Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry 9781407314303, 9781407344027

Table of contents :
Front Cover
Title Page
Copyright
TABLE OF CONTENTS
Preface
Part I. Archaeological Materials and Artefacts
Part II. Archaeological Landscapes

Citation preview

2016 PHOTOS-JONES (Ed.)

The 6th Symposium of the Hellenic Society for Archaeometry (HSA) took place in May 2013, at the Acropolis Museum in Athens. The HSA Symposium proceedings are aimed particularly at young researchers working on Greek materials and landscapes as a venue for presenting their work. This volume comprising thirty-two papers is divided in two parts: the first deals with materials (ceramics, glass, metal, paintings, paper) and the second with the landscape and its multifaceted aspects (dating, prospection, visualisation). Within each section issues of conservation, dating, and computer applications are interwoven together with aspects of intangible heritage.

BAR S2780

The Hellenic Society for Archaeometry (HSA) was founded in 1982 and is the professional body representing academics and research centers in Greece and within the Greek Archaeological Service who employ physical and earth sciences in archaeology. The HSA is a vibrant group of researchers who over the years have made a substantial contribution to the understanding of Greece’s Cultural Heritage.

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Edited by

6TH SYMPOSIUM OF THE HELLENIC SOCIETY FOR ARCHAEOMETRY

E. Photos-Jones in collaboration with: Y. Bassiakos, E. Filippaki, A. Hein, I. Karatasios, V. Kilikoglou and E. Kouloumpi

BAR International Series 2780 B A R

2016

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Edited by

E. Photos-Jones in collaboration with: Y. Bassiakos, E. Filippaki, A. Hein, I. Karatasios, V. Kilikoglou and E. Kouloumpi

BAR International Series 2780 2016

First Published in 2016 by British Archaeological Reports Ltd United Kingdom BAR International Series 2780 Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

© The editor and contributors severally 2016 The Authors’ moral rights under the 1988 UK Copyright, Designs and Patents Act, are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher.

ISBN 9781407314303 paperback ISBN 9781407344027 e-format DOI https://doi.org/10.30861/9781407314303 A catalogue record for this book is available from the British Library

Cover Images: Iron sword blade decorated with birds and fish, 4th century BC tomb at Katerini, Northern Greece (Chapter 15) Virtual reconstruction of a Neolithic village in Hungary based on geophysical survey (Chapter 28)

All BAR titles are available from: British Archaeological Reports Ltd Oxford United Kingdom Phone +44 (0)1865 310431 Fax +44 (0)1865 316916 Email: [email protected] www.barpublishing.com

Table of Contents Pages Preface

v

Part I: Archaeological Materials and Artefacts 1

2 3 4 5

6 7 8 9 10

11

12 13 14 15 16

17 18 19

20

21

The ‘management’ of painted and monochrome pottery of Neolithic Thessaly, Central Greece: technology and provenance Vasso Rondiri and Eleni Asderaki-Tzoumerkioti Craftsmanship of big storage pithoi in the Early Helladic settlement of Helike, Achaea Dora Katsonopoulou, Ioannis Iliopoulos, Stella Katsarou and Vayia Xanthopoulou The Riace bronzes: recent work on the clay cores R. Jones, D. Brunelli, V. Cannavò, S.T. Levi and M. Vidale Technical aspects of two Byzantine glazed relief pottery icons George Mastrotheodoros, Konstantinos Beltsios, Yannis Bassiakos and Varvara Papadopoulou Identification and diagnostic study of Neolithic beads from Kalyvia, Attica, applying Raman spectroscopy E. Kougemitrou, G. Moraitou, S. Raftopoulou, G. Chela and K. Polykreti Portable XRF analysis of glass tesserae from the House of Fourni in Delos, Greece Francesca Licenziati and Thomas Calligaro Research on the glass oinochoe in the Lambropoulos Private Collection Α. Thanos, S. Vivdenko, D. Lampakis and P. Manoudis Spectroscopic study of a historical glass collection from Thebes, Greece, by Raman and IR E. Palamara, N. Zacharias, E.I. Kamitsos, A. Oikonomou, D. Palles and D. Möncke Topographic multi-scale surface analysis for the study of stone celts’ polishing techniques A. Boleti, H. Procopiou, H. Zahouani, R. Vargiolu and S. Mezghani Did copper actually arrive in the Aeolian Islands in the fourth millennium BC? The evidence from a small but iconic fragment of vitreous material thought to be copper slag Maria Clara Martinelli, Effie Photos-Jones and Sara Tiziana Levi Selective use of arsenical copper during the Mycenaean period: the evidence form Pylia (Messenia, Greece) Charilaos E. Tselios, Eleni Filippaki and Georgios S. Korres Copper production during the Early Bronze Age at Aghios Antonios, Potos on Thasos Nerantzis Nerantzis and Stratis Papadopoulos A comparative study of Cypriot bronzes dated to the Late Bronze and the Early Iron Age Andreas Charalambous, Vasiliki Kassianidou and George Papasavvas Conservation of a silver coin hoard from the Geometric period Necropolis in Samos Ioannis P. Staikopoulos Gold inlay decoration on the blade of an iron sword from the Tomb A at Katerini Vasiliki Michalopoulou, Barbara Smit-Douna and Ioannis Karapanagiotis The so-called ‘hellicoidal’ ore washeries of Laurion: their actual function as circular mills in the process of beneficiation of silver and lead contained in old litharge stocks George D. Papadimitriou Technical and chemical examination of a small silver treasure dated to the 6 th-7th century AD D. Kotzamani and J.F. Merkel Study of gold and silver finds from ancient Akanthos, Chalkidiki, N Greece S. Vivdenko, E. Trakosopoulou-Salakidou and P. Manoudis Physico-chemical study of pigments from the prehistoric settlement of Ialysos, Rhodes Yorgos Facorellis, Stamatis Boyatzis, Τoula Marketou, Eleni Tziamourani and Eleni Ioakimoglou The grounds of portable icons: written sources and scientific examination of post-Byzantine icons from Epirus George P. Mastrotheodoros, Konstantinos G. Beltsios and Varvara Papadopoulou Ground layers: a promising dating factor E. Kouloumpi, A. Rousaki, A.V. Terlixi, A.P. Moutsatsou, G.S. Polymeris and K.M. Paraskevopoulos

3

13 21 29 33

41 49 59 65 73

81

89 95 101 107 113

119 127 133

141

147

22

23 24

25

26

27

28 29

30 31 32

Conservation methods of Edo period Japanese paper garment ‘Kamiko’ Kyriaki Lentzi and Yorgos Facorellis

Part II: Archaeological Landscapes Landscape evolution in the Kifissos floodplain E.D. Chiotis Aiding and abetting the archaeological enquiry: geochemical work-in-progress at the site of San Vincenzo, Stromboli, Aeolian Islands, Italy Andrea Di Renzoni, Gianna Ayala, Daniele Brunelli, Sara Tiziana Levi, Stefano Lugli, Effie Photos-Jones, Alberto Renzulli and Patrizia Santi New geoarchaeological and palaeoenvironmental studies in Mani Peninsula (S. Peloponnesus) by employing luminescence dating techniques I. Christodoulakis, C. Athanassas, Y. Bassiakos and C. Varotsos A.Sho.Re. 2011-2015, SE Kephallenia in the Ionian Sea: investigating the geoarchaeology of the coastal zone E.Yiannouli The relationship between early settlements in arid environments and sources of water supply: the case of the Bronze Age site of San Vincenzo, Stromboli, Italy Andrea Di Renzoni, Leandro Lopes, Maria Clara Martinelli and Effie Photos-Jones Reconstructing archaeo-landscapes: myth versus reality A. Sarris, A. Chliaoutakis, S. Dederix and J.C. Donati Luminescence dating and quartz grain surface features of aeolian sediments from Agia Napa, Cyprus E. Tsakalos, C. Athanassas and Y. Bassiakos Luminescence dating of Quaternary coastal deposits of Evoikos Gulf (Central Greece) M. Kazantzaki, C. Athanassas, Y. Bassiakos and E. Tsakalos Intangible Cultural Heritage and place Despoina Karavia and Andreas Georgopoulos The ayiasmata (holy springs) of Lemnos, NE Greece: an archaeo-hydrogeological study E. Photos-Jones, E. Zagana and P. Roumelioti

Pages 153

161 167

175

179

187

195 201

207 215 221

Preface The 6th Symposium of the Hellenic Society for Archaeometry (HSA) took place in May 2013 (May 1618) in a suitably inspiring location, at the Acropolis Museum in Athens. The Society, which was founded in 1982, is the representative professional body of all those individuals in academic and research centers in Greece, and within the Greek Archaeological Service, who practice archaeological science, or the application of the physical and earth sciences in archaeology. Since its inception the Society has promoted the study of Archaeometry in Greece through meetings, conferences and seminars, and, over the years, the publication of the proceedings of its conferences. The 6 th conference of the HSA was organized by its then governing Council comprising Y. Bassiakos, E. Filippaki, A. Hein, V. Kilikoglou, I. Karatasios, E. Kouloumpi and E. Photos-Jones. The members of the Scientific Committee were S. Boyatzis, P. Day, A. Hauptmann, C. Heron, V. Kassianidou, V.Kiriatzi, G. Kitis, Y. Maniatis, N. Maravelaki, A. Sarris, S. Sotiropoulou, P. Triantafyllidis and P. Vandenabeele. In Greece and abroad, the field of Archaeological Science has made significant contributions to the study of material culture of the past as a means of understanding the nature, role and function of ancient artifacts as well as the human agency underwriting their production and use. It has expanded considerably our understanding of the landscape surrounding the archaeological resource. Since its early and well-publicised contribution nearly years ago in the form of C-14 dating, Archaeological Science has provided tools, via geophysics and remote sensing, to establish the presence of buried remains prior to their excavation; it has spearheaded research in conservation; it has addressed issues of the palaeoenvironment, diet, disease and ancient DNA, and more recently it has entered the field of computer application to archaeology in issues of 3-D visualization and landscape reconstruction. In many respects Archaeological Science has woven itself within the archaeological enquiry, if not seamlessly, yet in a manner that makes its implementation a necessity rather than a desirable. Since its inception in the early 1980s, Αρχαιομετρία or Archaeological Science in Greece has continued to address the questions and complexities of our nation’s immensely rich material cultural heritage. This has indeed been the subject of the previous five HSA symposia. The theme of the 6th symposium was Craft-based Cultural Influences attempted to look beyond materials and into our nation’s equally rich Intangible Heritage. A craft is not simply about materials but also about the skills of the men and women behind the craft, and of course without those skills there would be no craft. An archaeological science approach to the study of the Intangible Heritage is still at its infancy. Although more than seven papers were presented in the course of the symposium, only two were submitted for publication, perhaps highlighting the difficulty in formulating a new methodology for its discussion. Nevertheless, it remains an intriguing prospect. In the event, some 120 oral papers and posters were presented during the course of the three-day symposium. These presentations, a quarter of which were subsequently submitted for publication in this volume of proceedings, covered very adequately current concerns with provenance and technology of materials, geoarchaeology and landscape studies. Notable and most encouraging were the number of presentations in conservation science, highlighting the fact that conservation and archaeological science have common directions. The HSA Symposium proceedings have always aimed to give young researchers, working on Greek materials and the landscape, a venue for presenting their work, while recognising that that work is frequently in progress. This volume has broken new ground on two fronts: first, in asking the contributors to present their work solely in English; secondly in being published in the BAR International Series. In that way the work of young (and older researchers) can be easily accessed and incorporated within the broad-based archaeological science bibliography. Strictly speaking, most of the contributions in this volume represent reports on specific topics rather than in-depth, long-term investigations which would be expected to be published in established academic journals. However, given the tremendous wealth of the archaeological resource in Greece, the need for archaeological scientists to play their part in understanding this resource, ‘as the story evolves’, the HSA may in the future consider an open- access venue for its conference proceedings.

v

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

This volume is divided in two parts: materials and the landscape. The first deals with materials (ceramics, glass, metal, paintings, paper) and the second with the landscape and its multifaceted aspects (dating, prospection, visualisation). Within each section issues of conservation, dating and computer applications are interwoven; also aspects of the Intangible Heritage. The Society is very grateful to its sponsors: NCSR “Demokritos”, S&B Industrial Minerals, Mistras Group Hellas, Bruker, Τέχνης Μέτρον, and Potingair Press. We are indebted to Richard Jones and Allan Hall, for their invaluable contribution in the preparation of this volume and to the many anonymous reviewers in Greece and abroad who read carefully and commented in detail on the papers submitted. Finally, the editor would like to extend her thanks to the authors for their patience in responding to the many queries raised by the reviewers and herself, alike. For all, it has been a long process, yet hopefully richer in experience and understanding as a result. The Organising Committee of the 6th Symposium of the Hellenic Society for Archaeometry Yannis Bassiakos, Eleni Filippaki, Anno Hein, Ioannis Karatasios, Vassilis Kilikoglou, Eleni Kouloumpi and Effie Photos-Jones

vi

Part I Archaeological Materials and Artefacts

Chapter 1

The ‘management’ of painted and monochrome pottery of Neolithic Thessaly, Central Greece: technology and provenance Vasso Rondiri1 and Eleni Asderaki-Tzoumerkioti2 1

Ephorate of Antiquities, Magnesia, Athanassaki 1, 380 01 Volos, Greece 2 UCL Qatar, Doha, Qatar Corresponding author: [email protected]

Abstract The research potentials of Neolithic ceramic material have been stressed repeatedly by archaeologists concerned with ancient pottery in Greece. In Thessaly, Central Greece, the density of Neolithic sites offers an additional research tool to the study of Neolithic pottery in matters of technology and provenance. The present study examines the results of macroscopic and chemical analysis of painted pottery of the Early and Middle Neolithic period deriving from excavations and surface surveys of prehistoric sites in Thessaly, having as a goal further investigation on matters of production, provenance and circulation of Neolithic pottery. An experimental non-destructive method was used for the chemical analysis of the painted sherds in order to examine the paste, pigments and slips applied on their surface. The analysis was performed with a pXRF device in the laboratories of the Ephorate of Antiquities in Volos and had as a result the identification of the pigments as well as specific elements which show a differentiation among pottery of different geographical areas of Thessaly and maybe a first indication of its provenance. Keywords: Middle Neolithic sherds, pigments, pXRF analysis

The work presented in this paper started in 2012 as an experiment and is still ongoing. It concerns the technological study of pottery from ten Neolithic settlements in Thessaly. The understanding of the decoration, slip and pigment technology of this specific chronological period is expected to add information to existing data from other areas of Greece and to shed light on the relationship of people who implemented them inside and outside the settlement. In the first stage of the study macroscopic analysis was performed followed by non-destructive analysis with pXRF which was mainly used as a forerunner of other techniques to follow. However, it was possible to understand the quality of clay and determine the white slip and the pigments which were used for the decoration on the surface.

Introduction The study of Neolithic pottery in Thessaly goes back more than one century and is directly related to research of the same period all over Greece. According to the history of research, archaeologists discovered the interpretative potentials of pottery at an early stage (Tsoundas 1908; Wace and Thompson 1912). Based on fabrics and their characteristics they have supported different theories or rejected others according to their scientific interests, their beliefs, identities or ideologies. These are exactly the issues traced in material culture as a basic part of Neolithic communities on a micro- or macro-scale. The way agency is expressed within the dynamics of material culture is based on how humans and things are intertwined as well as on the forms it takes in social practice (Latour 1999; Bourdieu 1977). On the other hand, the relationship between technology and agency is based on the agents’ products and their involvement with production and use (Dobres 2000). In pottery, the potters’ choices are interrelated and also interact with the communities which they express through their products (Lemonnier 1993). Since pottery production is dynamically expressed in specific social relations and practices, the choices concerning preservation or rejection of a certain production is a matter of social decision and not simply transfer of knowledge (Jones 2002).

The macroscopic analysis The pottery samples under examination belong to a wider group of ceramic material analyzed macroscopically for a research project concerning technology and spatial distribution of Neolithic pottery in Thessaly (Rondiri 2009). The original material of this project was selected using a sampling strategy according to which Thessaly was divided into five zones based on geomorphology (stratified sampling: Shennan 1988, 315-316) and using specific criteria according to which different sites were selected for study. The pottery was sampled based on qualitative criteria (ware type, colour, shape etc.) and

3

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry included sherds covering all typical characteristics of Thessalian Neolithic pottery.

Narthaki or Magoula at Ambelia is situated near Narthaki village 15km southeast of Pharsala. The only report of this settlement is made by Gimbutas (1974, 2 & 4, n. 2). Siaterli is a site near the village Dilofo, west of Narthaki. There are reports by Halstead (1984, nr. 605) and Kilian (1976).

The main purpose of the present study is to examine the painted so-called Α3β ware of the Early and Middle Neolithic period. 45 Early and Middle Neolithic sherds belonging to ten Neolithic sites were selected from the original sample (Rondiri 2009) and were examined macroscopically. The painted Α3β ware constitutes only a small amount of the whole pottery group (10-20%), a fact that explains the small number of sherds involved. For the purposes of chemical analysis it was important to select only the sherds that preserved their decoration best. These sites were chosen from four of the five geographical zones (zones A, B, D, E) for specific reasons concerning the sampling strategy used. The recorded technical characteristics of the pottery were integrated into fabric groups (see below).

Larisa area (Zone B) Rodia 3 is located at a distance of 15km northeast of Tirnavos and 1.5km southeast of Rodia by the River Pineios near its confluence with Titaresios. The pottery is surface material collected by Theocharis (1964, 263). There are reports by Arvanitopoulos (1910, 261) and Gallis (1992, 178). Karditsa area (Zone A) Magoulitsa 1 lies near the village Artesiano on the 5th km of the road Karditsa-Trikala on the right side. This site was explored by Papadopoulou (1958).

The sites (Fig. 1) Almyros area (Zone E) Kamara is situated in the Sourpi plain east of the village Drimon. The small-scale excavation (Rondiri 2004, 2005) brought to light an elliptical stone structure, a hearth and remains of houses with stone foundations and post holes. The finds derived from excavation and surface survey were rich in pottery and lithics (Reinders et al. 2009) and testify the existence of a settlement bearing all characteristics of Neolithic ‘tradition’ in Thessaly. Zerelia is situated west of Efxinoupoli near Almyros in a unique landscape surrounded by two lakes. The trial excavation (Wace and Thompson 1912, 150-166) revealed a settlement with a deposit 6-8m high starting in the Middle Neolithic period and going into historical times. Recent research (Malakassioti et al. 2012) has revealed Neolithic and Bronze Age structures confirming the results of the past. Aidiniotiki Magoulat lies within the territories of the 111th Hellenic Airforce near N. Anchialos, between Almyros and N. Anchialos near the village Aidini. Trial excavations were undertaken by Arvanitopoulos (1907, 171) and Tsoundas (1908, 1112, n.60, Fig.2, & 241, 244, 246-247, 249). Dautza or Perdika lies north of the stream Asprorema at the northern edge of the province of Almyros between the villages of Perdika and Argillochori. The site was first visited by Tsoundas (1908, 13, n. 63).

The painted ware Α3β Under this designation we mean the characteristic painted ware of all phases of the Middle Neolithic period in the whole of Thessaly. It is also well known in the bibliography as red on white or red on cream ware (Tsoundas 1908; Wace and Thompson 1912; Gallis 1992; Demoule et al. 1988; Schneider et al. 1991, 1994). Its fabric is mainly fine grained but is sometimes medium grained. Firing was done under oxidizing or mixed conditions and the surface’s hardness is either high or low. The vessel forms represented are skyphoi, basins and cups and to a lesser extent bowls and jars. The last two forms are not represented in the medium-grained ware. The decorated outer surface is either high or low burnished. There are only a few non-burnished samples. The colour of the clay is mainly red to pink and brown or reddish brown. The colour of the surface is white, pink or pale yellow for the slip and dark red or dark brown for the paint. The decorative motifs of the Middle Neolithic are either of the ‘solid style’ for the early phases or of the ‘linear style’ for the later phases. Quite common for the middle and late phases is the ‘flame pattern’.

The fabrics (Table 1) Fabric 1 (ΛΑ) of fine-grained pottery with white particles is represented by eight sherds only. It seems to be a common feature in sherds from Zerelia and Dautza in Almyros (zone E) and also from Tsangli in the area of Pharsala (zone D). Fabric 2 (ΛΜΑ+ΛΜΑΣ) of fine-grained pottery either with mica, white particles or with mica, white and dark particles seems to be the most common fabric group. This is a general observation concerning also the original pottery sample coming from all areas of Thessaly. Fabrics 3 (ΜΜΑ) and 4 (ΜΜΑΣ) are found in only four sherds of medium-grained pottery with mica, white particles and/or dark particles.

Pharsala area (Zone D) Tsangli is situated by the village Eretria near the road leading from Volos to Pharsala very close to the River Enipeas. Trial excavation, conducted by Tsoundas (1908, 8-9, n. 38) and Wace (Wace and Thompson 1912, 9, n. 38 & 86-130), revealed remains of two houses as well as many finds, especially pottery. The deposit was 10m high. Dasolofos was probably discovered by Giannopoulos (1902, 428). There are also reports by Halstead (1984, nr. 645) and Gimbutas in the publication of nearby Achilleion (Gimbutas 1974, 2 & 4).

4

V. Rondiri and E. Asderaki, The ‘management’ of painted and monochrome pottery of Neolithic Thessaly Generally speaking it seems that Α3β ware was widely used in many parts of Thessaly. Fabric 2 seems to be widely distributed in all four zones under study, and fabric 1 is detected at neighbouring sites, a matter that needs further examination.

certain elements were noted from one area to another, however they are not discussed further here pending the results of more refined techniques of analysis. White slip White-slipped vessels are rather rare in the Middle Neolithic period and thus the number of the examined samples is rather limited. In many cases the slip has worn off and is preserved only in small spots making analysis difficult. The white slip is usually used as a background for painted decoration (Yiouni 2001). When the clay body was cream, yellow, pale yellow or yellowish gray a red pigment was directly placed on the surface of the vase. When the clay body was red or reddish-yellow or even pink the surface was usually covered with a white slip and then was usually painted with a red pigment. From the macroscopic examination of the samples and from the recent literature, we can assume that the white slip was placed directly on the clay body of the vase whereas the red pigment was used on top of it as decoration (Kotsakis 1983; Yiouni 2001). Most of the time burnishing took place resulting in a shiny surface after firing. In the case that burnishing did not take place, a white slip was applied thickly on the surface giving the impression of a paste. At times of high temperature in firing the result was a very smooth and well burnished surface; in this case the distinction between the slip and the pigment was difficult. It seems quite certain though that the chaîne opératoire following the surface treatment had three stages: covering of the surface with a white slip, applying decoration in red pigment on the white slip, burnishing, and then firing.

Chemical analysis Here the aims were: a) To determine similarities or differences between clays of the same or different settlements. b) To identify the white slip. To determine possible similarities or differences in the slip within and between settlements. c) To identify the red pigment. To determine possible similarities or differences in the pigments within and between settlements. d) To determine possible differences between geographical areas. 47 samples were analysed non-destructively with the Xray Fluorescence technique.1 Two of them are dated to the Early Neolithic and the rest to the Middle Neolithic period. The analysis was performed in the Laboratories of the Department of Conservation of Antiquities of the Archaeological Ephorate of Magnesia, using the portable X-ray Fluorescence (pXRF) spectrometer which was set up at the NCSR ‘Demokritos’ (Karydas 2007). The analysis was essentially qualitative, but relative peak intensities were recorded. For the analysis of the sherds the pXRF operational conditions were set as follows: for the light elements 20kV high voltage, unfiltered mode and tube current 100μA, while for the major elements 40kV high voltage, filtered mode and tube current 200μA. The time of each measurement was 500sec. Two to three measurements were performed per area, i.e. the clay, the white slip and the red pigment.

Again here, the identification of the slip by means of pXRF analysis was based either on the detection of a specific single element or on a certain group of elements which determine a particular pigment compound. The relative XRF peak intensities among the detected elements were also taken into consideration and they should be in accordance either with theoretical estimated relative intensities, or with experimental ones determined by measuring standard pigments. These procedures were followed in our case, and in most cases allowed the confirmation of the slip identity.

Results-Discussion From the 47 samples analyzed so far 32 have a white slip, one has a yellowish slip and 41 preserve red or brown pigment. The lack of quantification is not a drawback at this stage at least as regards the slip and pigments analyses since they can be characterized by the presence either of a single inorganic element or by the identified chemical compound.

33 samples were examined and presented here as they preserve enough quantity of slip on their surface which varies in two colours. In most cases the slip is white and only in one case yellowish. A variety of raw materials were used for the production of this white slip which helped both the deposition of the red pigment and the emergence of colour hues; calcareous-rich material has been identified in 22 samples, selenite or gypsum in ten samples and lead-white in one sample (Table 2). Lead white has not previously been detected in such an early period, however its presence is not connected with contamination. The presence of high iron (Fe) peak is attributed to the clay (Figs. 3, 4). As this is the first time that lead white is detected in that early sample we intend to confirm its presence with SEM-EDS analyses.

Clay The colour of the clay of the selected samples varies from red, light red, reddish-yellow, yellowish-red, pink, reddish-brown, brown to dark brown. The clay is of two types: calcareous and non-calcareous (Fig. 2, Table 2). It is known that calcareous clays have been used since the Early Neolithic period and that in non-calcareous clays crystallization of Fe2O3 occurs which results in the presence of a red colour (Maniatis and Perdikatsis 1983, 339). Similarities and differences in the contents of

1

To the 45 samples in Table 1 two were added: one example of A3a ware, white on red (0180074) and one A3γ ware, red painted (0010099) from Tsangli.

5

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Knowing the limitations of pXRF the possibility of sulphur being a surface contamination in the case of selenite or gypsum slip cannot be considered as both the clay and the red pigment on top of it have been analyzed in all samples and no sulphur was found (Fig. 5). Furthermore, a similar XRF spectrum of slip has been received in Hellenistic ceramic figurines and has been confirmed with SEM-EDS (Asderaki-Tzoumerkioti et al. forthcoming). Calcareous-rich material is a very common white slip used in antiquity (Gumlia-Mair et al. 2010; Yiouni 2001) and has been identified in 22 samples (Table 2). It is not possible to confirm if the slips were added before or after firing. In some cases we assume they have been applied after firing as they are more vulnerable to environment while others are more resistant (Maniatis and Perdikatsis 1983).

decoration with red colour against that of white colour (Α3α ware) (Rondiri 2009). This variation may indicate the expression of the potters' particular preference in red colour and therefore different social identities, i.e. the conscious choice of red colour by potters in conjunction with conscious preference of the user/consumer; alternatively the common use of the red colour in micro-scale or macro-scale in a settlement or at community level which ascribes a different social meaning to the categories of painted pottery with red colour compared to those with white colour or even implies the easier access to the red colour against the white colour. Moreover, the macroscopic analysis of pottery samples and their incorporation into different fabric groups shows that there are common elements within the decorated Α3β ware among different areas or sites, so it is possible that potters used certain clays for this specific category within the same community but also outside it; therefore, there is a specific technique in manufacturing a particular product, here the painted Α3β ware. Within this frame it can be claimed that this similarity could imply a kind of affinity between members of one or more communities or even that the potter has a social restriction for the manufacture of such particular vessels. Furthermore, it might be possible that the proximity of different communities or the locality of their inhabitants is another reason for the specific choice. Another conclusion is that the use of the same ceramic material within some decorated Α3β fabric groups of the same area or site could indicate either a controlled restriction or a non-controlled access to raw materials, but also the social identity of the potters. It could even indicate the continuation of a tradition from one generation to another or from one family to another due to affinity or other relationship.

Red pigments The red colour varies from red to dark red, light red, reddish-yellow, yellowish-red and reddish-brown. Iron is present in all spectra as the main element and calcium as minor element. In some cases, there is a higher presence of potassium (K) than in the clay- a possible indication that it was used as a flux to aid the vitrification of the surface layer during firing and make it hard and durable (Yiouni 2001). All other elements detected are impurities linked to the basic raw materials. 38 pXRF spectra are compatible with an iron-rich pigment (Table 2), a pigment type which is the most common, abundant, widespread and easily available in antiquity (Domingo et al. 2012; Photos-Jones et al. 1997; Stratis et al. 2002; Brecoulaki et al. 2006; Yiouni 2001). Iron sulphide is detected in one sample (Fig. 6), while no cinnabar has been detected so far. However, cinnabar is present in the area and has been detected in figurines from Dimini dated to the Late Neolithic period (Topa and Skafida in press). The reddish brown colour which exists in one sample from Dasolofos could be attributed to ironrich manganese pigment due to the presence of manganese (Mn) which has not been detected in the clay matrix (Fig. 7).

Finally, the use of different clays within some of the decorated fabric groups in the same area or site could indicate the heterogeneity that exists within and outside the settlement as well as a possible competition in the access or use of some clays, or maybe a change or a transformation for social reasons in the production techniques of a specific category.

Conclusions and further work It has been claimed (Jones 2002, 73) that the use of analytical methods in the study of cultural heritage creates a distance between the social practices of the past and the material itself, by subjecting it to a series of transformations, instrumental or others. In order to be able to cover this distance we have to choose not only the kind of analysis to use but also how we are going to interpret the results we receive (Jones 2002, 68). Moreover, the more detailed an analytical method, the more we learn about the analyzed pottery. Also, one should choose which technical analysis is better to apply both in micro- scale and macro-scale.

At this preliminary stage of our study it is quite early to certify the above assumptions. Nevertheless some of the above conclusions can be documented with the examination of the samples using the pXRF technique which is becoming a prevalent technique for elemental analysis of materials. Being a non-invasive and nondestructive technique, it is an ideal analytical method for archaeological finds, very useful both to archaeologists and conservators of antiquities. Although it is not thought to be the most suitable technique for analysis of clay for provenance determination, as pXRF only reports what elements are present and does not give any information concerning the stoichiometry, several laboratories use it to analyze pottery (Shugar 2009; 2013). Being aware of the limitations of the technique, we would say that pXRF may well act as an ‘outrider’ of

The macroscopic analysis of the original sample of Α3β ware samples and its distribution in the Thessalian Neolithic sites has shown a broadening in the use of

6

V. Rondiri and E. Asderaki, The ‘management’ of painted and monochrome pottery of Neolithic Thessaly the sound scientific analysis to follow in order to achieve desired results.

does not exist in the clay. Could these differentiations be linked with a possible provenance of the red pigment? We still need to explore this possibility with the use of other techniques which will be able to give us more sound results.

In our case, the results obtained from our experimental survey in the small collection of the Early and Middle Neolithic Α3β ware enabled us to understand that different clays were used for the manufacture of the vessels and to identify the decoration materials which were used. Thus, we can now safely continue our research with the selection of specific samples for further analysis. The results of the pXRF analysis confirm in some cases the initial macroscopic examination of the sherds.

With this preliminary work, an attempt was made to give interpretations through the process of analytical data based on the painted pottery of the Early and Middle Neolithic period which for a long period were ‘in the dark’. In other words, its social dimension became obvious through a closer look at technology, while pottery itself as part of material culture acquired many different roles with the help of applied analytical methods. It is known that in the history of archaeological research in Greece, the study of Neolithic pottery in Thessaly has a past and it is certain that despite the existing difficulties it has a future as well.

The white slip is made of calcareous-rich material, gypsum or selenite and lead white. Calcareous-rich slip is detected at all settlements: at Rodia, Siaterli, Magoulitsa and Kamara only calcareous-rich slip is detected, while at Narthaki, Tsangli, Zerelia, Daoutza and Aidiniotiki gypsum or selenite is detected as well; moreover, lead white is detected in one sample at Dasolofos (Fig. 8). Some of the slips are very rich in potassium (K).

Having all the above in mind and having ‘good’ representative samples of ceramic material deriving from research performed during the 20th century in Thessaly, our future plans involve petrographic analysis of the clay in certain samples as well as SEM-EDS and XRD analyses in order to confirm the up to date results and get complementary information on the nature of the pigments and also see how the layers of the pigments are stratified on the pottery surface. Moreover, X-raying of the sherds has already begun and we hope that it will reveal details of the manufacturing techniques of the vessels (Berg 2008, 2009; Mirti 2000).

The red-brown colour is due mainly to an iron-rich pigment and in one case iron-manganese-rich pigment; also, in one case at Kamara iron sulphide has been detected (Table 2). Traces of nickel are present in two samples from Daoutza and in one sample from Magoulitsa; however, nickel cannot be attributed as a clay impurity as it has not been detected in the clay analysis of these certain samples. Also, copper is detected in one sample from Kamara and again here this element

References Arvanitopoulos A., 1907, Anaskafai kai erevnai en Thessalia, Praktika Archaiologikis Etaireias, 147182. Arvanitopoulos A., 1910, Anaskafai kai erevnai en Thessalia, Praktika Archaiologikis Etaireias, 168264. Asderaki-Tzoumerkioti E., Dionyssiou M., DoulgeriIntzesiloglou A. and Arachoviti P., forthcoming, Some new aspects on the materials used for the decoration of Hellenistic terracotta figurines in Pherai workshops, Greece, in Proc. Internat. Coroplastic Studies Conference (Cyprus, 3-5 June 2013). Berg I., 2008, Looking through pots: recent advances in ceramics X-radiography, Journal of Archaeological Science 35, 1177-1188. Berg I., 2009, X-Radiography of Knossian Bronze Age Vessels: Assessing our knowledge of primary forming techniques, Annual British School at Athens 104, 137-173. Bourdieu P., 1977, Outline of Theory and Practice, Cambridge. Brecoulaki H., Fiorin E. and Vigato P. A., 2006, The Funerary klinai of tomb 1 from Amphipolis and a sarcophagus from ancient Tragilos, eastern Macedonia: a physico-chemical investigation on the

painting materials, Journal of Cultural Heritage 7, 301-311. Demoule J.P., Gallis K. and Manolakakis L., 1988, Transition entre les cultures néolithiques de Sesklo et de Dimini: les catégories céramiques, Bulletin de Correspondance Hellénique 112, 1-58. Dobres M.A., 2000, Technology and social agency, Oxford. Domingo I., García-Borja P. and Roldán C., 2012, Identification, Processing and Use of Red Pigments (hematite and cinnabar) in the Valencian Early Neolithic Spain, Archaeometry 54 (5), 868–892. Gallis K., 1992, Atlas proïstorikon oikismon tis anatolikis thessalikis pediadas, Larisa. Giannopoulos E., 1902, Ekthesis archeologikis ekdromis ana ton dimo Skotoussis, Armonia Γ΄, 427-434. Gimbutas M., 1974, Achilleion: a Neolithic mound in Thessaly. Preliminary Report on the 1973/74 excavations, Journal of Field Archeology 1, 277303. Gumlia-Mair A., Albertson C., Boschian G., Giachi G., Iacomussi P., Pallecchi P., Rossi G., Shugar A.N. and Stock S., 2010, Surface characterization techniques in the study and conservation of art and archaeological artefacts: a review, Materials Technology 25 (5), 245-261.

7

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Halstead P., 1984. Strategies for Survival: An Ecological Approach to Social and Economic Change in the Early Farming Communities of Thessaly, N. Greece, Unpublished PhD thesis, University of Cambridge. Jones A., 2002, Archaeological theory and scientific practice, Cambridge. Karydas A.G., 2007, Application of a portable XRF spectrometer in the analysis of museum metal Collections, Annali di Chimica 97(7), 419–432. Kilian K., 1976, Der Siedlungshügel Bunar-Baschi bei Sikourion, in: Milojčić V. Von den Driesch Α. Enderle K. Milojčić-v. Zumbusch J. and Kilian K., Die Deutschen Ausgrabungen auf Magulen um Larissa in Thessalien, 1966, Beiträgezurur- und früh geschichtlichen Archäologie des MittelmeerKulturraumes 15, Bonn, 65-71. Kotsakis Κ., 1983, Keramiki technologia kai keramiki diaphoropoiisi: provlimata tis graptis keramikis tis Mesis Neolithikis Epochis tou Sesklou, PhD dissertation, Aristoteleio Panepistimio Thessalonikis. Latour B., 1999, Pandora’s hope: essays on the reality of science studies, Cambridge, Mass. Lemonnier P., 1993, Introduction, in P. Lemonnier (ed.), Technological choices: transformations in material culture since the Neolithic, London, 1-35. Malakassioti Z., Rondiri V., Vouzaxakis K. and Kalogianni A., 2012, To erevnitiko programma sti Magoula Zerelia, 3o Archeologiko Ergo Thessalias kai Stereas Elladas, Praktika epistimonikis sinantisis (Volos, March 2009). Maniatis Y. and Perdikatsis V., 1983, Technologiki kai proelefsiaki meleti tis kerameikis tou Sesklou, in K. Kotsakis, Keramiki technologia kai keramiki diaphoropoiisi: provlimata tis graptis keramikis tis Mesis Neolithikis Epochis tou Sesklou, PhD dissertation, Aristoteleio Panepistimio, Thessalonikis, 339-352. Mirti P., 2000, X-Ray Microanalysis Discloses the Secrets of Ancient Greek and Roman Potters, X-Ray Spectrometry 29, 63–72. Papadopoulou M., 1958, Magoulitsa: Neolithikos synoikismos para tin Karditsan, Thessalika 1, 3949. Photos-Jones E., Cottier A., Hall A. and Mendoni L., 1997, Kean Miltos: The Well-Known Iron Oxides of Antiquity, Annual British School Athens 92, 359371. Reinders H.R., Karimali L., Prummel W., Rondiri V., Tzevelekidi V. and Wijnen M., 2009, The Neolithic site of Kamára in the Sourpi plain (Thessaly, Greece), Pharos 15 (2007), 59-136.

Rondiri V., 2004, Erga Ethnikis Odou PATHE, Tmima Ag. Theodoroi-Almyros, Ch. Th. 271.300, Archeologiko Deltio 53 (1998) B1 Chronika, 427429. Rondiri V., 2005, Sourpi (ergo OTE) Ch.Th. 271.300Thesi Kamara. Archeologiko Deltio 54 (1999) Β1 Chronika, 403-404. Rondiri V., 2009, Thessaliki Neolithiki keramiki. Technologia kai katanomi sto choro, PhD dissertation, Aristoteleio Panepistimio Thessalonikis. Schneider G., Knoll H., Gallis K. and Demoule J.-P., 1991, Transition entre les cultures néolithiques de Sesklo et de Dimini: recherches minéralogiques, chimiques et technologiques sur les céramiques et les argiles, Bulletin de Correspondance Hellénique 115, 1-64. Schneider G., Knoll H., Gallis K. and Demoule J.-P., 1994, Production and circulation of Neolithic Thessalian pottery. Chemical and mineralogical analysis, in La Thessalie, Quinze années de recherches archéologiques, 1975-1990, Billanset perspectives, Actes du Colloque International, Lyon, 17-22 Avril 1990, vol. A, 61-70. Shennan, S., 1988, Quantifying Archaeology, Edinburgh. Shugar, A.N., 2009, Peaking your interest: an introductory explanation of how to interpret XRF data, WAAC Newsletter 31 (3), 8-10. Shugar, A.N., 2013, Portable X-ray Fluorescence and Archaeology: Limitations of the Instrument and Suggested Methods to Achieve Desired Results, Archaeological Chemistry VIII, 173-193. Stratis, I., Varella, E. A. and Vavelidis, M., 2002, Pigments from the Ochre Mines on Thasos Island, in M.A. Tiverios and D. Tsiafakis (eds.) Colour in Ancient Greece, The Role of Colour in Ancient Greek Art and Architecture 700-31 BC, Thessaloniki, 155-160. Theocharis D.R., 1964, Archeotites kai mnimeia Thessalias, Archeologiko Deltio 17, Β2 Chronika, 241-249 and 255-267. Topa Ch. and Skafida E., in press, Meleti zografikis epifaneias se keramika ostraka me epitheto chroma apo to Dimini, Archeologiko Deltio 64 (2009) / Chronika. Tsoundas Ch., 1908, Ai proïstorikai akropoleis Diminiou kai Sesklou, Athens. Wace A.J.B. and Thompson M.S., 1912, Prehistoric Thessaly, Cambridge. Yiouni P., 2001, Surface Treatment of Neolithic Vessels from Macedonia and Thrace, Annual British School at Athens 96, 1-25.

8

V. Rondiri and E. Asderaki, The ‘management’ of painted and monochrome pottery of Neolithic Thessaly

Table 1. The Neolithic pottery examined from Thessaly.

A/A

Site name

Site

Sample

Coarseness

1 1 3 1 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

Kamara Kamara Kamara Kamara Kamara Kamara Zerelia Zerelia Zerelia Zerelia Zerelia Aidiniotiki Aidiniotiki Aidiniotiki Dautza Dautza Dautza Dautza Tsangli Tsangli Tsangli Tsangli Tsangli Tsangli Tsangli Tsangli Tsangli Dasolofos Dasolofos Dasolofos Dasolofos Dasolofos Narthaki Narthaki Narthaki Siaterli Siaterli Rodia Rodia Rodia Rodia Rodia Rodia Magoulitsa Magoulitsa

1 1 1 1 1 1 2 2 2 2 2 3 3 3 4 4 4 4 5 5 5 5 5 5 5 5 5 6 6 6 6 6 7 7 7 8 8 9 9 9 9 9 9 10 10

0240022 0240023 0240024 0240198 0240128 0240129 0020014 0020015 0020018 0020084 0180045 0180048 0180048 0170033 0170034 0170107 0170109 0170112 0170113 0010059 0010098 0010100 TS 11/3 TS 11/4 TS 11/5 TS 11/6 TS 11/7 0140017 0140018 0140019 0140020 0140021 0150050 0150057 0150058 0160034 0160036 0220003 0220006 0220007 0220009 0220010 0220012 0050300 0050301

Medium Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Medium Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Medium Fine Fine Medium Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine Fine

9

Fabric group ΜΜΑΣ/4 ΛΜΑ/2 ΛΜΑΣ/2 ΛΜΑΣ/2 ΛΜΑ/2 ΛΑ/1 ΛΜΑ/2 ΛΑ/1 ΛΜΑΣ/2 ΛΜΑΣ/2 ΛΜΑ/2 ΛΜΑ/2 ΛΜΑ/2 ΛΑ/1 ΜΜΑ/3 ΛΜΑ/2 ΛΜΑΣ/2 ΛΜΑΣ/2 ΛΑ/1 ΛΜΑ/2 ΛΑ/1 ΛΑ/1 ΛΜΑ/1 ΛΑ/1 ΛΜΑΣ/2 ΜΜΑΖ/4 ΛΜΑΣ/2 ΛΜΑΣ/2 ΜΜΑΣ/4 ΛΜΑ/2 ΛΜΑ/2 ΛΜΑ/2 ΛΜΑΣ/2 ΛΜΑΣ/2 ΛΜΑ/2 ΛΜΑ/2 ΛΜΑΑ/2 ΛΜΑΣ/2 ΛΑ/1 ΛΜΑΣ/2 ΛΜΑΣ/2 -

Date MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN MN AN AN MN MN MN MN MN MN

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Table 2. The clay, slip and pigment characteristics.

A/A 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Area

Sherd ID

Narthaki 0150050 Narthaki 0150057 Narthaki 0150058 Kamara 0240022 Kamara 0240023 Kamara 0240024 Kamara 0240109 Kamara 0240128 Kamara 0240129 Tsangli TS 11/3 Tsangli TS 11/4 Tsangli TS 11/5 Tsangli TS 11/6 Tsangli TS 11/7 Tsangli 0010059 Tsangli 0010098 Tsangli 0010099 Tsangli 0010100 Zerelia 020014 Zerelia 020015 Zerelia 020018 Zerelia 020084 Aidiniotiki 0180045 Aidiniotiki 0180048 Aidiniotiki 0180050 Aidiniotiki 0180074 Daoutza 0170033 Daoutza 0170034 Daoutza 0170107 Daoutza 0170109 Daoutza 017011 Daoutza 0170113 Magoulitsa 0050300 Magoulitsa 0050301 Dasolofos 0140017 Dasolofos 0140018 Dasolofos 0140019 Dasolofos 0140020 Dasolofos 0140021 Rodia 0220003 Rodia 0220006 Rodia 0220007 Rodia 0220009 Rodia 02200010 Rodia 02200012 Siaterli 0160034 Siaterli 0160036

Sample ID 050 057 058 22 23 24 109 128 129 11-3 11-4 11-5 11-6 11-7 59 98 99 100 14 15 18 84 45 48 50 74 33 34 107 109 112 113 300 301 17 18 19 20 21 03 06 07 09 10 12 34 36

Clay

White & yellowish slip

Red colour

Non-calcareous Non-calcareous High K Non-calcareous Calcareous Non-calcareous Non-calcareous Calcareous Non-calcareous Calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous Calcareous Non-calcareous Calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous Calcareous Calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous Calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous Calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous Calcareous Calcareous Non-calcareous Non-calcareous Non-calcareous Non-calcareous

Calcareous rich Gypsum or Selenite Gypsum or Selenite Calcareous rich Calcareous rich Gypsum or Selenite Gypsum or Selenite Gypsum or Selenite Calcareous rich Calcareous rich Gypsum or Selenite Calcareous rich Gypsum or Selenite Calcareous rich Gypsum or Selenite Calcareous rich Gypsum or Selenite Calcareous rich Calcareous rich Calcareous rich Calcareous rich Gypsum or Selenite Calcareous rich Calcareous rich Calcareous rich Calcareous rich Calcareous rich Pb white Calcareous rich Calcareous rich Calcareous rich Calcareous rich Yellow ochre Calcareous rich

Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron sulphide? Iron-rich Iron-rich Same as clay Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich Iron-rich

10

V. Rondiri and E. Asderaki, The ‘management’ of painted and monochrome pottery of Neolithic Thessaly

Fig. 1. Map of Thessaly with the Neolithic sites in black whose pottery was analyzed both macroscopically and chemically (Rondiri 2009).

Fig. 2. XRF spectra showing calcareous versus noncalcareous clay, measured at 20kV.

Fig .3. Lead white is detected in sherd 0140019 from the Dasolofos area.

11

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

Fig. 4. XRF spectra of clay (dark spectrum) versus lead white (light spectrum) detected in Dasolofos 0140019, measured at 40kV.

Fig. 5. XRF spectrum of a gypsum slip measured at 20 kV in Tsangli sample TS 11/6. Sulphur is present in the pigment (black spectrum) while absent in the clay (grey-coloured spectrum).

Fig. 6. XRF spectra of possible iron sulphide slip (dark spectrum) in Kamara sample 0240024 and the clay (light spectrum).

Fig. 7. XRF spectra of iron-manganese-rich pigment (dark spectrum) and the clay (grey spectrum) in Dasolofos sample 0140024.

Fig. 8. Distribution of types of white slip in different settlements.

12

Chapter 2

Craftsmanship of big storage pithoi in the Early Helladic settlement of Helike, Achaea Dora Katsonopoulou1, Ioannis Iliopoulos2, Stella Katsarou3 and Vayia Xanthopoulou2 1

The Helike Project and The Helike Society, Athens Department of Geology, University of Patras, Patras, Greece 3 Ephorate of Palaeoanthropology-Spelaeology, Ministry of Culture, Athens, Greece 2

Corresponding author: [email protected]

Abstract The paper presents the first results of investigations on craftsmanship of big storage pithoi found in domestic contexts at the Early Helladic settlement of Helike, Achaea, in the NW Peloponnese, Greece. Beyond describing the conspicuously standardized morphological qualities in terms of profile, surface treatment and building techniques that pertain to these vessels, we have carried out detailed petrography, SEM and XRD analyses which suggest that regularized methods in their clay recipes and raw material distribution are equally distinguishable. The technological issue of low-fired conditions has been specifically investigated to meet the question of in-household construction of the pithoi. The overall assessment of the physicochemical results contributes to speculations about the jar builders constituting a different artisan group from ordinary potters. Ethnohistorical data of itinerant craftsmen from other Greek regions are also considered as a useful point of reference. Keywords: Helike, Early Helladic, storage jars, standardized fabric, specialization, itinerant pithos makers

cooking pots and a significant number of storage containers ranging from middle-sized jars to extra high, floor-based pithoi for long-term storage (Katsarou 2011).The distribution of the pithoi within the dwellings is crucial for exploring the communal and private level of surplus strategies in the settlement in association with the evidence on the emergence of local centralized authorities. In addition, their particular technological, functional and contextual properties, in comparison with the qualities featured by the smaller jars and other domestic containers, may imply the existence of a specialized group of potters for the manufacture of big size pithoi, practicing very particular methods of craftsmanship.

Introduction: the Early Helladic Settlement in Helike Excavations conducted by The Helike Project in the area of Rizomylos in Aigialeia revealed the remains of an Early Helladic II-III settlement extending over 1000 sq. m. at a distance of 1.1km from the present coastline (Katsonopoulou 2005; 2007; 2011) (Fig. 1). Following destruction by earthquake and fire the ruins, including their final occupation contexts in full, were sealed to the depth of 3-5m under lagoon and riverborne sediments generated by immediate geomorphological transformations of the area (AlvarezZarikian et al. 2008; Soter and Katsonopoulou 2011a, 2011b). The subsided town implies social complexity and elite authority on account of organized town planning, architectural differentiation and a high technological level of innovative building techniques (Fig. 2), as exemplified by the Corridor House, apsidal structures and multispaced rectilinear dwellings (Katsonopoulou 2011; Kormann et al. in press). The evidence on rich commodities for domestic crafts, weapons, as well as luxurious accessories in gold and silver inside the dwellings strengthens the picture for the existence of various levels of social and economic statuses within the local community (Earle 2002; Day et al. 2010).

Macroscopic Observations on the Pithoi To give an idea of the outline, the large storage pithoi from Helike exhibit standardized morphological features. Wall thickness ranges average from 2 to 4cm, rim diameter from 25 to 45cm, and their estimated height is usually exceeding 60cm to reach over 1.1m, implying that they constituted really large containers, as restoration of one pithos of the biggest size has shown (Fig. 3). They are formed with flaring necks attached on an ovoid-globular solid body standing on a wide flat base, and carry arched handles or non-perforated ledges (Fig. 4). Their most usual decorative feature is their horizontal roped or incised plastic bands along the neck-to-shoulder join, or placed in vertical, horseshoe or curved arrangement on

The pottery assemblages include a large array of complete monochrome and painted tableware from drinking and food-serving vessels, compounded with

13

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry the upper body, possibly demonstrating that individualized choices for ornamentation were employed. Their wide open rim enabling high accessibility and manipulation of the content suggests that the food material stored inside would be most possibly of a solid substance.

trace chemical clusters (Iliopoulos et al. 2011). Petrography has showed mudstone to be the predominant constituent of the Helike storage pithoi, and has assigned all examined fragments to the very same fabric, Group B, among the ceramic groups of Helike particularly characterized as ‘the bimodal calcite-chert-mudstone tempered fabric group’ (Fig. 7). Group B is mainly differentiated by Group A, characterized by the predominant presence of mudstone and radiolarian chert fragments (Iliopoulos et al. 2011), because of the greater particle size and the more calcareous character of the micromass. As indicated, common to frequent constituents include radiolarian chert, quartz and calcite, and bimodality is the typical feature of their grain size frequency distribution. Both groups of inclusions are distinctively angular to subangular. Clayslate and metaconglomerate fragments are also common, and together with mudstone and chert cover almost the entire size range from fine sand to pebble (0.125-8mm). The above data suggest that clayey material for the pithoi was exploited from the local area and was then further tempered mainly by chert and mudstone, the primary constituents found in the surrounding fluvial sands and crushed to serve for non-plastic temper grits.

Fragmentation of the pithoi very eloquently exhibits the hand-building technique applied by the EH potters at Helike (Fig. 5). They would primarily lay the base with the accumulation of several thick, sometimes even extremely thick, layers of coarse clay mass thrown in a ‘muddy’ shape. Evidence on the merged texture of the clay core testifies to the craftsmen building the pithos body walls by layering or overlapping thick flabby coils or easily shaped clay pieces (scored with more malleable clay), rather than solid and clear-joined coils or slabs. The flaring neck would be attached as a final piece on top of the shouldered body, and sometimes present detachment faces show that some preparation of the joining surfaces had occurred. Potters would secure their construction with minor additions of clay on the neck-to-shoulder join, the rim’s borderline or around the attached handles and applied bands. They would basically smooth the exterior surface, and rarely polish it or apply some slip or internal coating. The surfaces of the storage containers usually exhibit a light homogeneous colour throughout the cores, testifying to the prevalence of oxidized firing conditions (Molera et al. 1998). Colour variation from firing clouds is unusual on the surface, however cases of dark sandwich cores from incomplete firing may be present.

An additional diagnostic feature is the high calcareous character of its micromass which is optically active, pointing to firing temperature that is constantly kept at minimum levels, never exceeding 600°C. The mineralogical composition of the samples has been further evaluated through X-Ray diffractometry identifying pithoi fragments within the assemblage of quartz + calcite + illite + vermiculite ± kaolinite (Fig. 8). The presence of vermiculite (and kaolinite) in particular gives further support to the firing estimation inferred above on petrographic grounds (30 min 20 min 25 min

Olive-oil Olive-oil Water

Abrasive Naxos natural sand containing no corundum Naxos natural sand containing emery components + ground emery Ground corundum Drôme sand + 10% ground corundum Drôme sand +50% ground corundum

Fig. 9. Experimental protocol for polishing diasporite celts.

71

Chapter 10

Did copper actually arrive in the Aeolian Islands in the fourth millennium BC? The evidence from a small but iconic fragment of vitreous material thought to be copper slag Maria Clara Martinelli1, Effie Photos-Jones 2,3 and Sara Tiziana Levi 4,5 1

Museo Archeologico Luigi Bernabò Brea, Lipari (Messina), Italy Analytical Services for Art and Archaeology (Scotland) Ltd, Glasgow, UK 3 Archaeology, School of Humanities, University of Glasgow, Glasgow, UK 4 University of Modena e Reggio Emilia, Dept. Chemical and Geological Sciences, Modena, Italy 5 Hunter College, The City University of New York, New York, USA 2

Corresponding author: [email protected]

Abstract A singular fragment of ‘copper slag’ from Lipari thought until recently to be the key artefactual evidence for the arrival of the copper metallurgy in the Aeolian Islands, Sicily, in the 4 th millennium BC has now been reappraised, based on scientific analysis with the Scanning Electron Microscope (SEM-EDS). The results and comparative data sets point to the fragment being vitrified fuel ash (VFA). Preliminary experiments using local plant ash heated to high temperature in a laboratory furnace in the presence of SiO2-rich (but not necessarily quartz-rich) beach sand from the island of Stromboli, resulted in sintering but no fusion; this necessitates the closer scrutiny of the sourcing of the raw materials and heating conditions. However, it is suggested that the process is not associated with any attempt to melt or smelt copper. Keywords: early metallurgy, Diana-Spatarella, Italy, p-XRF, SEM-EDS

(Trento province) and Botteghino of Parma (Emilia Romagna) in Chassey levels (metal ‘drops’ in association with slag) (De Marinis and Pedrotti 1994; Pessina and Tinè 2008, 132); in Liguria, at Monte Loreto, there are indications that copper extraction began around 3500 cal. BC (Maggi 2002; Maggi and Pearce 2005). In Central Italy and on the Adriatic coast at Santa Maria in Selva e Fossacesia in Abruzzo, the presence of some copper fragments has been reported; in Southern Italy, in a late Serra d’Alto context of Mattinelle di Malvezzi, a copper object has been found in a burial inside a stone cist (Pessina and Tinè 2008, 134).

Introduction The first appearance of metals in Italy is ascribed to the Neolithic and is evident in contexts dated to the second half of the 5th millennium- first half of 4th millennium BC (Fig. 1). Metallurgy is thought to have been introduced to Northern Italy from the north as a result of the close connection and exchange network with the Munchshofen culture of Lower Bavaria and Austria (Pessina and Tinè 2008, 132). Objects (mainly awls) have been discovered in Northern and Central Italy in (a) Chassey levels in Northern Italy dated to the middle Neolithic, (b) VBQ (Vaso a Bocca Quadrata) levels, a middle Neolithic phase in northern Italy characterized by square-mouth pots, and (c) late Ripoli levels dated to middle-late Neolithic. Ripoli contexts occur in a large area of Central Italy and show close connections with other Neolithic cultures like those at Serra D’Alto (Southern Italy), Danilo (on the Dalmatian coast), Lagozza-Chassey (Northern Italy and France) and finally at Diana (Southern Italy). Ripoli pottery is fine and painted with red and brown geometric motifs organized in panels and, mainly in the latest phases, decorated with cordons, dots, circles and series of incised lines.

Evidence of metal manufacturing (slags and crucibles) is primarily concentrated in the metals-rich areas bordering the Tyrrhenian Sea: in Tuscany in late Ripoli (Final Neolithic) contexts at Neto-via Verga, Podere Pietrino, Orti Bottegone where there are crucibles and copper awls (Sarti and Volante 2002; Pessina and Tinè 2008, 134) and in Cagliari, Sardinia, in Ozieri contexts, at the sites of Anghelu Ruju, Cuccuru S’Arriu and Su Coddu di Selargius, some smelting slag of copper and silver has been recovered (Lo Schiavo 1989). Finally in the Aeolian Islands, at the Acropolis of Lipari (Figs. 2a, b) a single fragment of ‘slag’ has been found (Figs. 3a, b). This paper focuses on the technical examination of this find, currently on display in the Archaeological Museum at

In North Italy the archaeological traces of the earliest evidence of metal production are at the sites of Mezzocorona-Borgonuovo and Acquaviva di Besenello

73

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Lipari (MA no. 8993), and asks the question whether copper metallurgy did actually arrive on the Aeolian islands in the 4th millennium BC.

Following the publication of the results of the excavations at the Acropolis, this unique find became the critical evidence for arguing in favour of the introduction of copper metallurgy on the Aeolian Islands in the Early Chalcolithic, and furthermore for the insertion of Lipari within the network of commercial routes for the procurement of metals (Cocchi Genick 1994, 269) at the beginning of the Bronze Age. None of the other three settlements belonging to the same Diana-Spatarella phase have revealed evidence for metal objects nor indeed associated copper processing waste (Cavalier 1979; Cavalier 1981; Bernabò Brea and Cavalier 1995; Martinelli 2001).

In the Aeolian Islands, the Final Neolithic (4000-3400 BC) is represented by the Diana culture from a large settlement on the homonymous plain to the north of the town of Lipari; the Diana culture is characterized by a distinctive pottery made of a glossy red slip and with tubular handles (Bernabò Brea and Cavalier 1980). The same pottery has been found in Sicily and on the mainland and is thought to reflect the intensive exploitation of the Liparian obsidian sources across the Tyrrhenian Sea. The Diana-Spatarella represents the first phase of the Chalcolithic (3400-3000 BC). The pottery from this phase is made up of vessels with truncated cone surfaces of brown colour and decorated with rows of zigzag under the edge of the inner surface of the rim. There are currently four settlements in the archipelago known for their Diana-Spatarella phase. Two are on the island of Lipari and include (a) the settlement on the coastal hill of the Acropolis (Bernabò Brea and Cavalier 1980), and (b) the settlement at Monte-Spatarella, an inland plateau rising to c. 150-230m above sea level (Cavalier 1979; Martinelli 2001). The third settlement is on the island of Salina, at Serro Brigadiere, 50-55m above sea level (Bernabò Brea, Cavalier 1995). Finally, the fourth settlement is at San Vincenzo, on the island of Stromboli on the slopes of the volcano (Cavalier 1981; Levi et al. 2011, 2014).

Methodology This paper presents the first technical characterization of this small, but iconic fragment. The fragment measures c. 10cm (long axis) and is c. 2cm thick. It has a ‘top’ vitreous surface of pale green-blue colour with grey white accretions adhering to the surface (Fig. 3a). The opposite (‘bottom’) surface appears sintered and with a patches of ferruginous areas. It is likely that both ferruginous and non-ferruginous areas are of the same composition but not exposed to the same temperature and redox conditions. There is abundant evidence of white grains of ‘sand’, originally thought to be quartz but no quartz grains were observed in the sections analysed (see discussion below). Two sets of analyses were carried out: first, a non-destructive analysis with a portable X-Ray Fluorescence analyser (p-XRF) (Niton XL3t using the TestAllGeo calibration) was carried out at the Lipari Museum; second, invasive analysis on two small fragments (554.1 and 554.2), which were removed from the object at the spot indicated by the arrows in Figs. 3a, b, was carried out with the Scanning Electron Microscope with Energy Dispersive Analyser (SEM-EDS) in the SEM facility of Earth Sciences, University of Glasgow. We focus here on the results of the SEM-EDS analyses since they provide an insight into the structure of the fragment, rather than simply on its surface chemical composition (as analysed by the p-XRF). The two samples were mounted in metallographic resin, ground, polished and carbon coated prior to analysis.

The important archaeological excavations of the Acropolis of Lipari (Fig. 2b), a lava dome rising vertically above the sea, were conducted during the 1950s (Bernabò Brea and Cavalier 1980). The research uncovered about 10m of stratigraphy spanning the Neolithic, Eneolithic, Bronze Age, Greek and Roman periods, to the present. In the Acropolis, the DianaSpatarella phase was manifested in trench AP, cuts 10 and 11, at a depth of approximately 2m below present ground level and has been dated with C-14 on the basis of charcoal to 5000 ± 200 BP or 3050 BC (MASCA 40603640 BC, Bernabò Brea and Cavalier 1980, 839-844; Martinelli and Procelli 2011, 118). However, according to more recent investigations in the Archipelago and in Sicily, this single radiocarbon date, obtained from the 1952 excavation (sample R-180, Bernabò Brea and Cavalier 1989, 840-841), appears not fully reliable. Within these layers a blueish-green glassy fragment (Figs. 3a,b) was recovered and interpreted as ‘slag of copper’ (Bernabò Brea and Cavalier 1980, 337). In the excavation report there is no description of any associated structures or areas of heating. The authors simply mention ‘some shapeless items’ collected in the layer associated with cuts 10-11. Since today there is no evidence for any additional fragments of ‘shapeless items’ stored in the Lipari Museum’s collections, it is likely that MA no. 8993 from cut 10, published in Bernabò Brea and Cavalier (1980, 339), is the only relevant find which was interpreted as slag.

Results Images of the polished sections were taken at high magnification (see Figs. 4a, b). They reveal a highly vesicular structure and a fused matrix rather than one which cooled from the molten state. The observed porosity emanates from the burning of organic material. Following photography, three sets of analyses were undertaken: (a) area analyses over the entire surface of each sample, reflecting ‘bulk’ chemical composition; (b) spot analyses on the semi-fused matrix; and (c) spot analysis on individual grains. Area analyses (Table 1a) reveal a highly siliceous material consisting of aluminium silicates (SiO2/Al2O3 ratio c. 5) and about 10-15% of combined oxides (Na2O, CaO, K2O, and FeO). Spot analyses on the matrix revealed a similar composition.

74

M. C. Martinelli et al., Copper in the Aeolian Islands, Italy in the 4 th millennium BC? There are numerous inclusions which are predominantly iron rich, as shown in Table 1b, but also iron aluminium silicates and on one occasion barium sulphate. There was no evidence of pure quartz grains or indeed of any metallic inclusions within the sections analysed. Trace elements like Mn, Ba, V, Cr and S were below detection limits unless specified.

A mixture of Stromboli beach sand and poseidonia oceanica was subsequently heated in at 1000oC in a laboratory furnace for one hour, but there was no evidence of vitrification, only the production of a sintered mass forming at the bottom of a glass crucible (see Fig. 6) and resembling the ‘bottom’ surface of the Lipari VFA. We note that the high contents of titanium and iron in Stromboli beach sand and the high calcium content of the Stromboli plant ash do not correlate well with the low concentrations of the same elements in the Lipari VFA, although iron-rich unreacted mineral grains were indeed observed.

Discussion The results presented above preclude the identification of this fragment of vitreous material as metallurgical slag. Instead it is suggested that it resembles a vitrified fuel ash (VFA). VFA is a light, frothy, highly siliceous vesicular material with a predominantly glassy surface giving the appearance of having been fused and/or melted. There are many types of VFAs and they can derive from a number of very diverse processes, as for example from the burning of coal in the context of 19th century cast ironmaking furnaces; they can also originate from more ‘primitive’ technologies like the burning of kelp (seaweed) for the production of raw materials for glass and soap manufactories, as was the case in the 18 th century (Clow 1952).

In summary, based on the examination of 554.1 and 554.2, we can conclude that object MA 8993 in the Archaeological Museum on Lipari is most certainly not a copper smelting or melting slag, on account of the absence of non-ferrous metallic inclusions and/or the usual iron silicate phases normally associated with copper slag. Instead, it resembles a slag-like vitreous material or vitrified fuel ash (VFA). Further experimental work and under varying laboratory conditions is required to establish the nature of the raw materials that made up this object; a single simulation experiment carried out here for the purpose of checking the plant ash+beach sand hypothesis shows that, for example, no calcium-rich marine plants were used for the Lipari VFA. Therefore we suggest that this object is treated as a ‘unique’ fragment until such time as evidence for similar materials is forthcoming, either from the Aeolian Islands or beyond.

In the course of our research into this fragment we suggested that the Lipari VFA bears many similarities with VFAs found within Bronze Age cremation contexts in Northern Scotland (see Fig. 5b) (Photos-Jones et al. 2007). These materials recovered from within burial cists in association with small fragments of human bone are formed in the course of the cremation; they were recovered from the funerary pyre and buried within the cist. They were found to derive from the reaction of plant (usually seaweed) ash in the presence of (human) bone and a source of silica, normally the local quartz-rich sand. In comparing the Lipari VFA and the Scottish VFA, called in Orcadian dialect ‘cramp’ (Figs. 5a, b), we note the same extent of porosity and inclusion of unreacted mineral grains. What is completely absent from the Lipari VFA is the presence of bone.

On the evidence presented above, we can suggest that metallurgy did not arrive on the Aeolian Islands at this early date. It may be that self-sufficiency in the procurement of a ‘cutting edge’ via the exploitation and exchange of the valuable local obsidian source did not necessitate experimentation with or use of new materials like metals. We therefore conclude that early metallurgy in Italy appears to be concentrated in the metals-rich areas of the Central and North Tyrrhenian area and did not include the Aeolian Islands.

To test this hypothesis, seaweed (poseidonia oceanica) from the shores of the island of Stromboli was collected together with other marine plants, such as cystoseira, and crithmum maritimum, both native to the island. All three were burnt to ash and each was subsequently analysed with the p-XRF together with samples of Stromboli beach sand. The compositions (Table 2) reveal that the plants are rich in calcium and potassium, while the beach sand is rich in iron and titanium. Some trace elements in the plant ashes are interesting and perhaps merit further investigation in their potential association with local ‘pollution’, either man-made or derived from the volcano. For example, there are high contents of arsenic in the cystoseira and lead in the poseidonia, phosphorus in the crithnum and sulphur in all three plant species. Beach sands are iron and titanium-rich reflecting the composition of some of the inclusions analysed in the Lipari VFA.

Acknowledgements The authors are grateful to Madeleine Cavalier and Umberto Spigo for their continuous support in the Museum of Lipari. The authors are also grateful to Leandro Lopes for the collection and subsequent burning of seaweed and other marine plants in preparation for the laboratory-based experiments.

75

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

volcanological investigation at Stromboli, Aeolian Islands, Italy, Antiquity 88, issue 342. Lo Schiavo F., 1989, Le origini della metallurgia e il problema della metallurgia nella cultura di Ozieri, in L. Dettori Campus (ed.) La cultura di Ozieri. Problematiche e nuove acquisizioni, Ozieri, 279293. Maggi R., 2002, Pastori, miniere, metallurgia nella transizione fra Neolitico ed età del Rame: nuovi dati dalla Liguria, in in A. Ferrari A. and P. Visentini (eds.) Il declino del mondo Neolitico. Ricerche in Italia centro settentrionale fra aspetti peninsulari, occidentali e nord alpini (Atti convegno Pordenone 2001), 437- 440. Maggi R. and Pearce M., 2005, Mid-fourth millennium copper mining, north-west Italy: the earliest known mines in Western Europe, Antiquity 79, n. 303, 6677. Martinelli M.C., 2001, Un’altra capanna nella località Spatarella a Lipari, in M.C. Martinelli and U. Spigo (eds.) Studi in onore di Luigi Bernabò Brea, Palermo, 89-112. Martinelli M.C. and Procelli E., 2011, L’età del Rame in Sicilia: dalla facies ceramica alla facies archeologica, una strada difficile, in Atti XLIII Riunione Scientifica IIPP “L’età del rame in Italia” (Università di Bologna 26-29 Novembre 2008), Firenze, 113-120. Pessina A. and Tiné V., 2008, Archeologia del neolitico. L'Italia tra VI e IV millennio a.C., Roma: Carocci. Photos-Jones E., Ballin-Smith B., Hall A.J. and Jones R.E., 2007, On the intent to make cramp: an interpretation of vitreous seaweed cremation ‘waste’ from prehistoric burial sites in Orkney, Scotland, Oxford Journal of Archaeology 26(1), 1-23. Sarti L. and Volante N., 2002, Neto – via Verga – (Firenze): le produzioni del Neolitico tardo e finale e del passaggio all’Eneolitico, in A. Ferrari A. and P. Visentini (eds.) Il declino del mondo Neolitico. Ricerche in Italia centro settentrionale fra aspetti peninsulari, occidentali e nord alpini (Atti convegno Pordenone 2001), 441-446.

References Bernabò Brea L., 1988, L’età del rame nell’Italia peninsulare: la Sicilia e le Isole Eolie, Rassegna di Archeologia 7, 469-506. Bernabò Brea L. and Cavalier M., 1980, L’Acropoli di Lipari nella preistoria, Meligunìs Lípara IV, Flaccovio, Palermo. Bernabò Brea L. and Cavalier M., 1995, Salina, ricerche archeologiche (1989-1993), Meligunìs Lípara VIII, Accademia di Scienze, Lettere e Arti di Palermo. Cavalier M., 1979, Ricerche preistoriche nell’arcipelago Eoliano, Rivista di Scienze Preistoriche XXXIV, 1-2, 45-136. Cavalier M., 1981, Villaggio preistorico di San Vincenzo, Sicilia Archeologica 46, 27-54. Cipolloni Sampò M., 1993, Il Neolitico nell'Italia meridionale e in Sicilia, in A. Guidi and M. Piperno (eds.), Italia preistorica, Laterza, Bari, 334-365. Clow A. and Clow N.L., 1952, The chemical revolution: a contribution to social technology, London: Batchworth Press. Cocchi Genick D., 1994, Manuale di Preistoria. Neolitico II, Viareggio. De Marinis R.C. and Pedrotti A. L., 1994, L’età del Rame nel Versante italiano delle Alpi centro-Occidentali, in Atti XXXI Riunione Scientifica IIPP ”La Valle d’ Aosta nel quadro della preistoria e protostoria dell’arcoalpino centro-occidentale”, Firenze, 247299. Fiorentino G., Caracuta V., Martinelli M.C., 2012, Plant remains and AMS: a new tool for dating palaeoclimate change in the Aeolian Islands (N-E Sicily) during the II millennium BC, Radiocarbon 54, 689-700. Levi S.T., Bettelli M., Di Renzoni A., Ferranti F. and Martinelli M.C., 2011, 3500 anni fa sotto il vulcano. La ripresa delle indagini nel villaggio protostorico di San Vincenzo a Stromboli, Rivista di Scienze Preistoriche LXI, 157-172. Levi S.T., Ayala G., Bettelli M., Brunelli D., Cannavò V., Di Renzoni A., Ferranti F., Lugli S., Martinelli M.C., Mercuri A.M., Photos-Jones E., Renzulli A., Santi P. and Speranza, F., 2014, Archaeological and

76

M. C. Martinelli et al., Copper in the Aeolian Islands, Italy in the 4 th millennium BC?

1 Lipari Acropolis 2 Pizzica Pantanello 3 Fossacesia 4 Su Cuddu 5 Cuccurru S’Arriu 6 Anghero Rujo 7 Neto 8 Orti Bottegone 9 Podere Pietrino 10 Fontenoce 11 Santa Maria in Selva 12 Attiggio 13 Travo 14 Botteghino 15 Alba 16 Arene Candide 17 Rivoli-Rocca 18 Isera 19 Bannia 20 Palù di Livenza

Fig. 1. Distribution of main sites across Italy with evidence of early copper artefacts and metallurgical waste (adapted by A. Di Renzoni from Pessina and Tinè 2008, fig. 14).

Fig. 2a. Location map of the Aeolian Islands.

Fig. 2b. The Acropolis of Lipari, Lipari (ME).

77

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

Figs. 3a and 3b. Fragment referred to as ‘slag of copper’ (MA no. 8993), consisting of a ‘top’ vitreous surface of pale green colour with light grey surface accretions and a ‘bottom’ sintered surface with patches of ferruginous areas. Two small fragments (about 1.5 cm) were removed for analysis (see arrow), incorporating sections of the sintered surface, but not the ferruginous area.

Fig. 4a. Scanning electron microscope image of sample 554.1. Bar = 100 microns.

Fig. 4b. BS-SEM-EDAX image of sample 554.2 showing spot and area analyses.

Fig. 5a. SEM-EDAX image of sample 554.1 showing ferruginous inclusions – white – in a highly vesicular matrix.

Fig. 5b. SEM-EDAX image of a sample of Scottish VFA showing high degree of porosity (round black holes) and unreacted grains of quartz (dark grey and fractured) (after PhotosJones et al. 2007, fig. 9).

78

M. C. Martinelli et al., Copper in the Aeolian Islands, Italy in the 4 th millennium BC? Table 1a. Area analyses – mean of three areas per sample – giving the composition expressed as element oxides in weight %. Bdl is below detection limit. Areas analysed vary in size. See, for example, Fig. 4b for spectrum 25. spectrum number 6 14 15 mean 2 3 5 25 mean

sample 554.1

554.2

SiO2

Al2O3

FeO

K2O

MgO

Na2O

TiO2

CaO

Mn

Ba

S

V

Sb

Cr

75.3 75.7 74.6 75.2 74.9 75.1 75.1 58 75

14 14 14.5 14.2 14 13.8 13.6 19.7 13.8

2.3 2.1 2.3 2.2 2.1 2.1 2.2 9.5 2.1

5.4 5.4 5.3 5.4 6 6.5 6 4.9 6.2

bdl bdl bdl bdl bdl bdl bdl 1 bdl

4.5 5.6 bdl 5 4.3 3.1 5.8 5.4 4.4

bdl bdl bdl bdl bdl bdl bdl 1.2 bdl

1.8 1.8 bdl 1.8 1.4 1.5 1.5 3.5 1.5

bdl bdl bdl bdl bdl bdl bdl bdl bdl

bdl bdl bdl bdl bdl bdl bdl bdl bdl

bdl bdl bdl bdl bdl bdl bdl bdl bdl

bdl bdl bdl bdl bdl bdl bdl bdl bdl

bdl bdl bdl bdl bdl bdl bdl bdl bdl

bdl bdl bdl bdl bdl bdl bdl bdl bdl

Table 1b. Spot analyses on iron-rich, barium-rich and other inclusions; weight % element oxides.

Bdl below detection limit. spectrum number 7 8 10 12 16 17 27 28

sample

554.1

554.2

SiO2

Al2O3

FeO

K2O

MgO

Na2O

TiO2

CaO

Mn

Ba

S

V

Sb

Cr

0.6 1.9 33.4 0 2.4 1.1 53.9 1.3

0 0 8.1 4.2 7.6 7.6 2.3 2.8

76.4 0 48.4 81.2 72.9 72.7 10.7 94.7

bdl bdl 2.2 0 0.1 bdl bdl 0.2

5 bdl bdl 2 bdl 6.6 15.3 2.7

bdl bdl 4.9 bdl bdl bdl bdl bdl

8.5 bdl 4.8 8.5 6.5 7.5 0.5 9

bdl bdl bdl bdl 0.3 0.3 22.1 bdl

bdl bdl bdl bdl bdl 0.6 0.3 bdl

bdl 60.8 bdl bdl bdl bdl bdl bdl

bdl 14.5 bdl bdl bdl bdl bdl bdl

bdl bdl bdl bdl bdl bdl bdl 1

bdl bdl bdl bdl bdl bdl 1.1 bdl

bdl bdl bdl bdl bdl bdl bdl bdl

Table 2. p-XRF analyses of major oxides of plant ash deriving from the three marine plant species and also Stromboli beach sand. Major oxides are in %; minor and trace elements are in ppm.

Cystoseira Poseidonia Crithnum Beach ‘sand’

Fig. 6. Sintered mass produced in the laboratory from the reaction of marine plant ash and Stromboli beach sand.

Cystoseira Poseidonia Crithnum Beach ‘sand’ Cystoseira Poseidonia Crithnum Beach ‘sand'

SiO2 7.4 9.2 7.7 37 Sr 36340 6439 8647 393 Cu 26 87 86 52

79

CaO 16.3 14.3 25.2 9.3 Rb 125 71 86 20 Cr bdl bdl 85 281

K2O 10.8 7.9 10 1.7 Pb 17 73 bdl bdl V 140 198 137 394

Al2O3 4.3 6.2 5 14.7 As 1249 29 bdl 9 Ti 617 1090 1081 2970

Zn 5430 198 206 68 Mn 86 360 89 1907

Bi 122 57 14 11 S 45533 33258 bdl

P 3885 5076 20484 bdl Ba 230 bdl bdl 823

Chapter 11

Selective use of arsenical copper during the Mycenaean period: the evidence from Pylia, in Messenia, Greece Charilaos E. Tselios1, Eleni Filippaki2 and Georgios S. Korres3 1

Hellenic Ministry of Culture, Education and Religious Affairs, Directorate for the Administration of the National Archive of Monuments, Documentation & Protection of Cultural Property Agion Asomaton 11, 105 53 Athens, Greece 2 N.C.S.R. “Demokritos”, Institute for Advanced Materials, Physicochemical Processes, Nanotechnology & Microsystems, Laboratory of Plasma Physics, Agia Paraskevi Attikis, 153 10 Athens, Greece 3 National & Kapodistrian University of Athens, Faculty of Philosophy, Dept. of History and Archaeology, University Campus, Athens 157 84, Greece Corresponding author: [email protected] Abstract During the Late Bonze Age tin bronze became the prominent alloy for the manufacture of weapons, vessels, tools and implements in the Greek mainland, as well as in the rest of the Aegean. Apart from a small number of objects dated to the transitional phase from Middle Helladic I, the vast majority of bronze objects that have been examined so far are characterized as tin bronzes. A small number of tin bronzes also contain arsenic as a trace element, usually attributed to the ores or as minor element coming from the recycling of arsenical copper objects of previous periods. Despite this general picture, according to the analytical examination of two bronze artifacts from Pylia in southwestern Peloponnese, arsenical copper was still in use well within the Mycenaean period. In the course of a research project dealing with the study of the Mycenaean metalwork in Pylia, minor samples taken from a battle knife of unknown context from Metaxada, dated to the Late Helladic I or II period and from a razor uncovered in the chamber tomb cemetery at Volimidia, dated to Late Helladic IIIA, were examined under the Optical Microscope and SEM (at N.C.S.R. “Demokritos”). According to the analytical data, the battle-knife consists of a rare Cu-As-Ag alloy, while the tin bronze razor is plated with a thin sheet of arsenical bronze. In both cases arsenical bronze was used well within the chronological borders of the Mycenaean period for aesthetic purposes. In the case of the battle knife, the Mycenaean bronze smith managed, through cycles of hammering and annealing, to take advantage of surface enrichment phenomena in order to manufacture a silver-like of the original tin bronze artifact with a sheet of arsenical copper no more than 10μm thick. Since the two artifacts are dated to different phases of the Mycenaean era, it seems likely that the use of arsenical copper was not restricted to earlier periods and that the Mycenaean metalworkers had never forgotten the properties of arsenical copper and utilized them for limited applications. Moreover, the technical characteristics of the two artifacts reveal the high level of expertise of the Mycenaean bronze smiths who were capable of applying complex and in some cases, sophisticated metalworking practices.

Keywords: Mycenaean, arsenical copper, metalwork, metallography, Electron Microscopy

the rest of the Aegean islands at sites like Kastri on Syros (Bossert 1967, 75-76; Renfrew 1967) and the Greek mainland at Manika on Euboea (ΜcGeehan-Lyritzis 1988). During the same period strong cultural influences from the Northeastern Aegean made their appearance in Southern Greece (Sotirakopoulou 1997).

Introduction Arsenical copper, the natural or artificial alloy (Charles 1967; Budd & Ottaway 1991; Rapp 1988, 1999; Lechtman 1996; Papadimitriou 2001) consisting of Cu and As, in most cases up to 3-5%, had been the prominent copper alloy used for the manufacture of weapons, tools and implements in the Early Bronze Age Aegean, from around 3000 BC to 2100/1900 BC (Papadimitriou 2001; Mangou and Ioannou 1997, 1998, 1999). Arsenical copper remained in use even when tin bronzes came in circulation, during the early phases of the Early Bronze Age (3000-2500 BC) in the Northeastern Aegean, at Troy (Schmidt 1902, 226, 229; Bittel 1959), Poliochni on Lemnos (Pernicka et al. 1990) and Thermi on Lesbos (Lamb 1936, 214-215). During the middle to late phases of the same period, around 2200 to 1900 BC (Late Bronze Age II to III), tin bronze was introduced in

Apart from arsenical copper or tin bronze, another type of alloy containing both arsenic and was also in use during the same period in the Aegean, Asia Minor and Cyprus, though in lesser degree (Muhly 1985, 127-128; Gale et al. 1985, 148, tab. 1, 2; Pernicka et al. 1990; Stos-Gale et al. 1996, 52; Tselios 2008, 88). Cu-As-Sn alloys (with lower Sn content than As), according to some scholars, represent a transitional technology from arsenical copper to tin bronzes (Muhly 1985), while according to others Cu-As-Sn alloys were mere products of recycling

81

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry (Balthazar 1990, 73). In any case, Cu-As-Sn alloys survived until the early Late Bronze Age (16th c. BC) and in some occasions, until the Late Bronze Age II (15th c. BC) (Tselios 2013).

Analytical results and discussion The analytical investigation of both artifacts involved metallographic examination followed by observation and microanalyses under the SEM, according to the licence granted by the Hellenic Ministry of Culture and Sports. The analytical procedure took place at the Laboratory of Archaeometry of the N.C.S.R. “Demokritos” in Athens, under the supervision of Dr. Yannis Bassiakos. Minute samples measuring no more than 2 mm x 1.5 mm were removed from the artifacts with the use of a small saw at the Archaeological Museum in Chora. After mounting in polyester resin the samples were initially observed under a binocular microscope and then examined under a Leica DM polarizing microscope with parallel and vertical Nicol’s at magnifications from 25X to 200X, after etching with alcoholic FeCl3 in various times, according to the reaction of each metallic surface to the solution (Scott 1991, 8-10). The metallographic examination involved extensive observation of the microstructure and documentation with photomicrographs of high resolution. Examination of the samples at high magnifications (up to 50,000X) and elemental analyses were conducted under an e-spect FEI (Philips) Scanning Electron Microscope equipped with an EDAX PV 7760 microanalytical system.

During the Middle Bronze Age (20th-17th c. BC) arsenical copper remained in use in equal terms with tin bronze (Papadimitriou 2001), while from the late phases of the period and the early phases of the Late Bronze Age (17th 16th c. BC) onwards tin bronze replaced arsenical copper (Varoufakis 1973; Craddock 1976 Mangou and Ioannou 1999; Papadimitriou 2001). Nevertheless, a battle knife (Schlachtmesser) from Metaxada and a razor from Volimidia in Pylia, Southwestern Peloponnese (Fig. 2), bear clear evidence that arsenical copper had not been abandoned during the Mycenaean era, but remained in use until the Late Helladic IIIA sub period (14th c. BC). This paper focuses on the analytical investigation of these two objects. The 35 cm long and 2.3 cm wide battle knife CM 605, according to the Chora Archaeological Museum inventory catalogue, is a stray find from the area of Metaxada in Southwest Messenia (Figs. 2 and 4). Battle knives of this type with flanged handle, slightly curved back and straight edge are common between the funerary gifts of the Shaft Graves at Mycenae (Karo 1930, 103104, 162; Mylonas 1973, 322-324) and of the early tholos and chamber tombs in Messenia and the rest of the Peloponnese. So far, battle knives have been discovered in Vagenas tholos tomb (or grave circle) V at Ano Englianos (Blegen and Rawson 1973, 159, 160-161), in tholos tomb 2 at Routsi (Marinatos 1956) (Fig. 2), in chamber tomb 12 at Dendra (Verdelis 1977) and in chamber tomb I at Prosymna (Blegen 1931, 328-354), in most cases accompanying type A swords. Therefore, the battle knife CM 605, that may have resulted from a rich tholos or chamber tomb, can be securely dated to the LH I or to the early phases of the LH II, thus to the 16th or to the early 15th c. BC.

The battle knife consists of a rare Cu-As-Ag alloy. The As content ranges between 3.5% in the solid phases and 3% in the mineralized surface, resulting to a typical arsenical copper object (Caley 1949). On the other hand, the Ag content of the alloy ranges between 4-5% in the metallic core, gradually increasing in volume towards the surface, where it reaches 40%. This phenomenon is documented with EDS bulk analyses (in 5 consecutive frames) from surface to core (Figs. 1 and 5; Table 1) and semi-quantitative line scanning, performed under the SEM (Figs. 6 and 7). Both techniques have revealed high Ag concentrations on the surface of the battle knife and moderate to small Ag content in the core. The silver enriched surface was further revealed by the etching solution used during the metallographic examination (alcoholic FeCl3, 4x5sec). A shiny white silver-rich zone, penetrating no more than a few μm in the metallic core, was observed under the SEM and the polarizing microscope (Figs. 5 and 8). Underneath the silverenriched surface, well-shaped, polygonal twined crystallites of the α Cu-As-Ag phase (measuring 60 μm x 40 μm on average) under the overlaying cored structure are present, along with scattered Ag insoluble phases. Consequently, the mechanical treatment of the battle knife involved repeated circles of cold hammering and annealing at temperatures around 400-500oC and efficient time. The metallographic examination revealed the mechanical treatment of the body of the object, as the sample was removed from the middle of the back, while the edge must have been finally cold hammered. As weapons made of Cu-As alloys with 3% As content are resistant enough to be used in battle (Northover 1989; Scott 1991) the battle knife from Metaxada must have been a functional weapon and not a ceremonial one.

The 18.8 cm long and 4.5 cm wide ‘razor’ MC 578 from the chamber tomb Mastoraki 1 at Volimidia, near Chora in Southwestern Peloponnese (Marinatos 1953) (Figs. 2 and 3) seems to be another interesting occasion of use of arsenical copper well within the chronological borders of the Mycenaean era. The razor was discovered in the 1950s by Spyridon Marinatos, according to the inventory catalogue of the Chora Museum, inside the pit no. 2 of the chamber tomb, along with pottery of the LH IIIA1 (1st half of the 14th c. BC) (Kountouri 2002, 16-17). The shape with three horizontally placed nails at the edge of the butt represents a widespread weapon type that was in use from LH I to LH III (Weber 1983, 139), even though the particular sub type was mainly in circulation during the LH II and III, according to similar finds from tholos ii at Routsi (Marinatos 1956). Moreover, as, according to the pottery, the tomb remained in use from the late phases of LH II until LH IIIA1, the razor can be securely dated in the late 15th or in the early 14th c. BC.

82

Ch. Tselios et al., Selective use of arsenical copper during the Mycenaean period The silver-enriched surface can only be attributed to intentional depletion gilding through repeated circles of hammering and annealing (confirmed by the metallographic examination) that prompted the Ag present in the alloy from the core to the surface. The increasing Ag content towards the surface also points to the same direction. The depletion gilding technique had been used for the manufacture of daggers as early as the 3rd millennium at Arslantepe in central Asia Minor (Palmieri et al. 1998; Hauptmann and Palmieri 2000, 77-78), while daggers consisting of Cu-Ag alloy from EM Crete (Branigan 1968, 221; Tselios 2008) may have been formed with the same technique. In later times depletion gilding was used for the production of silver plated coins mainly during the Late Roman period (Ingo et al. 2007, 11).

since pure Ag is very malleable and fragile, especially in the case of hammering a cast plate down to thin foil (Demakopoulou et al. 1995), a small quantity of Cu could have been added to Ag in order to facilitate hammering. The question why this silver-like weapon had to be silver plated before inhumation remains open and will be discussed elsewhere. The razor CM 578 consists of tin bronze with ambiguous Sn content, as both the surface layers and the metallic core are nowadays totally mineralized. According to the analyses of similar artifacts from Papoulia tholos 1 (razor CM 607) and Volimidia Koroniou chamber tomb (razor CM 602) weapons of this type consist of tin bronze with Sn content around 8-10% (Table 3) (Tselios 2013). The surface of the weapon had been covered with a thin sheet of metal, probably just before inhumation, small fragments of which are visible only under the polarizing microscope (Fig. 10). Unfortunately, all the metallic finds from Volimidia had conservation treatment in the 1950s or 1960s without the use of a binocular microscope and solely with plain mechanical and chemical means. Since no conservation records were kept we cannot be certain whether any other traces of plating were still visible right after the excavation. The 10 μm thick sheet (about the thickness of a kitchen aluminum foil) consists of arsenical copper with As content around 5% (Table 4). As in the case of the silver-plated battle knife from Metaxada, the purpose of plating was probably the creation of a silver-like surface, solely for funerary purpose. If the razor was initially a non-utilitarian funerary gift, the manufacturer would not have to spend precious tin, but would use pure copper instead. Therefore it is quite obvious that the razor was utilitarian during the life of its owner and that plating took place just before its deposition in the grave. The selection of arsenical copper instead of pure or alloyed silver, that must have been more expensive, can only be explained either by the economic status of the family of the deceased owner of the razor or by the possible intention of the metalworker to deceive the mourning family. In any case, the local Mycenaean bronze smith who plated the razor was well aware of the colour properties of arsenical copper.

Surprisingly, in the case of the battle knife from Metaxada the Mycenaean bronze smith who hammered this magnificent weapon decided to use a Cu-As-Ag alloy, even though weapons and tools were manufactured with tin bronze during his time. Moreover, all the up to date examined contemporary battle knives and other weapons from Messenia and elsewhere consist of tin bronze (Mangou and Ioannou 1999; Kayafa 1999, 411412; Tselios 2013, 153, 374-380). In the case of this weapon, unique in terms of composition, the use of arsenical copper along with silver resulted to a silver enriched surface. As the use of tin bronze with considerable Sn content would have resulted to a golden like appearance of the object (Gillis 1999; Varoufakis 1991; Asderaki et al. 2006; Soles 2008, 154) the Mycenaean bronze smith used arsenical copper instead of tin bronze in order to hammer a utilitarian weapon with silver-like appearance and excellent mechanical properties. Moreover, with the addition of Ag he managed to take advantage of both the surface enrichment in As after casting and in Ag through circles of hammering and annealing. The selection of arsenical copper to be alloyed with silver instead of pure copper, as in the cases of the Arslantepe and the Early Minoan daggers, clearly demonstrates the progress achieved from the 3rd millennium onwards in the field of metalwork practices. The examples from Arslantepe and Crete must have been non-functional weapons according to the use of pure copper alloyed with silver for their manufacture. This malleable alloy could only have resulted in items for demonstration or ceremony or as funerary gifts. On the other hand, the Mycenaean bronze smith took a step forward forging a utilitarian weapon of arsenical copper with the addition of silver in the alloy.

Concluding remarks Plating of funerary gifts with metal or alloy other than silver was a common practice during the mature Mycenaean period (14th to 12th c. BC). Tin-plated clay vessels are quite common in Mycenaean tombs from the LH IIIA (14th c. BC) onwards and it has been clearly demonstrated that the purpose of plating vessels with thin tin foil was to create silver-like vases, exclusively for funerary use (Gillis 1999). Despite the fact that the practice of silver or false silver plating for funerary purpose is not yet attested for bronzes, there seem to be some references that point to the opposite direction. A type A sword from Vagenas tholos (or grave circle) V is mentioned to bear traces of silver on its blade (Blegen and Rawson 1973) while a bronze conical cup in the National Archaeological Museum in Athens (NAM 4546)

Strangely enough, despite the fact that the surface of the battle knife from Metaxada already had a silver like appearance, the artifact was plated with a thin sheet of almost pure Ag (Table 2), probably slightly prior to its inhumation in the tomb. Traces of the Ag metallic layer were observed in the form of small fragments around the surface, under the SEM (Fig. 9). Cu present around 3% in Ag could be considered an impurity. On the other hand,

83

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry from the tholos tomb of ‘Daemons’ at Mycenae was definitely tin plated, according to laboratory examination (Demakopoulou 2000).

Demakopoulou K.E., Mangou H., Jones R.E. and PhotosJones E., 1995, Mycenaean Black Inlaid Metal ware in the National Museum, Athens: A Technical Examination, BSA 90, 137–154. Demakopoulou K., 2000, Ethiniko Archailogiko Mouseio, Proistoriki Syllogi, AD B1 (1995), 2. Gale N.H., Stos-Gale Z.A. and Gilmore G.R., 1985, Alloy Types and Copper Sources of Anatolian Copper Alloy Artifacts, Anat. Studies 35, 143-173, Gillis C., 1999, Color Symbolism and Metallurgy, in P.P. Betancourt, V. Karageorghis, R. Laffineur and W-D. Niemeyer (eds.) Meletemata III: Studies in Aegean Archaeology presented to Malcolm H. Wiener as he enters his 65th year, Aegaeum 20, Liège-Austin, 289-298. Hauptmann A. and Palmieri A., 2000, Metal Production in the Eastern Mediterranean at the Transition of the 4th/3rd Millennium: Case Studies from Arslantepe, Anatolian Metal I, Der Anschnitt Beiheft 13, Bochum, 75-82. Ingo G.M., Padeletti G., De Caro T., Riccucci C., Guida G., Angelini E. and Grassini S., 2007, Microchemical Investigation of Ancient Silver and Gold Plated Objects: Coating Techniques and Degradation Mechanisms, in V. Argyropoulos, A. Hein and M. Abdel-Harith (eds.) Strategies for Saving Our Cultural Heritage, Papers presented at the International Conference on Strategies for Saving Indoor Metallic Collections with a Satellite Meeting on Legal Issues in the Conservation of Cultural Heritage (Cairo 25 February-1 March 2007), 9-13. Karo G., 1930, Die Schachtgräber von Mykenai, München. Kayafa M., 1999 Bronze Age Metallurgy in Peloponnese, Greece, Unpublished PhD Thesis, University of Birmingham. Κοuntour, Ε., 2002, I Ysteroelladiki IIIA Keramiki apo to Nekrotafeio ton Volimidion Choras kai i Sygchroni Keramiki Paragogi tis Messinias, Unpublished PhD Thesis, Αthens University. Lamb W., 1936, Excavations at Thermi in Lesbos, Cambridge. Lechtman H., 1996, Arsenic Bronze: Dirty Bronze or a Chosen Alloy? A View from the Americas, JFA 23 (4), 477-514. Mangou Η. and Ioannou P.V., 1997, On the Chemical Composition of Prehistoric Greek Copper-Based Artifacts from the Aegean Region, BSA 92, 59-72. Mangou H. and Ioannou P.V, 1998, Οn the Chemical Composition of Prehistoric Greek Copper-Based Artifacts from Crete, BSA 93, 91-102. Mangou Η. and Ioannou P.V., 1999, On the Chemical Composition of Prehistoric Greek Copper-Based Artifacts from Mainland Greece, BSA 94, 81-100. Marinatos S., 1953, Anaskafai en Pyloi, PAE, 238-250. Marinatos S., 1956, Anaskafai en Pyloi, PAE, 202-203. ΜcGeehan-Lyritzis V., 1988, The Early Bronze Age Metals from Manika, Euboea, in A. Sampson, Μanika, Ο Protoelladikos Oikismos kai to Nekrotafeio II, Athens, 105-112.

The analytical data of the battle knife from Metaxada and the razor from Volimidia suggest that the use of arsenical copper was not limited to the Early and the Middle Bronze Age. Mycenaean bronze smiths, as late as the 14th c. BC, made use of arsenical copper at least in two occasions, during a period when the use of tin bronze for the manufacture of weapons, vessels, tools and implements was the rule. Even though both weapons are probably dated in the Early Mycenaean period, the possibility that artifacts of similar techniques and alloys had been manufactured later than the 14 th c. BC cannot be excluded. Moreover, the access to sources of arsenical copper as late as the Mycenaean period and the distinction between arsenical copper and pure copper that was used for the manufacture of the majority of the metallic artifacts is quite interesting. In any case, the selective use of copper alloys according not only to their mechanical properties but to their visual characteristics as well, clearly demonstrates the high level of expertise of the Mycenaean bronze smiths.

References Asderaki E., Tsatsouli K., Karydas A.G., 2006, Manufactured Technology and Materials of an Early Hellenistic Funerary Bronze Urn, 34th Internat. Symp. Archaeometry (3-7 May 2004, Zaragoza Spain), Zaragoza, 137-144. Balthazar, J.W., 1990, Copper and Bronze Working in Early through Middle Bronze Age Cyprus, Jonsered. Bittel Κ., 1959, Beitrag zur Kenntnis Anatolischer Metallgefässe des dritten Jahrtausend v. Chr., Jdl 74, 1-34. Blegen C.W., 1937, Prosymna, The Helladic Settlement Preceding the Argive Heraeum I-II, Cambridge, Massachusetts. Blegen C.W. and Rawson M., 1973, The Palace of Nestor at Pylos in Western Messenia ΙΙΙ, Princeton. Βossert Ε.Α., 1967, Kastri auf Syros, AD 22 A, 53-76. Branigan K., 1968, Silver and Lead in Prepalatial Crete, AJA 72, 219-229. Budd P. and Ottaway B.S., 1991, The Properties of Arsenical Copper Alloys: Implications for the Development of Eneolithic Metallurgy, in P. Budd, C. Chapman, C. Jackson and B. Ottaway (eds.) Archaeological Sciences 1989: Proc. Conf. Application Scientific Techniques to Archaeology, (Bradford, September 1989), 132-142. Caley E.R., 1949, On the Prehistoric Use of Arsenical Copper in the Aegean Region, Commemorative Studies in Honor of Theodore Leslie Shear, Hesperia Supplement 8, 60-63. Charles J.A., 1967, Early Arsenical Bronzes, a Metallurgical View, AJA 71, 21-26. Craddock P.T., 1976, The Composition of the Copper Alloys Used by the Greek, Etruscan and Roman Civilizations. 1. The Greeks before the Archaic Period, JAS 3, 103-123.

84

Ch. Tselios et al., Selective use of arsenical copper during the Mycenaean period Muhly J.D., 1985, Beyond Typology: Aegean Bronze Age Metallurgy in its Historical Context, in N.C. Wilkie, and W.D.E. Coulson (eds.) Contributions to Aegean Archaeology, Studies in Honor of W.A. McDonald, Minneapolis. Mylonas, G.E., 1973, O Tafikos Kyklos B ton Mykenon, Αthens. Northover J.P., 1989, Properties and Use of Arsenical Copper-Alloys, in A. Hauptmann, E. Pernicka and G.A. Wagner (eds.), Old World Archaeometallurgy. Proc. Internat. Symp. (Heidelberg 1987), Der Anschnitt Beiheft 7, Bochum, 111-118. Palmieri A.M., Hauptmann A. and Hess K., 1988, The Μetal Οbjects in the Royal Tomb Dating from 3000 b.c. found at Arslantepe (Malatya): a New Alloy (Cu-Ag), in A. Basim (ed.) Arkeometri sonuclari toplantisi XIII. Kultur Bakanligi Anitlar ve Muzeler Genel Mudurlugu, Ankara, 115-121. Papadimitriou G.D., 2001, I Ekseliksi ton Kramaton Chalkou ston Elladiko Choro mehri to Telos tis Geometrikis Epochis: Kramatikes Prosmiksis kai Technologiki Eksiliksi, in Y. Bassiakos, E. Αloupi and G. Fakorellis (eds.) Archaeometrikes Meletes gia tin Elliniki Proistoria kai Archaiotita, Athens, 587-608. Pernicka E., Begemann F., Schmitt-Strecker S. and Grimanis A.P., 1990, On the Composition and Provenance of Metal Artifacts from Poliochni on Lemnos, OJA 9 (3), 263-298. Rapp G. Jr., 1988, On the Origins of Copper and Bronze Alloy, in R. Maddin (ed.) The Beginning of the Use of Metals and Alloys: Papers from the 2nd Internat. Conf. on Beginning of Use of Metals and Alloys, Cambridge, 21-27. Rapp G. Jr., 1999, Copper, Tin and Arsenic Sources in the Aegean Bronze Age, in P.P. Betancourt, V. Karageorghis, R. Laffineur and W-D. Niemeyer (eds.) Meletemata III: Studies in Aegean Archaeology presented to Malcolm H. Wiener as he enters his 65th year, Aeageum 20, Liège-Austin, 699-704. Renfrew C., 1967, Cycladic Metallurgy and the Aegean Early Bronze Age, AJA 71, 1-20.

Scott D.A., 1991, Metallography and Microstructure of Ancient and Historic Metals, The J. Paul Getty Museum, Singapore. Schmidt H., 1902, Heinrich Schliemann’s Sammlung. Trojaninischer Altertümer, Berlin. Sotirakopoulou P., 1997, Κυκλάδες και βόρειο Αιγαίο: Οι σχέσεις τους κατά το δεύτερο ήμισυ της 3ης χιλιετίας π.Χ., in C. Doumas and V. La Rosa (eds.) Poliochni e l'antica età del bronzo nell'Egeo settentrionale, Athens, 522-542. Stos-Gale Z.A., Sampson A. and Mangou E., 1996, Analyses of Metal Artifacts from the Early Helladic Cemetery of Manika on Euboea, AEA 3, 49-62. Soles J.S., 2008, Metal Hoards from LM IB Mochlos, Crete, in I. Tzahili (ed.) Aegean Metallurgy in the Bronze Age. Proc. Internat. Symposium University of Crete, Rethymnon (November 19-21, 2004), Athens, 143-56. Tselios H.C., 2013, I Metallourgia kai i Metallotechnia tou Chalkou kata tin Ysteroelladiki Periodo stin Notiodytiki Peloponniso: Archaeometriki Meleti ton Chalkon Technergon tis Ysteroelladikis Messinias, unpublished PhD thesis, Αthens Unniversity. Tselios Τh., 2008, I Metallourgia tou Chalkou stin Proanaktoriki Kriti: Technologikew Eksiliksis kai Koinonikes Opsis, PhD thesis, Athens University. Varoufakis I.G., 1973, Metalographiki Ereuna Antikimenon ek Chalkou ek tou Tafikou Perivolou B ton Mykinon, in G.E. Mylonas, O Tafikos Kyklos B ton Mykinon, A-B, Athens, 363-375. Varoufakis G., 1991, Metallurgical Investigation of the Bronze Crater of Derveni, in W.A. Οddy (ed.), Aspects of Early Metallurgy. Proc. Symp. British Museum (22-23 April 1977), British Museum Occasional Papers 17, 71-86. Verdelis N.M., 1977, The Metal Finds, in P. Åström, Τhe Cuirass Tomb and Other Finds at Dendra. Part I: The Chamber Tombs, SIMA 4, Göteborg, 28-65. Weber C., 1983, Die Bronzenen Rasiermesser in Südosteuropa, Teil I: Das Kretisch-mykenische Griechenland, Frankfurt am Main.

85

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Table 1. Battle knife CM 605. SEM-EDS bulk analyses in consecutive frames from surface to core. (25.00 KeV). n.d. not detected.

Table 2. Battle knife CM 605. SEM-EDS bulk analyses on fragment of the Ag sheet. n.d. not detected.

Table 3. Razors CM 578, 602, 607. SEM-EDS bulk analyses n.d. not detected.

Table 4. Razor CM 578. SEM-EDS bulk analyses on Cu-As sheet. n.d. not detected.

Fig. 1. Battle knife CM 605. Diagram depicting the growing Ag content from core to surface in correlation with the decreasing Cu content.

86

Ch. Tselios et al., Selective use of arsenical copper during the Mycenaean period

Fig. 2. Map of the Southwestern Peloponnese with marked locations mentioned in the text, 1:200.000.

Fig. 3 (top). The razor CM 578 from Mastoraki chamber tomb 1 at Volimidia. Fig. 4 (bottom). The battle knife (Schlachtmesser) CM 605 from Metaxada.

Fig. 5. Battle knife CM 605. Sample from the back of the blade. Bulk analyses in consecutive frames, marked 1 to 5. Scale 10 microns.

Fig. 6 (left). Battle knife CM 605. Sample from the back of the blade. SEM - EDS line scanning from surface to core. Fig. 7 (right). Battle knife CM 605. Sample from the back of the blade. SEM - EDS line scanning from surface to core.

87

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

Fig. 8. Battle Knife CM 605. Sample from the back of the blade. Crystalline structure. Scale 100 microns. Etchant FeCl3 in alcohol.

Fig. 9. Battle Knife CM 605. Fragments of Ag sheet near the surface. Scale 100 microns.

Fig. 10. Razor CM 578. Fragments of Cu-As sheet on the surface. Scale 100 microns.

88

Chapter 12

Copper production during the Early Bronze Age at Aghios Antonios, Potos on Thasos Nerantzis Nerantzis1 and Stratis Papadopoulos2 1

Ephorate of Antiquities Rhodope, Ministry of Culture, Greece Ephorate of Antiquities Drama, Ministry of Culture, Greece

2

Corresponding author: [email protected]

Abstract The emergence of copper metallurgy in the northern Aegean is regarded as a derivative from more southeasterly areas whence the innovation of metals technology had originated. In the case of Thasos the possibilities offered by the occurrence of extensive metal bearing deposits enabled its early inhabitants to become engaged with metal extraction at an early stage during the Final Neolithic/Early Bronze Age. A growing body of evidence from Limenaria and more recently from Aghios Antonios, Potos testifies to the reduction of local, secondary copper ores associated with the rich Pb/Zn/Ag mineralization on the island. A project of instrumental analysis that focused on metallurgical residues from Aghios Antonios revealed information on smelting during which low contents of zinc and arsenic had been incorporated into the copper prills found entrapped in the slag. Although zinc, which is residual, does not surpass 1%, the presence of arsenic in contents up to 4% is significant. The unintentional formation of arsenical copper owing to the polymetallic nature of the utilized ores and the resulting effects on the produced metal would have caused a reorientation of procurement strategies and further treatment of raw materials thereof. Such a finding might be taken as an indication of potentially deliberate selection of arsenic bearing ores to co-smelt with malachite, both of which could be found on the island. The evidence from Aghios Antonios puts Thasos into the wider discussion concerning the nature of copper reduction and emergence of alloying technologies in the Early Bronze Age Aegean. Keywords: Thasos, copper, copper ores, slag, Early Bronze Age

which is dominant on the western part of the island has been suggested as the most probable source of the ores being processed at Limenaria (Bassiakos et al. 2015).

Introduction The island of Thasos lies 10 km south of the eastern Macedonian littoral and represents a dominant feature in the seascape of the northern Aegean. Its geology is characterized by marble and gneiss and a diversity of mineral ore deposits that had been the focus of extraction at an early stage. Archaeological evidence testifies to an early habitation of the island by prehistoric communities, which goes back at least to the Middle Neolithic (Papadopoulos and Malamidou 2012). Yet the mining of ochre for use in pigments, dating to the Upper Palaeolithic, suggests that its mineral wealth attracted human groups in earlier periods for which there is only sporadic archaeological data (Koukouli-Chrysanthaki and Weisgerber 1999). Neolithic inhabitants on the island utilized malachite for the making of ornaments and beads, and collected pieces of limonite possibly for use as pigments. The earliest evidence for extractive metallurgy derives from Limenaria on the SW part of the island (Fig. 1) where residues of lead and silver separation were recovered from a Final Neolithic layer of the middle 4th millennium BC (Papadopoulos 2008; Bassiakos 2012). The similarity of the Thasian litharge pieces to those found at Lamprika, Koropi in Attica as well as their dating to a synchronous FN phase testify to the practice of a similar process for lead-silver separation at both sites (Kakavogianni et al. 2006). The Pb/Zn mineralization

As far as copper metallurgy is concerned the earliest evidence, in form of slag pieces and crucible fragments, derives from an Early Bronze Age layer at the same settlement (Bassiakos 2012). Since the iron mineralization on Thasos contains appreciable copper it has been suggested that copper reduction on the island had started since the Early Bronze Age, although no firm evidence for early copper mining has been recorded so far. More recent findings of Bronze Age copper reduction that came to light at Aghios Antonios near Potos have added substantial new data to the issue of processing local mineral ores (Nerantzis and Papadopoulos 2013). An ongoing study of these residues has produced the first results concerning the nature of copper ore reduction being preliminarily presented and discussed here. The site and metalworking finds Aghios Antonios is a coastal settlement founded on a hilltop near Potos, 4 km southeast of Limenaria (Fig. 2). A two year excavation project revealed a complex of six Early Bronze Age buildings, five rectangular and one apsidal with stone foundations and mud-brick superstructures (Papadopoulos et al. 2014). The earliest

89

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry phase of occupation falls within the first half of the 4th millennium (3900-3600 BC) while the Early Bronze Age houses were erected sometime during the first half of the 3rd millennium (2900-2400 BC).

(Fe2SiO4) and magnetite (Fe3O3) as well as shiny metallic inclusions in form of prills and opaque minerals often green in colour. The presence of cuprite (Cu2O) has also been established in a few instances. Large spheroid copper prills (50μm to 1mm) and smaller inclusions were observed in most of the samples some of which are weathered to malachite.

The single metal find from the excavation is the pointy edge of a copper dagger. This scarcity of copper, apparent during the EBA on Thasos, is possibly due to an uneven diffusion of its production technology and probably owes more to the limited investigations focused on Bronze Age settlements in the north Aegean. Despite the lack of metal artefacts, significant finds related to metalworking practices were recovered from the house interiors and courtyards at Aghios Antonios. These include fragments of two crucibles, a clay mould for the casting of chisels, a small tuyère used for the blowing of air during heating and possibly alloying in crucibles, and pieces of slag. Investigation of the metallurgical residues is ongoing at N.C.S.R. ‘Demokritos’ as part of an INSTAP-funded project aiming to examine the various stages of early copper production at Aghios Antonios (Nerantzis and Papadopoulos 2013).

Based on the data acquired from the SEM/EDX analysis (Table 1) all specimens were found to be rich in barite (BaSO4) which is a characteristic compound of the geology in Thasos. Two types of fayalitic phases are prevalent. The first one is a light grey coloured, Ca-rich fayalite which is microcrystalline with increased Ba contents. The second is a darker grey fayalitic phase forming dendrite systems with higher silica contents. Magnetites are present in clusters of fine grained systems with an average crystal size around 10-15μm. A glassy phase although present is generally more infrequent. Nuggets of copper and thin filaments were also noted in a few instances (Fig. 7a). These grew into surface pores and fissures or appear as bands at the interface of sulphide inclusions with the slag matrix. Copper prills are frequent in some of the specimens indicating the loss of copper into the slag due to inefficient smelting conditions (Fig. 7b).

A total of 127 slag pieces were recovered during excavation. The majority (i.e. 70 pieces) comes from the surface and the upper thin layer, covering the archaeological features. Three major groups of slag were identified based on their macroscopic features. The first and largest group consists of dark-brown and black specimens with a ropey flow texture, which appear to relate to the processing of ferrous minerals (Fig. 3). The second group is made up of porous, concave specimens, which appear to derive from bloomery and/or smithing activity related to the first group (Fig. 4). The third group includes smaller-sized, lighter slag with brown encrustations and occasional green nodules on their surface (Fig. 5). Only a portion of slag from the surface layer and half the amount deriving from layer 2 were analysed since these represent iron production residues of historical times (Papadopoulos and Nerantzis 2012). All slag pieces deriving from layers 3, 4, 5 and 6, representing the Bronze Age habitation phase were analysed instrumentally (Fig. 6a).

The green minerals were identified by SEM/EDX as inclusions of un-reacted malachite. In most instances malachite contained in the slag had been weathered to atakamite [Cu2Cl(OH)3] due to the presence of Cl, prevalent in coastal environments such as that at Aghios Antonios. Matte phases in form of spheroid prills were observed in most of the samples, a fact that indicates the use of secondary copper minerals which contain traces of residual sulphides. Alternatively the source of sulphur could be attributed to the presence of barite which decomposes at low temperatures, contributing to the formation of matte phases during the smelt. Of special importance is sample S1, which appears as a heavily slagged ceramic mass most probably representing a crucible or rather a furnace wall fragment. The interface of the ceramic material with slag is highly vitrified and shows extensive bloating with large pores. Large quartz inclusions were noted in the ceramic paste while the major phases in slag are fayalite, magnetite, copper and matte prills, barite crystals concentrated in fissures and a glassy matrix (Fig. 7c). The SEM/EDX analysis has shown that matte phases in the adhered slag contain 2.52% As and 0.59% Zn, while the common occurrence of barite in clusters points to the Thasian iron ores containing barite and secondary copper oxides.

Results For a qualitative determination of the chemical composition of slags, X-Ray Fluorescence Spectroscopy (XRF/EDS) has been applied. Thirty-nine samples were found to contain significant amounts of Fe while their Ti contents and the absence of Cu led to their identification as bloomery iron slags. The remaining twenty-four samples are characterized as copper slag based on the presence of Cu, and traces of As and Zn. The results are in accordance with the stratigraphic sequence since the distribution of iron slag is far greater across the surface and layer 2 while layers 3 to 7, i.e. the prehistoric phases, contained almost exclusively copper slag (Fig. 6b).

The presence of 2-4% As within prills of metallic copper contained in the slag is significant as it hints at the relative dating of the residues to the Early Bronze Age, a period when arsenical copper alloys were widespread across the Aegean (Rapp 1999, 700). By the middle of the 3rd millennium the use of such alloys diminishes as tin gradually replaced arsenic as an alloying agent (Muhly 1985). These preliminary observations suggest that

Microscopic examination has shown that most of the copper slags show a glassy matrix filling the voids of crystalline phases, which are mainly composed of fayalite

90

N. Nerantzis and S. Papadopoulos, Copper production during the Early Bronze Age copper slag from Aghios Antonios derives either from the smelting of copper ores rich in arsenic or the deliberate co-smelting of arsenic-bearing ores along with copper minerals, as has been suggested for the case of PorosKatsambas in Crete of the EBA (Doonan et al. 2007). It should be mentioned that only a portion of the copper prills in Aghios Antonios slag contains up to 4% arsenic, yet these levels do not provide indisputable evidence for deliberate alloying since accidentally produced alloys would also produce such concentrations (Doonan et al. 2007, 111).

Tzines, where Palaeolithic ochre mining has been confirmed (Koukouli-Chrysanthaki and Weisgerber 1999), might indicate that Mavrolakkas was also a focus for mineral extraction during the Bronze Age. Copperbearing minerals, mainly malachite and azurite, can be found in this rich iron mineralization deposit. Samples were taken for the creation of a reference materials collection and their analysis is still ongoing. At the site of Elia, 1.5 km east of Aghios Antonios, an outcrop deposit had been exploited both for ochre and lead-silver during antiquity as attested by the shape of adits present on the western slope of the hill. There is evidence of mining on three levels. From the upper two levels two types of ochre had been mined, namely red ochre (hematite and goethite) and yellow ochre (jarosite rich in barite). The third level consists of the Pb/Ag mineralization, mainly galena and cerussite that has been the subject of extraction possibly since the Classical period. Interestingly, previous analysis has detected traces of chalcopyrite and malachite (Vavelidis et al. 2001, 635) within ores samples, a fact that might suggest a possible source of copper that became exhausted. The proximity of this mine to Aghios Antonios makes it a possible source of copper in prehistory and iron for bloomery smelting in historic times, in addition to the iron-rich sands of southern Thasos (Photos 1992).

Experimental studies carried out previously have shown that sulphur, present as sulphide in the metallic mineral portion of a primary ore, can reduce copper oxide ore to metallic copper in a co-smelting operation. At sufficiently high smelting temperatures, the sulphur extracts oxygen from the oxide ore, thereby reducing the ore, and is eliminated as sulphur dioxide (Lechtmann and Klein 1999, 499). In a hypothetical co-smelting of oxide and/or suphide ores with arsenopyrite, which are present in the Thasian mineralisation, the product would be a copper– arsenic alloy according to the following equation: 3 CuCO3 + FeAsS  Cu,As + FeO + SO2 + 3 CO2  Whether accidental or deliberate, the maintenance of arsenic concentrations in the produced copper is difficult since arsenic is prone to volatilization. Based on the available evidence from Aghios Antonios co-smelting could not be ruled out, although further investigation is necessary for a better understanding of the smelting process.

Conclusions Based on the surviving evidence it appears that Aghios Antonios was participating in a network of exchange on a technological level during the 3rd millennium BC. It is important to examine the possibility of sharing technical knowledge on metal production within the context of increasing interaction of island communities that utilized mineral ores since the Final Neolithic/Early Bronze Age transition (Tzachili 2008; Zachos 2007). Furthermore, the technological change associated with a shift from arsenical copper to tin bronze should be examined on Thasos in the context of the earliest adoption of the new alloys in the north Aegean (Pernicka et al. 1990). The early presence of bronze jewellery and tools from Troy, Poliochni and Thermi does not imply that bronze metallurgy was restricted to those islands (Pernicka et al. 2003). Recent findings of tin bronze production at EBA Palamari on Skyros (Parlama 2007) indicate that the use of the new alloy might have been more widespread due to increased contacts with western Anatolia (Mehofer 2014) whence tin probably came from.

The possible sources of copper ores So far the data are inconclusive regarding the source of the secondary oxidized copper-bearing minerals as is the source of arsenic, although the local copper minerals found in iron ores and arsenopyrites represent the most likely sources. Nonetheless the presence of barium is further proof for the indigenous origin of the utilized ores at Aghios Antonios. With regards to copper ores, these were of limited availability on Thasos as reported by the team from Bochum Bergbau-Museum and the MaxPlank-Institute that carried out mineralogical investigations in the 1980s (Wagner and Weisgerber 1988). The presence of copper-bearing mineralization at Makryrachi, Mavrolakkas and Marlou has been taken as proof that mining could have potentially started during the Bronze Age, although no solid evidence for such practice has been ever reported in the literature. Furthermore, the presence of arsenopyrites and secondary copper minerals in the above mining locations (Vavelidis et al. 1988, 55) suggests that treatment of such ores in the same furnace could have been responsible for the unintentional production of arsenical copper.

Up to now early tin bronze production on Thasos was unknown. This owes to the small number of analysed bronze artefacts from securely dated contexts. The earliest bronze artefacts derive from the cemeteries around Kastri and date to the Late Bronze Age (Photos 1992). It is interesting to note that lead isotope analysis of these artefacts revealed exploitation of local deposits at a rate of only 15% (Stos-Gale and Gale 1992, 784). Out of a total of 59 copper alloy artefacts only one object of Final Neolithic and seven objects of Early Iron Age date are consistent with the Thasian lead isotope ore field. The provenance of all Late Bronze Age artefacts is not

Mavrolakkas shows extensive evidence of modern opencast mining on the hematite-limonite deposit which also contains secondary copper oxides and the possibility of ancient extraction is highly possible. Its proximity to

91

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Thasian and instead points to the Anatolian, Cypriot and possibly also the Lavrion ore fields (Stos-Gale and Gale 1992, 792). Whether the alleged ceasing of Thasian copper production during the Late Bronze Age is due to depletion of the local copper deposits, or of increasing Mycenaean trading, or indeed is a product of the analytical ambiguities inherent in lead isotope studies (Budd et al. 1993) is still an issue of conjecture. At any rate, our analytical data corroborate the use of indigenous mineral ores for copper production, at least for the earlier phases of the Bronze Age. Further investigation is needed to account for its development, expansion and eventual decline in later stages.

Journal of Archaeological Science 26, 497-526. Mehofer M., 2014, Metallurgy during the Chalcolithic and the Beginning of the Early Bronze Age in Western Anatolia, in B. Horejs and M. Mehofer (eds.) Western Anatolia before Troy, ProtoUrbanisation in the 4th millennium BC?, Proc. Int. Symp. Kunsthistorisches Museum Wien. Vienna: Austrian Academy of Sciences Press, 463-490. Muhly J.D., 1985, Sources of tin and the beginnings of bronze metallurgy, American Journal of Archaeology 89(2), 275-291. Nerantzis N. and Papadopoulos S., 2013, Reassessment and new data on the diachronic relationship of Thasos Island with its indigenous metal resources: a review, Archaeological and Anthropological Sciences 5(3), 183-196. Papadopoulos S., 2008, Silver and Copper Production Practices in the Prehistoric Settlement at Limenaria, Thasos, in I. Tzachili (ed.) op. cit., 59-67. Papadopoulos S., Nerantzis N. and Tsoutsoubei-Lioliou S., 2014, Μια νέα αρχαιολογική θέση στον Άγιο Αντώνιο Ποτού", in Κ. Χιόνης (επιμ.) Θασιακά 16 (2011-2014), ΣΤ΄Συμπόσιο Θασιακών Μελετών "Η Θάσος διά μέσου των αιώνων "Ιστορία-ΤέχνηΠολιτισμός, Δήμος Θάσου, 371-380. Papadopoulos S. and Malamidou D., 2012, Προϊστορικός Οικισμός Λιμεναρίων: Το χρονικό της Έρευνας, Πρακτικά Ημερίδας Δέκα Χρόνια Ανασκαφική Έρευνα στον Προϊστορικό Οικισμό Λιμεναρίων Θάσου, ΥΠΠΟΤ-ΙΗ’ ΕΠΚΑ, Θεσσαλονίκη. Papadopoulos S. and Nerantzis N., 2012, New evidence on ancient mining from Lekani and metal production from Aghios Antonios, Thasos, in Ν. Ζαχαριάς (ed.) Πρακτικά 2ου Συμποσίου ARC-RNT ‘Αρχαιολογική Έρευνα και Νέες Τεχνολογίες’, Πανεπιστήμιο Πελοποννήσου-ΤΙΑΔΠΑ, Καλαμάτα, 21-23/10/2010, 177-181. Parlama L., 2007, Παλαμάρι Σκύρου. Παρατηρήσεις στην εξέλιξη του οικισμού κατά την 3η π.Χ. χιλιετία και προβλήματα αστικοποίησης, in A.A. Laimou, L.G. Mendoni , N. Kourou and E. SimantoniBournia (eds.). Αμύμονα Έργα Τιμητικός τόμος για τον καθηγητή Βασίλη Κ. Λαμπρινουδάκη, Περιοδικό «Αρχαιογνωσία» αρ. 5, Athens, 25–48. Pernicka E., Begemann F., Schmitt-Strecker S. and Grimanis A.P., 1990, On the Composition and Provenance of Metal Artefacts from Poliochni on Lemnos, Oxford Journal of Archaeology 9(3), 263298. Pernicka E., Eibner C., Öztunali Ö. and Wagner G. A., 2003, Early Bronze Age Metallurgy in the Northeast Aegean, in G.A. Wagner, E. Pernicka and H.-P. Uerpmann (eds.) Troia and the Troad, Scientific Approaches, Natural Science in Archaeology, 143-172 Photos E., 1992, Late Bronze Age- Early Iron Age Copper and iron slags from Kastri and Paliokastro on Thasos, in Ch. Koukouli-Chrysanthaki (ed.) Πρωτοϊστορική Θάσος, Τα νεκροταφεία του οικισμού Καστρί, Αθήνα: Δημοσιεύματα του Αρχαιολογικού Δελτίου, αρ. 45, 795-801. Rapp Jr. G. 1999 Copper, tin and arsenic sources in the

Acknowledgments We are indebted to the Institute for Aegean Prehistory (INSTAP) for generously funding the investigation of metalworking at Aghios Antonios and our collaborator Dr. Y. Bassiakos. Thanks are due to the two anonymous reviewers for commenting on the draft.

References Bassiakos Y., 2012, Πρώιμη παραγωγή χαλκού και αργύρου στα Λιμενάρια Θάσου: Μια τεχνολογική προσέγγιση. Πρακτικά Ημερίδας Δέκα Χρόνια Ανασκαφική Έρευνα στον Προϊστορικό Οικισμό Λιμεναρίων Θάσου, ΥΠΠΟΤ-ΙΗ’ ΕΠΚΑ, Θεσσαλονίκη, 197-225 Bassiakos Y., Papadopoulos S. and Nerantzis N., 2015, Early copper and silver production at prehistoric Limenaria, Thasos: a technological approach, in Bassiakos Y. (ed.) Prehistoric Metal Production in the Aegean: Material Evidence and Analysis. Heidelberg: Springer-Verlag. Budd P., Gale D., Pollard A.M., Thomas R.G. and Williams P.A., 1993 Evaluating lead isotope data: further observations, Archaeometry 35, 241–263. Doonan R.C.P., Day P.M. and Dimopoulou-Rethemiotaki N., 2007, Lame excuses for emerging complexity in Early Bronze Age Crete: the metallurgical finds from Poros-Katsambas and their context, in P.M. Day and R.C.P. Doonan (eds.) Metallurgy in the Early Bronze Age Aegean. Sheffield Studies in Aegean Archaeology 7, Oxbow Books, 98-122. Kakavogianni O., Douni K., Nezeri F. and Georgakopoulou M., 2006, Απόπειρα τεχνολογικής προσέγγισης της παραγωγής αργύρου και μολύβδου κατά την Τελική Νεολιθική και Πρωτοελλαδική Ι Περίοδο στα Μεσόγεια, Πρακτικά Β΄ Διεθνούς Συνεδρίου Αρχαίας Ελληνικής Τεχνολογίας, Αθήνα, 77-83 Koukouli-Chrysanthaki Ch. and Weisgerber G., 1999, Prehistoric Ochre Mines on Thasos, in Ch. Koukouli-Chrysanthaki, A. Muller and S. Papadopoulos (eds.) Πρώτες Ύλες και Τεχνολογία από τους Προϊστορικούς Χρόνους ως Σήμερα, Πρακτικά Διεθνούς Συνεδρίου, Λιμενάρια Θάσου, ΥΠΠΟ-ΙΗ’ ΕΠΚΑ, 129-144. Lechtman H. and Klein S., 1999, The Production of Copper Arsenic Alloys (Arsenic Bronze) by Cosmelting: Modern Experiment, Ancient Practice,

92

N. Nerantzis and S. Papadopoulos, Copper production during the Early Bronze Age Aegean Bronze Age, in Betancourt P., Karageorghis V., Laffineur R. and Niemeier W-D. (eds.) Meletemata: Studies in Aegean Archaeology presented to Malcolm H. Wiener as he enters his 65th year. Vol III, Universite de Liège, Historie de l’ art et archéologie de la Grèce antique, University of Texas at Austin: Programs in Aegean Scripts and Prehistory, 699-704. Stos-Gale Z. and Gale N., 1992, Sources of copper used on Thasos in Late Bronze and Early Iron Age, in Ch. Koukouli-Chrysanthaki (ed.) Πρωτοϊστορική Θάσος, Τα νεκροταφεία του οικισμού Καστρί, Αθήνα: Δημοσιεύματα του Αρχαιολογικού Δελτίου, αρ. 45, 782-793. Tzachili I. (ed.), 2008, Αegean Metallurgy in the Bronze Age, Proc. Int. Symp. University of Crete, Rethymnon, Greece, Ta Pragmata. Vavelidis M., Gialoglou G., Pernicka E. and Wagner

G.A., 1988, Die Buntmetall und Eisen-ManganLagerstatten von Thasos, in G. A. Wagner and G. Weisgerber (eds.) op. cit., 40-58. Vavelidis M., Gialoglou G., Trontsios G., Melfos B. and Weisgerber G., 2001, Ένα αρχαίο μεταλλείο ώχρας και μολύβδου-αργύρου στην περιοχή Ελιά της Θάσου, in Μπασιάκος Ι., Αλούπη Ε., Φακορέλλη, Γ. (eds.) Αρχαιομετρικές μελέτες για την Ελληνική Προϊστορία και Αρχαιότητα, Γ Συνέδριο της ΕΑΕ & ΚΜΑΜ, Αθήνα. Wagner G.A. and Weisgerber G., 1988, Antike Edel- und Buntmetallgewinnung auf Thasos, Der Anschnitt, Beiheft 6, Bochum: Bergbau-Museum. Zachos K., 2007, The Neolithic Background. A Reassessment, in P.M. Day and R.C.P. Doonan (eds.) Metallurgy in the Early Bronze Age Aegean, Sheffield Studies in Aegean Archaeology 7, Oxbow Books, 168-206.

Fig. 1. Map of Thasos showing sites mentioned in the text.

Fig. 2. Aghios Antonios hill, view from the SE.

93

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

Fig. 3. Tapping slag (group A).

Fig. 4. Furnace slag (group B).

Fig. 6a. Number of slag pieces per layer.

Fig. 5. Copper slag (group C).

b. Iron and copper slag relative abundance per layer.

Fig. 7. SEM photomicrographs of slag sections: top left a. fayalitic phase (light grey forming the background), magnetite (dark grey angular crystals), copper filament (white, centre of image); top right b. metallic copper prill; bottom left c. fayalitic phase (light grey), magnetite crystals (dark grey), barite crystals (white cluster at centre).

94

Chapter 13

A comparative study of Cypriot bronzes dated to the Late Bronze and the Early Iron Age Andreas Charalambous, Vasiliki Kassianidou and George Papasavvas Archaeological Research Unit, Department of History and Archaeology, University of Cyprus, P.O. Box 20537, Nicosia, CY1678, Cyprus Corresponding author: [email protected]

Abstract This paper concerns the compositional analysis of a significant number of copper alloy artefacts from three sites in Cyprus, and aims to investigate the types of alloys used on the island during the Late Bronze and the Early Iron Age. The sites under study are the area of the modern city of Limassol (a number of Late Bronze Age tombs situated in several areas of the city), the Late Bronze Age settlement of Pyla Kokkinokremos and the Early Iron Age cemetery of Palaepaphos Skales. This research intends to provide significant information primarily regarding the use of tin, as well as the presence and use of other metals, for the production of the bronze artefacts. Keywords: p-XRF, chemical analysis, copper alloys, Late Bronze Age, Early Iron Age, Limassol, Pyla Kokkinokremos, Palaepaphos Skales, Cyprus

weapons (mainly daggers) and ornaments (rings and pins). The second group of 65 objects comes from the settlement of Pyla Kokkinokremos which dates to the beginning of the 12th century BC (Table 1) (Karageorghis and Demas 1984). Over half of this assemblage consists of various forms of scrap metal, such as wires, sheets and fragmentary artefacts of various categories, such as tools, weapons and ornaments. The third group of 157 artefacts comes from the necropolis of Palaepaphos Skales which dates to the early Iron Age (Karageorghis 1983). The assemblage consists of several types of tools, weapons, vessels, and ornaments (Table 1).

Introduction Although Cyprus is extremely rich in copper there are no tin deposits on the island and therefore this metal had to be imported (Kassianidou 2003). The first objects made of bronze appeared in Cyprus in the transition between the Early and the Middle Bronze Age (i.e. Early Cypriot III (2000-1900) to Middle Cypriot I (1900-1800 BC), but most of them are believed to have been imported as finished artefacts (Weinstein Balthazar 1990). Arsenical copper and unalloyed copper retained their predominance until the end of the Middle Bronze Age. It is in the Late Bronze Age (1600-1050 BC) that bronze is finally established and this is intimately associated with other major changes and developments that take place on the island (Kassianidou 2003). During the so-called ‘Crisis Years’ of the 12th century BC an important number of major cities in the Aegean and the Eastern Mediterranean were abandoned or destroyed and as a result trade networks collapsed (Muhly 1984; Snodgrass 2000). It has often been argued that this resulted in a shortage of tin which must also have affected Cyprus, but was this really the case? In order to answer this question we analyzed the copper alloy artefacts from three sites on the island which date to the beginning of the Late Bronze Age (Limassol), the end of the Late Bronze Age (Pyla Kokkinokremos), and the Early Iron Age necropolis of Palaepaphos Skales.

Method of Analysis A portable, handheld Innov-X Delta XRF analyser was used for the non-destructive analysis of the artefacts. The use of the pXRF was a requisite because it was not permitted to either take the objects out of the museum or remove samples which would enable the use of other analytical techniques. The specific instrument is equipped with a 4W, 50kV tantalum anode X-Rays tube and a high-performance Silicon Drift Detector with a resolution of 155 eV (Mo-Kα). The diameter of the XRays beam was 3 mm. The final reported composition for each object is the mean value of three to five measurements conducted on corrosion-free areas that were detected and chosen using a high-resolution handheld microscope (ProScope HR, Bodelin Technologies). The measurement time for each spot analysis was 70 seconds. Certified reference materials like CRM-875 (bronze standard) and BCR-691 (set of five copper alloys) were used for checking the accuracy of the applied analytical mode.

The studied artefacts The first group of 22 copper alloy artefacts that was analysed comes from tombs excavated in an area of the modern city of Limassol (Karageorghis and Violaris 2012) (Fig. 1). The assemblage, which dates to the Late Cypriot (LC) I-II (1600-1300 BC) consists of

95

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry behind the use of this alloy must have been to produce objects with a golden colour, probably in an attempt to create prestigious artefacts that imitate the appearance of similar artefacts made of gold (Keswani 2004; Ashkenazi et al. 2012; Papasavvas 2012). At the other extreme, there are two artefacts, an obelos and a bowl, with a tin content below 1% (0.3 and 0.7%, respectively), and only three artefacts, namely a pair of tweezers, a pin and a base of a fragmentary vessel, which contained no tin at all.

Results and Discussion The results of the EDXRF analyses indicated that almost all artefacts are made of copper-tin alloys, with lead, iron, arsenic and zinc being the major metallic impurities. The tin (Sn) content of the bronze artefacts is reported in Fig. 2 as histograms of the frequency distribution for each assemblage. The addition of tin improves the hardness and corrosion resistance of the alloy and reduces its melting temperature (Tylecote 1987; Klein and Hauptmann 1999; Ashkenazi et al. 2012). Although scholars do not always agree on the matter, because tin rarely occurs in concentrations higher than 0.1% in copper ores, anything above this limit should be considered a deliberate addition (Hall and Steadman 1991; Moorey 1994; Pernicka et al. 1990). However, tin levels below 2% may also be the result of mixing copper with recycled bronze.

Lead (Pb) had frequently been used in antiquity as an alloying element with copper, as it improved fluidity and cast ability (Klein and Hauptmann 1999), but it reduced the alloy’s hardness and toughness when added in higher amount than a few per cent (Craddock 1977; GiumliaMair 1992). Lead, in the assemblage from Limassol (Table 2), is detected in 15 objects, ranging from 0.1 to 0.6%, while two pins have a much higher amount of lead (1.9 and 6.2%). In the assemblage from Pyla Kokkinokremos, Pb is detected only in 23 objects, at levels ranging from 0.1 to 0.4% while in four cases, namely a pendant (1%), a dagger (1.3%), a tool (3.2%) and a small cylindrical fragment (19.9%), the amount of lead is much higher. Finally, in the assemblage from Palaepaphos Skales, Pb is detected in all objects, ranging from 0.1 to 2.6%. Four objects, namely a hemispherical bowl, a fibula and the two rod tripods, reveal much higher Pb concentrations of 5.2%, 6.7%, 4.6% and 11.9% respectively. In the objects with high Pb concentrations, there is a possibility that the results of the analysis do not accurately reflect the actual concentration of the metal. Lead, when added in significant amounts to molten copper, is not soluble, and has the tendency to produce a dispersion of fine particles on the surface of the object (Giumlia-Mair 1992; Ingo et al. 2004). Since this is a surface analysis, there is a possibility that in the areas analysed such globules of lead were present. Copper ores are often associated with lead (Hauptmann 2007) and thus when low amounts of lead are detected in copper alloys they are usually believed to be impurities. But Cypriot copper ores are unusually free of lead (Constantinou 1982); thus its presence in the studied artefacts, even at levels of 1% or lower, can only be interpreted as a deliberate or accidental addition.

In the assemblage from Limassol, the average Sn content is 7.5 ± 3.7% (average and one standard deviation). More specifically, 18 out of 22 artefacts have a tin concentration ranging between 5.4 and 14.2%. Significantly, all the objects which have a concentration of tin above 10% are daggers while the example with the highest concentration of tin had a bone hilt with rivets covered by gold sheet. Two artefacts, a spearhead and a ring, were found to have a very low percentage of tin (0.2% and 0.3% respectively). Such a low concentration is highly unlikely to have been the result of a deliberate addition of tin to copper. Furthermore, as Cypriot copper ores do not contain tin (Constantinou 1982), this small amount must have been introduced accidentally perhaps when copper metal was melted together with recycled bronze (Weinstein Balthazar 1990). The average tin content of the 65 objects from Pyla Kokkinokremos is 5.6 ± 3.9%. Because the assemblage consists mostly of scrap metal, it exhibits a great variability in the tin content. Three artefacts, namely a nail, a thin mass and a sheet, contain no tin, while four objects, namely an amorphous lump, a small fragment and two sheets, have a tin concentration ranging between 0.5 and 1.5%. Twenty nine objects, mainly scrap metal but also ornaments, tools and weapons have a tin concentration between 5.2 and 9.7 % and only 14 objects, including four needles, a spatula, a fragment of a bull figurine, the tip of a dagger and scrap metal in the form of thin masses and sheets have a tin concentration higher than 10%. As in the case of Limassol, the object with the highest concentration of tin (16.8 ± 0.5%) is a dagger.

Arsenic (As), when present concentrations higher than 2%, improves the properties of the alloy, resulting in the increase of its ductility and hardness (Pernicka et al. 1990; Hauptmann 2007). In the assemblage from Limassol, As is detected only in six objects, ranging from 0.2% to 0.4%, while in the assemblage from Pyla Kokkinokremos, is detected in 30 objects, ranging from 0.2 to 0.4%. Furthermore, in the assemblage from Palaepaphos Skales, arsenic is present in 25 artefacts, in a concentration of 0.2%. The small detected concentrations indicate a non-intentional addition of the specific element. Arsenic was normally introduced to the alloy through the smelting of polymetallic copper ores, which contained small amounts of arsenic (Giumlia-Mair 1992; Swiny 1982), or through the flux (Tylecote 1982). However, the only area in the island where ores with high enough concentrations of arsenic suitable for the

In the assemblage from the necropolis of Palaepaphos Skales, the average Sn content of the largest group of artefacts, which consists of needles, tweezers, fibulae, pins, rings, finger-rings, ear-rings and spearheads, is 8.1 ± 2.2%. However, there is another group of 32 artefacts which are made of a high-tin bronze. This group, comprised mainly of hemispherical bowls (27 out of the 32), has an average tin content of 18.9 ± 2.8%. Clearly there was a deliberate choice of alloy, one with a tin content which is consistently over 10% and usually much higher, for this type of artefact. We believe that the reason

96

A. Charalambous et al., A comparative study of Cypriot bronzes production of arsenical copper are Laxia tou Mavrou and Pevkos, which are located in the Limassol Forest, southwest of the Troodos massif (Gass et al. 1994). Much more likely is the possibility that these objects were partially made of metal deriving from the recycling of objects dating to the Early and/or Middle Cypriot which were made of arsenical copper. The fact that almost all objects that contain arsenic also contain tin confirms this possibility (Weinstein Balthazar 1990).

the alloys to improve their cast ability, although its low concentration in many cases indicated that they were possibly made of recycled metal with initially higher concentration of this additive. The presence of arsenic in a number of artefacts under study is interpreted as evidence for the use of recycled metal deriving from artefacts dating to the Early and Middle Bronze Age. Iron and zinc are believed to be non-intentional additions to the alloys, resulting from the smelting process and originating either in the copper ores or in the flux.

Iron (Fe) was found in all analysed objects in a concentration range of 0.05-1.6%. The vast majority of the objects have an iron content below 1% and only a small number exceed this concentration. The presence of iron can be interpreted as coming through the smelting of chalcopyrite, CuFeS2 (Tylecote 1982), the most common copper ore in Cyprus (Constantinou 1982). Also, the deliberate or accidental use of iron minerals as a fluxing agent during the smelting procedure could have resulted in the introduction of a small quantity of iron in the finished artefacts (Muhly 1984; Weinstein Balthazar 1990; Ashkenazi et al. 2012).

Acknowledgements The authors wish to thank Dr. Maria Hadjicosti, former Director of the Department of Antiquities of Cyprus, and the excavators of the sites, Prof. Vassos Karageorghis (for Pyla and Palaepaphos) and Yiannis Violaris (for Limassol), for the permission to analyse the bronze artefacts from the three areas. The analyses of the objects were carried out by one of the authors (ACh) as part of a research project entitled ‘A diachronic Study of Cypriot Metalwork’. This project formed part of the NARNIA (New Archaeological Research Network for Integrating Approaches to ancient material studies) Project (Marie Curie Initial Training Network). For more information please visit the NARNIA website: http://narnia-itn.eu/.

No zinc (Zn) was detected in the assemblage from Limassol. In the assemblage from Pyla Kokkinokremos, zinc is found only in 15 objects in the range 0.1% to 0.4%, while in the assemblage from Palaepaphos Skales, zinc is detected only in 20 objects in the range 0.2% to 1%. The presence of zinc can be justified as a nonintentional addition to the alloy, resulting from the smelting procedure (Hauptmann 2007), due to its occurrence in the Cypriot copper sulphide ores (Constantinou 1982).

References Ashkenazi D., Iddan, N. and Tal O., 2012, Archaeometallurgical characterization of Hellenistic metal objects: the contribution of the bronze objects from Rishon Le-Zion (Israel), Archaeometry 54 (3), 528-548. Constantinou G., 1982, Geological Features and Ancient Exploitation of the Cupriferous Sulphide Orebodies of Cyprus, in J.D. Muhly, R. Maddin and V. Karageorghis (eds.) Early Metallurgy in Cyprus, 4000-500 B.C., Pierides Foundation and Department of Antiquities of Cyprus, Nicosia, 13-23. Craddock P.T., 1977, The composition of the copper alloys used by the Greek, Etruscan and Roman civilisations: 2. the Archaic, Classical and Hellenistic Greeks, Journal of Archaeological Science 4, 103-23. Figueiredo E., Silva R.J.C., Senna-Martinez J.C., Fátima Araújo M., Braz Fernandes F.M. and InêsVaz J., 2010, Smelting and recycling evidences from the Late Bronze Age habitat site of Baiões (Viseu, Portugal), Journal of Archaeological Science 37, 1623-1634. Gass I.G., MacLeod C.J., Murton B.J., Panayiotou A., Simonian K.O. and Xenophontos C., 1994, The Geology of the Southern Troodos Transform Fault Zone, Cyprus, Geological Survey Department Memoir 9, Nicosia, 183-184. Giumlia-Mair A., 1992, The composition of copper-based small finds from a west Phoenician settlement site and from Nimrud compared with that of contemporary Mediterranean small finds, Archaeometry 34 (1), 107-119. Hall M.E. and Steadman S.R., 1991, Tin and Anatolia:

Conclusions The chemical analysis of a large number of copper alloy artefacts from the sites of Limassol, Pyla Kokkinokremos and Palaepaphos Skales, brought to light some interesting results. Already from the earliest phase of the Late Bronze Age, represented by the objects from Limassol, bronze was used to produce a variety of objects but it was the daggers which were made of alloys with a tin concentration over 10%. On the other hand, in the case of Pyla Kokkinokremos, the fact that more than half of the assemblage, mainly composed of different categories of scrap metal, has a tin concentration lower than 5% may indicate that tin was not available in abundance in the specific area, during that period (mid 12th century BC). As for Palaepaphos Skales, the analyses of the artefacts showed the tin content to be generally high, discouraging any suggestions for a shortage in this metal. The analysis of different categories of artefacts revealed the use of different alloys suitable for the specific type of artefact. Thus, tools and weapons were produced with the optimal amount 8-10% of tin indicating the very good knowledge of the properties of this precious additive. For the group of the hemispherical bowls, a different alloy with much higher percentage of tin was used in order to produce objects whose colour imitated that of gold. Regarding the other elements detected, lead was added deliberately into

97

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Another look, Journal of Mediterranean Archaeology 4, 217-234. Hauptmann A., 2007, The Archaeometallurgy of Copper. Evidence from Faynan, Jordan, Springer-Verlag, Berlin Heidelberg. Ingo G.M., Angelini E., De Caro T., Bultrini G. and Calliari I., 2004, Combined use of GDOES, SEM+EDS, XRD and OM for the microchemical study of the corrosion products on archaeological bronzes, Applied Physics A 79, 199-203. Karageorghis V., 1983, Palaepaphos-Skales: An Iron Age Cemetery in Cyprus. Ausgrabungen in Alt-Paphos auf Cypern, Band 3, Konstanz. Karageorghis V. and Demas M., 1984, PylaKokkinokremos A Late 13th-century B.C. Fortified Settlement in Cyprus, Department of Antiquities, Nicosia. Karageorghis V. and Violaris Y., 2012, Tombs of the Late Bronze Age in the Limassol area, Cyprus (17th-13th centuries BC), Nicosia. Kassianidou V., 2003, The trade of tin and the island of copper, in A. Giumlia-Mair and F. Lo Schiavo (eds.) Le problème de l'étain à l'origine de la métallurgie. The Problem of Early Tin, BAR International Series 1199, Oxford: BAR Publishing, 109-119. Keswani P., 2004, Mortuary Ritual and Society in Bronze Age Cyprus, Monographs in Mediterranean Archaeology, Equinox, London. Klein S. and Hauptmann A., 1999. Iron Age leaded tin bronzes from Khirbet Edh-Dharih, Jordan, Journal of Archaeological Science 26, 1075-1082. Moorey P. R. S., 1994, Ancient Mesopotamian Materials and Industries, Oxford: Clarendon Press. Muhly J.D., 1984, The role of the Sea Peoples in Cyprus

during the LC III period, in V. Karageorghis and J.D. Muhly (eds.) Cyprus at the Close of the Late Bronze Age, A.G. Leventis Foundation, Nicosia, 39-56. Papasavvas G., 2012, Profusion of Cypriot copper abroad, dearth of bronzes at home: a paradox in Late Bronze Age Cyprus, in V. Kassianidou and G. Papasavvas (eds.) Eastern Mediterranean Metallurgy and Metalwork in the Second Millennium BC, Oxford: Oxbow Books, 117-128. Pernicka E., Begemann F., Schmitt-Strecker S., and Grimanis A.P., 1990. On the composition and provenance of metal artefacts from Poliochni on Lemnos, Oxford Journal of Archaeology 9 (3), 263298. Snodgrass A.M., 2000, The Dark Age of Greece, Edinburgh: Edinburgh University Press. Swiny S., 1982, Correlations between the composition and function of Bronze Age metal types in Cyprus, in J.D. Muhly, R. Maddin and V. Karageorghis (eds.) Early Metallurgy in Cyprus 4000 - 500 BC, Nicosia, 69-80. Tylecote R.F., 1982, The Late Bronze Age: Copper and bronze metallurgy at Enkomi and Kition, in J.D. Muhly, R. Maddin and V. Karageorghis (eds.) Early Metallurgy in Cyprus, 4000-500 B.C., Nicosia, 81100. Tylecote R.F., 1987, The Early History of Metallurgy in Europe, New York: Longman. Weinstein Balthazar J.W., 1990, Copper and Bronze Working in Early through Middle Bronze Age Cyprus, Studies in Mediterranean Archaeology and Literature, Pocket-book 84, Paul Åströms Forlag, Jonsered.

Fig.1. Map of Cyprus showing the locations of Limassol, Pyla Kokkinokremos and Palaepaphos Skales. Map based on digital geological data from the Department of Geological Survey.

98

A. Charalambous et al., A comparative study of Cypriot bronzes

Fig. 2. Histograms showing the distribution of the Sn content in the studied bronze artefacts.

Table 1. The studied copper alloy artefacts Weapons

Site Limassol

Tools

Table-ware / Utensils

Jewellery

Daggers

11

Rings

7

Spearhead

1

Pins

3

Arrowheads

2 Needles Handles Spatula

7 2 1

Earrings Bracelet

2 1

Knife Nail

1 1

Attachments / pendants

2

Pins

2

Others

3

Scrap Metal

Total: 22

Pyla Kokkinokremos

Fragments Flat sheets Wires

28 11 2

Total: 65 Spearheads

6 Tweezers

3

Arrowhead

1 Needles Ladle

11 1

Palaepaphos Skales

Fibulae

47

Bowls

32

Pins Rings

12 12

Vessel Obeloi

1 3

Rod tripods

2

Awl

1

Finger-rings

10

Saw

1

Earrings

6

Spatula

1

Bracelet

1

Others

6 Total: 157 TOTAL: 244

Table 2. Pb, As, Fe and Zn content of the majority of the studied artefacts Site

Pb (%)

As (%)

Fe (%)

Zn (%)

Limassol

0.1 - 0.6

0.2 - 0.4

0.1 - 1.6

­

Pyla Kok k inok remos

0.1 - 0.4

0.2 - 0.4

0.1 - 1.5

0.1 - 0.4

Palaepaphos Sk ales

0.1 - 2.6

0.2

0.05 - 1.4

0.2 - 1

99

Chapter 14

Conservation of a silver coin hoard from the Geometric period Necropolis in Samos Ioannis P. Staikopoulos 21st Ephorate Ministry of Culture and Sports, Athens, Greece Author’s address: [email protected]

Abstract The present paper deals with the methodological approach that was adopted for the conservation of a hoard of twenty seven silver coins, discovered in a plain cylindrical lead pyxis with lid, at the Geometric cemetery of ancient Samos. A hard and insoluble homogenous corrosion layer, with a characteristic off-white colour, covered the coins as well as the body and the lid of the pyxis. This corrosion layer could not be removed by application of any materials and methods that can be found in the bibliographic records and are commonly being applied in such cases. The analysis of this layer indicated that it consisted of lead carbonate (cerussite). In order to remove the lead carbonate corrosion, the standard treatment methodology was adjusted and involved the combination of a series of cleaning agents and techniques aiming at the most effective cleaning while taking into account the principle of minimum intervention. During the course of the treatment, the use of a chelating solution of EDTA proved to be the most suitable and effective means of cleaning. The methodological approach to the problem allowed us to make significant observations and to draw important conclusions that can be applied in the removal of the cerussite layers that have been formed on lead and silver objects under a burial environment. Keywords: silver hoard, Samos, conservation, EDTA chelating solution, cerussite

followed daily in a metals conservation laboratory as described in the literature. After partial failure of the initial cleaning procedures, chemical analysis identified cerussite, which created the basic problem for cleaning the surface of the silver coins. This corrosion product encouraged the adoption of a new, gentle approach to cleaning operations, which was entirely successful and mostly gave useful conclusions for cases such as this.

Introduction During the 21st Ephorate’s archaeological research on Samos conducted in October 2001 a hoard of twenty seven silver coins was found inside an undecorated lead pyxis. This hoard was recovered above the level of the burials in the Necropolis of the Geometric period situated in what is nowadays Pythagorio and specifically below the stone-paved floor of a room that is possibly related to the complex of a gymnasium of the Hellenistic period. An Achaemenid-type lead flask (phiale) was used as the lid of the pyxis (Viglaki-Sophianou and Touratsoglou 2010, 283). The hoard was probably made in Karia, considering that the minting of Mausolus’ coins was done in workshops found at the narrow territory of Ekatomnedes. It was probably hidden at the end of the 4 th or early 3rd c. BC by an exiled citizen who was repatriated after the law that was imposed by Alexander the Great in 322 BC (Viglaki-Sophianou and Touratsoglou 2010, 287) (Fig. 1).

Initial state of preservation After macroscopic and microscopic observations and various measurements that took place at the workshop for all the individual coins as for the pyxis and the phiale, the following observations emerged: (a) All the coins were made of silver. They were completely covered by a particularly thick and hard layer of whitish-grey corrosion product with inclusions of calcite and quartz agglomerates from the burial environment. The depictions on both sides of the coins were in most cases impossible to identify. It was estimated that they were obscured by the chlorides and/or carbonate salts of silver and copper, while in some of these areas subtle green deposits were observed. The weight measurements that followed confirmed the existence of metallic core.

In the same year, the coins together with the pyxis and the phiale were transported to the Metals Conservation Laboratory of the Archaeological Museum of Vathy in Samos in order for the necessary conservation procedures to take place (Figs. 2, 3). The results though were far from adequate. The burial conditions and the long-term contact of the two different metals had resulted in the formation of a unique corrosion product for both the lead pyxis and the silver coins: cerrusite. This corrosion product was ‘unexpected’ in terms of the usual theoretical approaches and therefore the cleaning procedures

(b) The pyxis was made of lead. A substantial part of the ‘body’ was missing around the rim, so the initial height of the pot is not known. The surviving part had a diameter

101

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry of 71 mm, its walls are 1 mm thick and showed strong distortions, partly because of the ductility and thickness of the construction material and of the intense mechanical stresses possibly brought about either from the environment or human burial agent. The flask with a diameter of 56 mm and, a thickness of 2.5 mm, has also been distorted and is decorated both internally and externally. The corrosion products appeared to be lead carbonate.

aid of an ultrasonic bath, also with minimal results. Electrolytic cleaning (Cronyn 1990, 233; Stambolov 1985, 186; Plenderleith 1956, 189) was tested on another silver coin, which had better results than the previous process, but in no case was it sufficient, and essentially highlighted the need to search for a more successful cleaning technique.

Analyses of samples and results Sampling was performed in order to identify the metal alloys of the artifacts but also of their corrosion products and to tackle the problematic cleaning issue. Specifically, (a) a sample of minimum dimensions was taken from a detached fragment of the pyxis, (b) a sufficient amount of corrosion products from the pyxis was collected and (c) a sufficient amount of corrosion products of the silver coins was removed and together with the lead samples was examined in order to obtain all the necessary information concerning the chemical composition, and furthermore to revise the conservation interventions. The analytical methods used in this study were Scanning Electron Microscopy (SEM) coupled with EDX and X-ray diffraction (XRD) (Fig. 4).

Theory and Conservation Interventions The methodology of conservation interventions that followed was based on the theory of corrosion concerning objects made both of silver and lead, which are retrieved from long-term burial conditions. Specifically: (a) Concerning the lead pyxis and the flask coming from the burial environment in the presence of air and moisture, a layer of cerussite (PbCO3) or hydrocerussite (PbCO3.Pb(OH)2) was expected (Cronyn 1990, 202, 203). It had a protective effect on the surface of the artifacts. Mechanical cleaning (with scalpel, needle etc.) is not recommended for the removal of these corrosion products as it is very likely to cause scratches on the soft lead surface. Chemical cleaning is recommended with caution although these corrosion products are very hard and insoluble in several chemical solutions. Moreover, the danger of developing aggressive etching on the metal itself always exists. Thus it was decided to follow the method of electrolytic cleaning as the procedure is highly controlled (Cronyn 1990, 207) (Fig. 6: 3 and 4).

Analyses were performed at the research laboratories of IGME: (a) qualitative SEM analysis for alloy identification was applied to the fragment from the body of the pyxis and (b) analysis to identify the chemical composition of the corrosion products by XRD, which was applied to (i) a sample from the corrosion products of the pyxis and (ii) samples of the corrosion products of three different coins. The results of the analyses for each case testified that (a) the pyxis is made of pure lead, (b) the corrosion products of the pyxis (No. 14746) is cerussite and (c) the corrosion product of the silver coins (No.14753, 14764, 14748) is also cerussite (Fig. 5).

(b) Concerning the silver coins coming from the burial environment in the presence of air, moisture and soluble salts (Cronyn 1990, 202; Selwyn 2004, 119; Stambolov 1985, 167), a layer of silver chloride (AgCl, chlorargyrite or cerargyrite) or silver carbonate on the surface (Ag2CO3) was expected. It is also possible that the occurrence of a green layer of weak copper corrosion products within certain areas is usually due to a low content of copper in the alloy of silver. Along with these corrosion products the coexistence of calcite and quartz inclusions from the burial environment (Cronyn 1990, 231, 232) is expected.

In addition to the analyses, a series of chemical spot tests1 (Gilroy and Godfrey 1998, 173-175) were done at the Conservation Laboratory of Vathy, Samos, on the metallic core of three coins which certified the detection of silver and copper in the alloy. During the application of this method an electrical closed circuit was created. A specific indicator immersed in a K2CrO4 solution confirmed the presence of silver, and accordingly a specific indicator immersed in distilled water confirmed the presence of copper.

The use of organic acids such as ammonium thiosulfate ((NH4)2S2O3) is recommended for the removal of the corrosion products. Furthermore, the use of other acids or chemical pads is likely to initiate further corrosion to the object (Cronyn 1990, 233). In this case the use of electrolytic cleaning is adopted, which is considered a mild intervention for the removal of corrosion deposits from the surface (Plenderleith 1956, 223).

All of the tests in general showed that (a) the pyxis is made of pure lead and the corrosion products consist of cerussite as originally estimated, (b) the coins are made of silver and copper as initially estimated. Despite the initial macroscopic estimations, where silver corrosion products were expected for the coins microscopic analyses have indicated only the presence of lead corrosion products; the same corrosion layer existed on the surface of the lead pyxis.

Electrolytic cleaning (Cronyn 1990, 207; Stambolov 1985, 169, 170) was tentatively applied to fragments of the lead pyxis using a 10 %wt. sodium hydroxide (NaOH) solution. The cleaning process did not provide the expected results. 10 %wt. ammonium thiosulfate ((NH4)2S2O3) was then tested on a silver coin, with the

1

These were chosen because they are non-destructive. Furthermore, it was not possible to transfer the coins outside the Museum of Vathy for analyses.

102

I. Staikopoulos, The conservation of a silver coin hoard being a more energy reactive metal than silver, has eroded at a faster rate than silver, which has less energy potential to form corrosion. The whole process, namely the creation of a cerussite layer on the surface of silver coins, grew in this particular way because of the coins’ long-term contact to the pyxis and phiale in burial conditions. This resulted in the development of a local galvanic cell potential energy between the two metals and the corrosive burial environment (Fig. 7).

Final conservation interventions These observations and the data of the original conservation interventions resulted in the adjustment of the methodology for the cleaning operations. For each coin specifically, it was decided to adopt a gradual, but ultimately overall implementation of the cleaning interventions that were tested initially, adding as a final step the application of the salt of a weak organic acid. In this way, the accelerating action of the cleaning solutions would be progressively achieved and the desired result would be reached, which was simply the disclosure of the shiny surface of the silver coins. Therefore, with regard to Figs. 8 and 9:

The combined use of different cleaning processes was necessary in order to achieve the required result. The chemical cleaning with ammonium thiosulfate and the subsequent electrolytic cleaning with sodium hydroxide solution for each coin separately was proved to be the most effective method due to the increase of solubility, which made its removal easier.

(a) The following sequence of interventions was followed for each coin: (1) immersion in a 10%wt ammonium thiosulfate aqueous solution, (2) removal of the coin from the solution and rinsing with distilled water, (3) electrolytic cleaning of the same coin using 6V voltage and sodium hydroxide solution as electrolyte, (4) removal of the coin and rinsing while using liquid soap (Texapon), glass bristle brush and distilled water, (5) immersion in a solution of disodium EDTA 10% wt., while heating at 60°C to achieve the formation of a chelating solution2 and (6) simultaneous use of distilled water and neutral soap. After the completion of these interventions all the coins exhibited the expected shiny surface. The coins followed a series of water baths, while conductivity of the rinsing water was measured every time, desiccating at 40°C and finally a protective coating with Paraloid B 72 in acetone was applied.

Immersion in aqueous solution of EDTA, a chelating agent that can form stable complexes with almost all of the metal cations, proved to be another effective cleaning method. The stability of these complexes is due to several binding sites of the molecule of EDTA (lone electron pairs of nitrogen atoms), which is a typical multiplex (or polydentate) substituent. During the application of the chelate solution of EDTA, the corrosion product of cerussite was not removed from the coins but it was transformed into a loose layer/epidermis, which was then easily flaked off the surface by mechanical means, completely revealing the metallic silver surface.

(b) Respectively for the lead phiale and pyxis the interventions include: (1) electrolytic cleaning, using 6V voltage and sodium hydroxide solution 10 wt% as electrolyte, (2) removal of the object, washing with tap water and simultaneous use of neutral soap and glass bristle brush, (3) dipping in a solution of disodium EDTA 10% wt, while heating at 60°C, (4) removal of the object , washing with tap water and simultaneous use of neutral soap and glass bristle brush, (5) baths with tap water and continuous measurements of the rinsing water, (6) desiccating at 40°C, (7) application of a protection coating of Paraloid B-72 in acetone and finally, (8) gap filling of the pyxis body with a twocomponent epoxy resin, matting agent and earth pigments of a similar hue.

The initial chelating solution alone did not give satisfactory results as a cleaning agent for the silver coins or the lead pyxis with its lid. In order to achieve the desired result one has to combine all the afore-mentioned techniques following the methodology that was described above.

Conclusions: Description of the Coins After the completion of the conservation work the following information is obtained: (a) All the coins in the hoard were made by the ‘flipping’ method; they have specific weights and diameter. (b) The center of the internal surface of the phiale was decorated with a sitting female figure. Over her left shoulder is a male figure who possibly represents a slave. The coins were examined by Ioannis Touratsoglou who provided information regarding their provenance, the minters’ names and the year of minting. This information is summarized below. Since 2009 the hoard has been on exhibition at the archaeological Museum of the Pythagorion, Samos.

Conclusions: Conservation After the completion of the conservation procedure we have been able to draw some conclusions concerning the formation of the corrosion layers as well the effectiveness of various cleaning methods for its removal from excavated objects. The presence of cerussite on the surface of the silver coins is explained by the rate of corrosion that is determined from the difference in nobility between silver and lead. That can be expressed as a difference in voltage potential (Singley 1988, 30). Lead,

2



http://www.chem.uoa.gr/chemicals/chem_EDTA.htm

103

Nine silver didrachma of Samos (NM 649, NM 648, NM 650-656). Observe: lionhead. Reverse: semisection of bull, national of ΣΑ (Samians), olive branch. Minters: ΛΕΟΝΤΙΣΚΟΣ 321-312

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry













BC. ΒΑΤΤΟΣ (306-301 BC), ΑΛΕΞΗΣ (250 BC) Two silver didrachma of Magnesia (NM 657658). Observe: armed horseman. Reverse: bull, national of ΜΑΓΝΗ (Magnesians). Minters ΠΙΤΘΙΩΝ (350-190 BC) ΔΙΟΠΕΙΘΗΣ (350-325 BC) One silver didrachmon from Ephesus (ΝΜ 659). Observe: bee, national of ΕΦ (Ephesians). Reverse: palm tree and semisection of deer. Minter: ΑΡΙΣΤΕΥΣ (late 4th c. BC before 322 BC) Two silver drachmas from Miletus (ΝΜ 660661). Observe: laurel-crowned head of Apollo. Reverse: lion, star national of M (Miletians) and I (Ilesianns). Minters: ΠΡΟΞΕΝΟΣ (352-325 BC.) ΔΙΟΠΟΜΠΟΣ (352-325 BC) Seven silver drachmas of Priene (ΝΜ 662, ΝΜ 665 – ΝΜ 670). Observe: helmeted head of Athena. Reverse: trident, national of Prienians. Minter: ΛΥΣΑΓΟ (334-330 BC or 290-250 BC) Two silver didrachma of Priene (ΝΜ 664, ΝΜ 663).Observe: helmeted head of Athena. Reverse: trident, national of ΠΡΙΗ [Prienians]. Minters ΛΥΣΑ (334-330 BC), ΛΥΣΑΓΟ (334330 BC or 290-250 BC) Four silver drachmas of Miletus (ΝΜ 674, ΝΜ 671- ΝΜ 673). Observe: laurel-crowned head of Apollo-Sun. Reverse: Zeus wearing himation. Satrap ΜΑΥΣΩΛΛ (Mausolus) (377-353 BC)

granting of approval for the publication of this article, colleagues Konstantinos Alexiou for translating the text into English, Stefania Chlouveraki for editing and finally colleagues Maria Bia and Eleni Roditaki for image editing. The analyses at the Greek Institute of Geology and Mining Research (IGME) in Athens were carried out by Dr. G. Oikonomou to whom sincere thanks for the cooperation and patience are owed.

References Cronyn J.M., 1990, The Elements of Archaeological Conservation, London: Routledge. Gilroy D. and Godfrey I. (eds.), 1998, Conservation and Care of Collections, Western Australia Museum. Plenderleith H.J., 1956, The Conservation of Antiquities and Works of Art, Oxford: OUP. Selwyn L., 2004, Metals and Corrosion, Canadian Conservation Institute. Singley K., 1988, The Conservation of Archaeological Artifacts from Freshwater Environment, Lake Michigan Maritime Museum, Michigan. Stambolov T., 1985, The corrosion and conservation of metallic antiquities and works of art, Amsterdam: Central Research Laboratory for Objects of Art and Science. Viglaki-Sophianou M. and Touratsoglou I., 2010, The ‘hoard’ of Samos Gymnasium, 2001, in The Friends of Numismatic Museum, «Οβολός 9, Τα Νομίσματα στα Νησιά του Αιγαίου», Proc. Fifth Scientific Meeting (Mytilene, 16-19 September 2006), Mytilene, 283-299. http://www.chem.uoa.gr/chemicals/chem_EDTA.htm (visited: July 19, 2008).

Acknowledgements Special thanks are due to the archaeologist and excavator Maria Viglaki-Sophianou for her confidence and the

Fig. 1. The lead pyxis, the lead phiale and both sides of the silver coin hoard before conservation treatment.

104

I. Staikopoulos, The conservation of a silver coin hoard

Fig. 2. Achaemenid-type phiale (exterior) before conservation.

Fig. 4. Chemical characterization of the leaden pyxis in the scanning electron microscope (SEM) was performed with energy dispersive X-ray analysis (EDX).

Fig. 3. Achaemenid-type phiale (interior) before conservation.

Fig. 5. Crystalline phase characterization by XRD analysis identified the presence of cerussite (characteristic peaks are indicated with the letter C) as the corrosion layer at the surface of the silver coins.

Fig. 6 (left). Conservation treatments with ammonium thiosulfate (1 and 2) and electrolysis (3 and 4) on one of the silver coins were not adequate. Fig. 7 (right). Achaemenid-type phiale (interior) after conservation.

105

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

Fig. 8. The silver coins after the final conservation treatment. Part I.

Fig. 9. The silver coins after the final conservation treatment. Part II.

106

Chapter 15

Gold inlay decoration on the blade of an iron sword from the Tomb A at Katerini Vasiliki Michalopoulou1, Barbara Smit-Douna2 and Ioannis Karapanagiotis3 1

Archaeological Museum of Thessaloniki, 6 M. Andronikos, 54621 Thessaloniki,Greece School of History and Archaeology, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece 3 University Ecclesiastical Academy of Thessaloniki, Department of Management and Conservation of Ecclesiastical Cultural Heritage Objects, 65 N. Plastira, 542 50 Thessaloniki, Greece 2

Corresponding author: [email protected] Abstract The iron sword from Tomb A at Katerini preserves rare inlaid decorations on both sides of its blade. The inlays were made of gold and attached on the iron surface in an unusual way. In an attempt to reconstruct its manufacturing procedure, we examined the morphological characteristics preserved on the inlays and the substrate and analyzed their chemical composition. Keywords: Sword blade, inlaid iron, gold inlay, portable XRF, SEM-EDX

the inlays were attached to the iron blade. The blade was studied from a technical and analytical perspective in an attempt to enhance our knowledge about this rare find.

Introduction Tomb A of Katerini (Central Macedonia) was excavated in 1977 and dates to the second quarter of the 4 th century BC (Despini 1980). This tomb and its grave offerings are now being revisited. During the study of a heavily corroded iron blade, under the disfiguring corrosion products a small piece of gold inlay decoration was noted. The X-radiography revealed three parallel decorative zones running along the blade with miniature fish and aquatic birds on both sides of the blade (Fig. 1) (Smit-Douna et al. in press).

Methodology All aspects of the object and its inlays were examined visually and under the microscope. All the inlays from the one side and one fish from the other side of the blade have been numbered as shown in Fig. 2. They were analyzed by a portable XRF Fluorescence Analyser (Bruker, Tracer IV equipped with a Silicon Drift (SDD) detector) and Light Alloys FP software. In order to detect elements that could imply the use of soldering (lead and tin), the underside of the detached inlays and their corresponding substrate metal were also analysed by the portable X-Ray device. In addition, a thin layer found attached to the base of the recess no. 18 was analysed by SEM-EDX (Energy-Dispersive X-ray Spectroscopy) (Fig. 8).The surface composition and the microstructure of the fragment of the tail of fish no. 8 was also analyzed by SEM (Fig. 9). The analyses were performed at the laboratory of the University Ecclesiastical Academy of Thessaloniki.

Iron swords found in tombs of Macedonia are often decorated with gold sheets or other decorative figures/components, on their scabbard, hilt or on the cross-guards. There are no iron swords with gold decoration on theirs blades in Macedonia or elsewhere. An exception is a Sarmatian sword (5th or 4th century BC) decorated with quadruped animals and a deer sacrifice scene (Shemakhanskaya et al. 2009). From an iconographic point of view, as far as we are aware, blades decorated with animal friezes have not been found in Greece with the exception of the Mycenaean daggers (late 16th to the end of the 14th century BC); of course the technical aspects and the materials of the Mycenaean weapons’ inlaid decoration differ from those of the blade being examined here (Photos et al. 1994, 272, 273).

Results The blade was in fact an 8.7 cm long fragment with one end damaged. Its material, iron, is in mineralized condition. The inlays are preserved in good condition and in place, apart from three small fragments, one of them near the damaged area. More specifically at the upper and lower zone there are twelve well-preserved birds (nos. 16 and 12-17) and one being in fragmentary condition (no. 18 - missing body). In the central zone there are five fish (nos. 7-11) and a part of an empty recess for a sixth one.

Due to the rarity of the find, further examination was deemed necessary. In addition, intervention was kept to a minimum, being limited to the removal of the secondary corrosion products from the one side of the blade. The side that remains untreated is the one that preserved some residual of a mineralized fabric on the corrosion layer. The uncovered inlays raise questions concerning the manufacturing techniques used and in particular the way

107

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry The dimensions of the birds are between 0.7-0.8 cm, including body and tail, and the fish 0.8 to 0.9 cm without the tails. The maximum thickness is 0.04 cm. The feathers of birds nos. 14 and 18 were found detached, as was one part of the tail of fish no.8 that has been preserved (Fig. 5). All the front sides of the inlays are flattened (the birds are a bit concave) and polished. All side views that can be seen today are vertical or slightly concave. There are no signs of a cutting tool, such as striation lines, on any of the side or under views. When viewed from their undersides, fish no.8 and no.19 have a swelling around their bodies’ perimeter (Figs. 6, 7), while fish 19 displays yet another swelling where the tail meets the body (Fig. 7). The birds’ undersides that can be examined show two wires connected, especially in the case of bird no. 12 (Fig. 3) while it is a possibility in the case of bird no. 18 (Fig. 4).

metal by mechanical means a metallurgical bond between the two is needed as otherwise the inlay would fall out. For the joining of dissimilar metal soft-soldering and hard-soldering alloys (brazing) operations have been used. Τhe soft soldering involves the use of low-melting alloys (up to 350oC) and those were the alloys of tin and lead (Maryon 1971, 6). The hard soldering uses alloys with high melting point (over 350 oC). As tin was not found in the analysis the bonding of the gold inlays to the iron base must be sought in the ability of the two metals to form a bond (thus to be bondable). In this case, the inlay’s gold alloy is the hard solder (Maryon 1971, 158). As far as we know, it is unknown when this property was discovered. The earliest mention occurs in medieval texts (G. Della Porta 1658 Magia naturalis in Scott and Eggert 2004, 28). As there is no similar paradigm to compare to, we are forced to believe that the craftsman either knew or experimented in the issue of attaching gold to iron. Our observations follow this hypothesis in order to determine, to the extent possible, which procedure was followed.

The inlays consist of a gold alloy with a small amount of silver (or with no silver at all in some cases) and also a small but significant amount of copper (Table 1). From EDX analysis of the fragment of fish no. 8 the gold alloy was found to be very pure gold (99%) with a small amount of copper (0.3-0.7%) and no silver. The iron from the bulk of the blade contains titanium, copper and lead, elements of the initial ore which ancient smiths could not extract (Varoufakis 2010, 84; Scott 2011, 15-16). The elements tin or lead were not detected either on the underside of the inlays or on the thin layer found on the base of the recess 18. There was no tin on the iron substrate, while the detection of lead must be attributed to the initial ore as it also appears on the bulk of the blade (Table 2).

The morphology of the recesses as can be seen today in the form of the iron corrosion products in the imprint of the gold inlay with which it was in contact (Figs. 3-5). The reason for that is the accelerating galvanic corrosion that is expected to have started almost immediately after the decoration was completed in which the iron acts as the anode (Selwyn 2004, 30). Thus, as it appears from the undersides of the birds and their imprints as they had been formed on the mineralized iron surface, they must have been manufactured with the attachment of parts of thin wire, 0.2 cm in diameter (Figs. 3, 4). The fish would have been made by one wire starting from the upper stem of the fork tail continuing along the low part of body’s perimeter folded in two when reaching the mouth and turning to the opposite direction until the lower stem of the tail is reached, with the attachment of a part that was cut in the shape of the body to which a Vshaped wire for the tail was jointed (Figs. 6, 7). The welding seams should have been preserved in the underside. The fact that we cannot detect that is an indication that during the manufacturing procedure, high temperatures were used which were able to bring the surface into a liquid or semiliquid state. Similarly, it can be presumed that the line in the center of the inlay’s feathers of no. 18 must be a welding seam that due to the application of heat dies out toward the two ends and remains visible only at the middle of the fragment (Fig. 4).

Discussion The blade was made from wrought iron which is a relatively soft and malleable metal that can easily be embossed when heated. The recess and the lines that form the three zones are smooth without traces of engraving so they had to be made by chasing and punching. Punching requires a special tool which has its base designed in a suitable shape (one for the bird and another for the fish). The inlay material is a very pure gold alloy that is soft and can be easily formed even cold. In order to hold the inlay in place, the ancient craftsman created an interlock either by reverse bevelling the recesses’ walls or by passing the cutting tool twice to create a W-shaped profile on the recesses’ base. Then he filled the recess using the inlay metal in the form of wire. When a soft metal wire is forced to enter a recess that had its walls reverse bevelled (so it is wider at the base), it fills the recess’s undercut areas and becomes wider at its bottom and narrow towards the front (exposed surface).

As both sides of the blade are decorated, the sword must have been heated at least twice. When a number of sequential joints are needed the soldering operation starts with the alloy of the higher melting point and continues with the alloy with the lower melting point so as not to damage the bonding of the first operation (Demortier 2004, 499-500). The inlay composition as determined with non-destructive techniques gives very useful information about the technical requirements of the procedure and the craftsman’s choices, but it is not representative of the original composition especially

In cases where the base of the recesses had been marked (W shape) the inlay material will insert into and interlock with the base metal (Maryon 1971, 158, 159). In this case a protruding part is formed at the inlay’s underside. None of these figures are shown on the inlays that had been examined. When the inlay is not attached to the base

108

V. Michalopoulou et al., Gold inlay decoration on iron blade when the manufacturing procedure involves heating. Apart from the known phenomenon of surface depletion of the alloy elements due to corrosion in a burial environment, the soldering procedure is hiding its special figures. The amount of the alloy element that drastically reduces the melting temperature, that is copper, is eliminated during heating, mostly when repeated heating operations are performed (Ruvalcaba 2004, 47). So the original composition cannot be detected unless examined metallographically; consequently the melting point of the original alloy cannot be determined or how many heating operations took place.

like to thank the director of the Archaeological Museum of Thessaloniki, P. Adam-Veleni, for her support and permission for the necessary interventions as well the conservator Paraskevi Bilali for the x-ray of the fragment, which was taken at the Conservation Workshop of the Ephorate of Paleoanthropology and Speleology of Northern Greece in Thessaloniki.

References Despini Κ., 1980, Ο τάφος της Κατερίνης, ΑΑΑ 23, 198209 (in Greek). Demortier G., 2004, Precious metals artefacts, in K. Janssens (ed.) Non-Destructive microanalysis of Cultural Heritage materials, Elsevier, 493-564. Maryon H., 1971, Metalwork and Enamelling, London. Photos E., Jones R.E. and Papadopoulos Th., 1994, The black inlay decoration on Mycenaean bronze dagger, Archaeometry 36 (2), 267-275. Ruvalcaba L.J., Torres L., Franco F. and Diaz O., 2004, Artifacts rich gold surfaces: depletion gilding or natural surface corrosion? Study of corrosion and oxidation of gold alloys, in A. Perea, I. Montero Ruiz and Ó. García-Vuelta (eds.) Ancient gold technology: America and Europe, Madrid: Consejo Superior de Investigaciones Científicas, 41-48. Scott D. and Eggert G., 2004, Iron and Steel in Art: Corrosion, Colorants, Conservation, London. Selwyn L., 2004, Metals and corrosion: A handbook for the conservation professional, Ottawa. Shemakhanskaya M., Treister M. and Yablonsky L., 2009, The technique of gold inlaid decoration on the 5th-4th centuries BC silver and iron finds from the early Sarmatian barrows of Filippovka, Southern Urals, in M. Filomena Guerra and Th. Rehren (eds.) Authentication and Analysis of Goldwork, ArcheoSciences & Revue d’Archéometrie 33, 211220. Tylecote, R.F. 1978, The Solid Phase Bonding of Gold to Metals, Gold Bulletin 11(3), 74-80. Smit-Douna B., Michalopoulou V. and Nazlis I., in press, Ο τάφος Α της Κατερίνης: ένα παλαιό εύρημα επανεξετάζεται, in Το Αρχαιολογικό έργο στην Μακεδονία και Θράκη, ΚΕ΄ Scientific Meeting, 1-3 March 2012 (in Greek). Varoufakis G.A., 2010, Μεταλλουργική μελέτη σιδερένιων συνδέσμων του ναού της Τράπεζας υστεροαρχαικής εποχής, Αρχαιολογία και Τέχνες 114, 83-85 (in Greek).

Alternatively, the blade could have been heated during both procedures without reaching the gold alloy’s melting point. For example, on the one side, the inlays could have been heated until red hot, possibly beyond that as well until reaching a liquid state (close to 1000oC), while for the bonding of the inlays on the other side the temperature need not have reached the same high levels and the inlays could have stayed in a solid state throughout the entire heating operation (solid phase bonding). In this case, success in bonding depends on how good the contact was between the surfaces of the two materials (Tylecote 1978, 80).

Conclusion According to one possible scenario, first the blade was constructed then the recesses were formed on both sides and then they were filled on the one side and the sword was heated until red hot. The same procedure was conducted on the other side. The choice of a gold alloy that contains copper, for the inlays’ manufacture, was most probably deliberate since it facilitated the operation by lowering the melting point and improving the gold’s wetting ability. Both qualities are known to ancient goldsmiths. The fact that the inlays display differences on their undersides reinforces the hypothesis of applying different ranges of temperature. It is also natural that these were due to the craftsman’s inability to fully control conditions (temperature) as well as technical details such as the uniform distribution of the inlay within the recess.

Acknowledgments We would like to express our warm thanks to A. Despini who was kind enough to bestow to B. Smit-Douna the final publication of the tomb and its finds. We would also

109

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry Table 1. Results from the XRF analysis of the inlays (the high iron amount is from the inlay’s surrounding area). Inlay 1 2 3 4 5 6 7 8 front side 8 underside 9

Fe (%) 70.20 64.00 80.60 71.00 82.30 75.00 78.80 1.59 10.80 77.40

Cu (%) 2.36 2.74 1.66 2.52 1.68 1.93 0.91 5.79 5.49 0.96

Ag (%) 0.51 0.67 0.24 0.52 0.29 0.40 0 0 0 0

Au (%) 25.70 31.50 15.90 25.10 14.60 21.40 19.60 90.50 81.60 20.90

Inlay 10 11 12 13 14 15 16 17 19

Fe 86.40 85.20 81.60 60.60 75.50 69.60 72.00 62.00 50.00

Cu (%) 0.82 1.01 1.74 2.94 1.94 2.35 2.64 2.86 2.59

Ag (%) 0 0 0.35 0.88 0.42 0.54 0.52 0.70 0

Au (%) 11.30 12.30 15.10 34.80 21.30 26.60 24.00 32.70 46.10

Table 2. Results from the XRF analysis of the iron blade.

Blade Core Substrate area of recess no. 18 Substrate area of recess no. 8

Ti (%) 0.36 0.16 -

Fe (%) 78.30 98.10 97.40

Cu (%) 0.52 1.07 1.54

Pb (%) 3.31 0.14 -

Si (%) 10.30

Al (%) 3.31

-

Fig. 1. X-radiograph of the sword blade showing the inlaid birds and fish on both sides.

Fig. 2. The sword blade.

Fig. 3. The inlay 14 showing the feathers’ underside and the recess morphology.

Fig. 4. The inlay 18 showing the feathers’ underside and the recess morphology.

110

V. Michalopoulou et al., Gold inlay decoration on iron blade

Fig. 5. The recess in 8.

Fig. 6. The inlay in 8 (underside).

Fig. 7. The inlay in 19 (underside).

Fig. 8. SEM image of the layer from the recess in 18.

Fig. 9. SEM image of the inlay 8 (underside).

111

Chapter 16

The so-called ‘helicoidal’ ore washeries of Laurion: their actual function as circular mills in the process of beneficiation of silver and lead contained in old litharge stocks George D. Papadimitriou Professor Emeritus, Laboratory of Physical Metallurgy, National Technical University, Athens, Greece Author’s address:[email protected]

Abstract The so-called ‘helicoidal washeries of Laurion’ are actually circular edge runner mills, probably derived from the olive crushing mill trapetum (in Latin). They were used for fine grinding of litharge abandoned earlier near furnaces as waste, in order to recover its silver/lead inclusions by hydromechanical concentration in the rectangular washeries. They were invented probably in the early 3rd century BC and their operation in Laurion lasted until the middle of the 1 st century AD. Keywords: Laurion, circular mill, edge runner mill, helicoidal washeries, silver metallurgy, litharge recycling

installations were initially closed, but some blocks of marble were removed under some circumstances. Devastation of ancient monuments is not unusual, as they are subjected to physical wear and human interventions over time, such as for example the damage inflicted on the mining installations during the slave revolt, sometime between 150 and 100 B.C. (Athenaeus, Deipnosophistae VI, 272 e-f). On the other hand, there are striking similarities between open and closed constructions: apart from their similar shape and dimensions, they are usually found near places with metallurgical residues (fragments of furnace walls and slags) rather than being associated with mines, as is the case with ‘rectangular washeries’. Therefore, open and closed installations should be considered as identical. The initial hypothesis of helicoidal washeries should be accordingly abandoned, since the remains of installations confirm the absence of slope, which is necessary for water flow. The slope observed at Demoliaki is insignificant and might be caused by local subsidence. Furthermore, the joints between blocks are not sealed, as was the strict practice for all constructions of Laurion, in which water was involved. In consequence, these installations were not used for hydromechanical concentration, but for another dry process. The original hypothesis was also critically reviewed in a paper of Rehren et al. (2002). The authors did not refute the hypothesis of helicoidal washeries, but agreed that they were probably an early trial, which was “soon recognized as an error and abandoned as too expensive and not sufficiently productive”.

Introduction The so-called ‘helicoidal washeries’ were firstly described by C. Conophagos and H. Mussche (1970) in a communication concerning three circular structures in the mining area of Laurion (Megala Pefka, Dimoliaki, Bertseko). They were made from carefully fitted marble blocks with a circular trough on their upper surface, having a diameter of about six metres. The trough at Dimoliaki and Bertseko consists of a succession of roundish cavities, like shallow bowls. These installations and a fourth one discovered later at Megala Pefka, were greatly disturbed, with a part of their periphery missing (Fig. 1). Conophagos and Mussche (1970), based on the presence of bowls and on the small slope observed in the trough at Demoliaki, assumed that the structure was an helicoidal sluice serving for the concentration of fine silver bearing ores of lead and identified them as ‘helicoidal washeries’. K. Tsaimou (2004; 2006) excavated four similar installations at Ary of Lavreotiki, three of them (Ary II to IV) preserved in a good condition. These excavations are very important, because they unveiled closed circular structures without any missing stones and practically horizontal (Fig. 2). Tsaimou (2006; 2008) suggested that these closed structures were different from the open ones reported by Conophagos and Mussche and that they served as devices for mixing fine ore concentrates with clay and water, in order to prepare briquettes for feeding the smelting furnaces. Mixing would be performed by means of a sort of hoes moving in the circular trough.

As far as the hypothesis of the circular structures being used in the mixing of ore, water and clay is concerned (Tsaimou 2006; 2008), convincing arguments are missing. Mixing is an operation easy to achieve manually,

In my opinion there are no sufficient arguments to support that open and closed installations are different, on the contrary it is reasonable to accept that open

113

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry using for example a shovel and such a complex and expensive installation as that described by Tsaimou (2006; 2008) seems unnecessary.

many parts of the world, which have worked as edge runner mills for different agricultural applications: grinding of apples for preparing cider, grinding of oaks for leather tanning, breaking flax, threshing cereals, decorticating rice. Alternatively they have been used for fine grinding of ores of gold, silver and lead. They are known in China, India, Western Europe and in the American continent.

In the present communication I shall show that the round constructions cited above are circular grinding mills, belonging to a wider class of facilities actually known in technical literature as edge runner mills or Chilean mills. These mills were probably developed in Laurion and used for fine grinding of litharge, aiming at the recovery of its lead/silver inclusions by a subsequent concentration process.

Circular grinders are neither described in classical literature, nor mentioned by Agricola. In modern literature a discussion of this kind of grinders and a comparison between China and Europe is due to Needham (1965, 198-199, fig. 453). According to him, a circular mill of this kind is reported for the first time in a Chinese document of the 2 nd century AD. And the oldest picture is contained in a Chinese book of 1637 AD, showing a device decorticating rice or threshing cereals, with two millstones moved by oxen (Fig. 6). In the same book is also shown a photograph of a modern mill in Hupei-China, moved by an ox (Needham 1965, fig. 454).

The circular mills of Laurion Fig. 3 is a possible representation of the circular structure, which I shall call “the circular mill of Laurion”. It consists of a vertical millstone, freely mounted on a long horizontal axle which turns around a vertical shaft in the centre of the structure. The millstone revolves in a circular path within a trough, crushing the material which should be reduced in size. The mill is driven most probably manually or possibly by a donkey.

There are, however, several pictures in the Internet. In France remains of circular grinders from the Middle Age used for agricultural purposes exist mainly in Bretagne and Normandy (Fig. 7). A painting by Cezanne is also showing remains of a circular mill in front of the entrance of a mine gallery. In England such devices have been used in mines for crushing ores (Wood 2012) and belong to the 19th and early 20th century (Fig. 8). In Austria a mill of unknown age, shown in Fig. 9, was probably used for grinding oaks in a forest (Gross 2012).

Strong indications in favour of the above hypothesis are traces maintained on the blocks of marble of some installations, proving that a vertical millstone was running in the circular trough of the construction. At first, the wearing action of the rotating hard millstone is evidenced from the change of the dimensions and shape of the trough. As a matter of fact, the depth of the trough is not the same in all installations: in some of them the trough is shallow, in others it is deeper and the cavities are hardly observable on its bottom or even completely absent. This suggests that initially the trough in all installations was constituted from a succession of shallow cavities, but the cavities disappeared progressively due to wear and were shaped into a deep trough. This is clearly displayed in Fig. 4, showing that the depth of trough and depth of bowls vary inversely.

In the USA circular mills were used for crushing apples for cider and a painting in the Metropolitan Museum of Art shows a mill in operation. In South America circular mills (Chilean mills) were used in mines, in particular in Chile until recent times. In Greece, apart from Laurion, they were probably used also in the ancient gold mines of Pangaion, where some marble blocks with curved troughs on their surface remain near slag heaps (Photos et al. 1987).

Of course, it cannot be excluded that some of the circular mills, which exhibit a shallow trough without bowls, were constructed hastily, aiming at sparing effort and time at the expense of milling efficiency. Furthermore, the wearing action of the hard millstone running on the much softer marble is evidenced from wear tracks adjacent to the trough, which are similar to the traces of chariot wheels left on ancient stone pavements (Fig. 5).

The origin of the circular mills of Laurion The circular mill of Laurion derived probably from the trapetum; for information, the trapetum was an olive grinder and is probably a Greek invention, since, according to archaeological findings, appeared first in Northern Greece. The earliest specimen is from Pindakas at Chios (Boardman 1959) and dates from the mid of the 5th to the early 4th centuries BC. Millstones belonging to trapeta have also been found in the town of Olynthos, dating between 433 and 348 BC (Robinson and Graham 1938). In Lavriotiki a trapetum dating between the 4th and 2nd century BC was found in excavations near Legrena (Lohmann 1994). In Italy, the trapetum was used extensively and found in Pompeii and elsewhere. The trapetum is a rotary mill with two hemispherical millstones moving in a stone mortar, around a cylinder in its center (Fig. 10).

Their formation should be attributed to local displacements of the mill structure during operation. Under the heavy load of the running millstone, some blocks of marble might collapse or shift outwards. Then, since the radius of the trajectory was constant, the millstone deviated out of the trough and formed a shallow track. This irregular situation occurred probably in a period shortly before the installations were definitely abandoned. But the strongest argument in favour of the proposed interpretation is the existence of similar structures in

114

G. Papadimitriou, The so-called ‘helicoidal’ ore washeries of Laurion It was first described by Cato the Elder in its book De Agri Cultura. Its name in Greek is unknown, unless it is the elaiotropion mentioned in a Greek inscription of 224 BC from Antiochia of Syria (IGL Syr 2 376) and in Geoponika, 6.1 of the 10th century AD. However, the derivation of the word trapetum is associated with the Greek verb ‘trapein’, which is found in Homer (Odyssey 8.125) and in Hesiod (Shield of Heracles 301), meaning, according to Hesychius, ‘treading the grapes’ in a vessel. According to Pliny the Elder (Natural History VII, 199), the trapetum was invented by Aristaios, son of Apollo, and this means that Romans recognized it as a Greek invention.

A particular characteristic of the Laurion circular mill is the presence of cavities in the trough. I believe that these cavities served for entrapping the finest pieces produced during grinding and protecting them from further grinding. As a matter of fact, the finest pieces created just under the millstone passage, fall down in the bowls through the voids of the mineral layer, always leaving the largest pieces exposed on the surface. This is equivalent to screening, i.e. separation of grains with different size, increasing the efficiency of the grinding process and preventing overgrinding. Overgrinding is very annoying and would make difficult the subsequent concentration process in the rectangular washeries, by forming slimes and mud. It should be emphasized that the circular mill of Laurion was the ideal device for grinding litharge, due to its high productivity and strict control of the fineness, due to the presence of bowls in the trough. The older crushing devices of Laurion (saddle querns of marble blocks and Olynthus mills) were certainly not effective in accomplishing this task. As a matter of fact, large quantities of litharge, of the order of 1 to 1.5 million tons, were probably abandoned in more ancient times near the furnaces as waste. Small quantities of litharge were certainly treated before the invention of the circular mills, as it is testified by the presence of ground litharge in some ergastiria associated to the mines, but certainly in limited quantities.

A comparison of the trapetum with the circular mill of Laurion shows that the latter derived from the former, after it was adapted for crushing minerals instead of olives. As a matter of fact, both devices display as a common characteristic the movement of a vertical millstone (edge runner) inside a circular trough and their single difference is the diameter of the trough. In the trapetum the trough is formed between the central column and the walls of the mortar, whereas in the case of the circular mill of Laurion, the trough became of much higher diameter and was displaced away from the central column. This modification was certainly imposed by the properties of the material treated. Minerals are much stronger than olives and need therefore a heavier millstone for efficient crushing. This could be achieved by increasing the dimensions of the millstone and hence the diameter of the trough. This development, increased also the capacity, made easier the feeding and emptying of the mill and allowed driving with animals. It was, indeed, a revolution in the technology of milling exhibiting industrial characteristics, since it transformed grinding to an almost continuous process.

Date of invention and operation period of the circular mills of Laurion Tsaimou, based on sherds from Bertseko and Ary, considers that the circular constructions of Laurion date from the late classical period and extend to Roman times (Tsaimou 2004; 2006; 2008). Based on historical observations I agree with this estimation. It is well known that the mines of Laurion declined after about 300 BC (Crosby 1950). Since the circular mill - despite its efficiency and productivity - is completely absent from ergastiria associated with mines, it should be concluded that the circular mill was invented later than 300 BC, i.e. after investments in mining had stopped.

The use of the circular mills of Laurion The use of circular grinders is indicated by mineral residues remained in their trough. Visual observation, metallography and spectroscopy show that these residues are litharge, partially transformed to lead carbonate by weathering. Heaps of finely ground litharge is also present inside and outside the installations. The same observations have been made previously by Rehren et al. (2002). Now arises the question why litharge was ground, since grinding is an expensive and time consuming operation and should only be justified if it was highly profitable. The answer to this question is contained in a previous communication (Papadimitriou 2012). Litharge which was previously produced during the process of cupellation, contained metallic inclusions of a silver/lead alloy. These inclusions attained sometimes 0.5 or even 1 mm in size and were certainly perceived by the ancient metallurgists, who tried to recover them. Their beneficiation required fine grinding of litharge (below 0.5 mm) and thereafter subjecting the ground material to hydromechanical processing, in order to separate the heavier metallic inclusions from litharge. This was certainly done in the rectangular washeries, which are always present next to circular mills.

The decline of mining activities certainly pushed the old metallurgists to recycling old litharge stocks and obtaining silver from its inclusions by fine grinding and then by concentration in rectangular washeries. This process had eventually the advantage that the concentrate could be sent directly to the cupellation furnace, without smelting. This technique was certainly known earlier than the end of the 4th century BC, its application was however limited due to technical difficulties. The bottleneck was in fine grinding, which should be performed in the low productivity grinding stones and Olynthus mills of the ergastiria. The following example shows the extent of the problem: one ton of primary ore with 20% lead and 2,000 g Ag/ton of lead contains 400 g of silver. For the equivalent silver quantity, 2.5 tons of litharge with about 0.015% Ag, needed to be recycled. This means that disregarding the degree of recovery - a quantity of litharge 2.5 times as much as the quantity of the primary ore should be finely ground in order to

115

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry obtain the same quantity of silver. Therefore a massive exploitation of litharge was feasible only after the circular mills were invented, at around 300 BC. The owners of furnaces were probably also the owners of the immense stocks of litharge abandoned as waste near their furnaces and practiced the recycling activities. Recycling activities are reported by Strabo (Geographica 9.1, 23) and stopped in the 1st century AD. At 40 AD. Pomponius Mela (De situ orbis 2, 3) states that Thorikos and Brauron were previously towns and are now only names.

Structures rurales et Societés antiques, Actes du Colloque de Corfou (14-16 May 1992), Paris, 81132. Needham J., 1965, Pounding, Grinding and Milling, Science and Civilization in China, 4.2, Cambridge : Cambridge University Press, 183-225. Papadimitriou G.D., 2012, Litharge: Waste or useful material? An intriguing material revisited, in N. Zacharias, M. Georgakopoulou, K. Polykreti, G. Facorellis and Th. Vakoulis (eds.) Proc. 5th Symp. Hellenic Society for Archaeometry 2008, Athens: Papazisis, 799‐819 Photos, E., Koukouli-Chrysanthaki Ch., Tylecote R.F. and Gialoglou G., 1987, Precious Metal Extraction in Palaia Kavala in Proc. Internat. Symp. Old World Archaeometallurgy (Heidelberg 1987), Der Anschnitt, Beiheft 7, 178-190. Rehren Th., Vanhove D. and Mussche H., 2002, Ores from the ore washeries in the Lavriotiki, Metalla (Bochum) 9.1, 27-46. Robinson D.M. and Graham J.W., 1938, Excavations at Olynthus Part VIII, Baltimore: Johns Hopkins Press. Tsaimou K.G., 2004, The sixth helicoidal washery, a first presentation, Mineral Wealth 131, 37-41. Tsaimou K., 2006, New archaeological findings in Lavreotiki, relative to ore concentration, in Proc. 11th Scientific Meeting of South East Attica, Spata, 467-478. Tsaimou K.G., 2008, New information on the concentration process of silver bearing ores in Ancient Laurion, in Proc. 12th Scientific Meeting of South East Attica, Kalyvia, 435-451. Wood St., 2012, Millstone images-edge runners, www.peakscan.freeuk.com/millstone_images57.htm, (visited November 14, 2012).

Acknowledgments I am indebted to the 2nd Ephorate of Prehistoric and Classical Antiquities of Greece and to its Director Mrs Eleni Andrikou for the permission to publish photographs from the circular structures at Laurion. I also thank Dr G.N. Dermatis, Stephen Wood, Willi Gross and Patrick Loison for their permission to use their photographs, as well as Cambridge University Press for the permission to reproduce a photograph from Needham (1965).

References Boardman J., 1959, Excavations at Pindakas in Chios, Annual British School Athens 53-54, 295-309. Conophagos C., and Mussche A., 1970, The helicoidal washeries of Laurion, Essays of Athens Academy 292, 2-19. Crosby M., 1950, The Leases of the Laurion Mines, Hesperia XIX, no.3, 189-312. Gross W., 2012, Eine Steinmühle in Bruckmühl, www.ottnang.info/bruckmuehl/muelstein/bild.htm (visited December 10, 2012). Lohmann H., 1994, Ein alter Schaffstall in neuem Licht: Die Ruinen von Palaia Kopraisia bei Legrena (Attika), in P.N. Dukellis and L.G. Mendoni (eds.)

116

G. Papadimitriou, The so-called ‘helicoidal’ ore washeries of Laurion

Fig. 1. Greatly disturbed open circular structure at Demoliaki. Photograph by permission of Dr G.N. Dermatis.

Fig. 2. Well-preserved closed circular structure at Ary IV.

Fig. 3. Representation of the operation of a circular mill (edge runner mill) at Laurion.

Fig. 5. Tracks formed outside the trough by the wear action of the millstone on the marble blocks.

Fig. 4. As depth of trough increases due to wear, the cavities become shallower and finally disappear. The troughs shown are from Bertseko, Ary III and Ary II, respectively.

117

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

Fig. 6. Edge runner mill from China. From Needham,1965, Science and Civilization in China vol.4.2, fig.453, 198. By permission of Cambridge University Press.

Fig. 7. Grain mill at Blandy les Tour, France. By permission of Conseil Général de Seine et Marne and Patrick Loison.

Fig. 8. Edge runner mill for grinding lead ore. Odin mines, England. By permission of Stephen Wood, U.K.

Fig. 9. Edge runner mill at Ottnang, Austria. By permission of Willi Gross.

Fig. 10. Representation of the Greek Trapetum, probably the ancestor of the circular mill of Laurion.

118

Chapter 17

Technical and chemical examination of a small silver treasure dated to the 6th–7th century AD D. Kotzamani1 and J.F. Merkel2 1

Metals, Glass and Bony Materials Conservation Laboratory, Benaki Museum, Athens, Greece 2 Institute of Archaeology, University College London, London, United Kingdom Corresponding author: [email protected]

Abstract Eight plates from the Benaki Museum collection were technologically and chemically studied. Hammering was the major forming technique facilitated by cold working and annealing, and final polishing was completed on the lathe. The base was made independently and attached to the body with a hard solder while decoration details were produced by chasing. Silver gilt present on three dishes is characteristic of the mercury gilding process. The metal used is a silver alloy with high purity. Copper was added deliberately in order to increase its hardness and workability. Argentiferous ores seem the most possible ore type from which the Benaki silver was extracted. Provenance studies revealed North West Turkey as the provenance area as well as Laurion in Greece. Keywords: Early Byzantine, silver, metallography, Scanning Electron Microscope analysis, Lead Isotope Analysis, hammering, fire gilding, argentiferous ores

which led to several difficulties in handling (Kotzamani 1999).

Introduction This project involves the technological and compositional study of eight silver dishes dated to the early Christian period (6th-7th century AD). The small treasure, belonging to the Benaki Museum of Athens (Fig. 1), forms a part of a closely related group with central medallions standing on a low foot ring (base). Seven of the eight plates are closely related in size but not in iconography. Five of these are decorated with erotes accompanied with dolphins, while the other two just with a cross. Regarding the last plate, neither the size nor the iconography is compatible. Nevertheless, it is included in this group since it shows similar stylistic characteristics with the five dishes with eros.

Forming The initial estimations of the fabrication were somewhat confusing. According to the decreased thickness of the sides and rim in relation to the bottom, it was presumed that the plates were planished (hammered on the front). While comparing the wall thickness to the rim, which is thicker without any solder or folding indications (Lang et al. 1977), it was suggested that the plates were turned over and the rim was formed by knocking down (hammering), although it could have been strengthened on a lathe (Ward 1993). On the contrary, the centering mark observed on the bottom (at the exterior of the plates), some minute pores at the more solid areas revealed by radiography, the fine parallel lines and the forms of pointed and feathery 'pine trees' on the surface similar to dendrites, could indicate either casting where the molten metal has cooled slowly (Lang et al. 1977) or forming and final polishing on the lathe (Hodges 1976, 74-75).

Analytical procedure Technical examination (fabrication, joining and decoration) of the plates was carried out by stereoscopic and metallographic examination, micro-hardness tests (Vickers scale), physical dimensions measurements, Xray radiography as well as Scanning Electron Microscopy (SEM). Chemical analysis results including compositional variations between the surface and bulk were obtained by X-ray Fluorescence Spectroscopy (XRF) and SEM-EDS. Finally in order to obtain more precise clues about the plates’ provenance, four of them were subjected to Lead Isotope Analysis (LIA).

Therefore, eight samples (0.5-1 mm in size) were extracted from the more 'healthy' areas at the contour sides and subjected to cross-section examination and micro-hardness tests. The microstructure is an alpha silver-rich phase typical of hammering with recrystallized grains having great variability and annealing twins left in the re-crystallized condition (Fig. 2). Strain lines were observed only on the sample taken from the largest plate. This artifact required over working stages, while localized deformed and smaller sized crystals, revealed at the edge near the surface of the sections, are

Their planning and results were substantially determined by the thick and compact corrosion layer, indicating several degrees of preservation of the underlying metal and possible technological evidence; the weak silvery appearance either caused by the type of corrosion or due to post-excavation treatments as well as silver brittleness

119

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry related to more extensive working on the area due to polishing. By contrast, the body samples obtained from the plates with the cross showed a microstructure left in a deformed condition. A similar re-crystallized picture was given for seven samples cut from the foot of the plates but with slightly larger grains indicating a separately made part and a limited degree of cold working and annealing (Fig. 3). The same forming estimations were made according to the micro hardness values for the body (Hv average 42) and base (Hv average 68) with reference to the ones given by Scott (1991, 82) for cast pure, work hardened and annealed silver alloy (20%Cu/80%Ag), although the results in several cases have been influenced by internal corrosion.

according to its amounts (>0.5%), is almost certainly added deliberately to silver in order to increase its hardness and workability. Copper as a residue in silver from parent ores would not exceed 0.5% while refining which was practiced in those periods could easily have produced silver containing traces of copper (Cowell and La Niece 1991). Furthermore, if silver was incorporated as an economic necessity or monetary policy, the copper content would be higher (Moorey 1985). The other elements detected (Fig. 8a-d) were attributed either to the primitive sources of silver and copper or to the smelting and re-melting processes as well as the burial environment. Lead and bismuth (Fig. 9) at the 0.3-0.8% and 0.1-0.5% level respectively are related to silver exploited from argentiferous ores mainly galena, depending to native silver, dry ores, jarosite ores (Moorey 1985; Craddock 1995; Stos-Gale 1981) and deliberate additions (in the case of lead, Craddock 1985). Such concentrations suggest also extraction processes with more efficient reduction methods, known from the Roman periods since earlier smelting furnaces could not reach temperatures higher than 950-1200oC (Conofagos 1980, 311-327; Hughes and Hall 1972).

As far as the method of joining the body to the base is concerned, hard soldering seems the most probable technique in relation to autogenous welding or mechanical joining. Its application was recognized on two plates by the slight flowing of the solder around the attachment area and its silvery colour as well as its dark characteristics given by some radiographs (Fig. 4). Decoration The Benaki’s collection bears designs and gold plating. The designs are either low relief figures or decorative motifs such as floral, lines and crosses. In the first case the plates were turned upside down and the figures were drawn down on the front thus producing hollowing on the back. Then the plate was turned again and the background of the design was driven out leaving the figures in relief. The whole procedure and final details were completed by chasing tools due to the characteristic subsidence of the metal surface left by them (Fig. 5).

Argentiferous ores and in particular galena compared to the other silver sources (Craddock 1995; Gale and StosGale 1981) are supported also by the detection of gold between 0.0-1.0%, combined with the high amount of lead (Schubiger et al. 1976) as well as the absence of arsenic and antimony, which are more related to jarosite ores (Craddock 1991). On the other hand, zinc (2 cm, the thickness of the analyzed fragments of San Vincenzo pottery shows a very standard value around 0.9 cm. This suggests that at San Vincenzo there is no evidence for the use of this kind of large thickwalled vessels for water storage; therefore, we have to argue that another kind of storage system was used or that in the surrounding area there was a spring or an exploitable water table.

Table 1. Aeolian Islands: Relative size, presence and type of natural springs. A: Mammana (2006), The flow rate of the Schicciola (dripping water) is too low to be measured. B: authors’ observation. C: IGM topographic map. Island

Sq. Km

Lipari

37.6

Vulcano

21

Salina

26.8

Panarea

3.4

Stromboli

12.6

Filicudi Alicudi

9.5 5.2

On Stromboli natural source of water is to be found along the east coast of the island. One of these localities is known as La Schicciola (dripping water) since the water drips from a rock face. La Schicciola is located about 200 m above sea level, just below the sandy valley Rina Grande, overlooking the sea (Fig. 3, 8). The locality name is mentioned at the end of the 19th century, in the Nuovo Giornale Botanico Italiano (March 16, 1899), and in association with the collection of some species of plants that grow locally and in a wet environment (e.g. Lunularia cruciata).

Natural springs 2 main springs, 250 ltr / hr; 100 ltr / hr A Schiocciola Schicciola S. Onofrio, Schicciola U Puzzo A None reported A La Scicciola B, schicciola Pozzillo B, schicciola (?) La Petrazza C Schicciola Funtana B None reported B

A similar locality with water dripping from a rock (another ‘schicciola’) is to be found in the ‘Pizzillo’, near the present-day harbour (Fig. 3, location 6). Again the morphology of the area is strongly influenced by volcanic activity and the action of water. A deep incision flows in a small valley in which today is placed a terraced garden surrounded by cliffs, towering over 10 m, from which water drips constantly (Fig. 6). Immediately below the dripping point, there are the remains of a trough that collected the water, probably to water the garden, which contains some species of plants requiring large amounts of water, such as banana trees. The presence of a natural spring, albeit seasonal, is documented by the IGM (Military Geographic Institute) map (1:25000 V - 244 I SE.), in the locality ‘La Petrazza’ (Fig. 3, location 7). It is placed in relation to a deep incision created by the water flow, as in the case of the other dripping point.

Presently, there is no archaeological evidence pointing to water tanks or reservoirs in Stromboli. The porous lava stone, widely used in the construction of the Bronze Age structures, although easy to work with is not suitable for the storage of water unless it is coated with mortar. On the island of Salina, the excavation of the settlement of the Portella (belonging to the Milazzese phase that follows the Capo Graziano facies) has highlighted a system of channels perpendicular to the slope that converge towards a single point. During storms water is channeled to a single catch point. Another important characteristic of the site of Portella is the number of large vessels found. One of these, with the capacity of 500 liters, was found buried below the floor of the hut Z (Martinelli 2011, 115-121); this surely would serve as a cistern. In the villages of Montagnola of Filicudi and the Acropolis of Lipari, there was also a significant number of large jars. Their distribution indicates that some huts were intended for their storage.

Given the absence of large storage vessels we suggest that perhaps large quantities of water accumulated in specific areas, which would form natural reservoirs. To establish the likely location of these ‘wet’ areas we study the drainage system (Fig. 3) for the island. The area into which many of the highest order streams (order 3) flow appears to be the large flat area between Fico Grande and Punta Lena (Fig. 3, locations 4 and 5). The other order 3 stream is connected to the La Schicciola dripping point. It should be noted how the slope of the volcano side in this area decreases rapidly, creating a similar environment to that of Milazzo described above. Furthermore, Fico Grande can be reached from San Vincenzo without excessive slopes being exceeded; GIS analysis shows the shortest path to Fico Grande being a distance of only

Further archaeological evidence for the exploitation and management of water resources is provided by recent discoveries at Piazza XXV in Milazzo, Sicily, in the lower area of the Isthmus. The excavation brought to light seven ‘structures’ made by placing bottomless jars inside holes cut in the alluvial deposits to reach the water table to a depth of 1.5-2 m. (Fig. 5), a sort of well for collecting fresh water (Martinelli and Tigano 2006).

189

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry 1640 m. The same can be said for the dripping point of ‘Pizzillo’, which is a distance of only 1100 m from San Vincenzo. We emphasize that both of the paths linking San Vincenzo with these two water sources coincide more or less with the present-day roads.

Acknowledgements We would like to thank Sara Levi, Carmelo and Mario Triolo for their precious help during fieldwork and for their suggestions about the schicciolas. We also thank Marco Bettelli, Francesca Ferranti and Valentina Cannavò for all the work done together. Prof. M.A. Marsella in the Faculty of Engineering at the University of Rome “Sapienza” is thanked for access to the data required for creating the DTM.

Conclusions and future work This work has allowed us to test, with the aid of computer modeling, a number of hypotheses regarding early settlement allocation and water availability/proximity to water source in the Aeolian Islands, Italy. Although this work is still at its preliminary stages, some important conclusions have been drawn which we summarise as follows: (1) Relatively high-lying and level areas would have been ‘assessed’ and preferentially chosen, over and above others, for the establishment of a settlement during the Bronze Age; proximity, albeit not immediate proximity, to a natural water supply would have been the second criterion for choice of locality. (2) Of the possible high-lying and level areas available on Stromboli there are only three: one, Ginostra, presents traces of Late Neolithic/BA human activity; the other, San Vincenzo, has developed into settlements with (almost) continuous habitation till the present, and the third possibly from the Roman period. The most favoured of the three has been San Vincenzo, which evolved into the present-day main district on the island, on account of both superb visual command over the S. Tyrrhenian Sea and the Straits of Messina and nearest proximity to two of the schicciolas (dripping water sources) available on the island.

References AbdelKhaleq R.A. and Alhaj Ahmed I. 2007, Rainwater harvesting in ancient civilizations in Jordan, Insights into Water Management: Lessons from Water and Wastewater Technologies in Ancient Civilizations, Water Science & Technology: Water Supply 7 (1), 85-93. Berbabò Brea L., 1985, Gli Eoli e l’inizio dell’età del bronzo nelle isole Eolie e nell’Italia meridionale, Istituto universitario orientale, Dipartimento di studi del mondo classico e del Mediterraneo antico, Napoli. Cadogan G., 2007, Water management in Minoan Crete, Greece: the two cisterns of one Middle Bronze Age settlement, Insights into Water Management: Lessons from Water and Wastewater Technologies in Ancient Civilizations, Water Science & Technology: Water Supply 7 (1), 103-111. Chisholm M., 1968, Rural Settlement and Land Use, London: Transaction Publishers. Dawson H., 2010, A question of life and death? Seafaring and abandonment in the Mediterranean and Pacific Islands, in A. Anderson et al. (eds.) The global origins and development of seafaring, Cambridge: McDonald Institute Monographs, 203-212. Frumkin A., Bar-Matthews M. and Vaks A., 2008, Paleoenvironment of Jawa basalt plateau, Jordan, inferred from calcite speleothems from a lava tube, Quaternary Research 70, 358-367. Ferranti F., Bettelli M., Cannavo V., Di Renzoni A., Levi S.T. and Martinelli M.C. 2015, San Vincenzo, isola di Stromboli (Lipari, Prov.di Messina) – Campagna 2014. In Notiziario di Preistoria e Protostoria- 2.II, Sardegna e Sicilia. Firenze: Istituto Italiano di Preistoria e Protostoria, 55-61. Levi S.T., Bettelli M., Di Renzoni A., Ferranti F., and Martinelli M.C., 2011, 3500 anni fa sotto il vulcano. La ripresa delle indagini nel villaggio protostorico di San Vincenzo a Stromboli, Rivista di Scienze Preistoriche LXI, 157-172. Mammana A., 2006, Le Sorgenti delle Isole Eolie, Reggio Calabria, Città del Sole Edizioni. Martinelli M.C., 2011, Archeologia delle isole Eolie. Il villaggio dell’età del bronzo medio di Portella a Salina. Ricerche 2006 e 2008, Muggiò (MB), Rebus Edizioni. Martinelli M.C. and Tigano G., 2006, Milazzo (Me) - Via XX Settembre. Un villaggio in pianura dell’età del bronzo medio, Proc. XLI Riunione Scientifica dell’Istituto Italiano di Preistoria e Protostoria (San Cipirello (PA), 16-19 novembre), 1295-1299.

Future work will evolve along three directions, namely computer modeling, field work and laboratory analysis of materials. The research agenda is aimed to eventually include all Aeolian Islands and not just Stromboli, given the rich archaeological record and the ‘peculiarities’ of each island regarding type of water source and methods of storage and management. Regarding field work, this will involve survey of Stromboli Island to provide new archaeological and ethnographic data regarding water collection and storage on an individual (home based) and collective (communal) basis. From a hydro-geological perspective, field work will concentrate around the sites of the schicciollas but also places highlighted by the analysis, especially in the area of Fico Grande (Fig. 3, 5), where high rate streams flow into a flat area. Regarding laboratory-based work, investigations of the raw materials available on Stromboli will be undertaken to understand whether some kind of waterproofing was applied on the lava rock, and whether these materials can be traced in the archaeological record. Finally and with regards to computer modeling it is hoped that we can expand the GIS-based analysis to other Aeolian islands, as a means of establishing a comprehensive record of the relationship between early settlements in arid environments and water supply and management and a model for sustainability.

190

A. Di Renzoni et al., The relationship between early settlements and sources of water supply Miller R., 1980, Water use in Syria and Palestine from the Neolithic to the Bronze Age, World Archaeology 11 (3), 331-341. Saletta E., 2013, Processo di formazione degli strati e interpretazione del deposito archeologico nel sito di San Vincenzo a Stromboli: analisi quantitativa del quadrato AN124, Masters Thesis Università di Modena e Reggio Emilia, a.a. 2011-2012.

Shreve R.L., 1966, Statistical law of stream numbers, Journal of Geology 74, 17-37. Vagnetti L., 1982, Quindici anni di ricerche e studi tra mondo egeo e l’Italia protostorica, in L. Vagnetti (ed.) Magna Grecia e mondo miceneo. Nuovi documenti, Catalogo della mostra, Taranto, 9-40.

Fig. 1. Southern Tyrrhenian Sea and the Aeolian Islands. The line shows the probable route followed by the Mycenaeans on their way to the Bay of Naples, crossing the Straits of Messina and stopping at the Aeolian Islands where there is evidence of Mycenaean pottery.

Fig. 2. San Vincenzo site, West Area; arrows point to terrace walls and huts.

191

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

Fig. 3. Digital Terrain Modelling (DTM ) of Stromboli island with the reconstruction of the drainage system. Key localities and associated sites and features mentioned in the text are: 1 Scari landing point; 2 Fico Grande landing point; 3 San Vincenzo BA settlement; 4 Fico Grande flat area; 5 Fico Grande possible well point; 6 Pizzillo; 7 La Petrazza; 8 La Schicciola; 9 Ginostra; 10 Punta Labronzo.

Fig. 4. Analysis of the visual potential of the different Middle Bronze Age (Capo Graziano II) sites in the Aeolian Islands. A: Stromboli- San Vincenzo; B: Stromboli-Punta Labronzo (hypothetical BA site); C: Lipari, Acropolis; D: Filicudi, Montagnola.

192

A. Di Renzoni et al., The relationship between early settlements and sources of water supply

Fig. 5. Plan (left) and section (right) of the pit 2 (from Martinelli and Tigano 2006, fig 1, 1302).

Fig. 6. A schematic illustration (DTM (top left) and profile graph (bottom left) of the Pizzillo schicciola (dripping water). On the right there are views of the cliff from which the water drips. The presence of a large-sized trough built at the bottom of the rock face and currently silted suggests that it was built to collect the dripping water.

193

Chapter 28

Reconstructing archaeo-landscapes: myth versus reality A. Sarris1, 3, A. Chliaoutakis1, S. Dederix1, 2 and J.C. Donati1 1

Laboratory of Geophysical – Satellite Remote Sensing & Archaeo-environment, Institute for Mediterranean Studies (IMS), Foundation for Research & Technology, Hellas (F.O.R.T.H.) 2 Research Fellow of the F.R.S. – FNRS, Aegean Interdisciplinary Studies Research Group (AegIS), Institut des Civilisations, Arts et Lettres (INCAL), Université Catholique de Louvain, Belgium 3 Department of Anthropology, The Field Museum of Natural History of Chicago, Illinois, USA Corresponding author: [email protected]

Abstract Thanks to the diversification of scientific methods that can support archaeo -environmental studies, researchers have at their disposal an increasing amount of data that can be combined to place past human activities back into their contemporary environment. Virtual Reality, Geophysics, Geomorphology, Remote Sensing, Agent-Based Modelling and Artificial Intelligence provide new opportunities, but also new challenges, regarding the study of ancient landscapes. Making the proper assumptions, selecting the corresponding variables and using flexible and dynamic models creates a ‘living landscape’ according to individual mindscapes. The obstacles encountered in such a task are numerous, ranging from the ways of reconstruction to the various interpretation schemes. There is therefore a variety of possible scenarios of reconstruction, which are all objective to a certain extent, insomuch as they are tied to measureable parameters of space. Such an integrative approach can only benefit archaeologists a nd social scientists that rely on solid groundwork when trying to recover the social, political, economic and symbolic meanings of past landscapes. This paper deals directly with the above issues, providing specific examples of how such reconstruction processes can (and should) be achieved using Geoinformatics and technology-based approaches to landscape archaeology. These include examples from (1) an aerial and satellite remote sensing application at an ancient Greek urban center in the Peloponnese, (2) a geophysical survey aiding in the 3D virtual reconstruction of Neolithic settlements in Hungary, (3) a GIS analysis of the Bronze Age tombs on Crete within their surrounding environment, and (4) an agent-based model to study the social organization of land-use patterns of Minoan Crete. As a whole, these approaches help formulate a more holistic approach to fieldwork that will frame the direction of future archaeological research. Keywords: Archaeo-landscapes, Reconstruction, GIS, Remote Sensing, Geophysics, VR, Agent-based archaeology

Introduction Conventionally, the reconstruction of archaeolandscapes followed the same trends as the reconstruction of any modern landscape using specific computer graphic software and an abstracted modeling of conceptualized worlds. Obviously, such kinds of virtual models are trying to visualize space in a more simplified format based on the fragmentary archaeological record (limited excavations), the missing data and the corresponding interpretation of them. Moving a step further, virtual reality models emphasize the degree of freedom of the various interpretation schemes (Chalmers and Stoddart 1996). Such models emphasize more the accuracy of the spatial extent and dimensions of the monuments and the landscape, while the remaining reconstruction is left to a more or less impressionistic approach of the past playing with visual effects (illumination, day/night conditions, etc.). As long as the abstracted reconstructions satisfy purposes of exhibitions for the general public or educational intentions, they remain valid, but, on the other hand, they cannot advance archaeological research offering a new knowledge.

Dressed with a photo-realistic render, and sometimes with an interactive immersive way of interaction, the particular models are trying to overcome the uncertainty of the archaeological record without being able to offer a cognitive approach to archaeological questions (Slocum et al. 2001; De Boer 2010). So, how can we make breakthroughs in the inherent limitations in reconstructing historical landscapes? Geoinformatics can partially fill this gap by gathering information about the terrain and the subsurface distribution of monuments and modelling the activities and the usage of space. Specific spatial tools and processing algorithms can be used to provide a more solid base for the reconstruction of archaeolandscapes as a counter measure of the abstracted virtual modeling of them. This paper presents just a sample of such techniques in order to manifest the ways of operation and how one can retrieve specific information that can be used to make a more productive (and realistic) representation (and not just visualization) of past cultural landscapes in a cognitive way.

195

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry space and the type of activities in the area around the tell. Contrary to Mantinea, the study of satellite images, historical aerial photos and spectroradiometry measurements was not able to provide any substantial additional information to enrich the data derived from the aforementioned techniques (Salvi et. al. 2013).

Archaeo-landscapes from Space. Satellite Remote sensing at Mantinea Satellite remote sensing conducted at the Greek settlement of Mantinea in the Peloponnese has identified an extensive network of subsurface streets and cityblocks that were arranged in an orthogonal manner. This valuable information helps clarify the wider urban dynamics of the largely unexplored city and more broadly places Mantinea squarely within the traditions of Greek urban planning of the Classical and Hellenistic periods. Previously, these kinds of archaeological discoveries in Greece have only been possible through expensive and long-term excavations or geophysical survey.

On the other hand, the extensive use of geophysical surveying that managed to map more than 30ha around the tell (within an area of about 1sq. km) had a significant impact in terms of drawing conclusions about the extent of the site and the evolution of habitation patterns in the region. Following a manifold strategy (Sarris 2012), geophysical prospection was carried out employing magnetic and electromagnetic (EM) methods, GPR and ERT. ERT and EM provided valuable information regarding the extent and stratigraphy of the tell and the flowing dynamics of the palaeochannels. Szeghalom tell appears to belong to the lower size category of tells, according to the classification made by Kalicz and Raczky (1987, 16) for the Tisza period tells, covering a total area of 9,000sq. m. It seems that it was at least partially fortified with palisades and a system of meanders was running around it forming a natural defensive ditch.

The main data-set was a multispectral, high-resolution (panchromatic 0.63m; multispectral 2.50m) Quickbird satellite image from 13 September 2003 with an off-nadir angle of 9.4 degrees. We applied a combination of vegetation indices, band ratios, linear transformations and spatial filtering to the Quickbird image in ERDAS IMAGINE 13.0.2. A combination of spatial filters revealed more than 100 linear surface anomalies inside the fortification walls that range in length from as little as 6m to more than 600m. Although we detected a handful of diagonal anomalies, the majority of linear anomalies at Mantinea are oriented almost at true north-south and eastwest. To be more specific, we found the dominant orientations to be 179° north-south and 89° east-west by averaging the orientations of all linear surface anomalies apart from the diagonals. This uniform trend of northsouth and east-west is strikingly similar to excavated buildings and venues within the city, most notably in the agora. It is unlikely that all surface anomalies from the spectral filters represent the location of an ancient street or building, but many of them certainly do. For example, remote identified a series of four parallel and evenly spaced north-south roads in the southern region of the city (Fig. 1). Based on the evidence as a whole, a partial reconstruction of the street system of Mantinea can be reproduced. The application of additional multispectral, high-resolution satellite imagery with different extraction dates would likely reveal more features of the ancient landscape. Whatever the case, satellite remote sensing has a great potential to provide information regarding the internal details of an ancient settlement together with other characteristics of its natural settings.

The magnetic results (Fig. 2) were even more revealing in terms of the internal organization of the settlement. Leaving an empty buffer of about 150-200m around the tell, that in later periods probably denoted a landscape of memory and ancestry, two sections of built environments were identified. A dense, flat settlement expands to the south of the tell, while a dispersed type of settlement consisting mainly of Tisza type farmsteads is suggested towards the remaining directions where the paleo meanders seem to have been expanded. More than 60 structures of oblong ground plans with similar dimensions (around 17-25m x 8m) and orientation have been identified, and their magnetic signature indicates that intense heating/crafting activities took place inside. The magnetic features can be easily correlated to burnt daub walls, postholes or other thermal targets (e.g. kilns and fire hearths). Having also detailed information of the digital elevation model of the area, it was shown that most of the clustered areas were established on higher elevations, indicating habitation on small islands of soil within a wetlands environment. Based on the above evidence, virtual landscape reconstruction has been initiated using AutoCad 3D and 3d Studio Max. The vector interpretation of the geophysical anomalies and the digital terrain model were imported in the above software packages, and, based on the test excavations that were carried out in one of the farmsteads, it was possible to reconstruct the habitation of the region. Thanks to this reconstruction, it is conceivable for everyone to have a general vision of how a Hungarian village during the Late Neolithic period developed in a landscape filled with oak trees and streams, with mud, wood and straw huts surrounding a tell (http://www.youtube.com/watch?v=J_eRN20PPuM) (Fig. 2).

The Exploitation of Landscape in Neolithic Hungary. Geophysics and Virtual Reality From a different perspective, geophysical approaches have been used for the past three years under the scope of the Köros Regional Archaeological Project (KRAP) to study the establishment and development of agricultural settlements on the SE Great Hungarian Plain during the Neolithic and Early Copper Age (c. 6,000-4,000 BC) (Yerkes 2010; Parkinson and Gyucha 2012). Focusing on the site of Szeghalom, located in the agricultural region north of the Sebes Körös River and SW of the village of Szeghalom, surface surveying, test excavations, coring for chemical analyses and topographic mapping have been carried out systematically to characterize the use of

196

A. Sarris et al., Reconstructing archaeo-landscapes associated random points. Such tables allowed assessing whether individual cemeteries afford a low, medium or high visibility in comparison to their own surroundings. Each table was treated independently. The 100 random points were subdivided in 10 ordered classes defined in such a way that each comprised exactly 10% of the random points – class 1 corresponding to the locations that offer the lowest visibility over the surroundings and class 10 to those that afford the most dominant views. The burial site itself was then assigned to one of these classes according to the breaking values obtained by this means. The histograms shown in Figure 4.1 compare the relative extent of the area visible from the cemeteries and the random points, for each sequence of analyses. Both histograms illustrate a strong underrepresentation of the burial sites in the categories that correspond to low values, and a corresponding overrepresentation in high values. The impression gained by visually examining the histograms is further confirmed by the results of Kolmogorov-Smirnov tests which indicate that the difference between the two distributions is large enough to be statistically significant at the 0.01 confidence level. In other words, the cemeteries considered in the analyses offer a wider view over their surroundings than they would have if they had been placed randomly within their 500m radius neighborhood. This underlines the existence of a clear concern for visibility in the decision-making process that led to the location of circular tombs, suggesting in turn that the burial buildings functioned as landmarks in their local landscape.

Minoan Tombs as Landmarks? The Case of the Circular Tombs The Early Bronze Age record of South Central Crete is dominated by monumental circular tombs, often misleadingly called ‘tholos’ tombs. Such tombs were obviously of great importance to their users: considerable energy was invested in their construction, they remained in use for centuries, and they were the focus of funerary as well as non-funerary rituals. But did the significance of circular tombs translate spatially? Were the burial structures intentionally positioned so as to occupy prominent locations in the landscape and to function as landmarks? It is well acknowledged that circular tombs were not erected on the summit of dominant peaks (Branigan 1998). Still, this does not mean that they did not exert any spatial dominance over their surroundings. Basically, assessing whether a monument is topographically prominent consists in assessing whether it is located on top of a ridge or in the bottom of a valley. But between these two extremes, there is a whole range of topographic settings that afford different degrees of prominence (Llobera 1996; Gillings 2009; De Reu 2011). To use a metaphor, prominence is not either black or white; it is grey-scaled. In addition, if prominence is to be explored in its totality, the pattern of visibility in the landscape must be taken into account as well (De Reu et al. 2011; Christopherson 2003). Making use of GIS technologies, it is upon this latter issue that this casestudy focuses on. Prominence is relative and scale-dependent – i.e. it is related to a defined neighborhood. Burial structures are local features, as their location is more often than not influenced by the place of dwelling of the associated community, and they must therefore be studied locally, in relation to their close neighborhood. Given the contrasted topography of South Central Crete, this neighbourhood was conventionally chosen here to be 500m in radius, which seemed to be large enough to ensure meaningful comparison, while retaining the local character of the sites. 54 cemeteries of circular tombs were considered in the analyses. In order to define whether the burial sites were positioned so as to offer a commanding view over their own surroundings, each one of these 54 cemeteries was studied individually: the extent of the viewshed of the cemetery was compared to the extent of the viewsheds that were computed separately for each one of 100 random points generated within those portions of the land that fit inside a 500m radius buffer around the site, while being characterized by a slope index of maximum 16°. Indeed, even if Bronze Age people had sought to establish their cemeteries at important viewpoints, they would still have had to cope with the terrain and therefore restrict their choices to topographic units able to accommodate built tombs. Two sequences of analyses were carried out, the first with a viewing radius of 50km, the second with a viewing radius limited to 5km.

Agent-Based Modeling. Reviving the Cultural Dynamics of Bronze Age Malia Agent-based modeling (ABM) has been increasingly used in Archaeology during the past decade as a tool for assessing the plausibility of alternative hypotheses regarding ancient civilizations, their organization, and social and environmental processes at work in past ages (Dean et al. 2000; Janssen 2009; Kohler; 2000). Its emerging popularity is due to its ability to represent individuals and societies, and to engage the inherent uncertainties in archaeological theories or findings. Indeed, the unpredictability of interaction patterns within a simulated agent society, along with the strong possibility of emergent behaviour, can help archaeologists gain new insights into existing theories; or even come up with completely novel explanations and paradigms regarding the ancient societies being studied. Though our work here is inspired by existing case studies, our model itself is quite generic, and does not aim to prove or disprove a specific theory. Indeed, using agent-based models built on knowledge derived from archaeological records and evidence, but not trying to fit their results to a specific material culture, allows for the emergence of dynamics for different types of societies in different types of landscapes, and can help derive knowledge of socio-ecological systems that are applicable beyond a specific case study. In more detail, we present a functional ABM system prototype for simulating an artificial ancient society of agents residing

The calculation of all these viewsheds led to the creation of one attribute table per cemetery and per sequence of analyses, each table recording separately the amount of cells in-sight from the cemetery and every one of its

197

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry at Malia on the island of Crete during the Bronze Age (resolution of 100m x 100m within an environmental grid of 20km x 25km). The agents correspond to households, which are considered to be the main social unit of production for the period (Whitelaw 2007). The ABM allows us to explore the sustainability of intensive farming (‘garden’ cultivation with hand tillage, manuring, weeding, and watering) and extensive cultivation (largescale tillage by ox-drawn ploughs) agricultural technologies in use at the time (Halstead 1981; Isaakidou 2008), and to examine their impact on population dispersion (Fig. 4).

Geoinformatics in unfolding new dimensions in the interpretation of the archaeolandscapes allowing a more cognitive approach in their reconstruction. Dealing with questions regarding the plan and the extent of settlements, the exploitation of the land and the changes of the natural settings or the establishment of landmarks (just to present a few of the many questions that can be addressed), spatial technologies help us to reveal the traces of the past activities and sometimes human perceptions that have been retained in the landscape and have been conceived in various ways trough the time line. Even though no one can deny the difficulties encountered in the decoding of landscapes' memoirs, especially when dealing with memories of past events, routes, habitation, activities, territories, ancestry, etc., it is indisputable that the particular methods can be productive in providing a solid basis for more convincing reconstruction models.

In addition, the ABM attempts to assess the influence of different social organization paradigms on land-use patterns and population growth. Importantly, the model evaluates the social paradigm of agents self-organizing into a hierarchical social structure, and continuously readapting the emergent structure, if required. We examine whether the adoption of a self-organized agent settlement organization will indeed give rise to a stratified social structure, which will also be able to sustain itself through time. In effect, the self-organization technique presented here is one that results in the continuous targeted redistribution of wealth, so that resources flow from wealthy to poor agents when needed. Our selforganization model is inspired by the work of Kota et al. (2009); however, we modified that model in several important ways. We note that this is the first time a selforganization approach is incorporated in an ABM system used in archaeology.

Acknowledgments This publication has been funded with support from the European Commission (project Archaeolandscapes Europe, Culture 2007-2013). It reflects the views of the authors, and the Commission cannot be held responsible for any use which may be made of the information contained therein.

References Bintliff J., 1982, Settlement Pattern, Land Tenure and Social Structure: A Diachronic Model, in C. Renfrew and S. Shennan (eds.) Ranking, Resource and Exchange, Cambridge: Cambridge University Press, 106-111. Branigan K., 1998, The Nearness of You: Proximity and Distance in Early Minoan Funerary Landscapes, in K. Branigan (ed.) Cemetery and Society in the Aegean Bronze Age, Sheffield: Sheffield Academic Press, 1326. Chalmers A. and Stoddart S., 1996, Photo-realistic Graphics for Visualising Archaeological Site Reconstructions, in T. Higgins, P. Main and J. Lang (eds.) Imaging the Past: Electronic Imaging and Computer Graphics in Museums and Archaeology, London: British Museum Press, 85-93. Cherry J., 1986, Polities and Palaces: Some Problems in Minoan State Formation, in C. Renfrew and J. Cherry (eds.) Peer Polity Interaction and Socio–Political Change, Cambridge: Cambridge University Press, 19–45. Christopherson G.L., 2003, Using ARC/GRID to Calculate Topographic Prominence in an Archaeological Landscape, in Proc. 23rd Annual Esri Internat. User Conference (San Diego, USA), http://proceedings.esri.com/library/userconf/proc03/p 0530.pdf. De Boer A., 2010, Processing Old Maps and Drawings to Create Virtual Historic Landscapes, e-Perimetron 5(2), 49-57. De Reu J., Bourgeois J., De Smedt P., Zwertvaegher A., Antrop M., Bats M., De Maeyer P., Fine P., Van Meirvenne P. and Verniers J., 2011, Measuring the

The ABM model was developed using the NetLogo1 modeling environment. Various scenarios were taken into account for the experimental setup with different parameterization. Our simulation results confirm these intuitions, demonstrating that self-organizing agent populations not only survive for a two millennia period, but are also by far the most successful, growing much larger than populations employing different social paradigms. The success of this (dynamic) social paradigm that gives rise to ‘stratified’, that is, non-egalitarian societies, provides support for so-called ‘managerial’ archaeological theories which assume the existence of different social strata in Neolithic and Early Bronze Age Crete. This early stratification was likely a pre-requisite for the emergence of the Minoan Palaces and the hierarchical social structure evident in later periods (Bintliff 1982; Cherry 1986; Gilman 1981). The prototype ABM is based on archaeological evidence, and is meant to be used as a tool enabling archaeologists to assess the potential validity of competing hypotheses; or even consider aspects of the past that have not yet been considered. Finally, it can provide the basis for a fully interactive tool to help popularize archaeological theories.

Final Remarks The above case studies manifest the potential of

1

See http://ccl.northwestern.edu/netlogo

198

A. Sarris et al., Reconstructing archaeo-landscapes Hungary, 4th EARSeL Workshop on Cultural and Natural Heritage, Matera, Italy. Sarris A., 2012, Multi+ or Manifold Geophysical Prospection? Computer Applications and Quantitative methods in Archaeology 2012, University of Southampton, Great Britain. Slocum T., Blok C., Jiang N., Koussoulakou A., Montello D., Fuhrman S. and Hedley N., 2001, Cognitive and Usability Issues in Geovisualization, Cartography and Geographic Information Science 28, 61–75. Yerkes R., 2010, Prehistoric Farming Societies in Hungary, The Center for Slavic and East European Studies 37(2), 14. Whitelaw T., 2007, House, Households and Community at Early Minoan Fournou Korifi: Methods and Models for Interpretation, in R. Westgate, N. Fisher and J. Whitley (eds.) Building Communities: House, Settlement and Society in the Aegean and Beyond, London: British School at Athens Studies, 65-76.

Relative Topographic Position of Archaeological Sites in the Landscape, a Case Study on the Bronze Age Barrows in Northwest Belgium, Journal of Archaeological Science 38, 3435-3446. Dean J.S., Gumerman G.J., Epstein J.M., Axtell R.L., Swedlund A.C., Parket M.T. and McCarroll S., 2000, Understanding Anasazi Cultural Change through Agent-Based Modeling, in T.A. Kohler and G.J. Gumerman (eds.) Dynamics in Human and Primate Societies. Agent-based Modeling of Social and Spatial Processes, Oxford: OUP, 179–206. Gillings M., 2009, Visual affordance, landscape and the megaliths of Alderney, Oxford Journal of Archaeology 28 (4), 335-56. Gilman A., 1981, The Development of Social Stratification in Bronze Age Europe, Current Anthropology 22, 1-8. Halstead P., 1981, Counting Sheep in Neolithic and Bronze Age Greece, in I. Hodder, G. Isaac and N. Hammond (eds.) Pattern of the Past, Cambridge: CUP, 307-339. Isaakidou V., 2008, The Fauna and Economy of Neolithic Knossos Revisited, in V. Isaakidou and P.D. Tomkins, (eds.) Escaping the Labyrinth. The Cretan Neolithic in Context, Oxford: Oxbow Books, 90-114. Janssen M.A., 2009, Understanding Artificial Anasazi, Journal of Artificial Societies and Social Simulation, 12(4) 13, http://jasss.soc.surrey.ac.uk/12/4/13.html. Kalicz N. and Raczky P., 1987, Berettyóújfalu-Herpály. A Settlement of the Herpály Culture, in L. Tálas and P. Raczky (eds.) The Late Neolithic of the Tisza Region. A survey of recent excavations and their findings: Hódmezõvásárhely-Gorzsa, SzegvárTûzköves, Öcsöd-Kováshalom, Vésztõ-Mágor, Berettyóújfalu-Herpály, Budapest: Szolnok, 105-125. Kohler T.A., Kresl J., West C.V., Carr E. and Wilshusen R.H., 2000, Be there then: A modeling Approach to Settlement Determinants and Spatial Efficiency among Late Ancestral Pueblo Populations of the Mesa Verde Region, U.S. Southwest, in T.A. Kohler and G.J. Gumerman (eds.) Dynamics in Human and Primate Societies. Agent-based Modeling of Social and Spatial Processes, Oxford: OUP, 145–178. Kota R., Gibbins N. and Jennings N.R., 2009, SelfOrganising Agent Organisations, in AAMAS’09 (Budapest, Hungary, 10-15 May 2009), 797–804. Llobera M., 1996, Exploring the Topography of Mind: GIS, Social Space and Archaeology, Antiquity 70, 612-622. Parkinson W.A. and Gyucha A., 2012, Tells in Perspective: Long-Term Patterns of Settlement Nucleation and Dispersal in Central and Southeast Europe, in R. Hofmann, F.-K. Moetz and J. Müller (eds.) Tells: Social and Environmental Space, Kiel: Universitätsforschungen zur Prähistorischen Archäologie, 105-116. Salvi M.C., Sarris A., Papadopoulos N., Agapiou A., Parkinson W., Yerkes R.W. and Gyucha A., 2013, The Contribute of Aerial and Satellite Images for the Archaeological Prospection of Neolithic Sites: The Case Study of Szeghalom-Kovácshalom Tell,

Fig. 1. Spectral filter (combination of WDVI, TSAVI, and Green NDVI) of the southern region of Mantinea that likely identifies four parallel subsurface roads (indicated by arrows).

199

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry

Fig. 2. Results of the magnetic survey at Szeghalom (left) and part of the virtual reconstruction of the Neolithic village based on the results of the geophysical survey (right).

Fig. 3. Comparison between the extent of the viewsheds computed for the circular tombs and their associated random points with a viewing radius of 50km (left) and 5km (right).

Fig. 4. Graphical user interface of the agent-based modeling at Malia, Crete.

200

Chapter 29

Luminescence dating and quartz grain surface features of aeolian sediments from Agia Napa, Cyprus E. Tsakalos, C. Athanassas and Y. Bassiakos

Laboratory of Archaeometry, Institute of Nanoscience and Nanotechnology, National Centre for Scientific Research, N.C.S.R. ‘Demokritos’, Athens, Greece Corresponding author: [email protected]

Abstract The coastal zone of Southern Cyprus is of great archaeological and palaeoenvironmental significance. Despite the great deal of palaeoenvironmental data produced for other Eastern Mediterranean coastal sites, the abundance of field evidence on the southeast coasts of Cyprus has been exploited to only a limited extent to date. Archives of relative sealevel change and aeolian sand accumulation have been encoded in the coastal sedimentary deposits along the coast of Cyprus. The purpose of this research work is to examine the chronology of aeolianite deposits of Southeast Cyprus and provide preliminary comments on the Late Quaternary environmental change by employing up-to-date luminescence dating methods. Another aim of this study is the analysis of the microtextures that are present on quartz grains of coastal dunes from South East Cyprus to uncover their depositional history. Keywords: Cyprus; palaeoenvironment; sea level change; coastal deposits; aeolianites; luminescence dating; IRSL; quartz; microtextures

fading (a signal decrease with time leading to an age underestimation) (Wintle 1973; Spooner 1994); this is an issue that had previously prevented the use of feldspar.

Introduction The study of sediments deposited during the Quaternary can provide detailed information on palaeoenvironmental changes, and coastal sediment archives in particular are of great significance as they give information on the timing of past climate changes and associated interactions between coastal landscape-sea level change. Although a great number of studies have been conducted in many areas of the Mediterranean regarding its coastal stratigraphy and the associated palaeoclimate and sea level changes, little is known about the coasts of Cyprus. In Southeastern Cyprus, coastal dunes (aeolianites) now forming elongated ridges, appear as morphological features running parallel to the current shoreline presenting an indicator of sea-level and climate changes. These features consist of lithified wind-blown, fine-to medium-grained, well-sorted sand composed mainly of quartz and feldspar grains.

Further, the application of Scanning Electron Microscope (SEM) in examining irregularities on quartz sand grains has been employed into a common method in determining different sedimentary environments and deciphering their palaeoenvironmental-depositional history (e.g. Moral-Cardona et al. 1996; 1997; Mahaney 1998; Newsome and Ladd 1999). Krinsley and Donahue (1968) were among the first who made use of micromorphology of sand grain surfaces permitting the categorization of different sets of imprinted textures formed by chemical and mechanical processes. These textures and the frequency with which they appear can in turn be used to determine the sedimentary histories of quartz grains, allowing the clear distinctions between aeolian, marine, glacial and diagenetic depositional environments to be made. If grains have been through different environments, their surface imprints may encompass a combination of different textures produced during sediment transport, deposition and diagenesis. Principally, mechanical processes leaving signatures as different impact and abrasion marks on the grain surfaces during transportation in different dynamic environments are the profound diagnostic features which record those mechanical processes. Marks as a result of chemical processes consisting of a range of different overgrowth and etching types can further facilitate the identification of the diverse post-depositional course of action that grains have experienced. This paper presents inherited

Dating of aeolian formations and the establishment of a reliable chronological framework are key in analyzing and interpreting changes of the past. Luminescence dating as an absolute dating technique has proven to be suitable for dating various Quaternary deposits (aeolian, fluvial, glacial etc.) and there have been a number of recent reviews (Wallinga 2002; Singhvi and Porat 2008; Liritzis et al. 2013; Fuchs and Owen 2008). However until recently, luminescence dating had mainly made use of quartz. Lately, advancements in luminescence dating have included infrared stimulation of feldspar (e.g. Buylaert et al. 2011; 2012; Stevens et al. 2011) (IRSL) using an IR signal which is less affected by anomalous

201

Proceedings of the 6th Symposium of the Hellenic Society for Archaeometry quartz features observed on Quaternary dune quartz grains from the Agia Napa area in SE Cyprus.

Step 1 2

Give dose Preheat, 60 s @ 320 ˚C

Geological setting The studied area is located in the coastal zone of Southeast Cyprus (Fig. 1). Late Pleistocene deposits are well developed all along the southeastern coasts particularly in the area of Agia Napa and Cape Greco. These coastal deposits form a stratigraphic sequence of shallow marine to beach sediments covered by calcitecemented aeolian sediments, with the latter dominating in both thickness and spatial extent sitting on a marly limestone base of Miocene age. The shallow marine deposits found at the base of the late Pleistocene deposits are rich in biogenic material and marine shells. This marine member is superimposed by a beach member with the presence of some fragments of broken shells. This member is overlaid by carbonate rich coastal dunes containing a high percentage of shell fragments most probably a result of onshore transportation of marine exposed fauna by prevailing south, south-west winds during the later part of the Quaternary (McCallum 1989). Only aeolian samples were collected for feldspar luminescence dating and micromorphological analysis of their constituent quartz grains.

3 4 5 6

IR stimulation, 100 s @ 50 ˚C IR stimulation, 100 s @ 290 ˚C Give test dose Cutheat, 60 s @ 320 ˚C

7 8 9 10

IR stimulation, 100 s @ 50 ˚C IR stimulation, 100 s @ 290 ˚C IR stimulation, 40 s @ 325 ˚C Return to step 1

Treatment

Table 1. Post-IR IRSL SAR protocol for coarse-grain feldspar measurements (after Thiel et al. 2011). Prior to standard dating of the two samples using the SAR procedure, validation of the protocol parameters was verified by a pre-heat plateau test (Murray and Wintle 2000) using different temperatures and a dose recovery test (Murray and Wintle 2003). The pre-heat plateau test confirmed an appropriate pre-heat temperature at 320˚C, while the dose recovery test showed that the measured to given dose ratios were within 2 σ-level (0.9-1.1), indicating that the applied protocol is working as it should. A test to investigate the bleachability (Klasen et al. 2006) of the pIRIR290 signal by natural light was further performed in one sample. After 90 minutes of sun exposure the sample retained ~3 Gy as a residual of the pIRIR290 signal, which is in general ~3% of the De value (99.6 Gy). This retained signal percentage was subtracted from the derived De values of the two samples before age calculations. A fading test was also performed in one sample (Huntley and Lamothe 2001). A mean g-value of -0.4% ± 0.06% was obtained indicating that using this protocol, feldspar experiences no loss of signal. It is therefore secured to assume that De values derived by the pIRIR290-SAR protocol are accurately calculated.

Methods Samples Preparation for Luminescence dating Two aeolian samples were collected as rectangular blocks for luminescence dating. Samples were treated with 10% hydrochloric acid to remove carbonate cements and 10% hydrogen peroxide to remove organic content, drying and sieving (80-125 μm fraction was selected). Feldspar minerals were density separated from quartz and heavy minerals by a heavy liquid (sodium polytungstate). Further, feldspar fractions were treated with 10% hydrofluoric acid to avoid contribution of the alphairradiated outer part of the mineral. Luminescence measurements were performed by an automated Risø TL/OSL DA-15 luminescence reader. Beta Irradiation was from a calibrated 90Sr/90Y source and stimulation was by infrared light diodes emitting at 870 nm at 90% power. The IRSL signal was detected using a combination of Schott BG-39 and Corning 7-59 filters in the blue light spectrum between 320 and 450 nm. Measurements were made using the pIRIR single aliquot regenerative-dose protocol (Thiel et al. 2011) (Table 1). Using this protocol infrared light stimulation at 50°C is initially applied, followed by stimulation at 290°C for 100s. The equivalent dose (De) for each aliquot was calculated using the first 2 s of the luminescence signal recorded at 290°C, less a background based on the sum of the final 20 s. Equivalent doses were chosen for further analysis if they had (i) recycling ratios within 10% of unity, and (ii) thermal transfer