tangential surface are mainly bi-seriate and tri-seriate rays (average ray height: 10 to 15 cells). On the radial surface perforation plates are simple, rays are generally homogeneous, rarely heterogeneous. Crataegus sp. (hawthorn) charred woods have diffuse and regularly distributed pores on the transverse surface. On the tangential surface, rays are generally biseriate to triseriate, 5 to 15 cells high and are generally almost homogeneous on the radial surface. Perforation plates are simple. Fine spiral thickenings occur. Charred charcoal of deciduous Quercus (oak) has porous rings visible at naked eye on the transverse surface, with a flame-like distribution of groups of pores in latewood. In the tangential section, rays appear uniseriate and multiseriate. Multiseriate rays are 0.5 to 1 mm wide (up to 30 cells) and 1-5 cm high. On the radial surface simple perforation plates in the vessels and homogeneous rays are visible. Monocotyledon structures (atactosteles) possibly ascribable to cfr. Arundo, were also found. The wood identification of the more than 200 well recovered, but mainly fragmented, single charred remains has evidenced differences in the use of timber in the different rooms of the complex. Most charcoal remains were very large in size and their positioning has also permitted, once the wood had been recognised, to reconstruct single beams, parts of which had broken with the fall and had been collected as separate samples. These results are mapped in Fig. 11. Each large sample was generally constituted by a number of fragments of only one woody taxon confirming that the recovery of the charcoals was very precise and accurate, and the stratigraphic and surface sampling meticulous. Only two samples in fact resulted composed by wood of two taxa. Some wood was clearly used for the ceiling beams, whilst smaller pieces, found in the proximity of doors and in the windows of the temple, were possibly window and door sills or frames. Obviously, next to these structural parts were also elements of furniture; these, in most cases were in a very bad state of conservation, since more fragile and covered by the collapse of the whole building, and thus have been very difficult to collect, document and identify. Table 1 and figs 11 and 12 give a summary of all the results of the analyses carried out on the samples. These come from rooms A809, the entrance to the temple, almost destroyed by a uccessive pit, A805, the passage-way from A800 to the huge room A450, A810, a window sill between A450 and A809, and A796, the central corridor of the palace. Alnus (45.72 %) is the most abundant charcoal, followed by Pinus sylvestris/montana gr. (18.08%), Juniperus (17.91 %) and Populus (13.62 %) (table 1). In the main, widest room A450, about 12 m long and 5.7 m wide, the most abundant charcoals (table 2) were of Alnus sp. (70 %) and Populus sp. (19 %). Pinus sylvestris/mon-

Discussion Archaeological and ethnographical examples of architecture in the Near East shows that techniques and materials have not changed much in the millennia. Still in recent years traditional architecture followed very ancient rules and used the same materials that can be found in the excavated sites of the same region. This helps greatly in reconstructing the archaeological contexts. Of course, in this very case, climatic and environmental change has to be taken into consideration in evaluating the use of specific timbers in today’s constructions compared with those of the past, but we can quite securely hypothesise that techniques, dimensions and general materials have not undergone major changes. Wood was and is still used in buildings of the Near East for different structural purposes; roof and roof beams are of course the most frequent, but we see it used, horizontally or vertically, within mud brick and stone walls, as vertical posts sustaining roofs, as window and door frames or the doors themselves. And this is to nominate just the most common. Mats or reeds are often used in building the ceilings of the houses, above the beams, to contain the pack of soil used to build the actual roof (Aurenche 1981, p. 80). Many ethnographic studies have testified and recorded these construction systems, also underlining the type of wood preferred in the different regions (Hall et al. 1973; 243

L. Sadori, F. Susanna, F. Balossi Restelli

Fig. 11. Arslantepe, plan of temple B and corridor A796 with the position of the main recognized charred wood samples.

Fig. 12. Graph indicating the quantity (weight) of the different taxa for each of the investigated rooms.

244

Collapsed beams and wooden remains from Arslantepe 2002, p.149) and at Aşvan during the Early Bronze Age (Willcox 1974, p.126) there is testimony of this tree, but not in large quantities. Might its potentially greater length be the reason of its large use at Arslantepe? In fact, in none of the known contemporary sites have structures this size been discovered. The two longest remains of wooden beams that we have been able to measure on the floor of room A450 are of Alnus and measured 3.04 and 2.14 m, well below the 5.7 m width of the room. We should however consider that the beams probably broke in two or more pieces during the collapse of the roof and we thus do not know their initial size. We could also hypothesise that the abundant use of alder could be due to a possible greater value at the eyes of the inhabitants of the site, as its finding in the slightly later royal tomb might testify. In that case in fact the wooden plank on which the “chief” was laid was in alder (Frangipane et al. 2001, p.133). As for the possible species utilised, we know that Alnus glutinosa, a riverside tree, is nowadays growing in eastern Turkey, in mountain forest. Being at present very rare, it is included in the noble hardwood species and introduced in a program for the preservation of biodiversity (Alan 2002). Black alder is a fast growing tree with four natural subspecies in Turkey, one of them more important nowadays for timber production in the Eastern Black Sea Region (Atasoy et al. 2002), and it can grow up to 25 m in height along water ways. The other tree used quite abundantly in the palace construction is Pinus (18.08 %). This is mainly found in the corridor, whilst only one pine beam, broken in two long parts (1.3 m and 2.7 m), is present in the temple. Quite the opposite, in corridor A796 Alnus is not as common as in A450. The difference in use of these two trees might be due to the chronological gap between the wood of the corridor, belonging to a second phase of construction, and that of room A450. Pine was supposedly privileged because of its tallness and especially straightness and in fact the Assyrian kings mostly used it for their majestic palace and temple doors, as well as for roof beams (Postgate 1992, p. 180); in the case of Arslantepe, certainly height was not a necessary attribute for the horizontal beams of the roof of A796, whilst we do not know how high the vertical posts were. In two posts holes though charcoals of poplar were found, and in a third one deciduous oak, thus suggesting that here too, it was not height the ultimate cause for which pine had been chosen. Mountain pine (Pinus sylvestris/montana gr.) timber is easy to process and currently employed for supporting posts in buildings. Pine is not common amongst the finds of sites in these regions. The Aşvan settlements in fact have no evidence of pine neither in the Chalcolithic nor in the Early Bronze Age levels. This might be due to the fact that pine nowadays is only found in the mountains, thus rather far from the known archaeological sites. Arslantepe, compared to Aşvan, is only at a slightly higher altitude, thus the importance of the building might be mostly to blame for the use of a more precious and not so readily available wood. Nowadays Pinus nigra and Pi-

Kuban 1970; Koyunlu 1982). The vast majority of large wooden beams found in the Arslantepe temple and corridor was lying directly on the floor. According to their position and dimension these were most probably horizontal roof beams. This has been taken to indicate that the building collapsed under fire, thus the ceiling fell before the walls did. On top of this is a thick burnt pack of earth, with a great quantity of charcoal and traces of mat-like vegetal material. The latter were very difficult to sample since they mostly disintegrated at touch, but two samples have been analysed and identified as a reed, very similar in structure to Arundo. Thanks to the abundant ethnographic cases, this kind of roof construction is quite easy to reconstruct: the wooden beams sustain a layer of thin branches (sometimes preceded by a mat) (Kuban 1970, p. 182), on top of which is a thick layer (approximately 20 cm) of earth or clay, which forms the actual roof closure (Naumann 1971, p. 154). This can end with a thin layer of very fine and plastered clay, which gives a better impermeability to the roof. The smaller charcoals analysed in this pack of soil, which we hypothesise to be the layer of branches and mats sustaining the earth of the roof, turned out to belong to the same species of the wooden beams on the floor. This brings to hypothesise that most of the trees from which the beams were made were probably local, since it is difficult to imagine that wood brought from afar would have been brought together with its smaller branches (Aurenche 1981, p. 154). Dimensions of the rooms are crucial in understanding whether these needed central vertical posts to sustain the ceiling or whether the perimetric walls were sufficient to hold up the roof. Most ethnographic studies in the regions around Malatya agree in saying that 3, 3.5 m is the maximum length that one can find in which rooms need no central vertical posts (Hall et al. 1973, p. 257; Kuban 1970, p. 182; Koyunlu 1982, p. 262). The small lateral rooms of temple B and the corridor are below this size, but the central room A450 of the temple is well above, measuring 12 x 5.7m. Yet, on the floor of the room, no traces have been identified that would bring to suggest central vertical posts. In what remained of the walls too, were no traces of possible vertical posts that might sustain some alternative cover, as those hypothesised by Margueron (1992) for later period palaces in Mesopotamia or by Naumann for the Neo-Hittite buildings at Zincirli (Naumann 1971, p. 116). Amongst the species recovered from the charred beams in this study, only some species of alder and pine are supposed to exceed six metres in height. On the contrary, poplar and oak rarely span, in eastern Anatolia, much more than 5 m (Willcox 1992, p. 5). Peculiar is the predominance of Alnus (45.72 % in the temple area) in the temple structure and, above all, amongst the trees used for the beams of A450 (69.72%). Alder is not as well testified in Near Eastern archaeological contexts as the other trees found in the temple. At Tell Shiukh Fawqani (Middle Euphrates, Syria) during the Late Uruk (Pessin 245

L. Sadori, F. Susanna, F. Balossi Restelli nus sylvestris trees (Willcox 1992, p. 8), both with wood referable to Pinus sylvestris/montana gr., are found on the Taurus mountains, few kilometres from Arslantepe. Very particular, since limited only to the small room A800, is the finding of Juniperus. This too is commonly identified in archaeological contexts of the Near East, but mostly in the Levant and Central Anatolia and at earlier sites (Beidha, Çatal Höyük, Can Hasan III). Juniperus can grow quite high and the tallest modern trees in Turkey have been measured up to 40m, even though in the archaeological context of these Taurus regions scholars argue that it would not reach the height of Pinus (Miller 1999, p. 16), but would remain quite short. This would be confirmed by the fact that in temple B it was only used for the smallest of the rooms (3 x 3 m). Two juniper species, J. oxycedrus and J. excelsa are still found in the region of Arslantepe, at an elevation between 800 and 1500 m a.s.l. (Kaya et al. 2001, p. 134) even though nowadays J. excelsa is an endangered species, included in the list of forest species for in situ conservation of genetic diversity in Turkey (Alan 2002). Juniper shows different specific features but unfortunately wood can not be identified at a specific level. Juniperus oxycedrus timber has high resistance to decay, and big trunks provided in the past very long lasting frames (Giordano 1971); softness to work further justifies its use in architecture. Often, in antiquity (see the ancient fraud of 4000 years ago in Egypt; Corona 1991, p. 214), juniper was falsely disguised as cedar, a very precious wood, since only coming from the Lebanese coast. Caution should be thus used when interpreting historical founts, but this also tells us that the use of juniper in architecture might be probably more common than what appears. Annals of Sargon II of Assyria, even though many millennia after the Late Chalcolithic period, say that Juniperus was used for roof beams in the mountainous regions north-east of Assyria, thus testifying for its use in architecture in Eastern Anatolia, and Salmanassar III imports Juniperus for his palace from the Amanus, thus again Turkey, and both Sennacherib and Assurbanipal use it for their doors and columns (Postgate 1992, p. 180). Strangely enough though, Juniperus is not commonly found in East Anatolian Chalcolithic or Early Bronze Age sites; some comes from Aşvan Kale and Çayboyu (Willcox 1974, p. 126; Willcox 1992, p.14). It is indeed possible that, being there in the region availability of other species more resistant and adapt to architectural use, these were preferred instead of Juniperus. Even the juniper trees that were eventually used in temple B were particular, since long-lived (Kuniholm 1996, p. 330) and thus most probably expressly chosen among the most resistant and large junipers. It is not yet given to us to know whether this very specific and delimited ceiling of room A800 distinguishes itself so much from the rest because this room had a specific architectonic plan, thus for structural reasons, or whether it was built, or rather, rebuilt, in a different moment. Populus is found in a certain quantity in the central temple room, but it is mostly distributed around the edges of the

room, thus possibly used for other structural parts rather than for the roof beams themselves. In corridor A796 instead, Populus is surely used for the vertical posts and/or for the roof beams. Even though in Europe it is nowadays not considered to be a good timber for structures, being used only for wooden partitions and domestic tools of scant durability (Giordano 1971), both ethnographic and archaeological cases in the region indicate Populus (at times undistinguished from Salix) as the most common source of structural timber; this is still evident today in Aşvan (Hall et al. 1973, p. 257), just north east of Malatya, but also during the Neolithic period at Mureybet, where poplar is used both for beams and vertical posts, at Cafer Höyük (40 km from Malatya), where Salicaceae (comprehending thus both Populus and Salix) wood is the most used in architecture (Willcox 1991a, p. 143; Willcox 2002, p. 143), and in the Chalcolithic and Early Bronze Age at Tell Brak, where it was used for roof beams (Mallowan 1947, p. 15). Poplar was also found at Mefesh (Aurenche 1981, p. 81), at Hacinebi (where poplar is not distinguished from willow (Miller, 1994, p. 170), at Korucutepe (van Zeist, BakkerHeeres 1975, p. 245), and at Tulintepe (Aurenche 1981, p. 81). Indeed in most Anatolian and north Mesopotamian sites, Populus beams appear to be the most common (Willcox, 1992, p. 5); in all these areas this tree, growing exclusively along rivers, (Willcox 1991a, p. 143) is in general nowadays still available. In Assyrian times too, we know from various texts, that Populus was the main source for roofing and was often deliberately cultivated too (Postgate 1992, p. 183). This tree, together with Quercus, would appear to be preferred to the others available because of its hardness, straightness and relatively fast growth rate (Hall et al. 1973, p. 257). Furthermore, it is easily peeled of its bark, it dries without cracking and can also be used without seasoning (Aurenche 1981, p. 78). Its resistance to flexion, traction and pressure is not much lower than that of Pinus, which is considered one of the best trees for construction, but is not very abundant in these regions during the Chalcolithic and Early Bronze Age. Quercus is the other wood, decidedly the best wood for building, from our point of view, among those recovered in temple B (extremely rare, only 0.02 % of the samples examined in this paper), of which we have good testimony for its use in ancient architecture in the region (Willcox 1991b, p. 122; Willcox 1992, p. 5). All the sites that evidenced Populus or Salicaceae, in fact have testimony of deciduous Quercus, used in architecture, as well as for fuel (hearths), as also other sites in Eastern Anatolia, as Çayönü in the Prepottery Neolithic (van Zeist 1972, p. 16). Quercus is more resistant and robust than Populus and would thus be preferred to the first in architecture, but generally Populus, being a riparian tree, and thus probably found nearer to the sites (that were in general close to permanent watered places), and having a faster growth rate, was used more abundantly. Last of all, another possible reason for which poplar was more used than oak is given by the fact that in the Aşvan area deciduous oaks, 246

Collapsed beams and wooden remains from Arslantepe probably even of considerable age, did not reach, during the Chalcolthic and Early Bronze Age, a height of more than 2.5 m (Willcox 1974, p. 122), thus resulting too small for many constructions, especially those of the size of the one analysed here. Amongst the wood found in temple B and in corridor A796, we have not yet mentioned Fraxinus and Ulmus. Both are only minimally used; one single beam has been identified for each of the two species and the rest are small charcoal samples that might also derive from other kind of objects or might have broken off the two beams, as their position could, in most cases, suggest. Ash (Fraxinus) has very high mechanic resistance, in particular to dynamic inputs. The small pieces can be used for small posts, fuel or basketry. Elm (Ulmus) is generally used for manufacts to be put in water, wheels, and furniture (Giordano 1971). Use of these two trees in ancient architecture is testified, even though not as abundant as for the species seen above (Willcox 1992, p. 9). As usual, the Aşvan sites have them, as well as Cafer Höyük and Korucutepe, amongst the nearest to Malatya. Of all these species used in the construction of temple and corridor, some differences in choices have been evidenced, that might be linked to the properties and significance of the different woods or to their availability. First of all, even though containing the same species, there is an evident difference in their proportions between species used in the corridor and those of the temple. We have already pointed out that this might be due to the chronological difference of these two rooms, and to the fact that the wood, reused from other earlier structures, might have been taken from two distinct contexts. The position of the beams on the floor of room A450 leaves no doubts about the fact that they were put transversally to the structure, element this which is furthermore quite obvious, due to the great length they would have to cover in the other direction (12 m). Possibly, the beams in room A800 were set in the same direction too, as the majority of the juniper samples are oriented that way. Wood is quite obviously absent from rooms A812 and A813 as that was the space for the stairway going up to the roof. It was thus open at the end, with no ceiling, and the steps themselves were probably made in mud brick. One small juniper sample found on the altar along the long western side of room A450, together with the small ash sample on the same altar, might belong to some objects that were positioned there, rather than to some structural part of the temple. Next to the altars in A450 and in the corners of the room were many ceramic vessels, possibly used for offerings. In the corner, close to the a group of ceramics, hawthorn (Crataegus) charcoals have been found. Interesting is to note that in temple B, only Populus appears to be used for window sills. All poplar samples from the temple furthermore are found, as has been noted, in the peripheral areas and two pieces have been found just next to the two niches in the northern wall of A450. Clearly, this indicates a specific choice not to use poplar for the roof beams, but instead for other small structural elements.

The fact that the two windows looking into the temple (that of A811 was too ruined to be sampled) had a wooden windowsill or frame brings us to hypothesise that these might have had a wooden “window”, that could be closed. It might well be that in some moments, view to the inside of the temple was closed to all who did not have direct access. For the entrance to the temple (A805) unfortunately we have found no direct evidence of a door, but the presence of burnt material certainly cannot rule out this hypothesis.

Conclusions As we have seen, the most commonly used woods in the temple and corridor structure are, in decreasing order Alnus (by weight highly more abundant than the other woods), Pinus, Juniperus, and Populus. Interestingly, we have pointed out that neither Alnus nor Populus are generally considered, nowadays, as a good wood for construction; Pinus is viewed as the best, but we have hypothesised that it was probably less available than the other trees in the area around the site, reason for which it is not very abundant. Populus, still today the most common tree surrounding the mound and in the Late Chalcolithic too attested in the area, was not surprisingly used for the palatial building, but the fact that it is not the most commonly present suggests that it was either considered to have lower properties than the other trees, had minor value at the carpenters eyes, or didn’t reach the necessary length for such a large and important structure. Testimonies of the use of poplar for buildings is abundant in contemporary sites of the region and neighbouring areas, we thus tend to think that it must have been to the particular character of the palatial building, its size and significance, that the more limited use of poplar is to blame. All the woods identified in the Arslantepe temple and palace corridor have also been evidenced in more or less neighbouring sites. This is evident both in earlier sites (Cafer Höyük, being the nearest, but also Çayönü, Can Hasan, and Tulintepe) and in IV-III millennium BC settlements (Aşvan Kale, Çayboyu, Taşkun Kale, Taşkun Mevki, and Korucutepe, are the nearest, and Hacinebi is further south). The main difference between these sites and the Arslantepe palace lies in the relative abundance of the single taxa, especially in the overwhelming abundance of alder, very rare in contemporary sites of the region. We have suggested that the reason for which the latter was privileged might have been the greater height it could achieve (more than 10 m), making it thus preferential for the construction of such a large roof, or its possible particular significance. Only the presence of Pinus at Arslantepe appears to be rather singular and distinctive compared to the other sites; in pollen diagrams from east Anatolian lakes this taxa is present but not abundant; no samples have been found in the early levels of occupation in the Aşvan sites (Çayboyu, 247

L. Sadori, F. Susanna, F. Balossi Restelli prefer humid places (Willcox 1992, p. 8; Postgate 1992, p. 185), thus all are found in a similar environment. Arslantepe not only is not far from the Euphrates river, but at its feet runs a stream, and most probably it did in past times too. We could thus easily imagine that these particular species must have been readily available just next to the site and selected according to their technological features by the carpenters of Arslantepe. The fact that some of the juniper samples from A800 still have evidence of the bark, furthermore, argues in favour of the presence of the latter near the site, since we would imagine that beams brought from outside might have been at least partially worked and peeled (maybe also seasoned?) before transport.

Aşvan Kale, Taşkun Kale and Taşkun Mevkii), nor in the Altınova plain. This might be due to the greater importance of this building compared to all architectonic finds of the other sites; more difficult would be to blame a greater altitude and vicinity to the mountains of Malatya. It is furthermore probable that the actual pine trees were not exactly at the edge of the site, but rather along the slopes of the mountains; we should thus imagine medium distance transportation of this wood to the site. A surprising, almost complete, lack is oak charcoal. Well before the 4th millennium BC in fact had Quercus reached the region (Van Zeist, Bottema 1991, p. 67). It was by far the dominant woodland tree in Turkey during the Chalcolithic (Eastwood et al. 1999; Roberts et al. 2001; Wick et al. 2003) and was very likely found in the vicinities of the site, as it requires a minimum of 400 mm of rain per year, that in the Malatya plain was surely surpassed at that time (Bryson and Bryson 1999, p. 13; Willcox 1974, p. 132; Miller 1999, p. 19), at an altitude which is only slightly higher than that of the plain of Malatya (900 m a.s.l.). Nowadays oak woodland is found over 1200 m above sea level (Woldring and Cappers 2001) but sparse trees can grow at lower altitudes. The presence of oak during the late Chalcolithic near the sites of Hacinebi and Kurban, south of Malatya, at lower altitudes and certainly in a less humid environment (Miller 1994, p. 170), even more so confirms that this tree must have been growing in the Malatya region too. Pollen records furthermore indicate that open deciduous oak woodlands were at that time present both in regions to the west (Roberts et al. 2001; Woldring, Bottema 2001/2) and to the east (van Zeist, Woldring 1978; Bottema 1995; Wick et al. 2003) of the Malatya plain well before the Chalcolithic period, even if the prehistoric land use probably retarded oak’s spread from the west to the east of Anatolia (Roberts 2002). The fact that oak was not used for the construction of the palace and temple, might thus be to blame to possible small dimensions of oak at that time, or other properties or meanings that we, today, do not appreciate. During the Chalcolithic, East Anatolian vegetation was one of grassland/steppe formation with trees or open wood formations dominated by deciduous oaks and sparse juniper trees or shrubs, with pines probably growing on the mountains. Some angiosperm trees like ash, poplar, elm, and alder, shrubs like pistachio and some species of Rosaceae were more or less sporadically present in the different sites. Van Zeist and Bakker-Heeres furthermore reconstruct the Late Chalcolithic and Early Bronze Age vegetation of the Altınova plain, slightly north east of Arslantepe, as one of open woodland dominated by poplar, ash and elm, with oak in the surrounding mountains and pistachio, maple (Acer) and juniper as minor constituents (van Zeist, Bakker-Heeres 1975, p. 233). Most of the wood used in the palace and temple would have thus been more or less readily available. But how far from the site did people have to go to collect it? Some of the taxa (alder, poplar) need permanent water and others (ash, elm)

Acknowledgements

We are grateful to Marcella Frangipane for having permitted and discussed this work with us and with Corrado Alvaro, with whom we have analysed the architectonic characters of the Arslantepe palatial structures. Many thanks are due to Maria Follieri for her advice and stimulus and to Fabio Nocca which cooperated in the charcoals identification. To Gian Maria di Nocera goes a special thanks as he started and greatly contributed to the combination of the botanical and archaeological data. It should also be stated that the detailed level of this work has been possible thanks to the numerous archaeologists excavating at Arslantepe.

References ALAN, M., 2002. Turkey. In: J. Turok, G. Eriksson, K. Russell,

S. Borelli, eds. Noble Hardwoods Network. Report of the fourth meeting 46 September 1999 Gmunden, Austria and the fifth meeting 17-19 May 2001. Blessington, Ireland: International Plant Genetic Resources Institute, Rome, 33-39.

alessio, m., allegri, l., azzi, c., bella, f., calderoni, g., cortesi, c., improta, s., petrone, v., 1983. 14C Dating of Arslantepe. Origini, 12, 575-580.

alvaro, c., 2004. Il più antico complesso palatino: architettura

del palazzo di Arslantepe. In: M. Frangipane, ed. Alle origini del potere. Arslantepe, la collina dei leoni. Milano: Electa, Mondadori, 60-61. ATASOY, H., KUCUK, M., ULU, F., CETINER, S., EREN, N., 2002. Abstract of “Management of Fast Growing Plantations”. In: (Results of Seventeenth Year) IUFRO Meeting: Spacing Trials on Black Alder [Alnus glutinosa (L) Gaertn Subsp. barbata (C. A. Mey.) Yalt.] in Eastern Black Sea Region of Turkey. 11-13 September 2002, Izmit, Turkey. AURENCHE, O., 1981. La maison orientale: l’architecture du Proche Orient ancien des origines au millieu du quattrième millennaire. Paris: Geuthner. BOTTEMA, S., 1995. Holocene vegetation of the Van area: palynological and chronological evidence from Söğütlü, Turkey. Vegetation History and Archaeobotany, 4, 187193. BRYSON, R.A., BRYSON, R.U., 1999. Holocene Climates of Anatolia as Simulated with Archaeoclimatic Models. Tüba-ar, 2, 1-14.

CALDERONI, G., CANEVA., I., CAZZELLA., A., FRANGIPANE, M., PETRONE, V., 1994. Department of Earth Sciences at the 248

Collapsed beams and wooden remains from Arslantepe Malatya (Turkey). Paléorient, 27 (2), 105-139.

University of Rome Radiocarbon Dates III. Radiocarbon, 36, 143-152 CORONA, E., 1991. Xilologia: implicazioni tecniche, storiche, filologiche. Annali Accademia Italiana di Scienze Forestali, 40, 211-236. DI NOCERA, G.M., 2000a. Radiocarbon datings from Arslantepe and Norşuntepe: the fourth-third millennium absolute chronology in the upper Euphrates and transcaucasian region. In: C. Marro, H. Hauptmann, eds. From the Euphrates to the Caucasus: Chronologies for the IV-III Millennium B.C. Istanbul: Institut Français D’Etudes Anatoliennes, 73-93. DI NOCERA, G.M., 2000b. C14 Datings at Arslantepe and Bronze Age Chronology in the Upper and Middle Euphrates. In: P. MATTHIAE, A. ENEA, L. PEYRONEL, F. PINNOCK, eds. Proceedings of the First International Congress on the Archaeology of the Ancient Near East. Rome, May 18th‑23rd 1998. BAR International Series 1063. Oxford: BAR Publishing, 333-348. EASTWOOD, W.J., ROBERTS, N., LAMB, H.F., TIBBY, J.C., 1999. Holocene environmental change in southwest Turkey: a palaeoecological record of lake and catchment-related changes. Quaternary Science Review, 18, 671-695. FAHN, A., WERKER, E., BAAS, P., 1986. Wood Anatomy and Identification of Trees and Shrubs from Israel and Adjacent Regions. Jerusalem: The Israel Academy of Sciences and Humanities. FRANGIPANE, M., 1997. A 4th-millennium temple/palace complex at Arslantepe-Malatya: north-south relations and the formation of early state societies in the northern regions of greater Mesopotamia. Paléorient, 23 (1), 45-73. FRANGIPANE, M., 2001. Centralization Processes in Greater Mesopotamia. Uruk “Expansion” as the Climax of Systemic Interactions among Areas of the Greater Mesopotamian Region. In: M.S. Rothman, ed. Uruk Mesopotamia and Its Neighbors. Santa Fe, Oxford: School of American Research Press, James Currey, 307347. FRANGIPANE, M., 2002. “Non-Uruk” Developments and UrukLinked Features on the Northern Borders of Greater Mesopotamia. In: N. Postgate, ed. Artefacts of Complexity. Tracking the Uruk in the Near East. Iraq Archaeological Reports 5. Warminster: Aris & Phillips, 123-148. FRANGIPANE, M., 2003a. Origini ed evoluzione del sistema centralizzato ad Arslantepe: dal “Tempio” al “Palazzo” nel IV millennio a.C. ISIMU, 3-2000, 53-78. FRANGIPANE, M., 2003b. Developments in Fourth Millennium Public Architecture in The Malatya Plain: From Simple Tripartite To Complex and Bipartite Pattern. In: H. Hauptman, M. Özdoğan, eds. From Primary Villages to Cities, Essays in honour of Ufuk Esin. Istanbul: Arkeoloji ve Sanat Yayinlari, 147-169. FRANGIPANE, M., 2004. Alle Origini del Potere. Arslantepe, la collina dei leoni. Milano: Electa, Mondadori. FRANGIPANE, M., PALMIERI, A., 1983. Perspectives on Protourbanization in Eastern Anatolia: Arslantepe (Malatya). An Interim Report on 1975-1983 campaigns. Origini, XII, 287-668.

GIORDANO, G., 1971. Tecnologia del legno. vol. I. Torino: Unione tipografico-editrice torinese.

GREGUSS, P., 1955. Identification of living Gymnosperms on the basis of xylotomy. Budapest: Akadémiai Kiadó.

GREGUSS, P., 1959. Holzanatomie der europäischen laubhölzer und sträucher. Budapest: Akadémiai Kiadó.

HALL, G., MCBRIDE, S., RIDDEL, A., 1973. Asvan 1968-1972,

An interim report. Architectural study. Anatolian Studies, 23, 245-269. KAYA, Z., RAYNAL, D.J., 2001. Biodiversity and Conservation of Turkish Forests. Biological Conservation, 97, 131-134. KOYUNLU, A., 1982. The Village Settlement of Munzuroğlu - Observations on housing. In: S. Pekman, ed. Keban Project 1974-1975 Activities. Ankara: METU, 249-256. KUBAN, D., 1970. The first preliminary report on the rural architecture of the Keban dam area, 1968. In: S. Pekman, ed. Keban Project 1968 Activities. Ankara: METU, 179182. KUNIHOLM, P.I., 1996. The Prehistoric Aegean: Dendrochronological progress as of 1995. Acta Archaeologica, 67, 327-335. MALLOWAN, M.E.L., 1947. Excavations at Brak and Chagar Bazar. Iraq, 9, 1-266. MARGUERON, J.C., 1992. Le Bois dans l’architecture: premier essai pour une estimation des besoins dans le Bassin Mésopotamien. Bulletin on Sumerian Agriculture. 2, p. 79-96. MILLER N. F. 1994. Some archaeobotanical remains from the 1992 excavation season at Hacinebi Tepe. Anatolica, 20, 168-172. MILLER, N., 1999. Interpreting Ancient Environment and Patterns of Land Use: Seeds, Charcoal and Archaeological Context. Tüba-ar, 2, 15-30. NAUMANN, R., 1971. Architektur Kleinasiens von ihren Anfängen bis zum Ende der Hethitischen Zeit, 2nd edn. Tübingen. PESSIN, H., 2002. Preliminary report on the charcoal analysis of Tell Shiukh Fawqani (Middle-Euphrates, Syria), IIIrd- Ist millennia BC. In: S. Thiebault, ed. Charcoal analysis. Methodological Approaches, Palaeoecological Results and Wood Uses. Proceedings of the Second International Meeting of Anthracology, Paris, September 2000. BAR International Series 1063. Oxford: BAR Publishing, 147-150. POSTGATE, N.J., 1992. Trees and Timber in the Assyrian Texts. Bulletin of Sumerian Agriculture, VI, 177-191. ROBERTS, N., 2002. Did prehistoric landscape management retard the post-glacial spread of woodland in Southwest Asia? Antiquity, 76, 1002-1010.

ROBERTS, N., REED, J.M., LENG, M.J., KUZUCUOĞLU, C., FONTUGNE, M., BERTAUX, J., WOLDRING, H., BOTTEMA, S., BLACK, S., HUNT, E., KARABIYIKOĞLU, M., 2001.

The tempo of Holocene climatic change in the eastern Mediterranean region: new high-resolution crater-lake sediment data from central Turkey. The Holocene, 11, 721-736. ROTHMAN, M., 2001. Uruk Mesopotamia and Its Neighbors. Santa Fe, Oxford: School of American Research Press, James Currey. SCHOCH, W.H., PAWLIK, B., SCHWEINGRUBER, F.H., 1988. Botanische Makroreste. Bern, Stuttgart: Verlag Paul

FRANGIPANE, M., DI NOCERA, G.M., HAUPTMANN, A., MORBIDELLI, P., PALMIERI, A., SADORI, L., SCHULTZ, M., SCHMIDT-SCHULTZ, T.,

2001. New Symbols of a New Power in a “Royal” Tomb from 3000 BC Arslantepe,

249

L. Sadori, F. Susanna, F. Balossi Restelli Haupt.

SCHWEINGRUBER, F.H., 1978. Mikroskopische Holzanatomie.

Kommissionsverlag Zürcher AG, CH-6301 Zug. Birmensdorf: Eidgenössische Anstalt für das Forstliche Versuchswesen CH-8903. SCHWEINGRUBER, F.H., 1990. Anatomie europäischer Hölzer (Anatomy of European woods). Bern Stuttgart: Verlag Paul Haupt. WICK, L., LEMCKE, G., STURM, M., 2003. Evidence of Lateglacial and Holocene climatic change and human impact in eastern Anatolia: high-resolution pollen, charcoal, isotopic and geochemical records from the laminated sediments of Lake Van, Turkey. The Holocene, 13, 665-675. WILLCOX, G., 1974. A history of deforestation as indicated by charcoal analysis of four sites in Eastern Anatolia. Anatolian Studies, 24, 117-133. WILLCOX, G., 1991a. Cafer Höyük (Turquie): Les Charbons de bois Néolithiques. Cahiers de l’Euphrate, 5-6, 139-150. WILLCOX, G., 1991b. Exploitation des espèes ligneuses au proche-orient: données anthracologiques. Paléorient, 17, 117-126. WILLCOX, G., 1992. Timber and Trees: Ancient Exploitation in the Middle East: Evidence from Plant Remains. Bulletin of Sumerian Agriculture, VI, 1-31. WILLCOX, G., 2002. Evidence for ancient forest cover and deforestation from charcoal analysis of ten archaeological sites on the Euphrates. In: S. Thiebault, ed. Charcoal analysis. Methodological Approaches, Palaeoecological Results and Wood Uses. Proceedings of the Second International Meeting of Anthracology, Paris, September 2000. BAR International Series 1063. Oxford: BAR Publishing, 141-145. WOLDRING, H., BOTTEMA, S., 2001/2. The vegetation history of East-Central Anatolia in relation to archaeology: the Eski Acigöl pollen evidence compared with the Near Eastern environment. Palaeohistoria, 43/44, 1-35. WOLDRING, H., CAPPERS, R., 2001. The origin of the “Wild Orchards” of Central Anatolia. Turkish Journal of Botany, 25, 1-9. VAN ZEIST, W. 1972. Palaeobotanical results of the 1970 season at Çayönü, Turkey. Helinium, 12, 319. VAN ZEIST, W., BAKKER-HEERES, J., 1975. Wood identifications. In: M. Van Loon, ed. Korucutepe vol. I. Amsterdam, New York: North-Hollad publishing company, Elsevier, 233257. VAN ZEIST, W., BOTTEMA, S., 1991. Late Quaternary vegetation of the Near East. Wiesbaden: L. Reichert. VAN ZEIST, W., WOLDRING, H., 1978. A postglacial pollen diagram from lake Van in East Anatolia. Review of Palaeobotany and Palynology, 26, 249-276.

250

The Use of Wood: Traditional Building Techniques in the Swat Valley (Pakistan). An Ethno-Archaeological Research Ilaria E. Scerrato

Associazione Italiana Etnoarcheologia E-mail: [email protected]

Abstract

The Swat Valley in North-Eastern Pakistan, reached by the military expedition of Alexander the Great, was an important destination of Buddhist pilgrimages until the beginning of the XVII century. After a first Islamic penetration in the X century, a real Islamization took place only between the XVI and XVII centuries, when the Yusufzai coming from the Kabul Valley, established in the area. An interesting wooden religious and civil building technique was preserved until relatively recent times in most of the area. This usage of wood (both structural and decorative) was made by local craftsmen, who handed down from generation to generation their technical capabilities. In the last thirty years, the economical development and social transformations have caused the disappearance of this tradition, with a replacement of the traditional wooden building techniques with cement and sheet-metal structures. A rich photographic documentation has been collected by the Italian Archaeological Mission of the Istituto Italiano per l’Oriente e l’Africa (ISIAO). However, there are still extant examples of this traditional architecture, with buildings harmonious with the surroundings, employing several materials easily available in the area such as schist, river gravel, and wood. The hypothesis of this research is based on the identification of two main areas, the mountainous Northern area and the Lower Valley area, respectively, where different building techniques are used. In the Lower Valley the walls are made of layers of river pebbles alternated with wooden beams. In the mountainous Northern area, rich of Himalayan pine and cedar trees, walls are totally made of wood, excepting foundations, made of dry stone. This research aims at identifying the fundamental characteristics of these structures: static and construction principles, ground anchorage, walls and corner solution, structure and roof covering as well as the distribution of the living space, including both religious and civil structures like houses, hay stacks, grain storage facilities, mills, bridges, and irrigation channels. An ethnoarcheological approach is used, revealing the complementarity between ethnology and archaeology.

The ethnoarchaeological research project described in the following pages is going to become an important approach in the ample and intense studies conduct by the Italian Archaeological Mission of the Istituto Italiano per l’Oriente ed Africa (IsIAO) of the history, archaeology and culture of Pakistan Swat Valley. The urgency to proceed with the research is due to the unfortunate and well known abandonment of the use of the traditional building methods of religious and civil structures. However, recent preliminary surveys in the interested areas were able to highlight the enduring presence of dwellings, mosques and water mills built accordingly to the old traditional techniques. The project will be articulated in two stages: • A thorough study of the information collected by the Italian Archaeological Mission as well as other sources.

Fig. 1. The Swat Valley (1996, I.E. Scerrato).

• An additional field research especially focused on the less studied areas of the Swat Valley. Target of the project is to highlight the traditional building patterns, both in civil and religious structures, so that it would be possible to individuate the elements and techniques which survived economical, social and cultural changes and to study the figure of the carpenters who play a pivotal role in their society. Descending from the majestic Hindukush mountain range the Swat River gives his name to a luscious and fertile valley blessed by the Monsoons influence (Fig. 1). Its strategic location made the valley not only the heart of the Buddhist culture and religion, a pilgrimage venue as several standing monasteries remind us, but also an important commercial meeting point for India, Central Asia and China. Starting in 1956 Professor Giuseppe Tucci, of the Istituto Italiano per il Medio ed Estremo Oriente (IsMEO now IsIAO), promoted and implemented numerous archaeological researches in the ancient Uddyana region, now known as Swat. The research, originally interested in acquiring a deeper knowledge of the traces left by the invasion and conquest by Alexander the Great, also involved studies of Buddhist Art and Architecture (Tucci 1958; Faccenna 1964; Gullini 1962). Then the studies were extended to the Protostoric period (II and I millennium BC) and to the first Islamic penetration in the XI century accomplished by Mahmud of Ghazna (Scerrato 1986; Stacul 1987). 1 Later, the excavation activity was joined by an intense research aimed at collecting ample graphic and photographic data of the wooden mosques of the Swat region doomed

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

I. E. Scerrato

Fig. 2. Map of Northern Pakistan (Is.I.A.O. Dep.Cs. 12231).

to disappear due the aggressive economical and social changes. From the Swat Valley the survey was expanded to all the impervious area of the Northern Pakistan as far as in Chitral and through the Indus Khoistan to the far North East of the country in Baltistan and Gilgit regions, at the border with China.2 During the years of its activity, the Italian Archaeological Mission was able to collect from the bazaars of the Swat Valley a treasure of interesting wooden elements from abandoned mosques. This very important material, preserved in Saidu Sharif Mission Base, along with several other wooden objects gathered in the area, has become the starting point for the study of the local wooden handicraft, its techniques, privileged patterns and cultural contest. The research is still ongoing and is studing the background of the carpenters, who usually are not part of the Pukhtun ethnia but descend from the Dardic autochthon element like the Torwaly and the Gawri concentrated in the Swat Kohistan. The proposed study will focus on the techniques and the utensils they use and their training within their families: the youngsters will acquire from their elders the essential skills and knowledge of the art of carpentry. The handing down from generation to generation of the technical capabilities guarantees a cultural continuity. Artisans’ manual skills and their intimate knowledge of the wood allow them to use primitive tools to transform rough trunks into beams or pillars. To accomplish these tasks, the artisans must have deep knowledge of the materials they employ so that they can maximize their usefulness. It is significant that these artisans are part of the Dardic ethnic

group, a people that lived in this area before the Islamic expansion, in XVI century when the Yusufzai, coming from the Kabul Valley, established themselves in Swat and Dir and then in the Northern Areas, where the ruggedness of the region created peculiar political conditions with the absence of any strong central authorithy.3 The study will also highlight the important role that the carpenter plays in the society. It is indeed the carpenter who builds the traditional objects necessary to perform passage rites like weddings and deaths. It is the carpenter who builds the small bed where the bride sits during the ceremony as well as the wooden fence surrounding the graves of notable members of the society. Moreover, is the

Fig. 3. Manikial Bala, Darel Valley, Indus Kohistan. House. (Is.I.A.O. Dep. Ng.L 13778/25).

252

The Use of Wood: Traditional Building Techniques in the Swat Valley carpenter who places the keystone pole at the beginning of the construction of a new house, a true symbol of foundation, and he also carves pillars and capitals, doors and window frames, domestic utensils as well as the lectern for the Koran, the Prayer Tables and the Mimbar for the mosques. A preliminary field research was able to determine that to this date traditional building techniques are still in use and the wood artisans are involved in the building of modern structures. Nowadays artisans employ several materials easily available in the area such as schist, river gravel, and Himalayan pine and cedar trees. The blending of these materials, which is harmonious with the surroundings, is both beautiful and practical. The usage of wood makes the structures more elastic and thus resistant to earth quakes, a fairly common occurrence in Northern Pakistan which lies on the Indian-Asian fault line (Fig. 2).

Fig. 4. Water mill near Kalam, Swat Kohistan (1996 D. Rosati)

An ethnoarcheological approach is important because it will reveal the complementarity between ethnology and archaeology. It will offer the opportunity of a deeper knowledge of use, origin and socio-cultural context of specific architectural remains and material culture, through a critical use of analogy between the living pre-industrial society and the archaeological context. 4 The ethnographic fieldwork aims at enriching the archaeological research, which in turn will give an important and unreplaceable contribution to understanding the living traditions in a region like Swat Valley, that has been under the influence of Greek, Indian, Buddhist, Central Asian, and Islamic cultures.

Fig. 5. Bridge on Swat river near Kalam, Swat Kohistan (1996 D. Rosati)

A detail review of existing buildings and techniques used will help to gather relevant socio-cultural information. The research will include both religious and civil structures like houses, hay stacks, grain storage facilities, mills, bridges, and irrigation channels (Figs. 3, 4 and 5). Common characteristics as well as variations will be taken into consideration. The survey of wooden mosques in the 80s, enriched with data about dwellings, has identified two major categories corresponding to the two cultural areas of the Swat. The Lower valley where the main culture is Pukhtun, and the Upper Valley, area where the population is mainly Dardic (Scerrato 1981; for the Dardic culture see Biddulph 1977; Tucci 1977; Leitner 1978; Macmahon, Ramsay 1981). The proposed research will focus on well preserved buildings that present structures built using traditional techniques as well as recent ones that follow traditional patterns. Our hypothesis is that the differences identified in the mosques will also exist in dwellings since the layout of the religious buildings is ispired by the non religious ones. The research will aim at identifing the fundamental characteristics of these structures: static and construction principles, ground anchorage, walls and corner solution, structure and roof covering as well as the distribution of

Fig. 6. Laikot, Swat Kohistan. General view of the mosque. (Is.I.A.O. Dep. n.g. R 5291/7; F. Bonardi)

253

I. E. Scerrato the living space. Once these fundamental characteristics will have been identified, they will be correlated with climatic factors as well as cultural, social and economic ones. The significant archaeological data will provide a guide that will help trace the history of the construction techniques. Particular emphasis will be placed on analyzing those techniques that have been past from generation to generation, noting how they have changed and have been reintroduced due to cultural factors. This archaeological data derives from the research studies and the dig reports that illustrate the formation processes and the construction techniques that came to light with the excavations of sacred areas, living quarters and defensive structures. This data provides us with information on the history of archaeological context, on the choice of materials, on their use and on the construction techniques developed to best exploit the characteristics of the selected materials.

Fig. 7. Malam Jaba area. General view of the mosque (Is.I.A.O. Dep. n.g. L 14508/34).

The traditional building in the Swat Valley presents a horizontal layout. Its architectural scheme pillar-capitalbeam rarely reaches heights over 3 meters (approximately 10 feet). The basic pattern of the structure is composed of wooden horizontal elements on which rest horizontal beams held in place by capitals. All wooden junctures use a mortise and tenon connection without using dowels or pegs. Mortise and tenon connections are also employed in the making of domestic goods like chests to store fabrics or food, beds and the typical low chairs. As previously mentioned two principal construction approaches are found in the mountainous Northern Area and the Lower Valley area. In the North, which has greater

Fig. 8. Gurunai, Swat Kohistan. NW wall of the mosque (Is.I.A.O. Dep n.g. L 14930/5a U. Scerrato).

Fig. 9. Khwagakhela, Lower Swat. Bracket capital in Lora Jumat (Is.I.A.O. Dep n.g. L 14505/31 U. Scerrato).

254

The Use of Wood: Traditional Building Techniques in the Swat Valley

Fig. 11. House in Durushkhela, Lower Swat. River pebbles wall (1996 I. E. Scerrato).

Fig. 10. House in Durushkhela, Lower Swat. A wooden door (1996 I. E. Scerrato).

quantities of naturally occurring building materials, buildings’ walls as recently as 30 years ago were totally made of wood. Dry stone masonry was used only for foundations insuring a better stability and reducing the humidity coming from the soil (Figs 6 and 7). As time progressed, the usage of wood, even in these forest rich areas, became limited to the most significant structures such as capitals, columns, doors, and windows. These wooden structures are decorated with rich carvings made by able craftsmen. This newer architectural style is similar to the style which has been used in the Lower Swat Valley where the walls are made of layers of river pebbles alternated with wooden beams (Figs 8 and 11). The walls made of stones, even though do not follow a precise pattern, are compact so that it is possible to cover them with a mixture of mud and straw making them waterproof and insulated from heat and cold. However, even when the walls are made of stones, wood is the material of choice for pillars, columns, bracket capitals, doors and windows that are enhanced by elegant carved decorations. The carving style varies from region to region giving the buildings a unique flavor (Figs 9 and 10). The roofs are flat, composed of small beams that connect transversally with the beams below on which rest boards covered with branches, straw and mud, a top layer that needs to be periodically renewed. This roofing style is

Fig.12. Malam Jaba. Woman spreading mud to reinforce the roof’s protection (1996 I. E. Scerrato).

common in the mountains and in the plains. The flat roof provides an ideal space to lay out in the sun grains and clothes. In the mountainous areas the buildings are constructed on a series of terraces built on slopes. In these areas the roofs are the only available flat outdoor space making them the selected area where most of the social life is carried out (Fig. 12). The flat roofs are not unique to the Swat Region; they are also used in the Northern Areas of Pakistan and in the Afghanistan Nuristan Region. The social usage of 255

I. E. Scerrato roofs will be highlighted and analyzed to determine any significant differences among the various areas taken under consideration. 5 The project will also study the layout of the entrance, the fire place, the principle living spaces and their function. This will help to understand the hierarchy of the rooms which will reveal the social status of the people living in the house.6

behavior likely to leave identifiable residues in the archaeological record» (1996). And in David and Kramer we can read: « Ethnoarchaeology is neither a theory nor a method, but a research strategy embodying a range of approches to understanding the relationship of material culture to culture as a whole, both in living context and as it enters in archaeological record, and to exploiting such understanding in order to inform archaeological concepts and to improve interpretation» (2001: 2). For more interesting perspectives in ethnoarchaeology see Vidale 1992. 5 The usage of flat roofs is common in Northern Pakistan (Dainelli 1924; Hussam-ul Mulk, Staley 1968) as well as in the Afghani Nuristan territory that borders with Pakistan. This area in the past has also been called Kafiristan, land of the Kafirs, that is the infidels, since they were converted to Islam very late (1895-1896) subjugated by Abdur Rahman, an Afghani Emir. Information about the life style and habits of the population in this area prior to their conversion to Islam can be found in the writings of George Scott Robertson, a doctor of the Britannic Agency in Gilgit (Robertson 1974, Chapter XXVII). Roberson describes the houses in the Bushgal Valley covered by flat roofs. On one side he thought that it is bad choice because most of the time this style of roof offers poor protection from heavy rains; on the other side he states: «It is necessary that they should be flat, for contiguous roofs form the only level space which can be found in some villages where corn be winnowed or trashed, or fruit be spread out to dry» (Robertson 1974, pp. 486-487). For a more up to date study of the buildings in this area see Edelberg, 1984. 6 An example can be the hujra, the men guest house, which is characteristic of the area in which the Pukhtun woodwork is dominant and widely present in the Swat region. These types of dwellings have recurrent structures, but also have many variations that require additional studies. Ahmed, in his study about the Pukhtun in the Tribal Areas, considers the importance of the presence of the hujra: « The three key and prestige conferring symbols of tribal society, the male guest house, (hujra), the gun (topak) and the council of the elders (jirga), are exclusively the reserve of Pukhtun males.» (1980: 6); «The unity of the Pukhtunwaly and Islam is symbolized and expressed in village social life by the physical juxtaposition of the mosque and the hujra. These two institutions are the focus of life in every settlements and village» (ibid.: 106).

As it was said at the beginning of the present article the rich architectural and cultural patrimony is in danger of being corrupted and in many cases will disappear. The briefly outlined project of research will help to preserve the traditional art of carpentry and the still standing structures giving the opportunity to the future generations to maintain their traditions. The cultural isolation of these territories was largely responsible for the emergence of certain characteristic of their material culture and has allowed the preservation of extensive traditional forms. A hightly typical feature of this is the architecture- an architecture without architects- which has found in the mosque, with fine wooden carvings, its hightest form of expression.

Acknowledgments

I would like to express my heartfelt thanks to alls who have in some contributed to this work, to my sister Francesca R. Scerrato and to Piero Melchiorri, for their translation of the text and some useful suggestions; to Danilo Rosati, for his photos, for his friendly encouragements and remarks.

Endnotes 1

For a synthesis of the activity of the Italian Mission in Pakistan see Scerrato 1997, Taddei 1997. For an additional general review of the work of the Mission and an updated bibliography see Callieri, Filigenzi 2002. This latest publication was prepared in conjunction with the exhibit of the frieze of the principle Stupa of Saidu Sharif and of other artifacts from the Oriental Museum in Rome. 2 Next to the great Islamic architecture there is in Pakistan a very large patrimony of architecture less noble in the material used but of significant beauty and importance. It is the architecture that it is typified in the wooden Mosques of Northern Pakistan. A similar architecture is also found in homes and cemeteries. In the 1960s, this architecture was studied, thanks to Professor Tucci’s interest by architects Giuseppe Zander and Luca Mariani. In 1980, this research was resumed and expanded by Umberto Scerrato. It was urgent to document this style of architecture because many of the Mosques were thorn down to be replaced with cement ones and the beautiful carved elements in the original Mosques were being destroyed (Scerrato 1981, 1982, 1983: Mariani 1986). 3 The ethnic group of Pukhtun lives between South-Eastern Afghanistan and in the Pakistan North-Western Frontier Province. They are a tribal society with agnatic descent. (For more information about the Pukhtun economic and socio-political organization see Barth 1959; Ahmed 1980.) 4 Kramer defines the ethno-archaeology as « the the ethnographic fieldwork carried out with the express purpose of enhancing archaeological research by documenting aspects of sociocultural

References Ahmed, A.S. 1980. Pukhtun Economy and Society. Traditional Structure and Economic Develpoment in a Tribal Society. Routledge & Kegan Paul. London. Barth, F. 1956. Indus and Swat Kohistan. An Ethnographic Survey. Forende Trykkerier. Oslo. Barth, F. 1959. Political Leadership among Swat Pathans. The Athlone Press. University of London. Biddulph, J. 1977. Tribes of the Hinddo Koosh. Indus Pubblications. Karachi (1st ed. Calcutta 1880). Callieri, P. and. Filigenzi, A. 2002. Il Maestro di Saidu Sharif. Alle origine dell’arte del Gandhara. Museo Nazionale d’Arte Orientale, Istituto Italiano per l’Africa e l’Oriente. Roma. Dainelli, G., 1924. Le condizioni delle genti. Nicola Zanichelli Editore. Bologna. David, N. Kramer, C. (eds.) 2001. Ehnoarchaeology in Action.

256

The Use of Wood: Traditional Building Techniques in the Swat Valley Cambridge University Press. Edelberg, L. 1984. Nuristani Buildings. Jutland Archaeological Society Pubblications XVIII. Moesgård. Faccenna, D. 1980-81. Butkara I (Swat, Pakistan) 1956-1962 (IsMEO Reports and Memoirs, III). Rome. Gullini, G. 1962. Udegram. Is.MEO Reports and Memoirs I. Rome. Hussam-ul-Mulk, S., Staley, J. 1968. Houses in Chitral: Traditional Design and Function, Folklore, 79: 92-110. Jettmar, K. 1961. Urgent Tasks of Research among Dardic Peoples of Eastern Afghanistan and Northern Pakistan, Central Asiatic Journal,VI, 4: 85-96 (1st ed. 1959). Jettmar, K. 1980. Bolor & Dardistan. National Institute of Folk Heritage. Islamabad. Kalter, J. 1991. The Arts and Krafts of the Swat Valley. Living Traditions in the Hindu Kush. Thames and Hudson. London. Kramer, C. (ed) 1972. Ethnoarchaeology. Implications of Ethnography for Archaeology. Columbia University Press. New York. Kramer, C. 1996. Ethnoarchaeology. In Levinson, D. and Ember, M. (eds.), Encyclopedia of Cultural Anthropology. Henry Holt and Co. New York:396-9. Leitner, G.W. 1978. Dardistan in 1866, 1886 and 1893. Manjuri Publishing House. New Delhi (1st ed. 1894?). Mariani, L. 1986. Tecniche costruttive nelle moschee lignee dello Swat, Annali Istituto Orientale. Napoli, vol. 46: 63-75. McMahon, A.H., Ramsay, A.D.G. 1981. Report on the Tribes of Dir, Swat and Bajour together with the Utman-Khel and Sam Ranizai. Saeed Book Bank. Peshawar (1st ed. 1901). Olivieri, L.M. 2003. Bir-kot- Ghwandai Interim Reports I. The Survey of the Bir-kot hill. Archaeological Map and Photographic Documentation. Rome. Scerrato, U. 1981. Survey of wooden mosques and related wood carving in the Swat valley, IsMEO Activities, East and West, 31: 178-181. Scerrato, U. 1983. Survey of wooden mosques and related wood carvings in North-West Frontier Province, 3rd and 4th Reports, IsMEO Activities, East and West, 33: 325-328. Scerrato, U. 1984. The Wooden Architecture of Swat and the Northern Areas of Pakistan: a Report on the Research carried out in 1984, East &West, 34: 501-515. Scerrato, U. 1985. More on Wooden Mosque in North Pakistan, Journal of Central Asia, VIII,1: 105-109. Scerrato, U. 1986. Excavation of the Ghaznavide Mosque. Second Preliminary Report, East & West, 36,4: Scerrato, U. 1986. Excavations on Raja Gira, Swat, Pakistan Archaeology, 10-22, 1974-1986. Scerrato, U. 1997. Ricerche di Archeologia, Storia dell’Arte e Architettura Islamica in Pakistan, in Missioni Archeologiche Italiane. Ministero degli Affari Esteri. Roma: 243-248. Stacul, G. 1987. Prehistoric and Protohistoric Swat. IsMEO Reports and Memoirs, XX. Rome. Taddei, M. 1997. Missione Archeologica Italiana nella Valle dello swat e aree limitrofe, in Missioni Archeologiche Italiane. Ministero degli Affari Esteri. Roma:249-252. Tucci, G. 1958. A Preliminary Report on an Archaeological Survey in Swat, East and West, 9, 279-328. Tucci, G. 1977. On Swat. The Dards and Connetted Problems, East and West, 27, 94-103. Tucci, G. 1997. On Swat. Historical and Archaeological Notes. Istituto Italiano per l’Africa e l’Oriente, Italian Archaeological Mission in Pakistan. Rome.

Vidale, M. 1992. Produzione artigianale protostorica. Etnoarcheologia e archeologia. Saltuarie dal laboratorio del Piovego 4, Università degli studi di Padova.

257

Environmental and cultural history of South American temperate forests: an interdisciplinary approach M.E. Solari

Instituto de Ciencias Sociales, Laboratorio de Arqueobotánica, Facultad de Filosofía y Humanidades, Universidad Austral de Chile. E-mail: [email protected] Abstract To understand the interactions between culture and the forested ecosystems of southern Chile we have had to study recent and past issues, such as the contrasting visions that natural and social sciences have about the environment or the mistaken dichotomy between environment and culture that prevails in the social sciences. Our main goal has been to establish links between these two sciences in order to study the Holocenic temperate forest of southern Chile and the numerous cultural strategies in them originated. This has required the combination of results and analysis from diverse disciplines such as anthracology, carpology, environmental history and ethnobotany.

Introduction To establish interactions between the social and natural sciences, in the last five years we have had to go beyond the interdisciplinary discussion between archaeology and palaeo-archaeobotany of our earlier studies in Patagonia (Solari 1990; 1993; 1994). Now are reassessing forested ecosystems by determining their species composition, creating comparative collections of seeds and fruits and woods, generating data bases and identification keys for its carpological and wood anatomy elements, and to lesser extent, developing a palynological component. These efforts have been supplemented by adding cultural and environmental history records obtained from numerous disciplines such as archaeology, anthracology, carpology, palynology, ethnography and the study of historical documents. In the last 15 years our interests have been focused on the following questions: - Why should we study the environmental history and Holocenic culture in temperate forests? - What methods can we use to study them? Are they suitable? And at last, - What are the inhabiting/use models that prevailed in these temperate forests of southern Chile and the ChileanArgentinean Patagonia for the last 10000 years? Due to the novelty of our studies in Chile, we had to establish some basic principles to bring together the confronting perspective that the natural and social sciences have about “human communities-environment-climate”. We are certain that only by combining and conciliating both views, a systematic and suitable synthesis about the use/inhabiting of the temperate forests will be obtained. This will allow us to assess our region from a transdisciplinary point of view. The main disciplines involved in our efforts have been archaeology, palaeobotany, archaeobotany; and lately, we

have also included ethnobotany. The systematic study of contemporary Huilliche communities in southern Chile (39-40º S) exemplifies the inclusion of the latter. This study pretends to understand what these communities think about plant communities, how they classify them, and the use they give to the plant species. All this has been supplemented by the study of phenological patterns, abundance/presence of edible species, how wood and other primary forest products are gathered and processed, and the route these products follow from the forest to the cities (i.e. the “wood culture” that characterises this region) (Solari and Poblete 1999; Ordoñez 2004, unpublished). The use of bioecological disciplines by the social sciences such as carpology (the study of seed/fruit remains), anthracology (charcoal wood) and dendrochronology (tree-rings of 2 submerged wood canoes) has provided suitable tools to study plant macroremains in archaeological contexts. We consider these disciplines as linking disciplines, since they create a nexus between the natural and social sciences. Also, numerous ecological, anthropological and ecological history studies have greatly contributed to recreate the past climate, environment and the use/inhabiting models. Ecological characterisation Temperate forests occur above the 30° latitude on both hemispheres; from sea level to the tree lines. In Chile, temperate forests form a continuous belt between the 35°S and 55°S (Patagonia). They also extend towards the Andes, colliding with the Argentinean steppe (Armesto et al. 1995). Our study area is mainly restricted to the temperate rain forests in the west slope of the Andes; from c. 40°S (Valdivia) to 56°S (Archipelago of Cape Horn) (fig 1). We have chosen temperate forests because they are fragile and threatened ecosystems and secondly, because of the cultural link with their inhabitants still remains and has gave birth to numerous adaptive strategies.

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

M.E. Solari Cultural characterization

Anthracology

A long settlement record that accounts for the last 13000 years has been observed in southern-austral Chile and numerous examples of adaptive strategies have been documented during this period. During the Paleo-Indian Period (late Pleistocene), records of hunter-gatherers groups in Monte Verde (41°S) suggest they have modified their local landscape and co-existed with fauna nowadays extinct. A great diversity of uncharred seeds, edible fruits, medicinal plants and wood tools has been retrieved from this site in which preservation of plant remains has been particularly favoured by the formation of a peat bog on top of the site (Dillehay 1989, 1997, 2004).

For the last 15 years the study of charcoal wood or anthracology has been rarely included in the archaeological studies. Charcoal wood analyses have been restricted to southern Chile and the Chilean Patagonia. These few studies have helped to depict the following activities and issues;

Stone tools associated to extinct mega-fauna, have been found in Austral Patagonia (e.g. caves of Fell, Pali Aike in Chile and Los Toldos in Argentina). These findings suggest a “nomadic” culture almost 11000 BP (Massone 1989, 1996). Studies on sediment, pollen and charcoal wood on other Patagonian sites (i.e. Ponsonby; Isla Riesco, Magallanes) have shown local changes in the habitat by 7000 BP and also confirmed the arrival, after marine routes were opened e.g. Fitz-Roy Channel, of marine hunter-gatherer groups to a terrestrial hunter’s habitat (in Legoupil 2003). From the studies conducted at the site Marifilo (39ºS, Calafquen Lake) the presence of a tradition to the temperate forest can be proposed for the last 10000 years. Marifilo is located in the interface of the Araucanean lakes (sensu Thomasson 1963) and the Andes. This tradition consists in a seasonal occupation of different ecological niches; from the Araucaria araucana forests above the tree line, allowing the gathering of its edible seed, to ecotone habitats besides the valley lakes (Adan 2000, unpublished; 2003). Later, about 5500 BP, hunter-fisher and marine gatherer groups populated the Southern and Austral Coastal areas. Large accumulations of shells and other marine debris in sites such as Chan-Chan, Piedra Azul, Seno de Otway, Seno de Skyring, and Archipelago of Cape Horn still remain as proof of their occupation (Navarro 1995, 1997, 1999, Gaete et al. 2004; Legoupil 1989, 1995, 1997). Finally, about 300 AC, during the Formative period, many groups were settled in the territory, examples such as the Pitren Ceramic and cultivation of maize and potato account for their establishment in the region. The presence of ceremonial soil build-ups (Kuel) are evidence of the great complexity of these societies. The Vergel and Valdivian culture were truncated by the arrival of the Europeans (Dillehay 1986, 1990, 2003). 260

• Forest clearing by settling groups The large chronology (6000 to 680 BP) obtained from shell accumulations in the Archipelago of Cape Horn (56°S) and surrounding areas shows that forest modification occurred in several stages; probably a set of progressive occupational steps with ecological consequences for the less valued plant species, e.g. heliophilous species occurring in the site. These steps could include first a colonization stage for “cleaning and pruning” the vegetation, an intermediate stage in which resources depletion occurs (and likely partial fertilization of the soils) and a final stage in which the forest regenerates towards its primary composition (Solari 1993, 1994). • Linking palinology, archaeobotany (leaf remains), dendrocronoly, xylology and antrachology. To study Ponsonby site (Riesco Island; Magallanes) a combination of palaeoecological disciplines was required. Charcoal wood analyses confirmed 2 ecological changes; one ca. 7000 BP in which many of the plant taxa in the site indicated moist conditions, presence of lakes and small forests of Nothofagus betuloides. Contrary, in the second period ca. 4000 BP, steppe species such as Maytenus magellanica dominated the landscape; as is currently observed. This change in the vegetation seemed to have occurred after the opening of the FitzRoy channel, joining the Skyring and Otway Seas and the dry and marine conditions that prevailed afterwards (Solari 2003; Solari and Richard 2003; Solari et al. 2002). • Physiognomy of the southern temperate rain forests (39-42° S). The vision of a stable and heavily dense temperate rain forest in southern Chile during Holocene is common between many archaeologists. Nevertheless, anthracological studies in Marifilo (39°S, Calafquen Lake) have demonstrated changes in the vegetation of Andean regions, probably due to volcanic events (ash and fires) and the anthropization of the habitat by horticultural-agricultural groups during the Formative Period (Lehnebach et al. MS). Also it is important to point out the apparent bond between archaeological sites and deciduous forests of Nothofagus obliqua and ecotone habitats. These habitats are generally more exposed, resources are more abundant and they are easier to live in and to “domesticate” (Lehnebach et al. 2005, unpublished).

Figure 1: Distribution of temperate forests in the world (A) and study area in Chile (b).and cultural history of South American temperate forests Environmental be re-assessed. Questions such as how well seed and fruit remains found in these archaeological sites (e.g. Marifilo) illustrate the plants included in the diet? What is the origin of the seed/fruit remains (human/natural) and how did they get in the site (intentionally/accidentally)? Is there a differential durability of the remains? Many non-edible species with durable seeds are over-represented in the site (e.g. stone of Aextoxicon punctatum) while edible species with fragile seeds are scanty.

A

We expect that building a large comparative collection and conducting taphonomic and experimental studies with plant species native to the south of Chile at the Archaeobotanical Lab at the Universidad Astral de Chile will help us to understand many of these matters. Moreover, we are looking forward for extending our data collection to sources currently unavailable such as the study of phytolites, starch grain in stone instruments such as mortars, pestles, knives, etc and other residues in cooking/storing vessels.

B

Dendrochronology Dendrochronology studies have mainly used conifers such as Fitzroya cupressoides, Pilgerondendrom uviferum, Austrocedus chilensis and southern beech species such as Nothofagus pumilio, due to its long life and good response to climatic changes and the potential to create long chronologies (Lara et al. 2001, 2005a; WolodarskyFranke 2002; Aravena et al. 2002). Dendrochronological studies on F. cupressoides have helped to build chronologies of 5666 years ca. However, these studies need to be extended to other evergreen species such as Nothofagus obliqua, N. dombeyi and Persea lingue (Lara et al. 2005b). These species are particularly important since many of them have been used by groups inhabiting these forests and can be used to study archaeological wood artefacts. An immediate application would be the study of two wood canoes found in the bottom of the Calafquen Lake (Cuq 2005, unpublished). This project is the first of its kind in Chile and combines dendrochronology and subaquatic archaeology (Carabias et al. 2005, unpublished). By applying dendrochronological methods we expect to date sailing technologies. Records regarding sailing technologies are well documented by the historians but hardly seen in the archaeological record. This will also help to understand forest resources and wood use, in particular, and the use of water and other natural routes.

Fig. 1. Distribution of temperate forests in the world (A) and study area in Chile (B).

Carpology The study of fruits and seeds from archaeological sites is scanty in southern Chile. The main insights into this discipline come from previous experiences in agroceramic sites in central Chile (Planella, Tagle 2004), Mediterranean Europe (Buxo 1997). Due to the many differences between these sites and the many local periods (especially hunter-gatherers groups), patterns of habitat use and inhabiting, soil types and preservation issues, many of the previously established principles needed to

Historical ecology and climatology This amalgamation of disciplines is not new in Chile. As suggested by Le Goff (1995) 20 years ago, we have noticed that historical studies have evolved and diversified in Chile, and are clearly pointing towards environmental history. Considering this trend, we have combined information from several sources; e.g. climatic history, economic 261

M.E. Solari and over-exploitation observed nowadays (Catalan and Ramos 1999; Donoso and Lara 1997; Lara et al. 1997; Fuentes 1994).

history, natural ecosystems evolution and agriculture and population history in our studies. The extent of environmental records (vegetation-climate) for the region comprising southern and Austral Chile and Argentina covers at least the last 450 years. These records have been compiled from chronicles, traveller and naturalist’s narratives, public archives, newspapers and magazines as well as maps, old letters, photographs, drawings and many other documents. However, while rereading these documents the following matters have to be kept in mind (Prieto et al. 2004, unpublished): 1-The territory was not settled and inhabited permanently, and many cities were destroyed or abandoned between the XVI and XVIII century and even later (c. XIX) for Magallanes and Coyhaique. This has caused a discontinuity in the historical record. 2-The country was not systematically explored until mid XIX century. 3-The main access to the southern territory was by sea and from the central valley; therefore many Andean regions are not included in the chronicles.

In summary, we expect that our studies will provide an interface between two sciences normally distanced; which in turn will help us to understand the evolution of the temperate forests of southern South America during the last 10000 years taking into account a multicultural intervention and perception.

Acknowledgments

I would like to thank the organising committee of the 3rd Meeting of Anthracology (Lecce, Italia), French Archaeological Mission in Patagonia (Dr. D. Legoupil), C.A. Lehnebach, A. WolodarskyFranke, A. Lara and L. Adan and the staff of the Museo Histórico Antropológico (UACh). These studies have been supported by the grants Fondecyt Grant 1040326, 1010200, 1000445, DIDUACh 199917 and 200154, WWF FB49.

References Adán, L., 2000. Citation of unpublished document. Tradición arqueológica de bosques templados en el centro-sur de Chile. Poblaciones arcaicas y formativas adaptadas a los sistemas lacustres andinos (Lago Calafquén, regiones IXª y Xª). Proyecto Fondecyt 2000-2002 Adan, L., Reyes, V., Mera, R., 2003. Ocupación humana de los bosques templados del centro-sur de Chile. Proposiciones acerca de un modo de vida tradicional. In: Colegio de Antropólogos, eds. Actas del IV Congreso de Antropología Chilena. Santiago: Lom ediciones, 1444-1455. Aravena, J.C., Lara, A, Wolodarsky-Franke, A., Villalba, R., Cuq, E., 2002. Tree-ring growth patterns and temperature reconstruction for Nothofagus pumilio (Fagaceae) forests in the upper tree line of southern Chilean Patagonia. Revista Chilena de Historia Natural, 75, 361-376. Armesto, J.J., Leon-Lobos, P., Kalin Arroyo, M., 1995. Los bosques templados del sur de Chile y Argentina: una isla biogeografica. In: J.J. Armersto, C. Villagran, M. Kalin Arroyo, eds. Ecología de los bosques nativos de Chile. Santiago: Editorial Universitaria, 23-28. Buxo, R., 1997. Arqueología de las plantas. Barcelona : Editorial Crítica. Carabias, D., Chapanoff, M., Adán, L., 2005. Citation of unpublished document. Informe de Arqueología subacuática “Sitio Dos Canoas”, lago Calafquen. Proyecto Fondecyt 1040326. Catalán, R., Ramos, R., 1999. Pueblo mapuche, bosque nativo y plantaciones forestales. Las causas subyacentes de la deforestación en Chile. Temuco: Editorial CDS-ConadiCet. Cuq, E., 2005. Citation of unpublished document. Informe dendrocronológico de muestras de canoa Lago Calafquén. Proyecto Fondecyt 1040326. Dillehay, T., 1986. Cuel: observaciones y comentarios sobre los túmulos en la cultura mapuche. Actas X Congreso Nacional de Arqueología Chilena, Arica. Chungara, 1617, 181-194.

Historiography has been helpful to understand how the original interactions “native communities – habitat” were modified in southern Chile by a combination of exogenous elements, mainly after the arrival of European settlers (Spanish, German, Yugoslavian and Italian). European settlers brought along new models of territorial occupation, transforming the primary landscape and the resources exploitation. This mixture of exogenous and local models resulted in a “hybrid” pattern of occupation, which is currently employed. Nowadays this hybrid model may be inclined towards the earlier local or native model or go to extremes, generating severe changes in the landscape such as mono-plantations, forest clear-cutting and erosion (Fuentes 1994; Catalan, Ramos 1999).

Conclusions Our studies pretend to determine the “route(s) followed by forest products” (or cultural itinerary) such as wood, fruits, seeds, etc, during the last 10000 years from an etic point of view of the social and natural sciences to the emic cultural meanings from ethnographic studies on human communities and their forested ecosystems, e.g. in the Valdivian basin. This route comprises culturally determined models that have a differential effect on the conservation of temperate forests biodiversity. These models of use/inhabiting the landscape can be local, exogenous or hybrid. This process started with the use/inhabiting of the territory by hunter-gatherers during the Archaic Period, domestication of the habitat by agro-cattle farmers groups during the Formative Period (Dillehay 1990; Adan 2000, unpublished; et al. 2003), and followed by the destruction 262

Environmental and cultural history of South American temperate forests Dillehay, T., 1989. Monte Verde: a late pleistocene settlement in Chile. Paleoenvironmental and site context. Washington: Smithsonian Institution. Dillehay, T., 1990. Araucanía. Presente y pasado. Santiago: Editorial Andres Bello. Dillehay, T., 1997 Monte Verde: a late pleistocene settlement in Chile. The archaeological context. Washington: Smithsonian Institution. Dillehay, T., 2003. Tracking the firsts Americans. Nature, 425, 23-24. Dillehay, T., 2004. Monte Verde, un asentamiento humano del Pleistoceno tardío en el sur de Chile. Santiago: Editorial Lom. Donoso, C., Lara, A., 1997. Utilización de los bosques nativos en Chile: pasado, presente y futuro. In: J.J. Armersto, C. Villagran, M. Kalin Arroyo, eds. Ecología de los bosques nativos de Chile. Santiago: Editorial Universitaria, 363386. Fuentes, E., 1994. ¿Qué futuro tienen nuestros bosques? Hacia la gestión sustentable del paisaje del centro y sur de Chile. Santiago: Editorial Universidad Católica de Chile. Gaete, N., Navarro, X., Constantinescu, F., Mera, C., Selles D., Solari, M.E., Vargas, M.L., Oliva, D., Durán, L., 2004. Una mirada al modo de vida canoero del mar interior desde Piedra Azul. Chungará, special vol. (1), 333-346. Lara, A., Donoso, C., Aravena, J.C., 1997. La conservación del bosque nativo de Chile: problemas y desafíos. In: J.J. Armersto, C. Villagran, M. Kalin Arroyo, eds. Ecología de los bosques nativos de Chile. Santiago: Editorial Universitaria, 335-362. Lara, A., Aravena, J.C., Wolodarsky-Franke, A., Villalba, R., Luckman, B., Wilson, R., 2001. Dendroclimatology of high-elevation Nothofagus pumilio forests in the Central Andes of Chile. Canadian Journal of Forest Research, 31, 925-936. Lara, A., Villalba, R., Wolodarsky-Franke, A., Aravena, J.C., Luckman, B., Cuq, E., 2005a. Spatial and temporal variation in Nothofagus pumilio growth at tree line along its latitudinal range (35º40’ – 55º S) in the Chilean Andes. Journal of Biogeography, 32, 879-893. Lara, A., Wolodarsky-Franke, A., Aravena, J.C., Villalba, R., Solari, M.E., Pezoa, L., Rivera, A., LeQuesne, C., 2005b. Climate fluctuations derived from tree-rings and other proxy-records in the Chilean Andes: state of the art and future prospects. In: U. Huber, H. Bugmann, M. Reasoner, eds. Global change and mountain regions: a state of knowledge overview. Germany: Kluwer Academic Press, 145-156. Le Goff, J., 1995. Pensar la historia. Barcelona: Editorial Altaya. Legoupil, D., 1989. Ethno-archéologie dans les archipels de Patagonie: les nomads marins de Punta Baja. Paris  : Editorial Recherche sur les Civilisations. Legoupil, D., 1995. Des indigènes au Cap Horn: conquête d’un territoire et modèle de peuplement aux confins du continent sud-américain. Journal de la Société des Américanistes, 81, 9-45. Legoupil, D., 1997. Bahía Colrada: les premiers chasseurs de mammifères marins de Patagonie australe. Paris  : Editorial Recherche sur les Civilisations. Legoupil, D., 2003. Cazadores-Recolectores de Ponsonby (Patagonia Austral) y su paleoambiente desde VI al III Milenio A.C. Punta Arenas : Magallania, 31, 327-335.

Lehnebach, C, Solari, M.E., Adán, L.I., 2005. Citation of unpublished document. Archaeological plant remains from a rock shelter in the temperate forest of Chile. Proyecto Fondecyt 1040326. Massone, M., 1989. Los cazadores de Tierra del Fuego. In: J. Hidalgo, V. Schiappacasse, H. Niemeyer, C. Aldunate, I. Solimano ed. Prehistoria de Chile, desde sus albores hasta la Conquista. Santiago: Editorial Andrés Bello. Massone, M., 1996. Hombre temprano y paleoambiente en la región de Magallanes: evaluación crítica y perspectivas. Anales del Instituto de la Patagonia, 24, 81-98. Navarro, X., 1995. Interpretación de ocupaciones precerámicas en los distintos microambientes de la costa de Chan Chan, Valdivia. Actas del XIII Congreso Nacional de Arqueología Chilena, Antofagasta. Universidad Antofagasta, 9 (1), 127-134. Navarro, X., 1997. Uso del espacio costero y de los recursos por recolectores del área extremo sur andina (Xª región, Chile), una aproximación etno-arqueológica. Actas del Segundo Congreso Chileno de Antropología, Valdivia: Colegio de Antropόlogos, tomo I, 147-154. Navarro, X., Pino, M., 1999. Estrategias adaptativas en ambientes costeros del bosque templado lluvioso de la zona mapuche. Una reflexión desde el precerámico. Actas de las III Jornadas de Arqueología de la Patagonia, Neuquén: Universidad del Comahue, 65-82. Ordoñez, A.M., 2004. Citation of unpublished document. Sociedad y bosque: su utilización y relación a través del tiempo, el caso de la población de Lanco, comuna de Valdivia, Xª región de Los Lagos. Planella, M.T., Tagle, B., 2004. Inicios de presencia de cultígenos en la zona central de Chile, Períodos Arcaico y Agroalfarero Temprano. Chungará, tomo I, 387-399. Prieto, M.R., Solari, M.E., Crouchet, J., Larroucau, A., 2004. Citation of unpublished document. Fuentes documentales para el conocimiento del clima de los últimos siglos en la región sur de Chile (latitudes 40 a 51ºS). Solari, M.E., Poblete, M.P., 1999. Este se llama, esto sirve para, esto no sé, en mi casa lo usamos, el machi lo usa: trabajando sobre plantas medicinales con los niños de la escuela de Tralahuapi. Valdivia: Editorial NuevaColor. Solari, M.E., 1990. Estudio antracológico del sitio Punta Baja 1 (Mar de Otway). Anales del Instituto de la Patagonia 19, 115-120. Solari, M.E., 1993. L’homme et le bois en Patagonie et Terre de Feu au cours des six derniers millénaires: recherches anthracologiques au Chili et en Argentine. Thesis (PhD), Université de Montpellier II. Solari, M.E., 1994. Estudio antracológico del archipiélago del Cabo de Hornos y Seno Grandi. Anales del Instituto de la Patagonia, Serie Ciencias Humanas, 22, 137-148. Solari, M.E., 1997. Détermination de charbons de bois du site Bahía Colorada. In: D.Legoupil, ed. Bahía Colorada: les chasseurs de mamifères marins préhistoriques de l’Ile Englefield (Patagonie). Paris: Editions Recherches sur la Civilisation. Solari, M.E., 2003. Analyse anthracologique du locus 1. In: Cazadores-Recolectores de Ponsonby (Patagonia Austral) y su paleoambiente desde VI al III Milenio A.C. Magallania, 31, 327-335. Solari, M.E., Bernard, V., Legoupil, D., Richard, R.P., Schoellammer, H., 2002. Palaeoenvironmental approach of Ponsonby archaeological site (Riesco Island, Chilean

263

M.E. Solari Patagonia. BAR International Series 1063. Oxford: BAR Publishing. solari, m.e., richarD, r., 2003. Les restes végétaux du site et la flore actuelle. In: Cazadores-Recolectores de Ponsonby (Patagonia Austral) y su paleoambiente desde VI al III Milenio A.C. Magallania, 31, 301-307. Solari, M.E., Lehnebach, C., 2004. Pensando la antracología para el centro-sur de Chile: sitios arqueológicos y bosque en el lago Calafquén (IX-X región). Chungará, special vol. (I), 373-380. Solari, M.E., Lehnebach, C.A., Adan, L.I., 2004. Estudios antracológicos y carpológicos en el sitio Alero “Marifilo 1” (Lago Calafquen, 39ºS). Boletin del Museo Nacional de Historia 53, 125-138. Thomasson, K., 1963. Araucanian Lakes. Acta Phytogeographica Swedish 47, 1-139. Wolodarsky-Franke, A., 2002. Fluctuaciones ambientales de los últimos 1000 años a partir de anillos de crecimiento de Fitzroya cupressoides en el área del Volcán Apagado, Décima Región, Chile. Thesis (MSc). Universidad Austral de Chile.

264

A critical assessment and experimental comparison of microscopic charcoal extraction methods R. Turner, A. Kelly, N. Roberts

School of Geography, University of Plymouth, Drakes Circus, Plymouth, PL4 8AA, UK E-mail: [email protected] Abstract Microscopic charcoal analysis has been used to reconstruct past fire activity over a range of spatial and temporal scales in Europe, the Americas and Australasia. Despite this, no uniform method of microscopic charcoal analysis has been adopted. This paper presents the results from a systematic experimental investigation into a range of charcoal extraction and quantification techniques. Existing techniques including a standard pollen preparation and the Oregon Sieving Technique were rigorously tested on a series of “control” lake sediments that contained a known added weight of charcoal. Two further techniques were developed during the course of this investigation that used heavy liquid to extract charcoal from lake sediments. Particles were quantified using the Point Count Method, total abundances and size measurements. The results suggest the need for a two-stage preparation procedure, the first stage removing the more fragile meso and macroscopic particles and the second stage utilising heavy liquid separation to extract microscopic charcoal. The separation of the different sized particles reflects their contrasting vulnerability to fragmentation during processing. Heavy liquid separation extracted the highest quantity of charcoal from the sediments. The quantification techniques highlighted for the spatial resolution required by the study to be considered when deciding the quantification procedure. Total abundance counts are more suited to the analysis of microscopic charcoal, whereas area measurements are better suited to the analysis of meso and macroscopic charcoal particles.

Introduction

Charcoal extraction

A major issue in microscopic charcoal analysis is the lack of a universally accepted method of charcoal extraction and quantification (Patterson et al. 1987; Rhodes 1998). Unlike other palaeoenvironmental proxies, e.g. pollen or testate amoebae, no specific and widely agreed extraction procedure has been developed or adopted. Micro-charcoal is often counted by palynologists interested primarily in vegetation change rather than fire activity. Charcoal analysts have realised the constraints of analysing charcoal on pollen slides, due to the fragility of charcoal and the mechanical and physical severity of this procedure (Whitlock and Larson 2001). Yet, despite other extraction procedures having been developed e.g. thin section analysis and Oregon Sieving Technique, pollen preparation remains the most frequently used extraction procedure. Furthermore, the lack of universal agreement on charcoal extraction procedures means that a range of techniques (alongside the pollen preparation) are now applied to the extraction of microscopic charcoal. Each of these techniques varies in the size range of particles analysed and the chemical and mechanical stress particles experience during processing. As a consequence, fire history records can be produced that may not be directly comparable. To date, a thorough investigation into charcoal extracting procedures has been called for but not conducted (Carcaillet et al. 2001; Tinner and Hu 2003). This paper presents an investigation based on controlled laboratory experiments conducted using published extraction procedures, and tests a new method of charcoal extraction based on density separation. Issues of charcoal identification and quantification are also addressed.

Three commonly used extraction procedures have been selected from the published literature to test the effects of preparation procedures on the fire history that is reconstructed. The methods selected are a standard pollen preparation (Moore et al. 1991), bleaching and filtering (Rhodes 1998) and the Oregon Sieving Technique (Millspaugh, Whitlock 1995). Other methods of charcoal analysis are available e.g. thin section analysis, but they were not investigated as they are not directly commensurable and therefore not amenable to comparison under controlled laboratory experimental conditions. 1. Pollen preparation. The extraction and quantification of microscopic charcoal on pollen slides is widely used as it allows pollen and charcoal counts to be obtained simultaneously with no extra sample preparation. However, this technique does have several limitations. The pollen preparation procedure focuses on particles between 10µm to 180µm. Particles larger than 180µm are removed by sieving in the early stages of processing. This creates a size bias toward the smaller size fractions and therefore over-represents extra local and regional source areas of charcoal inputs. The pollen preparation procedure is also chemically and mechanically rigorous which can result in the fragmentation of charcoal particles (Clark 1984). This increases the abundance of smaller size fractions resulting in an overestimation of the total charcoal concentration and further shifting the size bias (Clark 1984; Sarmaja-Kojonen 1991). 2. Oregon Sieving technique. Millspaugh, Whitlock (1995) developed this technique to extract charcoal from lake sediments. A sample is disaggregated and

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

R. Turner, A. Kelly, N. Roberts

Fig. 1. Diagnostic characteristics of charcoal.

then washed through a series of nested sieves. This gentle extraction technique limits potential sources of damage that can fragment charcoal particles. Likewise the emphasis on larger charcoal particles means the background component has been removed and therefore a reconstruction of localised fire activity is produced.

placed in predetermined size classes. Area measurements allow a greater emphasis to be placed on the importance of larger vs. smaller particles due to their contrasting transport potential (Batterson, Cawker 1983). However, these measures are strongly influenced by the extraction procedure used. Mechanically and chemically rigorous techniques, e.g. pollen preparations, potentially alter the size distribution of particles (Clark 1984). This may result in the misrepresentation of the charcoal record particularly if the analysis primary interest is to reconstruct the spatial scale of past fire events.

3. Bleaching & Filtering. Rhodes (1996; 1998) developed this technique of charcoal extraction based on the bleaching and then filtering of lake sediments. It aimed to provide a cheap, quick and gentle method of charcoal extraction by minimising physical and chemical stress. This method has no upper size limit and therefore no size bias, hence it produces a record of fire activity spanning a range of spatial scales. The technique is also gentle and so limits the amount of mechanical damage.

Charcoal identification Correct identification of charcoal is a problem associated with the optical analysis of microscopic charcoal. Although charcoal is described by many analysts as being jet black, angular and opaque, (Clark 1984; Clark 1988b; Patterson et al. 1987; Swain 1973; 1978; Tinner, Hu 2003), the actual identification of charcoal is subjective depending strongly on the experience of the analyst. Accurate identification of charcoal can be difficult for several reasons. There are many particles present in sediments that can resemble charcoal e.g. pyrite, dark plant fragments and insect cuticles (Clark 1984). Partial combustion also hinders identification problems; it results in a spectrum of particles present that span from charcoal to unburnt. Therefore to ensure accuracy in charcoal identification diagnostic criteria were applied to each particle, namely 1) jet black (Figs 1a, b and c), 2) angular, straight edges (Figs 1a and 1b), 3) straight but fuzzy edges (Figs 1c), 4) blue hue (Fig. 1a and 1c) and 5) the presence of cellular structure (Fig. 1b) and represented the most commonly observed characteristics (see Fig. 1). Particles that included a combination of these characteristics were classified as charcoal. All other particles were ignored.

Quantification procedures Similar to extraction procedures several quantification techniques exist which convey different information concerning past fire activity. 1. Absolute abundance measures involves all charcoal particles, regardless of their size, being quantified (Patterson et al. 1987). This assumes that all particles contribute an equal level of information to the fire history record. 2. Point Count. Charcoal abundance is recorded based on the number of “hits” of particles scored on a standard number of points on an eyepiece graticule (Clark 1982). Particles present in the field of view, but which do not “hit” a point are ignored. Again this assumes that all particles contribute an equal level of information to the fire history record. 3. Particle Area measurements can be used to infer transport distance of particles. Particle measurements e.g. longest axis or area can be taken using an eye-piece graticule or using image analysis techniques, and then particles can be 266

A critical assessment of microscopic charcoal extraction methods Sample material Fossil samples contain an unknown quantity of charcoal, and therefore the record of charcoal produced could reflect either the processing techniques or the abundance of charcoal in the sample. Therefore, for this investigation artificial lake sediment samples were produced. The control samples were comprised of a known volume of modern charcoal combined with sterile lake sediments. The charcoal used in the control samples had been manufactured from a mixture of hardwoods, using traditional methods, by the Forest of Dean Museum Heritage Trust, UK. The charcoal was sieved through a series of nested sieves (mesh sizes: 250µm, 125µm, 63µm, 50µm and 25µm) for five minutes using a Wretsch Vibro sieve shaker. These size classes were selected to be representative of the charcoal particles preserved in lake sediments that had extra-local to regional source area. Each sample contained 0.7g of sterile lake sediment. The sediments used were late Pleistocene lake marls from the Konya basin, Turkey, taken from near the archaeological site of Çatalhöyük. These sediments had previously been sieved to 1000 m asl and correspond to the Picos de Europa Eastern Massif. The cave is sheltered from the full wet west-wind regime. The resulting reduced precipitation permits the development of the Cantabrian evergreen oak formations along the calcareous slopes of the La Hermida Defile. El Castillo cave opens on the NE slopes of the Sierra del Escudo de Cabuérniga mountains at 195 m asl and 25 Km from the present shoreline. This horizontal mountain range with an altitudinal average of 500 m asl, separates the coastal plain from the mountain valley areas with a front oceanic climate. A strong rainfall regime and extensive Eucaliptus plantations are responsible for the severe degradation of the Deciduous and Evergreen oak forest formations, heathland and spiny deciduous shrub community being the main present day vegetation of the Puente Viesgo area. Covalejos cave opens on the South slopes of the Peñajorao massif calcareous outcrops at 105 m asl. This karstic formation of low altitude (300 m) separates the littoral area of the Bahía de Santander from the lower valley of

the Pas river area, 7 Km distant from the sea. Heathland formations, deciduous oak clumps with the deciduous spiny shrub community and Eucaliptus plantations are the main characters of the Peñajorao massif environment where Covalejos cave is located. Cobrante cave opens on the eastern slopes of the karstic massifs of San Miguel de Aras located at the confluence of the Clarín and Clarión rivers which are tributaries of the Ason river. This area is close to the Bahía de Santoña,12 Km from the present shoreline (eastern Cantabria). Evergreen oak vegetation developes along the lower Ason valley calcareous outcrops.

Chronostratigraphy of sites El Esquilleu El Esquilleu cave has yielded 30 stratigraphic levels in which relevant Mousterian lithic assemblages have been found becoming a cantabrian reference site for the Middle 274

Domestic fires and vegetation cover among Neanderthalians Palaeolithic (Baena et al., in press). Charcoal has been sistematically studied from level XI through level XXVII for the moment. The anthracological study thus spans from 36 kyr BP (based on AMS data from level XI) to 53 kyr BP (TL data obtained in level XXII). Lithic assemblages present several Mousterian technologies (debitage): Quina debitage was found in the Upper part of the stratigraphic record, Levels XI to XIV; Discoide debitage appears in levels XV and XVI; Levels XVII and XVIII are Levallois; Discoide debitage appears again in levels XIX and XX; and levels XXII to XXVII are Levallois debitage sensu lato with occasionally Quina debitage features appearing in some levels (J. Baena pers. com. and under study).

The levels concerned in this work (B, C, D and J-H) correspond to the Mousterian-Early Upper Palaeolithic transition. Levels B and C are respectively Aurignacian I dated to 30.3 Kyr BP and basal Aurignacian with a date as 32.8 Kyr BP (Montes et al., in press and under study). Level D is Mousterian dated to 41.6 Kyr BP. The Middle/ Early Upper Palaeolithic transition has been placed between levels C and D (Montes et al., in press), but it is impossible to know exactly the date at which this transition took place. Solifluxion processes occured at the edge of both levels and the date of 40.6 Kyr BP obtained at the D and C conctat zone indicates an hiatus of 8 Kyr. H-J complex is Mousterian with a TL datation of > 42 Kyr BP (R. Montes pers. com.).

El Castillo

Cobrante

El Castillo cave has yielded a very important stratigraphic record (26 levels) in which the most important Palaeolithic cultural periods have been found (Cabrera Valdés, 1984). This study concerns level 18 subdivided in 18B and 18C, and level 20 partially excavated. Level 18 corresponds to the transitional Aurignacian, first cultural manifestation of the Early Upper Palaeolithic, called “Auriñaciense de transición de tipo Castillo” according to Cabrera et al. (2001). AMS data obtained has yielded an average age of 40 kyr BP for the Middle/Early Upper Palaeolithic cultural transition (Cabrera Valdés, 1989, 1993; Bernardo de Quirós, Cabrera, 1993, 1996; Cabrera, Bichoff, 1989; Cabrera et al., 1996; Hedges et al., 1994; Rink et al., 1996). Level 20 corresponds to the Mousterian. AMS data have yielded an age around 42 kyr BP (H. Valladas, pers. com.). The interpretation of the archaeological context has been made only in level 18. The strong bone concentration and lithic assemblages found in level 18B, located at the entrance of the cave, has suggested a primary butchering area where shredding activities involving hunted animals were carried out (Pike Tay et al., 1999). Level 18C, located towards the interior of the cave in a marginal zone, has yielded a concentration of bones, charcoal and lithic assemblages very similar to those that were found in the central area of the cave (characterized by the presence of domestic hearths), during excavations made at the beginning of XX century, permitting correlations to be established. The 18C occupation level has been interpreted as a hearth sweeping area or “Aire de rejet”, according A. Leroi-Gourhan’s ethnoarchaeological nomenclature (Cabrera et al., 2001; Pike Tay et al., 1999), in which different assemblages (bones, charcoal and lithic) were accumulated as a result of the sweeping of several hearths located in the central area of the cave.

Seven stratigraphic levels with different Upper Palaeolithic human occupations have been found in this cave (Rasines, under study). The present study only concerns the lower part of this record. The information concerning Level 7, the base of the stratigraphy, is unknown at the moment of writing this paper. Levels 6 and 5 are Aurignacian with radiometric dates of 30 Kyr BP and 33 Kyr BP respectively. Most of the anthracological information comes from level 6.

Material and Methods Charcoal has been systematically sampled in all sites by manual flotation and sieving of all sediment collected in the excavation. Charcoal sampling techniques have been standardized and integrated in the methodology of excavation of caves and rockshelters along the whole Iberian Territory (Uzquiano, 1992a, 1994, 1997; Badal et al., 2003). Charcoals were manually fractured to provide transversal, tangential and radial sections. Botanical determination follows the keys for fresh wood (Greguss, 1959; Jacquiot, 1955; Jacquiot et al., 1973; Schweingruber, 1977, 1990) as well as the Charcoal Identification Guide (Vernet et al., 2001). A reference collection of modern charcoal aided the determination in cases where the identification based on the keys was unclear. Charcoals were observed by reflected light microscopy. Some specimens were further examined using scanning electron microscopy. The nomenclature employed follows Flora europaea’s criteria (Tutin et al., 1964). The number of charcoal fragments determined for each site depends on the degree of conservation of charcoal in the sediment, on the surface area excavated and on the nature of the human occupations of each level excavated. El Esquilleu cave has yielded 1341 fragments of charcoal with an irregular distribution by level. The most relevant information comes from 7 occupation levels: XI, XIII, XIV, XVIII, XIX, XX and XXII respectively. El Castillo cave has yielded 3939 fragments of charcoal, 2967 of which correspond to level 18; 896 fragments of

Covalejos The entire stratigraphic record has yielded 12 levels of human occupation. 275

P. Uzquiano charcoal have been sampled in level 20. Covalejos cave, 579 fragments (levels B, C, D and H-J), and finally, 309 fragments of charcoal have been counted in the lower occupation levels of Cobrante cave (7, 6 and 5), 222 of which correspond to level 6.

Results

unedo, Laurus nobilis, Prunus spp. and Hippophae rhamnoides calciphilous shrubs (Fig. 4). C and D transitional levels show Betula dominant while a clear alternation of Pinus-Betula has been observed in level B Aurignacian and H-J Mousterian, indicating a duality of the management of plant territories: very humid soils areas (Siliceous) and other zones with drier substrate (Calcareous).

El Esquilleu

Cobrante

Charcoal information has been subdivided in the upper levels XI, XII and XIV called the Quina complex (Fig. 2A), and the lower levels XVII, XVIII, XIX, XX and XXII called the Levallois complex (Fig. 2B) according to the floristic differences observed between upper and lower levels. Pinus sylvestris is largely dominant in the lower levels. From level XIV onwards Pinus decreases in favour of other tree species like Sorbus aria that becomes the main woody taxon in level XI. A great variety of shrubs appears along all these levels. This variety has been grouped by substrate-type affinity. On one hand there are heathland species: Leguminosae, Ulex, Cytisus and Erica tetralix, that probably came from the siliceous elevations surrounding the site; on the other hand, Arbutus unedo, Rhamnus alaternus, Laurus nobilis, Prunus spp. and Crataegus monogyna, came from the calcareous slopes. Both communities appear in important amounts in the Quina upper levels, while their presence is masked by high amounts of Pinus in lower levels. The alternation of Calcareous substrate and Siliceous substrate shrubs has been clearly recorded in the Quina levels, while Heathland species are more characteristic and regular in the lower levels (Figs. 2A and 2B). Corylus and Fraxinus as well as sporadic occurences of Salix indicate the proximity of the Deva river.

The anthracological results obtained in this cave come only from level 6 Aurignacian occupation (Fig. 5). Again high Betula masks the presence of other taxa: Pinus sylvestris, Sorbus aria, Hippophae rhamnoides, Arbutus unedo, Pistacia sp. and Leguminosae. All these taxa indicate the existence, frequentation and exploitation of different ecological biotopes to obtain the firewood that humans needed for their domestic fires carried out during the several seasonal occupations recorded in all these caves. Betula or/and Pinus have been systematically employed as main firewood in all fireplaces. Heathland species Cytisus, Ulex, also appears systematically but in lower amounts. These taxa alternate with the calcareous shrubs in some occupations.

Discussion A- Vegetal landscape during the Middle Pleniglacial Pinus and Betula indicate the existence of open environments where these pioneer species extended on poor soils. The variety of shurbs confirms the co-existence of areas with very humid soils adjacent to others territories with a drier substrate over the catchment area of prehistoric sites. Both substrates have been exploited by humans alternately in each occupation of these caves. East-West position of mountain ranges and North-South orientation of hydrologic system cutting transversally the elevations, result in a strong compartimentation of the territory. This discontinuity of substrat has produced a strong contrast of vegetation communities over very short distances. This geographical proximity of different vegetation communities has permitted their alternate exploitation along prehistoric times in different territories of the Iberian Peninsula (Uzquiano, 1992a, 1997, 2002, and in press). Taking into account the ecology of the main taxa identified, Birch communities develope along siliceous substrates on the northern slopes of the cantabrian mountain range between 800 and 1600 m asl with Sorbus aria and Quercus petraea (Mayor, Díaz, 1977; Cendrero et al., 1986; Aseguinolaza et al., 1989). Natural Pinus sylvestris forests

El Castillo Charcoal results obtained at this site show Betula dominant in all levels (Fig. 3), followed by Sorbus aria with significant amounts in 18C. This level shows a particular alternation of these two siliceous taxa. Pinus sylvestris is insignificant throughout level 18, but slightly higher in level 20. Leguminosae and Hippophae rhamnoides are present in low amounts. Covalejos The results show Pinus and Betula as main taxa. Sorbus aria and some deciduous species such as Quercus robur, Castanea sativa and Corylus avellana appear in low and sporadic amounts. The great variety of shrubs has been grouped by substrate affinity. Leguminosae, Ulex, Cytisus, Erica sp., acidophilous taxa, alternate with Cornus sanguinea, Sambucus nigra, Rhamnus alaternus, Arbutus 276

Domestic fires and vegetation cover among Neanderthalians ESQUILLEU : Upper levels' anthracoanalytical results 70

60

50

Pinus sylv. Betula Sorbus Lande Shrub. Karst. Shrub. Deciduou sp.

40

30

20

10

0

XI-F

XIII Mousterian QUINA type complex

XIV

Fig. 2A . El Esquilleu’s anthracological results from Quina Upper levels.

ESQUILLEU :Lower levels' anthracoanalytical results 100 90 80 70 60

Pinus sylv. Betula Sorbus Lande Shrub. Karst. Shrub. Deciduou sp.

50 40 30 20 10 0

XVII

XVIII

XIX

XX

Mousterian LEVALLOIS type complex

Fig. 2B. El Esquilleu’s anthracological results from the Lower levels.

277

XXII

P. Uzquiano are absent in the Cantabrian region (Northern slopes of the Cantabrian mountain range), but some natural Pinus sylvestris communities remain on several Southern slopes areas of the Cantabrian Cordillera (around 1000m asl) with optimal position on reliefs and good edaphic conditions (Bertrand 1974). Pinus sylvestris formations are also located in the Southern slopes of the Pyrenees mountain range between 1000-2000 m asl (Dupias 1985). Thus we have a more than 1000m of altitudinal shift of the charcoal flora with respect to the present vegetation period. Why this flora and ecology types? Radiometric data obtained in these prehistoric caves has permitted the assignment of the charcoal flora and the human occupations to the Middle Pleniglacial. The western European continental and marine palaeoecological records as well as the Greenland ice core record have revealed stadial-interstadial oscillations alternations. Interstadial oscillations have been variable, of more or less short duration (Behre, van der Plitch 1992; Johnsen et al. 1992; Dansgaard et al. 1993; Sánchez Goñi et al., 2000). Western European pollen vegetation dynamics recorded in continental deposits (Behre 1989; Woillard 1978; Woillard, Mook 1982; de Beaulieu, Reille, 1984, 1992; Reille and de Beaulieu, 1990; Pons et al. 1992; Jalut 1990; Jalut et al. 1988; 1992) have yielded Betula and Pinus dominance indicating a pioneer and heliophilous vegetation that developed on poor and gradually ice-free soils. The flora recorded in the Cantabrian prehistoric habitations present similar trends and could be also the consequence of environmental oscillations directly related to the variations in humidity supply recorded in marine record (Bond et al. 1993; Sánchez Goñi et al., 2000). Cantabrian Prehistoric occupations developed between 30 and >40 Kyr BP are contemporaneous with mountain glaciers (Butzer 1981; Moniño et al., 1988). Southwestern European glacial dynamics (northern and southern slopes of Pyrenees and Cantabrian mountains) have indicated that the beginning of deglaciation processes occurred before 30-35 Kyr BP (Hubsman, Jalut 1989; Jalut 1990; Jalut et al. 1982, 1988, 1992; Andrieu et al. 1988; Montserrat 1992; Vilaplana, 1983; Castañón, Frochoso 1996, 1998; Serrano 1995; Serrano, Martínez de Pisón 1994; García Ruiz et al. 2003; Jiménez, Farias 2002). The open and discontinous vegetation cover is the consequence of the presence of these mountain glaciers in proximity of prehistoric occupations. Glaciological research carried out in the Castro Valnera Mountains located in the southeastern part of Cantabria (1707 m asl), has yielded evidence of moraine deposits at low altitudes (600-340 m) in the headwater of the Asón, Miera and Pas rivers (Moñino et al. 1988; Castañón, Frochoso 1996). This extension of ices at low altitudes has been related to a high snowfall regime that would have fed the glacier, stopping consequently the permafrost phenomena (Serrano 1995; Castañón, Frochoso 1996). The Asón, Miera and Pas rivers (Fig. 1) would have remained under

humid periglacial conditions along the middle and lower parts of their repective valleys. Most of the CentralEastern Cantabrian middle mountain territories, as well as most of coastal plain areas, would have benefitted from these high environmental and edaphic humidity amounts coming from snowfall. Some snow amounts would have not remained in summer producing an important melting water amounts and consecuently the existence of very humid and swamp soils (M. Frochoso, pers. com.). This high moisture supply could explain the very abundant Betula in the environment developing on all these humid and swamp soil areas. Most of the catchment areas of El Castillo, Covalejos and Cobrante caves (Fig. 1) would present these palaeobiogeographic characteristics which would determine the way in which these plant communities were managed by humans groups (Figs 3, 4, 5). Taking into account cyclicity and irregularity of climatic oscillations and the state of development of the soils, these Betula plant communities would have constituted a climax vegetation at different times during the Middle Pleniglacial period, becoming the essential fuelwood for human groups that lived and moved along the middle and lower valleys of Central-Eastern Cantabria. Middle Pleniglacial Cantabrian palaeoenvironmental conditions could be compared to present day Northern Norway environmental trends, where Birch communities are extensively exploited by their inhabitants, the Sapmi. The glacial and periglacial processes in Western Cantabria studied by Castañón, Frochoso (1998), do not seem contemporaneous with the period considered in the present study. Glacial dynamics of this area indicate a higher altitude of icefield, and a more extensive permafrost. The existence of a strong rain shadow effect along the la Hermida Defile and La Liébana region has also been noticed by these authors, taking into account the peculiar disposition of the mountain range. Humidity was much lower in all this western area than in the Central-Eastern Cantabria. The greater karstic development (drier than the siliceous substrate) and a lower environmental humidity in this entire area would have made possible the development of Pinus sylvestris along the slopes of the La Hermida defile, which became the essential fuelwood for El Esquilleu cave’s inhabitants during most of the human occupations recorded in this prehistoric settlement, especially in the lower occupation levels (Fig. 2B). Pinus sylvestris communities developed along karstic slopes of the La Hermida defile (Western Cantabria), while Betula communities were the vegetation cover that mostly extended along large areas of the middle and lower part of the Pas river valley (Central Cantabria) and the lower part of the Asón basin (Eastern Cantabria), at different times during the Middle Pleniglacial. Nevertheless the presence of other humid siliceous substrate species has been recorded in this western Cantabrian site. Sorbus aria become significant in the El Esquilleu cave throughout the Quina upper levels (Fig. 278

Domestic fires and vegetation cover among Neanderthalians 2A), heathland species have been regularly recorded in all human occupations of this cave (Figs 2A and 2B). On the other hand some Covalejos’s occupations levels have recorded an alternation of Pinus and Betula (Fig. 4) that indicates a greater importance of Pinus sylvestris in the catchment area of this site. This indicates the existence of drier areas where Pinus and Betula could have lived together (Fig. 4).

in these upper occupation levels indicate a higher floristic diversity of firewood employed in fireplaces as well as an increase in the firewood source areas. The decrease in Pinus is probably related to palaeoenvironmental events that would have caused the fragmentation of their regular distribution in Northern Spain during the Middle Pleniglacial. Taking into account this possible environmental event, we want to outline the readjustments produced in the firewood management modality of this site. The decrease in Pinus supposes an important loss of firewood biomass availability around El Esquilleu. This loss was compensated by increasing the variety of woodfire species, specially those of ignition (shrubs). El Esquilleu’s upper levels record a greater exploration and exploitation of slopes and valleys surrounding this site (the alternation of calcareous and siliceous substrates along the La Hermida Defile), coinciding with changes in lithic technologies as well as in hunting strategies carried out also in these upper levels (Baena et al., in press). Mobility and human occupation functionality differs with regard to the lower levels. The dominance of Betula seems to be the main floristic trend of Central-Eastern Cantabrian sites. Nevertheless significant amounts of Pinus sylvestris have been recorded in some Covalejos cave human occupations: the alternation of Pinus-Betula recorded in the Mousterian lower levels H-J as well as in the Aurignacian level B (Fig. 4). In like manner, more abundant Pinus sylvestris amounts have been also recorded in the Mousterian level 20 of El Castillo cave (Fig. 3).

B- Woodfire management and woodfire supply areas B-1. Pinus sylvestris variations Differences in the amounts of the woody species indicate changes in the way of collecting wood and in the firewood source areas. These changes in the firewood management are related not only to palaeoenvironmental events but to the way of life and the activities recorded in each human occupation as well as the mobility around the sites too. The type of firewood collected and the firewood source areas recorded in all these human occupations will be considered in this chapter. The domestic fires in the lower human occupation levels at El Esquilleu used Pinus sylvestris firewood exclusively (Fig. 2B). The dominance of Pinus seems to be unchanging until level XIV. From this level to the top Pinus values gradually decrease while those of Sorbus aria increase, the latter becoming dominant in level XI (Fig. 2A). The greater values for Betula, Juniperus and shrubs recorded

EL CASTILLO : Anthracoanalytical results 90 80 70 60

Pinus sylv. Betula spp. Sorbus spp. Lande Shrub. Karst. Shrub. Hippophae Corylus

50 40 30 20 10 0

Level 18B

Aurignacian Levels

Level 18C

Level 20 Mousterian

Fig. 3. Charcoal results obtained at El Castillo cave’s Middle-Early Upper Palaeolithic levels.

279

P. Uzquiano All these variations in the firewood manegement would be directly related to changes in the subsistence strategies carried out at these sites: changes on the pattern of species hunted pattern noted for exemple in some levels of the Covalejos cave (R. Montes, pers. com.). But we want to underline that such variations in the firewood management have shown some Pine clumps in co-habitation with the dominant Birch communities in two Pas valley areas. These traces of Pine recorded between 40-30 Kyr BP, perhaps represent relict communities resulting from prior palaeoenvironmental events. Charcoal analysis carried out in level 22 of El Castillo cave (Uzquiano 1992a), undated but with a reference U/Th datation of 89 Kyr BP obtained in the subjacent stalagmitic layer, has yielded an alternation of Pinus-Betula clearly favourable to Pine. Environmental and Human factors have been interlacing constantly. It is possible that the alternation in the firewood management carried out along these human occcupations was oriented towards the preservation of all the firewood species availability in the catchment area of the sites, specially in those moments of the reduction extent of one of the main arboreal taxa recorded in charcoal spectrum of the caves. B-2. Birch economy ethnobotanical aspects This tree is at present extensively exploited in all Scandinavia (Polunin 1976; Noël, Bocquet,1987; Rivera, Obón de Castro 1991). Its wood has good qualities for the manufacture of all kind of objets and tools; its bark has been employed to insulate and preserve houses from humidity, and it can also produce a kind of flour to nourish populations during periods of famine. Resins can be obtained from its wood. Birch sap is employed for the fabrication of some beverages and potions. Birch wood as firewood has high caloric values and produces a strong and white smoke with antiseptic properties. It is extensively employed in some regions of Northern Norway (Finnmark) by Sapmi in order to stop the abundant plagues of mosquitos resulting from the very humid environmental conditions of these regions. Birch fires remain lit permanently and the smoke produced acts as a barrier protecting the habitat from these troublesome plagues. Taking into acount the relative abundance of Betula during the Middle Pleniglacial we can suppose a larger use of this tree by palaeolithic human groups that moved around the Pas and the Asón valleys even if we only have charcoal and not other Birch archaeobotanical remains. The ethnoarchaeological interpretation of El Castillo’s level 18B (Cabrera et al. 2001) lets us suppose that the more external fireplaces would have worked as a barrier in order to ameliorate the conditions of work carried out at the entrance of the cave: shredding activities around dead animals, and to keep at a distance of all predators and other external agents that could have made this work

difficult (Fig. 3). Strong smoke would produce good hygienic conditions for all these butchering activities permitting a longer conservation of the flesh. Level 18’s inhabitants would have also benefitted from these high caloric fires to ameliorate the conditions of life in this cave (Uzquiano in press). Strong birch smoke could be toxic, resulting inappropriate in some areas with deficient ventilation. This particularity could be related to the significant amounts of Sorbus aria recorded in El Castillo’s level 18C (Fig. 3). The alternative use of Sorbus would have served to attenuate the intensity of smoke in the most internal areas of this habitat (Uzquiano in press). Covalejos and Cobrante caves are not provided with such ethnoarchaeological context information but taking into account all Birch smoke properties it could be possible that some fires developed in these habitats (Figs 4 and 5) could also have worked as a barrier against all external agents that could have made difficult the ways of life developed in these sites as well as protected these occupations from the own cold and humid conditions of caves. Birch management radiometric data span between 42-40 Kyr BP (levels 20 and 18 of El Castillo cave and level D of Covalejos cave), and between 33-30 Kyr. BP (Level C of Covalejos and level 6 of Cobrante cave). B-3. Management of Shrub communities The Shrubs identified reflect better the calcareoussiliceous alternation of substrates and also confirm the open character of the Middle Pleniglacial vegetation around these caves. Covalejos (Fig. 4) and El Esquilleu (Figs 2A and 2B) caves are the best examples of this alternation of shrubs while their presence is practically irrelevant at El Castillo (Fig. 3) and Cobrante (Fig. 5) caves taking into account the high amounts of the main arboreal taxa. The H-J Mousterian occupation levels of Covalejos cave have recorded a significant presence of calcareous substrate shrubs in addition to pine and birch. Heath shrubs are more characteristic in the level D final Mousterian occupation, indicating an inversion of calcareous-siliceous shrubs values with regard to the precedent level. Calcareous-siliceous shrubs values of level C are similar to those of level D. Overall shrubs frequencies are very low in these two levels (C and D) where the highest Betula amounts have been recorded. Heath shrubs are clearly dominant in level B, besides Pine and Birch. This site has recorded a gradual preference for siliceous substrats (Heathland species, Betula and Sorbus) and acid soils (Quercus robur, Castanea sativa) during the cultural transition to the Upper Palaeolithic (Fig. 4). The presence of Quercus and Castanea confirms not only the preference for these very humid soils but the existence of mesophilous refugia in the catchment area

280

Domestic fires and vegetation cover among Neanderthalians COVALEJOS : Anthracoanalytical results 90 80 70 60 Pinus

50

Betula Sorbus

40

Lande Shrub Karst. Shrub

30

Hippophae Deciduou sp.

20 10 0

B

C

D(+E)

H-J

Cultural Levels

Fig. 4. Covalejos cave : Transitional Upper levels floristic data.

COBRANTE : Anthracoanalytical results

90 80 70 60 50

Pinus sylv. Betula Sorbus

40

Lande Shrub. Karst. Shrub.

30 20 10 0 Niv. 6 Aurignacian

Cultural Level

Fig. 5. Cobrante cave’s level 6 charcoal results.

281

P. Uzquiano Conclusions

of this site too. The calcareous-siliceous alternation is clearly marked in the Quina upper levels of El Esquilleu cave (Fig. 2A). This ensemble has also recorded an increase of firewood shrubs coinciding with gradual Pinus decrease. High amounts of calcareous substrate shrubs has been recorded in the XIV and XI levels, almost reaching those of Sorbus in this last level. Heath shrubs are more characteristic in level XIII. The increase in exploitation of shrubs in these upper levels has served to compensate a possible deficiency in the firewood supply, as indicated by the gradual decrease in Pinus. The exploitation of shrubs doesn’t seem important in the lower levels of this site (Fig. 2B) taking into account the Pinus high amounts recorded. Nevertheless heath shrubs appear more regularly throughout all this lower occupations. The regular occurence of heath shrubs in all these humans occupations suggests how relevant siliceous substrate is around the catchment area of these caves. This characteristic has appeared along other palaeolithic and postpalaeolithic anthracological studies carried out in the Cantabrian region (Uzquiano 1992a, 1995). This acid- substrate exploitation “preference” is related to the very humid environmental conditions that have determined the state of development of the soils. The shrub taxa identified from both substrates are considered good kindling species. Their use to start fires by cantabrian mountain shepherds has always been noted in Northern Spain ethnographic record (Aseguinolaza et al., 1989; Rivera, Obón de Castro 1991). Arbutus, Prunus spp., Crataegus, Hippophae rhamnoides produce edible fruits, and their presence in the anthracological spectrum of these sites would be also related to the collection of fruits. The period of fructification of them, Fall-Winter, almost coincides with seasonality studies carried out in some Cantabrian caves such as El Castillo (Pike Take et al. 1999). Hippophae rhamnoides charcoal remains coincide with Pike Take’s seasonal estimations. Their fruits are edible and rich in vitamine C (Bazile-Robert 1983). In a palaeoecological point of view this taxon is a good indicator of open environment vegetation. Its presence not only suggest an immediate catchment area (located on calcareous massifs sourounding the sites) but also a more distant catchment area, related to the winter frequentation of littoral areas to collect molluscs (Uzquiano 1992b). The shrubs communities of the cantabrian territory constituted an important firewood biomass that would have permitted the existence and development of human occupations located along these valleys during the Middle Pleniglacial period. Domestic fires were started with different shrub taxa coming from both calcareous and siliceous substrates, geographically closed inside the catchment area of the sites. The increase in their use recorded in some occupation levels has been related to a decrease in the biomass of other arboreal firewood taxa.

The flora of open environments determined by charcoal analysis has indicated that the development of prehistoric occupations was contemporaneous with the nearby mountain glaciers, some of which had started their deglaciation processes. The retreat of ices from middle altitudes of southwestern european mountain areas (prior to 30 Kyr BP) would have produced series of oscillations in the humidity supply, and consequently a climatic instability. Different preforest vegetation cycles would take place repeatedly between >53 and 30 Kyr BP. Taking into account the Middle Pleniglacial climatic instability and the state of development of soils, this vegetation type could be considered as climax for this period. Some different palaeoenvironmental and palaeofloristic trends have been observed from west to east of the Cantabrian territory which are responsible for the geographic distribution of the main tree species in the catchment area of prehistoric sites concerned. The western Cantabria sheltered conditions (Liebana region and the Deva Valley) would have stopped the wet west wind causing a marked decrease in the precipitations and in the humidity supply. Pinus sylvestris communities extended along a drier calcareous substrate. Nevertheless some edaphic humidity has also been noted (bottom of valleys) based on the presence of mesophilous taxa and the regular presence of heath shrubs. The Calcareous and siliceous substrate discontinuity, geographically close to each other, have partially explained these floristic contrasts and their availability to be managed. The Central-Eastern areas of Cantabria (Pas, Miera and Asón Valleys) have remained under the influence of the Castro Valnera glacier. The greatest humidity (environmental and edaphic) of this area was analogous to that recorded at the present day. The Castro Valnera massifs and the upper part of these valleys, record the highest rainfall index in Cantabria (Cendrero et al. 1986). The only difference is the nature of the precipitation: present day precipitations are rainfall while they would be snowfall during the Middle Pleniglacial. The snow melting would produce high environmental and edaphic humidity, explaining the great extent of Birch communities and the development of heath species in these areas. Nevertheless, significant calcareous outcrops extended in this central-eastern part of Cantabria, permitting the development of shrubs communities with such affinities and offering perhaps more optimal conditions to the dispersal of Pinus clumps. The latter would also be mixed into the Birch communities extent. Cantabrian Upper Pleistocene oscillations in the humidity supply has been indicated in previous works (Uzquiano, 1992a). The Greenland icecore and the North Atlantic marine records carried out during the last decade of XXth century (Dansgaard et al., 1993; Johnsen et al., 1992; Bond et al., 1993; Sánchez Goñi et al., 2000) confirm s 282

Domestic fires and vegetation cover among Neanderthalians it, permitting easier correlations between continental and marine records today. These variations of the humidity supply have influenced the seasonal variations of economic resources searched by humans as well as the changes in subsistence strategies carried out along the Mousterian and at the transition to the Aurignacian occupations. Human groups modus operandi concerning the woodfire management has been characterised some times by the alternative exploitation of different vegetal communities catchment areas (Levels B and H-J of Covalejos cave as well as the Upper Mousterian levels of El Esquilleu cave). Some other times, the woodfire management was charactericed by the strong exploitation of Betula (Covalejos C and D transitional levels; El Castillo’s levels 18 and 20 and Cobrante level 6) or Pinus (El Esquilleu’ s lower levels). The selection of wood is not casual but it is always influenced by a clear human will that is also under the influence of environmental factors (Uzquiano 1992a, 1994, 1997, 1999). Humans know which resources they want, where to find and how use them, but humans also depend on the potentiality of resources that an environment can offer them, developing different adaptation strategies to it. The nature of environment during Middle Pleniglacial has been changing as demonstrated in other cantabrian upper Pleniglacial and Tardiglacial prehistoric sites (Uzquiano 1992a, 1995). Resources have been in deplacement or have suffered seasonal variations, and humans have moved along the cantabrian territory in search of them. The collection of wood as human activity was also under the influence of all these factors, and that explains the different woodfire management strategies carried out by humans in the catchment area of each site. An anthracological study has always a double interpretation in which environmental and human factors are interlacing constantly. Middle Pleniglacial charcoal data have revealed open environments perfectly adapted to the faunistic assemblages (Large mammiferes) that lived and moved there, as well as to all different adaptation strategies employed by human groups for hunting, collecting wood, fruits, etc., even if they were Neanderthalians or Anatomically Modern Humans.

information concerning the Pyrenean and Cantabrian glaciation/ deglaciation patterns. Thanks finally to PhD Alicia Lesnikowska for her assistance with the english text.

References ALLORGE, P. 1941. Essai de synthèse phytosociologique sur le Pays Basque. Bull. Soc. Bot. Fr., 88, p. 291-356.

ANDRIEU, V., HUBSMAN, J., JALUT, G., HERAIL, G. 1988.

Chronologie de la déglaciation des Pyrénées françaises. Dynamique de sédimentation et contenu pollinique des paléolacs: application à l’interprétation du retrait glaciaire. Bull. A.F.E.Q., 2-3, p. 55-67. ASEGUINOLAZA, C., GOMEZ, D., LIZAUR, X., MONTSERRAT, G.; MORANTE, G., SALAVERRIA, M.R., URIBE-ECHEBARRIA, P.M. 1989. Vegetación de la Comunidad Autónoma del País Vasco. Vitoria-Gasteiz, Ed. Gobierno Vasco. BADAL, E.; CARRIÓN, Y., RIVERA, D., UZQUIANO, P. 2003. La Arqueobotánica en cuevas y abrigos : objetivos y métodos de muestreo. En R. Buxó y R. Piqué dir. : La recogida de muestras en Arqueobotánica : objetivos y propuestas metodológicas. La gestión de los recursos vegetales y la transformación del paleopaisaje en el Mediterráneo Occidental. Encuentro del Grupo de Trabajo de Arqueobotánica de la Península Ibérica (G.T.A.P.I.), p. 19-29. Barcelona/Bellaterra, 29-30 Nov./1 Dic. 2000. BAENA, J., CARRIÓN, E., MANZANO, I., REQUEJO, V., VELÁZQUEZ,

R., RUIZ, B., SANZ, E., SÁNCHEZ, S., UZQUIANO, P., YRAVEDRA, J. in press. Ocupaciones musterienses en la

comarca de Liébana (Occidente de Cantabria, España). Actas de la IV Reunión Nacional de Geoarqueología. Almazán, Soria. 16-18 Septiembre 2002.

BAENA, J., CARRION, E., RUIZ ZAPATA, B., SESE, C., YRAVEDRA, J., JORDA, J., BREZ, S., UZQUIANO, P., VELAZQUEZ, R., MANZANO, E. in press. Paleoecología y comportamiento humano durante el Pleistoceno superior en la comarca de La Liébana: la secuencia de la cueva de El Esquilleu (Occidente de Cantabria, España). Actas de la Mesa de Trabajo Neandertales cántabros, estado de la cuestión. Museo de Altamira, 20-22 Octubre 2004. BAZILE-ROBERT, E. 1983. Flore et végétation du Magdalénien final de la grotte des Eglises (Ussat, Ariège) d’après l’analyse anthracologique. Bulletin de la Société Préhistorique de l’Ariège, XXXVIII, p. 87-90. BEAULIEU, J.L., REILLE. M. 1984. The pollen sequence of Les Echets (France): a new element for the chronology of the Upper Pleistocene. Géographie physique et Quaternaire, vol. XXXVIII, nº 1, p. 3-9. BEAULIEU, J.L., REILLE. M. 1992. The last climatic cycle at La Grqande Pile (Vosges, France). A new pollen profile. Quaternary Science Reviews, vol. 11, p. 431-438. BEHRE, K.E. 1989- Biostratigraphy of the Last Glacial Periodmin Europe. Quat. Sci. Rev., 8, p. 25-44. BEHRE, K.E., van der PLICHT, J. 1992. Towards an absolute chronology for the Last Glacial Period in Europe : radiocarbon dates from Oerel, Northern Germany. Veget. Hist. And archaeobot., 1, p.111-117. BERNALDO DE QUIRÓS, F. CABRERA VALDÉS, V. 1993. Early Upper Palaeolithic industries of Cantabrian Spain. In H. Knecht; A. Pike-Tay and R. White eds. : Before Lascaux: the complex record of Early Upper Palaeolithic. Boca

Acknowledgements

I would like to express my gratitude to the persons and institutions that have contributed to the making of this work. The Altamira Museum have financed the Cobrante and Covalejos charcoal studies. Thanks to Ramón Montes, Pedro Rasines, Javier Baena, Victoria Cabrera and Federico Bernaldo de Quirós, responsibles of the Covalejos, Cobrante, El Esquilleu and El Castillo’s archaeological excavations, for making available the charcoal remains to study and for all helping information about the archaeological context of sites concerned. Thanks also to Pr. Manuel Frochoso and PhD Penélope González for their useful

283

P. Uzquiano à l’étude du sécteur Ibéro-Atlantique. Toulouse, Doc. Carte Prod. Végétation.

Ratón, CRC Press.

BERNALDO

DE

QUIRÓS, F., CABRERA VALDÉS, V. 1996. Raw

GARCIA RUIZ, J.M.,VALERO GARCÉS, B., MARTÍ BONO, C., GONZÁLEZ SAMPÉRIZ, P. 2003. Asynchroneity of

material in the Palaeolithic of cueva del Castillo and in the Cantabrian region. In N. Moloney; L. Raposo; M. Santonja eds.: Non-flint stone tools and the Palaeolithic occupation of the Iberian Peninsula. BAR Publishing, B.A.R. International Series, 649, p. 21-32. BERTRAND, G. 1974. Essai sur la systématique du paysage : les montagnes cantabriques. PhD. Dissertation. Université de Toulouse. 3vols.

maximum glacier advances in the central Spanish Pyrenees. Journal of Quaternary Science, 18, fasc. 1, p. 61-72. GREGUSS, P. 1959. Holzanatomie der Europaïschen Laubhölzer und Sträucher. Budapest, Akademikai Kiado.

HEDGES, R.E.M., HOUSLEY, R.A., BRONK RAMSEY, C., van KLINEN, G.J. 1994. Radiocarbon dates from the Oxford

BOND, G., BROECKER, W., JOHNSEN, S., McMANUS, J., LABEYRIE, L., JOUZEL, J., BONANI, G. 1993. Correlations

AMS System: Archaeometry Datelist 18. Archaeometry, 36, fasc. 2, p. 337-374. HUBSMAN, J., JALUT, G. 1989. Glacier pyrénéen, versant nord/versant sud (Ossau-Gállego; Garonne-Noguera Ribagorçana). Livre-guide de l’excursion de l’AFEQ. Toulouse, 3-7 Mai 1989. JACQUIOT, C. 1955. Atlas d’Anatomie des Bois de Conifères. Texte et Planches. Paris, Centre Technique du Bois, CNRS. JACQUIOT, C., TRENARD, Y., DIROL, D. 1973. Atlas d’Anatomie des Bois des Angiospermes (Essences Feuillues). Texte et Planches. Paris, Centre Technique du Bois, CNRS. JALUT, G. 1990. Le paléoenvironnement de la moitié occidentale des Pyrénées de 40000 BP à l’Actuel  : étapes de la déglaciation et histoire de la végétation In F. Mª Ugarte and A. Cearreta eds. : The Late Quaternary in the Western Pyrenean region, p. 67-78. Vitoria-Gasteiz, Universidad del País Vasco.

between climate records from North Atlantic sediments and Greenland ice. Nature, 365, p. 143-147. BUTZER, K.W. 1981. Cave sediments, upper Pleistocene stratigraphy and Mousterian facies in Cantabrian Spain. Journal of Archaeol. Sc., 8, p. 133-183. CABRERA VALDES, V. 1984. El yacimiento de la Cueva de El Castillo. Madrid, Biblioteca Praehistórica Hispana, XXII. CABRERA VALDES, V. 1989. El Paleolítico medio cantábrico y sus relaciones con el área pirenaica. Espacio, Tiempo y Forma. Serie I Prehistoria y Arqueología, 2, p.49-59. CABRERA VALDES, V. Ed. 1993. El origen del hombre moderno en el suroeste de Europa. Madrid, U.N.E.D. CABRERA VALDES, V., BISCHOFF, J.L. 1989. Accelerator 14C dates for Early Upper Palaeolithic (Basal Aurignacian) at El Castillo cave (Spain). Journal Archaeol. Sc., 19, p. 577-584.

CABRERA VALDES, V., VALLADAS, H., BERNALDO DE QUIROS, F., HOYOS, M. 1996. La transition Paléolithique moyen-

JALUT, G.  , DELIBRIAS, G., DARGNAC, J., MARDONES, M.,

BONHOURS, M. 1982. A palaeoecological approach to the last 21000 years in the Pyrenees : the peatbog of Freychènede (alt. 1350m, Ariège, south France). Palaeogeogr., Palaeoclim., Palaeoecol., 40, p. 321359.

Paléolithique supérieur à El Castillo (Cantabrie)  : nouvelles datations par le carbonne-14. C.R. Acad. Sci., Paris 322 Série IIa, p. 1093-1098.

CABRERA VALDES, V., MAILLO, J.M., LLORET, M., BERNALDO DE QUIROS, F. 2001. La transition vers le Paléolithique

JALUT, G., ANDRIEU, V., DELIBRIAS, G., FONTUGNE, M., PAGES, P. 1988. Palaeoenvironment of the Valley of Ossau

supérieur dans la grotte du Castillo (Cantabrie, Espagne) : la couche 18. L’Anthropologie, 105, p. 505-532. CASTAÑÓN, J.C., FROCHOSO, M. 1996. Hugo Obermaier y el glaciarismo. En A. Moure ed.: El Hombre fósil 80 años después, p.153-175. Universidad de Cantabria. CASTAÑÓN, J.C., FROCHOSO, M. 1998. La Alta Montaña Cantábrica : condiciones térmicas y morfodinámica en los Picos de Europa. En E. Gómez Ortiz et al. (eds.) : Procesos bióticos actuales en medios fríos, p.113-132. Universitat de Barcelona.

(Western French Pyrenees) during the last 27000 years. Pollen et Spores, XXX, fasc. 3-4, p. 357-394.

JALUT, G., MONTSERRAT MARTI, J., FONTUGNE, M., DELIBRIAS, G., VILAPLANA, J.M., JULIA, R. 1992. Glacial to

interglacial vegetation changes in the northern and southern Pyrenees : Deglaciation, vegetation cover and chronology. Quat. Sci. Rev., vol. 11, p. 449-480. JIMÉNEZ SÁNCHEZ, M. and FARIAS ARQUER, P. 2002. New radiometricand geomorphologic evidences of a last glacial maximum older than 18 Ka. In SW European mountains : the exemple of Redes Natural Park (Cantabrian Mountains, NW Spain). Geodinámica Acta, 15, p. 93-101.

CENDRERO, A., DIAZ DE TERAN, J.R., FLOR, E., FRANCES, E. GONZALEZ LASTRA, J.R. 1986. Guía de la naturaleza de Cantabria. Santander, Ed. Librería Estudio.

DANSGAARD, W., JOHNSEN, S.J., CLAUSEN, H.B., DAHL-JENSEN, D., GUNDESTRUP, N.S., HAMMER, C.U., HVIDBERG, C.S., STEFFENSEN, J.P., SVEINBJÖRNSDOTTIR, A.E., JOUZEL, J., BOND, G. 1993. Evidence for general

JOHNSEN, S.J.; CLAUSEN, H.B. ; DANSGAARD, W. ; FUHRER, K.; GUNDESTRUP, N.; HAMMER, C.U.; IVERSEN, P.; JOUZEL, J.; STAUFFER, B. and STEFFENSEN, J.P. 1992. Irregular

glacial interstadials recorded in a new Greenland ice core. Nature, 359, p. 311-313. MAYOR, M.; DIAZ, T.E. 1977. La flora asturiana. Col. Popular Asturiana, 27. Oviedo, Ayalga ed. MONTSERRAT, J. 1992. Evolución glaciar y postglaciar del clima y la vegetación en la vertiente sur del Pirineo : estudio palinológico. Monogr. IPE, 6, Zaragoza, CSIC. MONIÑO, M.; CENDRERO, A.; DIAZ DE TERAN, J.R. 1988. Dinámica glaciar cuaternaria en la vertiente norte de

instability of past climate from 250 Kyr. Ice core record. Nature, 364, p. 218-220. DUCHAUFOUR, P. 1948. Recherches écologiques sur la chênaie atlantique française. Paris, Ann. Ec. Nat. Eaux et Forêts. DUPIAS, G. 1985. Végétation des Pyrénées. Carte de la végétation de la France 1 : 200000. Paris, Ed. CNRS. DUPONT, P. 1962. La flore atlantique européenne : introduction

284

Domestic fires and vegetation cover among Neanderthalians UZQUIANO, P. 1997. Antracología y métodos : implicaciones en la

Castro Valnera, Cantabria. II Congreso Geológico de España, vol. I, p. 398-402. Universidad de Granada. NOËL, M.; BOCQUET, A. 1987. Les hommes et le bois. Histoire et technologie du bois de la Préhistoire à nos jours. Paris, Hachette Litérature.

PIKE TAY, A.; CABRERA VALDES, V.; BERNALDO

DE

Economía prehistórica, Etnoarqueología y Paleoecología. Trabajos de Prehistoria, t. 54 fasc. 1, p. 145-154. UZQUIANO, P. in press. Firewood management and Vegetation in the Cantabrian region during 40-37 Kyr. BP. Charcoal analysis of Transitional Aurignacian (level 18) of El Castillo cave (Puente Viesgo, Cantabria, Northern Spain). Actas del Symposio En el Centenario de la cueva de El Castillo, el ocaso de los Neandertales. Santoña, 16-18 Septiembre, 2003, U.N.E.D. UZQUIANO, P., CABRERA VALDÉS, V. 1999. Paleoecología y gestión del combustible en la ocupación del Auriñaciense arcaico de la cueva de El Castillo (Puente Viesgo, Cantabria). Espacio, Tiempo y Forma. Serie I Prehistoria y Arqueología, 12, p. 129-147. Madrid, U.N.E.D.

QUIROS,

1999. Seasonal variation of the Middle-Upper Palaeolithic transition at el Castillo, cueva Morín and El Pendo (Cantabria, Spain). J. Hum. Evol., 36, p. 283317. POLLUNIN, O. 1976. Arboles y Arbustos de Europa. Barcelona, Ed. Omega. PONS, A., GUIOT, J., de BEAULIEU, J.L., REILLE, M. 1992. Recent contributions to the climatology of the Last GlacialInterglacial cycle based on french pollen sequences. Quaternary Science Reviews, vol. 11, p. 439-448. RIVERA, D., OBÓN DE CASTRO, C. 1991. La guía Incafo de las plantas útiles y venenosas de la Península Ibérica y Baleares (excluídas las medicinales). Madrid, Ed. Incafo. REILLE, M., de BEAULIEU, J.L. 1990. Pollen analysis of a long upper Pleistocene continental sequence in a Velay maar (Massif Central, France). Palaeogeography, Palaeoclimatology, Palaeoecology, 80, p. 35-48. F.

VERNET, J.L., OGEREAU, P., FIGUEIRAL, I., MACHADO, C., UZQUIANO, P. 2002. Guide d’identification des

charbons de bois préhistoriques et récents. Sud-Ouest de l’Europe : France, Péninsule Ibérique, Îles Canaries. Paris, C.N.R.S. éd. p. 395 VILAPLANA, J.M. 1983. Quaternary Glacial Geology of Alta Ribagorça Basin (Central-Southern Pyrenees). Acta Geologia Hispana, vol. 18 fasc. 3/4, p. 217-233. WOILLARD, G. 1978. Grande Pile peat bog : a continuous pollen record for the last 140 000 years. Quaternary Research, 9, p. 1-21. WOILLARD, G., MOOK, W.G. 1982. Carbon dates at Grande Pile : correlation of land-sea chronologies. Science, 215, p. 159-161.

SANCHEZ GOÑI, M.F., TURON, J.L., EYNAUD, F., GENDRAU, S. 2000. European climatic response to Millenial-scale change in the Atmosphere-Ocean System during the Last Glacial Period. Quaternary Research, 54, p. 394-403. SERRANO, E. 1995. Geomorfología glaciar del Alto Trueba (Burgos). In G. Meaza; J.C. García cordón; J. Arañes, eds. : Libro-guía de las excursiones de campo de las XI Jornadas de Campo de Geografía Física, p.91-102. Universidad de Cantabria, Santander. SERRANO, E., MARTINEZ DE PISON, E. 1994. Geomorfología y evolución glaciar del Pirineo aragonés oriental. In C. Martí Bono and J.M. García Ruiz eds. : El Glaciarismo surpirenaico : nuevas aportaciones, p. 33-64.. Logroño, Geoforma Ediciones. SCHWEINGRUBER, F. 1978. Mikroskopische holzanatomie. Zücher, A.G. Zug. SCHWEINGRUBER, F. 1990. Anatomie europäischer Hölzer. Zücher, WSL/FNP Haupt.

TUTIN, T.G., HEYWOOD, V.H., BURGES, N.A., VALENTINE, D.H., WALTERS, S.M., WEBB, D.A. 1964. Flora europaea vol I. Cambridge, University press.

UZQUIANO, P. 1992a. Recherches anthracologiques dans le

secteur Pyréneo-cantabrique (Pays Basque, Cantabria et Asturias) : environnements et relations homme-milieu au Pléistocène supérieur et début de l’Holocène. PhD. Dissertation Université de Montpellier II. UZQUIANO, P. 1992b. L’homme et le bois au Paléolithique en Region Cantabrique, Espagne. Les exemples d’Altamira et El Buxu. In J.L. Vernet ed. Les charbons de bois, les anciens écosystèmes et le rôle de l’homme. Actual. Bot., 2/3/4, p. 361-372. UZQUIANO, P. 1994. Ciencias del Cuaternario. La antracología. Investigación y Ciencia (Spanish ed. of Scientific American), 210, p. 32-34. UZQUIANO, P. 1995. L’évolution de la végétation à l’Holocène initial dans le nord de l’Espagne à partir de trois sites archéologiques. Quaternaire, vol. 6, fasc. 2,  p. 77-83.

285

Fuel Supplies for Pompeii Pre-Roman and Roman charcoals of the Casa delle Vestali R. Veal¹, G. Thompson²

Macquarie University, Sydney, 2109, Australia E-mail: [email protected] 2 University of Bradford, Bradford, BD71DP, UK 1

Abstract Charcoal reports from prehistoric times provide us with large scale, extended diachronic studies which predominantly address the question of environmental reconstruction. In contrast, this paper examines charcoal recovered from the excavations of one Pompeiian dwelling from its modest beginnings circa 350 BC to its final destruction in AD 79. The Casa delle Vestali, excavated 1995-2000 by the Anglo-American Project in Pompeii, forms part of ongoing excavations of an entire city block in Regio VI.i. Preliminary anthracological analysis of the property, combined with ancient historic and modern phytosociological resources demonstrate that even small quantities of data can shed light on cultural and economic phenomena in pre-Roman and Roman Pompeii. Early results indicate strong selection of high calorific fuels throughout the period. It is postulated that most of the taxa identified grew in lower level managed montane deciduous forests or higher altitude natural forests, thus demonstrating the likelihood of a developed market structure for wood supply to the city.

Introduction and significance of the study Pompeii is renowned for the excellent preservation of the destruction phase from August AD 79, and the opportunities this offers to view Roman life frozen in just one day. In contrast, there are sometimes opportunities to excavate below the AD 79 levels and investigate the growth of the city itself. This is the approach of the Anglo-American Project in Pompeii which has been excavating an entire city block (Insula VI.i), researching the city’s social and economic development from its origins to its destruction in AD 79. Scholarship about Pompeii’s economy has been dominated by discussion of its status as a consumer city tied to its agricultural hinterland (Finley 1973, p. 131,139). Emphasis has been placed on outputs (food, textiles, ceramics, etc.) and their destinations, for example: Moeller (1976) (Pompeiian wool production and processing); Peacock (1977) (Pompeiian red ware pottery); and Mayeske (1972) (Pompeiian bakeries). More recently Laurence (1994) has threaded these studies together to describe Pompeii’s economy. Input analysis has therefore examined (slave) labour and raw materials for manufacturing production and service industries. The nature and supply of energy to fuel the urban economy has tended to be understated. Evaluating the wood fuel supply to the city could therefore contribute usefully to the analysis of its economy. This paper examines the preliminary results from a diachronic analysis of the charcoals excavated from one Roman dwelling, the Casa delle Vestali (House of the Vestals). It differs substantially from pre- or proto-historic charcoal studies, which utilise large numbers of charcoal fragments and typically reconstruct the palaeovegetation over long time periods. The scale is smaller in every way: it examines contexts from one single dwelling, and over time covers only several hundred years. Charcoal is sparse within the excavated deposits and a total of 712 fragments have been analysed to date. Nevertheless, the results have provided some immediate insights into fuel provision for Pompeii and it forms the basis of an extended and ongoing examination of charcoals from the entire insula.

Pompeii–history and palaeogeography Pompeii, located near the sea on the mouth of the river Sarno, in the fertile Campanian plain, was a logical place for human settlement (Carafa 1997, pp. 25-26) (fig. 1 shows the plain and the ancient Campanian cities). Trace remains of the early Iron Age Italic peoples have been found and evidence for both Greek and Etruscan influence exists for the period 8th - 6th centuries BC culminating in the construction of the city walls (Wallace-Hadrill 1998, p. 21). In the 4th century, an Italic people identified as the ‘Samnites’ who originated from the Apennines, descended from the mountains, gradually dominating the cities of the plain, including Pompeii. The earliest traces of a dwelling below Vestali date from this period (Jones, Robinson 2004, p. 109). Rome began her ascendancy over the next couple of centuries and gradually took over the Italian peninsula (David 1996, pp. 145-147). While Pompeii fought against Rome, and was in 89 BC besieged by the Roman general Sulla, the city was not destroyed, and eventually became a Roman colony in 80 BC. From early in the 2nd century BC, Pompeii had experienced a surge in development (Zanker 1998, pp. 34-35). In AD 62 the city’s expansion came to a sudden halt with a large earthquake. The damage was so severe that rebuilding was still underway in AD 79 when Vesuvius erupted. The then navigable Sarno River (Laurence 1999, p. 119; Mastroroberto 1990a, p. 21), enabled efficient trade between Pompeii and the inland cities of Nola and Nuceria. The coast was considerably further inland than today (Mastroroberto 1990b, p. 10), and probably less than 1km from the city of Pompeii (Pescatore et al. 2001), with extensive salt pans and marshes behind the ancient dunes and beach ridges just outside Pompeii (Stefani 1990, p. 12). The Campanian plain is low-lying, with some parts below sea level, and thus areas around the Sarno were prone to flooding. Pompeii is set on the only landform of any height (at 30m

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

R. Veal, G. Thompson altitude) - on a small ancient lava plateau (Stefani 1990, p. 11). The city’s development from the 2nd century BC is reflected in the development of Vestali which shows clear evidence of the Sullan siege; large expansion in the subsequent period and adjustments following the AD 62 earthquake (Jones, Robinson 2004, pp. 111-122).

well explored by Meiggs (1982). They provide the broad context on woodland management and which woods were favoured for specific purposes, however precise details are lacking. Ecology The Mediterranean climate consists of hot dry summers and mild winters, although microclimate variation is considerable (Grove, Rackham 2001, pp. 24-25, 57). In Campania, cold climate trees which typically exist at lower altitudes in northern Europe, are restricted to the surrounding mountainous areas, i.e. Mt. Vesuvius, the central Apennine range and the Lattari Mountains. Examples of such taxa include: Abies alba (alpine fir, or silver fir), Picea abies (alpine pine), both growing above 1000 m, and Fagus sylvatica (beech) – from about 8001000 m (Pignatti 1982, p. 112). These three taxa are now reasonably rare in the mountain areas surrounding the plain due to anthropogenic pressures (Ciarallo 2002, p. 174). There is some information about the regional vegetation based on pollen studies from Lake Avernus (north of Naples) (Grüger et al. 2002, pp. 240-273); the Bay of Salerno (Ermolli, Di Pasquale 2002, pp. 211-219); and from Pompeii itself (Dimbleby, Grüger 2002, pp. 181216). The mid-montane altitudes – 400m - 800m, were thought to have been dominated by mixed deciduous forests of Quercus sp. (oak), Castanea sativa (chestnut), Alnus

Overview Analysis of the ecological and socio-economic environment from which a charcoal assemblage originates is a necessary prerequisite to charcoal analysis (Smart, Hoffman 1988, pp. 167-205).

Roman wood cultivation and exploitation In the Roman world, wood was virtually the only source of fuel, with coal and peat almost unknown in the Mediterranean, and liquid petroleum products poorly understood (Humphrey et al. 1998, pp. 41-44). The Romans required wood and charcoal to smelt metals, make glass and ceramics, fuel their public baths, cremate their dead – and build warships and machines (Meiggs 1982, p. 237, 259). In cities like Pompeii, wood and wood charcoal were employed for cooking and heating in bakeries, inns, homes and baths, and to fuel the fulleries and dye plants of the woollen industry. The historic sources have been

Fig. 1. Campanian Plain: 500 m Contours (adapted from Wallace-Hadrill (1998: 21).

288

Fuel Supplies for Pompeii sp. (alder), Fraxinus sp. (ash), Ulmus sp. (elm), Acer sp. (maple), and members of the Corylaceae family: Carpinus sp. (hornbeam), Ostrya carpinifolia (hop hornbeam) and Corylus avellana (hazel). Evergreen oak was also present. Below 400m, deciduous forests continued in part, with some commercial Olea europaea (olive) and Vitis vinifera (grapevine) (Borgongino 1993; Ciarallo 2002; Ciarallo, De Carolis 1998; S.A.P. 1999-2001). Riverine taxa such as Alnus (alder), and members of the Salicaceae: Salix sp. (willow) and Populus sp. (poplar) would have spread along the river flats, wherever agriculture did not occupy the fertile soil of Campania Felix1. On the lower hillslopes and the plain and woodland margins, more degraded vegetation may also have occurred i.e. macchia (Foss et al. 2002, p. 66), including Erica arborea (tree heather), Arbutus unedo (strawberry tree) and Pistacia sp. On the coast the vegetation probably included various types of Pinus sp. (pine) and Cupressus sempervirens (cypress).

including industrial hearths for iron working; large tanks, interpreted as possibly being linked to the making of garum (fish sauce); a range of smaller domestic structures, and in the earliest years of settlement, a considerable area of open space. Fig. 3 depicts the standing remains of the front of the Casa delle Vestali (on the via Consolare) as they appear today.

Depositional environment Vestali exhibits complex stratigraphy where trace architectural features are observed through time, but in the main, the deposits indicate repeated episodes of destruction of walls, opening or closing of doors, and the levelling of floors for rebuilding (Jones, Robinson 2004, p. 108). While kitchen/cooking activities are the primary ones that would introduce charcoal into the archaeological record in a domestic environment, other activities to be considered are: heating – of rooms and/or water for the late bathing suite(s); possibly building activities (the stratigraphic notes suggest evidence of rough kilns or temporary hearths – possibly part of the rebuilding process (Anderson 1999); and major fire events (although there is no evidence for this in Vestali).

Previous archaeological research In Italy, archaeological research to describe and quantify wood resource use is as yet in its infancy. In Pompeii, examination of charred plant remains has been limited to studies of food plants (e.g. Ciaraldi 2001; Ciaraldi, Richardson 2000; Meyer 1980) and ritual offerings (Robinson 2002); woods used in building, agricultural tools and machinery (Hatcher 2002); and woods for furniture-making (Mols 2002). The archaeological evidence complements the historic sources with regard to orchards and gardens (Jashemski 1979; Jashemski et al. 2002) but with such a range of materials and taxa to consider, at this point little attention has been focussed specifically on wood fuel. There has been one published anthracological study from Pompeii which reported the analysis of around 200 fragments from a single archaic horizon in one Pompeiian house (Castelletti, 1984). Fagus sylvatica (beech) was the primary taxon, and it was concluded that the samples probably represented burning of manufactured charcoal.

Ongoing research: Project in Pompeii

the

Sampling issues All excavated deposits are sieved through 5mm mesh, and 20 litres/context are sampled for flotation. Charcoal is collected by hand and also recovered by dry sieving. Only small quantities of tiny fragments are generally recovered through flotation and this material has not been examined in this study. A preliminary audit of the Vestali charcoal assemblage demonstrated a relative dearth of charcoal, since most contexts contained fewer than 80 fragments. This study focuses on 25 contexts chosen because they contained the greatest concentrations of charcoal. They are primarily interpreted as kitchen/service areas of the house and there is some confidence that at least some of the charcoal recovered here is refuse from repeated cooking episodes where charcoal not swept up by cleaning was gradually absorbed into the floor surface. Some deposits, however contained charcoal/ ash – and could reflect waste as a result of levelling and building processes, or even industrial activities. The selected archaeological areas afford some spatial analysis – but this is limited at this point. Fig. 4 shows a plan of the house in AD 79, including the archaeological areas (AAs) chosen for charcoal analysis. Contexts were chosen to sample the site through time and most material analysed was from mixed secondary fills. Many contain traces of food scraps (those in and around the nominated kitchen areas: AAs 84, 85, 86 and 101 - the second kitchen area; and to a lesser extent AA74 - the first kitchen). These traces include: fragments of mammal bones, fish scales and bones, sea urchin remains and shell fragments. All

Anglo-American

The Anglo-American Project in Pompeii began excavating in 1995. The team has been granted the unusual right to excavate a whole suburban block, or insula (Insula VI.i) down to its origins. fig. 2 shows Insula VI.i – in the northwest of the city, located close to the Herculaneum Gate. The insula is complex, showing at the AD 79 level, a number of different types of buildings with commercial and domestic activities including bars, a stable, an inn, and a large public well. It is dominated by two large houses, the Casa delle Vestali and the Casa del Chirurgo, which predates Vestali (Jones, Robinson 2004, p. 109). Activities below the AD 79 level are also complex – 289

R. Veal, G. Thompson

Fig. 2. Location and Details of Insula VI.i in Pompeii (AD 79). (Diagram : courtesy D. Robinson, AAPP).

Fig. 3. Casa delle Vestali (via Consolare) AD 79 Standing Remains (Photo : AAPP archive).

290

Fuel Supplies for Pompeii Charcoal identification

contain rubble, plaster and broken ceramics, sometimes with glass and metals in the form of coins, nails, or even waste slag. A few contexts contained lead shot from the Sullan siege (89 BC). AA130, interpreted as having an industrial function (early) and storage function (later), differs somewhat from the other areas showing a lack of archaeobotanical remains, except charcoal. There is no specific evidence that materials for the fills have been brought in from elsewhere although this cannot be completely discounted. The secondary nature of the deposits suggests the charcoal excavated is likely to represent a number of different burning events and thus be more likely to represent the ‘average’ taxa used in domestic wood consumption over the time period represented by the context.

Method Identification was carried out by light microscopic examination according to Leney and Casteel (1975, p.153159) using modern wood atlases and reference charcoal. Schweingrüber (1990) was the primary reference atlas, together with Gale and Cutler (2000), Hather (2000), Miles (1978), Schoch et al. (2004) and Schweingrüber (1978). The modern charcoal reference collection of the University of Bradford was augmented with woods collected from L’Orto Napoli (Naples Botanic Gardens); the Royal Botanic Gardens in Sydney, Mt Tomah and Melbourne; Pompeii’s ancient and modern cities; and mountain areas surrounding the Campanian plain, in particular the Lattari Mountains and the Apennines. Modern woods were charred according to Pearsall (2000, pp. 129-131), at a temperature of 350° C.

Since the number of fragments from most contexts was less than 80, all fragments were analysed, down to approximately 2 or 3mm, until size precluded fracturing to view the relevant sections. For four contexts with exceptionally large quantities of charcoal (more than 150 fragments), 50% of the available material was analysed. These were samples: 84/4, 84/29, 86/14 and 130/42. The decision to identify a sub-sample allowed the evaluation of more contexts and thus maximisation of the number of taxa retrieved (Asouti 2001, p. 96; Keepax 1988, p. 45). The 50% was selected by spreading the charcoal fragments (all size categories

Overview of the Charcoal Assemblages Fragments successfully identified to at least family level varied in size from 2mm to 30mm in the longest dimension (most often the longitudinal), but the average size was about 10mm. Fungal hyphae were observed in most contexts, in at least some fragments, demonstrating collection of dead wood, or wood stored before burning. Much of the charcoal also showed post-depositional crystallisation of inorganic mineral salts and (microscopic) root penetration. Small branches and stem wood were both observed.

Results A total of 712 fragments was examined and 15 different taxa were identified, five of these to species level and six to genus. A few fragments were only identifiable to family level, and 17.7% were indeterminate. For the indeterminate fragments, in many cases a probable identification has been noted and some may be resolved in the future. Ubiquity and fragment count were adopted as the primary quantification measures. While the validity of quantification by weight is acknowledged, in the case of this assemblage, the fragile state of the charcoal and its relatively high contamination by post-depositional crystallisation suggested fragment count would be the most time effective and suitable method of quantification.

Fig. 4. Casa delle Vestali – Archaeological Areas Used in Charcoal Analysis (adapted from archive of D. Robinson)

Adequacy of Sampling

mixed) on paper and visually dividing the assemblage in half, and then check weighing - a method similar to Thompson (1988).

The Cumulative Taxa Curve in fig. 5 indicates that the sampling strategy has been reasonable, and that it is likely most taxa have been detected in the assemblage.

For analytical purposes, the contexts were grouped into centuries based on stratigraphic data, and ceramic and coin spot dates spanning the 2nd century BC (nine contexts); 1st century BC (ten contexts); and the 1st century AD (six contexts). No contexts from the 3rd century BC have yet been examined.

Summary of Results: Ubiquity The ubiquity table, fig. 6, lists the taxa by their general ecological zones, from highest altitude to lowest, with contexts grouped by century. The most prominent feature of the charcoal from the Casa 291

R. Veal, G. Thompson Percentage Frequency Analysis

delle Vestali is the dominance of Fagus sylvatica (beech) throughout every century – and in every context. Of the Fagaceae, the fragments are either Quercus sp. or Fagus sylvatica. Fragments identified to the Corylaceae family will be one of the other identified species: Carpinus (probably betula or possibly orientalis) (hornbeam), Corylus avellana (hazel) or Ostrya carpinifolia (hop hornbeam). Prunus spp. include cherries, plums, almonds and peach but many of these were only just being introduced into Roman cultivation in this period – most Prunus fragments are probably P. avium (wild cherry). The material identified as Rosaceae is probably also Prunus.

The individual taxa found in the ubiquity table were grouped into their broad ecological and economic zones and graphed (fig. 7 – excludes indeterminates). The four groupings in order of significance (in terms of fragment count) are: Fagus sylvatica; the Corylaceae (Carpinus spp., Ostrya sp. and Corylus avellana); Quercus spp. and other woods (Maloideae, Prunus spp., Vitis vinifera, Pinus spp. and Acer sp.). The pie charts clearly depict the overall dominance of beech (Fagus sylvatica) and its reduction over time to be replaced by wood from the Corylaceae and small quantities of oak (Quercus spp.). Modern phytosociological groupings (Pignatti 1997, pp. 342-392) show Carpinus generally grows slightly lower down the mountain than Fagus, and higher than Corylus or Ostrya – the latter two appearing across a range of low to mid altitudes like Quercus. The Corylaceae may all be coppiced and make good charcoal. By the 1st century AD the Corylaceae have slightly diminished in importance but Other Woods (mostly fruits and vines) increase further. Cuttings from orchards or vineyards are relatively rare in the 2nd and 1st centuries BC (less than 4%) – increasing to 20% by the 1st century AD.

The Maloideae include fruits such as Pyrus sp. (pear) or Malus sp. (apple), but also some non-commercial forest margin taxa such as Crataegus sp. (hawthorn). At least some of the Pinus spp. (pine) are possibly P. nigra (a montane pine). There were a few fragments of Vitis vinifera (grapevine). A second feature of the assemblage is that taxonomic diversity rises over time, but the average number of taxa per context remains fairly low overall (an average of about five taxa). Diversity in contexts is predominantly a reflection of environmental availability of woods - and cultural selection. Since Pompeiians had at hand many types of wood on the fertile plain and in the mountains, the low diversity observed shows high selection behaviour. There are six different taxa overall in the 2nd century BC, rising to nine in the 1st century BC, and ten in the 1st century AD (Material identified to family level is excluded from these counts where representatives from each of the species or genus are also present.).

Missing from the assemblage, perhaps surprisingly, are the riverine taxa Salix sp., Ulmus sp. or Alnus sp. (willow, elm and alder), which probably all grew close to the city (and are noted by the ancient historians as being used for wood fuel). There is also very little Acer sp. (maple - also considered good as both fuel and charcoal). There is no Olea sp. (olive).

Fig. 5. Cumulative Taxa Curve.

292

Fuel Supplies for Pompeii

Fig. 6. Ubiquity Table Showing the Charcoals from the Casa delle Vestali (25 contexts).

Other Analyses

Discussion

More detailed analyses of the data are continuing including a breakdown of indeterminate patterns, tree ring analysis to examine cropping regimes, and tests to examine preservational bias. An examination of spatial differences is also underway to determine if particular charcoal residues are associated with specific activity areas within the house.

The overall pattern of charcoal correlates with the increasing penetration of Roman influence into the area over time. It is probable that this is broad evidence of ‘Romanisation’ of agricultural practice as found in the historic sources. The Roman writers on agronomy, Cato, Columella and Varro, present a picture of mixed farming – moving to larger and more specialised (and more profitable) vine and olive cultivation between the 2nd century BC and the 1st century AD. Archaeological evidence tends to support this trend, e.g. Lomas (1993, pp. 117-121) discusses a trend in landholding from dominance by smaller farms in the early 2nd century BC to an increasing number of larger and more specialised estates by the 1st century AD, but with significant numbers of smaller mixed farmsteads surviving.

Bias Bias from sampling constraints is considered to be minimal in terms of the overall results however insufficient contexts have been processed for the 1st century AD at this point. It is possible that the charcoals from the main taxa are relatively resistant to breakdown, especially beech. It is difficult however, to estimate the degree of over-representation without further experiment. Despite this potential bias, significant cultural and economic implications are inferred from the sheer volume of beech (with its preferred growth at 800m+), as the nearest geographic location of this altitude lies at least 15 km away from Pompeii.

Beech, which grew at an altitude just below the alpine conifers (and sometimes mixed with them) was and is, the highest growing hardwood in Italy (Pignatti 1997, p. 225). Beech is generally a slow-growing tree but can sometimes be coppiced on a short rotation and has more commonly been pollarded at longer intervals (Grove, Rackham 2001, pp. 48-52). The use of mature beech wood, and/ 293

R. Veal, G. Thompson

Fig. 7. Percentage Frequencies: 2nd century BC, 1st century BC and 1st century AD (based on numbers of fragments for each taxon).

or long period pollarding appears (from early observations of the ring patterns in the Vestali charcoal) to have been limited to the earlier period in the 2nd century BC. Beech was cropped in the 1st centuries BC and AD also but at this point it is difficult to tell if it was formally coppiced (perhaps for charcoal making) or if the fuel wood was collected from fallen branches. Possibly both occurred.

winter and the prunings burnt then or shortly after. It is also possible that as kindling these cuttings could have predominantly burnt to completion. We do not see any Olea europaea (olive) at all. If used as kindling, olive would have been the most likely to burn completely due to its oily character. The remaining pine fragments, other than those thought to be Pinus nigra, could have come from virtually any area on the mountain or plain since various pine species inhabit most niches.

It is interesting that the overall use of Quercus (oak), either deciduous or evergreen, is not all that significant in the assemblage from Vestali, nor does it vary much over the period. Oak had many functions. Besides being one of the woods with a high calorific value, and making excellent charcoal, it was also used for furniture and building (Gale, Cutler 2000, pp. 204-205); its leaves were highly palatable to all herd animals, and its acorns were used to fatten pigs in particular - and the acorn flour was used in cooking (Grove, Rackham 2001, p. 195). It was common in the environment according to pollen studies. Similarly, Corylus (hazel) had many uses: nuts were food for animals and humans, stakes made agricultural supports, and wood was used as a fuel and for making charcoal (Grove, Rackham 2001, pp. 88-89). Fruits and vines grew in the foothills and on the plain (including just outside and inside the city walls) (Ciarallo 2002; Ciarallo, De Carolis 1998). The trees and vines would have been pruned in

The Dominance of Beech Beech requires high altitude for optimum growth in southern Italy. Consequently, its dominance in the charcoal assemblage at Pompeii provokes several questions. The city is located on the Campanian plain at an altitude today of only 30 m - some distance from any mountains. Where did the beech grow? How was it cropped and transported? How was it traded? Future research will address these questions in more detail, but for now only a brief discussion follows. Pollen evidence suggests beech was plentiful in the classical period and may have even been increasing (Grove, Rackham 2001, pp. 172-173). If this were the case, it follows that growth conditions were probably good and that beech-dominated forests may have 294

Fuel Supplies for Pompeii with acknowledged caveats). The Apennines were the origin of the Samnite occupants of Pompeii in the 2nd century BC. The Sarno River, which originates at the foot of the Apennines was a major trade route linking Pompeii to the inland cities, thus transport down the Sarno of either raw timber, and/or charcoal is quite likely. The observed increase in the taxonomic diversity of the charcoal over time, along with the nature of that increase reflecting both the availability and use of more highly managed wood taxa and orchard and vine cuttings, together demonstrate increased sophistication of consumption, and imply possibly greater efficiencies of production. These are markers of a more advanced economy. More work is required however, to confirm these trends in order to provide a substantial contribution to the debate on the Pompeiian economy.

been extensive at high altitudes. Modern ecologists have researched the optimum conditions for growth of beech trees in Italy, demonstrating that, in the current climate, photosynthesis is optimised between 1200 and 1500 m (Pignatti 1997, pp. 487-490). The area supporting beech growth is further restricted by considering its preference for calcareous to weakly acidic soils (Pignatti 1997, p. 213). Three mountain areas can be considered: Mt Vesuvius; the Apennines, and the Lattari Mountains. Vesuvius is volcanic, but before the AD 79 eruption, the soils on Vesuvius were alkaline, with a pH of about 8.2 (Dimbleby, Grüger 2002, p. 189) and stable for about 500 years since the previous volcanic event (Foss et al., 2002, p. 72). Beech may have been present on Vesuvius, but this may not have been the ideal environment. Both the Lattaris to the south and the Apennines to the northeast of Pompeii, are calcareous, and both could have supplied beech to the city. Wood, or charcoal, would have been transported by ox or mule and cart down the mountain to the main road – or to the river Sarno. Laurence (1999, p. 134-135) describes collegia of mules and muleteers which were used for short distance land transport. From there it would have been transported to the city, either along the riverside road, or on the river itself. River transport, at perhaps 10 km/hour, and maybe a greater carrying capacity, was thought to have been cheaper, but the difference is difficult to estimate (Laurence 1999, p. 120). The ecological requirements of beech preclude the growth of large amounts of this wood on the plain, and thus farms nearby could not have supplied the beech in the amounts implied by the assemblage. They may however, have supplied kindling and faggots of other types of fuel wood, including orchard and vine cuttings – and even oak and hazel, since these woods were valuable for many purposes, eminently managed by coppicing, and probably grew on the plain to some extent. Most of the plain however, would likely have been given over to cereals (Lomas 1993, p. 118-119).

Future Work This study has provided only a brief look at both the historic sources and complementary archaeological evidence. Future work will include a more thorough review of these sources together with the analysis of the remaining charcoals from the insula, and possibly comparative material from elsewhere in the city. More contexts from the 1st century AD need to be analysed. No contexts from the 3 rd century BC have yet been examined but the aim is to extend the diachronic picture. A cknowledgements

This study could not have been prepared without the assistance of the staff and students of the Anglo-American Project in Pompeii, in particular Directors Rick Jones (University of Bradford, UK) and Damian Robinson (Oxford University, UK). Special thanks also to staff of the Soprintendenza di Pompei, especially, Dr. Piero Giovanni Guzzo, (Soprintendente), and Dr. Annamaria Ciarallo whose gracious permission to use laboratory facilities at Pompeii, to gather reference material and to provide access to forthcoming publications and library facilities were much appreciated. Dr. Gaetano di Pasquale (l’Università di Napoli) assisted with early discussions on local ecology.

Conclusion: Fuel Supplies to the City The pattern of wood use in the Casa delle Vestali changes over the centuries with beech declining in importance, particularly from the 1st century BC onwards. Although beech still dominates the later deposits, lower montane managed taxa such as hazel and oak were used and evidence of cuttings from orchards and vineyards increases. This correlates well with ancient sources on Roman agricultural and wood management practices and other Pompeiian archaeological studies which reflect the increasingly sophisticated town that Pompeii had developed into by the 1st century AD. Since the study covers a period of nearly 300 years, a number of generations of owners and several construction phases of Vestali, the results are considered a possible indicator of the general wood supply to Pompeii (albeit

Endnotes

1 Campania Felix was the name Roman writers gave to the Campanian plain, its volcanic soil providing one of the most fertile areas for agriculture in the republic and empire.

References A nderson, M., 1999. AAPP 1999 Supervisor’s Report: AA84, 85, 86, 101. Department of Archaeological Sciences, University of Bradford, Bradford. Asouti, E., 2001. Charcoal Analysis from Çatalhöyük and

295

R. Veal, G. Thompson Pinarbasi, two Neolithic sites in the Konya Plain, South-Central Anatolia, Turkey, PhD Thesis, Institute of Archaeology, University College London, London. Borgongino, M., 1993. La Flora Vesuviana del 79 d.C. In: M. Borgongino, A. Ciarallo, C. Mazza, S. Striano, eds. Parchi e Giardini Storici, Parchi Letterari. Atti del III Convegno (primo internazionale), “Paesaggi e Giardini del Mediterraneo.” Salerno: GRG ipolitografica srl, 115-140. Carafa, P., 1997. What was Pompeii before 200 BC? In S.E. Bon, R. Jones, eds. Sequence and Space in Pompeii. Oxford: Oxbow, 13-31. Castelletti, L., 1984. Annalisi dei Legni. In: M. Bonghi Jovino, ed. Ricerche a Pompei l’Insula 5 della Regio VI dalle Origini al 79 d.C. (campagne di scavo 1976-1979). Roma: L’Erma di Bretschneider, Ciaraldi, M., 2001. Food and Fodder, Religion and Medicine at Pompeii. The use of plant resources at Pompeii and in the Pompeian area from 6th century BC to AD 79. PhD Thesis, Department of Archaeological Sciences, University of Bradford, Bradford. Ciaraldi, M., Richardson, J., 2000. Food, Ritual and Rubbish in the Making of Pompeii. In: G. Fincham, G. Harrison, R.R. Holland, L. Revel, eds. TRAC99. Proceedings of the Ninth Annual Theoretical Roman Archaeology Conference Durham April 1999. Oxford: Oxbow Books, 74-82. Ciarallo, A., 2002. Colture e Habitat del Territorio Vesuviano nel 79 d.C., Rivista di Studi Pompeiani, (XII-XIII) 167176. Ciarallo, A., De Carolis, E. (eds.), 1998, Around the Walls of Pompeii. The Ancient City in its Natural Environment, Milano: Electa, for the Soprintendenza Archeologica di Pompei,. David, J.M., 1996. Ch 7: The Social War. In: The Roman Conquest of Italy. Oxford: Blackwell Publishers Ltd, 140-149. Dimbleby, G.W., Grüger, E., 2002. Pollen Analysis of Soil Samples from the A.D. 79 Level. Pompeii, Oplontis, and Boscoreale. In: W.F. Jashemski, F.G. Meyer, eds. The Natural History of Pompeii, Cambridge: Cambridge University Press, 181-216. Ermolli, E.R., di Pasquale, G., 2002. Vegetation Dynamics of South-western Italy in the Last 28kyr Inferred from Pollen Analysis of a Tyrrhenian Sea Core, Vegetation History and Archaeobotany, (11), 211-219. Finley, M.I., 1973. Ancient Economy. London: University of California Press Ltd. Foss, J.E., Timpson, M.E., Ammons, J.T., Lee, S.Y. 2002. Paleosols of the Pompeii Area. In: W.F. Jashemski, F.G. Meyer, eds. The Natural History of Pompeii, Cambridge: Cambridge University Press, 65-79. Gale, R., Cutler, D, 2000. Plants in Archaeology. Identification manual of vegetative plant materials used in Europe and the southern Mediterranean to c. 1500. Kew, West Yorkshire: Westbury Publishing and Royal Botanic Gardens, Grove, A.T., Rackham, O., 2001. The Nature of Mediterranean Europe. An Ecological History, New Haven London:Yale University Press. Grüger, E., Thulin, B., Müller, J., Schneider, J., Alefs, J., Welter-Schultes, F.W., 2002. Environmental Changes in and Around Lake Avernus in Greek and Roman Times: A Study of the Plant and Animal Remains Preserved in the

296

Lake’s Sediments. In: W.F. Jashemski, F.G. Meyer, eds. The Natural History of Pompeii, Cambridge: Cambridge University Press, 240-273. Hatcher, P.G., 2002. Wood Associated with the A.D. 79 Eruption. In: W.F. Jashemski, F.G. Meyer, eds. The Natural History of Pompeii, Cambridge: Cambridge University Press, 217-224. Hather, J.G., 2000. The Identification of the Northern European Woods. A Guide for Archaeologists and Conservators. London: Archetype Publications Ltd. Humphrey, J.W., Oleson, J.P., Sherwood, A.N.E., 1998. Greek and Roman Technology: A Sourcebook. Annotated Translations of Greek and Latin Texts and Documents. London: Routledge. Jashemski, W.F., 1979. The Gardens of Pompeii, Herculaneum and the Villas Destroyed by Vesuvius. New Rochelle, N.Y.:Caratzas. Jashemski, W.F., Meyer, F.G., Ricciardi, M. 2002, Catalogue of Plants. In: W.F. Jashemski, F.G. Meyer, eds. The Natural History of Pompeii, Cambridge: Cambridge University Press, 84-180. Jones, R., Robinson, D., 2004. The Making of an Elite House: the House of the Vestals at Pompeii. Journal of Roman Archaeology, (17). Keepax, C.A., 1988. Charcoal Analysis with Particular Reference to Archaeological Sites in Britain. PhD Thesis, University of London, London. Laurence, R., 1994. Production and Consumption. In: R.Laurence, ed. Roman Pompeii Space and Society. London: Routledge, 51-69. Laurence, R., 1999. The Roads of Roman Italy. Mobility and cultural change. London - New York: Routledge. Leney, L., Casteel, R.W., 1975. Simplified Procedure for Examining Charcoal Specimens for Identification. Journal of Archaeological Science, (2), 153-159. Lomas, K., 1993. Rome and the Western Greeks 350 BC - AD 200. Conquest and acculturation in South Italy. London - New York: Routledge. Mastroroberto, M. 1990a, Il Sarno. In Uomo e Ambiente nel Territorio Vesuviano, (Ed: Stefani, G.), Marius Edizioni, Pompei, pp. 21. Mastroroberto, M., 1990b. Il Mare e la Costa. In: G. Stefani, ed. Uomo e Ambiente nel Territorio Vesuviano. Pompei: Marius Edizioni, 10-19. Mayeske, B.J.B., 1972. Bakeries, Bakers and Bread at Pompeii: a Study in Social and Economic History, PhD Thesis, University of Maryland, Maryland. Meiggs, R., 1982. Trees and Timber in the Ancient Mediterranean World. Oxford: Oxford University Press. Meyer, F.G., 1980. Carbonized Food Plants of Pompeii, Herculaneum and the Villa at Torre Annunziata, Economic Botany, (34) 4, 401-437. Miles, A., 1978. Photomicrographs of World Woods. Department of the Environment Building Research Establishment, Her Majesty’s Stationery Office, London. Moeller, W.O., 1976. The Wool Trade of Ancient Pompeii. Leiden: Brill. Mols, S.T.A.M., 2002. Identification of the Woods Used in the Furniture at Herculaneum. In: W.F. Jashemski, F.G. Meyer, eds. The Natural History of Pompeii, Cambridge: Cambridge University Press, 84-180. Peacock, D.P.S., 1977. Pompeiian Red Ware. In: D.P. S. Peacock, ed. Pottery and Early Commerce: Characterisation

Fuel Supplies for Pompeii and Trade in Roman and Later Ceramics. London: Longman. Pearsall, D.M., 2000. Paleoethnobotany. A Handbook of Procedures. San Diego: Academic Press. Pescatore, T., Senatore, M.R., Capretto, G. and Lerro, G., 2001. Holocene coastal environments near Pompeii before the A.D. 79 eruption of Mount Vesuvius, Italy., Quaternary Research, 55, 77-85. Pignatti, S., 1982. Flora d’Italia. Bologna: Edagricole. Pignatti, S., 1997. I Boschi d’Italia. Sinecologia e Biodiversità. Torino: Unione Tipografico-Editrice Torinese. Robinson, M., 2002. Domestic Burnt Offerings and Sacrifices at Roman and pre-Roman Pompeii, Italy, Vegetation History and Archaeobotany, 11, 93-99. S.A.P., 1999-2001. Il Territorio - L’ager pompeianus, Produced by: Soprintendenza Archeologica di Pompei. http:// www2.pompeiisites.org/database/pompei/Pompei2.nsf, Accessed: June 17, 2004. Schoch, W., Heller, I., Schweingrüber, F.H., Kienast, F., 2004. Wood Anatomy of Central European Species. Produced by: Swiss Federal Research Institute WSL, http://www. woodanatomy.ch/, Accessed: June, 2004. Schweingrüber, F.H., 1978. Mikroskopische Holzanatomie (Microscopic Wood Anatomy) Zurcher AG, Zug. Schweingrüber, F.H., 1990. Anatomie Europaischer Holzer, Verlag Paul Haupt, Bern. Smart, T.L., Hoffman, E.S., 1988. Environmental Interpretation of Archaeological Charcoal. In: C.A. Hastorf, V.S. Popper, eds. Current Paleoethnobotany. Analytical Methods and Cultural Interpretations of Archaeological Plant Remains. Chicago: University of Chicago Press, 167-205. Stefani, G., (ed.) 1990. Uomo e Ambiente nel Territorio Vesuviano. Pompei: Marius Edizioni. Thompson, G.B., 1988. Wood Charcoals and Intra-site Spatial Analysis. In: P.J., Fasham, R.S. Kelly, M.A. Mason, R.B. White, eds. The Graeanog Ridge: the Evolution of a Farming Landscape and its Settlement in North-west Wales. Cardiff: Cadw, Welsh Historic Monuments, 7277. Wallace-Hadrill, A., 1998. Storia della Città. In: J. Berry, ed. Sotto I Lapilli. Studi nella Regio I di Pompei. Milano: Electa, 21. Zanker, P., 1998. Pompeii: Public and Private Life. Cambridge, Massachussetts: Harvard University Press.

297

Anthracology and Mediterranean landscape, classical and new approaches J.L. Vernet

UMR 5059, University of Montpellier II, France, 163 rue A. Broussonnet 34000 Montpellier E-mail: [email protected] Abstract The Mediterranean landscape developed during different stages, especially since Neolithic. Progress of anthracology allows today to propose a complex relationship man-environment model under climatic change. Bases of Mediterranean anthracology are summarized then is discussed the transition from primary forest to secondary formations in balance with man influence. An example significant is proposed.

A short historic as introduction At the beginning of the XXth century, charcoal were soon noticed in France, for instance by Abate Breuil, in prehistoric settlements for the identification of plants used by man in its environment. Charcoal are quite visible in the hearths, they are taken carefully and results give the evidence of a selection by prehistoric man because they recorded only some species; the results appear in appendix of the monographs. During this time, palynology becomes the best tool (vs. charcoal) to reconstitute the environment of prehistoric man. In the years 1950, the methods of excavations improve under the impulse of André LeroiGourhan, and then radiocarbon invention gives to chronology a calendar sense. Because need of radiocarbon, all the charcoal are recorded in the hearths. But during dating, charcoal is destroyed becoming gas. Fortunately, Saint-Laurent (1934), Le Du, Saccardy (1948), Balout (1952), Momot (1955), Santa (1961), Couvert (1976), in the Mediterranean area, insisted on charcoal potentialities to reconstitute prehistoric environment in countries where sediments are not able to record pollen. Now, part of charcoal is thus reserved to identification. That time sees rise of charcoal analyses1 with, in Mediterranean area, M. Escalon de Fonton, H. de Lumley, J. Guilaine, J. Courtin, M. Lorblanchet, J.-L. Roudil, F. Bazile, L. Meignen, D. Sacchi, B. Marti Oliver, J. Bernabeu, J. Molina, J. Vaquer, M. Py, F. Laubenheimer, J. Maury, P. Bintz, M. Paccard, L. Wengler, G. Fiorentino, but it is impossible to quote them all. In 1967, under the impulse of the Hungarian Josef Stieber, charcoal observation gain in speed and reliability by the method of manual sections. This method who does not require preliminary preparation, can moreover apply to very small charcoal of millimetric dimension2. An approach more exhaustive of prehistoric settlements show that charcoal are present, not only in the hearths, but also in occupation levels or soils habitat. Sifting systematic of occupation levels allows consequently quantitative approach of anthracology. Occupation levels if they have functioned over a long period (several years) give a more complete picture of the environment than hearths. These structures of combustion represent only last fires before site abandonment, with collecting specific by prehistoric man. The question of the choice seem thus solved: only levels with dispersed charcoal recording many “sampling”

for fuel wood are now considered. Reference charcoal sequences can then be studied on all the French Mediterranean area then in Iberian peninsula, we quote Font Juvénal, l’Abeurador, la Balma Margineda, la Salpêtrière, Unang, Fontbrégoua, Châteauneuf-les-Martigues, Cova de Cendres, Cova de l’Or (Bazile-Robert 1981; KraussMarguet 1981; Heinz 1988; Thiébault 1999; Chabal 1991; Badal Garcia 1990; Vernet 1971; Badal et al. 1994, see also review in Vernet 1997). At this beginning of XXIth century quantitative wood anatomy applied to charcoal progress. This method gives possible recording types of growing and climatic parameters, allowing thus a better understand in processes of domestication (example of olive-tree, Terral 1999). In addition, a new promising approach is now the anthracology “out of site”, or pedoanthracology in relation with prehistoric and historic sites (Delhon 2005).

General data on charcoal methodology From field to laboratory On archaeological excavations, charcoal sampling is done by square and by artificial stratigraphic unit. We have to sample squares in order to have the best record of dispersed charcoal. In soil or sediments, charcoal sampling are too made on the basis of artificial stratigraphic unit horizontally and in-depth or on section upwards. The volume of sediment or soil must be known. Sediments are filtered with water at a sufficient mesh of 2 mm who allows recovering all determinable charcoal. After drying and sorting, charcoal are identified according to the method of manual sections. Charcoal study Charcoal is identified as any wood carbonized fragment whatever its size. For more precision we distinguish three categories of distinct taphonomy significance (fig.1): -Charcoal, recorded in archaeological sites with usually centimetric sizes, brought by man are named archaeocharcoal (archaeoanthracology),

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

J.L. Vernet - Charcoal, size less than 1 cm to 2 mm (limit of identification), brought by surface waters with mineral particles are pedo-charcoal (pedoanthracology), -Charcoal, less than 1 mm, brought usually by wind, are micro-charcoal (microanthracology). These categories with different modes of deposits cannot have the same significance. Thus, archaeo-charcoal are present in special primary deposits as archaeological sites; pedo-charcoal are recorded outside of sites in various deposit conditions not always primaries; micro-charcoal are associated with pollen dispersed by wind and they are not identifiable at any systematic level. They are using as fire signal by pollen analyst. Charcoal as wood remain must be studied on the basis of analytic characters. The level of identification depends much of the wood systematic position. Some woods are easily identified, thus boxwood. But, generally, specific identification is difficult. It allows, however, distinguishing between Mediterranean pines and mountain pines. On the other hand, it allows with difficulty the distinction between oaks species. A general discussed panorama of the prehistoric anthracologic flora is however possible and realized recently (Vernet et al. 2001).

centred deposit area well-known by palaeontologists. The archaeological site strictly respect these two principles: the concentration is here depending of man activity. For pedoanthracology, interpretation is quite different if charcoal are present in diffuse soils (Thinon, 1992) or in small concentred basin, for instance in karstic deposits as we will see below (chap. 3). These principles are those of fuel-wood in archaeo-anthacology or fire-wood in pedo- anthracology. In archaeoanthracology, therefore, wood may be selected on the basis of its technological quality in special uses as beams, roofs, posts (structural timber). For fuel-wood (Chabal 1991) prehistoric men collected especially wood biomass rather than species. This collect were dependent of the vegetation structuring. Thus, for example, we know that in a temperate forest 20% of species give 80% of the biomass. This ratio can be calculated for charcoal and allows knowing the state of transformation of the vegetation. For example, in a vegetation transform by selection of some forest species, the index of Pareto is ca 15/85 even 10/90.

The present Mediterranean landscape in the south of France and its Neolithic origin

Principles of Anthracology

The Mediterranean primary forest As other palaeoenvironmental science, anthracology folIn south of France, primary forest include sclero-therlows a general rule; reconstitute the ancient production area III International Symposium on Anthracology, 2004mediterranean trees and shrubs, with various fremophil through the deposit area only accessible with the observer Lecce, quencies (fig. 3), between Mesolithic to Early Neolithic (fig. 2). A principle must be respected, the concept of con-

cm

Archaeological Charcoal Man

mm

0.5 mm) in forest soils provides evidence for local fire (Ohlson, Tryterud 2000). There is a good correlation between charcoal and vegetation, when we studied modern fires, in absence of any transport (Scott et al. 2000). There is a relationship between the occurrence of local fires and peaks in macroscopic charcoal (Gardner, Whitlock 2001). The karstic depressions are closed deposits; charcoal transport is always very reduced generally at an hectometric scale and often less than 1 km (Quilès et al. 2002).

Generalities Since the years 1970, archaeological charcoal give numerous and important contribution to the knowledge of the prehistoric environment, in Mediterranean France. But, the research always limited to archaeological sites was not well adapted to a spatial approach, necessary to understand “in situ” mechanisms of anthropisation and climatic change. The spatial researches, in pedoanthracology started in mediterranean area under the impulse of Thinon (1992) and especially out of archaeological sites11 with work of Berger, Thiébault and Delhon (Berger, Thiébault 2002; Delhon 2005). In the submediterranean mountain of the south of central massive in France, the discovery of charcoal in karstic depressions (Quilès et al. 2002 ; Vernet et al. 2005), renews the debate concerning the history of the vegetation with researches “in situ” in the ancient ecosystems. Charcoal is generally present in soils or superficial formations, but it is necessary to have a good control of deposit conditions (Gouveia, Pessenda 2000; Thinon 1992; Ver-

Taphonomy of charcoal Charcoal is present in natural deposits because of water transport (Thinon 1992; Carcaillet, Talon 1996), colluvial or sandy transport (Nichols et al. 2000). Water concentrates charcoal whereas wind only transports small carbonised elements, flowers, buds etc (Scott et al. 2000), or small coals which would not incorporate themselves on the ground (Blackford 2000). Charcoal burying is the consequence of biological activity: worms ingest particles more than 400 µm to 2mm and the deposit of charcoal would not have a regular stratification (Carcaillet, Talon 1996). On 303

J.L. Vernet the present measurements ca –26‰ (Vernet 2006); this imply a certain climatic dryness (Vernet et al. 1996) in particular from the XIIIth to XIX th centuries. These results can be assigned to the Little ice age, and compared with those obtained on archaeological sediments from the VIIth to XIIth centuries (Vernet et al. 1996), with the δ 18O data from stalagmites near Clamouse cave (Mac Dermott et al. 1999), with results from sediments and charcoal from the low Rhône valley, which record a dry spell during the IXth century (Berger, Thiébault 2002).

the contrary, in the tropical area, charcoals are stratified as a result of biological activity (Vernet et al. 1994; Boulet et al. 1995; Tardy 1999; Gouveia, Pessenda 2000). In the submediterranean mountain, a good stratification was also recorded (Quilès et al. 2002)

The sections from St Guilhem-le-Désert (Hérault, France) Generalities

Conclusion

Saint-Guilhem-le-Désert, located near the river Hérault, at 35 km the North-West of Montpellier, is known for two main reasons: its abbey founded in IXth century (804) and its remarkable forest of black pines (Pinus nigra Arnold ssp. salzmanni (Dunal) Franco), with the northern limit of its area, extending on 1000 at 2000 ha between 300 an 600 m altitude on calcareous soils constituted primarily of Bathonien dolomites4.

Anthracology, of which the goal initial was the reconstitution of the environment of prehistoric men, also allowed contributing with a « dynamic » knowledge of the Mediterranean forest during the last 10 millennia. Three main periods will be considered: - The first is the ancient ecosystems time without true forest but pre-steppic forest, during the end of the last glacial until towards 9000 BP, under semi-arid and fresh climate, - The second is the meso-sclerophil primary forest period, between 9000 and 5000 BP. Stability prevails (biostasy). Climate is Mediterranean sub-humid little contrasted, - The third is the present period, since 5000 years, with a Mediterranean sub humid climate very contrasted with period of dryness well recorded. It is the time of the regressive dynamic, leading to strong ecological heterogeneities6, to the matorralisation and the fragmentation of the ecosystem. It is thus that creates the landscape: a memory complex and obstinate of human influence.

Fires and forest changes Three sites have been studied by pits reaching the bed rock. They are located in closed dolomitic depressions and belong to supramediterranean stage (Vernet et al. 2005; Vernet 2006). Fourteen samplings of 12 l were taken, each 20 cm, from bottom to top on the basis of artificial strata. Ten calibrated C14 AMS were obtained5 on the three sampling sites (fig 4). We have extracted charcoal by sieving under moderate water jet with a two millimetres mesh (Vernet accepted). The main results show a regular presence of Pinus nigra ssp. salzmanni with Juniperus, from the middle early Holocene at 7000 BP, testifying to a permanent pre-forest. Other elements Quercus cf. pubescens, Q. cf. ilex, Acer, Ilex, Rosa, Populus, cf. Rhamnus, Ruscus and a Fern record certain diversity. Sediments deposited after the foundation of the abbey (804 AD) record an increase of Buxus sempervirens, evidence of the opening of the mature forest and its exploitation for sheep - goats, litter and fodder (Durand Tullou 1972), in agreement with historical data (Durand, Ruas in press). The presence of scales of pine cones, especially in the historical levels, could result from different modes of fires, more violent, making the cones burst, while more current fires only affect the under storey vegetation. The abbey of St Guilhem thus has benefited in a way intensive of a forest settlement relictual. The current state of the forest must be understood like a dynamic balance. This conclusion must be taken into account for protection.

Endnotes 1 From an epistemological point of view, charcoal analysis precedes anthracology who will become recently science of charcoal. 2 It will be necessary to wait the end of XXth century with AMS radiocarbon dating to be able to exploit completely theses millimetric charcoal. 3 Minimum transport, concentration maximum. 4 (43°43 ‘ N, 3°31 ‘ E). 5 C14 (AMS) dating were made by Dr A. Scharf, Universität Erlangen-Nürnberg with a financial contribution of the Office National des Forêts. 6 For instance open landscape beside forests, forest structures caused by fire etc.

References

Climate

Badal Garcia, E., 1990. Aportaciones de la antracologia al estudio del paisage vegetal y su evolucion en el Cuaternario reciente en la costa mediterranea del Pais valenciano y Andalucia (18000-3000 B.P.). Tesis doctoral. Universitat de Valencia.

Measuring of δ C13 on charcoal was coupled with C14 AMS dates. These new results provide interesting data. δ 13C values fluctuate between -22,8 et -24 ‰, heavier than 304

Anthracology and Mediterranean landscape, classical and new approaches Badal E., Bernabeu J., Vernet J.L., 1994. Vegetation changes and human action from the Neolithic to the Bronze Age (7000-4000 BP) in Alicante, Spain, based on charcoal analysis. Vegetation History and Archaeobotany, 3, 155166. Balout L., 1952 . A propos des charbons préhistoriques. Bullétin Societé d’Histoire Naturelle de l’Afrique du Nord, 43,160-163. Barry J.P., Le Roy Ladurie E., 1962. Histoire agricole et phytogéographie. Annales, 17, 434-447. Bazile-Robert E., 1981. Flore et Végétation des gorges du Gardon à la moyenne vallée de l’Hérault, de 40000 à 9500 BP d’après l’anthracoanalyse. Paléobiologie continentale, 12(1), 79-90. Berger J.F., Thiébault S., 2002. The study and significance of charcoal as an indicator of ancient fires: an application to the middle Rhône valley (France). BAR, International Series, 1063. Oxford: BAR Publishing, 25-41. Boulet R., Pessenda L., Telles E., Melfi A., 1995. Une évaluation de la vitesse de l’accumulation superficielle de matière par la faune du sol à partir de la datation des charbons et de l’humine du sol. Exemple des latosols des versants du lac Campestre, Salitre, Minas Gerais, Brésil. C.R. Académie des Sciences Paris, IIa, 320, 287-294. Blackford, J.J., 2000. Charcoal fragments in surface samples following a fire and the implications for interpretation of subfossil charcoal data. Palaeogeography, Palaeoclimatology, Palaeoecology, 164 (1-4), 33-42. Braun-Blanquet J., 1952. Les Groupements Végétaux de la France méditerranéenne. Paris :CNRS. Carcaillet C., 1998. A spatially precise study of fires, climate and human impact within the Maurienne valley, North French Alps. Journal of Ecology 86, 384-396. Carcaillet C., Talon B., 1996. Stratigraphie et datations de charbons de bois dans les Alpes: quelques aspects taphonomiques. Géographie physique et Quaternaire, 50, 233-244 Chabal L., Laubenheimer F., 1994. L’atelier gallo-romain de Sallèles d’Aude, les potiers et le bois. XIV Rencontres international. Antibes : ADPCA, 99-129. Chabal, L., 1991. L’homme et l’évolution de la végétation méditerranéenne, des âges des métaux à la période romaine: recherches anthracologiques théoriques, appliquées principalement à des sites du Bas-Languedoc. Thèse de Doctorat. Université des Sciences Montpellier II. 2 vols. Couvert M., 1976. Traduction des éléments de la flore préhistorique en facteurs climatiques. Libyca, 24, 9-20. Delhon C., 2005. Anthropisation et paléoclimats du Tardiglaciaire à l’Holocène en moyenne vallée du Rhône, études pluridisciplinaires des spectres phytolithiques et pédoanthracologiques de séquences naturelles et de sites archéologiques. Thèse Université. Paris I, Panthéon/ Sorbonne, 2 vols. Durand Tullou A., 1972. Rôle des végétaux dans la vie de l’homme au temps de la civilisation traditionnelle, étude ethnobotanique sur le causse de Blandas, Gard. Journal d’agriculture tropicale et de botanique appliquée, 19 (6-7), 222-246. Durand A., Ruas M.P, in press. La forêt languedocienne IXèmeXIème siècles. In Corvol A., Les forêts de l’Europe occidentale du Moyen Age à aujourd’hui. Actes 24èmes journées internationales d’histoire de Flaran. Toulouse:

Presses universitaire du Mirail. Gardner, J.J., Whitlock, C., 2001. Charcoal accumulation following a recent fire in the Cascade Range, northwestern USA, and its relevance for fire-history studies. The Holocene, 11 (5), 541-549. Gouveia L., Pessenda L., 2000. Datation par le C14 de charbons inclus dans le sol pour l’étude du rôle de la remontée biologique de matière et du colluvionnement dans la formation de latosols de l’état de Sao Paulo, Brésil. C.R. Académie des Sciences Paris, 330, 133-138. Heinz C., 1988. Dynamique des végétations holocènes en Méditerranée nord-occidentale d’après l’anthracoanalyse de sites préhistoriques. Thèse Université. Montpellier II. Jalut G., Esteban Amat A., Bonnet L., Gauquelin T., Fontugne M., 2000. Holocene climatic changes in the Western Mediterranean, from south-east France to southeast Spain. Palaeogeography, Palaeoclimatology, Palaeoecology, 160, 255-290. Jalut G., Esteban Amat A., Riera i Mora S., Fontugne M., Mook R., Bonnet L., Gauquelin T., 1997. Holocene climatic changes in the western Mediterranean: installation of the Mediterranean climate. C.R. Académie des Sciences Paris, Sciences de la terre et des planètes, 325, 327334. Krauss Marguet I., 1981. Analyse anthracologique du gisement postglaciaire de La Poujade (Millau, Aveyron). Paléobiologie continentale, XII (1), 93-110. Le Du R., Saccardy L., 1948. Etude de quelques charbons préhistoriques de la région de Tébessa. Revue africaine, 92, 111-119. McDermott F., Frisia S., Huang Y., Longinelli A., Spiro B., Heaton T., Hawkesworth C., Borsato A., Keppens E., Fairchild I., Van der Borg K., Verheyden S., Selmo E., 1999. Holocene climate variability in Europe: evidence from δ18 O, textural and extension-rate variation in three speleothems. Quaternary Science Review, 18, 10211038. Marquet J.C., 1987. Les modifications de l’environnement postglaciaire en France méditerranéenne d’après les rongeurs de l’Abeurador et Font-Juvénal. In: Premières communautés paysannes en Méditerranée occidentale. Paris : CNRS, 155-163. Momot J., 1955. Méthode pour l’étude des charbons de bois. Bullétin Société. Préhistorique de France, 52,141-143. Nichols, G.J., Cripps, A., Collinson, M.E., Scott, A.C., 2000. Experiments in waterlogging and sedimentology of charcoal: results and implications. Palaeogeography, Palaeoclimatology, Palaeoecology, 164 (1-4), 43-56. Ohlson, M., Tryterud, E., 2000. Interpretation of the charcoal record in forest soils: forest fires and their production and deposition of macroscopic charcoal. The Holocene, 10 (4), 519-525. Pons A., Quézel P.,1998. A propos de la mise en place du climat méditerranéen. C.R. Académie des Sciences Paris, 327, 755-760. Quézel P., 1999. Les grandes structures de végétation en région méditerranéenne: facteurs déterminants dans leur mise en place post-glaciaire. Géobios, 32 (1), 19-32. Quilès D., Rohr V., Joly K., Lhuillier S., Ogereau P., Martin A., Bazile F., Vernet J.L., 2002 . Paléofeux et Végétations des Grands Causses, le cas du Causse Méjean (Lozère, France) Premiers résultats. C.R. Palevo, I, p.59-65.

305

J.L. Vernet Saint-Laurent J. de, 1934. Etude anatomique des rameaux chez les essences forestières d’Algérie avec quelques observations sur les modifications amenées par l’âge. Bulletin Station de recherches forestières du nord de l’Afrique, 2 (2), 61-201. Santa S., 1961. Essai de reconstitution de paysages végétaux quaternaires d’Afrique du Nord. Libyca, 6-7, 37-77. Scott, A.C., Cripps, J.A., Collinson, M.E., Nichols, G., 2000. The taphonomy of charcoal following a recent heathland fire and some implications for the interpretation of fossil charcoal deposits. Palaeogeography, Palaeoclimatology, Palaeoecology, 164 (1-4), 1-31. Solari M.E., Vernet J.L., 1992. Late Glacial and Holocene vegetation of the Corbières based on charcoal analysis at the Cove de l’Espérit (Salses, Pyrénées-orientales, France). Review of Palaeobotany and Palynology, 71, 111-120. Tardy, C.,1999 . Recherches sur les paléofeux en forêt tropicale humide, implications climatiques en Guyane. Thèse. Université Montpellier II. Terral J.F., Mengüal X., 1999. Reconstruction of Holocene climate in southern France and eastern Spain using quantitative anatomy of olive wood and archaeological charcoal. Palaeogeography, Palaeoclimatology, Palaeoecology,153, 71-92. Terral J.F.,1999. La morphométrie à la recherche des origines de la domestication et de la culture des arbres fruitiers : le modèle de l’olivier en Méditerranée nord-occidentale. Revue d’Archéométrie, 23, 101-126. Thiébault S., Vernet J.L., 1992. Végétation méditerranéenne et civilisations préhistoriques, le cas de Font-Juvénal. Bulletin Société Botanique de France, 139, Act. botaniques (2,3,4), 441-450. Thiébault S., 1999.  Dynamique des paysages et intervention humaine, du Tardiglaciaire à l’Holocène, de la Méditerranée aux Préalpes sud-occidentales – Apports de l’analyse anthracologique. Thèse HDR. Université Paris I. Thinon M., 1992. L’analyse pédoanthracologique: aspects méthodologiques et applications. Thèse. Université AixMarseille. Vernet J.L., 1971. Analyse des charbons de bois des niveaux Boréal et Atlantique de l’abri de Châteauneuf les Martigues. Bulletin Museum Histoire naturelle de Marseille, 31, 100-115. Vernet J.L., 1980. Premières données sur l’histoire de la végétation postglaciaire de la Provence centrale d’après l’analyse anthracologique. C.R. Académie des Sciences Paris, 291 D, 853-856. Vernet J.L., 1997. L’homme et la forêt méditerranéenne, de la préhistoire à nos jours. Paris: Errance. Vernet J.L., 2006. History of the Pinus sylvestris and Pinus nigra ssp. Salzmanni forest in the Sub-Mediterranean mountains (Grands Causses, St Guilhem, southern Massif Central, France) based on charcoal from limestone and dolomitic deposits. Vegetation History and Archaeobotany, 16, 2342. Vernet J.L., Meter A., Zeraïa L., 2005. Premières datations de feux holocènes dans les Monts de St-Guilhem-le-Désert (Hérault, France), contribution à l’histoire de la forêt relique de Pinus nigra Arnold ssp. Salzmanni (Dun.) Franco. Geoscience 337, 533-537. Vernet J.L., Ogereau P., Martin A., Bazile F., Zeraia L., 2002.

Anthropisation de la chênaie d’yeuse (Quercus ilex L.), équilibres et déséquilibres. Annales littéraires, série «  Environnement, sociétés et archéologie  ». Presses universitaires franc-comtoises, 423-430. Vernet J.L., Ogereau P., Figueiral I., Machado Yanes C. et Uzquiano P., 2001. Guide d’identification des charbons de bois préhistoriques du sud-ouest de l’Europe. Paris : CNRS. Vernet J.L., Pachiaudi C., Bazile F., Durand A., Fabre L., Heinz C., Solari M.E., Thiébault S., 1996. Le delta C13 des charbons de bois préhistoriques et historiques méditerranéens de 35000 BP à l’actuel. Premiers résultats. C. R. Académie des Sciences Paris, 323, IIa, 319-324. Vernet J.L., Thiébault S., 1987. An approach to northwestern Mediterranean recent prehistoric vegetation and ecologic implications. Journal of Biogeography, 14,117-127. Vernet J.L., Wengler L., Solari M.E., Ceccantini G., Fournier M., Ledru M.P., Soubiès F., 1994. Feux, climats et végétations au Brésil central durant l’Holocène. C.R. Académie des Sciences Paris, 319, III,1391-1397. Villaverde V., Martinez Valle R., Badal E., Guillem P.M., Garcia R., Menargues J., 1999. El Paleolitico superior de la Cova de les Cendres (Teulada-Moraira, Alicante), datos proporcionados por el sondeo efectuado en los cuadros A/B 17. Archivo de Prehistoria Levantina, XXIII, 9-65.

306

Solar influence on Holocene fire history K.J. Willis

Oxford University Centre for the Environment, Dyson Perrins Building, South Parks Road, Oxford, OX1 3QY;

[email protected]

K.D. Bennett

¨ SE-752 36, Uppsala, Sweden Palaeobiology, Uppsala University, Geocentrum, Villavagen 16,

S.G. Haberle

Research School of Pacific and Asian Studies, Building 9, HC Coombs Bldg, The Australian National University, ACT 0200, Australia Abstract: It has been suggested that climates during the Holocene have varied on timescales of about 1500 years, forced by oscillations of a solar cycle. These climatic changes have been implicated as the cause of changes in geological records from, mostly, marine sediments. If there were such changes, and if they were of great enough amplitude to force changes in marine sedimentation, it would be reasonable to expect to see some changes in terrestrial environments. One record that is available across most terrestrial environments is the record of charcoal that results from the burning of vegetation. As burning might be expected to be more frequent under hot dry climates, this might react well to the climatic changes expected for maxima in the solar cycle, and thus provide a contrast to the reverse behaviour during solar minima. We examined eight records of microscopic charcoal from Europe and Australia, from temperate and more arid environments, using a range of statistical techniques to look for behaviour that might be caused by oscillating climates. None of the tests was positive. However, we did identify a range of behaviours indicative of specific structure in burning patterns (they are not random in time or intensity). Apart from an influence in some sites of late Holocene anthropogenic buring, there also appears to be little regional correlation in burning patters. We suggest that, for these eight records at least, burning events appear to be, on the whole, site-specific.

Introduction

biomass burning identified in sedimentary microfossil charcoal records and climate change (e.g. Carcaillet et al. 2002; Pierce et al. 2004; Whitlock 2004). However, no attempts have been made to directly examine the relationship between the solar forcing associated with Bond cycles and Holocene burning regimes. An obvious advantage of the charcoal record is that the response should not be timetransgressive. Variations may occur between regions, however, due to the flammability of the vegetation, regional climate etc. (Camill et al. 2003; Huber et al. 2004; Lynch et al. 2004) and on this basis it can be hypothesised that if there is a millennial-scale burning signature it might be more apparent in Mediterranean than north-west European records. Similarly in regions that have had significant prehistoric presence the record might become distorted — or this may even be a good method of distinguishing between human and natural fires in the Holocene.

There is increasing acceptance amongst palaeoclimatologists that a millennial-scale climatic cycle, operating independently of the glacial-interglacial climate state, is apparent in the Holocene (Bond et al. 1997). A series of climatic shifts oscillating every 1470 ± 500 years is detected by various fossil proxies in marine and terrestrial sequences. Bond et al. (2001) have suggested that solar forcing is the main driving mechanism behind these climate changes. Broad teleconnective responses associated with these Bond cycles are suggested to have resulted in global climate changes including intervals of cooler climates in mid to high latitudes (Bond et al. 2001; Bj¨orck et al. 2005; Hu et al. 2003), reduced rainfall and aridification events in mid-latitudes and reduced monsoonal activity in low latitudes (Fleitmann et al. 2003; Wang et al. 2005). However the evidence for vegetation change associated with these Bond cycles is far from ubiquitous in time or space and identifying the influence of a millennial-scale climatic cycle on the Holocene vegetation dynamics is not straightforward (Heiri et al. 2004). Perceived changes may be region specific, time-transgressive (lags due to migration rates) or influenced by other abiotic and biotic factors.

In this study we aim to test statistically various hypotheses relating to millennial-scale climate oscillation and to examine the influence of solar variability on Holocene fire history using microfossil charcoal records from lake sedimentary sequences. Are there distinguishable burning patterns apparent in Holocene microfossil charcoal records? Are these related in time and frequency to the Bond cycles? Is there a latitudinal spread apparent from the Mediterranean region to northern Europe? Do these burning cy-

A number of links have been made between Holocene

g. fiorentino, d. magri (eds). Charcoals From The Past, BAR Int. Ser., 2008.

1

K.J. Willis, K.D. Bennett, S.G. Haberle

cles become less distinguishable with the onset of anthropogenic burning? Failing to find a correlation between Bond cycles and burning cycles would enable us to reject a hypothesis that there is a causal relationship. Finding a correlation would leave that hypothesis open for further investigation. All ages are given as calibrated years before present.

Methods Charcoal data

Whitlock & Millspaugh 1996). Small lake basins (?

!"#$%&'$()'$

$@ 6% #6 A

Fig. 4. The eight microfossil charcoal sequences with intervals of solar maxima (taken from Bond et al. 2001) indicated.

*

+***

,***

-***

.***

/***

0***

1***

2***

3***

+****

++***

+,***

+-*** *

.

2

+,

+0 *

:&+*,&&;',&;'!,&76$%!+

.

:&+*!-

2

+,

+0 *

.

:&+*!,

2

+,

* &

*J. *J2 +J, +J0 *

,

:&+*!,

9

315

.

0

2

+*

*

+*

*

:&+*!,&&;',&;'!,&76$%!+

,

.

0

2

&

+

,

-

.

/

*

,

:&+*!,

.

0

2

+*

K.J. Willis, K.D. Bennett, S.G. Haberle Table 5. Results of the chi-squared test between solar maximum/minimum and charcoal peaks (a) >3 standard deviations and (b) charcoal peaks >mean where values greater than 3.84 indicate that there is less than 5% probability that the association could have arisen by chance. Site

Solar min

(a) Solar max

Solar min

(b) Solar max

Euramoo

0.13

4.50

0.13

0.02

Pardillas

1.03

1.02

0.21

0.91

Kis Mohos To´

1.02

1.02

0.06

0.35

Mlaka

1.04

1.03

0.44

1.09

Dallican

1.04

1.03

0.09

0.31

Hockham Mere

1.09

2.96

0.43

2.41

High Lake

1.02

1.02

0.06

0.02

Jezero

1.02

1.02

0.48

0.03

also no statistically significant correlation.

demonstrated in this paper should be employed to examine associations between charcoal peaks and i) intervals of vegetation change (measured through rates of change), ii) species composition and iii) known intervals of human activity. By examining these associations it should then be possible to start to critically assess and test statistically the driving mechanisms responsible for Holocene burning regimes.

Conclusions The preliminary conclusion that can be drawn from this study is that at the eight sites examined there is no statistically significant relationship between Holocene solar variability (maxima or minima) and burning. But in drawing this preliminary conclusion, a number of other interesting results are also highlighted: First, that at every site examined the charcoal records indicate a specific structure suggesting distinctive burning events. Second, with the exception of two sites there is no obvious increase in burning frequency or intensity with increased human presence in the region. Third, that there appears to be little correlation between charcoal records from different regions. These three results together therefore suggest that burning, at least for the eight records studied, appears to be more strongly influenced by local abiotic and biotic factors than regional factors, and that burning events appear to be, on the whole, site specific (cf. Bennett et al. 1990).

ACKNOWLEDGEMENTS

We thank Maja Andriˇc, Janice Fuller, Adam Gardner and Maria Fernanda Sanchez-Go˜ni for making their charcoal data freely available for these analyses.

References Andresen, C.S., Bond, G., Kuijpers, A., Knutz, P.C. & Bj¨orck, S. (2005). Holocene climate variability at multidecadal time scales detected by sedimentological indicators in a shelf core NW off Iceland. Marine Geology 214, 323–338.

So is it possible to identify predominant mechanisms responsible for burning regimes apparent in the Holocene sedimentary sequences? In order to rule out totally the influence of solar variability, it would probably be prudent to examine more ‘ecologically sensitive’ regions, for example the Near East and Mediterranean. Other palaeo-proxies for the 4.2 ka aridification event (one of the Bond minina) in the Mediterranean, for example, have indicated significant environmental change occurring around this event (Weiss & Bradley 2001) and this probably warrants further investigation. In terms of other factors influencing the burning regime, however, a similar approach to that

Andriˇc, M. & Willis, K.J. (2003). The phytogeographical regions of Slovenia: a consequence of natural environmental variation of prehistoric human activity? Journal of Ecology 91, 807–821. Bennett, K.D., Boreham, S., Sharp, M.J. & Switsur, V.R. (1992). Holocene history of environment, vegetation and human settlement on Catta Ness, Lunnasting, Shetland. Journal of Ecology 80, 241–273. Bennett, K.D., Simonson, W.D. & Peglar, S.M. (1990).

316 10

Solar influence on Holocene fire history Fire and man in post-glacial woodlands of eastern England. Journal of Archaeological Science 17, 635–642.

forcing of the Indian monsoon recorded in a stalagmite from southern Oman. Science 300, 1737–1739.

Birks, H.J.B. (1997). Reconstructing environmental impacts of fire from the Holocene sedimentary record. In J.S. Clark, H. Cachier, J.G. Goldhammer & B. Stocks (Eds.), Sediment Records of Biomass Burning and Global Change, NATO ASI Series, volume I 51, pp. 295–311.

Fuller, J.L. (1997). Holocene forest dynamics in southern Ontario, Canada: fine-resolution pollen data. Canadian Journal of Botany 75, 1714–1727. Gardner, J.J. & Whitlock, C. (2001). Charcoal accumulation following a recent fire in the Cascade Range, northwestern USA., and its relevance for fire-history studies. The Holocene 11, 541–549.

Bj¨orck, S., Muscheler, R., Kromer, B., Andresen, C.S., Heinemeier, J., Johnsen, S.J., Conley, D., Koc¸, N., Spurk, M. & Veski, S. (2005). High-resolution analyses of an early Holocene climate event may imply decreased solar forcing as an important climate trigger. Geology 29, 1107–1110.

Heiri, O., Tinner, W. & Lotter, A.F. (2004). Evidence for cooler European summers during periods of changing meltwater flux to the North Atlantic. Proceedings of the National Academy of Sciences of the USA 101, 15285–15288.

Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M.N., Showers, W., Hoffman, S., Lotti-Bond, R., Hajdas, I. & Bonani, G. (2001). Persistent solar influence on north Atlantic climate during the Holocene. Science 294, 2130– 2136.

Hu, F.S., Kaufman, D., Yoneji, S., Nelson, D., Shemesh, A., Huang, Y., Tian, J., Bond, G., Clegg, B. & Brown, T. (2003). Cyclic variation and solar forcing of Holocene climate in the Alaskan subarctic. Science 301, 1166–1171.

Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi, P., deMenocal, P., Priore, P., Cullen, H., Hajdas, I. & Bonani, G. (1997). A pervasive millennial-scale cycle in North Atlantic Holocene and glacial climates. Science 278, 1257– 1266.

Huber, U.M., Markgraf, V. & Sch¨abitz, F. (2004). Geographical and temporal trends in Late Quaternary fire histories of Fuego-Patagonia, South America. Quaternary Science Reviews 23, 1079–1097.

Camill, P., Charles E. Umbanhowar, J., Teed, R., Geiss, C.E., Aldinger, J., Dvorak, L., Kenning, J., Limmer, J. & Walkup, K. (2003). Late-glacial and Holocene climatic effects on fire and vegetation dynamics at the prairieforest ecotone in south-central Minnesota. Journal of Ecology 91, 822–836.

Lynch, J.A., Hollis, J.L. & Hu, F.S. (2004). Climatic and landscape controls of the boreal forest fire regime: Holocene records from Alaska. Journal of Ecology 92, 477–489. Maxwell, A.L. (2004). Fire regimes in north-eastern Cambodian monsoonal forests, with a 9300-year sediment charcoal record. Journal of Biogeography 31, 225–239.

Carcaillet, C., Almquist, H., Asnong, H., Bradshaw, R.H.W., Carri´on, J.S., Gaillard, M.J., Gajewski, K., Haas, J.N., Haberle, S.G., Hadorn, P., M¨uller, S.D., Richard, P.J.H., Richoz, I., R¨osch, M., S´anchez Go˜ni, M.F., von Stedingk, H., Stevenson, A.C., Talon, B., Tardy, C., Tinner, W., Tryterud, E., Wick, L. & Willis, K.J. (2002). Holocene biomass burning and global dynamics of the carbon cycle. Chemosphere 49, 845–863.

Ohlson, M. & Tryterud, E. (2000). Interpretation of the charcoal record in forest soils: forest fires and their production and deposition of macroscopic charcoal. The Holocene 10, 519–525. Pierce, J.L., Meyer, G.A. & Jull, A.J.T. (2004). Fireinduced erosion and millennial-scale climate change in northern ponderosa pine forests. Nature 432, 87–90.

Clark, J.S., Lynch, J., Stocks, B.J. & Goldammer, J.G. (1998). Relationships between charcoal particles in air and sediments in west-central Siberia. The Holocene 8, 19–29.

Press, W.H., Teukolsky, S.A., Vetterling, W.T. & Flannery, B.P. (1992). Numerical Recipes in C: the Art of Scientific Computing (Second ed.). Cambridge: Cambridge University Press.

Clark, J.S. & Patterson, III, W.A. (1997). Background and local charcoal in sediments: scales of fire evidence in the paleorecord. In J.S. Clark, H. Cachier, J.G. Goldhammer & B. Stocks (Eds.), Sediment Records of Biomass Burning and Global Change, NATO ASI Series, volume I 51, pp. 23–48.

S´anchez Go˜ni, M.F. & Hannon, G.E. (1999). High-altitude vegetational pattern on the Iberian Mountain Chain (northcentral Spain) during the Holocene. The Holocene 9, 39– 57.

Clark, J.S. & Royall, P.D. (1995). Transformation of a northern hardwood forest by aboriginal (Iroquois) fire: charcoal evidence from Crawford Lake, Ontario, Canada. The Holocene 5, 1–9.

Wang, Y., Cheng, H., Edwards, R.L., He, Y., Kong, X., An, Z., Wu, J., Kelly, M.J., Dykoski, C.A. & Li, X. (2005). The Holocene Asian monsoon: links to solar changes and North Atlantic climate. Science 308, 854–857.

Clark, R.L. (1982). Point count estimation of charcoal in pollen preparations and thin sections of sediments. Pollen et Spores 24, 523–535.

Weiss, H. & Bradley, R.S. (2001). What drives societal collapse? Science 291, 609–610. Whitlock, C. (2004). Forests, fires and climate. Nature 432, 28–29.

Fleitmann, D., Burns, S.J., Mudelsee, M., Neff, U., Kramers, J., Mangini, A. & Matter, A. (2003). Holocene

317 11

K.J. Willis, K.D. Bennett, S.G. Haberle Whitlock, C. & Larsen, C.P.S. (2001). Charcoal as a fire proxy. In J.P. Smol, H.J.B. Birks & W.M. Last (Eds.), Tracking Environmental Change Using Lake Sediments. Volume 3: Terrestrial, algal, and siliceous indicators, pp. 75–97. Dordrecht: Kluwer Academic Publishers. Whitlock, C. & Millspaugh, S.H. (1996). Testing the as-

318 12

sumptions of fire-history studies: an examination of modern charcoal accumulation in Yellowstone National Park, USA. The Holocene 6, 7–15. Willis, K.J., S¨umegi, P., Braun, M., Bennett, K.D. & T´oth, A. (1998). Prehistoric land degradation in Hungary: who, how and why? Antiquity 72, 101–113.