Bronze Age Warfare: Manufacture and Use of Weaponry 9781407308227, 9781407338118

Table of contents :
Cover Page
Title Page
Copyright
PREFACE
INTRODUCTION TO ‘BRONZE AGE WARFARE: MANUFACTURE AND USE OF WEAPONRY’
SPEARHEADS AND SWORDS – THE MAKING OF BRONZE OBJECTS
USE-WEAR ON BRONZE AGE HALBERDS: THE CASE OF IBERIA
CEREMONIAL OR DEADLY SERIOUS? NEW INSIGHT INTO THE FUNCTION OF IRISH EARLY BRONZE AGE HALBERDS
DELIBERATE DESTRUCION OF HALBERDS
USE-WEAR ANALYSIS AND USE-PATTERNS OF BRONZE AGE SWORDS
CHELSEA AND BALLINTOBER SWORDS: TYPOLOGY, CHRONOLOGY AND USE
IRISH SWORDS: USE AND ABUSE
INTEGRATING FORM, FUNCTION AND TECHNOLOGY IN ANCIENT SWORDS. THE CONCEPT OF QUALITY
SLASH AND STAB: SOME REMARKS ON THE MANUFACTURE, FUNCTION AND DEPOSITION OF CARP’S-TONGUE SWORDS (SUMMARY)
THE DISCOVERY OF A NEW TACHLOVICE SWORD AT PONT-SUR-SEINE, AUBE, FRANCE
RITUAL OBJECT OR POWERFUL WEAPON – THE USAGE OF CENTRAL EUROPE BRONZE AGE SWORDS
NEUTRON RESONANCE CAPTURE ANALYSIS AND THE BRONZE AGE
THE SINTASHTA BOW OF THE BRONZE AGE OF THE SOUTH TRANS-URALS, RUSSIA
THE FUNCTION OF BRONZE AGE SHIELDS
METALLURGICAL AND ARCHAEOLOGICAL INVESTIGATIONS ON THE MACEDONIAN SHIELDS OF STARO BONČE (3RD CENTURY BC)
THE RESULTS OF THE X-RAYS EXAMINATIONS OF SWORDS FROM POLAND
EXPERIMENTAL COMPARISON OF SPEARHEADS USE-WEAR ANALYSIS

Citation preview

BAR S2255 2011

Bronze Age Warfare: Manufacture and Use of Weaponry

UCKELMANN & MÖDLINGER (Eds)

Edited by

Marion Uckelmann Marianne Mödlinger

BRONZE AGE WARFARE

B A R

BAR International Series 2255 2011

Bronze Age Warfare: Manufacture and Use of Weaponry Edited by

Marion Uckelmann Marianne Mödlinger

BAR International Series 2255 2011

Published in 2016 by BAR Publishing, Oxford BAR International Series 2255 Bronze Age Warfare: Manufacture and Use of Weaponry © The editors and contributors severally and the Publisher 2011 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 9781407308227 paperback ISBN 9781407338118 e-format DOI https://doi.org/10.30861/9781407308227 A catalogue record for this book is available from the British Library BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 1974 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by Archaeopress in conjunction with British Archaeological Reports (Oxford) Ltd / Hadrian Books Ltd, the Series principal publisher, in 2011. This present volume is published by BAR Publishing, 2016.

BAR PUBLISHING BAR titles are available from:

E MAIL P HONE F AX

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

The participants of the workshop in the Natural History Museum, Vienna – November 1st, 2009

PREFACE Marion Uckelmann, Marianne Mödlinger

these ongoing studies show the need for co-ordinated perspectives at European scale in order to find answers to the multitude of new questions that have arisen. The original idea was to bring together scientists from different professional backgrounds, yet working on the same topic, to exchange research and develop new approaches. The focus was to discuss: the application of modern metallurgical methods and techniques used to investigate ancient weapons; experimentation relating to the manufacture and utilisation of the weapons; and the archaeological approaches to weapons, particularly, the use wear on objects. This idea needed funding as it involved bringing specialists from across Europe together. We found this a much harder and longer road than we expected. However, we are both optimists by nature as well as very persistent and benefitted greatly from the support of many individuals who would later become participants. Thank you for bearing with us! We received funding two years after conceiving the idea and subsequently gathered in the beautiful Riva del Garda and later in the stunning Natural History Museum of Vienna. Our utmost gratitude extends to the Österreichische Forschungsgesellschaft and the Österreichisches Bundesministerium für Wissenschaft und Forschung for making the workshop possible, and to the Natural History Museum in Vienna and its wonderful staff for providing the perfect location and catering, including the unforgettable wine tasting! Our thanks go out as well to the EAA committee for selecting our session for their programme. Furthermore, we would like to thank the Landesmuseum Kärnten for supporting the work on the publication. Finally, we would like to thank all participants of the EAA session in Riva del Garda and the workshop in Vienna, who made both meetings so productive and inspiring and created such an open and collegiate atmosphere, and especially to those who contributed to this volume.

This volume presents the outcome of two meetings: the international workshop on Warfare in Bronze Age Europe: Manufacture and Use of Weaponry. An Interdisciplinary Research on Technology and Utilisation of Archaeological Finds, held in Vienna at the Natural History Museum, from the 30th October to 1st November 2009 and the session New Approaches on Studying Weaponry of the European Bronze Age held at the 15th annual meeting of the European Association of Archaeologists (EAA) in Riva del Garda, Italy, on the 17th September 2009. The articles cover aspects relating to archaeometallurgy, functional analyses, experimental work and archaeology and are focussing on multidisciplinary approaches for studying archaeological artefacts.

From the very beginning of research on early antiquity, weapons have always been a source of fascination and have been studied and explored from various angles throughout this time. The reasons for the attraction of this subject are the universal interest in looking at how war and warfare affected human life, a topic still pertinent in this day and age, as well as the fine craftsmanship needed to ‘create’ these weapons. In most cases, the weapons being produced are the result of the application of the most advanced technologies of a culture. Therefore, they can be used as a good example of the scientific, technical and ethical advances of a culture or society. Warfare and weapons are a prominent topic in Bronze Age research and ‘heroes’ of the period were depicted as warriors in contemporary images. These ‘men of war’ are traditionally characterised in scholarship by their weaponry, but despite this many fundamental aspects of weapon manufacture, use and functionality have not been comprehensively analysed. The consequence is that we are left with a still incomplete picture of this part of Bronze Age life. For instance, many weapons are believed to be of non-practical use in combat and are therefore regarded simply as weapons of display or weapons of a purely ceremonial character. How should the Bronze Age warrior be interpreted in such a narrative? Over the last few years, several European countries (e.g. Ireland, Great Britain, Germany, Austria and Lithuania) have launched national research projects on Bronze Age weaponry. All

Marion Uckelmann and Marianne Mödlinger London/Klagenfurt, March 2011 I

WARFARE IN BRONZE AGE EUROPE: MANUFACTURE AND USE OF WEAPONRY CONTENT Introduction Marianne Mödlinger, Marion Uckelmann and Steven Matthews

1

The smith’s approach Spearheads and swords – The making of bronze objects Markus Binggeli

11

Halberds Use-wear on Bronze Age halberds: The case of Iberia Dirk Brandherm

23

Ceremonial or deadly serious? New insight into the function of Irish Early Bronze Age halberds Ronan O’Flaherty, Michael D. Gilchrist and Trevor Cowie

39

Deliberate destruction of halberds Christian Horn

53

Swords Use-wear analysis and use-patterns of Bronze Age swords Barry Molloy

67

Chelsea and Ballintober swords: Typology, chronology and use Steven Matthews

85

Irish swords: Use and abuse Ian Colquhoun

107

Integrating form, function and technology in ancient swords. The concept of quality Marc Gener

117

Slash and stab: Some remarks on the manufacture, function and deposition of carp’s-tongue swords (summary) Dirk Brandherm and Magdalena Moskal-del Hoyo The discovery of a new Tachlovice sword at Pont-sur-Seine, Aube, France Sylvie Boulud-Gazo

125 133

Swords and new analysis methods Ritual object or powerful weapon – The usage of Central Europe Bronze Age swords Marianne Mödlinger

153

Neutron resonance capture analysis and the Bronze Age Hans Postma, Peter Schillebeeckx and Marianne Mödlinger

167

III

Bow and arrow The Sintashta bow of the Bronze Age of the South Trans-Urals, Russia Andrey Bersenev, Andrey Epimakhov and Dmitry Zdanovich

175

Shields The function of Bronze Age shields Marion Uckelmann

187

Metallurgical and archaeological investigations on the Macedonian shields of Staro Bonče (3rd century BC) Paolo Piccardo and Dusko Temelkoski

201

Abstracts from the workshop in Vienna 2009 The results of the X-rays examinations of swords from Poland Michał Bugaj

213

‘By Weapons made Worthy’? Production, use and deposition of Bronze Age weaponry David Fontijn

213

Metal alloy and texture analysis of Bronze Age swords by neutron diffraction Evelyne Godfrey, Winfried Kockelmann and Marianne Mödlinger

213

Late Bronze Age crested helmets in the Eastern Alps Andreas Lippert

214

Weapons grade bronze Peter Northover

215

Computed tomography for the 3D analysis of metallic archaeological findings Dietmar Salaberger and Johann Kastner

215

Examples of imagery in ornamentation on Late Bronze Age weapons Stefan Wirth

216

Studies on the production and function of Late Bronze Age swords Michael Siedlaczek

217

Abstracts from the Session at the 15th EAA Annual Meeting in Riva Del Garda 2009 Experimental comparison of spearheads use-wear analysis Kate Anderson

218

Halberds and Red Herrings. Weaponry and warfare in the Earlier Bronze Age of Southeast Spain 218 Dirk Brandherm, Sandra Montón, Margarita Sánchez and Gonzalo Aranda Weapons or tools? Insights into the usage of metalwork in Copper Age Italy Andrea Dolfini

219

The spirit of the sword and spear Mark Pearce

219 IV

INTRODUCTION TO ‘BRONZE AGE WARFARE: MANUFACTURE AND USE OF WEAPONRY’ Marianne Mödlinger, Marion Uckelmann and Steven Matthews produced significant and largely pragmatic perspectives on the subject of Bronze Age warfare and weaponry (e.g. Bridgford 2002), demonstrating the importance of bringing together multiple strands of knowledge. To this end, this introduction seeks to provide a brief framework of the studies undertaken so far concerning use wear analysis and the manufacture of Bronze Age weaponry. The field is broad, encompassing numerous and often disparate strands of archaeology across many countries, and consequently the discussion below is far from exhaustive. However, contemporary studies would suggest that our studies can be defined according to three different research perspectives: traditional artefact studies, with a developing trend in macro-scale use wear analysis, archaeometallurgical approaches, and finally experimental archaeology. The recognition of these three different but inter-related emergent approaches means that it is difficult to identify any single beginning for the study of wear and tear on Bronze Age weapons or tools. Each of these approaches has largely developed independently of the concerns of the others and, with the exception of artefact studies, have not been concerned exclusively with the study of the Bronze Age. All three of these approaches are represented by the papers given at the workshop and the EAA session and are collected in this volume. Whilst this suggests that there continues to be little in the way of any convergence between the three approaches, this is not in itself overly problematic, as there is clearly increasing degrees of symmetry, overlap and awareness.

The chapters in this volume all relate to research into Bronze Age weapons which are analysed using archaeometallurgical, experimental and artefact-based approaches. This encompasses not only traditional typologies but also more recent developments such as use wear analysis, which is a particular emphasis in current research. The need for an inter-disciplinary approach is essential since prehistoric warfare has been relatively fashionable in the last few decades, leading to many publications but also many assumptions, especially when it comes to the past use of weaponry. To interpret successfully combat and warfare we need to understand the archaeological evidence, the mechanical properties of the weaponry in question and the social milieu in which these events took place. The aim of the workshop in Vienna on Warfare in Bronze Age Europe: Manufacture and Use of Weaponry. An Interdisciplinary Research on Technology and Utilisation of Archaeological Finds in 2009 and the session New Approaches on Studying Weaponry of the European Bronze Age at the 15th annual meeting of the European Association of Archaeologists (EAA) in Riva del Garda 2009 was to create an opportunity where different interpretative perspectives and analytical techniques could be brought together, and to share information regarding all aspects of the use and manufacture of European Bronze Age weaponry. This included though was by no means limited to, swords, halberds, bows and arrows, as well as defensive objects like shields and helmets. The results of these two inter-disciplinary meetings are presented here. Attempts to move beyond early functionalist approaches of European Bronze Age weaponry, concerned as they were with invasions, hillforts and anachronistic sources such as the Homeric tales, saw the important identification of categories of ceremonial and symbolic weaponry (e.g. shields: Coles 1962; spearheads: Burgess, Coombs and Davis 1972; also Needham 1990). This unfortunately led some to misunderstand this development as a pacification of the past, the production of a bloodless historical narrative, resulting in the re-introduction of a primitive functionalism: weapons equal warfare. However, holistic interpretations, considering weapons within their regional archaeological contexts, have also

ARTEFACT STUDIES The study of Bronze Age metalwork represents the very foundation, not only of the study of the historical epoch with which we are concerned, but was also a fundamental and ubiquitous aspect of the origins of archaeological antiquarianism and the study of the prehistoric past itself. These origins have been, and by necessity continue to be, related to issues pertaining to chronology, technological developments and innovation, socio-economic formations, and regional identities, 1

Marianne Mödlinger, Marion Uckelmann and Steven Matthews

Fen but was more descriptive and less qualitative, and based largely on the same mode of analysis used in her earlier study of the Irish weapons. These researches were then incorporated into one of the very few detailed interpretative syntheses of Bronze Age weapons and their uses (Bridgford 2002). These published works have had a great influence on subsequent studies, including the approach used by J. York (2002) in her study of weapons and tools from the Thames, and of the few studies of use wear on spearheads, by R. Davis (2006). The study of swords, typologically and in terms of use wear analysis, remains a major research focus in Bronze Age studies, which is reflected in the contributions found in this volume. M. Mödlinger (2011; this volume) has successfully combined aspects of artefact-based approaches to use wear studies on swords with the detailed archaeometallurgical examination of traces of manufacture and wear. S. Matthews (in prep., this volume), using macro-scale analysis, has studied the long term regional use wear of rapiers and swords of the Atlantic Bronze Age, as well as advocating new methodologies of recording and dissemination of use wear. B. Molloy (e.g. 2006; 2007) has also combined these artefact based approaches with experimental archaeology in a number of studies, including swords, of which more will be said below, and in this volume has provided an overview of the study of use wear on swords. I. Colquhoun provides a discussion of current work on Irish swords, including use wear through macro-scale visual examination (this volume). S. Boulud-Gazo studies in detail a recently found sword from France of Type Tachlovice, including its manufacture and, in particular, its ornamentation. The later aspect is recently been studied as well by S. Wirth (abstract this volume). D. Brandherm and M. Moskal-del Hoyo give a summary on their combined work on Carpe’s Tongue swords (this volume). Two Masters theses have recently been finished on swords: J. H. Bunnefeld on Westphalian swords and spearheads; and M. Siedlaczek on manufacture and function on Late Bronze Age swords in Germany, based on visual examinations as well as radiographs (abstract this volume). In recent years an increasing number of studies on halberds have also integrated use wear analysis, and part of these results are represented here (Brandherm; O’Flaherty, Gilchrist and Cowie; Horn this volume). The work by R. O’Flaherty et al. combines all three approaches to explore the use of the halberds in depth. D. Brandherm delivers a detailed study on the use wear found

or in short, and despite post-processual criticisms to the contrary, the study of a particular manifestation of Bronze Age society. Artefact studies have traditionally been concerned with building the necessary material framework upon which much of the study of Bronze Age society is based, including the two other approaches discussed below. This is largely done through the application of the typological method and the production of extensive corpuses, best represented by the Prähistorische Bronzefunde series. However, an implicit concern for the condition of bronze objects has always existed. For example, frequent observations and remarks on the extent of damage can be found in the study of British Bronze Age swords by I. Colquhoun and C. Burgess (1988). Indeed, the use of these objects has always been at the forefront of their arrangement, in terms of both typology and terminology, such as in differentiating between rapiers, as thrusting weapons, and swords, as stabbing weapons, and all that this entails about society and techniques (cf. Burgess and Gerloff 1981, 113). These observations, however, were rarely presented systematically or quantitatively. This was first undertaken by K. Kristiansen (1978; 1984) who, in his study of Danish ornaments and swords, prioritised the condition of the objects themselves, and more importantly, the causes of that condition, as a central aspect of his research. Since then, discussion of damage, whether anecdotally or systematically, have become an increasingly common aspect of artefact studies, and the majority of major bronze types, both weapons and tools, have been studied. In terms of weaponry, J. Wall (1987) published a brief study of Early Bronze Age daggers in southern England. Irish swords, as well as other metalworking types, were studied using similar macro-scale approaches by G. Ramsey (1993; 1995). Daggers and swords were also included in M. Pearce’s (1998) study of Italian metalworking types, and combined both artefact based and archaeometallurgical approaches. The publication of S. Bridgford’s (1997) study concerning the context, damage and function of Irish Bronze Age swords, clearly represents a seminal study in this tradition. A more wide-ranging study dealing with British swords, as well as spearheads, was also undertaken by Bridgford (2000) in her doctoral research but was never published. The breadth of Bridgford’s research was clearly influenced by the detailed work of R. D. Savage (e.g. 1979). A further detailed study by Bridgford (2001) concerned weapons from Flag 2

Introduction

on Iberian halberds and C. Horn proposes new interpretations on the destruction and deposition of halberds across Europe. A. Dolfini’s recent work on the Italian material is also noteworthy (2011, abstract this volume). All these works represent detailed studies of the objects and combine them with one or two other approaches to change our perception of halberds. Only a few studies are concerned with the use wear on Bronze Age defensive armour, but the works describing the finds generally provide an overview of the objects condition and include traces of use (for example Coles 1962; Hencken 1971; Schauer 1980; Clausing 2005). Significant amongst these are studies by H. Born and S. Hansen (2001) looking at the manufacture and use wear of helmets from throughout Europe, and H. Born (2009) on helmets from Olympia. Recent research by one of the authors incorporated use wear and usage by macro-scale examinations of Bronze Age shields, as well as combining archaeometallurgical and experimental approaches (Uckelmann forthcoming, this volume). There is an increasing useful body of literature on spearheads, particularly in Britain; on both typology (e.g. Tarot 2000; Davis 2006; in prep.; Gedl 2009) and context (e.g. Appleby, forthcoming), but with little emphasis on use wear. Exceptions include those in G. Ramsey (1995) and J. York (2002), with Davis’s (2006) assessment of the use of Middle Bronze Age basallooped spearheads currently being the only large scale detailed study of a specific type of bronze spearhead at present. T. L. Kienlin and B. Ottaway (1998) studied the use wear and manufacture of Copper and Early Bronze Age flat and flanged axes in Central Europe whilst M. Pearce (1998) examined similar axes in Italy. A study by B. Roberts and B. Ottaway (2003) is one of the few, and one of the earliest, to have combined aspects of all three approaches. Using the typology of socketed axes established by K. P. Schmidt and C. Burgess (1982) for Northern England and Scotland, a replica was produced and tested under varying working conditions. Recent work on Italian copper-based artefacts, axes, halberds and daggers, has addressed the same questions of use, circulation and meaning (Dolfini 2011). Some more general studies looking at the condition of different objects in order to give a better interpretation of their meaning or deposition, including L. Bourke’s (2001) study of Irish rivers, J. York’s study of the metalwork from the Thames (2002), R. Taylor in his work of the hoards of Southern England (1993), and F. Falkenstein on

the river and watery finds in Southern Germany (2005). Increasingly important in these studies are issues of fragmentation, particularly with respect to use wear studies, as much of the known metal dated later in the Bronze Age appears in the form of what has been traditionally viewed as scrap. The subtleties of purposeful treatment and fragmentation of weapons and tools are important aspects of the use and condition of Bronze Age weapons and tools, particularly from hoards (cf. Needham 1990). Beginning with the seminal work by J. Chapman (2000; and later Chapman and Gaydarska 2007) on fragmentation, which included Copper Age axes from eastern Europe, purposeful fragmentation and damage has also been studied by L. Nebelsick (2000), P. Northover and S. Bridgford (2002), R. Bradley and D. Ford (2004), M. Gabillot and C. Lagarde (2008), L. Turner (2010). ARCHAEOMETALLURGY AND ARCHAEOMETRY The workshop showed the importance of using new scientific techniques (e.g. 3D computer tomography, electron probe micro-analysis, neutron diffraction, scanning electron microscopy, energy dispersive X-ray analysis, etc.) presented by material scientists who have been working with Bronze Age weaponry. Rather than simply presenting the techniques, they sought to explain the kinds of answers that we can expect and also which questions can be addressed using the different techniques. The microstructure of the objects reveals the material’s possibilities and limitations. X-rays and alloy composition analyses stand at the beginning of the use of the scientific methods in analysing Bronze Age artefacts. The first studies to use X-radiography in order to examine the manufacture of Bronze Age weaponry were conducted on German and Austrian metal hilted swords by H. Drescher (1958), J. Drie-haus (1961), H.-J. Hundt (1965) and D. Ankner (1977), the latter discussing radiographs of more than 60 Type Riegsee and Achtkant swords. Most studies still concentrated mainly on typological aspects, not on the manufacture technique itself. This changed with the application of X-radiography on swords by H. Wüstemann (1992; 1994/95), D. Brandherm and B. Sicherl (2001), H. Born and S. Hansen (2001) and M. Mödlinger (2011), focussing on the production techniques of swords. Currently, M. Bugay (abstract this volume) is working on producing radiographs of Polish swords. Further radiographs 3

Marianne Mödlinger, Marion Uckelmann and Steven Matthews

Age necropolis from Olmo di Nogara, Verona, Italy – where skeletons with sword injuries have been found. Further work is also being done on helmets (Born and Hansen 2001; P. Brun, E. Ghesquières, C. Marcigny and B. Mille, work in progress). There still is a lack of analyses on greaves or cuirasses and the number of analysed helmets remains very low, an Austrian recent find has been analysed extensively (Lippert 2010; abstract this volume). A forthcoming project by one of the authors, M. Mödlinger, on the manufacture and use of Bronze Age defensive armour in Eastern Europe will start to fill this gap. The research by I. M. Allen et al. (1970) remains the only one covering all types of Bronze Age offensive weaponry, encompassing more than 100 objects such as swords, rapiers, daggers, axes, halberds, spearheads and tools such as chisels and gouges as well. However, the majority of the metallographic research on Bronze Age weaponry are focuses on swords with some of the most extensive work was carried out by S. Bridgford (1993; 2000), and further studies e.g. by M. Longauerová et al. (1999) and M. Mödlinger (2011). Studying the structure of swords and sword-blades in particular, and other objects as well, revealed insights into the post-casting treatment – cold working, annealing – its intensity, the total amount of deformation, the hardness and the final step of production. It also allowed the investigation of the decoration to analyse whether it was cast or applied after casting. The metallographic analyses on shields are partly still in progress and will be published in the near future. They are revealing insights such as the heavily cold-worked structure and hardness of the bronze. This helps to reconstruct the process of manufacture as well as their possible properties in use in combat (Uckelmann, forthcoming). The work of P. Piccardo and D. Temelkoski (this volume) is dealing with the manufacture technique of bronze shield production using new techniques, especially on the question if a backing of leather was attached; though much younger than the Bronze Age, we found this work to be useful to be incorporated in this volume. The structure of Bronze Age weapons has been studied non-invasively using neutron diffraction. G. Artioli studied the axe of the Iceman (Ötzi) (Artoli 2007). W. Kockelmann and E. Pantos carried out analyses on a Corinthiantype bronze helmet (2006) and L. Bartoli et al. (2007), E. Godfrey (abstract this volume) as well as H. Postma et al. (2010; this volume) looked at

of other Bronze Age weapons are still scarce (e.g. but see Schwenzer 2004 on full-hilted daggers). Swords are the most researched of the Bronze Age weapons so far. The number of swords being X-rayed as well as the number of analyses – metallographic or compositional – exceeds the number of analyses carried out on spearheads, axes, daggers or armour. In the archaeological literature, the reconstruction of manufacturing techniques of Bronze Age weapons has focussed mainly on casting techniques, primarily of swords (Drescher 1961; Hundt 1965; Liebel 1985). Today, with a more practical approach and including the expertise of practicing craftsmen, other weapons are also being reconstructed (see below). Even knowing the distribution, dating and cultural context of the weaponry and documenting the traces of manufacture and usage, we are still left with an incomplete picture of the production process. Analysing the alloy composition as well as the micro-structure of the metal gives us further information on the materials the metalworker used, and if the this was due to specific demands (e.g. optimum cold working of the edges which would result in sharp and hard edges for fighting). Applying the new techniques for analysing the bronze facilitates our understanding of the casting technology and quality and thus goes beyond earlier research which focussed mainly on the joining of the hilt and blade. Only a few studies using these new techniques on Bronze Age weapons have been carried out (see below, e.g. Mödlinger 2004, on Early Bronze Age daggers, Austria; Trampuž-Orel et al. 2004, on a sword from Peggau, Austria). The composition of swords has been studied invasively using techniques such as X-ray fluorescence, scanning electron microscopy, electron probe micro-analysis (Bridgford 2000; Brown 1982; Mödlinger 2011; Northover 1988; Riederer 2004; Rovira 2007) or in more recent times with new non-invasive methods such as neutron resonance capture analysis or time of flight neutron diffraction (Mödlinger and Postma 2010; Mödlinger 2011; Postma et. al. 2010; this volume). Furthermore, the composition of other objects has been analysed including shields (Uckelmann, forthcoming), cuirasses (Michel and Mohen 1970; Lehoёrff 2008), axes (Allen et al. 1970; Kienlin and Ottaway 1998; Roberts and Ottaway 2003; Kienlin 2008; Dolfini 2011; some also including metallographic analyses), daggers (Allen et al. 1970; Schwenzer 2004; Kienlin 2005) and halberds (Rovira 2007; Dolfini 2011). In addition, I. Angelini is currently working on the daggers and swords from the Middle Bronze 4

Introduction

experiments carried out by J. Coles, S. Bridgford, R. O’Flaherty, B. Molloy, R. Davis and most recently K. Anderson (abstract this volume) still remain the only ones for European Bronze Age swords, spearheads, halberds and shields. No experiments on European Bronze Age helmets, cuirasses, greaves or non-metal weapons have yet been carried out. Due to their rareness, very few studies on organic weapons have addressed the use wear on the objects. Due to recent experiments by Molloy (2009), traces on the leather shield from Cloonbrin, Ireland, can now be identified as weapon marks. In this volume, new light is shed on the use of a composite bow in the Sintashta culture (Bersenev, Epimakhov and Zdanovich, this volume). One very good reason for the lack of further important and much needed experiments might be the difficulties in the accurate reproduction of Bronze Age weaponry, especially using Bronze Age manufacturing methods. This relies on the results of the material analyses of the original Bronze Age weapons (see above). After assembling the metallurgical data and combining the detailed study of the objects for manufacture marks, a good understanding of how the object was made might have been achieved. The challenge is then to reproduce the object. This can only be done by a skilled craftsperson who knows the metal and its properties, but these individuals are relatively rare. Several such individuals have worked and helped the people above with their experiments: Neil Burridge (www.bronzeage-craft.com), Markus Binggeli (this volume), Jeroen Zuiderwijk (www.1501bc.com/metalworking) and Frank Trommer (www.trommerarchaeotechnik.de). There are undoubted more craftspeople out there! Many of them come from a re-enactment background, were such weapons are also crafted, but generally with a loose Bronze Age framing of the production and the material. The products often look very good and close to Bronze Age originals but unfortunately are not fit for experimental testing. However, the combined knowledge and experience of both archaeologists and skilled craftspeople will help us to understand better the practical aspects of Bronze Age metalworking, particularly in terms of manufacturing processes.

Bronze Age swords, measuring different areas on the blade and gaining a broader insight into the intensity of the post-casting treatment. EXPERIMENTAL ARCHAEOLOGY Experimental archaeology on Bronze Age weaponry focuses on two main and inter-dependent areas: manufacture and usage. The experimental manufacture of weapons has a long history of research, especially as many of the casting techniques are the same for other objects. It is beyond the scope of this introduction to give a full account of the relevant literature and only a few examples are listed (Fasnacht 2010; Jantzen 2008; Ó Faoláin 2004; Ó Faoláin and Northover 1998; Jochum-Zimmermann et al. 2002; Giese et al. 2002; Tylecote 1973; Drescher 1957; 1958; 1961; Holdermann and Trommer, forthcoming; Bingelli this volume). The testing of weapons in combat is far younger and still under developed. Perhaps one of the earliest examples of the testing of Bronze Age objects was recorded by Peregrine Bertie in the Spalding Gentleman’s Society magazine of 1741 who noted that the workmen who discovered the weapons of the Ambleside hoard found that the objects were ‘So Sharp as to Cut their fingers, But since, They have Tried the metal of ‘em so often on Every Oak Gable in the Parish, That they have a Little Notched & Blunted ‘em’ (in Needham 1982, 5). The first experiments published on the usage of Bronze Age arms and armour were carried out by J. M. Coles in the 1960s and gave the field of Experimental Archaeology both its form and validity. His famous work on the testing of Bronze Age shields (e. g. Coles 1962; 1977; 1979) laid the foundations for all experimental work that was to follow, and even today his conclusions are still quoted and disputed (e.g. Molloy 2004; Uckelmann, forthcoming). For almost 40 years after Coles there was very little Experimental Archaeology addressing the usage of weaponry. However, just over a decade ago, several projects began that would reinvigorate the field: S. Bridgford started to work on use wear on sword edges (2000), B. Molloy explored the usage possibilities of swords, shields and spearheads (2004; 2008; 2009) and R. O’Flaherty (2007; this volume) experimented with replica Irish halberds (on sheep heads); and R. Davis, in his study of the Middle Bronze Age basal-looped spearheads of Britain, also conducted research into their use (2006). However, these

CONCLUSION The study of weapon use, whether directly from the damage and wear exhibited by bronzes or 5

through analysing the material and experimentation, has made an important contribution to our understanding of Bronze Age society. The research on Bronze Age weaponry so far, and especially in this volume, show the enormous potential of new answers when the three approaches of studying the material of artefact studies, archaeometallurgy and experimental archaeology are combined with each other. Future works will hopefully use this potential. The detailed study and recording of damage on weapons, as well as research and experimentation on the design and mechanical properties of bronzes, have demonstrated that practical and utilitarian as well as ceremonial and symbolic uses and categories of weapons existed simultaneously, often drawing on the same forms and designs, and clearly played an important role in European Bronze Age society. In order to move forward we must also go beyond approaches based largely on anecdote. Our studies must be tempered with both quantitative and qualitative data, accessible and verifiable, just as our interpretations must begin to be more complex and subtle so as to avoid caricature. This volume seeks to contribute to this aim. We have many sources of evidence waiting in our local, national and international museum collections to be studied.

flight neutron diffraction. Il Nuovo Cimento C 30 (1) 21–26. Born, H. and Hansen, S. 2001. Helme und Waffen Alteuropas. Sammlung Axel Guttmann IX. Mainz, von Zabern. Born, H. 2009. Die Helme des Hephaistos. Handwerk und Technik griechischer Bronzen in Olympia. München, Minerva. Bourke, L. 2001. Crossing the Rubicon: Bronze Age Metalworking from Irish Rivers. Galway, National University of Ireland, Galway. Bradley, R. and Ford, D. 2004. A long-distance connection in the Bronze Age. Joining fragments of a Ewart Park sword from two sites in England. In H. Roche, E. Grogan, J. Bradley, B. Raftery, and J. Coles (eds), From Megaliths to Metals. Essays in Honour of George Eogan, 174–177. Oxford, Oxbow. Bridgford, S. D. 1993. The Swords of the Late Bronze Age in Ireland. Unpublished MA thesis, Sheffield University. Bridgford, S. D. 1997. Mightier than the pen? (An edgewise look at Irish Bronze Age swords). In J. Carman (ed.), Material harm: Archaeological Studies of War and Violence, 95–115. Glasgow, Cruithne Press. Bridgford, S. D. 2000. Weapons, Warfare and Society in Britain 1250–750 BC. Unpublished PhD thesis, University of Sheffield. Bridgford, S. D. 2001. Late Bronze Age swords and spears from the Power Station and Flag Fen. In F. Pryor (ed.), The Flag Fen Basin: Archaeology and Environment of a Fenland Landscape, 309–317. London, English Heritage. Bridgford, S. 2002. Bronze and the first arms race – cause, effect or coincidence? In B. S. Ottaway and E. C. Wager (eds), Metals and Society. British Archaeological Reports, International Series 1061, 123–132. Oxford, BAR Publishing. Brown, M. A. 1982. Swords and sequence in the British Bronze Age. Archaeologia 107, 1–42. Burgess, C. B. and Gerloff, S. 1981. The Dirks and Rapiers of Great Britain and Ireland. Prähistorische Bronzefunde IV, 7. München, C. H. Beck. Chapman, J. 2000. Fragmentation in Archaeology: People, Places, and Broken Objects in the Prehistory of south-eastern Europe. London, Routledge. Chapman, J. and Gaydarska, B. 2007. Parts and Wholes: Fragmentation in Prehistoric Context. Oxford, Oxbow. Clausing, C. 2005. Zwei neue urnenfelderzeitliche Bronzehelme mit Scheitelknauf.

BIBLIOGRAPHY Allen I. M., Britton, D., Coghlan, H. H. 1970. Metallurgical Reports on British and Irish Bronze Age Implements and Weapons in the Pitt Rivers Museum. Occasional Papers on Technology 10. Oxford, Pitt Rivers Museum. Anderson, K. forthcoming. Slashing and thrusting with Late Bronze Age spears: analyses and experiment. Antiquity (2011). Ankner, D. 1977. Röntgenuntersuchungen an Riegseeschwertern. Ein Beitrag zur Typologie. Archäologie und Naturwissenschaften 1, 269– 459. Appleby, G., forthcoming. British Bronze Age spears. Hoard pattering and significance. Bronze Age Review 2. Artioli, G. 2007. Crystallographic texture analysis of archaeological metals: interpretation of manufacturing techniques. Applied Physics A 89, 899–908. Bartoli, L., Siano, S., Kockelmann, W., Santisteban, J., James, J., Miccio, M. and de Marinis, G. 2007. Non-destructive characterisation of a Villanovan sword using time-of6

Introduction

Neolithic to the Metal Ages: Technical and Codified Practices. British Archaeological Reports, International Series 1758, 59–65. Oxford, BAR Publishing. Gedl, M. 2009. Die Lanzenspitzen in Polen. Prähistorische Bronzefunde V, 3. Stuttgart, Steiner. Giese E., Schwämmle, K. and Trommer, F. 2002. Bronzeguss – Eine Versuchsreihe zur Technik des prähistorischen Bronzegusses unter dem besonderen Aspekt des Formenmaterials. Experimentelle Archäologie in Europa, Bilanz 2002, Heft 1, 93–106. Hundt, H.-J. 1965. Produktionsgeschichtliche Untersuchungen über den bronzezeitlichen Schwertguss. Jahrbuch des RömischGermanischen Zentralmuseums Mainz, 41– 58. Jantzen, D. 2008. Quellen zur Metallbearbeitung im nordischen Kreis der Bronzezeit, Prähistorische Bronzefunde XIX, 2. Stuttgart, Steiner. Jochum-Zimmerman, E., Künzler Wagner, N. and Kunnert, U. 2002. Zurück zur Gussform! Zum Einfluss des Gussformmaterials auf die Mikrostruktur eines gegossenen Bronzeobjektes. Experimentelle Archäologie in Europa, Bilanz 2002, Heft 1, 79–91. Hencken, H. 1971. The earliest European helmets. Bronze Age and Early Iron Age. American School of Prehistoric Research, Bulletin 28. Cambridge/Mass., Peabody Museum. Holdermann, C.-S. and Trommer, F., forthcoming (2011). Organisation, Verfahrenstechnik und Arbeitsaufwand im spätbronzezeitlichen Metallhandwerk. Prähistorische Bronzefunde XX, 13. Stuttgart, Steiner. Kienlin, T. L. and Ottaway, B. 1998. Flanged Axes of the North-alpine Region. An Assessment of the Possibilities of Use Wear Analysis on Metal Artifacts. In: C. Mordant, M. Pernot and V. Rychner (eds), L’Atelier du bronzier du XXe au VIIIe siecle avant notre ere. Actes du colloque international Bronze’96. Neuchatel et Dijon II. Du mineral au metal, du metal a l’objet, 271–286. Paris, CTHS. Kienlin, T. L. 2005. Frühbronzezeitliche Vollgriffdolche und Randleistenbeile: Zu Herstellungstechnik, Zusammensetzung und Materialwahrnehmung. Archäologisches Korrespondenzblatt 35, 175–190. Kienlin, T. L. 2008. Frühes Metall im nordalpinen Raum. Eine Untersuchung zu technologischen und kognitiven Aspekten früher Metallurgie

Archäologisches Korrespondenzblatt 35, 31– 38. Coles, J. M. 1962. European Bronze Age shields. Proceedings of the Prehistoric Society 28, 156–190. Coles, J. M. 1977. Parade and Display Experiments in Bronze Age Europe. In V. Markotic (ed.), Ancient Europe and the Mediterranean: Studies presented in Honour of Hugh Hencken, 51–58. Warminster, Aris and Phillips Ltd. Coles, J. M. 1979. Experimental Archaeology. London, Academic Press. Colquhoun, I. and Burgess, C. B. 1988. The Swords of Britain. Prähistorische Bronzefunde IV, 5. München, C. H. Beck. Davis, R. 2006. Basal-Looped Spearheads: Typology, Chronology, Context and Use. British Archaeological Reports, International Series 1497. Oxford, BAR Publishing. Davis, R. in prep. The Early and Middle Bronze Age spearheads of Britain (with a contribution by P. Northover). Prähistorische Bronzefunde V. Stuttgart, Steiner. Dolfini, A. 2011. The function of Chalcolithic metalwork in Italy: an assessment based on use-wear analysis. Journal of Archaeological Science 38, 1037–1049. Drescher, H. 1957. Der Bronzeguss in Formen aus Bronze, Die Kunde 8, 52–75. Drescher, H. 1958. Der Überfangguss. Ein Beitrag zur vorgeschichtlichen Metalltechnik. Mainz, RömischGermanisches Zentralmuseum. Drescher, H. 1961. Die Gusstechnik des Vollgriffschwertes aus Meckelfeld, Kreis Harburg, Hammaburg 7, 57–66. Driehaus, J. 1961. Röntgenuntersuchungen an bronzenen Vollgriffschwertern, Germania 39, 1961, 22–31. Falkenstein, F. 2005. Zu den Gewässerfunden der älteren Urnenfelderzeit in Süddeutschland. In B. Horejs, R. Jung, E. Kaiser and B. Teržan (eds), Interpretationsraum Bronzezeit. B. Hänsel von seinen Schülern gewidmet, 491–504. Bonn, Habelt. Fasnacht, W. 2010. 20 Jahre Experimente in der Bronzetechnologie – eine Standortbestimmung. Experimentelle Archäologie in Europa, Bilanz 2010 [Europäische Vereinigung zur Förderung der Experimentellen Archäologie e.V.], Heft 9, 117–126. Gabillot, M. and Lagarde, C. 2008. Voluntary destructions of objects in Middle and Late Bronze Age hoards in France. In C. Hamon and B. Quilliec (eds), Hoards from the 7

Marianne Mödlinger, Marion Uckelmann and Steven Matthews

anhand der Gefüge frühbronzezeitlicher Beile. Universitätsforschungen zur prähistorischen Archäologie 162. Bonn, Habelt. Kristiansen, K. 1978. The consumption of wealth in Bronze Age Denmark: a study in the dynamics of economic processes in tribal societies. In K. Kristiansen and C. Paludan-Müller (eds), New Directions in Scandinavian Archaeology, 158–191. Copenhagen, National Museum of Denmark. Kristiansen, K. 1984. Krieger und Häuptlinge in der Bronzezeit Dänemarks. Ein Betrag zur Geschichte des bronzezeitlichen Schwertes. Jahrbuch des Römisch-Germanischen Zentralmuseums Mainz 31, 187–208. Lehoёrff, A. 2008. Les cuirasses de Marmesse (Haute-Marne), un artisanat d‘exception. Antiquites Nationales 39, 95–106. Liebel, D. 1985. Rekonstruktion des Bronzeschwertes von Stenn in originalgetreuer Technik. Restaurierung und Museumstechnik 6, 39–53. Lippert, A. 2010. Spätbronzezeitliche Kammhelme an transalpinen Fernhandelsrouten (mit einem Beitrag von M. Mehofer). Archaologisches Korrespondenzblatt 40,49–65. Longauerová, M., Longauer, S. and Makarius, P. 1999. Metallography of ancient Bronze Swords. In: L. Mihok and E. Miroššayová (eds), Východoslovenský Pravek, Special Issue: Archaeometallurgy in the Central Europe, 77–95. Košice, Archeologický ústav SAV. Matthews, S. in prep. Techniques and Society: the Middle and Late Bronze Age weapons of northwest Europe. PhD thesis, University of Groningen. Mödlinger, M. 2004. Metallographische Untersuchungen an ausgewählten Bronzefunden aus den frühbronzezeitlichen Gräberfeldern von Hainburg/Teichtal und Mannersdorf a. Lgb., Niederösterreich. Abschlussbericht für das ÖNB-Projekt 9677. Sozio-ökonomische, bevölkerungsbiologische und chronologische Parameter im rituellen Kontext des Bestattungswesens der Frühbronzezeit am Beispiel der Wieselburg-Kultur’ (Projektleitung: Alexandra Krenn-Leeb), Wien 2004. Mödlinger, M. 2011. Herstellung und Verwendung bronzezeitlicher Schwerter Mitteleuropas. Eine vertiefende Studie zur mittelbronzeund urnenfelderzeitlichen Bewaffnung und Sozialstruktur, Universitätsforschungen zur prähistorischen Archäologie 193. Bonn, Habelt.

Mödlinger, M. and Postma, H. 2010. Nicht-invasive Legierungsbestimmungen mittels Neutro nenresonanzabsorption (NRCA) an zwei bronzezeitlichen Schwertern aus dem Joanneum Graz. Schild von Steyr 23, 132–137. Michel, F. and Mohen, J. P., 1970. Étude comparée de deux cuirasses hallstattiennes, la cuirasse „de Grenoble“et la cuirasse „de Naples“. Annales du Laboratoire de recherche des musées de France, 65–80. Molloy, B. P. C. 2004. Experimental Combat with Bronze Age Weapons. Archaeology Ireland 18, 32–34. Molloy, B. P. C. 2006. The role of combat weaponry in Bronze Age society: The cases of Aegean and Ireland in the Middle and Late Bronze Age. Unpublished PhD thesis, University College Dublin. Molloy, B. P. C. 2007. What’s the bloody point?: Bronze Age swordsmanship in Ireland and Britain. In. B. P. C. Molloy (ed.), The Cutting Edge: Studies in Ancient and Medieval Combat. 90–111. Stroud, Tempus. Molloy, B. P. C. 2009. For Gods or men? A reappraisal of the function of European Bronze Age shields. Antiquity 83, 1052–1064. Nebelsick, L. 2000. Rent asunder: Ritual violence in Late Bronze Age hoards. In C. F. E. Pare (ed.), Metals Make the World Go Round, 160–175. Oxford, Oxbow. Needham, S. 1982. The Ambleside Hoard: A Discovery in the Royal Collections. British Museum, Occasional Paper No. 39. London, British Museum. Needham, S., 1990. Middle Bronze Age ceremonial weapons: new finds from Oxborough, Norfolk and Essex/Kent. The Antiquaries Journal 70, 239–252. Northover, J. P. 1988. The analysis and metallurgy of British Bronze Age swords. In I. Colquhoun and C. Burgess, The swords of Britain. Prähistorische Bronzefunde IV, 5, 130–146. München, C. H. Beck. Northover, J. P. and Bridgford, S. D. 2002. The characterization of a Bronze Age weapon hoard. In P. B. Vandiver, M. Goodway and J. L. Mass (eds), Materials Issues in Art and Archaeology VI, 159–166. Warrendale, Materials Research Society. Ó Faoláin, S. and Northover, J. P. 1998. The Technology of Late Bronze Age Sword Production in Ireland. Journal of Irish Archaeology 9, 69–88. Ó Faoláin, S. 2004. Bronze Artefact Production in Late Bronze Age Ireland. A survey. British 8

Introduction

Archaeological Reports, British Series 382. Oxford, BAR Publishing. Pearce, M. 1998. Reconstructing prehistoric metallurgical knowledge: The northern Italy Copper Age and Bronze Ages. European Journal of Archaeology 1, 51-70. Postma, H., Armkreutz, L., Borella, A., Clarijs, M., Kamermans, H., Kockelmann, W., Paradowska, A., Schillebeeckx, P. and Visser, D. 2010. Non-destructive bulk analysis of the Buggenum sword by neutron resonance cap-ture analysis and neutron diffraction. Journal of Radioanalytical and Nuclear Chemistry 238/3, 641–652. Ramsey, G. 1993. Damaged butts and torn rivet holes: The hafting and function of Middle Bronze Age “dirks“ and “rapiers“. Archeomaterials 7, 127–138. Ramsey, G. 1995. Middle Bronze Age metalwork: Are artefact studies dead and buried? In J. Waddell and E. Shee Twohig (eds), Ireland in the Bronze Age, 49–62. Dublin, The Stationary Office. Riederer, J. 2004. Ergebnisse metallanalytischer Untersuchungen von Bronzeschwertern. In H. Wüstemann, Die Schwerter in Ostdeutschland. Prähistorische Bronzefund IV, 15, 259–329. Stuttgart, Steiner. Roberts, B. and Ottaway, B. S. 2003. The use and significance of socketed axes during the Late Bronze Age. European Journal of Archaeology 6 (2), 119–140. Rovira Llorens, S. 2007. Las espadas del Broncefinal de la Peninsula Iberica: estudio arqueometalurgico. In D. Brandherm, Las espadas del Bronce Final en la Peninsula Iberica y Baleares. Prähistorische Bronzefunde IV, 16, 155–175. Stuttgart, Steiner. Savage, R. D. A. 1979. Technical notes on the Watford sword fragments. In C. Burgess and D. Coombs (eds), Bronze Age Hoards: Some Finds Old and New. British Archaeological Reports, British Series 67, 221–228. Oxford, BAR. Schauer, P. 1980 (1982). Der Rundschild der Bronze- und frühen Eisenzeit. Jahrbuch des Römisch-Germanischen Zentralmuseums Mainz 27, 196–248. Schmidt, K. P. and Burgess, C. 1982. The axes of Scotland and Northern England.

Prähistorische Bronzefunde IX, 7. München, C. H. Beck. Schwenzer, S. 2004. Frühbronzezeitliche Vollgriffdolche. Typologische, chronologische und technische Studien auf der Grundlage einer Materialaufnahme von Hans-Jürgen Hundt. Kataloge vor- und frühgeschichtlicher Altertümer 36. Mainz, Römisch-Germanisches Zentralmuseum. Tarot, J. 2000. Die bronzezeitlichen Lanzenspitzen der Schweiz. Universitätsforschungen zur Prähistorischen Archäologie 66. Bonn, Habelt. Taylor, R. 1993. Hoards of the Bronze Age in Southern Britain: Analysis and Interpretation. British Archaeological Reports, British Series 228. Oxford, BAR. Trampuž-Orel, N., Drglin, T., Urankar, R. and Orel, B. 2004 Inductively coupled plasma – atomic emission spectroscopy analyses of the Peggau hoard. In A. Weihs, Der urnenfelderzeitliche Depotfund von Peggau (Steiermark). Universitätsforschungen zur Prähistorischen Archäologie 114, 203–227, Bonn, Habelt. Tylecote, R. 1973. Casting copper and bronze into stone moulds. Journal of the Historical Metallurgy Society 7, 1–5. Turner, L. 2010. A re-interpretation of the later Bronze Age metalwork hoards of Essex and Kent. British Archaeological Reports, British Series 507. Oxford, BAR Publishing. Uckelmann, M. forthcoming. Die Schilde der Bronzezeit in Nord-, West- und Zentraleuropa. Prähistorische Bronzefunde III, 4. Stuttgart, Steiner. [PhD thesis, University of Münster 2008] Wall, J. 1987. The role of daggers in Early Bronze Age Britain: The evidence of wear analysis. Oxford Journal of Archaeology 6 (1), 115– 118. Wüstemann, H. 1992. Jungbronzezeitliche Vollgriffschwerter mit Bleifüllung, Arbeitsund Forschungsberichte zur sächsischen Bodendenkmalpflege 35, 39–49. Wüstemann, H. 1994/95. Die radiographische Auswertung der Schwertfunde aus dem Hortfund Berlin-Buch, Acta Praehistorica et Archaelogica 26/27, 122–128. York, J. 2002. The life cycle of Bronze Age metalwork from the Thames. Oxford Journal of Archaeology 21 (1), 77–92.

9

10

SPEARHEADS AND SWORDS – THE MAKING OF BRONZE OBJECTS Markus Binggeli

the original workshop.These projects were carried out on the complete spectrum of metals used in prehistory: bronze, silver, gold, and iron. The starting point is always an original prehistoric object. This chapter briefly describes aspects of Bronze Age weapon production techniques and discusses two examples: firstly spearheads cast in clay-moulds and secondly blades cast in sandstone moulds.

ABSTRACT Fragments of clay moulds for spearheads were found at Zug-Sumpf. After these original finds a mould was made and used for experimental testing. With the help of a model negatives were imprinted in the two halves of the mould. The composition of the clay material was as similar as possible to the composition of the moulds found. In the moulds used in the experiments spearheads could successfully be cast in possibly the same procedure as at Zug-Sumpf in the Bronze Age. Attempts to cast swords in a sandstone mould similar to one from Neckargartach were successful as well. During the experiments the premises and problems of good castings were explored. The casting and also the further processes involved in finishing the cast blade were researched through experiments. The elaborate steps of the hammering, grinding and polishing as well as decorating the blades are described. Unfortunately our knowledge about the tools (except about the grinding stones) used for these processes is scarce, especially for the Early Bronze Age. The archaeological evidence is not sufficient to give a clear picture about the further working processes of the blades in this period.

THE CLAY MOULDS OF SPEARHEADS FROM ZUG-SUMPF Description of the moulds The surviving mould fragments of clay from the site of Zug-Sumpf, Switzerland, were all, where still discernable, used for making spearheads (fig. 1). T. Weidmann (1982, 72–76) published an exact description of the mould’s construction, which I summarise here: - The clay mould was formed from a model. - They are composed of four main parts: two fitting half sections, with the negative imprint of the spearhead to cast; a conical shaped core, which during casting kept the socket-opening free; and an outer cover, which kept these three parts together. A cord for fixing the two half sections and a copper bolt which held the clay core in place have not survived, but can be deduced through imprints in the moulds. - The formed clay is well fired to a high temperature. The outer parts are brick-red or grey, and the inner lying parts are black.

KEYWORDS Zug-Sumpf – Neckargartach – sword production – spearhead production – sandstone mould – clay mould – Bronze Age – experimental casting – grinding – chasing

The shaping material

INTRODUCTION

The material from which the casting moulds was made was analysed by J. Bonzon, Institut für Mineralogie und Petrographie der Universität Freiburg (2004, 144–149). Following these results the material for the casting mould was reconstructed, which is as close as possible to the composition of the original moulds from ZugSumpf. Since Bonzon does not state the shares of different tempering materials, except for the bone splinters, I also used the specifications made by M. Fröhlich (1981; 1995), K. R. Hedegaard

The following remarks are based on experiences and results from 18 years of practising with prehistoric metalworking techniques. The underlying aim has been the exploration of the manufacture processes of single objects, reaching from the Copper Age to Early Medieval times. In each case the study comprised the whole production procedure of the object, from the making of an original form through casting and/or forging, and following the steps till completion, using (if possible) the required tools and infrastructure of 11

Markus Binggeli

1

2

3

4

5

6

Figure 1: Fragments of the clay mould from Zug-Sumpf Figure 2: The clay mould is formed from a model Figure 3: The two halves of the mould with inserted core Figure 4: The mould pieces are fixed with a cord Figure 5: The mould pieces are covered with a layer of clay and ready for casting Figure 6: Opened mould with cast 12

Spearheads and swords – The making of bronze objects

(1995) and A. Michaels (1985; 1986), and reconstructed the composition of the material as follows: 50 % clay, 30 % cow dung, 10 % cow hair, 8 % granite, and 2 % bones. The different components were dried, grinded, and sieved through a kitchen sieve, and a maximum particle size of 1.5 mm was achieved. To crush the clay and the granite was rather easy, but working the bones was quite challenging. Our experiments showed that even without the bonemeal, the forming material is able to produce flawless casting results, and it leaves the question why such a material, which required so much of an effort to make, was used. From different sites in the region of Bern, Switzerland, I collected three types of lean, sandy clay. One of these was very useful in producing, together with the other parts described in the above composition, a forming material that was malleable enough and as well visually very close to the originals from Zug-Sumpf. The components were mixed dry, stirred together with dark beer, and kneaded into a formable dough, which could be used instantly without prior storage. The usage of beer gives the advantage of increasing the dry-break-resistence.

the socket can be formed, by rolling a piece of clay, and cutting in the rills corresponding to those in the halves for pouring in the casting metal. A copper pin is layered with a clay slip and put through the core to fix it in the assembled mould (fig. 3), and forms through the holes in the cast which are used to fasten the spearhead with nails to the shaft. These three pieces of the mould were assembled and fixed with a cord, until they are almost completely dried (fig. 4). The mould pieces have to be controlled during the drying process in case they distort or shrink, and have to be corrected if neccessary. When all parts are well dried to a leather-hard stage, the parts are bound again firmly together with the fitted core and the whole assemblage is coated with a 3–8 mm thick clay layer (fig. 5). The thickness of the clay layer is up to 10–12 mm in the lower part of the mould and 13–15 mm in the upper part. It is important that the layer is not too thick, so that during the casting the enclosed air in the mould cavity can escape through the coating layer. The mould is fitted without any venting channels. At the final stage the mould has to dry completely. The manufacturing of multi-piece moulds must be considered as time intensive and at a difficult level of precision! The amount of work invested per spearhead is considerably greater compared to the use of moulds of sandstone. However, in my experience the success of casting is greater using clay moulds instead of sandstone, which might be connected to the better venting in the clay mould. This could have made up the disadvantage of the time intensive production of the clay mould.

The making of the moulds The negatives in the moulds are shaped from a model (fig. 2). This can be formed in wood or by an original bronze spearhead, in which socket a piece of wood is fitted. The piece of wood helps to form the area of the feeder where in the mould would be attached to the casting core. The feeder can be formed of course freehand as well, but it is much more difficult. In Ireland wooden models for the production of bronze moulds have survived (Hodges 1954), but bronze originals could also have been used. Using models from different materials does not affect the manufacture significantly. The tempered forming material does not sticks to the metal models as well. The two valves, with the negatives of the model, are the central piece of the mould. The halves found in Zug-Sumpf were manufactured without any connecting parts to keep the moulds attached during the cast, but comparable moulds from Mörigen, Switzerland, show such fittings (Viollier 1924). Once these halves are in a leather-hard dried stage, rills for fixating the pin that holds the core in place are carved in as well as in the upper part of the mould the channels in the feeder for the casting material. As soon as the halves are firm enough, a conical core for

The Firing After the complete drying the moulds are burnt in an open fire. This can be done without any great precaution, since the tempered clay is able to withstand such heat shocks. The moulds stay in the fire, until the inner side is glowing red. They are then at 750°C. It is important to keep the feeder free from charcoal pieces during the firing, since this could lead to a spoiled cast. When the moulds are annealed, they should be taken from the fire and left to cool down until no annealing colour is visible. In the experiments the moulds were immediately filled with the cast, at an estimated temperature of 300°C. After the experiment these moulds showed the same colouring as the original pieces from Zug-Sumpf, grey to brick-red on the outside and black on the inside (fig. 6). The light coloured areas are a bit 13

Markus Binggeli

7

9

8

Figure 7: Spearhead, finished (after the moulds from Zug) Figure 8: Sandstone mould from Neckargartach – the longitudinal sides show the negatives of different swords (after Seidel 1995) Figure 9 Sandstone mould for a short sword from Castione – the blades profile is carved into the mould (after Bernabo 1997)

14

Spearheads and swords – The making of bronze objects

and the tip of the spearhead showed blisters on the cast surface, but otherwise the cast had no flaws. The second spearhead was cast after the bronze had cooled down and a perfect cast was retrieved. The working time to make such a cast spearhead can take up to about four hours (fig. 7), without taking into account the time to obtain the raw material and infrastructure and as well without taking the spoiled casts into account. The casting procedure that was used by the casters from the Zug-Sumpf area can therefore be described as very efficient.

thicker on the original moulds, due to a stronger oxidised firing. The different coloured areas on the mould originate from differing conditions of oxidation of the clay material, which develops according to the atmosphere of the firing. The clay contains iron hydroxide which under oxidising firing absorbs oxygen and becomes Fe2O3, which has a red colour. Under reducing firing it dispenses oxygen and becomes FeO, which has a black colour. The organic tempering components of the forming clay behave differently: they burn completely to ashes in areas with an oxidised atmosphere and have no influence on the colouring; in contrast under a reduced atmosphere, where due to missing oxygen they cannot burn, they only carbonise and amplify the black colouring in the reduction areas. These pieces of charcoal in the clay form are important: during casting they burn the oxygen in the hollow space of the mould and allow an oxid-free cast. This delivers a significant improvement on the casting results, since no oxidised layer, which could constrain the feeding, can form on the surface of the cast metal. Actually, the cast spearheads in my experiments (after they cooled down a bit in the mould) left the mould almost completely blank (fig. 6)! This has of course a considerable advantage for any further working of the piece.

SANDSTONE MOULDS FOR SWORDS1 Observations on the moulds from Neckargartach, Germany, and Castione, Italy Moulds made out of sandstone survive better than clay moulds, but such finds are still rare and pose a challenge to everyone experimenting with Bronze Age casting technologies. There is the sword mould from Neckargartach (Seidel 1995) which is superbly well preserved and has survived with the careful cleaning of the conservators (fig. 8). The mould consists of a rectangular stone with a square section and in the lateral surfaces the negative imprints of swords are carved. The red-grey colouring of these negatives provides evidence that the mould was used for casting the bronze at a temperature of 1100°C. Peculiar are the black colourings on the rim of the negative imprints, which would not appear on an untreated mould. They must derive from carbon, which is generated when the mould is treated before the cast with an organic substance. When it comes in contact with the liquid metal this organic substance burns and produces a reducing atmosphere in the mould space. This has a positive effect on the casting result because the metal cannot form an oxidised layer and thin parts in the mould can therefore be filled more easily. Possible methods used could be the sooting of the negatives before the casting or a treatment with oil or grease. This treatment is still recommended in modern handbooks for goldsmiths for wire or ingot castings. My experience affirms the positive effect of such a treatment of a sandstone mould. The negative imprints in the stone are of simple form, and show only a rough profiling of the intended for blade. Edges, ribs or groo-

CASTING The alloy CuSn9 (81 % Cu, 9 % Sn) was used in all experiments. Of the six cast moulds three were successful, in one cast the crucible did not have enough bronze and the socket could not be cast fully, and two spearheads showed casting flaws. Therefore there was a 50 % successful outcome. This error rate can be decreased through gaining experiences. Besides an adequate forming clay and composition of the mould, in my experience, two factors are crucial fo a successful cast: the temperature of the mould and the cast metal. Is one or both of these temperatures too high, gases are accumulating during the casting in the mould, the bronze ‘boils’ and the casted piece shows flat, superficial blisters, as the gases could not discharge from the mould. I would like to describe two observations: two identically manufactured and fired moulds were filled from the same crucible with heavily over-heated cast metal. During the cast of the first spearhead the bronze boiled in the mould 1

The following text is a reworked version of Bingelli 2004.

15

Markus Binggeli

10

11

Figure 10: Sandstone mould for four swords, next to it the cast and finished blade; sword after an original from Hallau (Griffplattenschwert) Figure 11: The same mould (as in fig. 10), different side, next to it the cast and finished blade; sword after an original Type Rixheim sword from Sutz Figure 12: Not filled out part on the cast made in the replica of the mould from Castione

12 16

Spearheads and swords – The making of bronze objects

ves, which are present on most of the Bronze Age blades, are not carved into the negatives. An exceptional case is the mould from Castione, Italy (Bernabo 1997), which has a complicated blade profile worked into the stone (fig. 9). To my knowledge this kind of mould negative is unique in Europe. It means that the type of sword can be seen in the mould, but not the final detailed form of the sword, which is shaped after the casting. That means also that the same mould can deliver different blades that cannot be traced back to the one mould. The blade edge is similar, although in the mould negative it is not sharp, which results in an edge thickness of 1–2 mm for the cast blade. Along the edges of the negatives small channels or grooves, so called vents, allow the locked-in air and cast gases out of the mould hollow during the casting. On the mould from Neckargartach the runner is at the handle end (Griffangel), and on the mould from Castione the runner is placed at the blade’s tip. The casters fitted the runner after their experiences and where it appeared most advantageous in regard of the durability of the mould and for the best casting results. In this case compromises often have to be taken, since the mould is often damaged through the shrinkage of the bronze while cooling down. For example, if a blade with a complicated handle part for a grip-tongue-sword (Griffzungenschwert) is cast from the blade’s tip, the chances for a good cast of the handle part are quite high. Due to the shrinkage of the metal in the stone mould, the area of the handle parts can be destroyed and the mould would become useless. When casted from the handle part, there is a lower pressure of the metal during the casting, and the risk of casting uncompleted handle part is high. A well constructed and treated mould of sandstone can perform ten to twenty and plus casts and is therefore more economical than a clay mould, which has to be manufactured for each casting.

Premises for a successful casting A preheating of the mould before the cast is not necessary and it is enough to warm the mould next to the fire until all humidity is discharged from the stone. This has two advantages: one can assemble and bind the mould together without burning your hands, and secondly the large stone mould does not distort, which often happens during preheating, and would result in a leaking mould and a flawed casting. It is also important to have the right control of the fusing of the bronze and to choose the right temperature of the bronze for the casting. The bronze should be fused reductively and its surface has to be specular before casting, i.e. to be free from slags and oxide layers. The question arises at this point if reductive fluxing agents (for example salt, potash, soda) where already used in the Bronze Age (a subject which has not been researched as far as I know). Through practice and experience one learns to read the cast temperature without a thermometer alone due to appearance of the metal surface and the annealing colour. The same holds true for the right fire control. If the casting temperature of the metal is too low it will result in the casting of overly hot metal in a flawed cast, with the same outcome as described while casting in clay moulds. Problems The height of a mould such as the one from Neckargartach is about 70 cm, which has the consequence of a rather high pressure in the lower part of the mould during casting. The quality of the casting is therefore mostly, especially in this lower part, very good. Dammed gases often cause problems while casting blades in the area where the mould narrows towards the blade edges. This becomes visible through flawed areas along the blade edge of the casted piece (fig. 12). Sometimes these flaws can be beaten out to regain a functional blade. These problems appear more often, when casting daggers, since the pressure during casting is lower within to the shorter mould. Further problems include gas bubbles and porous areas in the unfinished casting. Both phenomena appear mainly in the upper third of the mould and can also be found on Bronze Age blades. An increased casting temperature encourages their formation. The gas bubbles are not visible in the unfinished casting, and become only apparent when the casted piece is polished and enough material from the surface is removed. This is due to the properties of copper, which in a melted con-

The casting of blades in a sandstone mould I used tin-bronze with 8–10 % tin as this alloy has a good castability and is good for forging, which is important for the finishing of the blades. Bronze alloys of comparable composition were used as well in the Bronze Age for producing blades. I prepared a mould of sandstone like the one from Neckargartach to cast sword blades and tested it with good results (fig. 10. 11). 17

Markus Binggeli

13

14

15

16

17

18

Figure 13: Heating distorted the mould halves, and they gape open, the cast becomes much too thick; sword after an original from Gaggio Figure 14: Sword blade with thickened and rounded midrib, blade edges are stepless; sword after an original from Brügg Figure 15: Sword blade with a rhombic section and narrow, stepped blade edges; sword after an original from Sutz Figure 16: Sword blade with pronounced midrib and concave edges, small stepped blade edges with rim along the edge; sword after an original from Hallau Figure 17: Sword blade with pronounced midrib, stepped mid part, broad edge rim; sword after the original from Castione Figure 18: Not removed hammer-marks in the hilt part of a finished blade; sword after an original from Brügg 18

Spearheads and swords – The making of bronze objects

19

20 Figure 19: Late Bronze Age grinding stones from Zug-Sumpf Figure 20: Late Bronze Age grinding plates made of sandstone from Zug-Sumpf

time consuming part of the manufacture of a bronze sword. The casting of the sword is the most spectacular, but also the shortest part concerning the work time. The finishing is the part that is least well known and explored. What happens to the metal structure has been researched in detail, but little is known of the tools which performed the work. Only for the Later Bronze Age are some metal working tools apparently established (Wyss 1971). Tools from earlier periods in the Bronze Age are rather uncommon or not as such identified. It can be assumed that the forging work was carried out with stone tools.

dition, absorbs gases. This decreases with higher tin levels in the melted bronze. Therefore, especially for low tin-bronze alloys, melting has to be conducted using a strong reductive firing process. Porous areas formed through the shrinkage of the metal in the mould happen when the metal in the runner solidifies before the metal in the mould’s inner part. This occurrence can be rarely avoided completely while using Bronze Age moulds, due to their short and narrow runner section where the metal is poured in – especially with the long sword moulds. Both ways, pouring from the hilt or the tip, can create a potential weakening of the finished blade, and the caster can choose when making the mould and the runner position if they prefer a potential weak spot in the blade tip or in the hilt part. Such a long mould of sandstone provides another difficulty during casting: the metal poured in heats the mould interior causes it to expand, which subsequently can lead to the bending apart of the mould halves. After casting the mould halves can diverge up to 5 mm. In such a condition the mould is useless, because the mould halves do not fit together anymore and the metal would flow out of the mould (fig. 13). They have to be left to cool down. The mould halves also bend apart during the casting when the metal is poured in too slowly. This results in the upper part of the sword cast being too thick, which means that the sword blank is potentially useless for any further processing.

Forging Firstly the waster and the casting seams are removed from the casted blade and the contact points abraded. Then the whole blade is forged, the coarse cast surface is smoothed over, the profile of the blade is further worked out and the density of the metal structure is increased. In this point it is important that the anvil or bedding layer and the profile of the hammer are in accordance with the envisioned blade profile. Afterwards it is advised to anneal the blade and quench it in water, so the metal structure can recrystallise leaving the blade softer, which is advantageous for any further working. The annealing can be skipped, but then careful further working has to take place, the metal structure has to be kept under the breaking limit, otherwise fractures might appear on the blade. If the annealing is undertaken the whole blade should be wrought again giving it a heightened hardness and more resilient properties. A freshly cast blade is soft, not resilient and will not spring

The finishing of sword blades The further processing from the unfinished casting to the finished sword blade is the most 19

Markus Binggeli

lines and rills are chiselled into the blade for decorative purposes. Through the final polishing and grinding the blade achieves its finished bright and gleaming appearance. To what extent the chasing has been performed with tools made from iron in the latest phases of the Bronze Age is still undefined. Since iron was used as inlays on jewellery and on the hilts of weapons, it can be assumed that some of the tools were made of iron as well (Speck 1981). Evidence is provided indirectly through traces of work on bronze objects that could only be undertaken with a harder material than the bronze used (Museum Schwab Biel 2004, 86).

back after bending in its original form as a worked-over blade will. After this, the shoulders and rims of the handle part are worked into their final stage. Then the blade edges are intensively hammered to gain a maximum hardness, sharpness and cutting ability – all attributes of the blade that are defined during this step. There are two main types known for working the blade edges of Bronze Age blades: a narrow stepped rim along the edge or a step-less thinning to the edge (fig. 14–17). The length and width of the blade change through over-working as much as 2–3 mm in length and 5 mm in width. Therefore, the blade has a differing form from the negative in the mould after the further processing meaning that the smith could impact on the blade form after casting.

*** The smiths of the Bronze Age were skilled craftsmen and they knew very well which methods would produce the best possible blades. Our knowledge of which tools were used to shape Bronze Age blades is very fragmentary, especially for the Early Bronze Age. The only exception are the different types of grinding tools made of stone, which are often found and recorded in excavations. The subject of used tools should be a priority in forthcoming research and their usage should be tested and explained through experimental approaches, especially the forging and hammering tools, stone hammers and anvils, as well as tools which were used to work on the blade’s shape and decoration.

Grinding work Well preserved Bronze Age blades show an evenly polished surface and a very regular worked layout of edges and chiselled rills. Hammer imprints are only found in places which were covered by the organic or metal hilt and were not visible (fig. 18). Abrasion work was sometimes substantial and decreased the material thickness, not only superficially, as becomes obvious from abraded blow holes. The grinding work which followed the forging was therefore carried out very carefully and elaborately. In this step the blade receives its final form and profile. Only the areas which are later covered by the hilt are not polished. The grinding work was undertaken with natural stones. Grinding stones of disk or rod shape have been recovered from Late Bronze Age settlements (fig. 19. 20). The comparison with the traditional manufacture of Japanese swords, where a blade is ground and polished with stones for weeks to bring it in perfect shape, demonstrates to what an extensive piece of work such a process could have been meant in Bronze Age times. It can be assumed that this working process was performed on a very high technological level, since the production of polished stone axes is known and practised thousands of years before the Bronze Age.

Translation: Marion Uckelmann BIBLIOGRAPHY Bernabo, M. B., et al. 1997. La terramare: la piu antica civiltà padana [esposizione Modena, Foro Boario, 15 marzo -1 giugno 1997]. Milano, Electa. Binggeli, M. 2004. Die Tongussformen – Herstellungen und Gebrauch im praktischen Nachvollzug. In I. Bauer, B. Ruckstuhl and J. Speck (eds), Die spätbronzezeitlichen Ufersiedlungen von Zug-Sumpf, Band 3/1: Die Funde der Grabungen 1923–37, 149–153. Zug, Kantonales Museum für Urgeschichte Zug. Bonzon, J. 2004. Die spätbronzezeitlichen Lanzenspitzen-Gussformen von Zug-Sumpf – eine petrographische Studie. In I. Bauer, B. Ruckstuhl and J. Speck (eds), Die spätbronzezeitlichen Ufersiedlungen von Zug-Sumpf.

Chasing The last step in the further processing is the chasing, i.e. the working with punches. The details of the blade profile are accentuated and 20

Spearheads and swords – The making of bronze objects

Band 3/1: Die Funde der Grabungen 1923– 37, 144–149. Zug, Kantonales Museum für Urgeschichte Zug. Fröhlich, M. 1981. Das Gold in der Kunst Westafrikas. Zürich, Museum Rietberg. Fröhlich, M. 1995. Cire-Perdue-Guss in Anlehnung an das Verfahren der Ashanti in Ghana. Zürich, Max Fröhlich. Hedegaard, K. R. 1995. Die Herstellung von wikingerzeitlichen tierkopfförmigen und dosenförmigen Fibeln aus Gotland. Archäologische Mitteilungen aus Nordwestdeutschland. Beiheft 8, Experimentelle Archäologie, Bilanz 1994, 311–345. Hodges, H. W. M. 1954. Studies in the late Bronze Age in Ireland. 1. Stone and Clay Moulds and wooden Models for Bronze Implements. Ulster Journal of Archaeology 17, 62–80. Michaels, A. 1985. Der Cire-perdue-Guss in Nepal, Teil 1. Mitteilungen aus dem Museum für Völkerkunde Hamburg, NF 15 [Festschrift Heller], 77–105.

Michaels, A. 1986. Der Cire-perdue-Guss in Nepal, Teil 2. Mitteilungen aus dem Museum für Völkerkunde Hamburg, NF 16, 173–202. Museum Schwab Biel 2004. Das archäologische Fenster der Region. Biel, Museum Schwab. Seidel, U. 1995. Die Bronzezeit. Stuttgart, Württembergisches Landesmuseum. Speck, J. 1981. Frühes Eisen in den Ufersiedlungen der Spätbronzezeit. Schriften des Kantonalen Museums für Urgeschichte in Zug 27, 265–271. Viollier, D. 1924. Moules à pièces de l‘âge du bronze. Jahresbericht des schweizerischen Landesmuseums in Zürich 33, 51–55. Weidmann, T. 1982. Keramische Gussformen aus der spätbronzezeitlichen Seerandsiedlung Zug “Sumpf“. Jahrbuch der schweizerischen Gesellschaft für Ur- und Frühgeschichte 65, 69–81. Wyss, R. 1971. Technik, Wirtschaft und Handel. In Ur- und frühgeschichtliche Archäologie der Schweiz, Band 3, Die Bronzezeit, 123– 130. Basel, Archäologie Schweiz. Markus Bingelli Gaselstrasse 30 3098 Köniz Switzerland

21

22

USE-WEAR ON BRONZE AGE HALBERDS: THE CASE OF IBERIA Dirk Brandherm ABSTRACT

2004, 280–291; cf. also Schwenzer 2002, 82), and in a recent review of Bronze Age chronology in northwestern Europe J. N. Lanting and J. van der Plicht indeed advocate an earliest date for Irish and British halberds not before the 21st century BC.3 Thus, the chronology presented here should be read as indicating the earliest dates for Early Bronze Age (EBA) halberd types plausible on present evidence, but no matter if this or a somewhat lower chronology is preferred, the relative sequence in which halberds seem to have made their first appearance in different parts of Europe would remain unchanged. An Italian origin with a dispersal into western Europe via Iberia and then from the Atlantic fringe into north-central Europe continues to be the most likely model.4 It should be particularly noted that what has been said above regarding the possibility of a somewhat later beginning of Early Bronze Age halberd types does not apply to the chalcolithic halberds of Italy and Iberia. While it would seem difficult to push the date for the first appearance of halberds in most other regions further back by any significant time span, the Italian types Gambara and Villafranca-Tivoli, as well as the Iberian Alcalar type halberds might actually date considerably earlier than indicated in our figure 1. Should this be the case, one might be left with a significant chronological gap between these Chalcolithic halberd types on the one hand and the much more widely distributed EBA halberds on the other. In the present contribution, however, it is not problems of chronology and distribution, but rather the question of function which is to receive particular attention. Like swords, and unlike other weapons used in Bronze Age armed conflict, such as daggers, axes, spears or bow and arrow, whose designs only differ gradually from

Different types of use-wear detected on Bronze Age halberds from the Iberian Peninsula are described and a tentative interpretation is given of the use-wear patterns observed. The latter clearly demonstrate the practical use of Iberian Bronze Age halberds as combat weapons, and some hypotheses are formulated regarding the specific way in which these weapons were employed. Evidence from outside the Iberian Peninsula would seem to support our interpretation. Without sufficient experimental data to test the hypotheses put forward, however, any interpretation of the observed use-wear patterns must remain provisional. KEYWORDS Bronze Age – Iberia – El Argar – halberd – use-wear – edge damage INTRODUCTION Next to swords, halberds are probably the most conspicuous category of Bronze Age weapons. A full survey of these pieces undertaken by the present author on a previous occasion was mainly aimed at solving questions regarding the chronology and distribution of these artefacts at a European level (Brandherm 2004). The outcome of that exercise with regard to chronology is presented in figure 1.1 Some authors have suggested even earlier dates for the first appearance of halberds in the British Isles and in north-central Europe.2 However, there is a number of problems that come with a higher chronology for halberds in those parts (Brandherm 1

though. Even if one shares their doubts about the character of the alleged Faversham assemblage as a closed find, several similar associations of tanged flat daggers and ribbed EBA halberd blades are known from Iberia (cf. Brandherm 2003, pl. 185 B. C; 186 A. B), so that their synchronous co-existence also in other parts of Europe should not a priori be ruled out. 4 Brandherm 2004, 323–326; cf. also Burgess 1978, 213. Similar conclusions are reached by K. Rassmann (forthcoming), which is particularly encouraging, given that in his study he followed an entirely different methodological approach, using correspondence analysis to define typological relationships.

The entry seen in an earlier version (Brandherm 2004, fig. 24) of the chronological overview presented here in Figure 1 that had Type Gambara as the earliest halberd type known from Ireland has been removed from the present figure, as the respective piece (Harbison 1969, no. 312) has to be considered a modern import (pers. comm. R. O’Flaherty). 2 Needham 1996, 130 fig. 2; Müller 1999, 71 fig. 19 b; Schuhmacher 2002, 274–276; O’Flaherty 2003, 371–376. 3 Lanting and van der Plicht 2001/02, 141 f. Their argument that in the British Isles halberd blades with midribs are unlikely to predate the first appearance of ribbed daggers (Type Butterwick) is not necessarily a compelling one,

23

Dirk Brandherm

Figure 1: Chronology of halberd types across Europe

that of similarly-shaped implements or hunting weapons, the original design of halberds as weapons of war, intended from the beginning for use in interpersonal combat, remains largely unquestioned. The question if most of the halberds that survive from the Bronze Age in practice actually functioned as combat weapons or if they must be regarded largely as ritual objects is quite a different matter though, and both views have had their fervent defenders (Lenerz-de Wilde 1991; Falkenstein 2006/07, 37–39; Consuegra 2007, 12). Very little of this debate, however, has been based on actual evidence. So far, we are still largely lacking systematic surveys of use-wear traces on Bronze Age halberds, with

the notable exception of recent work carried out on the Irish material.5 Also, while ongoing experimental work headed by R. O’Flaherty (see this volume) has clearly demonstrated the potential effectiveness of halberds as combat weapons,6 earlier studies concerned with usewear on bronze axes would seem to indicate that much more extensive series of experiments with replica pieces will be needed to provide us with sufficient evidence for reliable interpretations of use-wear on Bronze Age halberds (cf. Kienlin and Ottaway 1998; Roberts and Ottaway 2003). Ultimately, only the systematic study of usewear on archaeological pieces on the one hand and more comprehensive experimental work

5 O’Flaherty et al. 2008, 22 f. The percentage of pieces in which edge damage was observed among the Irish sample (40 %) considerably exceeds that recorded by the present author for Iberian halberds (10 %). Part of this discrepancy is probably owed to the higher proportion of grave finds among the Iberian material, as these are generally much more affected by corrosion damage likely to obliterate traces of use-wear. Also, when we carried out our survey of Iberian halberds, its primary purpose was to establish a typological classification for these pieces. Specific attention was given to use-wear only with growing experience acquired over the course of our survey. Thus, especially during the earlier part of our touring the respective museum collections, we

may easily have missed less obvious edge damage, so that light bowing or denting in particular is most likely underrepresented in the sample introduced here. This makes the total lack of injuries caused by halberds among the skeletal material studied by G. Aranda-Jiménez et al. (2009, 1045– 1047) seem even more surprising. Clearly there are issues of sample size and representativeness involved here that need to be addressed in future work. 6 O’Flaherty 2007; O’Flaherty et al. 2008; cf. also O’Flaherty et al. this volume. Some useful comparative data are also provided by a series of experiments carried out with different replicas of an Middle Bronze Age sword from La Perla (Carrión et al. 2002).

24

Use-wear on Bronze Age halberds: The case of Iberia

Figure 2: Overview of El Argar halberd types: Type San Antón (1), Form Arrayanes (2), Form Monteagudo (3), Form Barranquete (4) and Form Las Laderas (5) (after Brandherm 2003, pl. 50, 807; 100, 1397. 1399; 102, 1416; 105, 1434) Scale 1:3

following the general classification established by C. Gutiérrez (Gutiérrez 2002, 262–265; cf. also Gutiérrez and Soriano 2008, 440–442). The most common trace of use encountered in Iberian halberds is hilt marks, preserving the outline of the original wooden shaft where it came into direct contact with the blade, as a consequence of differential corrosion processes affecting those parts of the blade that were covered by the shaft and those that were left exposed. Occasionally two or even more different hilt marks can be observed in one piece, providing valuable clues to the re-hafting of individual pieces. That it was not uncommon for halberd blades to have their wooden shaft replaced would seem to indicate a practical use for them rather than a role as mere prestige items, although it is difficult to draw any more detailed conclusions based on this observation alone.

on the other will allow us to arrive at definitive conclusions. The present contribution focusses on the first of these two aspects. It is based on a survey of more than a hundred halberds with provenances in the Iberian Peninsula, about half of which come from the area of the El Argar culture in southeast Spain (fig. 2), the other half mostly from central and western Iberia (fig. 3). Use-wear observed in Iberian EBA halberds

Given the lack of a sufficiently comprehensive series of experiments that would allow us to relate use-wear observed in Bronze Age halberds to experimentally produced patterns of use, any interpretation proposed here for use-wear in archaeological pieces can only be provisional. In our description of specific use-wear traces we are 25

Dirk Brandherm

Figure 3: Overview of halberd types from the Iberian Peninsula (not including types specific to the El Argar culture): Type Carrapatas (1), Type Jaén (2), Type Baútas (3), Type Finca de la Paloma (4), Type Villamiel (5), Type Cano (6), Type Vale de Carvalhal (7) and Type Montejícar (8) (after Brandherm 2003, pl. 46, 772; 49, 803; 52, 832; 92, 1361; 96, 1378; 97, 1380; 98, 1383. 1387). Scale 1:3 26

Use-wear on Bronze Age halberds: The case of Iberia

The presence of specific hilt marks does, however, allow us to identify a number of pieces as halberds in the first place which otherwise would be hard to distinguish from ‘regular’ daggers.7 For this purpose we assume that hilt marks cutting across the blade diagonally identify a blade as a halberd, while omega-shaped or triangular recesses in the base line of the hilt are likely to identify that blade as a dagger rather than a halberd. This is generally confirmed by the orientation of the grain in any wood surviving from shafts (fig. 5. 6) and also by the length of the rivets in the respective pieces, as the insertion of a blade into a halberd shaft tends to require longer rivets than those normally used in hafting dagger blades. On a number of occasions, diagonal and recessed hilt marks can be observed in the same piece (e.g. Brandherm 2003, no. 783. 792). This is interpreted as evidence for these blades being used as halberds during one stage of their lives and as daggers during another (fig. 2, 2). Such a change has significant implications for use-wear analysis carried out on these pieces, as the re-hafting of a halberd blade as a dagger or vice versa obviously would lead to quite different use-wear patterns from those that could be expected to occur in blades which throughout their life were hafted exclusively as halberds. While such a change in function is quite common among Iberian halberds in general, for the pieces with pronouncedly winged necks that are typical of the El Argar culture, as well as for particularly broad and heavy ‘Atlantic’ halberds it is quite unlikely that they could have ever been usefully attached to a dagger hilt, and this is borne out by the evidence from hilt marks and rivets.8 The further analysis of use-wear then concentrates particularly on those pieces that show no sign of ever having been hafted as anything but halberds. Use-wear traces providing more specific evidence on the practical use of halberds are much less common than hilt marks. Of the various types of use-wear categorised by Gutiérrez, the following can be identified in Iberian halberds: notches, gaps, bluntness, and asymmetry, to be discussed in more detail in the following paragraphs.

Notches Where notches can be observed in the edge of a blade, they frequently give the impression of having been caused by blows against a hard, narrow edge, possibly of another blade. More pronounced notches tend to concentrate on the inner edge of the blade, which in cases where a diagonal or asymmetrically curved hilt mark survives is easily identifiable from the angle at which the hilt mark crosses the blade (fig. 4. 5. 7. 9. 10. 12).9 Gaps In contrast to notches, where the edge of a blade has been dented and bent out of shape, but no metal is lost, gaps occur where metal has been broken or torn from the edge. Gap damage is quite frequent, particularly in halberds from El Argar burials, where it would appear to be mostly post-depositional. Some cases of gap damage, however, can likely be identified as actual Bronze Age use-wear. Even then, without sufficient experimental data to draw upon, its interpretation is not always straightforward. Bluntness Pronounced blunting of edges is rarely observed in pieces among our sample, although a considerable number of pieces has suffered from corrosion, which makes it difficult to recognise any blunting of the edges. Where bluntness can be observed, it is invariably found on the point rather than on the lateral edges of the blade. A number of factors might contribute to this picture. First of all, by design a halberd is not primarily a cutting implement, thus the lateral edges may be much less likely to be exposed to blunting damage than the point. Also, the point of most halberd blades is notably thicker than the lateral edges, as it is here where the mid-rib meets the edge. Re-sharpening might thus be notably more effective on the lateral edges than on the point. On the other hand, blows to soft tissue are unlikely to cause blunting of any noticeable degree in a

7

just because of their geographical distribution. From a typological point of view, however, their affinities in many cases are no closer to pieces from western France or the British Isles than to central European halberd types. 9 This category also includes use-wear traces which O’Flaherty et al. (2008, 23) would classify as ‘denting’.

The use of the term ‘dagger’ here follows conventional archaeological usage, to denote double-edged blades which are too short to qualify as swords, whereas the term ‘knife’ remains restricted to single-edged blades exclusively. 8 In specialist literature, non-El Argar halberds from the Iberian Peninsula are frequently referred to as ‘Atlantic’

27

Dirk Brandherm

4 5 Figure 4: El Argar (tomb 999): poorly visible hilt mark (A); immediately below its line there is some gap damage on both edges (B); among the numerous smaller gaps and notches visible on the edges, only one particular notch on the inner edge matches the angle and characteristics expected from pre-depositional combat damage (C), while the rest of the damage on the edges would appear to result mainly from corrosion (L. 17.4 cm) Figure 5: Unprovenanced (Gómez Moreno collection): clearly visible, asymmetrically curved hilt mark (A); line of small gaps in the upper part of the inner edge (B); very small notches (C) on the lower half of the outer edge, caused by the impact of a very fine edge, seemingly not post-depositional (L. 16.0 cm)

halberd. Thus, only blows hitting bones – or other sufficiently hard material – would be capable of causing blunting damage to the blade.10

identify halberds in the first place, in the Iberian sample strongly asymmetrical halberd designs are much less common. Design asymmetries in Iberian halberds are generally limited to the positioning of the rivet holes and slight asymmetries of the neck. Blades with slightly asymmetrical outline of the edges also exist, but while they certainly permit conclusions regarding the function of a piece, they cannot be considered usewear sensu stricto. The main reason for dwelling on asymmetrical designs here is that with some of the pieces from our sample it is not entirely clear whether the asymmetry in the outline of the blade is an original feature or whether it was caused by differential re-sharpening.

Asymmetry Here one needs to distinguish between the original design of the blade on the one hand and secondary asymmetries resulting from differential re-sharpening on the other. While strongly asymmetrical designs occur in some Italian halberd types and while an asymmetrically designed blade is one of the main criteria to 10

While generally following the classification of usewear traces according to Gutiérrez (2002, 262–265), in the

present category we would also provisionally include the type of use-wear that O’Flaherty et al. (2008, 23) define as ‘bowing’.

28

Use-wear on Bronze Age halberds: The case of Iberia

Discussion of individual pieces

small notches that could only have been caused by the impact of a very fine edge and which do not seem to be post-depositional in nature. Hence, we would provisionally identify these as use-wear traces as well (fig. 5; Brandherm 2003, no. 812).

In this section those pieces that present clear evidence of one or several of the categories of use-wear described above are discussed in more detail. The following are not the only, but the most conspicuous examples of use-wear among the Iberian material. In other instances where likely traces of use-wear can be identified, these are more difficult to discern, mostly as a consequence of post-depositional corrosion or damage from modern cleaning attempts. The objective of our taking a close look at these pieces on the one hand is to instigate more systematic study of use-wear in halberds from other parts of Europe, on the other hand we hope to build some working hypotheses regarding the interpretation of specific use-wear patterns that can be tested in future experimental studies. Such experimental work hopefully would not only provide us with data to test any hypotheses, but through a better understanding of use-wear on Bronze Age halberds would also allow the identification of use-wear in pieces where it is less apparent, thus expanding the sample that can be meaningfully interpreted.

3. El Argar (tomb 449), Antas, Almería. This Form Arrayanes halberd has suffered considerable post-excavation damage. Its most important use-wear-induced features, however, appear largely unaltered when comparing its current state with that recorded in the original 19th century watercolours by L. Siret. Most of the remains of the wooden shaft then still preserved have since been removed, but the hilt mark remains clearly visible. As the hilt mark runs perpendicular to the longitudinal axis of the blade, the latter’s inner and outer edge can only be distinguished based on its slightly asymmetrical shape which most likely has to be interpreted as a result of differential re-sharpening. The inner edge immediately below the hafting point features a pronounced gap which we would interpret as use-wear. The same interpretation may apply to a number of notches further down the inner edge. Some of the gap damage on the outer edge clearly seems post-depositional, given its cutting at sharp angles through the corrosion layer, but some of the smaller gaps and notches that are found on the outer edge close to the shaft are quite likely to be pre-depositional (fig. 6; Brandherm 2003, no. 1393).

1. El Argar (tomb 999), Antas, Almería. This San Antón type halberd preserves a slightly diagonal hilt mark. On both edges of the blade, immediately below the hafting point, gaps can be identified that at least in part seem the result of pre-depositional use, although some loss of substance here may also be due to later corrosion processes. A series of smaller gaps and notches particularly along the outer edge would appear to result mainly from corrosion, and the loss of metal at both the tip and the neck is almost certainly post-depositional. Only one particular notch on the inner edge by its angle and general characteristics would appear to qualify as potential combat damage. It is not readily discernible whether a slight asymmetry in the outline of the blade should be interpreted as a result of differential sharpening (fig. 4; Brandherm 2003, no. 804).

4. Puerto de Gumial, Aller, Asturias. Its rounded neck sets this piece apart from most other halberds of Form Barranquete, to which it is attributed here nonetheless. Modern cleaning has left the surface of this piece covered in scratches and has almost obliterated the slightly diagonal hilt mark. While the identification of other potential use-wear is much hampered by the effects of this cleaning attempt, a shallow gap on the inner edge of the blade, immediately below the hafting point, would appear to be predepositional in nature. In the light of evidence from other pieces we would interpret this gap as resulting from differential re-sharpening to eliminate notches on this stretch of the edge (fig. 7; Brandherm 2003, no.1420).

2. Unprovenanced, from the Gómez-Moreno collection. Some wood from the shaft is still attached to the neck of this San Antón type halberd which displays an asymmetrically curved hilt mark. The upper stretch of the inner edge is lined with a number of small gaps that do not appear to have been caused by corrosion and which we would interpret as use-wear. The lower half of the outer edge shows some very

5. San Antón (?), Orihuela, Alicante. Despite some idiosyncratic features, in typological terms this piece also is included with our Form Barranquete. Corrosion damage to the blade is quite extensive, rendering the identification of 29

Dirk Brandherm

6

7

Figure 6: El Argar (tomb 449): clearly visible hilt mark (A); inner and outer edge distinguished by slightly asymmetrical shape, probably as a result of differential re-sharpening; possibly pre-depositional gap on the inner edge immediately below the hilt mark (B). Interpretation of further notches on both edges (C) as use-wear remains doubtful (L. 14.1 cm) Figure 7: Puerto de Gumial: barely visible, slightly diagonal hilt mark (A); shallow gap on the inner edge of the blade, immediately below the line of the hilt mark, possibly from differential re-sharpening (B). Use-wear traces on this piece are obscured by a modern cleaning attempt which has left the surface of this piece covered in scratches (L. 18.5 cm)

use-wear along its edges virtually impossible. However, the tip of the blade has suffered severe blunting damage, and the entire blade appears warped from heavy impact, upsetting the metal along the longitudinal axis of the piece. While we cannot entirely rule out post-depositional damage here – most of the original patina has flaked off –, it would appear much more likely that this was inflicted by dealing one or several heavy blows against a hard surface while a wooden shaft was still in place. Inflicting this kind of damage after the wood had decayed or the shaft had been removed would have required ham-

mering blows to the neck of the blade. No signs of any impacts can be discerned there, however (Brandherm 2003, no. 1421). 6. Monte do Castro, Carrapatas, Trás-osMontes. This Carrapatas type halberd, like many other pieces from wet contexts has survived in pristine condition, preserving traces of the original surface treatment, such as hammer marks and fine scratches from grinding and polishing. No hilt mark can be discerned, but the slightly asymmetrical shape of the neck allows to clearly distinguish between inner and outer edge. While 30

Use-wear on Bronze Age halberds: The case of Iberia

8

9

Figure 8: Monte do Castro: no hilt mark, but the slightly asymmetrical shape of the neck allows to distinguish between inner (A) and outer (B) edge; a very slight concave curvature of the inner edge may be interpreted as indicative of differential re-sharpening, and a series of notches and small nicks with some degree of bowing (C) is clearly visible in the upper part of the inner edge (L. 34.8 cm) Figure 9: Baútas: the slightly diagonal hilt mark (A) does not correspond to the asymmetrical position of the rivet holes, but the more intensive damage to the inner edge (B) as well as its somewhat more pronounced concave curvature, most likely resulting from differential re-sharpening, are in line with the identification of inner and outer edge as indicated by the hilt mark. Gap damage on the inner edge (C), immediately above the line of the hilt mark could be related to an earlier state of hafting that might have better corresponded to the position of the rivet holes, but might also be peri-depositional, as are likely some gap damage on the outer edge (D) and the two torn (E, F) rivet holes (L. 19.7 cm)

pret as indicative of differential re-sharpening (fig. 8; Brandherm 2003, no. 1364).

on the outer edge hardly any traces of use-wear are visible, the inner edge presents a number of notches, particularly in its upper part, immediately below the likely hafting point. A certain asymmetry can also be observed between the two edges, with the inner edge presenting a very slight concave curvature which we would inter-

7. Baútas, Mina, Estremadura. A slightly diagonal hilt mark can be discerned on this eponymous type specimen of Type Baútas. The tearing damage affecting the central rivet hole seems 31

Dirk Brandherm

as combat damage, one would have to assume that the inflicting blow either was sufficiently heavy to cut up to 1 cm deep into the wood of the shaft, or else that it was inflicted while a presumably earlier wooden shaft was attached in a slightly different position, of which no discernible hilt mark survives, but whose original presence might be indicated by the position of the rivet holes, witch does not match the angle of the hilt mark. Alternatively, this might be peridepositional damage, inflicted when the shaft was removed. This certainly is the most likely explanation for the torn rivet holes and perhaps also for the large gap in the outer edge (fig. 3, 3; 9; Brandherm 2003, no. 1380). 8. Cortijo de Jaduglón, Arcos de la Frontera, Cádiz. The diagonal hilt mark and the slightly asymmetrical shape of the neck allow to clearly distinguish between the outer and inner edge of the blade in this Type Baútas halberd. The inner edge bears a series of deep notches in the area closest to the hafting point, while the smaller notches that can be observed on its outer edge are much less pronounced. As in some of the previously described cases, this pattern might be indicative of a specific use of this weapon in combat, although we would not entirely rule out the possibility that what we are dealing with here is marks left by forcibly loosening the wooden shaft prior to deposition. The bent tip of the blade, in any case, is a feature shared by several halberds from non-funerary contexts in the middle and lower Guadalquivir valley; this definitely is not pre-depositional use-wear, but testimony to a specific peri-depositional treatment these pieces received. Some deep scratches running across the blade, including the area that would have been covered by the wooden shaft, appear to be post-depositional in nature (fig. 10; Brandherm 2003, no. 1380 A).

Figure 10: Cortijo de Jaduglón: clearly visible hilt mark (A) and series of notches on the inner edge of the blade, immediately below the line of the hilt mark (B); the notches that can be observed on the outer edge of the blade are considerably less pronounced (C). The folded tip (D) is a feature of several halberds from non-funerary context in the Guadalquivir valley; it clearly is testimony to a specific treatment these pieces received when deposited in the ground and is unconnected to their practical role as combat weapons (L. 24.4 cm)

peri-depositional, given that a secure hafting would have been impossible with the neck of the halberd bent out of shape, and the same may be true for a large gap in the upper section of the outer edge. Otherwise, the outer edge is in much better condition than its inner counterpart which displays a slightly more concave curvature, probably as a result from differential re-sharpening, and a series of notches, most of which are concentrated in its upper reaches. A gap on the inner edge, however, is situated directly above the line of the hilt mark. If this were to be interpreted

9. Manzanares Valley, Central Spain (exact provenance unknown). A diagonal hilt mark is clearly visible on this Type Villamiel halberd. Apart from a pronounced scratch running across one half of the blade and a couple of dents or gaps in its outer edge, the most clearly discernible use-wear on this piece is some blunting damage visible at the tip. Without autopsy, however, it is not possible to decide if any or all of these traces are pre- or post-depositional (Fig. 3, 5).11

11 Blasco et al. 2001, fig. 3, 2; Brandherm 2003, no. 1387. At the time the author conducted his museum survey of Iberian halberds, it was not possible to gain

access to this particular piece, which is kept in the teaching collection of the Archaeology department at Oviedo University.

32

Use-wear on Bronze Age halberds: The case of Iberia

11

12

Figure 11: Beluso: if the slightly asymmetrical position of the rivet holes is accepted as indicative of the distinction between the inner and the outer edge, this piece is exceptional in that most of the notches observed then would be located in the lower part of the outer edge (A); clearly recognizable blunting damage (B) to the tip (L. 25.3 cm) Figure 12: Sierra de Alta Coloma: slightly diagonal back hilt mark (A) and asymmetrically curved front hilt mark (B) both clearly discernible, as are some small notches on the outer (C) and some deeper notches on the inner edge (D), some of the latter possibly smoothed over during re-sharpening (L. 30.0 cm)

10. Beluso, Boiro, La Coruña. As far as the typological classification of Iberian halberds is concerned, this is a rather idiosyncratic piece which could be described as loosely attached to the Villamiel type. No hilt mark survives, but the somewhat asymmetrical position of the rivet holes provides a clue for distinguishing between

the inner and outer edge of the blade. Assuming that our distinction of inner and outer edge is correct, this piece would be exceptional in that most of the notches that can be observed are located in the lower part of the outer edge. This is also one of the comparatively rare pieces that suffered clearly recognizable blunting damage to 33

Dirk Brandherm

There are two main questions raised by the usewear detected in our sample. Does the occurrence of notches, gaps, bluntness and asymmetries in these pieces reveal any significant patterns, and if so, what possible interpretations are there for any such patterns? In trying to answer these questions, we shall concentrate on pre-depositional use-wear. Likely peri-depositional damage (e.g. folded blades, torn-out rivets) does occur in a substantial percentage of halberds from the central and western parts of the Iberian Peninsula, providing evidence for specific – presumably ritual – treatment at the time of deposition, but at

present such damage is largely lacking in El Argar halberds. The latter, on the other hand, tend to be much more severely affected by post-depositional corrosion damage, due no doubt to them being deposited as part of funerary assemblages in cists, rock-cut tombs or pithoi together with decomposing organic matter, while outside the area of El Argar burial customs halberds upon deposition were much more likely to be immediately embedded in rather homogeneous sediments or even in waterlogged conditions. This in itself of course constitutes a clear pattern, but in the present contribution our main attention shall go not to peri-depositional treatment or the effects of post-depositional conditions on preservation, but rather to pre-depositional usewear and the information it conveys regarding the practical function of Bronze Age halberds. Aspects of peri- or post-depositional damage are only taken into account as far as they have potential effects on the reading of earlier use-wear. A general assessment of the broader implications from use-wear, find circumstances and preservation of Iberian Bronze Age halberds has already been undertaken elsewhere (Brandherm 2003, 387–399). Overall, with regard to pre-depositional usewear a surprisingly clear picture emerges from our survey. Based on the distribution of notches, gaps and asymmetries on the halberds under scrutiny here, the inner edge of the blade clearly would appear to have seen more active use than the outer edge. According to the specific character of the notches and gaps involved, the inner edge regularly would have suffered not only more, but also heavier impacts from a hard and narrow edge, most likely that of an opponent’s weapon. The angle and general characteristics of the notches that can be identified on the outer edge of some specific pieces are distinctly different from those on the inner edge (e.g. unprovenanced piece from the Gómez Moreno collection: no. 5; halberd from the Sierra de Alta Coloma: no. 11). We would tentatively interpret these very small notches that usually cut into the blade more or less at a right angle to its longitudinal axis as evidence for parrying, possibly of sword blows, but the sample under scrutiny here is certainly too small to draw any definitive conclusions.12 Chronology at least does not seem to be an obstacle here. While Type San Antón

12 Notches possibly from such blows on a sword blade among the Iberian material so far have only been identified on the piece from La Perla (Gómez 2001, 6). While experimental work carried out with a replica of the La Perla

sword has given rise to doubts regarding the interpretation of the respective notches as actual combat damage, it must not be forgotten that the original piece contained up to 50 % more arsenic than the respective replica, and that no

its tip. The interpretation of this damage poses a number of problems; as this piece appears to have been dredged from the bottom of the Ría de Arousa, some or all of this damage might well be post-depositional. The very uneven and pockmarked saltwater patina does not provide any criteria to resolve this particular imponderable (fig. 11; Brandherm 2003, no. 1390). 11. Sierra de Alta Coloma, Montejícar, Granada. As usual with Type Montejícar halberds, both the asymmetrically curved front hilt mark and the slightly diagonal back hilt mark can be discerned in this piece. One particularly deep notch and a number of somewhat shallower ones are found on the inner edge of the blade. While the former could only have been caused by the impact of a very hard and narrow edge, the character of the latter is less unequivocal. Some of these may have been smoothed over in a re-sharpening attempt. All of the notches on the inner edge of the blade, however, apparently were caused by blows cutting diagonally into the blade, against the direction of any blow that would have been dealt with the piece under scrutiny here. On the outer edge, a series of much smaller notches can be observed that cut into the edge in a direction which does not run diagonally, but perpendicular to the longitudinal axis of this blade. There is a clear pattern in this differential distribution of notches, which given the fine preservation of this piece in general and the condition of its patina most likely have to be interpreted as pre-depositional use-wear (fig. 3, 8; 12; Brandherm 2003, no. 831). CONCLUSIONS

34

Use-wear on Bronze Age halberds: The case of Iberia

halberds generally predate the introduction of swords to the Iberian Peninsula at the beginning of the Middle Bronze Age, in at least one case a piece of this type was associated with an early rapier, and the same holds true for the Sierra de Alta Coloma halberd whose Middle Bronze Age date has long been recognised (Schubart 1973, 255–263; Brandherm 2003, 247 f. 255). The majority of notches, and among them particularly the more pronounced ones, tend to cluster in the upper part of the blade, close to the hafting point, and again the heavier concentrations are found on the inner edge of the blade. For gaps it is sometimes difficult to tell predepositional wear apart from later damage. Generally, however, likely pre-depositional gap damage shows a distribution pattern not unlike that of notches: mostly on the inner edge and mostly in its upper reaches. In stark contrast to this, pronounced bluntness is comparatively rare, and where it occurs it remains largely limited to the tip of the blade. Finally, while an asymmetrical outline in halberds is not at all uncommon, and in most cases would appear part of their original design, in a number of instances slight asymmetries can be perceived in the outline of the edges that seem to result from differential re-sharpening. It is usually the inner edge whose outline is somewhat more concave, indicating a loss of metal as a consequence of more intensive or more frequent sharpening. Overall, the concentration of damage from notches and gaps, as well as some of the asymmetries observed in the upper part of the inner edge constitutes a clear pattern that is provisionally interpreted here as indicative of a particular combat style, in which an opponent’s blade was frequently caught in the angle between the shaft and the blade of one’s own weapon. An alternative explanation for the concentration of more pronounced notches in this particular area could be that particularly with more steeply inclined blades the angle between the shaft and the inner edge of the blade was more difficult to access when it came to re-sharpening, so notches that in other sections of the edge would have been obliterated might survive here. However, this is not supported by any direct evidence for differential re-sharpening along the inner edge of the respective pieces.

The same holds true for a further possible interpretation of notches and gap damage to the edges of the blade close to the hafting point. While it is certainly feasible that halberd blades which had their wooden shafts removed prior to being deposited in the ground would have suffered damage to their edges close to the shaft from levering action to loosen the latter, the same kind of damage is found in some pieces that were deposited with their shaft intact (fig. 4–6). This would seem to corroborate the hypothesis formulated above, regarding a particular combat style which involved cutting blows with the inner edge and catching the opponent’s blade in the angle between blade and shaft. Such a use is consistent with documentary evidence from ancient China, where bronze halberds continued to be employed well into the early historical period. No use-wear analysis has been conducted on Chinese halberds, however, and without a proper series of experiments no adequate testing of the above hypothesis will be possible. In any reconstruction of the handling of Bronze Age halberds in combat, other considerations also have to be taken into account. A piece like the Carrapatas type halberd from Monte do Castro, with a length of 34.8 cm and a weight of 605 g may at first glance simply appear too heavy and unwieldy for practical use, which is probably one of the reasons that have contributed to the widespread belief that Bronze Age halberds were essentially ritual items, used for display rather than actual combat. This, however, is not consistent with the use-wear observed also on such large pieces, which hence we would interpret as double-handed weapons, employed perhaps in a manner not entirely unlike mediaeval Danish axes – unfortunately Bronze Age rock-art depicting halberds is not usually quite as explicit as the Bayeux tapestry. There are figurative images showing double-handed use of halberds among the rock art from Mont Bego, but those are clearly no combat scences (Lenerz-de Wilde 1991, 41 fig. 18). Again, the documentary evidence from ancient China that mentions halberds with extra-long shafts for boat and chariot combat as well as the use of halberds for decapitating prisoners may prove helpful here (cf. So 1980, 316 fig. 107; Kolb 1991, 77–79).

determination of the hardness of edges was undertaken on either the original or the replica (cf. Carrión et al. 2002, 293 fig. 6). In our view, the most likely explanation for the notches on the blade of the La Perla sword is still

combat damage from a parrying blow by or to another sword, although in the light of the evidence presented here, we would not want to rule out that they could have been inflicted by striking a halberd blade.

35

Dirk Brandherm

However, for a number of very long and slender halberd blades a practical function as a combat weapon can indeed most likely be ruled out. Just a handful of such pieces is known from different parts of Europe, and from among the Iberian material a halberd with considerable, probably peri-depositional damage and a likely provenance from Valdepeñas (Ciudad Real) is the only piece that can be included in this group.13 The same most likely holds true for halberds with metal shafts from north-central Europe, as well as obviously for the exceptional halberd blades from precious metal so far found exclusively in the Carpathian Basin.14 A practical function as a combat weapon more than likely can not only be ruled out for a number of excessively long and preciously crafted halberds, but also for some very small and inconspicuous pieces. An example for this is a piece recovered from tomb no. 10 at the Cerro de Culantrillo (Gorafe, Granada; García 1963, 76. 80 pl. 4, 1). Although no experimental work has been conducted to test its potential usefulness in a combat, it must be strongly doubted that with a length of just 9.8 cm and its flat section this blade could have served any practical purpose as a weapon, and no traces of use-wear are visible on the piece. Still, the length of the rivets and the remains of the wooden shaft clearly identify this find as a halberd. A symbolic function has already been suggested elsewhere, and it may be no coincidence that this find was recovered from the burial of a juvenile, possibly even female individual (Brandherm 2003, 69. 205 no. 546). Finally, when hypothesising about the effectiveness of halberds as combat weapons, one must nor forget that the wooden shaft of such a weapon is not necessarily less effective than the bronze blade. A well aimed blow or thrust with its butt to many parts of an unprotected human body will end any combat just as quickly as a direct hit with the blade. There are clearly limits here for what experimental archaeology can do with regard to determining the actual effectiveness of halberds in combat, and it will of course never be possible to accurately reconstruct Bronze Age combat styles, but there certainly is still enormous potential in experimental work to further our understanding of use-wear on Bronze Age halberds.

Only empirical data from systematic and adequately recorded experiments will eventually allow us to move on from the current state of affairs, where discussion about the function of these items is informed more often by preconceived ideas about the role of armed conflict in Bronze Age societies rather than by direct evidence obtained from the archaeological record. BIBLIOGRAPHY Aranda-Jiménez, G., Montón-Subías, S. and Jiménez-Brobeil, S. 2009. Conflicting Evidence? Weapons and Skeletons in the Bronze Age of South-east Iberia. Antiquity 83, 2009, 1038–1051. Brandherm D. 2003. Die Dolche und Stabdolche der Steinkupfer- und der älteren Bronzezeit auf der Iberischen Halbinsel. Prähistorische Bronzefunde VI, 12. Stuttgart, Steiner. Brandherm, D. 2004. Porteurs de hallebardes? Überlegungen zur Herkunft, Entwicklung und Funktion der bronzezeitlichen Stabklingen. In H.-J. Beier and R. Einicke (eds), Varia Neolithica III, 279–334. Langenweissbach, Beier and Beran. Burgess, C. 1978. The Background of Early Metalworking in Ireland and Britain. In: M. Ryan (ed.), The Origins of Metallurgy in Atlantic Europe. Proceedings of the fifth Atlantic Colloquium, Dublin 30th March to 4th April 1978, 207–214. Dublin, The Stationary Office. Carrión Santafé, E., Baena Preysler, J. and Blasco Bosqued, C. 2002. Efectismo y efectividad de las espadas argáricas a partir de una réplica experimental del ejemplar de La Perla (Madrid) depositado en el museo arqueológico del Cataluña. In: I. Clemente, R. Risch and J. F. Gibaja (eds), Análisis funcional. Su aplicación al estudio de sociedades prehistóricas. British Archaeological Reports, International series 1073, 285–293. Oxford, BAR Publishing. Consuegra Rodríguez, S. 2007. Alabarda del depósito de Vélez Blanco (Almería) en el Museo Arqueológico Nacional. Revista Velezana 26, 2007, 8–12. Falkenstein, F. 2006/07. Gewalt und Krieg in der Bronzezeit Mitteleuropas. Bericht der bayer-

13

halberd blades found with metal shafts, however, might previously have been attached to regular wooden shafts and may have suffered use-wear during such an earlier part of their life.

Brandherm 2003, no. 1384; cf. Meier-Arendt 1969, fig. 3; pl. 11. 12; Schauer 1971, no. 180. 14 Wüstemann 1995, no. 97–132; Brandherm 2004, 307 fig. 15. Some of the central European

36

Use-wear on Bronze Age halberds: The case of Iberia

pologie – Raumanalysen. Beiträge zum Neolithikum und zur Frühbronzezeit im Mittelelbe-Saale-Gebiet, Teil 1. Bericht der Römisch-Germanischen Kommission 80, 1999, 25–212. Needham, S. 1996. Chronology and Periodisation in the British Bronze Age. Acta Archaeologica (København) 67, 121–140. O’Flaherty, R. 2003. A Consideration of the Early Bronze Age Halberd in Ireland. Unpublished PhD thesis University College Dublin. O’Flaherty, R. 2007. A Weapon of Choice – Experiments with a Replica Irish Early Bronze Age Halberd. Antiquity 81, 423–434. O’Flaherty, R., Bright, B., Gahan, J. and Gilchrist, M. D. 2008. Up Close and Personal. Archaeology Ireland (4), 22–25. Rassmann, K. 2010. Die frühbronzezeitlichen Stabdolche Ostmitteleuropas. Anmerkungen zu Chronologie, Typologie, Technik und Archäometallurgie. In H. Meller and F. Bertemes (eds), Der Griff nach den Sternen. Wie Europas Eliten zu Macht und Reichtum kamen. Internationales Symposium in Halle (Saale) 16.–21. Februar 2005. Tagungen des Landesmuseums für Vorgeschichte Halle Bd. 5, 807–822. Halle (Saale). Roberts, B. and Ottaway, B., 2003. The Use and Significance of Socketed Axes During the Late Bronze Age. European Journal of Archaeology 6 (2), 119–140. Schauer, P. 1971. Die Schwerter in Süddeutschland, Österreich und der Schweiz I (Griffplatten-, Griffangel- und Griffzungenschwerter). Prähistorische Bronzefunde IV, 2. München, C. H. Beck. Schubart, H. 1973. Las alabardas tipo Montejícar. In: J. Maluquer de Motes Nicolau (ed.), Estudios dedicados al Prof. Dr. Luis Pericot, 247–269. Barcelona, Universidad de Barcelona. Schuhmacher, T. X. 2002. Some Remarks on the Origin and Chronology of Halberds in Europe. Oxford Journal of Archaeology 21 (3), 263–288. So, J. F. 1980. The Inlaid Bronzes of the Warring States Period. In: W. Fong (ed.), The Great Bronze Age of China, 305–320. New York, Knopf. Soriano Llopis, I. and Gutiérrez Sáez, C. 2007. Use-wear Analysis on Metal. The Influence of Raw Material and Metallurgical Production Processes. In: Proceedings of the 2nd International Conference ‘Archaeometallurgy in Europe’, Aquileia, 17–21 June, 2007. Milano, AIM.

ischen Bodendenkmalpflege 47/48, 2006/07, 33–52. García Sánchez, M. 1963. El poblado argárico del Cerro del Culantrillo, en Gorafe (Granada). Archivo de Prehistoria Levantina 10, 1963, 69–96. Gómez Ramos, P. 2001. La espada de La Perla. Estudio de las empuñaduras de remaches con doble arco: un unicum en la serie de armas europeas de la Edad del Bronce. Gladius 21, 2001, 5–30. Gutiérrez Sáez, C. 2002. Traceología aplicada al material metálico: limites y posibilidades. In: I. Clemente, R. Risch and J. F. Gibaja (eds), Análisis funcional. Su aplicación al estudio de sociedades prehistóricas. British Archaeological Reports, International series 1073, 261–272. Oxford, BAR Publishing. Gutiérrez Sáez, C. and Soriano López, I. 2008. La funcionalidad sobre material metálico. Bases y aplicaciones de estudio. In: S. Rovira Llorens, M. García Heras, M. Gener Moret and I. Montero Ruiz (eds), Actas del VII Congreso Ibérico de Arqueometría, Madrid, 8–10 de octubre 2007, 432–446. Madrid, Quadro. [Internet publication (ISBN 978-84612-8598-3)]. Harbison, P. 1969. The Daggers and the Halberds of the Early Bronze Age in Ireland. Prähistorische Bronzefunde VI, 1. München, C. H. Beck. Kienlin, T. L. and Ottaway, B. 1998. Flanged Axes of the North-alpine Region. An Assessment of the Possibilities of Use Wear Analysis on Metal Artifacts. In: C. Mordant, M. Pernot and V. Rychner (eds), L’Atelier du bronzier du XXe au VIIIe siècle avant notre ère. Actes du colloque international Bronze’96. Neuchâtel et Dijon II. Du mineral au métal, du métal à l’objet, 271–286. Paris, CTHS. Kolb, R. T. 1991. Die Infanterie im Alten China. Materialien zur Allgemeinen und Vergleichenden Archäologie 43. Mainz, von Zabern. Lanting, J. N. and van der Plicht, J. 2001/02. De 14 C-cronologie van de Nederlandse pre- en protohistorie, IV: bronstijd en vroege ijzertijd. Palaeohistoria 43/44, 117–262. Lenerz-de Wilde, M. 1991. Überlegungen zur Funktion der frühbronzezeitlichen Stabdolche. Germania 69, 25–48. Meier-Arendt, W. 1969. Ein frühbronzezeitlicher Stabdolch im Römisch-Germanischen Museum Köln. Germania 47, 53–62. Müller, J. 1999. Radiokarbonchronologie – Keramiktechnologie – Osteologie – Anthro37

Dirk Brandherm

Schwenzer, S. 2002. Zur Frage der Datierung der Melzer Stabdolche. Prähistorische Zeitschrift 77, 76–83.

Wüstemann, H. 1995. Die Dolche und Stabdolche in Ostdeutschland. Prähistorische Bronzefunde VI, 8. Stuttgart, Steiner. Dirk Brandherm School of Geography, Archaeology and Palaeocology (GAP) Queen‘s University Belfast Belfast, BT7 1NN Northern Ireland, Great Britain

38

CEREMONIAL OR DEADLY SERIOUS? NEW INSIGHT INTO THE FUNCTION OF IRISH EARLY BRONZE AGE HALBERDS Ronan O’Flaherty, Michael D. Gilchrist and Trevor Cowie very well-designed from a functional perspective, easily penetrating skull-bone and proving remarkably resilient to damage (O’Flaherty 2007a; 2007b). The supposed ‘weakness’ of hafting of the Irish halberds is simply not borne out by the results of practical trials. The question now is not if they could have been used as a weapon, but rather whether they were so used. The key to answering this lies in the halberds themselves. A careful study of the museum specimens reveals that a great many Irish halberds display clear evidence of impact damage, which includes edge damage, damage to the rivet-holes, and buckling or actual breaks to the blade. In terms of edge damage, which is the focus of the current paper, what one finds is very similar to what occurs on the swords of later periods, i.e. notching, nicks and dents, and these are often taken as indicators of combat-use on the swords (Bridgford 1997; Kristiansen 2002; Molloy 2007). While it might be tempting to simply transfer this interpretation to the halberds, the authors consider that one must first identify the degree of force involved in the creation of these notches and dents, as well as identifying the most likely source, before drawing any firm conclusions. The legal maxim that ‘force goes to intent’ is a good touchstone and will form a key element in our considerations. In a short paper published elsewhere (O’Flaherty et al. 2008) two of the authors described and characterised for the first time the types of edge damage which occur on Irish halberds. They also set out an agenda for a focused programme of impact replication to be undertaken under laboratory conditions at the School of Mechanical Engineering and Materials, University College Dublin (hereafter, simply ‘UCD’). The purpose of this programme was to attempt to replicate the conditions under which the types of edge damage observed on the museum specimens were created, what other artefacts were involved and at what levels of force. In this paper, we set out the results of that study and consider its implications for the function of Irish and other halberds. This work is being carried out as part of a joint halberd-research project involving experts in

ABSTRACT This paper describes a series of highly focused experiments, carried out under laboratory conditions, as part of a wider Irish-Scottish project which is investigating the function of the Early Bronze Age halberd. Earlier trials showed that these artefacts, when hafted properly, were capable of piercing sheep-skulls without suffering significant damage and that by extension they were therefore capable of being used as weapons. Having observed that up to 50 % of museum specimens bore evidence of edge damage, a new set of laboratory trials was undertaken to replicate this damage and to determine how it occurred and under what circumstances. These trials indicate that most of the edge damage on Irish halberds results from impact with other halberds, in a yielding environment (suggesting that both halberds could move) and at tightly-controlled energy-levels which remained capable of lethal effect. It is argued that this is consistent with skilled combatuse and that many of the Irish halberds must have been employed to such end. KEYWORDS Halberds – Bronze Age – function – edge damage – replica – experimental testing INTRODUCTION The halberds of the Early Bronze Age period in Europe have tended to be regarded as largely ceremonial objects. In Ireland, archaeologists have pointed to a perceived weakness in the hafting technique in support of a view of these objects as non-utilitarian, as well as a presumed clumsiness in the hand.1 Similar views have also been expressed in relation to the British and European material up to quite recently (Osgood 2000, 86; Toms 2000, 99). However, that view is beginning to change (Brandherm 2003; 2004; this volume) and practical experiment by one of the authors has shown that the Irish halberds, at least, are 1 Ó Ríordáin 1937, 241; Macalister 1949, 132–134; Herity and Eogan 1977, 137; O’Kelly 1989, 164 f.; Mallory

and McNeill 1991, 102; Waddell 1991, 70; 1998, 129–131.

39

Denting/Bowing Notching Both Total Impacts None

high medium low (burnt)

Total None Possible Notching Grand Total

0 1 0 1 3 2 6 1 12 0 13

4 3 4 11 5 4 15 0 24 2 37

1 3 2 6 2 2 4 0 8 0 14

6 15 12 33 6 4 25 0 35 1 69

0 1 0 1 0 0 1 0 1 0 2

Grand Total

Misc

Cotton

Clonard or similar type

Carn

Impacts

Breaghwy or similar type

Corrosion

Ronan O’Flaherty, Michael D. Gilchrist and Trevor Cowie

11 23 18 52 16 12 51 1 80 3 135

Figure 1: Overall summary of impact damage on Irish halberds

60 50 40 % 30

20 10 0

Breaghwy

Carn

Cotton

Clonard

Figure 2: Percentage of each halberd type with impact

Ireland and Scotland which hopes to throw some light on just how these fascinating artefacts may have been used. In Scotland, the aim of the project has also been to combine conventional archaeological research, metal analysis and experimental archaeology, in order to better interpret the Scottish data. The halberds in the Scottish dataset differ in a number of ways from their Irish counterparts and one particular aim of the Scottish arm of the project is to investigate the functionality of a series of halberds mainly found in the North-East of Scotland and recently termed the ‘Auchingoul’ type after an important find in

Banffshire (Needham 2004, 231–234). The nature of these distinctively Scottish halberds, with their weaker-looking rivet arrangements, thin hafting plates and slender, triangular blades, makes them ideal for further testing of the ‘ceremonial versus functional’ argument; moreover, the experimental results will provide a good contrast with the considerably more robust Irish halberd of the type replicated by one of the authors and his Irish colleagues (O’Flaherty 2007a; b). If shown to be as fully functional, there is a strong possibility that such blades would have been employed in a markedly different fighting style. 40

Ceremonial or deadly serious? New insight into the function of Irish Early Bronze Age halberds cas of a generic ‘Auchingoul’ type halberd have been commissioned using as authentic a metal composition and production technique as possible. Analysis of use-wear and impact damage observed on actual halberds will be undertaken and compared with the condition of the replicas (although it may be noted that, on the whole, soil conditions appear to have been less conducive to the survival of Scottish blades, which are frequently considerably more corroded than the Irish series). As well as investigating whether Scottish halberds were used, a major goal of the project is to explore whether and how they could have been used in actual combat. The key element of the final stage of the project will therefore be to bring the practical experience and skill of re-enactors familiar with traditional weaponry to bear on the question of the combat functionality of the different types of Irish and Scottish halberds. This will involve the design and commissioning of accurate replicas, the briefing and observation of re-enactors (suitably protected!) hopefully with the aid of digital video-tagging to help identify fighting techniques. In addition to the academic research, the project will therefore lend itself to outreach and the popular presentation of archaeology. After some unforeseen setbacks, the experimental component of the Scottish project work is now back on track – and as this paper will make clear, the remaining stages of the project can now draw on the sterling experimental results gleaned from the Irish research. It is to that aspect of the project that we now return.

In Scotland, our main strands of work to date have involved re-assessment of all the recorded halberds. As in Ireland, the survey of S. Ó Ríordáin (1937) provides a baseline until superseded by J. Coles’ wider-ranging review of Early Bronze Age metalwork in Scotland (1969), which included a revised list of all the known Scottish halberds, line drawings of many of them and a catalogue of the hoards. More recently, as noted above, halberds have figured in S. Needham’s excellent re-assessment of the Early Bronze Age in the North-East of Scotland (2004). The inventory of material from Scotland is much smaller than that from Ireland, making reappraisal a practical proposition: taking recent unpublished finds into account, the number of halberds from Scotland amounts to nearly 60 specimens, nearly all in collections readily accessible for research purposes (compared to nearly 190 examples from Ireland: O’Flaherty 2002). New discoveries and antiquarian research are also offering opportunities to assess the circumstances of discovery and context of deposition. The value of this source criticism as an essential first stage in the project has become apparent with the recent re-provenancing of a fine halberd in the collections of the National Museums of Scotland, for long believed to be from the Kilmartin area of Argyll in the West of Scotland. As one of our finest and most complete halberds, this blade was initially earmarked for replication as part of the experimental element of the project! Recently however, we have been able to show the true provenance of the ‘Poltalloch’ halberd was actually the Channel Island of Alderney! (O’Connor et al. 2009; O’Connor et al. 2010). A fresh look at archival sources has also shed light on the circumstances of discovery of another Scottish halberd: manuscript collections of the Society of Antiquaries of Scotland appear to show that a halberd found in 1826 at Portmoak Moss in Kinross was discovered with the remains of its wooden haft – the first indication that any of the Scottish halberds was actually deposited while hafted. The project is also providing an opportunity to augment the available metal analyses for Scottish halberds by sampling recent finds and by filling in the gaps in the SAM dataset (Junghans et al. 1968). As in Ireland, experimental techniques will be employed to investigate whether Scottish halberds were used, whether and how they could have been used in actual combat, and to investigate whether any observed blade damage presenting on museum specimens is best explained in this manner. After some initial problems, repli-

EDGE DAMAGE ON IRISH HALBERDS Examination of 135 museum specimens in the hand showed that 52 or nearly 40 % displayed evidence for edge damage in the form of notching and/or denting. Figure 1 summarises the overall situation by halberd type and in addition records the relative level of corrosion present on specimens where no edge damage was recorded. The first thing to be said is that these are minimum figures since many halberds do not survive in good enough condition to allow assessment for this type of damage. As the table at figure 1 shows, 28 of the 80 halberds for which no evidence of edge damage was noted also display medium-to-high levels of corrosion. If we are to allow in some measure for this and express the proportion of edge-damaged halberds as a percentage of those which can be accurately assessed (i.e. excluding those with medium-to-high 41

Ronan O’Flaherty, Michael D. Gilchrist and Trevor Cowie

3

4

5

6

7

Figure 3: Impact damage, V-notching (halberd, National Museum of Ireland, Ref. R1978) Figure 4: Impact damage, U-notching (halberd, National Museum of Ireland, Ref. 1929:1499) Figure 5: Impact damage, denting (halberd, National Museum of Ireland, Ref. P252) Figure 6: Impact damage, denting (halberd, National Museum of Ireland, Ref. W234) Figure 7: Impact damage, bowing (halberd, National Museum of Ireland, Ref. W247) 42

Ceremonial or deadly serious? New insight into the function of Irish Early Bronze Age halberds levels of corrosion), we might think in terms of just under 50 % of Irish halberds as showing evidence for edge damage. On the other hand, and once again excluding those with medium-to-high levels of corrosion, it is equally clear that a very similar percentage displays no evidence of edge damage at all. Whether this means that these were never actually ‘used’ is another matter – it depends very much on the type of use to which the artefact may have been put. For example, the replica halberd employed to test the effectiveness of the artefact on sheep-heads (O’Flaherty 2007a) shows virtually no damage whatsoever, despite having pierced some 20 sheep-heads. That said, the possibility that some Irish halberds were never put to any utilitarian purpose is strongly suggested by an examination of the way in which impact-damage is represented across the various halberd types, as the chart at figure 2 illustrates. It is immediately apparent that Type Breaghwy halberds are not seeing the same sort of use as other halberds: in fact halberds of this type are THREE TIMES less likely to display such damage as the next lowest of the halberd types, Type Carn. In this respect, it is worth noting that Type Breaghwy halberds are quite different from the other halberd types – in the first place they are made of bronze, not copper, and appear to be quite late in the sequence (O’Flaherty 2002). Secondly, they tend to be flatter, wider and more weakly hafted than other types and, overall, give the impression of an artefact more concerned with display than utility. The fact that such halberds show so little evidence of the impact damage found on other types might even be regarded as negative evidence in support of the view that its occurrence on other halberd-types does in fact result from their being put to more utilitarian purpose.

Each type is illustrated by an example from one of the museum specimens (all images reproduced courtesy of the National Museum of Ireland). Notching (fig. 3 and 4) This is one of the classic impact types on Irish halberds and is also observed on later prehistoric sword-blades.2 There appear to be two main types on the halberds, i.e. ‘V’ notching, so-called because the profile of the impact is v-shaped, and ‘U’ notching where the profile is u-shaped. Both types of notching can appear anywhere along the blade edges, either at right angles to the edge or acute angles, and are often associated together or with denting on the same blade. The first type, V-notching, is strongly suggestive of impact with another blade and this is normally how it would be interpreted on swords. It is almost always deeper than it is wide, with typical depths between 1–6 mm. U-notching, on the other hand, tends to be quite regular in its proportions, e.g. 1x1 mm or 2x2 mm and generally not exceeding 2.5 mm. The regularity of its profile is particularly interesting. Initial examination uncovered no clues as to what might have created this type of impact. Where it occurs on swords, experiments carried out by B. Molloy (2006; 2007) suggested impact on a metal shield rim, but this is not a realistic explanation for such damage on the much earlier halberds. However, the experimental work to be described later in this paper quite unexpectedly provided the explanation which eluded us at this early stage. Denting (fig. 5 and 6) This category includes quite a wide range of impacts, all of which would generally be described as ‘denting’ of some sort or other. This type of impact tends to be shallow, but generally greater than 1.5 mm in depth, and capable of extending over significant lengths of the blade edge. It can occur anywhere along the blade and is associated on some halberds with V and Unotching. In a small number of cases, the denting can be continuous, extending for up to 125 mm and presenting a slightly ‘hammered’ effect to the eye (fig. 5). In one case (National Museum of Ireland Ref: W234), a series of regular clustered impacts about 1.5 mm deep and 4.5 mm wide on the underside of the halberd presents an almost

The types of edge damage on Irish halberds The types of edge damage found on Irish halberds have been described elsewhere (O’Flaherty et al. 2008) but for ease of reference are summarised here. There appear to be three broad categories of impact damage, i.e. notching, denting and bowing. A number of sub-categories can also be identified, but these are easily accommodated within the three broad categories mentioned. Each of these is discussed below in terms of form, breadth, depth, angle, location and association. 2

Bridgford 1997; Molloy 2007; Kristiansen 2002; Quilliec 2008.

43

Ronan O’Flaherty, Michael D. Gilchrist and Trevor Cowie

45

B owing

40

Number of Impacts

Dent 35

Not ch

30 25 20 15 10 5

9 W 26

R1 97 8

P2 52

W 26 2

7 W 24

W 23 4

53

91 3: 11 2

SA 1

R2 5

17 R1 3

14 99

19 29 :

19 3

5: 44 8

0

Halberd Ref

Figure 8: Combination and incidence of impacts on halberds (condensed study-group)

B

A Figure 9: Replica artefacts used in the impact tests: halberds, axe and daggers (A) and stone-axe (B). Metal Replica artefacts used in the impact tests – halberds, axe and dagger (A) and stone-axe (B). Metal artefacts are shown as removed from mould with casting accretions visible at edges (subsequently removed by post-casting treatment)

Figure 10: The test facility 44

Ceremonial or deadly serious? New insight into the function of Irish Early Bronze Age halberds serrated effect (fig. 6), extending some 35 mm from just below the hafting-point. The regularity of the damage in this case, and in others, is suggestive of some form of deliberate destruction.

then designed to see what type of edge damage resulted when each was impacted against the target halberd at differing levels of force. The only known contemporary metal artefacts from the period that could possibly be considered as a source of impact damage are (a) other halberds, (b) flat axes, or much less likely (c) simple daggers. Accordingly the replica metal artefacts created consisted of five Type Cotton halberd blades, one Lough Ravel type axe, and a simple triangular dagger of Type Corkey. All were cast in copper with additional arsenic added at the proportion of 2 % to recreate the quality of arsenical copper used for the vast majority of Irish halberds. The casting work was carried out by B. Rankin of Irish Arms (www.irisharms.ie). Post-casting treatment and edge-hardening was carried out by N. Burridge of Bronze Age Craft in Cornwall (www.bronze-age-craft.com). A number of real and replica stone axes were very kindly provided by the School of Archaeology, UCD, as it was our intention to test the impact signatures created by contemporary stone artefacts as well. The real axes had already been damaged by sampling under a previous programme, but maintained good cutting edges and were more than sufficient for our experiments. The final piece of the assemblage was a timber shaft, replicating the deduced proportions of a real halberd shaft (see O’Flaherty et al. 2002 for further details). Our purpose here was to see whether any of the damage which had been observed on the museum specimens might have resulted from impact with a timber shaft, perhaps when a blow was being parried. It was felt that some of the denting which had been observed might conceivably have occurred in this way. Testing the potential of timber to create edge damage on a metal blade would also give some indication as to whether impact with a wooden shield rim should also be considered and, if necessary, a replica created for testing.

Bowing (fig. 7) This occurs on a number of Irish halberds and has also been identified on later sword-types. Rather than being notched or dented, the thin blade edge is bent or bowed away suggesting a different type of impact. It is often interpreted as combat damage on swords, and some experimental work has shown that impact on bone can produce this effect (Molloy 2007). Depth is typically 1–1.5 mm, but the area of bowing can extend for as much as 8 mm. To date, on the halberds examined, it appears on the distal half of the blade, on both upper and lower edges. Combination and incidence Figure 8 shows the incidence and combinations of the various impact types on the reduced studygroup of eleven halberds selected for the characterisation study. It is the authors’ intention to extend this analysis of incidence and combination to the wider population of Irish halberds but for the present this chart gives some indication of the likely patterns. The table at figure 1 showed very clearly that notching was by far the most common form of edge damage, occurring on some 45 % of halberds displaying edge damage. The incidence is higher in the condensed study group as this is made up of halberds selected for the purpose of characterising different types of damage but the combination with denting/bowing at 36 % is strikingly similar to the figure of 35 % obtained for the population as a whole in earlier studies (O’Flaherty 2002).

Post-casting treatment THE REPLICATION PROCESS The edge-hardening of the replica metal artefacts posed a particular challenge as the standard of hardness displayed by the original copper halberds is remarkably high, rivalling some of the best Late Bronze Age work. In order to ensure that these standards would be achieved, experiments were undertaken with one halberd first to identify the process that seemed best able to deliver the required level of hardness. The hardened halberd was delivered to UCD for testing.

Having characterised the main types of edge damage found on the prehistoric halberds, the next stage in the process was to try to find out just how this damage occurred in antiquity. It was decided to assemble a set of real and replica artefacts from the period, any of which might conceivably have been used in a combat-situation and which were capable of inflicting some form of edge damage to a halberd (fig. 9). A set of experiments was 45

Ronan O’Flaherty, Michael D. Gilchrist and Trevor Cowie

Test Data Energy

2 Joule

Mass

6.5 kg

Velocity

0.79 m/s

Swing speed

2.96 m/s

Energy

5 Joule

Mass

6.5 kg

Velocity

1.24 m/s

Swing speed

4.67 m/s

Energy

10 Joule

Mass

6.5 kg

Velocity

1.76 m/s

Swing speed

6.61 m/s

Energy

20 Joule

Mass

6.5 kg

Velocity

2.48 m/s

Swing speed

9.3 5m/s

Energy

30 Joule

Mass

6.5 kg

Velocity

3.04 m/s

Swing speed

11.45 m/s

Solid Base

Yielding Base

No test

Figure 11: Impact tests: halberd on halberd (all replicas). Scalebars in mm

A cross-section was excised by cutting with a water-cooled cut-off wheel. This cross-section was then mounted in resin and prepared for metallurgical examination. Hardness tests were carried out on this cross-section, both in the middle section and near the edge, to determine the effect of work hardening at the cutting edge of the halberd.

In the past, the Brinell hardness test method has often been employed to determine the hardness of Bronze Age artefacts. This method uses a 10 mm hardened steel ball indenter and loads of up to 3000 kg. This method results in very large indentations in the tested material. Furthermore, the results obtained are not independent of load. For the purpose of testing the replica halberds, the 46

Ceremonial or deadly serious? New insight into the function of Irish Early Bronze Age halberds Test Data Energy Mass Velocity Swing

10 Joule 6.5 kg 1.76 m/s 6.61 m/s

Energy Mass Velocity Swing

20 Joule 6.5 kg 2.48 m/s 9.35 m/s

Energy Mass Velocity Swing

30 joule 6.5 kg 3.04 m/s 11.45 m/s

Timber Shaft

Stone Axe

Copper Axe

No test

No test

Figure 12: Impact tests: replica halberds and other artefacts. Scalebars in mm

13

14

Figure 13: Wide V-notch on museum specimen (Halberd NMI 1935:448 - Impact A) Figure 14: Massive damage inflicted by full-force halberd-on-halberd impact (102.8J) against a rigid base. Scalebars in mm

Vickers hardness test method was selected, using a pyramidal diamond indenter and a load of 1 kg, as a combination of low loads and very small indentation size enables valid measurements to be carried out very close to the edge of the section. The value of the hardness obtained by the Vickers method is independent of the load employed. The hardness measurement obtained at the centre of the section in the unhardened

(as-cast) region was 49.0 HV1.0, and near the blade edge, 142 HV1.0. This represents a considerable increase in hardness following post-casting treatment and is equivalent to the hardness levels recorded for Early Bronze Age pieces.3 Having established that this process would deliver the required level of hardness, the remaining replica metal artefacts were finished to this standard.

3

Penniman and Allen 1960 recorded that the edge hardness of museum specimens had been increased from

62 HB to 150 HB by cold hammering.

47

Ronan O’Flaherty, Michael D. Gilchrist and Trevor Cowie

results for halberd-on-halberd action, while the second table shows the edge damage resulting from impact with other artefacts. The effect of allowing some degree of ‘give’ in response to the blow is immediately apparent in the halberd-on-halberd results. Although different zoom-imagery has been employed in order to record the damage to best effect, the scale bar – which represents millimeters in every case – clearly shows the much deeper and broader impact damage occurring when the target piece was not allowed to yield at all. It should be noted that the images reproduced here represent just a sample of those taken: each test was carried out a number of times, and the resulting damage recorded in each case. The samples reproduced here are representative of the type and scale of damage occurring. An equivalent ‘swing-speed’ for an actual halberd in use is also given for each test: thus, for example, a force of 2 joules is equivalent to a halberd swing of just under 3 m/s, which is quite slow and certainly too slow to indicate combat use. However at 20–30 Joules of force, the equivalent swing-speed is between 9–12 m/s: a reasonably fast but measured blow. As regards the effect of impact by other artefacts, tests at the lower levels of force produced little evidence for damage, so only the results of tests taken at higher levels of force are reproduced here. No tests have as yet been undertaken using the small dagger, as it proved difficult to imagine (and therefore reproduce) the circumstances where such an artefact might have been brought into play with a level of force which might actually damage a hardened halberd blade. A dagger is used at close-quarters, in a stabbing or cutting action. Without the leverage, reach and added weight of a shaft of some length, it is difficult to see how a dagger could be employed against a halberd in circumstances which would result in the type of damage observed on the museum blades. However, we continue to reflect upon this point.

Impact replication Having characterised the various types of edge damage and assembled a set of real and replica artefacts for testing, the next step was to design the test-environment. The impact replication was carried out using a Rosand high strain rate test facility, i.e. an instrumented drop-tower impact machine. The basic procedure employed was very simple: the target piece (halberd) was fixed in a vice and then the impacting piece (halberd, copper axe, stone axe, etc.) was allowed to drop at various impact velocities (fig. 10). Previous experimental and computational work by one of the authors had analysed similar impact scenarios but involving biological tissue (McCarthy et al. 2007; 2009). The resulting edge damage was recorded against each of the relevant velocities, with force calculated as a function of impact energy, and the velocity and mass of the impactor. The impacts were designed to correlate levels of energy against observed damage on the halberds which, in turn, was compared directly against that found on the museum specimens. Due to the large carriage mass of 6.5 kg, the Rosand-rig operates at lower than actual impact velocities, but the results achieved were validated using a separate and independent wire-guided drop test rig, which operated on actual impact velocities and which confirmed the Rosand-rig results as accurate. The integrated instrumentation on the Rosand test machine provided more complete data than could be obtained solely from this wire guided rig and it is for this reason that the following results are based on this more accurate data. In order to allow the maximum number of tests, the halberds used were chopped into a series of segments and mounted separately, as required. Initial tests were carried out with the target piece fixed upon a rigid, unyielding base which resulted in very serious levels of edge damage at quite low levels of force. Such an environment is highly unlikely to occur in a combat situation, where the target would be expected to yield and move with the blow of impact, but would not be inconsistent with what might be expected of deliberate destruction. The results obtained are accordingly of interest in terms of considering deliberate destruction as a source of edge damage. However, in order to better represent a combat environment, the test environment was redesigned to incorporated a foam-cushioned base to allow the target piece to ‘yield’ somewhat with each blow. The results obtained are shown in the two tables at figures 11 and 12. The first table shows the

DISCUSSION The results of the various tests proved most informative. There is no doubt that the most common form of edge damage observed on the museum specimens, i.e. notching, was regularly reproduced by a halberd-on-halberd action. What was even more interesting, and completely unanticipated, was the fact that both V-notching and U-notching were created in this way. All that was different was the environment. In a yielding environment, the blade-edges moved slightly on impact, creating 48

Ceremonial or deadly serious? New insight into the function of Irish Early Bronze Age halberds

Figure 15: Penetration by halberd of sheepskull during on-site trials

deep V-notch, but this is of a wide-mouthed type that was not created by the halberd-impacters. The polished stone axe also delivers a wide-mouthed notch, which is less angular than the copper axe, but again quite unlike anything created by the replica halberd-impacters. This type of widemouthed V-notch is found on some of the museum specimens (see fig. 13 for an example) and as our tests seem to rule out impact with another halberd as the likely source, the real possibility exists that we are looking at impact with an axe, whether that be stone or copper. The actual specimens displaying this type of damage would have to be re-examined to see whether this represents some form of deliberate destruction rather than combat, but the instance reproduced at figure 13 would not seem to fall into that particular category. The timber-shaft was quickly ruled out as a possible source of edge damage. Bowing or denting had been considered as possible results, but in fact the blade edge showed practically no damage whatsoever from impact with the timber shaft, even at a force of 20 joules. Damage to the shaft, on the other hand, was considerable, comprising of deep cuts into the timber. In fact, none of the tests reproduced the type of denting or bowing found on the museum specimens, apart from some impacts against the sides

that characteristic U-shaped profile found on so many of the museum specimens. There was no need to seek an artefact with a U-shaped profile as the source of this type of damage (such as a hypothetical shield-rim): it results from impact with a sharp-bladed halberd in circumstances where both are in motion. On the other hand, the deeper V-notching seems to occur when both blades are held more rigidly and where there is therefore less scope for lateral movement. Overall, the edge damage observed on the museum specimens seemed best reproduced by impacts at energy levels of between 20–30 joules. A second, and once more completely unanticipated result, was the realization that the ‘double-impact’ signature observed frequently on museum specimens and consisting of two (or sometimes more) closely grouped notches (generally u-shaped), was also reproduced when two halberds were struck in a yielding environment. Examples of these can be seen in the halberd-onhalberd table (fig. 11), at force levels of 5 and 10 joules. Each of the tests represents a single ‘drop’ through the apparatus, and this double-impact signature is created when the impacter rebounds to strike again rapidly, but at less force, close to its original impact. Turning to look at some of the other artefacts, it is clear that the copper axe can also deliver a 49

Ronan O’Flaherty, Michael D. Gilchrist and Trevor Cowie

Test No. 1 2 3 4

Impact Energy (J) 32.28 22.73 17.21 15.59

Drop height (cm) 253 215 155 126

Velocity (m/s) 5.41 4.54 3.95 3.76

Force (N) 1388 1454 2363 883

Penetration (mm) 35 17 13 0

Figure 16: Test results for halberd impacts on sheep heads

whether the level of force associated with those impacts constituted a ‘lethal’ level of force. If it did, then all the evidence would point towards the use of halberds in some form of combat situation; if it did not (and while not ignoring the anthropological evidence for sub-lethal forms of combat), it simply could not be said with any degree of certainty that the impacts observed on the museum specimens had resulted from combat use. In order to establish the lethal capacity of the force-levels in question, one further test was conducted. On 17th June 2010, a team led by the second-named author returned to ICM Camolin, the sheep-killing plant in Co. Wexford, Ireland where the original trials which gave birth to this project had been carried out back in 2002. Once more, freshly killed sheep-heads were selected for the tests, these providing reasonably good parallels for human skulls in terms of bone thickness (if anything, being slightly thicker). The halberd blade was fixed in a movable rig designed by UCD which allowed similar data-capture to what had been achieved in the laboratory. The halberd was then allowed drop on the sheep skulls at differing levels of energy equivalent to those which had been identified as responsible for the impact damage on the museum specimens (fig. 15). The results are shown in the table below (fig. 16) It will be recalled that the laboratory tests had shown that the impact damage present on the museum specimens had most likely been caused by impact with other halberds and energy levels of between roughly 20–30 Joules. The results clearly show that this level of energy routinely results in penetration of the skull: in other words, it is a potentially lethal level of force.

of the stone and copper axes. Impact against bone is another possibility for bowing in particular, and the project-team are examining its options in this regard. Practical experiments with replica swords of the Late Bronze Age indicate this as the source of bowing observed on sword-edges (see Molloy 2007) so there is every possibility that this may prove to be the case also for the halberds. Certainly none of the other options examined to date seem to explain this. In terms of considering deliberate destruction as a possible source for edge damage on Irish halberds, the hypothesis seems unlikely. Our experiments indicate that a forceful, swinging blow – such as might be imagined should one attempt destruction of the artefact by striking it off something else – would produce much higher levels of edge damage than are observed on the museum specimens. Similarly, should the halberd be placed on a rigid base and then struck forcefully with another object, again our experiments show that the resulting edge damage would exceed anything observed on the Irish halberds to date. By way of illustration, figure 14 shows the effect of hitting one halberd against another on a rigid base at a force of 102.8 joules – a swing speed of 17.8 m/s, which would represent a full force blow by a six-foot man. The resulting edge damage is substantially deeper and wider than anything observed on surviving museum specimens. This is not to say that no Irish halberd was the subject of deliberate destruction. On the contrary, it is possible to identify a small number whose condition is strongly suggestive of deliberate destruction of some sort prior to deposition. However, these are unusual: the vast majority of Irish halberds either display the sort of small-scale edge damage discussed already, or display no edge damage at all.

CONCLUSIONS What conclusions can we draw from all this about the way Irish halberds were used in the past? It has been shown already by one of the authors (O’Flaherty 2007a; b) that the Irish halberds were certainly capable of being used with lethal force. In tests using the heads of freshly-killed sheep, a replica Type Cotton halberd repeatedly pierced the

A lethal level of force Having established that the most likely source for most of the impacts observed on the museum specimens was impact with another halberd, the last remaining question to be answered was 50

Ceremonial or deadly serious? New insight into the function of Irish Early Bronze Age halberds with the impact replication module is most gratefully acknowledged. The Scottish arm of the project is being co-ordinated by T. Cowie and is being undertaken in collaboration with Dr B. O’Connor and Dr P. Northover, University of Oxford (metal analyses), with additional valuable input from Dr. C. Horn (post-doctoral researcher, University of Göteborg). We are indebted to the Society of Antiquaries of Scotland and the Russell Trust for their generous support for the project.

skulls of 20 separate animals, and gave no indication that it could not continue doing so indefinitely. The halberd itself suffered no damage and was easy and effective to handle. The comments traditionally made about the unwieldy nature of the halberd as a weapon and the weak hafting technique are simply not supported by the evidence. There is no doubt that the halberd could have been used very effectively as a weapon and need not be considered as ceremonial in nature. The next question to be asked was whether these halberds had in fact been so used. The evidence would seem to suggest that some were and some were not, in roughly equal proportions, but probably with some differentiation in terms of halberd type. Certainly, the later Breaghwy types seem much more likely to have served a ceremonial or display purpose, having very light rivets and displaying practically none of the edge wear found on other halberd types. As regards the edge wear itself, our experiments show that the type of impact observed on the museum specimens is best reproduced by halberd on halberd action. Halberd on halberd action is capable of reproducing both V- and U-notching, which are by far the most common types of impact damage occurring on the museum specimens. The experiments also strongly suggest that this took place in a ‘yielding’ environment, best imagined by both weapons being held in the hand. In terms of force, the morphology of the impacts observed on the museum specimens was best reproduced at energy levels of not more than 30 joules. This equates to an average swing speed of not more than 11 m/s. This is a ‘reasonable’ speed, indicating a high degree of control, but the question remained as to whether this amounted to a ‘lethal’ level of force, consistent with use in a combat situation. A specific test was designed to answer this question and this showed, unequivocally, that the levels of force involved are indeed sufficient to penetrate the skull. If we return to that legal precept referred to at the start of this paper, that ‘force goes to intent’, then it would seem that at least some of these Irish Early Bronze Age halberds were indeed intended to hurt, maim and kill. It is a sobering thought that amongst the very first artefacts that humans in Western Europe chose to make from metal was an object which apparently could be used for one purpose only, and that was the killing of other humans.

BIBLIOGRAPHY Brandherm, D. 2003. Die Dolche und Stabdolche der Steinkupfer- und der älteren Bronzezeit auf der Iberischen Halbinsel. Prähistorische Bronzefunde, VI, 2. Stuttgart, Steiner. Brandherm, D. 2004. Porteurs de hallebardes? Überlegungen zur Herkunft, Entwicklung und Funktion der bronzezeitlichen Stabklingen. In H.-J. Beier and R. Einicke (eds), Varia Neolithica III, 279–334. Langenweissbach, Beier and Beran. Bridgford, S. 1997 Mightier than the pen? (An edgewise look at Irish Bronze Age swords). In J. Carman (ed.), Material harm: archaeological studies of war and violence, 95–115. Glasgow, Cruithne Press. Coles, J. M. 1969. Scottish early bronze age metalwork. Proceedings of the Society of Antiquaries of Scotland 101 (1968–69), 1– 110. Herity, M. and Eogan, G. 1977. Ireland in Prehistory. London, Routledge. Junghans, S., Sangmeister, E. and Schröder, M. 1968. Kupfer in der frühen Metallzeit Europas. Studien zu den Anfängen der Metallurgie 2. Berlin, Gebr. Mann. Kristiansen, K. 2002. The tale of the sword – swords and swordfighters in Bronze Age Europe. Oxford Journal of Archaeology 21, 319–332. Macalister, R. A. S. 1949. The Archaeology of Ireland (second edition). London, Methuen. Mallory, J. P. and McNeill, T. E. 1991. The Archaeology of Ulster from Colonization to Plantation. Belfast, Institute of Irish Studies, Queen’s University. McCarthy, C. T., Hussey, M. and Gilchrist, M. D. 2007. On the sharpness of straight edge blades in cutting soft solids: Part I: Indentation experiments. Engineering Fracture Mechanics 74, 2205–2224. McCarthy, C. T., Ní Annaidh, A. and Gilchrist, M. D. 2009. On the sharpness of straight edge

ACKNOWLEDGEMENTS The generous assistance of the Royal Irish Academy in meeting some of the costs associated 51

Ronan O’Flaherty, Michael D. Gilchrist and Trevor Cowie

blades in cutting soft solids. Part II: Analysis of blade geometry. Engineering Fracture Mechanics, 77, 437–451. Molloy, B. P. C. 2006. The role of combat weaponry in Bronze Age societies: the cases of the Aegean and Ireland in the Middle and Late Bronze Age. Unpublished PhD thesis, University College Dublin. Molloy, B. P. C. 2007. What’s the bloody point? Bronze Age swordsmanship in Ireland and Britain. In B. P. C. Molloy (ed.), The Cutting Edge: Archaeological Studies in Combat and weaponry, 90–111. Stroud, Tempus. Needham, S. 2004. Migdale-Marnoch: sunburst of Scottish metallurgy. In I. A. G. Shepherd and G. J. Barclay (eds), Scotland in Ancient Europe. The Neolithic and Early Bronze Age of Scotland in their European Context, 217– 245. Edinburgh, Society of Antiquaries of Scotland. O’Connor, B., Cowie, T. and Horn, C. 2010. Une trouvaille oubliée de l’ancien duché de Normandie et sa redécouverte. Bulletin de l’Association pour la Promotion des Recherches sur l’Âge du Bronze, no. 7, mars 2010, 4–5. O’Connor, B., Cowie, T., Horn, C. and Sebire, H. 2009. A famous general’s forgotten find: the provenance of the halberd from the Poltalloch Collection rediscovered. Kilmartin House Museum Newsletter 17 Autumn/Winter, 3. O’Flaherty, R. 2002. A consideration of the Early Bronze Age Halberd in Ireland – Function and Context. Unpublished PhD thesis, University College Dublin.

O’Flaherty, R. 2007a. A weapon of choice – experiments with a replica Irish Early Bronze Age halberd. Antiquity 81, 423–434. O’Flaherty, R. 2007b. The Irish Early Bronze Age Halberd: Practical experiment and combat possibilities. In B. P.C. Molloy (ed.), The Cutting Edge: Archaeological Studies in Combat and weaponry, 77–89. Stroud, Tempus. O’Flaherty, R., Bright, P. Gahan, J. and Gilchrist, M. D. 2008. Up close and personal. Archaeology Ireland 22, No. 4, 22–25. O’Kelly, M. J. 1989. Early Ireland. Cambridge, Cambridge University Press. Ó Ríordáin, S. P. 1937. The halberd in Bronze Age Europe. Archaeologia 86, 195–321. Osgood, R. 2000. Central and Eastern Europe. In R. Osgood, S. Monks and J. Toms (eds), Bronze Age Warfare, 65–88. Sutton, Stroud. Penniman, T. and Allen, I. 1960. A metallurgical study of four Irish Early Bronze Age ribbed halberds in the Pitt Rivers Museum, Oxford. Man 60, 85–89. Quilliec, B. T. 2008. Use, wear and damage: treatment of bronze swords before deposition. In C. Hamon and B. Quilliec (eds), Hoards from the Neolithic to the Metal Ages, 67–78. British Archaeological Reports, International series 1758. Oxford, BAR Publishing. Toms, J. 2000. Italy. In R. Osgood, S. Monks and J. Toms (eds), Bronze Age Warfare, 89–114. Sutton, Stroud. Waddell, J. 1991 The earlier Bronze Age. In M. Ryan (ed.), The Illustrated Archaeology of Ireland, 68–71. Dublin, Country House. Ronan O’Flaherty Crane Bag Consulting Ballyhitt Lane Barntown Co. Wexford Ireland Michael D. Gilchrist School of Mechanical & Materials Engineering University College Dublin Belfield Dublin 4 Ireland Trevor Cowie Department of Archaeology National Museums Scotland Chambers Street Edinburgh EH1 1JF Scotland 52

DELIBERATE DESTRUCION OF HALBERDS Christian Horn

ABSTRACT

are much better, because they were usually riveted to their handle, hence a more or less violent removal would certainly leave some traces. And indeed, damage that most likely derives from such a treatment could be observed during the research on several original halberds. In an extension of these studies drawings of all European halberds were examined in order to spot indicators of similar damage. Due to the quality of the drawings and reproductions it is often linked to several difficulties and uncertainties. Therefore this article is to be seen as a hypothesis and as a suggestion to explain some of the severe damage seen on halberds. To explain not just this destructive treatment some remarks are given in the concluding part, in order to provide a wider context to destruction as a sacrificial act.

The frequent deliberate destruction in hoards that L. Nebelsick has observed did not start at earliest in the Middle Bronze Age. In fact it can be traced back to far earlier times. It occurs frequently on halberds not just in hoards, but in graves and single finds as well. The method to destroy the object was the violent separation of handle and blade. There are various causes to carry out, but the underlying psychological structure for destruction as sacrifice remains the same. When the latter appears on weapons especially in pure weapon hoards it is most likely interlinked with prehistoric warfare. The destruction is just one element of the sacrifice and deposition is another. If the latter element was not part of the ritual the whole sacrifice escapes the archaeologists’ attention. This might explain why these hoards are missing in some areas. Halberds that sometimes appear in graves in these areas confirm this proposal.

CRITERIA The decision whether a handle was removed or not is based on a range of indicators. Some of them are strong other are weak. Strong indicators for the detachment of handle and blade are twisted and/or bent hafting plates (fig. 1) as well as twisted blades. The bending of a blade is of course a deliberate destruction, but it does not indicate the separation of the handle and the blade (fig. 2). When several rivet-holes are broken, a violent separation most likely took place. However, the twisting of the blade is hardly visible in any drawing, because a longitudinal section through the blade has to focus of the midrib and tend to conventionalise the run of the cutting edges. But it is either not visible in a cross-section, because this is just a small section and at any point the derivation is almost always just minimal. The downsizing of drawings for publications is also a conventionalising factor (fig. 3). Several halberds with metal shaft-heads had also a metal coating around the wooden handle. These two components were casted separately. They were connected through a ring which was casted over the two components (Drescher 1958). It is a clear indicator for the separation of the handle and the shaft-head, when this connection ring is broken (fig. 4). This can be observed for example on pieces from the famous hoard

KEYWORDS Halberd – destruction – sacrifice – hoard – grave – psychoanalysis INTRODUCTION In 1997, and then again in 2000, L. Nebelsick stated in articles concerning the deliberate destruction of objects in Late Bronze Age hoards that an axe or a spear is also destroyed when the shaft is broken, even though the axe or the spearhead itself may appear fairly intact. He considers the destruction of these objects to be a violent process that was undertaken in an ecstatic state of mind. According to him this process appears at the earliest in the Middle Bronze Age (Nebelsick 1997, 36 ff.; 2000, 160 ff.). Due to obvious reasons the separation of the handle and the axe in respect to the spearhead is hardly recognizable. An axe is simply too massive to gain damage even if the handle is torn away violently. The spearhead might receive some recognizable damage, but due to its general form it is more likely that the shaft is simply broken. For halberds the chances to observe such a treatment 53

Christian Horn

1

2 Figure 1: Side view of the hafting plate from Crott, Co. Longford, Ireland (National Museum of Ireland, Dublin 1935.448) Figure 2: Side view of a halberd from ‘Prov. Jaén’, Spain (British Museum, London 1909.7-14.7)

of Melz II (Müritz, Germany). The hoard was discovered during land improvement works. In fact this hoard consists of two depositions. Five metal halberd-blades together with a flanged axe casted on a metal shaft lay 1.75 m deep in the peat bordering on a silt-sand layer. Just two metres away at the same depth eight metal halberdshafts were discovered (Schoknecht 1971, 233). Later a sixth halberd-blade was discovered in the excavated material. Two shaft-heads can be related to two of the shafts. The other shafts belonged seemingly to other halberds (Schoknecht 1971, 236). It is apparent that the halberd-blades and the shafts were separated and sunk into a lake at nearly the same time. The shaft of the axe was broken too (Rassmann/Schoknecht 1997, 44). It is possible that the remains discovered belonged to a total number of twelve halberds. However, in case of a completely missing connection ring only a groove at the lower end of the shaft-head

proves that it had once a metal shaft (fig. 5). In Deszczno (Gorzów Wielkopolski, Poland) just a part of the metal coating of the shaft was deposited in a hoard (Gedl 1980, 37 no. 62 pl. 10, 62). The complete loss of rivets is just a weak indicator, because most finds were made purely by chance. Hence, they might be missing accidentally. Although in some cases this was considered as a sufficient indicator, because certain find circumstances1 or the comparison with other finds in the surrounding area2 proved that the separation took place. When all rivets are missing and more than two-thirds of the rivetholes are broken the separation can be assumed as certain. However, present rivets are a strong indicator for a deposition with attached handle, except when they are extremely bent (fig. 6). Another quite strong indicator for a handle left in place are wooden remains on the hafting-

1 For example, the five blades from the hoard of Hofkirchen-Unterschöllnach (Passau, Germany) were found together in one vessel. It is obvious that there was no room for the handles, especially because the vessel was closed with a bowl. In fact this is one of the rare occasions where

it is perfectly clear that the handle of the associated axe was removed as well (Wandling/Engelhardt 2007, 38 ff.). 2 In the north-western Iberian Peninsula for example, just one halberd out of nineteen was discovered with rivets. It is highly unlikely that all the others are missing by accident.

54

Deliberate destrucion of halberds

3

5

4

6

Figure 3: Comparison of the cross section from the halberd from Killaha East, Co. Kerry, Ireland; top: Harbison 1969, 46, no. 301, pl. 23, 301; bottom: cross section graphic rendering of the original halberd (National Museum of Ireland, Dublin 1939.396) Figure 4: Halberd from Pustohl-Radegast, Bad Doberan, Germany (Wüstemann 1995, 73 no. 107, pl. 14, 107) Figure 5: Halberd from Melz II, Müritz, Germany (Wüstemann 1995, 75 no. 116, pl. 18, 116.) Figure 6: Halberd from San Antón (?), Orihuela, Alicante, Spain (Brandherm 2003, 246 no. 809, pl. 50, 809) 55

Christian Horn

plate (fig. 7). When small pieces broken off from the hafting-plate are preserved and even corroded onto their former joint parts, it can be assumed that they were broken off during use, but remained in the wooden shaft head (fig. 8). On some rare occasions the deposition with attached handle is fairly certain due to the observed find circumstances, even if there are no rivets preserved. In Grave 2 from Spilamberto (Emilia Romagna, Modena, Italy) the position of the halberd in relation to the hands of the deceased proves that the halberd was put undamaged into the grave (fig. 9). In this special case a hafting without rivets is suggested. Instead it might be assumed that the halberd was secured to the handle with a wrapping (Horn 2010). However, there are of course several difficulties and some should be mentioned. Rivets can get lost during discovery or later during storage. A halberd could on the other hand be deposited without a handle, but may appear fairly intact with preserved rivets and an undamaged hafting-plate. In this case the handle could be detached in another non-violent way, but it remains of course a destruction of the object. These are in short the criteria under which the original material and the drawings were examined and assigned to the different ways of peridepositional treatment. There were of course several cases in which no distinctive interpretations could be made. However, there is certainly a process of destroying objects through the tearing of the handle, but the observations on halberds cannot confirm an ecstatic component during this process. Things done in ecstasy are more likely done with a certain speed, but, for example, the bending and twisting of the hafting-plate seems to require a slow movement. It must be done with a subsequently stronger force, so the wood did not simply give way before the halberd got ‘properly’ damaged.

The distribution of the detached handles scatters across Europe. This seems to be largely insignificant. (fig. 10) But when this distribution is integrated with the distribution of the other phenomenon (fig. 11) the picture becomes quite structured. Both ways of treatment – detached and attached shafts viewed together3 – trace existing cultural regions as for example the El Argar region in Spain,4 but they also connect regions that are seen more or less as separated such as the areas of the Remedello and the Rinaldone culture in Italy.5 It is not possible to address all factors involved here, since it would be too extensive for this chapter. The picture becomes clearer when the halberds are grouped according to their find circumstances. Before this is undertaken, two important points on the context of the European halberds will be described. For a large part of the halberds the find circumstances are simply unknown and one third are single finds. Nearly a quarter of them were deposited in hoards. At least 15 % were discovered in graves. Settlement finds are insignificant, under 0.5 % and even in these cases the relation to the settlement is mostly not secure (Fig. 12, a). In this whole dataset6 almost onethird of the halberds are simply not assignable to any form of treatment. The deposition of fragments from halberds or shafts are only special cases, but their handles were certainly detached beforehand. Therefore, they are viewed here together. About 40 % of the halberds are relatively certain to have been deposited in the ground with the handle detached, and about 22 % percent retained it and 4 % respectively 3% cannot be assigned with enough certainty (fig. 12, b). If the finds are analysed according to their find circumstances opposing trends in hoards and graves become obvious. 54 % of the halberds in hoards were deposited without their handle (fig. 13, a), whereas 53 % came into the graves complete (fig. 13, b). On the other hand just 16 % of the graves had a halberd with a detached handle. In hoards the amount of complete halberds is at 12 % even smaller. Both find circumstances provide around a third of uncertain objects. The single finds repeat this general pattern of the specimen from hoards (fig. 13, c) and shows again that they are closely related to each other. The slight blurring of this diagram might be due to some of them coming from unrecognised graves as well

MAPS AND STATISTICS The following part shows these two phenomena as reflected in their spatial distributions and in the statistics concerning the find circumstances. This will demonstrate whether these occurrences happen just randomly or if a certain pattern becomes obvious. 3

5

The distribution is here however represented in two maps, because the overlay of symbols in areas with a large amount of finds would obscure the view again. 4 For a summary of this culture see Lull 1983.

For a critique of this view see Steiniger 2005,

287 ff. 6

By now 705 halberds can be identified in Europe (Horn 2010).

56

Deliberate destrucion of halberds

that some of them possibly being the remains of unrecognized hoards. The distribution of the two peri-depositional treatments according to their find circumstances (fig. 14, 15) show that in the northern Apennine Peninsula almost only halberds with detached handles got into the graves, while they usually put attached ones in the graves of the El Argar and the Aunjetitz regions. At least in relative chronological terms the graves of the Apennine Peninsula predate those of the El Argar-culture. Halberds started being burial offerings in Aunjetitz at an even younger age. If this is compared with the hoards several shifts can be observed: one is geographical, one relates to the sacrificial customs and at least in the Apennine Peninsula it appears to be a time shift. The Circum-Harz-group of the Aunjetitz-culture shows a rather mixed picture (Zich 1996, 26 ff.; 34 ff.). Here hoards were discovered in which both phenomena reside, but also such that represent one way only. This seems to be influenced by the treatment of the halberds within graves, which is no surprise because it is the only region where halberds appear frequently in hoards and graves. In Ireland both forms of treatment are commonly used in the single finds and also appear in mixed hoards. The only grave in Ireland that provides a halberd was discovered in Moylough (Sligo, Ireland; Harbison 1969, 46 no. 303), where the halberd were burned with the dead, but the handle seems not to be removed beforehand. Scotland shows quite another picture. The halberds in hoards were exclusively put into the ground without handles. This is somewhat surprising, because Ireland and Scotland are always seen as closely related to each other in the Copper and Early Bronze Ages (Coles 1971, 68 ff.; Walker 1974, 71). But it must be mentioned that Scotland also provides some uncertain cases. However, one grave was discovered in Bishop’s Mill, Elgin (Morayshire, Scotland; Ó’Ríordáin 1937, 203 312, no. 20) in which a halberd was buried with its handle detached. In the northwest of the Iberian Peninsula halberds without handles were exclusively discovered in hoards. On the opposite in the south eastern part, in the area of the El Argar culture, almost only graves with attached handles were discovered. The few pieces which got buried without appear to be rather late and date to El

Argar B, for example the stone packing grave from Cerillo de Ciavieja, Dalías (Almería, Spain; Brandherm 2003, 383 no. 1417). In the southern to south-western regions an actual trend is hard to determine, because here halberds with rivet notches appear, e.g. in grave 3 from Alcalar, Mexilhoeira Grande (Portimão, Algarve, Portugal; Brandherm 2003, 79 no. 23–26). Southern Scandinavia and France are not represented in this observation, this is due to the fact that these regions provide almost only single finds of halberds. The above described study showed the spatial distribution of the two peri-depositional treatments – at least for the halberds –in Europe. These two phenomena seem to follow certain rules. It becomes clear that the destruction of objects in hoards started earlier than proposed by Nebelsick. The chronology cannot be discussed in detail here,8 but for the halberds it can be said, that the destructions occurred as early as at least the first phase of the northern Italian Copper Age (Remedello Phase 1; Dolfini 2004, 185 ff.), here of course in graves and in single depositions but not in depositions with multiple objects. INTERPRETATION The following remarks are a general outline for a psychological framework that might underlie not only these two peri-depositional ways of treatment, but in a wider context, the destruction of objects as a sacrificial act in general or in fact their absence. This broader view has to be chosen, because the detachment of the handle is just one way to destroy the object. To leave the handle on the object is a necessity in order to keep the object undamaged. In this sense both ways are nothing exceptional within their ritual setting. Therefore, the detachment of the handle needs no other psychological background than any other sacrificial destruction. And as far as it is necessary for the object to retain its handle to be complete the same is true for the non-destruction of them. The custom of burial offerings can relatively securely be interpreted as a sign for the belief in some kind of afterlife.9 This can be suggested in any way, whether the objects are intact or not. What the different ways of peri-depositional

8

Horn 2010. Afterlife is used in this paper in a very wide sense. It means not necessarily a new world in which the dead

transcends, but also the process of becoming an ancestor or the feared returning of the dead in order to haunt the living. (Parker Pearson 2009, 7 ff.).

9

57

Christian Horn

7

9

8

Figure 7: Wooden remains on the halberd from Grave 42 in El Oficio, Cuevas de Almanzora, Almería, Spain (British Museum, London 1889.7 – 4.221) Figure 8: Small broken off but still connected parts of the hafting plate on the halberd from Grave 449 from El Árgar, Antas, Almería, Spain (Brandherm 2003, 379 no. 1393, pl. 99, 1393) Figure 9: Grave 2 from Spilamberto (Emilia Romagna, Modena, Italy) with reconstructed shaft and accentuation of the blade and the forearm (Bagolini 1981, 109; reconstruction and accentuation by the author)

hand an object needed no special treatment for the transition to the Netherworld it might be seen as a ‘dead thing’. In this case the object possesses a dualistic nature as it was present in the realm of the dead as well as in the realm of the living. Grave 2 of Spilamberto, Emilia Romagna (Modena, Italy) supports this hypothesis quite strongly because the halberd was held by the deceased in a ‘ready to use’ position (fig. 9). Here it is clear the dead can use the same object as the living, without the requirement for the object to pass over to this world. In graves the purpose for the deposition of objects is quite clear. They are gifts for the dead, markers and symbols for the living as well as inherently a reproduction of the society as a whole. For hoards and single finds however a huge

treatment show is how the object itself is seen and how it makes the transition to the Netherworld. With its destruction the object has to undergo the same procedure as the dead. The deceased is ripped out of the realm of the living and disappears out of his surroundings. Anthropological examples show that the objects which accompany the dead equally had to ‘die’ in order to pass away with him (Parker Pearson 2009, 26; Horn 2010). Despite the fact that the halberd from the grave in Moylough (Sligo, Ireland) did not lose its handle it got equally burned with the dead to make the transition to the realm of the dead with him. As far as it concerns the halberds, this treatment is unique. However, it might be assumed that the object was, when it needed to be, destroyed to pass as a ‘living thing’. But when on the other 58

Deliberate destrucion of halberds

violently reversed in the form of the detachment of the handle. In the case of the handle a similar violent process becomes apparent. The wood got chopped from a tree with an axe and afterwards cut as well as sawed into shape. These things happen on a daily basis and do not always cause for a big sacrifice. This violence is experienced on a more or less subconscious level, this is why it is expressed for example in myths. Hence it might be enough to do it in the ‘right order’ or do something simple alongside the normal work. For a sacrifice like a deposition something extraordinary must have happened in addition so that the normal routine was no longer sufficient to compensate the experience of this violence. However, the violence and forces used during the manufacturing process are apparent. The reflection of the violence experienced during the production process can be suggested as an underlying structure especially for the mixed objects hoards with jewellery, tools and weapons. It might even explain the deposition of unfinished objects, which are considered possibly as destroyed from the very first moment of their existence on. But this almost never happens to halberds, except in the exceptional hoard from Perşinari (Văcăreşti, Dîmbovita, Romania; Vulpe 1995, 43 ff.). In the case of the mixed material hoards the gratitude as an act of balancing is most likely the underlying structure as well as the plea. For the plea the scenario could be that some object got destroyed which is a misfortune, as time and labour was put into them. This misfortune could be caused in order to ensure the supply of ore and metal or ongoing failure during the casting might trigger the sacrifice to secure the fortune of good casting again. Especially in the latter case it would make sense to sacrifice unfinished objects, because like miscasts they are ‘born dead’ and they even might be miscasts themselves. There are of course several other imaginable interpretations. For the pure weaponry hoards including those consisting only of halberds, which appear quite frequently in Europe another aspect of the psychological framework can be suggested. A plea could be the reason for a sacrificial destruction of weapons before combat action. Here the causation of the misfortune to lose weapons and therefore actual fighting power might secure victory or even the avoidance of any actual combat. The sacrifice triggered by gratitude might occur after fighting, e.g. for an achieved victory, low

variety of reasons exists for their sacrifice. In this the destruction of – at least – halberds play the major role, hence the underlying structure of this is to be examined in a general outline. Psychoanalysis provide three specific reasons for destruction as sacrificial act. It can occur as a plea, which is a wilful causation of a smaller misfortune in order to prevent a bigger one. The second reason is gratitude for example for the healing of a person, or the expected loss had to be compensated. The third reason is the execution of an object, which is experienced as a burden (Freud 1999a, 187 ff.). It is most likely that these reasons do not occur in a pure form. Despite the fact the sacrifice itself may be carried out as a conscious act these reasons remain often unconscious. As a general psychological framework this applies to the variations of the immediate cause of the sacrifice, the different objects that got destroyed peri-depositional and the regional or local varieties of the sacrificial act itself. Indigenous myths collected and analysed by C. Lévi-Strauss suggest, that the extraction of copper from the earth was experienced as a violent act.10 Something is ripped out of the nature. These myths are always related in certain ways to violence and humans becoming – at least temporally – at one with nature in order to bring the copper to the human culture, which is always something opposite to nature. Hence the initial obtaining of copper is quite frequently linked to violence, for example it comes through martyring a frog or it is Lady Wealth that tortures people with her copper fingernails. The wounds caused by this later provide the richness of copper (Leví-Strauss 1999, 121). In another example the woman who showed the copper to the man got raped, which causes the copper to disappear from the surface. Afterwards it had to be ripped out of the earth (Leví-Strauss 1999, 109 ff.). Violence induces the disappearance and the reappearance of copper. Nebelsick interprets the so-called scrap- or founder-hoards of the Late Bronze Age in the same way. He deduces that the destruction of the objects before they were deposited mirror the violent act of extracting the ore from the earth and in a second step the smelting of the ore in a raging fire to produce copper, which brings the metal into the cultural realm (Nebelsick 2000, 171). Here it can be added the latter happens through forcing them into a specific form and attaching something which is not naturally attached to metal. The latter is 10 Lévi-Strauss (1999, 53) provided a myth in which it is explained that one must know how ‘to kill the live copper’

in order to make it workable.

59

Christian Horn

10

11 Figure 10: Distribution of halberds deposited with the shaft detached Figure 11: Distribution of halberds deposited with the shaft attached 60

Deliberate destrucion of halberds

12 b

13 b

13 c Figure 12: a: Pie chart of the general find circumstance of European halberds. – b: Pie chart of the general conditions in which halberds were deposited Figure 13: a: Pie chart of the conditions of halberds deposited in hoards. – b: Pie chart of the conditions of halberds deposited in graves. – c Pie chart of the conditions of halberds from single finds

casualties on the own side or the healing of wounded warriors. There are of course again several other variations possible. The third reason – the execution of an object – is in respect of weapons especially interesting, because weapons as tools of death and destruction11 are certainly experi-

enced as a burden as is war as a whole (Keeley 1996, 144 ff.). Not so much because killing was seen as something especially irksome, in fact it is most likely something that happened on a regular and frequent basis.12 This does not exclude the interpretation of weapons having a prestige cha-

11 The numerous notches and other damages observable on halberds confirm that some of them were indeed used as weapons (pers. comm. R. O’Flaherty; see also O’Flaherty

2007, 423 ff.; and this volume; Horn 2010). 12 Keeley 1996, 68; Harding 2006, 511; Weinberger 2008, 69 ff.

61

Christian Horn

14

15 Figure 14: Distribution of the two treatments of halberds deposited in hoards Figure 15: Distribution of the two treatments of halberds deposited in graves 62

Deliberate destrucion of halberds

racter or being an heirloom. A weapon can be all this at the same time, but the violence somebody can carry out with a weapon provides the background for its characteristics. They are symbols of status because one can kill with them (Keeley 1996, 50). However, the spirits of the dead are always feared (Freud 1999b, 50. 66). This is due to the fact that the hostility against the recently deceased is projected (Freud 1999b, 79). Additionally it is expected that the dead have the power to drag the living into their realm, hence they turn into vengeful demons.13 Against these the living have to protect themselves with a large variety of rules (Freud 1999b, 66 ff.). In this respect it can be assumed that the killed enemy warriors were feared even more, because the hostility is even more visible and they have a special eligibility to retaliate upon the warrior for killing them (Freud 1999b, 75). Hence the killing act has to be projected into something, to keep the revenge of the demon away from oneself. Then it can be hypothesised it was projected into the weapon. So the weapon becomes a constant reminder and evidence of what happened. As an act of balancing, the object that brings death has to die itself. That the projection took place and the weapons were seen as perpetrators of the killing and destruction enabled philosophers of several periods to come to the revolutionary conclusion that it is not the weapon that kills, but the human who wielded it.14 This theory can only be made, when it is thought otherwise by a majority. However this phenomenon happens because the ‘destruction drive’ is centred onto the object (Freud 1999c, 480). In respect of the killed enemies Freud analysed that water is seen as separation from the dead (Freud 1999b, 75). It was noticed that native warriors had to undergo certain rituals (‘taboo-customs’) after bloodshed, which included a cleansing act (Freud 1999b, 47 ff.). In cleansing acts in taboo-customs water or lustration is very common (Freud 1999b, 38. 51 ff.). When the above assumption that the killing was – at least partially – projected into the weapon is correct, these appears a strong link between weapons and water. This is an astonishing parallel to the watery context in which halberds and other weapons are frequently found. When it is assumed that

Figure 16: a: Burning of firearm in Nairobi. – b: Destruction of weapons in Honduras. – c: Weapon sheared by the MAG

not every weapon involved in a killing had to be sacrificed, then it is possibly enough to sacrifice one weapon, most likely of the one who leads the raid or fighting during which a killing of enemies occurred or of the weapons ‘that killed the most’. In addition it is suggested that a killing had to be severe enough in order to trigger a sacrifice. This would be a model which could easily explain the frequent appearance of halberds with detached handles in single depositions and halberd deposition in rivers or bogs.15

13 This fear does even appear directly in archaeological context for one example (the fear of revenants) see Trachsel 2005, 53ff. 14 ‘L’arme […] uccidon gli uomini molte volte, non per malizia di loro, ma di coloro che malvagiamente l’adoperano.‘ trans.: ‚Arms [ … ] by them are men not seldom maliciously slain, albeit the malice is not in them,

but in those that use them for a malicious purpose.‘ (Rolli 1725, 283.); ‚Quemadmodum gladius neminem occidit, occidentis telum est.‘ trans.:‚[ … ] a sword by itself does not slay ; it is merely the weapon used by the slayer.’ (Gummere 1920, 341.). 15 Which were most likely lake depositions; see i.e. the hoard from Melz (Röbel, Germany).

63

Christian Horn

It remains somewhat difficult to explain the deposition of not-destroyed objects in hoards and genuine single finds. They are by definition excluded from this framework. Possibly it was not applied this strictly, especially because it is an underlying structure and as such it is subconscious. Therefore, the immediate occurrence may vary. Another possibility can be drawn from Leví-Strauss observations. Certain groups feel the need to stand out against other groups. Hence they differ in small steps from neighbouring groups. One small step is simply the inversion. This could be one solution to this interpretative problem. However, in respect to the halberds it seems to be an influx from the treatment in graves, because halberds with attached handles and without any other deliberately caused damage almost only appear in areas where they also were put undamaged in graves. It was deduced that the undamaged deposited object in graves is seen as a ‘living thing’. As a fact of this it might be enough to bury it before it was ‘dead’. The closing outlook is in a way anachronistic, but it shows that even today the deliberate destruction of weapons is still practiced in a public display of violence against the object. In Kenya four hundred thousand weapons from rival tribal groups got burned, during an annual ceremony in which the signing of the ‘Nairobi Declaration on small arms’ was celebrated (fig. 16, a). 16 In Honduras five thousand weapons from criminals were torn apart via welding and sawing (fig. 16, b)17 and after the secession of the Kosovo several thousand weapons were collected and got publicly destroyed.18 These sacrifices are made as a plea. They cause a smaller misfortune, which is considered in modern times to be an economic one, in order to prevent a bigger one. The latter is in these cases the uprising of civil war and organized crime. But it is as well the execution of an object seen as burden. The most striking evidence is provided by the Mines Advisory Group (MAG) which destroys weapons in several African countries with hydraulic shears.19 The result shows an astonishing similarity to the material of the so-called scrap hoards of the Late Bronze Age (fig. 16, c).20 Hence just in this few examples it becomes apparent that the same psychological structure survived, but is of course

today expressed in a different way. In modern secularized times the religious aspect is of course pushed more or less to the back. As a final remark it should be strongly emphasised again that this is a hypothesis to stimulate the discussion about deposition habits in prehistoric Europe. It is therefore of course open to debate and critique.

16 The „Regional Centre on Small Arms“ provides a photo gallery in which pictures from several ceremonial destructions of weapons can be found from several African countries. http://www.recsasec.org/photos.htm. 17 http://cache.daylife.com/imageserve/

0fEE5KVaZi1Ll/610x.jpg. 18 www.dw-world.de/dw/article/0,,2731515,00.html. 19 http://www.maginternational.org/cwmd/. 20 http://www.maginternational.org/silo/images/ mags-approach-to-cwmd-3_446x205.jpg.

BIBLIOGRAPHY Bagolini, B. 1981. Il neolitico e l’età del Rame. Ricerca a Spilamberto - S. Cesario 1977–1980. Vignola, Cassa di Risparmio. Brandherm, D. 2003. Die Dolche und Stabdolche der Steinkupfer- und älteren Bronzezeit auf der Iberischen Halbinsel. Prähistorische Bronzefunde VI, 12. Stuttgart, Steiner. Coles, J. M. 1971. Scottish Early Bronze Age Metalwork. Proceedings of the Society of Antiquaries of Scotland 101, 20–110. Dolfini, A. 2004. La necropoli di Rinaldone (Montefiascone, Viterbo). Rituale funerario e dinamiche sociali di una comunità eneolitica in Italia centrale. Bullettino di Paletnologia Italiana 95, 127–278. Drescher, H. 1958. Der Überfangguß. Ein Beitrag zur vorgeschichtlichen Metallurgie. Mainz, Römisch-Germanisches Zentralmuseum. Freud, S. 1999a. Zur Psychopathologie des Alttagslebens. Über Vergessen, Versprechen, Vergreifen, Aberglaube und Irrtum. In A. Freud (ed.), Gesammelte Werke, Vol. 17, 5–324. Frankfurt am Main, Fischer Taschenbuch. Freud, S. 1999b. Totem und Tabu. Einige Übereinstimmungen im Seelenleben der Wilden und der Neurotiker. In A. Freud (ed.), Gesammelte Werke,Vol. 9, 3–207. Frankfurt am Main, Fischer Taschenbuch. Freud, S. 1999c. Das Unbehagen in der Kultur. In: A. Freud (ed.), Gesammelte Werke, Vol. 14, 419–506. Frankfurt am Main, Fischer Taschenbuch. Gedl, M. 1980. Die Dolche und Stabdolche in Polen. Prähistorische Bronzefunde VI, 4. München, C. H. Beck. Gummere, R. M. (ed.) 1920. Seneca: Ad Lucilium epistulae morales. Vol. II, 341. Cambridge, Harvard University.

64

Deliberate destrucion of halberds

Rassmann, K. and Schoknecht, U. 1997. Insignien der Macht. Die Stabdolche aus dem Depot von Melz II. In A. Hänsel and B. Hänsel (eds), Gaben an die Götter. Schätze der Bronzezeit Europas, 43–47. Berlin, Unze. Rolli, P. (ed.) 1725. Giovanni Boccaccio: Decamerone, 283. Londra. Schoknecht, U. 1972. Ein neuer Hortfund von Melz, Kr. Röbel, und die mecklenburgischen Stabdolche. Bodendenkmalpflege in Mecklenburg, Jahrbuch 1971, 233–253. Steiniger, D. 2005. L‘Énéolithique en Italie. In P. Ambert and J. Vaquer (eds), La première métallurgie en France et dans les pays limitrophes, 287–301. Paris, Société préhistorique française. Trachsel, M. 2005. Kriegergräber? Schwertbeigabe und Praktiken ritueller Bannung in Gräbern der frühen Eisenzeit. In R. Karl and J. Leskovar (eds), Interpretierte Eisenzeiten. Fallstudien, Methoden, Theorie, Folge 18, 53–82. Linz, Oberösterreichisches Landesmuseum. Vulpe, A. 1995. Der Schatz von Perşinari in Südrumänien. In A. Jockenhövel (ed.), Festschrift für Hermann Müller-Karpe zum 70. Geburtstag, 43–62. Bonn, Habelt. Walker, I. C. 1974. The counties of Nairnshire, Moray and Banffshire in the Bronze Age, Part II. Proceedings of the Society of Antiquaries of Scotland 104, 71–120. Engelhardt, B. and Wandling, W. 2007. Ein frühbronzezeitliches Waffendepot von Unterschöllnach. Das archäologische Jahr in Bayern 2006, 38–40. Weinberger, S. 2008. Warfare in the Austrian Weinviertel during the Early Bronze Age. Wien, Österreichischen Akademie der Wissenschaften. Zich, B. 1996. Studien zur regionalen und chronologischen Gliederung der nördlichen Aunjetitzer Kultur. Berlin, Walter de Gruyter.

Harbison, P. 1969. The Daggers and the Halberds of the Early Bronze Age in Ireland. Prähistorische Bronzefunde VI, 1. München, C. H. Beck. Harding, A. 2006. What does the context of deposition and frequency of Bronze Age weaponry tell us about the function of weapons? In T. Otto, H. Thrane and H. Vandkilde (eds), Warfare and Society. Archaeological and social anthropological perspectives, 505–513. Aarhus, Aarhus University. Horn, C. 2010. Studien zu den europäischen Stabdolchen. Unpublished PhD thesis, RuhrUniversität Bochum. Keeley, L. H. 1996. War before civilization. The myth of the peaceful savage. Oxford, Oxford University. Krause, R. 2003. Studien zur kupfer- und frühbronzezeitlichen Metallurgie zwischen Karpatenbecken und Ostsee. Rahden, Marie Leidorf. Lévi-Strauss, C. 1999. The way of the masks. Seattle, University of Washington. Lull, V. 1983. La “Cultura” de El Argar. Madrid, Akal. Nebelsick, L. 1997. Auf Biegen und Brechen. Ekstatische Elemente bronzezeitlicher Materialopfer – ein Deutungsversuch. In A. Hänsel and B. Hänsel (eds), Gaben an die Götter. Schätze der Bronzezeit Europas, 35– 41. Berlin, Unze. Nebelsick, L. 2000. Rent asunder. Ritual violence in Late Bronze Age hoards. In C. F. E. Pare (ed.), Metals make the world go round. The supply and circulation of metals in Bronze Age Europe, 160–175. Oxford, Oxford University. Ó’Ríordáin, S. P. 1937. The halberd in Bronze Age Europe. A study in prehistoric origins, Evolution, Distribution and Chronology. Archaeologia 86, 196–317. Parker Pearson, M. 2009. The archaeology of death and burial. Stroud, The History.

Christian Horn Göteborgs Universitet Historiska studier Box 200 405 30 Göteborg Sweden

65

66

USE-WEAR ANALYSIS AND USE-PATTERNS OF BRONZE AGE SWORDS Barry Molloy

ABSTRACT

moved on from such romantic visions and strive to understand the reality of day-to-day life in prehistory, be it subsistence, ritual or violence related (Bradley 2005). A renewed interest in the study of war that began in anthropology in the early 1980’s and took hold in archaeology by the middle of that decade (see Thorpe 2005; Harding 2007, 15 f.). This rejuvenation of research saved the weapons of our museum display cases from being intellectually resigned to out-moded visions of prehistory and has on the contrary seen many new research projects develop, as exemplified in this present volume. R. Bradley’s call (2005, 145) for archaeologists to move beyond compiling detailed records of bronze artefacts and return them to the sphere of social interpretation is increasingly being met through this new work. It is within this context of reconceptualising the study of Bronze Age weapons that this paper redresses some misleading factoids and methodological conservatism that has developed over the course of a century of research and offers new directions in deriving social meaning from the material culture of war. To begin, we may take an expanded chaîne opératoire model that considers the entire lifecycle of weapons in biographical terms (Kopytoff 2000). This should take account of the conception, birth, life, death, burial and archaeological recovery of weapons. Through this model, we can explain reasons why weapons were created in the manner and form that they were, how this was done, what happened to them when they were being used by prehistoric persons, why they were removed from circulation, how were they interred, and eventually, how were they recovered. First, however, a brief characterisation of the ‘intellectual biography’ of bronze weapon research is required to contextualise the current state of our knowledge and research practice.

The study of bronze weaponry is older than the field of archaeology and modern research has inherited both benefits and problems associated with this chronological breadth of research. Bronze weapons occur in relatively similar forms throughout Europe, making them one of the few categories of artefact to receive similar academic treatment on such a wide scale, and in various academic traditions. This paper addresses terminological and methodological complications that have arisen in no small part due to the natural application of unifying language and functional interpretations that such a broad scale of research history has attracted. It is argued that such innocuous things as generic names for sword types or broad statements on spear use have a more profound impact than may be expected in determining research methodologies and results. Some paths towards creating a methodological consensus, or complementary strands thereof, are suggested and potential impacts of related changes are considered. A final purpose is to suggest methods to create greater synthesis between use-wear, taxonomic, experimental and archaeometric analyses. KEYWORDS Rapier – cut-and-thrust – slashing – stabbing swords – experimental – archaeology INTRODUCTION Walking into the National Museums of most countries in Europe, one may be forgiven for thinking the Bronze Age to have been one of the bloodiest epochs of our past. Display cases are laden with swords, spears, axes, lances, war-horns, shields, armour, helmets and all the fossils of past battles. To the modern archaeologist, this lure of bronze is perhaps better seen as a fossil of our antiquarian forbearers, drawn by mythology to a vision of the nasty, brutish and short nature of prehistoric life (Parker-Pearson 2005; Vandkilde 2006a, 419). As a discipline, we have long-since

HISTORY OF RESEARCH The recovery and publication of Bronze Age weapons is far older than the discipline of archaeology, with the first records dating back to early antiquarianism (Colquhoun and Burgess 1988, 5 f.) during the Enlightenment, a time when secular man was seeking to make sense of our origins and character. C. J. Thomsen’s division of objects 67

Barry Molloy

new field of archaeology, and studies of bronze weapons grew in numbers and complexity (e.g. Coffey 1894; Naue 1903). In this milieu, W. P. Brewis (1923) stands out as having offered a relatively unique study that addressed differences in modes of use alongside categorisations used to create an archaeological lingua franca for weapons. By the time weapon studies became the subject of more systematic analysis in culture historical and ‘new archaeology’ frameworks in the 1950’s and 1960’s,1 the basic tools for this research had long been established and accepted. In cold scientific terms, a weapon was treated the same as any artefact would be, and so meticulous records of its metric attributes were created that allowed cross-referencing with similar objects. Through this, the evolution of object forms could be charted and dissemination of physical characteristics could be charted across wide distances. The Prähistorische Bronzefunde series has been instrumental in codifying this approach and bringing vast quantities of objects to publication. Building on this long-established heritage, further questions have begun to be asked in the past fifteen years that pertain to how and why weapons were made, why they changed over time, how were they traded/exchanged, how were they used to fight and what were their nonmartial functions in the societies that used them.2 New techniques of analysis have been developed, foremost being metallurgical, use-wear and experimental approaches, as well as synthesis between these.3 The above descriptive system of recording weapons may be regarded as a ‘typological approach’ while the latter may be seen as an ‘interpretative approach’. Synthesis between the two strands is increasingly popular, and draws benefit from the former providing raw material for discussion, while the latter investigates the social and technological strategies behind the metric variability charted.

1 2

3 4

5

6

Figure 1: Zones of regional difference in sword proportions in Ireland

into the three-age system (1820) built on a century of intellectual development, and was one of the first systematic steps for meaningfully segregating prehistoric material culture. The growth from reports of curiosities into the production of detailed catalogues occurred around the same time as C. Darwin’s (1859) The Origin of Species emerged, reflecting a new scientific inflection to the study of cultural objects (e.g. Wilde 1863; Evans 1881). In the following decades, this system of pigeon-holing objects into types and subtypes to make order from chaos solidified into a sub-field of social inquiry in itself. Bronze weapons are non-ferrous objects that were frequently deposited in anaerobic environments, and this made them primary targets of such research as they survived in good condition, in great quantities and carried an implicit air of excitement and wonder. From the late nineteenth century, archaeology was becoming more self-aware and formalised as a discipline, utilising its own distinct research tools alongside those borrowed from other disciplines. Artefact studies were a core aspect to this

THE LIFE-CYCLE OF A WEAPON Conception The first stage in the life of artefacts is one that all hold in common: the design or conception stage. Before creating an artefact, a craftsman

1 E.g. Cowen 1951; 1955; Catling: 1956; 1961; Sandars 1961; 1963; Coles 1962; Eogan 1965. 2 Peatfield 1999; Bridgford 1997; 2000; Bradley 2005; Molloy 2006.

3 E.g. Kristiansen 1984; 2002; Bridgford 1997; 2000; York 2002; Tselios 2004; O’ Flaherty 2002; Molloy 2006; Mödlinger 2011; Uckelmann 2006; various contributions, this volume.

68

Use-wear analysis and use-patterns of Bronze Age swords

sign represent real-world variation in the combat requirements of the warriors that used them in battle; requirements that were (reflexively) voiced to smiths at the outset of the chaîne opératoire of production. This highlights the need to address why specific weapons or groups of weapons were made in the manner they were, in order to effectively employ data relating to how they were made. Changes in these functional requirements can furthermore be seen as a major driving force behind the technological evolution and refinement of bronze-craft traditions.

establishes the criteria and parameters of its design in relation to its intended function and existing technological and aesthetic trends. This may or may not include consultation with the intended end-user, but it involves instilling personalised variations on an established theme that will dictate both the functional and aesthetic qualities of the finished product. If we look at the example of Irish Bronze Age spearheads, the breadth of their size and typological range (Ramsey 1989) demonstrate that craftsmen producing them made wildly varying responses to design considerations. The technological choices in weapon design reflexively link the work of the craftsman and the requirements of the combatant, and the two would meet at this design stage in particular.4 In order to understand weapons as weapons, and not simply generic ‘artefacts’, it is necessary to characterise design elements that result from decisions made prior to production. In typological analyses, morphological consideration focuses primarily on hilt forms at the expense of the business end of the weapon – the blade (Peatfield 1999). Irrespective of hilt configuration, lengths and weights of weapons are related to both regional and local patterns, and relate to variations in the functionality of weapons and, by extension, combat systems. In Ireland for example, where so many bronze weapons survive (Waddell 2000; Molloy 2006), there are distinct regional preferences in the way that weapons were made and used (fig. 1–3). There is a broad band from the east to the west coast in the centre of the island where light and short swords were preferred. These could be deployed rapidly and required close proximity to a potential target to make a strike. In the northeast, the preference was for longer and heavier swords that had longer reach and greater impact force, but were slower to deploy. On a more micro-scale, the Type B swords from the two Grave Circles at Mycenae in Greece have typological homogeneity in hilt form, but in combat terms there is a graduation in length steadily from ca. 300 mm to nearly 700 mm (fig. 4; Fortenberry 1990). It can further be noted that these swords were very much a local tradition, and are rare outside of the territory of Mycenae itself. While enshrined during the production phase of a weapon, these differences in weapon de-

Several factors affect the quality of a weapon, most of which are encoded into it at the time of manufacture. The alloy chosen is of particular importance, though as P. Northover (abstract this volume) notes, there was no specific alloy unique to weapons. By the time tin-bronzes developed, in many areas of Europe the typical content was around 8–12 % tin.5 In Atlantic Europe, some weapons had a deliberate inclusion of lead (Northover 1988) which lowered the viscosity of molten bronze to ensure better filling of the mould and better surface quality as a result, but it lowered the toughness/durability of the alloy. Closely allied to alloy choice was the quality of mould preparation. For ceramic bi-valve moulds in particular, the smoothness of the interior surfaces was of fundamental importance. While this may have been perfected prior to preheating and pouring, a good mould-maker needed to ensure that degradation of the fine interior surface did not occur in the moments surrounding the pour, as experiments have shown that this severely damages the quality of the finished object by increasing porosity and causing larger pitting. Likewise excess lime in the clay used for the ceramic mould can cause gasses to generate during the pour resulting in internal and surface bubbles/flaws (J. Zuiderwijk, pers. comm.). Following the pour, the post-casting treatment of a weapon involves removing the flashing and polishing the surface of the object, repairing imperfections when possible. The most important post-casting step, however, is hardening and annealing (or in some cases tempering) of weapons.

4 This need not be weapons being ‘made to order’, but rather represents a feedback process between those who create weapons in a workshop and those who used them in battle.

5 E.g. Eogan 1965; Northover 1988; Ó Faoláin and Northover 1998; Mangou and Ioannou 1999; Mödlinger 2011.

Birth

69

Barry Molloy

100

90

80

70

60

400 ! 475mm

50

476 ! 550mm 551 ! 625mm

40

626mm> 30

20

10

0 1

2

3

4

5

6

Figure 2: Regional variations in length of swords in the 6 zones in Ireland marked in figure 1

100

90

80

70

60 grams

30

20

10

0 1

2

3

4

5

6

Figure 3: Regional variations in weight of swords in the 6 zones in Ireland marked in figure 1 70

Use-wear analysis and use-patterns of Bronze Age swords

Figure 4: Lengths of Type B swords from Greece

but not soft core to the blade. This was functionally similar7 to case-hardening of iron swords as practiced from Roman times (Bishop and Coulston 1993; Sim 2002). In terms of archaeometallurgical research, this reveals a shortcoming in the preponderance of samples taken from the edges of the weapon.8 The hardness values and microstructures revealed create a bias towards one component in the weapon’s design at the expense of others, a point recognised and beginning to be redressed through the use of newer non-destructive technologies (Mödlinger, this volume; Godfrey et. al. abstract this volume). While measurement of toughness on archaeological samples is prohibitively destructive, interpretation of the hardness values needs to be offset against consideration of toughness if this strand of scientific research is to be translated coherently into social analysis, in this case weapon design and functionality. In the case of sheet metal objects, shaping the material required hammering, which increased

Simply put, a hard blade-edge is required to maximise the cutting potential of a weapon. In theory, the harder an edge is, the better it will cut. With bronze, however, increasing hardness commensurately decreases toughness, which (for the purposes of this discussion) is the ability of a material to resist fracture through impact stress. Hardening of copper-alloys was achieved through hammering along the blade edge only, thus compressing the bronze and deforming the microstructure.6 By reheating the blade, re-crystallisation reduces stress and tempers it so that adequate functional toughness is retained. In this process, the body of the blade would not be hardened, as retaining higher toughness here allowed a degree of stress-absorption and plastic flexibility. The balance which the smith had therefore to reach was a blade edge that would be hard enough to cut effectively while retaining a degree of toughness to offset risk of fracture or chipping along the edge and at the same time maintaining a tough 6

while the edges are hardened. 8 This bias is generally created by museum policies in relation to permissible sampling procedures, rather than research agendas by archaeometallurgists.

Allen et al. 1970; Coghlan 1975; Higgins 1983; Bridgford 2000. 7 It was the result of a very different metallurgical technique, though in both cases a tough core is produced

71

Barry Molloy

from being held in the hand. This has important implications for how the weapon served to embody identity, but reflexively, how weaponry could receive symbolic potency through a symbiotic relationship with a person who also used other means to characterise martial components to their identity (Treherne 1995; Fowler 2004; Molloy 2006). In terms of how traces of use reflect their raison d’être in the form of combat damage, much has been written on this subject in recent years.9 This evidence of wear occurs in the form of combat scars and subsequent repairs. Damage concerns the patterns of action that placed a particular object into direct physical opposition to other objects, typically in the form of impacts. The relationship between design, functionality and evidence of use informs us on the active life of the object and how it played a very real role in shaping society. Emphasis to date in usewear studies has more often been placed on the death and burial stage of a weapon’s life-cycle. Accounting for use-wear, however, is as important as counting it. Recording that it exists therefore still calls for appreciation of its cause and character if it is to reveal the story of the ‘life’ of a weapon as well as its character in ‘death’, as will be elaborated on below.

hardness, though this could be tempered or annealed. As with bladed-weapons, hardness increased rigidity, and also increased resistance to cutting, but excess hardness would cause the metal to break or split through impact force (rather than cutting in strict terms), and so a balance had also to be met in terms of alloy chosen and hardness level instilled. The final step for bladed-weapons was to haft the blade, and this had considerable importance for functionality. In the case of spears or halberds, the surviving bronze component reflects just one element of a weapon. The length, thickness and weight of the wooden shaft would have heavily influenced how the weapon could be used. For example, a short shaft would transform a spear into a sword-spear akin to a Zulu (iklwa) Assegai. A long and heavy shaft would make a ‘spear’ functionally comparable to a medieval polearm whereby the wooden component constituted the majority of the weapon in terms of combat practice as well as proportions (Anglo 2000). For swords or daggers, some ornate types with gold sheeted hilts (Kilian-Dirlmeier 1993) were non-utilitarian whereas others with bronze handles required organic components to allow them to be gripped (Kristiansen 2002; Molloy 2010). In the case of Aegean swords in particular, it can be demonstrated that the varying forms of hilt design, as well as blade form, imposed widely varying modes of use (Molloy 2008). We need further note that superficially similar objects need not have been manufactured in the same manner, so differences in alloys or finishing techniques (such as hardening) had ramifications for how an object would perform in battle.

Death As with the warriors that bore them, there was no typical ‘death’ of a weapon. Some never made it to completion, others were removed from circulation with no evidence of having ever been used, some were used and repaired on several occasions before they met their end, others were used lightly and retired while others still were used to their utter destruction in combat or the smithy.10 Weapons shed their practical functionality when they are removed from circulation, though they retain their intrinsic symbolic and metallic worth. The decision in most European Bronze Age societies to remove a weapon from circulation would have usually been due to mechanical factors such as damage beyond repair. Occasional finds of ‘founders hoards’ are thought to be the raw materials of smiths buried under specific circumstances, be they temporary measures in times of trouble or ritual offerings (Bradley 1990). Objects that we have recovered that had

Life The period of time that a weapon was the possession of a warrior was the greatest duration it materially interacted with society. Its creation lasted days or weeks, the event of its interment perhaps merely hours, but its life would in many cases have lasted for years, even decades. K. Kristiansen (2002) and S. Bridgford (1997) have highlighted that traces of wear and tear on a blade can come from a variety of contexts of use. In particular some differentially worn-hilts on solid-hilt swords indicate that they received more wear from being worn at the waist than 9 Bridgford 2000; O’Flaherty 2002; Kristiansen 2002; York 2002; Molloy 2006; Mödlinger 2011; Matthews this volume.

10

2005.

72

Bradley 1990; Bridgford 1997; York 2002; Fontijn

Use-wear analysis and use-patterns of Bronze Age swords

Along with interpretative biases derived from our methodological tradition, there are some

terminological issues13 that confound systematic treatment of the evidence. The academic framework for analysis is currently undergoing substantial remodelling; yet, superficially harmless elements in the language used to describe weapons inhibit systematic development of this field. W. Wilde in 1862 introduced the term ‘rapier’ and W. P. Brewis (1923) the term ‘cutand-thrust’ sword, both of which are peppered throughout the literature along with references to apparently self-explanatory phrases such as ‘thrusting swords’ or ‘slashing swords‘. Shafted weapons of any proportion are grouped together as spears (Molloy 2006), and shields incorporate everything from small-diameter bronze bucklers to broad wooden shields (Molloy 2009). We also feel a need to create a division between daggers and swords when the only identifiable difference in an artefact group may be length, so that the material itself often resists partition; e.g. Aegean Type B or C daggers and swords (Kilian-Dirlmeier 1993; Papadopoulos 1998). In Atlantic Europe, establishing a difference in the so-called daggers, dirks and rapiers (Burgess and Gerloff 1981) has been the product of trying to force material to fit an inappropriately borrowed terminological framework. Replicating such terminology without qualification is problematic because segregation of objects can occur even where this bears no real cultural significance; the lack of recognisable boundaries today may often be the result of a lack of perceived boundaries in antiquity. As argued in Molloy (2007), the use of objects such as the Lissane Rapier (Burgess and Gerloff 1981), Type Sauerbrunn swords (Schauer 1971) or Mycenaean Type A swords (Kilian-Dirlmeier 1993; Molloy 2010) as archetypal early forms of sword has led to a common perception that early swords were for thrusting attacks. These indeed are typically called thrusting-swords or rapiers in the literature. A perceived next generation of swords (disregarding protracted periods of transition; Molloy 2007; 2010) are the broad families of grip-tongue and solid-hilted swords that occur across much of Europe. These receive equally unqualified generic names such as ‘slashing swords’, ‘cut-and-thrust’ swords or worse still

11 See for example Bradley 1990; Bridgford 1997; York 2002; Fontijn 2005; Harding 2006; 2007. 12 Many are recovered by archaeologists involved either in monitoring extracted materials or through rescue

excavation. 13 In English language literature in particular. 14 E.g. Burton 1884; Oakeshott 1960; Amberger 1998; Clements 1998; 1999; 2007; Wagner and Hand 2003.

been prepared for recycling are exceptional because they survive and do not reflect the typical death of a weapon, the countless examples that were either not buried or were recovered in antiquity of course leave no trace. We can postulate that the majority of bronze was recycled and remained in circulation, perhaps beyond the Bronze Age itself. The majority of weapons that made it to the burial stage of a lifecycle ended their functional life due to social, not mechanical, factors. The material available for us to study today is thus poorly representative of how the majority of weapons in a society reached the end of their lifecycle. Burial In-depth analysis of the death and burial of objects goes beyond the scope of this paper.11 It should be clear from the above discussion, however, that the patterns we find in depositional practice reflect social practices surrounding the symbolic or economic roles of martial objects. They are less applicable in understanding the longer period when weapons were active participants in negotiating social relationships through combat and personal display. Archaeological recovery Archaeological recovery of objects is the final factor that influences the character of our dataset, and it is simply noted here that a majority of finds in Europe have been discovered as the result of non-archaeological work programs such as chance finds, river dredging, peat-extraction or infrastructural development.12 This in turn presents a randomised character to our datasets that we must deal with systematically. SKELETONS IN OUR CLOSET (OR, TERMINOLOGICAL CONSERVATISM THAT WON’T GO AWAY)

73

Barry Molloy

DESIGN AND DAMAGE

‘true swords’ (Heath 2009, 99). Reference must be made to the broad corpus of literature that deals with sword design and use in other periods or places, from a martial art perspective, as it demonstrates that our adherence to archaic archaeological terms is a genuine methodological problem.14 The common perception of the evolution of bronze sword forms is that they went from stabbing to slashing designs, a framework so facile that it actually defies application of martial arts theory (Clements 2007). Too many of our ‘stabbing swords’ have well-defined cutting edges, and too many of our ‘slashing swords’ are well-balanced for effective point attacks for these categorisations to be considered valid in any real sense. This goes beyond semantic issues because the unfounded view that ‘thrusting’ swords were used for duelling is still widely held (Heath 2009, 98 f.). This is based in part on unhelpful comparisons to renaissance society15 and also frequent reference to ‘duelling scenes’ from the Aegean, despite the fact that scenes with long-swords exclusively depict swordsmen fighting spearmen (Papadopoulos 2006; Molloy 2010). De-loading our terminology is the most systematic way of proceeding, along with an acceptance of blurred boundaries between categories, particularly daggers and short swords. There are already terms in existence that can be applied more comprehensively so that we may look at grip-plate swords or Naue ii swords in place of dirks, rapiers or cut-and-thrust swords. For spears, the distinction is not so easy and needs further research to create acceptable ways of distinguishing a heavy, single-handed stabbing or throwing spear from a heavy polearm suited to cutting and stabbing attacks. For shields, the typological categories are restricted enough (Coles 1962; Uckelmann, in prep.) to allow differentiation of modes of use based on size and fabric (Molloy 2009). Terminological change therefore does not so much require a paradigm shift as it does a more sensible application of non-functionally loaded terms from the existing typological lexicon. Change must be made in our specialist archaeological literature which may then filter into more general studies.

The example of swords is taken in this section because the majority of the original weapon survives, due to their minimal organic components. With the typological diversity of these weapons, any generalising statement is problematic, but we can broadly say that the majority were between 350 and 800 mm in length.16 A modal weight range between 300 and 800 grams has broad applicability, and the vast majority of swords can be described as having two clearly defined sharp edges and an acutely angled point. Many swords fall outside of this metric range and the breadth of these parameters is in itself very coarse, but it serves to give a general impression of the majority of weapons from this period in relation to swords from later periods. In terms of the wider history of swords, most Bronze Age types can be considered to be short, light and manoeuvrable weapons that were multipurpose in terms of potential attacks – suited to various forms of cutting and thrusting. When we take specific categories, such as a French Carp’s Tongue or an Aegean Type Fii sword, these parameters can be significantly tightened and discussion made more specific. The purpose in this section, however, is to create the general context into which these specific studies can be best located. To this end, we may note that even the longest and heaviest bronze swords were generally within the parameters of a Roman Gladius (Bishop and Coulston 1993), the archetypal short sword, rather than a medieval long-sword (Clements 1998; 2007). The point of balance on a sword is an issue of considerable importance in relation to handling, and by extension modes of use. However, for a great many bronze swords, their light weight and short length reduces the importance of balance points as they do not gain significant momentum or kinetic energy during a cutting attack. Most of these swords require elliptical cutting arcs that draw the blade along a target, as opposed to striking down onto it with force, so that there is minimal commitment of balance of both weapon and user in making a strike. For heavier, longer examples the issue of balance becomes more important, but with the relative inflexibility of bronze, cutting attacks retain this

15

16

This is discussed at length in Molloy 2007. Here it is noted that renaissance rapiers had nothing whatsoever functionally in common with those bronze swords termed rapiers, apart from their length. There is furthermore no basis for comparing the social context of combats in these two societies.

For grip-plate swords, ca. 60 mm must be added to the length to account for the missing organic handle. This length is derived from the surviving wooden handle on the sword from Shower, Co. Tipperary, Ireland. Burgess and Gerloff 1991, 116; Molloy 2006, no. 8; National Museum no. 1904: 5604.

74

Use-wear analysis and use-patterns of Bronze Age swords

as Atlantic Group III Class Lissane grip-plate swords or some early Aegean swords were used differently (Molloy 2007; 2010), though they represent a small percentage of the corpus of European bronze swords. The effectiveness of cutting attacks with a sword is increased when the blade strikes at an area called the centre-of-percussion, a part of the blade where the harmonics are such that maximum force is transferred to the target. This is typically ca. two-thirds the way from shoulders to tip on bronze blades. As cutting attacks were rarely robust strikes, we would not expect frequent heavy damage on blade edges20 and might expect proportionately more damage to be concentrated around the centre-of-percussion. Damage is inflicted either when a strike is intercepted by another object or from making contact with an intended target. In the quarter of the blade closest to the hilt, damage may be inflicted when a potential incoming attack is blocked or stifled by intercepting the attacking blade with this area prior to it gaining full momentum (fig. 6). Parries or re-direction of an attack would make use of the trajectory and relative motion of both blades on interception and thus result in minor damage to any location along the blade. All of the above activities would result in a range of types of damage to the blade. V-shaped nicks occur as the result of a brief point of contact when blades impacted on each other and were retrieved/rebounded almost instantaneously.21 Burred nicks, notches and u-shaped notches were usually the result of twisting movement of either or both blades immediately following impact. Shallow nocks can occur from a variety of impacts and defy hard and fast explanation. Bowing of the blade edge occurs when an attack strikes a spear shaft, the flat of another blade or bone at an

bias towards incising rather than percussive cuts (see Amberger 1999). The difference in sword designs therefore most often relates to the balance between incision and percussion in cuts, rather than a bias towards either cutting or thrusting. All Bronze Age swords (except the rare bronze ‘scimitars’ from Scandinavia) were suited to thrusting attacks. Along with balance, the toughness and flexibility of a sword will dictate the forces that can be applied in making attacks. For example a very soft sword will bend when impacting a target whereas a very hard sword may snap at the point of impact or the leverage point close to the hand.17 Bronze swords have plastic rather than elastic deformation properties, which means that they will bend but not naturally spring back to their original form. Their relative lack of toughness18 also means that when bent, they can only be straightened once or twice without the aid of heat treatment before fatigue and internal flaws cause a fracture. The hardness value of the edge of bronze swords was typically 100–200 HV (Bridgford 2000; Mödlinger 2011) and consequently they would not retain a very sharp angled cutting-edge,19 and this is why they typically have a wide edge-angle to best retain a level of moderate sharpness. A concise general statement may therefore be that bronze swords were short, light, unforgiving in terms of flexibility and could not hold a very sharp edge. They had neither the weight nor length required to make broad, sweeping cutting attacks and they were likewise restricted from making heavy-impact cutting/cleaving strokes. In general terms they were suited to tighter cuttingarcs and/or making cuts using a drawing motion and thrusting attacks could easily be performed (Amerger 1998; Molloy 2010). Exceptions such 17

which include: Nocks – shallow and narrow ‘dents’ in a blade edge, often burred parallel to the blade edge; Vshaped nicks – sharply defined impact creating a wedge shaped impact mark commensurate with the profile of an attacking blade-edge; Notches – impact marks whereby a portion of metal has been chipped or broken away from the blade on impact; Burred nicks or notches – a sliver of metal is broken from its original position creating a gap in the blade, but this is attached still on the edge of the impact mark at ca. 90 degrees to the blade edge; Rippling – this is where the edge forms a wavy pattern so that it is no longer a straight line; Bowing – an impact literally folds a small section of the blade out of position, similar to a nock but proportionately wider; Rolling – when the blade edge folds back on itself at the very edge; Cut – when a thin section of the blade is cut away, typically leaving a bur at the end of the cut-line.

When a blade strikes a target the force of the impact is transferred both to the target and to the hand of the user. When the centre of percussion is effectively used, the force to the target is maximised and that to the hand minimised due to the natural harmonics of the metal. On less ideal strikes, more force is returned to the user and the natural point of leverage occurs typically in the area of the hilt. 18 Due to the mechanical impact testing required to measure toughness, archaeometallurgical research has had to focus solely on hardness with little reference to toughness, a significant methodological problem in terms of assessing combat worthiness of a weapon. 19 Determined by examining ancient pieces and through experimental work. 20 Except when deliberately inflicted such as ‘killing’ a weapon. 21 Forms of damage have been discussed by S. Bridgford (2000) and B. Molloy (2006) the latter system is used here

75

Barry Molloy

5

6 Figure 5: Edge damage on long and narrow Irish grip-plate sword (National Museum of Ireland, no. W104). Some taphonomic damage is evident, though the general coarse character to the edges appears to have been in this condition when the weapon was used: A – Shallow nicks and nocks; B – U-shaped notch; C – Shallow burred notch; D – Over-sharpened and rolled-edge Figure 6: Edge damage immediately on the lower blade on Irish grip-plate sword (National Museum of Ireland, no. P251A)

oblique angle (off-line from the trajectory of a clean cutting-line) so that the edge buckles instead of cuts. When parries result in a sliding motion, the relative lack of hardness can cause rippling or slight rolling of the blade edge. Occasional examples of a sharply defined u-shaped notch ca. 1 mm wide occur, which it is suggested may be the result of a light impact against the edge of an unrolled bronze shield. In order to investigate these different areas of damage on ancient swords, the author divided complete blades into three equal zones and recorded differential damage (Molloy 2006). The case-study of complete and near-complete Irish Bronze Age swords was used during the author’s PhD research.22 During choreogra-

phed test-sparring and test-cutting with fully sharpened replica swords, the above forms of damage were encountered and recorded. This allowed causation to be charted and compared against damage on ancient weapons. It was noted that the damage on Irish Bronze Age swords was in many cases frequent, though in very few cases it was severe. Damage along the blades was usually in the form of shallow nicks, and in only in exceptional cases was combat damage recorded along the flat of the blade, indicating that blocks and parries used the blade edges (fig. 5). This indicates that the users were conscious of the mechanical qualities of their weapons and used them accordingly.

22 During subsequent consideration of this data it was noted that the centre of percussion spanned the two zones on the point side of the blade, and so it is suggested that in future

research, a blade is better divided into four equal parts, the centre-of-percussion falling into the second quarter of the blade from the point.

76

Use-wear analysis and use-patterns of Bronze Age swords

indicative of varying skill levels of their original owners but could also be the result of the duration of use in single battle. The character of combat damage, as well as its form or frequency, has largely untapped potential to reveal more about ancient martial arts.

In combat, most forms of weapon would not exclusively meet similar types of weapon in exchanges of blows. When we focus on one particular category of weapon, such as swords, it is insufficient to consider their use solely in relation to their fellows; the sword must be placed in the multi-weapon martial milieu to which it belonged. Terms like ‘swordsmanship’, while far from redundant, must not dominate how we consider swords to have been used and certainly not exclusively reflect the weapons they faced and received damage from/inflicted damage upon. For example, the typical defensive weapon used in sword fighting appears to have been shields, whereby the majority of interceptions would have used this in favour of the sword (Molloy 2009). We therefore need not expect extensive edge damage on swords when used according to their combat expectations. The Hollywood myth of swords clanging against swords in combats is very much a piece of showmanship and does not reflect real combat practice, and so we should not be looking for damage on a sword to have been inflicted by another sword in all cases. Whether a swordsman had a shield or not, whether he23 faced another sword or a spear, we can be certain that he was conscious of the tradeoff between inflicting injury or damage on an opponent and causing damage to his own weapon (as well as exposure to personal injury). In cases where we get heavy damage to a sword, this was the result of it being used ‘inappropriately’. Given that battle was dynamic and that luck was as important as skill in many contexts, the risk of damage to a weapon needs to be considered as subordinated to risk of injury to the person and so heavy-handed use occurred when absolutely necessary. Through this the weapon may fail, and the user may escape impending attacks or be killed, but that in itself was the nature of combat. Furthermore, while warriors were aware of the mechanical limits of their weapons, protracted use could lead to degradation and breakage even in a single engagement. There was no ideal sword, spear or shield and the effectiveness of each was subject to the skills of the user and the context of use. When we look for damage on weaponry we can therefore characterise forms of damage broadly as those inflicted from ‘proper’ use of a weapon and those which were the result of less proper, perhaps desperate, use. Differences in damage may be

WEAPONS IN SOCIETY In order to consider how weapons were used in the larger scale of battles, the numbers that were in circulation must be considered. As discussed above, the way weapons enter the archaeological record was heavily linked to biographies, and the easy recyclability of bronze means that survival rate speaks of certain social practices, making it of little benefit in estimating original numbers in circulation. The most direct sources we have are from prehistoric Crete, where surviving archival records list swords. These Linear B records were inscribed on clay tablets, some small few of which survive due to their being partially ceramicised in fires that destroyed their storage rooms. While the resultant record is therefore very partial and selective, we can note the Ra Linear B tablets from Knossos include Ra 1540 that lists 50 swords and Ra 7498 that lists 18 and 99 swords; 20 fragmentary tablets list an uncertain number of swords (Driessen and Macdonald 1984, 64). Considering the 24 swords known archaeologically for the entire Bronze Age from this area, the implications are considerable. While this tells us little of who actually owned these swords, it does tell us that the facility for mass production and distribution of armaments was within the technological and social parameters of this society. Speculating along the lines that ‘it is unlikely that...swords were the normal weapon of foot soldiers, or “infantry”’ (Heath 2009, 100) is harmful in that it has no factual foundation. We cannot estimate the number of weapons that were in circulation, but through direct and circumstantial evidence (through technological argument), we know that weapon production was a well-developed industry that produced vastly more pieces than survive today. Using this technological approach, we can extend the argument derived from Crete to other areas of Europe. The creation of bronze weapons required the carving of a wooden template for bladed weapons and a wooden former for

23

The assumption that warriors were male and the relevance of the term warrior in itself are made here,

following the rationale of Harding (2006).

77

Barry Molloy

leather and bronze shields (Ó Faoláin 2004). Once created, the template can be re-used repeatedly to create multiple copies of similar swords or spears at the same time or in alternate cycles by creating clay moulds based on the former. For shields, sheets of bronze or leather can be repeatedly pressed into the same former. This would produce virtually identical objects that can be recognised archaeologically. From Ireland, three identical swords found in the same context at Ballycroghan and were made from the same template and bear evidence of various different steps in the production process (Jope 1953). From Djursland, Denmark (Jensen 2002, 73, after Kristiansen and Larsson 2005, 207), we have eight swords that were produced from the same mould, again not completely finished. From the site of Arkalochori in Crete (Marinatos 1935), nine swords occur, at least five of which were made from the same wooden original and yet again none of these were completed. These are not exceptional cases, but they are also absolutely not the rule. In truth, there are few if any known examples of swords or spears found in different contexts that irrefutably came from the same template. This may indicate that one template was more usually used for just one weapon, but the caches of unfinished weapons made from a single pattern denote this was unlikely and that separation of similar pieces occurred after weapons were completed in a workshop. The technological ease of mass production and Cretan texts further hint that very few of the weapons originally produced have made it into the archaeological record. One can also note that the hundreds of thousands of Greek Hoplite warriors historically attested along with the millions of Roman soldiers known from our sources have left fewer weapons than the Bronze Age inhabitants of Europe.24 In the case of shields, the dataset throughout Europe is too small to determine significant patterns, but finds such as the sixteen shields from Fröslunda in Sweden (Hagberg 1988) indicate that shields were not as rare in society as indicated by their survival rates. The choice of weapons for deposit in most areas was clearly biased towards the offensive rather than defensive ones. The imitation of elements of bronze body armour on gold ornaments in Ireland (Cahill 2005)

is noteworthy as various smiths clearly had ready access to this armour, yet not one piece of it survives archaeologically. The likelihood that we have recovered a mere fraction of the weapons originally in use has important implications for understanding how groups of combatants worked together to engage in warfare. It is not the purpose of this paper to speculate on forms of warfare or actual numbers involved, this has been the subject of many insightful analyses.25 Further consideration of population estimates based on settlement patterns may reveal more than the weaponry can in relation to each specific region. Of more immediate concern in this paper is understanding how variability in forms of weaponry might be used to illuminate modes of combat practice. Weapon availability may have fluctuated in any given region from times when they were freely available to those entitled to bear them, to times when they were of more restricted circulation (Kristiansen 2002). At all times in most regions there was a wide variety of weapons available, and even variety in the forms of any given category. For example, for the later Bronze Age in Central Europe, rod-tanged swords were in use alongside solid-hilted swords and grip-tongue swords; in Atlantic Europe, Carp’s Tongue swords were used alongside leafshaped grip-tongue forms; and in the Aegean Naue ii swords were in use alongside Type Fii, Gii and Dii. Looking at shields, in the latter region tower and figure-of-eight shields were used alongside each other and in northern Europe, shields varied in size from around 300–700 mm diameter and could be made from wood, leather, bronze or combinations of any. Spears, as mentioned above, ranged from throwing forms, to short-hafted sword-like types, to single-handed stabbing forms to two-handed pole-arms. Axes ranged from miniature tools to lethal weapons. One can also add archers and slingers to this mix. The battlefield was thus an environment where a host of different weapon forms would come into direct opposition in combat. Should this be conceived as a chaotic ensemble of mismatched warriors fighting in haphazard fury, or can it be seen to reflect a much more sophisticated martial system underpinning battles? Some brief observations from early literature can prove revealing, as while they do not provide a generic guide to prehistoric warfare, they

24 Bishop and Coulston 1993; Snodgrass 1999; Connolly 1998; Prähistorische Bronzefunde (multiple volumes).

25 Ferguson 1993; Osgood 1998; Bridgford 2000; Harding 2007; O’Brien 2009; Wylie 1995; Vandkilde 2006b, 515; Kristiansen and Larsson 2005, 229.

78

Use-wear analysis and use-patterns of Bronze Age swords

relate to peoples in the same geoclimatic regions utilising similarly complex weapon panoplies, and thus provide useful analogies.26 These indeed provide much more relevant datasets in terms of material culture and settlement patterns than the often found comparisons to tribal societies in far-flung reaches of the earth (Osgood 1998). In the Irish epic ‘The Táin’, we read of a host of weapon types and there is a recognition that different forms of spear had different names. We also repeatedly hear in these epics of specialised training regimes for youths who were to become warriors and the specialised skills that they learned are repeated throughout the text. On the training of Cuchulainn, we hear that ‘if he really wanted to learn heroic deeds, he must go to where Scáthach was teaching her two sons’ and that ‘Scáthach taught him brave deeds and craft of arms’ (Kinsella 1969, 31). We later read of the single-combat between Cuchulainn and Ferdia, where they repeatedly ask at each stage of the long battle ‘what weapons will we use today?’ (Kinsella 1969, 187–192), indicating wide variation in available weapons and related modes of use. In terms of comparing this to the Bronze Age, we can recognise that the material culture of war was equally rich. The physiological requirements to use weapons with due attention to their mechanical limitations requires skill-development through repetitive-task execution (Malafouris 2004; Molloy 2008). It also requires stress-inoculation to permit execution of these tasks while experiencing stress-induced physiological distortion of the senses that occurs in a combat environment (Molloy and Grossman 2007). These are all the product of a martial art training regime, not occasional fracas or beat-‘em-ups. Use-wear analysis (Molloy 2006) provides evidence for prehistoric weapons being used by skilled martial art practitioners, as the damage is in many cases commensurate with controlled application of force and trajectories of attack. Contemporary to the Bronze Age throughout much of Europe, the epic Odyssey was written in the Greek world and tells of the journeys of the hero Odysseus. In this work, we read about the protagonist deliberately masking his martial art skills, as they would reveal his knowledge of specialised training and thus his elite status.

‘Now both contenders put their hands up. Royal Odysseus pondered if he should hit him with all he had and drop the man dead on the spot, or only spar, with force enough to knock him down. Better that way he thought – a gentle blow, ‘else he might give himself away’ (The Odyssey xviii 80–120 [my emphasis]). It is interesting to note that from the Aegean Bronze Age we have comparatively abundant iconographic evidence of martial arts training and practice in the form of boxing as well as evidence of application of martial arts in deadly combats (Hiller 1999; Peatfield 2007). From Roman through to Medieval times, we repeatedly read of, and see in art, wooden training swords being used to hone the skills of warriors (Clements 1998). From Ireland and Britain in particular, a number of such wooden swords survive from prehistory (e.g. from Cappagh in Ireland and Groatsetter in the Orkneys), once again indicating that the development of combat skills was part of the expectation of those who would engage in combat. The Cappagh sword is broken, possibly through use and the handle of the Groatsetter sword is polished through repeated handling.27 Examples citing the use of wooden swords can be found widely when literary and artistic source material in Europe increase in quantity and quality.28 In general, it remains consistent throughout European history that at times of rich martial assemblages, there was a commensurately rich martial art tradition. Without recourse to details of frequency, duration, scale, intensity, strategy, tactics, logistics, differential armaments, causation or conflict resolution mechanisms, there is little we can say about the conduct of war or who combatants were. J. Whitley (2002) and S. K. Smith (2009) and others have argued that we equally cannot use the mortuary record to ascertain the status or identity of warriors in most areas of Europe. The often cited concept of a ‘warrior elite’ is thus at odds with the resolution we possess on social structures. There was differential wealth distribution in all known societies from Bronze Age Europe, and in societies that lack the rule-of-law, protection of acquired wealth required deterrents for would-be raiders/thieves from taking this wealth by force. The materialisation of this power, wealth and influence was typically through

26 Wylie 1995; Vandkilde 2006b, 515; Kristiansen and Larsson 2005, 229. 27 The lack of damage to the edges indicates that it was not used in heavy contact sparring, but it would have served well to practice forms (choreographed sequences of

attack). Its greater proportions than a typical sword can be accounted for by the requirement of it weighing the same as a bronze type. 28 E.g. several references in the Tain (Kinsella 2002, 104. 118. 119); Vegetius (Reeve 2003); Talhoffer 2000.

79

Barry Molloy

reference to the widely cited notion of thrusting and slashing swords, for example. A ‘bottom-up approach’ that builds from primary artefact research through to addressing how combinations of weapons work together in combat has potential to add new interpretative dynamics to a traditional field of archaeological research. The relationship between typological and interpretative approaches need not be divisive, as synthesis complements the objectives of both. Of particular importance is addressing cultural issues that shape the emergence of new artefact forms, particularly through the sphere of long-distance movement and interaction of warriors as well as craftsmen. Combat practitioners would favour functional attributes of various weapons and these ideas were brought together by craftsman in different regions, resulting in the sequential development of weapon forms through time. The biological overtones of conceptualising artefact change as ‘evolution’, whereby a new form derives from pre-existing ‘parent’ forms, risks prioritising aesthetic over functional elements in processes of change. While dominant aesthetics derived from craft traditions, functional changes relate to the requirements of warriors, and so we must bear in mind both of these characters when considering changes in forms. Use-wear analysis, archaeometallurgical research and experimental archaeology should thus inform typological research in the same fashion that typological research provides metric data on forms that can inform functional analyses. Similarly, metallurgical research must remain cognisant of why manufacturing processes were enacted in relation to the intended functions of a weapon, bringing ever greater unity to this specific field and affording it resonance in addressing social and technological issues on a wider scale. A biographical / chaîne opératoire model30 can further serve to coherently bring together the varying traditions of archaeological scholarship that look at these datasets – those concerned with metallurgy, typology, craft-traditions and trade/communication, non-martial roles such as display and deposition, use-wear analysis and finally combat applications. Moving beyond discussion of differences between thrusting and slashing swords, from functional versus ceremonial shields or from throwing to stabbing spears, we can begin to better appreciate the dynamic multidisciplinary quality of our data and

portable assets, be it a cow or a cauldron. Warriors may have been free men that commanded material benefits from authority structures, but we have no evidence that they had a ‘controlling stake’ in socio-political decision making. The term ‘warrior elite’ therefore remains unhelpful, and while we may replace it with ‘elite warriors’ in many cases, tying the two terms together seems unnecessary. Nonetheless, it would be an untenable paradox if those who controlled wealth were not those who controlled the power to physically protect it, and so we may consider a direct relationship between those who sanctioned violence and those who were sanctioned to enact it. This places warriors as key players within elite power systems even when they were not at the helm. The authority structure of Anglo-Saxon Britain is not entirely inappropriate as an analogy for many areas of Bronze Age Europe.29 In this case, we have warrior bands who pledged loyalty to a lord or king (Pollington 1996), who then negotiated social relationships within the group and with other social groups. All free men may have had the right to bear arms when they came of age, and slaves or bondsmen were maintained for undertaking mundane labour. Where is the evidence? There is none, yet we have clear evidence in the differential access to wealth that these were not societies of equals and that violence, as well as wealth, was a potential means to assert influence. Neither history nor anthropology can provide a single model (Keeley 1996) where a similar milieu was sustained for centuries without enforced social differentiation, and so we may present the above as fitting our material evidence for existence of different practitioners of wealth management, extreme violence and mundane farming. CONCLUSION – WEAPONS AS WEAPONS The study of Bronze Age weapons is one of the few areas in prehistoric archaeology where a panEuropean analysis has obvious resonance, mainly due to the functional and typological parity of weapon designs. The long history of research has led to the development of common functional interpretations that have worked through this panEuropean vision so that concepts developed, right or wrong, in one region can have a direct impact on research in another area. One can make direct 29

Other examples of similar social structures are common thoughout Europe, and this is thus one of many analogies. In particular, the settlement systems and hierarchies fit well for many prehistoric regions, as do the

cases of individuals characterised as warriors through their funerary assemblages. 30 Particularly through a Combat Archaeology framework (Molloy 2008).

80

Use-wear analysis and use-patterns of Bronze Age swords

better explore the life and times of Bronze Age weapons.

Cahill, M. 2005. Ornaments: Cuirass to Gorget? An Interpretation of the Structure and Decorative Elements of Some Gold Ornaments from the Irish Late Bronze Age. Archaeology Ireland 19. 4, 26–30. Catling, H. 1956. Bronze Cut-and-Thrust Swords in the Eastern Mediterranean. Proceedings of the Prehistoric Society 22, 102–125. Catling, H. 1961. A new sword from Cyprus. Antiquity 35, 115–123. Clements, J. 1998. Medieval swordsmanship. Bolder, Paladin Press. Clements, J. 1999. Renaissance swordsmanship. Bolder, Paladin Press. Clements, J. 2007. The Myth of Thrusting versus Cutting Swords. In B. P. C. Molloy (ed.), The Cutting Edge: Studies in Ancient and Medieval Combat, 168–76. Stroud, Tempus. Coffey, G. 1894. Notes on the classification of spearheads of the Bronze Age found in Ireland. Proceedings of the Royal Society of Antiquaries of Ireland 111, 486–510. Coles, J. M. 1962. European Bronze Age shields. Proceedings of the Prehistoric Society 28, 156–190. Connolly, P. 1998. Greece and Rome at War. London, Greenhill. Coghlan, H. H. 1975. Notes on the prehistoric metallurgy of copper and bronze in the Old World (2nd edition). Oxford, Pitt Rivers Museum. Cowen, J. D. 1951. The earliest swords in Britain and their origins on the continent of Europe. Proceedings of the Prehistoric Society 17, 195–213. Cowen, J. D. 1955. Eine Einführung in die Geschichte der bronzenen Griffzungenschwerter in Süddeutschland und den angrenzenden Gebieten. Bericht der Römisch-Germanischen Kommission 36, 53– 155. Darwin, C. 1998. The origin of species. Ware, Wordsworth Editions Limited [reprint of 1859]. Driessen, J. and Macdonald, C. 1984. Some military aspects of the Aegean in the late fifteenth and early fourteenth centuries BC. Annual of the British School of Athens 79, 49–75. Eogan, G. 1965. Catalogue of Irish Bronze Swords. Dublin, National Museum. Evans, J. 1881. The ancient bronze implements, weapons and ornaments of Great Britain and Ireland. London, Longmans, Green and Co. Ltd. Ferguson R. B. 1993. Explaining War. In J. Haas (ed.), The Anthropology of War. Cambridge, Cambridge University Press.

ACKNOWLEDGEMENTS First and foremost I would like to thank the organisers for bringing us all together for the conference and for inviting me to participate. I am grateful to M. Cahill and M. Lannin at the National Museum of Ireland, S. MacCartin in the Ulster Museum, J. Swaddling in the British Museum and M. Vickers at the Ashmolean Museum for advice, help and access to materials. My thanks to M. Milić for sensible and helpful comments on this paper. Thanks also to A. Pearfield for discussions on combat applications of weapons. This research was carried out under funding from the Irish Research Council for Humanities and Social Sciences and the Department of Social Welfare, Ireland. BIBLIOGRAPHY Allen, I. M., Britton, D. and Coghlan, H. H. 1970. Metallurgical reports on British and Irish implements and weapons in the Pitt Rivers Museum. Oxford, Pitt Rivers Museum. Amberger, J. C. 1998. The Secret History of the Sword. California, Multi-Media Books. Anglo, S. 2000. The Martial Arts of Renaissance Europe. London, Yale University Press. Bishop, M. C. and Coulston, J. C. N. 1993. Roman Military Equipment from the Punic Wars to the Fall of Rome. London, Oxbow. Bradley, R. 1990. The Passage of Arms. Cambridge, Cambridge University Press. Bradley, R. 2005. Ritual and domestic life in prehistoric Europe. London, Routledge. Brewis, W. P. 1923. The sword in Great Britain. Archaeologia 23, 253–265. Bridgford, S. 1997. Mightier than the pen? (An edgewise look at Irish Bronze Age swords). In J. Carman (ed.), Material Harm: Archaeological Studies of War and Violence, 95–115. Glasgow, Cruithne Press. Bridgford, S. 2000. Weapons, Warfare and Society in Britain: 1250–750 BC. Unpublished PhD thesis, University of Sheffield. Burgess, C. and Gerloff, S. 1981. The dirks and rapiers of Great Britain and Ireland. Prähistorische Bronzefunde IV, 7. München, C. H. Beck. Burton, R. F. 1997. The book of the sword. Ontario, Dover Publications [reprint from 1884]. 81

Barry Molloy

Fontijn, D. 2005. Giving up weapons. In M. Parker-Pearson and I. J. N. Thorpe (eds), Warfare, Violence and Slavery in Prehistory. British Archaeological Reports, International series 1375, 145–154. Oxford, BAR. Fortenberry, C. D. 1990. Elements of Mycenaean Warfare. Unpublished PhD thesis, University of Cincinnati. Fowler, C. 2004. The Archaeology of Personhood. London, Routledge. Hagberg, U. E. 1988. The bronze shields from Fröslunda near Lake Vanern, West Sweden. In B. Hardh, L. Larsson, D. Olausson and R. Petre (eds), Trade and exchange in prehistory. Studies in honour of Berta Stjernquist. Acta Archaeologica Lundensia, 8. 16. 119–126. Lund, Historiska Museum. Harding, A. F. 1999. Warfare: a defining characteristic of Bronze Age Europe? In J. Carman and A. F. Harding (eds), Ancient Warfare, 157–175. Stroud, Sutton. Harding, A. F. 2006. What does the context of deposition and frequency of Bronze Age weaponry tell us about the function of weapons. In T. Otto, H. Thrane and H. Vandkilde (eds), Warfare and Society: Archaeological and Social Anthropological Perspectives, 505– 514. Aarhus, Aarhus University Press. Harding, A. F. 2007. Warriors and weapons in Bronze Age Europe. Budapest, Archaeolingua. Heath, J. 2009. Warfare in Prehistoric Britain. Stroud, Amberley. Higgins, R. A. 1983. Engineering Metallurgy (5th edition). London, English University Press. Hiller, S. 1999. Scenes of warfare and combat in the arts of the Aegean Late Bronze Age. Reflections on typology and development. In R. Laffineur (ed.), Polemos: Le Contexte Guerrier en Égée á l’Âge du Bronze, 319–331. Liege, Aegeum. Jope, E. M. 1953. Three Late Bronze Age swords from Ballycroghan, near Bangor, Co. Down. Ulster Journal of Archaeology 16, 37–40. Keeley, L. 1996. War before Civilisation. Oxford, Oxford University Press Kilian-Dirlmeier, I. 1993. Die Schwerter in Griechenland (ausserhalb der Peloponnes), Bulgarien und Albanien. Prähistorische Bronzefunde IV, 12. München, C. H. Beck. Kinsella, T. 1969. The Tain. Oxford, Oxford Paperbacks. Kopytoff, I. 2000. The cultural biography of things. In J. Thomas (ed.), Interpretive archaeology, 377–393. Leicester, Leicester University Press.

Kristiansen, K. 1984. Krieger und Häuptlinge in der Bronzezeit Dänemarks. Ein Beitrag zur Geschichte des bronzezeitlichen Schwertes. Jahrbuch des Römisch-Germanischen Zentralmuseums 31, 187–208. Kristiansen, K. 2002. The tale of the sword – swords and swordfighters in Bronze Age Europe. Oxford Journal of Archaeology 21, 4, 319–332. Kristiansen, K. and Larsson, T. B. 2005. The rise of Bronze Age society. Cambridge, Cambridge University Press. Malafouris, L. 2004. The cognitive basis of material engagement: where brain, body and culture conflate. In E. DeMarrais, C. Gosden and C. Renfrew (eds), Rethinking Materiality: The Engagement of Mind with the Material World, 53–62. Cambridge, McDonald Institute Monographs. Mangou, H. and Ioannou, P. V. 1999. On the chemical composition of prehistoric Greek copper-based artefacts from Mainland Greece. Annual of the British School at Athens 94, 81–101. Marinatos, S. 1935. Ausgrabungen und Funde auf Kreta, 1934–1935. Archäologischer Anzeiger 50, 159–244. Molloy, B. P. C. 2006. The role of combat weaponry in Bronze Age societies: The cases of the Aegean and Ireland in the Middle and Late Bronze Age. Unpublished PhD thesis, University College Dublin. Molloy, B. P. C. 2007. What’s the bloody point?: Fighting with Irish Bronze Age weapons. In B. P. C. Molloy (ed.), The Cutting Edge: Studies in Ancient and Medieval Combat, 90–112. Stroud, Tempus. Molloy, B. P. C, 2008. Martial Arts and Materiality: A Combat Archaeology perspective on Aegean swords of the fifteenth and fourteenth centuries BC. World Archaeology 40 (1), 116–134. Molloy, B. P. C. 2009. For Gods or Men? A reappraisal of the function of European Bronze Age shields. Antiquity 83, 1052–1064. Molloy, B. P. C. 2010. Swords and swordsmanship in the Aegean Bronze Age. American Journal of Archaeology 114. 3, TBC, 403–428. Molloy, B. P. C. and Grossman, D. 2007. Why can’t Johnny kill? The psychology and physiology of interpersonal combat. In B. P. C. Molloy (ed.), The Cutting Edge: Studies in Ancient and Medieval Combat, 188–202. Stroud, Tempus. Mödlinger, M. 2011. Herstellung und Verwendung bronzezeitlicher Schwerter Mitteleuropas. 82

Use-wear analysis and use-patterns of Bronze Age swords

Eine vertiefende Studie zur mittelbronzeund urnenfelderzeitlichen Bewaffnung und Sozialstruktur, Universitätsforschungen zur prähistorischen Archäologie 193. Bonn, Habelt. Naue, J. 1903. Die vorrömischen Schwerter aus Kupfer, Bronze und Eisen. München, Piloty and Loehle. Northover, P. 1988. The analysis and metallurgy of British Bronze Age swords. In I. Colquhoun and C. Burgess, The swords of Britain. Prähistorische Bronzefunde IV, 5, 130–148. München, C. H. Beck. Oakeshott, E. 1999. The Archaeology of Weapons. Woodbridge, Boydell Press [reprint of 1960]. O’Brien, S. 2009. Beyond the sharp bronze: warfare and society in Mycenaean Greece. Unpublished PhD thesis, University of Liverpool. Ó Faoláin, S. and Northover, J. P. 1998. The technology of Late Bronze Age sword production in Ireland. The Journal of Irish Archaeology IX, 69–88. Ó Faoláin, S. 2004. Bronze artefact production in Late Bronze Age Ireland. British Archaeological Reports, British series 382. Oxford, BAR Publishing. O’Flaherty, R. 2002. A consideration of the Early Bronze Age Halberd in Ireland – Function and Context. Unpublished PhD thesis, University College Dublin. Osgood, R. 1998. Warfare in the Late Bronze Age of North Europe. British Archaeological Reports, International series 695. Oxford, BAR Publishing. Papadopoulos, A. 2006. The Iconography of Warfare in the Bronze Age Aegean. Unpublished PhD thesis, University of Liverpool. Papadopoulos, T. J. 1998. The Late Bronze Age Daggers of the Aegean I: The Greek Mainland. Prähistorische Bronzefunde VI, 11. Stuttgart, Steiner. Parker-Pearson, M. 2005. Warfare, violence and slavery in later Prehistory: and introduction. In M. Parker-Pearson and I. J. N. Thorpe (eds), Warfare, Violence and Slavery in Prehistory. British Archaeological Reports, International series 1375, 19–35. Oxford, BAR Publishing. Peatfield, A. D. 1999. The Paradox of Violence: Weaponry and Martial Art in Minoan Crete. In R. Laffineur (ed.), Polemos: Le Contexte Guerrier en Égée á l’Âge du Bronze, 67–74. Liege, Aegeum. Pollington, S. 1996. The English Warrior from earliest times to 1066. Norfolk, Anglo-Saxon Books. 83

Ramsey, W. G. 1989. Middle Bronze Age weapons in Ireland. Unpublished PhD thesis, Queens University Belfast. Reeve, M. D. 2003. Vegetius’ Epitoma rei militaris. Oxford, Clarendon Press. Sandars, N. K. 1961. The First Aegean swords and their ancestry. American Journal of Archaeology 65, 17–29. Sandars, N. K. 1963. Later Aegean Bronze Swords. American Journal of Archaeology 67, 117–53. Schauer, P. 1971. Die Schwerter in Süddeutschland, Österreich und der Schweiz I (Griffplatten-, Griffangel-, und Griffzungenschwerter). Prähistorische Bronzefunde IV, 2. München, C. H. Beck. Sim, D. 2002. Iron for the Eagles. Stroud, Tempus. Smith, S. K. 2009. Skeletal evidence for militarism in Mycenaean Athens. Hesperia, Supplement 43, 99–111. Snodgrass, A. 1999. Arms and Armour of the Greeks. Baltimore, John Hopkins University Press. Talhoffer, H. 2000. Fechtbuch. Edited by M. Rector. London, Greenhill Books. Thorpe. I. J. N. 2005. The origins of warfare and violence. In M. Parker-Pearson and I. J. N. Thorpe (eds), Warfare, Violence and Slavery in Prehistory. British Archaeological Reports, International series 1375, 1–19. Oxford, BAR Publishing. Treherne, P. 1995. The warrior’s beauty: the masculine body and self-identity in Bronze Age Europe. Journal of European Archaeology 3, 105–144. Tselios, T. 2006. New Look at Minoan Metalworking Techniques. In J. Day, C. Greenlaw, H. Hall, A. Kelly, L. Matassa, K. McAleese, E. Saunders and D. Stritch (eds), SOMA 2004 Symposium on Mediterranean Archaeology. Proceedings of the eighth annualmeeting of postgraduate researchers. School of Classics, Trinity College Dublin. 20-22 february 2004, 193–197. British Archaeological Reports, International series 1514. Oxford, BAR Publishing. Uckelmann, M. 2006. Schutz, Prunk und Kult – Zur Funktion bronzezeitlicher Schilde. Anodos. Studies of the Ancient World 4/5, 2004/2005, 243–249. Uckelmann, M. in prep. Die Schilde der Bronzezeit in Nord-,West- und Zentraleuropa. Prähistorische Bronzefunde III, 4. Stuttgart, Steiner. Vandkilde, H. 2006a. Warriors and warrior institutions in Copper Age Europe. In T. Otto, H.

Thrane and H. Vandkilde (eds), Warfare and Society: Archaeological and Social Anthro-pological Perspectives, 393–422. Aarhus, Aarhus University Press. Vandkilde, H. 2006b. Warfare and gender according to Homer: An archaeology of an Aristocratic Warrior Culture. In T. Otto, H. Thrane and H. Vandkilde (eds), Warfare and Society: Archaeological and Social Anthro-pological Perspectives 515–528. Aarhus: Aarhus University Press. Waddell, J. 2000. The Prehistoric Archaeology of Ireland. Wicklow, Wordwell. Wagner, P. and Hand, S. 2003. Medieval sword and shield: The combat system of Royal Armouries MS I.33. Highland Village, Chivalry Bookshelf.

Whitley, J. 2002. Objects with Attitude: Biographical Facts and Fallacies in the Study of Late Bronze Age and Early Iron Age Warrior Graves. Cambridge Journal of Archaeology, 12 (02), 217–232. Wilde, W. R. 1861. Catalogue of the antiquities of animal materials and bronze in the museum of the Royal Irish Academy. Dublin, Royal Irish Academy. Wylie, A. 1995. The Reaction against analogy, Advances in Archaeological theory and method, Vol. 8, 63–111. New York, Academic Press. York, J. 2002. The life cycle of Bronze Age metalwork from the Thames. Oxford Journal of Archaeology, 21, 1, 77–92. Barry Molloy UCD School of Archaeology University College Dublin Belfield Dublin 4 Ireland

84

CHELSEA AND BALLINTOBER SWORDS: TYPOLOGY, CHRONOLOGY AND USE Steven Matthews

historically and culturally situated artefacts, and integral to wider discussions of Bronze Age societies. Typology and chronology should have an enormous bearing on any discussion of usewear, and in particular edge damage on weapons. Where once existed reasonably neat series of successive weapon types, with only minimal degrees of overlap (cf. Colquhoun and Burgess 1988), reassessment of these has created a far more subtle and complex landscape of rapier and sword types during the later Bronze Age.1 Such a complex theatre must, whenever possible, be given serious critical attention when looking at the use of different weapon types. In this paper issues of typology, chronology and use, will be discussed in relation to Chelsea and Ballintober swords, as a means of demonstrating aspects of the above agenda.

ABSTRACT This paper lays out an agenda for the detailed recording and dissemination of edge damage on Bronze Age weapons, combined with detailed typology and chronology. This approach is applied to swords of Type Chelsea and Ballintober in southern England. Their relationship to other swords of related types in northwest Europe is discussed, where their origin is critically reviewed in light of an earlier emergence under the influence of rod-tanged swords of Type Arco/Terontola. The identification of a number of chronologically significant variants highlights the later modification of Chelsea and Ballintober swords under the influence of early Urnfield and Atlantic swords types, particularly in Ireland and France. KEYWORDS

HISTORY AND TYPOLOGY OF CHELSEA AND BALLINTOBER SWORDS

Chelsea – Ballintober – Arco/Terontola – Twickenham – Mixnam – Battersea – typology – chronology – edge damage – use wear

English series These early tanged swords were first identified by H. W. M. Hodges (1956), and named for the find from Ballintober, Co. Mayo, which he regarded as intermediary between rapiers and the earliest indigenous flange-hilted swords. It was B. Trump (1962, 93) who first discussed their origins, which she saw occurring amongst the multitude of different weapon types in the Seine valley, before being introduced to England. G. Eogan (1965, 7 f.) similarly regarded rapier and sword development on the near Continent as being important but argued that Ballintober swords developed instead in England, in the Thames valley. C. Burgess (1968b, 15 f.) recognised that this series incorporated a number of different types, predicated on the form of the blade, of which there were three: Rosnoën rapiers, with parallel-sided blades and the same flat midsection as Group IV Atlantic rapiers, swiftly followed by Chelsea swords, retaining the same flat blade section but incorporating a leaf-shaped blade, and finally the Ballintober sword, which

INTRODUCTION This paper presents tentative steps toward exploring the development of two important aspects of usewear analysis: 1) The detailed recording of individual edge damage and, most importantly, its dissemination so that comparisons with other corpora can be made; 2) The critical integration of both typological and chronological frameworks so that more detailed comparisons can be made spatially and temporally. Thus far, usewear analysis has largely failed in both of these respects. These aspects are important if we are to achieve a greater degree of integration between metalwork studies and other aspects of Bronze Age research, particularly when addressing regional material manifestations and their interpretation. Without these we risk our studies remaining at the level of mere anecdotal observation, at best testing the limits of the mechanical properties of these objects rather than presenting them as 1

Cf. Needham 1982; Burgess 1991; Bridgford 2001; Brandherm and Burgess 2008.

85

Chelsea and Ballintober swords: Typology, chronology and use

Figure 1: Distribution of swords of Types Chelsea and Ballintober and their variants in Southern Britain

saw the adoption of the flat-lozenge blade section seen on the earliest leaf-shaped Urnfield swords. Burgess,2 disagreed that its origins were to be found in relation to the earliest Urnfield swords, specifically Type Hemigkofen, arguing instead that it was a combination of influences involving Rosnoën rapiers and leaf-shaped rod-tanged swords of Type Arco/Terontola. Of the latter, there is a noticeable concentration in the Seine valley,3 and a single example from the Thames valley (Colquhoun and Burgess 1988, no. 3). M. Rowlands (1976, 79 f.) and S. Needham (1982, 28) lend credence to this argument, as both saw Chelsea as the earliest form and Ballintober undergoing constant later modification, with the arrival of these early Urnfield swords, though the emphasis here was upon Rosnoën rather than Arco/Terontola origins. With the publication by I. Colquhoun and C. Burgess (1988) of ‘The Swords of Britain’, in the Prähistorische Bronzefunde series, the mixed Rosnoën-Arco/ Terontola origins for Chelsea and Ballintober swords was maintained, only now Ballintober swords were early and Chelsea relegated to being a derivative local variant. The primary distribution of these swords is found in northwest Europe, with the greatest concentration being in southeast England 2

(fig. 1), with notable outliers in the southwest and in southern Wales (Colquhoun and Burgess 1988, pl. 118. 119), provenanced examples demonstrate a liberal spread across the north of Ireland (Eogan 1965, fig. 85), and on the Continent concentrations in northern and northwest France (Gomez 1987, fig. 1). For Britain Colquhoun and Burgess (1988, 19–24) list a total of 39 swords, consisting of 23 Type Ballintober swords (no. 22–43. 776), 12 variant Chelsea swords (no. 44–54. 775), and four believed to be Irish imports (no. 55–58). However, given the significance of the Chelsea sword, particularly as a restricted regional series, we believe it should be termed a type rather than a variant, as it has been latterly assigned. These two types (fig. 2), Chelsea with its flat-sectioned blade and Ballintober with its lozenge-sectioned blade, form the basis of the typological scheme used in this paper, which will be further subdivided into variants, which are discussed below. Colquhoun and Burgess (1988, 19) established that Irish Ballintober swords differ from these English swords in having shoulders that are far more widely splayed, and identify four swords as being probable Irish imports. These are all Ballintober swords and are included amongst this group in the discussion here, being regarded as merely late forms. The sword from 3

Burgess 1968a, 44, postscript; 1974, 205, 318

Cf. McArdle 1969, 81–91 fig. 9; partially illustrated by O’Connor 1980, map 30.

no. 270.

86

Steven Matthews                                                                      

  5 2  

3 6  

1

4

 

  Figure 2: Rapier and sword types discussed in the text. – 1. Type Rosnoën rapier from Kingston on Thames, R. Thames (no. 4). – 2. Type Arco/Terontola sword from Corbeil, Essonne; – 3.early Type Hemigkofen from the R. Thames (no. 64). – 4. Type Ballintober, variant Twickenham sword from Twickenham, R. Thames (no. 23). – 5. Type Chelsea, variant Mixnam sword from Kingston on Thames, R. Thames (no. 48). – 6. Type Ballintober, variant Battersea sword from Battersea, R. Thames (1. 3–6 after Colquhoun and Burgess 1988; 2 after Mohen 1977, no. 234)  

Sandford, Oxford (Colquhoun and Burgess 1988, no. 77) has been included amongst those Chelsea swords. Two unpublished Ballintober swords have come to the author’s attention, both from the River Thames at London, and are listed in the figure 10 as swords A and B.4 As well as a further Ballintober sword recovered from the gravel pits at Mixnam, Egham, Surrey, and

listed here as sword C.5 An unusual northern example of a Ballintober sword from Barnhills Farm, Corsewall, Wigtownshire in Scotland, has recently been identified (Cowie and O’Connor 2007, 316–318 fig. 28. 2) but has not been included here. Though it is worth noting that the blade and hilt shape bear more resemblance to the Irish than English Ballintober series.

4 A: British Museum, London, acc. no. 1957.5-3.1. B: Salisbury and South Wiltshire Museum, Salisbury, acc. no. 1c4A.12.

5 Tomalin 1982, fig. 2. 1; not included in Colquhoun and Burgess 1988.

 

 

  87

Chelsea and Ballintober swords: Typology, chronology and use

Figure 3: Chronology and sequence of the sword development

Gomez (1987). Coffyn (1985, carte 13) lists 16 examples, which he divides into two groups. Type classique being the characteristic Ballintober sword with lozenge sectioned blade and subrectangular tang with four rivet-holes, and Type évolvé having notched tangs and blade outlines. Gomez (1987) similarly divides the Continental series in two groups. Type classique is again the Ballintober sword, of which there are seven examples, and a continentale variant, again based on notched tangs and the presence of blade outlines, of which there are 11 examples. The typologies of Coffyn (1985) and Gomez (1987) do not, however, correspond with that of Colquhoun and Burgess (1988). There are both Ballintober (e.g. Gomez 1987, no. 5. 8) and Chelsea (Gomez 1987, no. 1. 3) swords in Gomez’s Type classique, and the variant continentale consists of both riveted and notched (Gomez 1987, no. 6. 7. 13. 14) tanged swords, plain and grooved blades (Gomez 1987, no. 15. 16. 18), and examples both with (Gomez 1987, no. 9. 16) and without ricassi. A review of the entire series from northwest Europe suggests that the difference in blade section, largely ignored in discussions of these swords in Ireland and on the Continent, is indeed the major typological distinction. It is possible to identify further variants on the basis of the shape

Irish series For Ireland, Eogan (1965) listed 24 weapons and they remain undifferentiated within his Class 1 swords. Trump (1962, 93) was correct in her assessment that the Irish series was derivative of and later than the English series. The blades of the Irish swords are generally thinner, having a more sinuous outline, the tang is often more trapezoidal than any found in England, and the ends of the shoulders are noticeably extended. Tentatively, it would appear that the Irish series comprises 21 Ballintober6 and three Chelsea swords (Eogan 1965, no. 3. 10. 20). Colquhoun and Burgess (1988, 19, n. 2) suggest that nine of these Irish swords (Eogan 1965, no. 1–7. 12. 23) might be English imports, with all but one (no. 3) being of Type Ballintober. However, this author is inclined to view Eogan’s no. 4 and 5 as being of Irish form. Equally, all could simply be early examples of Class 1 swords. French series The similarity of the French series to those of England and Ireland was first discussed by J. Briard (1965, 164–166 fig. 55, 1–4), and more recently by A. Coffyn (1985, carte 13) and J. 6

Eogan 1965, no. 1. 2. 4–9. 11–19. 21–23. 618.

88

Steven Matthews

and Needham (1982, 28) in emphasising an earlier primary role being played by Rosnoën rapiers in the development of Chelsea and Ballintober swords, with early Urnfield swords impacting only on later Ballintober forms, also leave open the possibility for influence from rod-tanged swords of Type Arco/Terontola. Based on the then long Penard chronology, Needham suggested a two stage development for Chelsea and Ballintober swords, with the earliest examples emerging during Ha A1, and therefore contemporary with Arco/Terontola swords and Rosnoën rapiers, followed by secondary influence from early Urnfield swords of Types Hemigkofen and Erbenheim during Ha A2, with the development of a fuller leaf-shaped blade and more pronounced drooping shoulders, mirroring those of the flange-hilted swords. Rowlands (1976, 79 f.) similarly envisioned a two stage development, with Chelsea swords emerging first followed by Ballintober swords, which exhibit greater degrees of experimentation in hilt form, as a consequence of influence from early Urnfield swords. The two stage development suggested by Needham has considerable merit but misses the crucial characteristic that might be used to identify late examples, and that is presence of ricasso notches, absent on swords of the English series but common on Irish Class I weapons,10 where the bevelled edges have been drawn out to form primitive ricasso notches, and examples of which are all lozenge sectioned Ballintober swords. If the illustrations are reliable in this respect,11 their occurrence is significant, as the earliest ricasso notches only begin to appear regularly in southern England on Atlantic swords of Type Limehouse.12 This suggests that Irish Class 1 swords were still current when the earliest Atlantic swords began to develop there during Ha A2, suggesting a later date for these Irish Chelsea and Ballintober swords than in England. A late date is also confirmed by the occurrence of a Ballintober hilt fragment with ricasso notches from the hoard at Kerguerou en Rédené, Finistère (Briard 1965, fig, 57, B), which was found associated with an Ha A2 hilt fragment of an Atlantic sword of Type Essonne, the northern French equivalent of Type Limehouse (Burgess and O’Connor 2004, 191). We have so far suggested that there were three stages of development in Chelsea and Ballintober

of the tang and butt plate, which will be described below. CHRONOLOGY There is much debate concerning the dating of Chelsea and Ballintober swords (e.g. Gerloff 2007, tab. 13. 2; 2010, tab. 3), and whether they should be positioned earlier, in the mature stages of the Penard phase at the transition between Bz D and Ha A1, or later in Ha A2 during the early stages of Late Bronze Age 2, developing under the influence of early leaf-shaped Urnfield swords. The introduction of a short Penard phase, c. 1300–1150 cal. BC, (cf. Needham et al. 1996, 87. 90 ill. 15), means that the earlier divisions of Penard I and II, and their associated metalwork types (cf. Burgess 1974, 205; 1980, 266), are no longer valid, and has wider implications for the relative position of early sword types. The early date for Chelsea and Ballintober swords is based on their emerging in association with Rosnoën rapiers under influence from leaf-shaped rodtanged swords of Type Arco/Terontola,7 of which there are a number in the Seine valley, and a single English example from the Thames.8 An early date was also suggested by Briard (1965, 298), who saw them emerging prior to the arrival of the early Urnfield swords, and similarly in conjunction with Rosnoën rapiers during the later stages of his Atlantic Bronze Final I, our Penard phase. As our chronologies have altered significantly since Colquhoun and Burgess (1988) established the chronology of these different types, the relative position of the weapons discussed here have been compared and updated in figure 3. There remains scepticism as to the role that these rod-tanged swords played in the development of Chelsea and Ballintober swords, given the general paucity of examples in southern England (cf. O’Connor 1989, 126), with the early Urnfield swords still being seen as the more likely origin for the adoption of the leaf-shaped blade.9 Similar to Burgess, a number of early detailed discussions (e.g. McArdle 1969, 52; Coombs 1972, 173) also saw a significant role for these leaf-shaped rodtanged swords in the development of Chelsea and Ballintober swords. Unfortunately these remain unpublished. However, Rowlands (1976, 79 f.) 7 Burgess 1974, 318, n. 270; 1980, 266; Colquhoun and Burgess 1988, 22 f. 8 Burgess 1968a, 44, postscript; Colquhoun and Burgess 1988, no. 9. 9 Lawson 1982, 282; Gerloff 2007, tab. 13. 2; 2010, tab. 3.

10

Eogan 1965, 7, no. 5. 8. 9. 11. 13. 14. 20. 21. See also Burgess 1968a, fig. 3. 6; 1968b, fig. 3. 6. 12 E.g. Colquhoun and Burgess 1988, no. 97. 100. 101. 103. 104. 11

89

Chelsea and Ballintober swords: Typology, chronology and use

ving previously maintained an Atlantic rapier tradition of comparable scale to that of Britain and Ireland (Matthews 2010, 84). The Seine and Thames valleys invariably kept pace with each other, and likely represent a significant interaction zone, what Rowlands (1998, 162–166) termed the Channel Core Area. Therefore, whilst only one Arco/Terontola sword has been found in the Thames valley, it is not unreasonable to assume that they once had a comparable presence to that of the Seine valley. D. Coombs (1972, 172 f.) and I. Colquhoun and C. Burgess (1988, 21) have suggested the means by which the development of the tang on these rod-tanged swords could have progressed to that found on contemporary Chelsea and Ballintober swords. This would begin with the removal of the rod itself, and relocating the rivet-holes from the top of the blade onto the tang itself, which would have similarly been accompanied by a change in blade shape to accommodate changes in balance (Colquhoun and Burgess 1988, 21). These changes would presumably have occurred in the Thames valley rather than in France. Were this not the case, the French series would otherwise constitute an apparent chronological problem in terms of the typological development of Chelsea and Ballintober swords, as this region is populated by swords with primarily later rather than earlier typological traits. The absence of typologically early Chelsea and Ballintober swords was likely due to the overwhelming dominance of Rosnoën rapiers in this region, and also to continuing experimentation with swords of Ambleside/Bardouville types. Production in the Thames valley, however, has always orientated toward increased standardisation rather than experimentation and difference, a regional distinction also found amongst subsequent Atlantic swords. With the Rosnoën tradition largely absent from southern England, it is less surprising then that it was here, in the Thames valley rather than the Seine valley, that the earliest Chelsea and Ballintober swords emerged, took root and flourished. In terms of form there are noticeable differences between Arco/Terontola and Chelsea and Ballintober swords. The necessity of these changes has already been remarked upon above. However, it is worth noting that a number of Chelsea swords have a less than sinuous blade outline, one that errs close to parallel sided, and is highly reminiscent of some Arco/Terontola swords. It is true that on some, but not in fact

swords: 1) their emergence in relation to Rosnoën rapiers and rod-tanged Arco/Terontola swords, providing the characteristic tang shape and leafshaped blade, followed by 2) secondary influence from early Urnfield swords, notably Hemigkofen swords, contributing further changes to the shape of the blade, tang and shoulders, and finally 3) the adoption of traits otherwise found only on early indigenous Atlantic swords, such as those of Type Limehouse/Essonne, in the form of ricasso notches. It is worth briefly considering the conditions under which the earliest development of Chelsea and Ballintober swords took place, particularly in respect of their relationship to Arco/Terontola swords, as this remains the greatest point of contention in both typological and chronological terms. Although mistaken in seeing the earliest origins of Ballintober swords occurring in the Seine valley, and then being introduced into England and then Ireland, Trump (1962, 93) was correct in alluding to the importance of this region in terms of early sword development. This region has not only the greatest variety of weapon types during this period (Needham 1982) but also a noticeable concentration of leaf-shaped rod-tanged Arco/Terontola swords.13 N. K. Sandars (1976, 237) is alone in observing that there was real, credible alternatives to the flange-hilted swords of Urnfield origin during this period, in the form of these rod- and tanged weapons, which are to be found liberally distributed across France and northern Italy. Indeed, the earlier straight-bladed rod-tanged swords had a significant impact on local Atlantic rapier traditions, resulting in the Ambleside/Bardouville series of swords (Needham 1982). In his description of this eclectic series of swords, Needham provided much of the necessary groundwork as to the possible origins of Chelsea and Ballintober swords. The tanged straight-bladed Ambleside/ Bardouville series of swords belong to early Penard, contemporary with Bz D, and emerged alongside Rosnoën rapiers, with leaf-shaped rodtanged Arco/Terontola swords belong to the same period, but emerging in Ha A1, late in the Penard phase. All three weapons types are prominent in the Seine valley, and in the case of Arco/Terontola swords, represent the earliest occurrence of leafshaped blades in the Atlantic region. The impact of these weapons on contemporary rapier traditions should not be underestimated, as contemporary Group IV riveted and notched Atlantic rapiers of Type Appleby are, unlike the rest of northwest Europe, largely absent from France, despite ha13

Cf. McArdle 1969, Class B1 swords, 81–91 fig. 9; partially illustrated by O’Connor 1980, map 30.

90

Steven Matthews

all, Arco/Terontola swords, the greatest width of the blade is located closer to the tip than on any Chelsea or Ballintober sword. However, as Colquhoun and Burgess (1988, 21) have suggested, the absence of the rod would have required that the greatest width of the blade be moved further back. The significance of this difference in blade shape between Arco/Terontola and Chelsea swords may in fact be more apparent than real, and one that should be considered in terms of the overall blade proportions or ratios, as we shall see below.

Variant Mixnam These swords have noticeably elongated tangs with subtle trapezoidal or almost parallel sides. The spacing of the rivets can be of two forms, being either close together or wide apart but generally they are not placed close to the terminal of the tang. This terminal can be either rounded or flat. It is likely that variant Mixnam represents a later rather than earlier form, the tang being the longest and the blade the most sinuous. Weapons similar to those of variant Mixnam are common amongst Irish Class, and this is not surprising given that the Irish series likely represents a later development, subsequent to the development of Ballintober swords in the Thames Valley. These Irish weapons differ in a number of respects from their English counterparts. First, the top pair of rivet holes can be set far closer to the terminal of the tang than is usually found in England, and the tang is often more noticeably squared than in England. The Irish series likely requires the identification of subtly different variants than those in England but at present this is not possible given the difficulty in using the poor quality of the illustrations provided by Eogan (1965).

CHELSEA AND BALLINTOBER VARIANTS Having identified probable early and later traits amongst Chelsea and Ballintober swords throughout northwest Europe, we can use these to establish a limited number of variants (fig. 2), largely based on the shape of the tang, that appear to have a greater degree of chronological validity than does the basic distinction in blade form between the Chelsea and Ballintober types. It should be noted that the below descriptions are based on ideal forms and hybrids, as is so often the case, are common, particularly amongst such a short lived series. Uniquely, the same variants also appear to be valid for both types.

Variant Battersea Variant Twickenham

Variant Battersea are those swords with a broad tang, usually with the appearance of being square. These swords clearly represent a different approach to strengthening the hilt attachment to that found on swords of variant Mixnam. It is possible that they are slightly earlier than Mixnam but not significantly so and both likely ran contemporaneously until Ballintober swords were no longer fashionable or viable. Battersea swords are also prevalent amongst the Irish series but here they often appear combined with the lengthened tang of variant Mixnam, demonstrating their contemporaneity. That a number of Battersea swords in Ireland have primitive ricasso notches again confirms their late date. Whether the variants recognized for England have any relevance to the local development of Irish Ballintober swords remains to be seen. However, as the greater part of the Irish series is here considered to be later than the majority of the English series, it is surely no coincidence that the majority of the Irish series conform to our Mixnam and Battersea variants, with a great number appearing to be a hybrid of both, but few of variant Twickenham.

Variant Twickenham swords stand out readily from the more developed Chelsea and Ballintober variants. The tangs are usually short with an oval terminal and subtle trapezoidal sides that, on presumably early swords, sweep outwards in a gently sloping diagonal line about the end of much restricted shoulders where they join the butt plate or upper part of the blade. Already mentioned, the sword from Sandford, Oxford (Colquhoun and Burgess 1988, no. 7), has been included amongst our Type Chelsea, variant Twickenham. Although not a reliable association, this sword is also reputed to have been found with an Atlantic rapier (Burgess and Gerloff 1981, 106 f. no. 624) of Group IV Type Appleby, variant Weybridge. This type of rapier is tentatively assigned to very early in the Penard phase, as it retains the more archaic trapezoidal shaped butt, the rivet-holes placed towards the upper corners of the hilt plate rather than at the sides, the latter being more characteristic of later Type Appleby rapiers. This suggests an earlier rather than later date for variant Twickenham swords. 91

Chelsea and Ballintober swords: Typology, chronology and use

Figure 4: L/W ratio of the discussed sword types 92

Steven Matthews

Figure 5: The blade sections and units of the used terminology

Variant Nantes

some of the typological and chronological claims made above.

The types and variants recognised in southern England also appear to be relevant to the French series, demonstrating a mixture of Twickenham, Mixnam and Battersea swords. A long, thin and almost universally notched variant is also found (Briard 1965, fig. 55, 1. 2) and differs from any English example, and is therefore perhaps deserving of a local designation. Although blade outlines have no typological significance they may at least be chronologically late, given their almost exclusive association with the Mixnam variant in England and with similar swords in France. This unusual French variant forms a fairly homogenous grouping in lower northwest France and has therefore been named variant Nantes, again irrespective of type. Confirmation of a late date for this form is provided by the hilt fragment of an undecorated example from the hoard at Kerguerou en Rédené, Finistère (Briard 1965, fig. 57, B), which also bears primitive ricasso notches.

Blade comparisons M. A. Brown (1982) and S. Needham (1982) have demonstrated the validity of testing individual blade proportions against typological groups. The blade proportion is calculated on the basis of the blade length to maximum width ratio (blade width ÷ blade length). The blade length was measured from the tip to the bottom of any ricasso that might be present but does not include the ricasso itself. This allows comparison between a wider range of weapon types regardless of different hilting methods or the presence of a ricasso, which is particularly important for the Penard phase where a large number of different weapon forms co-existed contemporaneously. The individual blade length/width (L/W) ratios of a number of types relevant to the above typological and chronological discussion are plotted in figure 4, where we have compared Chelsea swords against Ballintober, Arco/Terontola, Hemigkofen and Rosnoën weapons. A clear but subtle difference between Chelsea and Ballintober swords in terms of their blade proportions is evident. Chelsea swords have a narrow L/W distribution of 8–11, whereas Ballintober swords have a much broader L/W distribution of 9–14. Moreover, Chelsea swords generally range much lower in their L/W measurements than Ballintober swords. The degree of overlap, however, justifies the identification of the same variants amongst both types. With Chelsea and Ballintober swords having identifiably distinct value ranges in the distribution of their L/W blade ratios, and Chelsea swords having been argued to be chronologically earliest,

TYPOLOGICAL COMPARISON OF WEAPONS Having established the significance of a series of typo-chronological traits amongst Chelsea and Ballintober swords, represented by these different variants, we might further explore the possible chronologically successive influence of developed leaf-shaped rod-tanged Arco/Terontola swords and the early Urnfield Hemigkofen swords. Much of what has been discussed above can be substantiated by comparing a limited number of quantitative morphological characteristics between Chelsea and Ballintober swords and those of other contemporary weapon types. Two aspects of form will be compared, the blade and hilt, which substantiate 93

Chelsea and Ballintober swords: Typology, chronology and use

Class

Description

Class 1

Cut: takes the form of a V-shaped indentation, with tapering sides that end in a narrow point rather than a U-shaped terminal, usually with no loss of material, forming a lip either at the terminal or on either side of the cut.

Class 2

Notch: takes the form of a U-shaped indentation that generally has parallel sides, ending in a U-shaped terminal that is usually the same width as the beginning of the notch. The excess material from a notch is always missing.

Class 3

Dent: takes the same form as Class 2 edge damage – a U-shaped indentation that generally has parallel sides, ending in a U-shaped terminal that is usually the same width as the beginning of the notch. There is no loss of material and a lip the terminal of the dent, or a bulb running the entire length of the damage, occurs on one side of the flat face of the blade.

Class 4

Broad shallow notch: takes the same form as Class 2 edge damage but wherein the length is always greater than its depth and is bracketed by ends that are usually parallel sided. Excess material, in the form of a lip, is not present.

Class 5

Broad shallow dent: takes the same form as Class 3 edge damage but wherein the length is always greater than its depth and is bracketed by gently sloping edges. There is no loss of material and a lip the terminal of the dent, or a bulb running the entire length of the damage, occurs on one side of the flat face of the blade. Figure 6: Description of the different use wear classes

it is this group that we have compared with the remaining weapon types. As only a single Arco/Terontola sword is known from England, examples from the Seine valley were also included. The L/W ratio of these swords has the same restricted distribution to that of Chelsea swords, between 8–11/12, with some examples similarly ranging low in L/W measurements. Their overall distribution, however, entirely encapsulate the distribution of L/W ratios of Chelsea swords. The fact that, despite some differences in blade form, they share the same overall blade proportions lends greater credence to the view that Chelsea swords were indeed influenced by Arco/Terontola swords. The distribution of Hemigkofen swords is more complicated, and appears to fall into two discrete groups. The first have L/W rations between 8–9, and are represented by those swords with short broad blades. The second group have ratios between 10–14, and have long slender blades. Whilst the first group has a distribution similar to that of Chelsea swords their range is so high they fall almost entirely outside of their distribution, unlike the Arco/ Terontola swords. Both groups of Hemigkofen swords, however, have a distribution similar to that of Ballintober swords. The differences between Chelsea and Hemigkofen swords, and the similarity between the latter and Ballintober swords, appears to substantiate Needham (1982, 28) and Rowlands (1976, 79) view of a continual process of development amongst Chelsea and Ballintober swords in relation to new influences, with Hemigkofen representing secondary rather

than primary influence. The two distinct distributions also lend considerable weight to the argument for earlier and later Hemigkofen forms (cf. Cowen 1951). Finally, Chelsea swords were compared with five English Rosnoën rapiers. These were found to have quite different distributions suggesting that, whilst they may share a common link in terms of their hilt arrangement and blade section, any further similarity ends there. These comparisons substantiate the primacy of Arco/Terontola swords in the development of Chelsea swords. The significant difference between the L/W ratios of Rosnoën rapiers and Chelsea, Hemigkofen and Arco/Terontola swords (fig. 4) demonstrate that they represent entirely different classes of weapon, as has always been suggested. Whilst the different L/W ratios suggest subtle but important differences between sword types, they indicate a real qualitative difference between rapiers and swords, substantiating the argument that they represent distinct suites of gestures and techniques (Burgess and Gerloff 1981, 113), and that the traditional terminological distinction of ‘rapier’ and ‘sword’ should be maintained. Hilt comparisons A single variable – the width of the shoulders – also provides a reliable basis on which to compare both variants and types within the Chelsea and Ballintober series, and again with other contemporary weapon types, particularly as changes 94

Steven Matthews

of 6.66 cm (5 ex.); Type Taplow 5.81 cm (8 ex.); Type Limehouse/Essonne 5.45 cm (4 ex.). In summary, whilst the subtle difference in the English types and variants may not be wholly significant alone their chronological arrangement is confirmed by the difference in shoulder width found on the Irish Class I and French variant Nantes swords, both of which have similar shoulder widths to Type Hemigkofen swords and the earliest Atlantic swords of Type Limehouse/ Essonne.

in this aspect of their form has also been argued to have a degree of typo-chronological significance. Rowlands (1976, 79 f.) and Needham (1982, 28) discussed changes in the shape of the shoulders of Ballintober swords in relation to the form of shoulders found on early Urnfield swords. However, unlike flange-hilted swords, where the shoulders would be visible through the attached hilt-plates, the shoulders of Chelsea and Ballintober swords would not, being obscured entirely by the hilt in the same way they would have been on Atlantic rapiers and AmblesideBardouville swords. The angle of the upper face of the shoulders, whether high and straight, or short and drooping, would therefore have been stylistically inconsequential. The conjectural reconstruction of hilts for a number of weapon types found during Bz D–Ha A1 by Needham (1982, 28–36 fig. 12), including Ambleside/ Bardouville swords, Rosnoën rapiers and a number of rod-tanged swords similarly provide a likely model for the hilting arrangement of Chelsea and Ballintober swords. Whilst Needham suggested that these drooping shoulders might be a late trait, we believe it is instead the width of the shoulders. Attempts to widen the width of the butt as whole, most likely to mimic the greater width of shoulders on early Urnfield swords, had the unintended consequence of rendering almost horizontal the upper face of the shoulders on some later Ballintober swords. That this is a later trait is confirmed by the presence of such shoulders on those Ballintober swords with ricasso notches, as well as amongst the later Mixnam and Battersea variants, and on some French notched swords (e.g. Briard 1965, fig. 55, 1. 2), and is all but ubiquitous on the Irish series. We can see this trend in more detail if we compare the shoulder width of Chelsea and Ballintober swords with other contemporary weapon types. If we look at the shoulder width of Chelsea and Ballintober swords by type and variant they and contemporary swords have a mean average of: Ballintober 4.26 cm (19 examples); Chelsea 4.23 cm (12 ex.); Twickenham 4.11 cm (15 ex.); Battersea 4.57 cm (seven ex.); Mixnam 4.35 cm (12 ex.); Irish Class 1 swords 4.95 cm (22 ex.); French Ballintober 4,08 cm (17 ex.);14 Type Hemigkofen 5.15 cm (15 ex.); Type Erbenheim 5,5 cm (3 ex.); Atlantic swords: Type Clewer 5.78 cm (9 ex.); Type Limehouse 6.36 cm (13 ex.); Type Limehouse, variant Mugdrum swords

EDGE DAMAGE ON CHELSEA AND BALLINTOBER SWORDS IN ENGLAND AND WALES Having substantiated our typo-chronological framework for Chelsea and Ballintober swords we now turn our attention to certain aspects of their use, through a discussion of edge damage. The socio-functional aspects of Ballintober swords, in particular their depositional circumstances, have been briefly discussed by R. Bradley (1998, 125 f. fig. 28). We are concerned, however, with those aspects of their use life prior to deposition and after production, a part of the life-biography (Gosden and Marshall 1999) that is difficult to ascertain and often neglected. This section will introduce limited aspects of the recording, categorisation, dissemination and analysis of edge damage currently being prepared by the author (Matthews, in prep.) of the Atlantic rapiers and swords from northwest Europe, from the Taunton/Portrieux/ Mont-Saint-Aignan/Boix-Saint-Croix phase until the end of the late Wilburton/Blackmoor/St. Nazaire phase (c. 1400–950 cal. BC; fig. 3). Two aspects of edge damage will be discussed in this study. The first is the type of edge damage, and the second is the spatial distribution of each of these different types of damage (fig. 5. 6). These are presented according to the different typological groups in figures 7 and 8, and in tabular form by each individual sword in figures 9–12. Of the 43 Chelsea and Ballintober swords from southwest and southeast England and southern Wales (fig. 1), 21 have been examined and traces of edge damage recorded and analysed. This is presented in figures 9–12 by the catalogue numbers used by Colquhoun and Burgess (1988). Two Ballintober swords not included in that work have also been examined and are listed in figure 10 as A and B. Excluding

14

However, as an eclectic group this last number increases dramatically amongst the French series if we

consider only those of variant Nantes.

95

Chelsea and Ballintober swords: Typology, chronology and use

Figure 7: The compared edge wear damage on the swords of Types Chelsea and Ballintober 96

Steven Matthews

Figure 8: The compared edge wear damage on the swords of Types Chelsea, Rosnoën and Arco/Terontola 97

Chelsea and Ballintober swords: Typology, chronology and use

Type Chelsea (8 examples) No. Var. 44 M. 48 M.

Name

Upper 1

near Battersea, R. Thames Kingston on Thames, R. Th. Kingston on Thames, R. Th. Wandsworth, R. Thames Chelsea, R. Thames

49

T.

51

B.

52

B.

54

T.

Thorpe Hall

77

T.

Sandford, R. Thames

775 T.

Oystermouth

2 3

Middle 3

1

2

Lower 3 2

1

2

3

2 2

2.2 2 3

3

3.2

2

2 3.2.2

2.2

R L R L R L R L R L R L R L R L

Figure 9: Edge damage on Type Chelsea swords (after Colquhoun and Burgess 1988). – Variants: T. = Twickenham; M. = Mixnam; B. = Battersea

Type Ballintober (13 examples) No. Var.

Name Brentford, R. Thames Twickenham, R. Thames

22

M.

23

T.

24

M.

Egham

25

M.

Peterborough

36 37

M. M.

Barking, R. Thames Kingston Bridge, R. Th. Battersea, R. Thames

38

T.

41

M.

Worth

43

M.

London

55

B.

Unknown

56

B.

Shepperton, R. Thames

A

T.

R. Thames

B

T.

R. Thames

Upper 1

2

3.3

5

Middle 3

3 3.1 3 3.5 5

1

2

1 3.5.5

3.4 2 3 3

Lower 3 1 2.4

1 2 2.3

2.3.3

3 2.2 2.2.5

3.3 4

3.5 3

2 2 1 4 1.3.3.3 3

3 2 4.4 3 3 3 3

3.5

2.2.2.2

3 2

3

2.2.2

3

2.2.2 2(x4).3

2 2

1 3

1.2 3

4.4

3.4 2.2 5

3 3 2.2

3 3 3 4 2

2.2

4.2 2

4.2 2.2

2.4 2 2

3

R L R L R L R L R L R L R L R L R L R L R L R L R L

Figure 10: Edge damage on Type Ballintober swords: (no. all after Colquhoun and Burgess 1988; A. B after author, see footnote 4). – Variants: T. = Twickenham; M. = Mixnam; B. = Battersea 98

Steven Matthews

Type Arco/Terontola (1 example) No.

Name

3

London, R. Thames

Upper 1 2

2

Middle 3 2

1

2 2.2.2

Lower 3

2

1 2.4 2

2 2.2

3 2

R L

Figure 11: Edge damage on a Type Arco/Terontola sword (after Colquhoun and Burgess 1988) Type Rosnoën (4 examples) No. 4 5 8 14

Name

Upper 1

2

Kingston on Thames, R. Th. Lambeth, R. Thames Putney, R. Thames Methwold

2

2.2.2.2 3

Middle 3 3.3

3

1 1

5 1.3 3 1.3.3.3

Lower

2 3.3 3

3

2.2 2

3 1

1

1.3 2.3 3.3.3

2 3

3

2

R L R L R L R L

Figure 12: Edge damage on Type Rosnoën rapiers (after Colquhoun and Burgess 1988)

on this basis. Similar to the approach adopted by R. D. A. Savage (1979, pl. 1) in his analysis of two Ewart Park swords from the Carp’s-Tongue hoard at Watford, Kent, this cutting blade is divided into three equal sections, which are then further subdivided into three. These divisions and the terminology employed are illustrated in figure 5. These divisions allow the frequency and distribution of different types of edge damage to be analysed and compared relatively, irrespective of the actual length differences exhibited not only by individual weapons within a series but also between different weapon series and types. Five different classes of prehistoric edge damage have been identified and these are described in figure 6. Although similar, there are some differences between the categories used here and those established by Bridgford (2000, 105) on the basis of her experimental damage types. Those described by Bridgford occur in significant quantities on swords of Type Ewart Park but rarely earlier. Violent and excessive edge damage is not common even on Type Wilburton swords, except where purposefully enacted (Matthews 2011). Moreover, the long term effects of depositional circumstances on individual swords, in terms of corrosion and bronze disease, have had a significant effect on the nature of edge damage. The simplification of the classes of damage used

fragments, this represents just below 50 % of all Chelsea and Ballintober swords in England and Wales. The single English Arco/Terontola sword and four English Rosnoën rapiers have also been examined for edge damage and are included in this study.15 Figure 7 presents the edge damage from both Chelsea and Ballintober swords, and is arranged in a series of separate tables according to the five different classes of edge damage discussed below. The edge damage from these Chelsea swords is also presented in figure 8 alongside the small sample of English Rosnoën and Arco/Terontola weapons. Edge damage: definitions and terminology The basic unit of analysis is obviously the individual weapon. In this study, these have been divided into a number of different analytical parts and grouped accordingly. As the research this paper is drawn from is concerned with both rapiers and swords (Matthews, in prep.), where ricassi are entirely absent from the former group, this element has been differentiated from the cutting blade and grouped with the hilt, so that comparisons between different weapon forms could be achieved. Measurements given in the tables below for the length of blade are therefore made 15

Colquhoun and Burgess 1988, no. 3–5. 8. 14; Needham 1982, fig. 6 no. 7.

99

Chelsea and Ballintober swords: Typology, chronology and use

Balance point

Ballintober

Chelsea

Sword no.

Weight

cm/gms

Corresponding blade

Length from Blade unit tip

Top

Bottom

44

591 g

8,924.1

15.1

28

Middle 1

28.8

24

48

309 g

3677.1

11.9

22

Middle 1

22.6

18.9

51

299 g

3,348.8

11.2

27.4

Middle 1

28.2

23.5

49

417 g

6,171.6

14.8

20.1

Middle 1

20.9

17.4

52

385 g

4,966.5

12.9

23.1

Middle 1

24

20

54

–

–

–

–

–

–

–

775

353 g

5,506.8

15.6

23.7

Middle 1

26.2

21.9

77

386 g

5,944.4

15.4

25.8

Middle 1

27.6

23

22

663 g

9,613.5

14.5

25.6

Middle 1

26.8

22.4

23

683 g

11611

17

29.3

Middle 1

30.8

25.6

24

587 g

8,687.6

14.8

26.8

Middle 1

27.8

23.1

25

517 g

7,444.8

14.4

26.5

Middle 1

27.2

22.6

36

393 g

7,427.7

18.9

29.2

Middle 1

32

26.6

37

583 g

8,628.4

14.8

25.3

Middle 1

26.8

22.4

38

332 g

5,046.4

15.2

22.9

Middle 1

25.4

21.1

41

261 g

3,053.7

11.7

19.2

Middle 1

20.6

17.1

43

605 g

8,772.5

14.5

27.8

Middle 1

28.2

23.5

55

560 g

8,568

15.3

29.6

Middle 1

30

25

56

338 g

5,577

16.5

24.3

Middle 1

27.2

22.6

A

387 g

7,314.3

18.9

27.8

Middle 1

31.2

26

B

492 g

6,199.2

12.6

23.5

Middle 1

27.2

22.6

Figure 13: The length/weight ratios (cm/gms) relating to the effective balance of individual Types Ballintober and Chelsea swords, as calculated from their weight x the distance of the balance point to the hilt; measurements in cm

here have had to reflect these circumstances, as has the necessity of devising categories that have universal applicability to both the rapier and sword series, which themselves reflect a very real difference in both technical craftsmanship and use in combat. Edge damage: data It is clear from the tables in figure 7 that the overall volume of edge damage on Ballintober swords both as a whole, and as individual objects (fig. 10), is far greater than that found on Chelsea swords (fig. 9). Amongst Ballintober swords, every blade unit has some evidence of damage but usually only one or two instances, meaning that the edge damage is unequally distributed between the swords examined. For example, only one sword 100

(fig. 10 no. A) was found to have only one blade unit undamaged, whilst three swords (fig. 10 no. 38. 43. 56) had damage only on one blade unit. All Ballintober swords have at least some edge damage on half or more of their blade units, with the majority of this damage occurring between blade units Middle 3 and Lower 1, primarily the lower half of the cutting blade, and constituted by edge damage of Classes 1 to 3, with the greatest volume being Classes 2 and 3. Half of the Chelsea swords exhibited no edge damage (fig. 9 no. 51. 52. 54. 77), and of the remaining three had edge damage on only three blade units (fig. 9 no. 44. 49. 775) and the remaining one had damage on only two blade units (fig. 9 no. 48). Whilst there is a small concentration of edge damage of Class 2 on blade unit Lower 2 amongst Chelsea swords, there were otherwise no significant concentrations. What small amount of

Steven Matthews

Corresponding blade unit Sword Centre of no. percussion Blade Top Bottom unit

14,000

13,000

12,000

11,000

Medium 10,000

9,000

7,000 8,000

Good 6,000

5,000

4,000

3,000

Light

44

44

18.9

Middle 3 19.2

14.4

48

12.2

Middle 3 15.1

11.3

49

13.3

Middle 3 13.9

10.4

51

16.5

Middle 3 18.8

14.1

52

15.2

Middle 3

16

12

54

–

–

–

–

775

14.7

Middle 3 17.5

13.1

77

15.2

Middle 3 18.4

13.8

22

16.6

Middle 3 17.9

13.4

23

20.3

Middle 3 20.5

15.4

24

16.6

Middle 3 18.5

13.9

25

25

16.3

Middle 3 18.1

13.6

36

36

17.9

Middle 3 21.3

16

37

17.5

Middle 3 17.9

13.4

38

15.3

Middle 3 16.9

12.7

41

12.6

Middle 3 13.7

10.3

43

18.3

Middle 3 18.8

14.1

55

16.5

Middle 3

56

15.9

Middle 3 18.1

13.6

A

23.2

Middle 2

20.8

B

17.5

Middle 3 18.1

49

Chelsea

Chelsea

48 51 52 775 77 22 23

Ballintober

Ballintober

24

37 38 41 43 A B 55 56

14

20 26

15

13.6

15

Figure 14: The distribution of Types Chelsea and Ballintober swords according to their balance category, calculated according to their ‘turning moment’ (weight x distance of the hilt from the centre of gravity = cm/gms), with the categories of Good (< 10,000 cm/gms), Medium (10,000–15,000 cm/gms) according to Bridgford’s experiments (1997, 104 f.) Figure 15: Length from tip to the centre of percussion, as represented by the widest part of the blade, and the upper and lower measurements for the corresponding blade unit; measurements in cm

damage there was, restricted to edge damage of Class 2 and 3, was spread across the remaining blade units, except for Upper 1 for which there was no damage found on any individual Chelsea sword. The greatest number of instances across all blade units of a single type of edge damage was of Class 3, with 41 instances of damage on Ballintober swords and only three instances on Chelsea swords, with Class 2 damage being the next most significant with 39 instances across all blade units amongst Ballintober swords and 14 instances amongst Chelsea swords. The most significant volume of any one type of edge 101

damage was found to be Class 2 of which there 16 instances on Ballintober swords but only one instance amongst Chelsea and occurred on blade unit Middle 3. The next significant volume was also Class 2 of which there were nine instances on Ballintober swords and again only one instance on Chelsea swords, which occurred on blade unit Lower 1. The sample of Rosnoën and Arco/Terontola weapons (fig. 8. 11. 12) is not significant enough to draw any firm conclusions regarding their edge damage. Comparatively, however, it is noticeable that there is a more even spread of the number of instances of edge damage along the length of the

Chelsea and Ballintober swords: Typology, chronology and use

cutting blade, similar to that on Chelsea swords, compared to the noticeable bias in the distribution of edge damage toward the lower half of the cutting blade on Ballintober swords. Of the four Rosnoën rapiers examined one was found to have no edge damage at all (fig. 12 no. 5), whilst the remaining rapiers had damage on at least six of their blade units (fig. 12 no. 8), five of their blade units (fig. 12 no. 14) and four of their blade units (fig. 12 no. 14). The Arco/Terontola sword (fig. 11 no. 3) was found to have edge damage on all but 2 of its blade units. Like Chelsea swords, the majority of its edge damage was of Class 2 and 3, however, Rosnoën rapiers also had five instances of Class 1 edge damage, of which Chelsea swords exhibited no evidence of. More significantly, although having only half the number of examples, Rosnoën rapiers exhibited 35 instances of different classes of edge wear compared to only 17 instances amongst Chelsea swords. There is clearly a marked difference in both the quantity and the distribution of edge damage between Chelsea and Ballintober swords (fig. 7). Edge damage of all types show their densest concentration on the bottom half of the blade on Ballintober swords, particularly in respect of Class 2 and 3 damage, with notable concentrations in blade units Middle 3, and Lower 1 and 2. With the exception of blade unit Middle 3, all edge damage on Chelsea swords is concentrated in the centre and higher parts of the blade, with the only noticeable concentration of edge damage being of Class 2 in blade unit Middle 1. The damage on Chelsea swords is confined entirely to an even spread from Upper 2 to Middle 3 and only of Class 2 and 3. On Ballintober swords the distribution of edge damage is much wider, with at least every blade unit being affected across the entire group, although the Upper part of the blade demonstrates significantly less damage than the Middle Lower areas of the blade. The sort of heavy damage represented by edge damage of Class 1 is infrequent on Ballintober swords when compared to the far more numerous instances of Class 2 and 3, whilst damage of Class 1 is entirely absent on Chelsea swords. In order to assess the significance of these patterns concerning the spatial distribution of the edge damage on the two sword types, we must first establish a number of technical characteristics, in particular the position of the balance point and the centre of percussion, in relation to individual blade units, as both are likely to have been a contributing factor in the weapons effective capability, and consequently the occurrence of edge damage. 102

Balance Establishing the relative position of both the balance point and the centre of percussion requires that the swords studied were largely complete, and therefore broken swords are excluded. Of the twenty one swords examined in this study only twenty were therefore viable for further examination. The balance point of each weapon was recorded and measured from the top of the weapon. When calculated according to the relative position of individual blade units on each Chelsea and Ballintober sword, the balance point was found to correspond consistently to area Middle 1 (fig. 13). In her study of the Irish Bronze Age swords, S. Bridgford (1997, 104) assessed the effective balance of each sword by calculating the turning moment, being the weight x the distance of the hilt from the centre of gravity or balance point. These measurements presented here are approximate as no allowance has been made for the missing hilt plates, though this difference is likely to be negligible (cf. Bridgford 1997, 105), and unlikely to affect its assignation to a particular blade unit. The location of the balance point, as measured from the bottom of the hilt, the weight, and the approximate length/weight ratio for these swords is given in figure 13 by group. Bridgford arranged these cm/gms results into three balance categories based on her own subjective handling of these weapons. The position of these measurements in relation to their balance categories are shown in figure 14. As Bridgford did not present a category below that of Good we have termed this group Light. The majority of our Ballintober swords fell into the Good category. No. 38, a particularly short example, ranged below this, and no. 23, an especially large broad bladed example, ranged above in the Medium category. Only one Chelsea sword placed in the Medium category, whilst the remaining swords ranged throughout the Light category. Bridgford (1997, fig. 6) lists six Irish Class 1 swords in her study and all fall within the Good category. No corpus was provided by Bridgford but as only three Class 1 weapons are possibly Type Chelsea it is likely that the swords studied were Ballintober, and therefore match our distribution within the different balance categories. Finally, much in the same way that we have with the balance point, it is necessary to establish the relative position of the centre of percussion. This has been identified as that part of the sword which, when swung, would yield the greatest velocity and weight, and is therefore that part of the

Steven Matthews

blade most likely intended to come into contact with another object (cf. Bridgford 1997, 103; Karasulas 2004, 510). Different from the balance point, the centre of percussion is located forward of or rather below this point, approximately a third of the length of the blade back from the tip and ideally corresponds with the widest part of the blade (Brewis 1923, 255, n. 2). Figure 15 shows that for every individual sword examined the centre of percussion was consistently found to correspond with blade unit Middle 3, with only one exception (no. A).

by the presence of twice as much edge damage on Rosnoën as Chelsea weapons, despite an even greater difference in blade form and the sample examined being half the size again. The greater distribution of edge damage on the latter weapons was more similar, however, compared to the more concentrated damage found on Ballintober swords. Although only one example was included in this study, it is of note that instances of damage on the Arco/Terontola sword displayed only edge damage of Classes 2 and 4 and Chelsea those of Classes 2 and 3.

Discussion of edge damage

CONCLUSION

It is hard to account for the notable differences between the occurrence and spatial distribution of edge damage on Chelsea and Ballintober swords. There is no appreciable difference in the length of the two different sword groupings, with the median length of all British Ballintober swords being c. 47.8 cm and the median length of Chelsea swords being c. 46.1 cm. However, as already discussed, the overall blade proportions between the two groups differ (fig. 4), and clearly relate to their differing positioning within the balance categories in relation to their cm/gms ratio. Although slightly more than half the number of Chelsea swords were examined compared to Ballintober swords, it is clear that the occurrence of less edge damage on this group cannot be explained simply as a consequence of sampling, especially as there is almost twice as much edge damage on the Rosnoën rapiers compared to Chelsea swords, despite the sample size again being only half as big. The greatest degree of edge damage on the Ballintober swords, whilst not corresponding exactly to blade unit Middle 3, the location of the centre of percussion, the highest number of instances of edge damage at least occur consistently on the lower half of the cutting blade, on blade units Middle 3, Lower 1 and Lower 3. The spatial distribution of edge damage found on Chelsea swords, however, does not correspond at all with that found on Ballintober swords. On the former group almost no edge damage occurs at all on the Lower part of the blade. Instead, the highest number of instances of edge damage corresponds with blade unit Middle 1, which relates not to the centre of percussion but rather the balance point of these weapons. Despite the blade proportions of Chelsea swords being only slightly different from Ballintober weapons, it is highly likely that this significant quantitative difference in edge damage was a product of technique, as indicated

In some way, both British and Irish writers have been correct in their differing interpretations of the origins of Chelsea and Ballintober swords. The three step model suggested here accommodates the basic divergent views on the subject, with both rod-tanged Arco/Terontola and Hemigkofen swords playing some role in their development. It seems to this author that the swords of Chelsea type do in fact stand at the head of the Ballintober family proper, closely followed by those of Ballintober type with their lozenge section (cf. Burgess 1968b, 15 f). It is this later type that was then transmitted to Ireland, and which eventually incorporated ricasso notches. There are clearly notable regional differences in the three series of swords. In France notched tangs, although more numerous amongst this small group of Continental swords, are equally numerous amongst the Irish and southern English groups. These notched tangs, however, only occur on Ballintober swords and never on Chelsea weapons, whilst blade outlines and primitive ricasso notches can be considered a late rather than typological trait amongst Ballintober swords. Short, slim line, sinuous blades are characteristic of Irish Class 1 swords. A notably trapezoidal tang, widest closest to the blade and narrowing toward the terminal, is also common in Ireland and France but not in England. There is a noticeable regional difference in the positioning of the rivets on these trapezoidal tangs, being more traditionally positioned in Ireland but situated at opposite ends of the tang in France, with the lower set often being notched. This suggests that, as valid as we believe the above discussion of variants to be to the wider series, there is a need to develop local manifestations of these variants for each region. For example, there appears to be no obvious variant Twickenham swords in Ireland, and the few probable Chelsea swords that are present

103

Chelsea and Ballintober swords: Typology, chronology and use

are all of very late form given the shape of the tang. There is no reason to assume that any of the Irish Class 1 swords need be significantly earlier than Ha A2. Similarly, it seems likely that variant Nantes developed later rather than earlier. Early Urnfield swords of Type Hemigkofen will have arrived in northern France first and likely upset development of Chelsea and Ballintober series much earlier than in England, by which time the series had already taken hold. The Irish series appears to be entirely dominated by swords similar to variants Mixnam and Battersea, and various hybrids of both. Ricasso notches are few in France compared to Ireland and we can therefore assume, given that so few early flange-hilted swords have been found in Ireland, that they survived latest there. As for the edge damage, only the briefest and most tentative interpretation can at this point be offered. The devil, as they say, is in the details, and the subtle differences in both the numerical and spatial distribution of the classes of edge damage found on each of the weapon types discussed above suggests that the typological differences alluded to, whilst significant, provide only a partial picture of use. Instead, it is those techniques, those learned ways of doing things (cf. Mauss 1973), that will likely provide greater interpretative depth in this respect. This paper has sought to lay the foundations for such an approach, based on comparison, which although lies beyond the scope of the current paper, the author intends to explore elsewhere at greater length (Matthews, in prep.). BIBLIOGRAPHY Bradley, R. 1998. The Passage of Arms (2nd edition). Oxford, Oxbow. Brewis, W. P. 1923. The bronze sword in Great Britain. Archaeologia XXIII, 253–265. Bridgford, S. D. 1997. Mightier than the pen? (An edgewise look at Irish Bronze Age swords). In J. Carman (ed.), Material harm: Archaeological Studies of War and Violence, 95–115. Glasgow, Cruithne Press. Bridgford, S. D. 2000. Weapons, Warfare and Society in Britain 1250–750 BC. Unpublished PhD thesis, University of Sheffield. Bridgford, S. D. 2001. Late Bronze Age swords and spears from the Power Station and Flag Fen. In F. Pryor (ed.), The Flag Fen Basin: Archaeology and Environment of a Fenland Landscape, 309–317. London, English Heritage. 104

Briard, J. 1965. Les Dépôts Bretons et l’Age du Bronze Atlantique. Rennes, Travaux Rennes: Travaux du laboratoire ‘Anthropologie – Préhistoire – Protohistoire – Quaternaire Armoricains’, Université de Rennes. Brown, M. A. 1982. Swords and sequence in the British Bronze Age. Archaeologia 107, 1–42. Burgess, C. 1968a. The Later Bronze Age in the British Isles and North-Western France. Archaeological Journal 125, 1–45. Burgess, C. 1968b. Bronze Age dirks and rapiers as illustrated by examples from Durham and Northumberland. Transactions of the Architectural and Archaeological Society of Durham and Northumberland 1, 3–26. Burgess, C. 1968c. Bronze Age Metalwork in Northern England, c. 1000 to 700 BC. Newcastle, Oriel Press. Burgess, C. 1974. The Bronze Age. In C. Renfrew (ed.), British Prehistory: A New Outline, 165– 232, 291–329. London, Duckworth. Burgess, C. 1980. The Bronze Age in Wales. In J. A. Taylor (ed.), Culture and Environment in Prehistoric Wales. British Archaeological Reports, British series 76, 243–286. Oxford, BAR. Burgess, C. B. and Gerloff, S. 1981. The Dirks and Rapiers of Great Britain and Ireland. Prähistorische Bronzefunde, IV. 7. München, C. H. Beck. Coffyn, A. 1985. Le Bronze Final Atlantique dans la Peninsula Iberique. Paris, Publications du Centre Pierre Paris no. 11. Colquhoun, I. and Burgess, C. B. 1988. The Swords of Britain. Prähistorische Bronzefunde IV, 5. München, C. H. Beck. Coombs, D. G. 1972. Late Bronze Age metalwork in the south of England: Typology, associations, distribution, chronology and industrial traditions. Unpublished PhD thesis, University of Cambridge. Cowen, J. D. 1951. The earliest bronze swords in Britain and their origin on the Continent of Europe. Proceedings of the Prehistoric Society 17, 195–213. Cowie, T. and O’Connor, B. 2007. Late Bronze Age swords from Scotland: Some finds old and new. In C. Burgess, P. Topping and F. Lynch (eds), Beyond Stonehenge: Essays on the Bronze Age in Honour of Colin Burgess, 316–344. Oxford, Oxbow Books. Eogan, G. 1965. Catalogue of Irish Bronze Swords. Dublin, National Museum of Ireland. Gerloff, S. 2007. Reinecke’s ABC and the chronology of the British Bronze Age. In C. Burgess, P. Topping and F. Lynch (eds), Beyond Stonehenge:

Steven Matthews

Essays on the Bronze Age in Honour of Colin Burgess, 117–161. Oxford, Oxbow Books. Gerloff, S. 2010. Atlantic Cauldrons and Buckets of the Late Bronze Age and Early Iron Ages in Western Europe. Prähistorische Bronzefunde II, 18. Stuttgart, Steiner. Gomez, J. 1987. Les épées du Cognaçais (Charente) et la chronologie des épées du type de ChelseaBallintober en France. In J-C. Blanchet, G. Bailloud, J. Briard, C. Burgess, G. Gaucher, J-P. Mohen and C. Mordant (eds), Les relations entre le continent et les Iles Britanniques à l‘âge du Bronze. Congrès Actes du colloque tenu dans le cadre du 22ème Congrès Préhistorique de France (1984), 125–131. Amiens, Supplément a la Revue Archéologique de Picardie. Gosden, C. and Marshall, Y. 1999. The cultural biography of objects. World Archaeology 31 (2), 169–178. Hodges, H. W. M. 1956. Studies in the Late Bronze Age of Ireland: 2. the typology and distribution of bronze implements. Ulster Journal of Archaeology 19, 29–56. Karasluas, A. 2004. Zaimokuza reconsidered: The forensic evidence, and classical Japanese swordsmanship. World Archaeology 36 (4), 507–518. Lawson, A. J. 1985. An unusual Bronze Age sword from Highclere, Hampshire. Proceedings of the Hampshire Field Club and Archaeological Society 41, 281–284. McArdle, T. D. 1969. Personal Armament in Middle and Late Bronze Age France. Unpublished PhD thesis, Edinburgh University. Matthews, S. 2010. Notes on a typological scheme for Atlantic Rapiers in France. Association pour la promotion des researches sur l’âge du bronze (APRAB) bulletin 7, 82–85. Matthews, S. 2011. Wrapping bronzes: Pottery encased metalwork in southern England and northern France during the Late Bronze Age. Association pour la promotion des researches sur l’âge du bronze (APRAB) bulletin 8. Matthews, S. in prep. Techniques and Society: the Middle and Late Bronze Age weapons of northwest Europe. PhD thesis, University of Groningen.

Mauss, M. 1973. Techniques of the body. Economy and Society, 2, 70–88. Needham, S. 1982. The Ambleside Hoard: A Discovery in the Royal Collections. British Museum Occasional Paper no. 39. London, British Museum. Needham, S. 1990. The Penard-Wilburton succession: new metalwork finds from Croxton (Norfolk) and Thirsk (Yorkshire). Antiquaries Journal 70 (2), 253–270. O’Connor, B. 1980. Cross-Channel Relations in the Later Bronze Age: Relations between Britain, North-Eastern France, and the Low Countries during the Later Bronze Age and the Early Iron Age, with Particular Reference to the Metalwork. British Archaeological Reports, International series 91. Oxford, BAR. O’Connor, B. 1989. The swords of Britain: Ian Colquhoun and Colin Burgess. Scottish Archaeological Review 6, 126–128. Rowlands, M. J. 1976. The Production and Distribution of Metalwork in the Middle Bronze Age in Southern Britain. British Archaeological Reports, British series 31. Oxford, BAR. Rowlands, M. 1998. Kinship, alliance and exchange in the European Bronze Age. In K. Kristiansen and M. Rowlands (eds), Social Transformations in Archaeology: Global and Local Perspectives, 142–177. London, Routledge. Sandars, N. K. 1976. Review: Hartmann Reim, Die spätbronzezeitlichen Griffplatten-, Griffdorn- und Griffangelschwerter in Ostfrankreich. Germania 54 (1), 237–239. Savage, R. D. A. 1979. Technical notes on the Watford sword fragments. In C. Burgess and D. Coombs (eds), Bronze Age Hoards: Some Finds Old and New. British Archaeological Reports, British series 67, 221–228. Oxford, BAR. Tomalin, D. 1982. Bronze Age swords of the Ballintober type found at Mixnams Pit, Thorpe, Surrey. Surrey Archaeological Collections 73, 163–167. Trump, B. 1962. The origin and development of Middle Bronze Age rapiers. Proceedings of the Prehistoric Society 28, 80–102. Steven Matthews Groningen Institute of Archaeology (GIA) University of Groningen Poststraat 6 9712 ER Groningen The Netherlands

105

106

IRISH SWORDS: USE AND ABUSE Ian Colquhoun

ABSTRACT

PREVIOUS RESEARCH AND TYPOLOGY

There are over 600 Late Bronze Age swords from Ireland, each with a different story to tell. To many archaeologists these weapons epitomise the warlike nature of a period dominated by a new aggressive technology. This paper discusses the relevant typology, examines the survival of these prehistoric weapons, and then discusses the reasons why so many swords show damage to the hilt, blade edges or the blade.

The Bronze Age swords of Ireland were drawn and researched by G. Eogan for his doctoral thesis and published as a catalogue by the National Museum of Ireland in 1965. This was a large undertaking, and, almost half a century later, the catalogue remains the basis for any study of these weapons. Eogan divided the swords into six classes, a typology which continues in use for those working on the Bronze Age in Ireland. There are over 600 swords or sword fragments in the catalogue. To guide those less familiar with the typology, I summarise the different types below, and equate them both with the typology in use for the British swords and types defined by earlier authors. Terminology will be considered first. Features are described starting with the hilt. On the very earliest swords this consisted of a simple tang, which developed into a flanged hilt section surmounted by a splayed terminal. Organic plates, made of wood, horn, or bone were then riveted to the hilt using bronze rivets through holes or slots in the hilt and shoulders. In addition globular or hat like pommels were attached to the terminal. These too were made of organic material; evidence for this comes in the form of surviving cast bronze pommels and grips from Britain and Europe, but none, unfortunately, from Ireland. One find of interest here is the wooden sword from Cappagh, Co. Kerry (Waddell 2000, fig. 131. 5) with its integral carved pommel. Late Bronze Age Irish swords did not have a hilt guard; instead, the shoulders splayed out to form a U or, more commonly, a V shape. Between the shoulders and the top of the blade edge the sides are often indented or flattened (the ricasso), sometimes with the top of the blade edge forming a clear notch (the ricasso notch). This feature can also be seen on later swords from historical periods as well as some modern knives, and is designed as a safety issue either to stop the user’s hand sliding onto the blade or to catch and slow down the opponent’s blade before it damages the user’s hand. Doubtless it had the same function in prehistory. The blade forms the major part of any sword, and extends from the ricasso to the point. The majority of Late Bronze Age swords from

KEYWORDS Late Bronze Age – swords – edge damage – hilt damage – sword typology INTRODUCTION Holding a bronze sword made around three thousand years ago is a magical experience. Each particular weapon has its own history, its own story of manufacture, use and deposition – of its birth, its life and its death. Each carries the individual marks of the parts it played within the lives and the deaths of people and communities in a distant Ireland of the Late Bronze Age (c. 1200–700 BC). Each would have been recognised by its maker and the owner, or owners, who may have carried it around with them every waking hour, and guarded it jealously during their sleeping hours. This paper explores some aspects of the evidence for the use – the life – of these bronze swords of the Late Bronze Age in Ireland and examines the possible reasons why they show particular and commonly occurring types of damage. It also examines why swords made three thousand years ago should survive to find their way into modern collections. I am currently undertaking research on these weapons, and hope to draw and record every example, especially those discovered since the 1960s. Observations discussed here are based largely on those swords I have already drawn and examined at the National Museum of Ireland in Dublin and the National Museum of Scotland in Edinburgh. There are many more yet to see as this is very much work in progress.

107

Ian Colquhoun

throughout western Europe have leaf shaped blades, with the widest part of the blade around two thirds down from the terminal. From here the blade narrows to a point. Swords were worked or annealed to produce a sharp edge defined by a bevel. The exact shape of the blade and the bevel would have been influenced by the preferences of the owner and later resharpening. Blades sometimes have a deviation to left or right, either as an accidental feature or a preference of the user (Kristiansen 2002, 320). Blade sections vary from lozenge shaped in the earlier swords to lenticular (or elliptical) in the majority of flange hilted swords. Eogan’s Class 1 equates with the Ballintober type, first defined and mapped by H. W. M. Hodges (1956, 37). These are tanged swords with wide shoulders. The tangs are lozenge shaped in cross section and flangeless, similar in shape to those of earlier rapiers. Four rivets would have secured organic hilt plates to the tang, although no such plates have survived for this type. The blades are leaf shaped with either a lozenge shaped, flattened lozenge or lenticular section. Interestingly, the distribution is split largely between the Thames valley and the northern half of Ireland (Hodges 1956, 32 fig. 3: Colquhoun and Burgess 1988, pl. 118, 119; Waddell 2000, fig. 82) suggesting close links between the two areas during the earlier part of the Late Bronze Age. Eogan’s Class 2 encompasses the few early flange hilted swords found in Ireland. These have characteristic wide convex ‘U shaped’ shoulders and wide leaf shaped blades with a lozenge cross section. Their clear continental ancestry was recognised by W. P. Brewis (1923) and J. D. Cowen (1951). Many of the continental blades carry ornamentation, in the form of inscribed lines or milling, notably either below the shoulders or on the ricasso. Eogan’s Class 3 swords have more pronounced and graceful leaf shaped blades with lenticular sections. The shoulders are narrower than Class 2 and ‘V shaped’. The hilts are flanged and the rivets were secured to the sword through slotted hilts or large holes. They can be compared with the British Wilburton type, most common in south eastern England (Burgess and Colquhoun 1988, pl. 124–126). These are clearly ancestral to the Ewart Park type, a development seen most clearly in the swords in the Blackmoor hoard (Colquhoun 1979). In Ireland swords of Class 3 are not as numerous as their British equivalents. Most Irish swords – 440 out of the 624 listed by Eogan – fall within his definition of Class 108

4. These are paralleled closely in Britain by the Ewart Park type first recognised by Cowen (1933). Examples from throughout Ireland can be compared directly with the hundreds from England, Scotland and Wales, where they also appear to have formed the standard weapon for the Late Bronze Age warrior. While there are regional differences the basic style is strikingly homogeneous. The terminal is always splayed. Hilts are generally bowed, and are normally narrower below the terminal than at the top of the shoulders. Hilt edges are always flanged, and in cross section the hilt often shows strengthening ribs or projections. Rivet holes normally number two or three on the hilt and two to six on the V shaped shoulders. There is almost always a ricasso, either straight or slightly concave. The absence of a ricasso on any weapon is likely to be due to extreme wear or post deposition corrosion. Blades are always undecorated and leaf shaped with a lenticular cross section. One particularly noticeable feature in Ireland is that the blades vary widely in length, with many being noticeably short. For example, the complete and unaltered sword from Lough Gur, Co. Limerick (Eogan 1965, no. 208) is only 43.5 cm long, while the sword from Inishmore, Co. Cavan (Eogan 1965, no. 545), is just under 46 cm long (fig. 4). By comparison, few Ewart Park swords with a verifiable British provenance are shorter than 50 cm. Eogan subdivided Class 4 into four subdivisions, with Classes 4c and 4d showing some Hallstatt influence as described in Class 5 below. Similar influences can be seen on Ewart Park swords of Step 4 in south east England (Colquhoun and Burgess 1988, 74 f.). This development was first recognised by Cowen (1967, 412) as a devolved version of the Ewart Park series and called by him the ‘Thames Type’. Class 5 comprise the classic Hallstatt C swords found throughout western and central Europe as well as the Atlantic fringes. These are often beautifully made and finished, with neatly defined blade edges. Cowen (1967, 391) described them as the ‘Hallstatt sword of bronze par excellence’. The terminals are normally notched and have rivet holes, probably designed to hold large, possibly hat shaped, pommels. Hilts are wide, bowed and flat or dished in section, with only slight flanges. Rivet holes are generally small and, where the rivets survive, these are often pin like, with decorated heads, and less than 2 mm in diameter. Shoulders are more widely splayed than on Class 4 swords and may be straight or slightly concave. The ricasso is always concave,

Irish swords: Use and abuse

Figure 1: Unprovenanced, probably Ireland (National Museum of Ireland, no. R372) – Figure 2: Knockadoo, Co. Roscommon (National Museum of Ireland, no. W86) – Figure 3: Meelick, Co. Roscommon (National Museum of Ireland, no. 1985.88) – Figure 4: Inishmore, Co. Cavan (National Museum of Ireland, no. 1937.3647) – Figure 5: Keelogue Ford, Co. Galway (National Museum of Ireland, no. W47) 109

Ian Colquhoun

and the top of the blade edge often forms a distinctive prominent swept back notch, in itself a decorative feature. A further distinguishing feature is the defined central rib extending from the upper blade to just below the lowest rivet hole on the hilt. These are long swords by the standards of the time; the sword from near Athlone, Co. Westmeath (Eogan 1965, no. 532), for example, measures 75.6 cm. The leaf shape of the long blades is not only less pronounced than Class 4, but also narrower. The maximum blade width is nearer the hilt, and often noticeably less than the shoulder width. The bevel edges are frequently well defined with raised beading and the blade section is generally slightly flattened along the centre, though some Irish swords have a simpler cross section with wide bevels. Although the point bevels are always blunted the main mass of the blade normally comes to a clearly defined point. Class 6 swords are basically weapons with wide Class 4 blades and hilts with features closer to Class 5. These form a small group, the main distinguishing difference between Class 4c and Class 6 being that the latter have shoulders that are more splayed and even concave. SWORD SURVIVAL Before looking at what sword damage signifies, it is worth considering the possible reasons why, firstly, the inhabitants of Late Bronze Age Ireland should have placed their prized weapons within a river, a bog or the soil, and why, secondly, these swords should have been recovered within the last two hundred years or so. There are more swords per square kilometre with an Irish provenance or labelled ‘Ireland’ in museums and collections throughout the world than from any other European country, as many as 7.61/km2 as opposed to 2.87/km2 for Britain as a whole (Harding 2000, tab. 8. 1). Precise find spots are, however, often difficult or impossible to pinpoint. Of those with a known provenance to some reasonable degree of accuracy, around 35 % are from rivers or loughs and about 25 % are from bogs, indicating that the majority of weapons have come from either ‘wet’ locations or locations near to water. Of those from rivers and loughs, many if not most were found during dredging designed to improve the navigation of a relatively small number of rivers such as the Shannon, the Bann and the Erne. Published distribution maps clearly show the concentration of swords in these locations (Eogan 1965, fig. 110

84; Cooney and Grogan 1994, fig. 9. 1; Waddell 2000, fig. 132). There is no reliable record of any Irish sword being found in association with a burial. The paucity of burials in the archaeological record for this period has been discussed elsewhere (Mallory and McNeill 1991, 138; Cooney and Grogan 1994, 144) and it is conceivable that cremated bone and ashes may simply have been placed in rivers or on the fringes of settlements, possibly with bronze metalwork. Whether deposited with or without a cremation, the placing of a sword or swords in a chosen place would have been, in most cases, a deliberate act with the objects permanently destined for the earth. Placing a sword or swords in the ground or in water at a particular location is a practice documented throughout Ireland and Britain. For example, deliberate permanent deposition was identified at Flag Fen in Cambridgeshire where Wilburton and Ewart Park swords were placed in the ground at the edges of settlements (Pryor 2003). Placing bronzes at liminal locations of contemporary significance may be the reason why there are so many swords from rivers and other possible boundary or prominent features within a local landscape. It must be stressed that the landscape of Bronze Age Ireland, like anywhere in western Europe, was very different from that of today. Rivers followed their natural courses and were left to flood without artificial impediment. Many riverside hamlets or small settlements on islands would have disappeared as the river changed course. One area which flooded regularly in the past is Toome, on the boundary of counties Antrim and Derry, where the Lower Bann drains Lough Neagh northwards towards the coastline of Ulster. At least 13 swords, of Classes 2, 4 and 5, have been found here, at or near spots known to be fording points prior to the dredging which took place during the nineteenth and twentieth centuries (Bourke 2001, 20). As well as marking boundaries, such fording points would have been places of local or regional political significance, important trading route ways, and perhaps centres of population. Today Toome is at the juncture of three modern administrative areas. Fording points may also have been scenes of conflict, as they were in historical times. Such an idea initially suggests large scale battles rather than the skirmishes that are normally envisaged as the everyday reality of Late Bronze Age warfare. This explanation of ‘battle loss’ should not be dismissed completely, although retriev-

Irish swords: Use and abuse

ing weapons post conflict from shallow fording places may not have been difficult. Such clashes may have been as relevant in prehistory as the Battle of the Boyne has been for the last three hundred years. Deliberate deposition could have been designed to be temporary. It would be possible to literally drop a sword, point first, into a bog at a location measured from a fixed landscape feature, with the intention of recovering it. Such a sword may even have been interred in a similar fashion many times over its life, perhaps as part of a ritual for the death of its owner or change of ownership for some other reason. Deposition for intended retrieval would have taken place in a landscape that was embedded in peoples’ minds and where specific features, perhaps as simple as a small island within a bog, held meaning (Needham 2001; Becker 2007). The location of such a deposition, whether ritual or for recycling, could become forgotten or the sword may not have been retrieved for some other reason. The same fate could have been suffered by bronze intended for recycling. Bronze survives remarkably well in conditions where organic material or iron deteriorates or disappears completely. Doubtless swords were recovered by farmers and turf cutters (and melted down) over the centuries, but it was only with the rise of interest in the distant past that the antiquity of these weapons was realised. Many, if not most, nineteenth century finds were by manual workers, whether in the fields, cutting turf, or dredging rivers such as the Shannon and Bann for ‘improving’ landlords and agencies. Whether or not they found their way to private or museum collections depended greatly on local contacts, and these varied tremendously throughout Ireland. Nineteenth century antiquarians such as William Wilde, father of Oscar, collected widely from a number of sources. Dealers were, if not unscrupulous, often cavalier with their attributions of provenances, and many collectors were more interested in the mere possession of an attractive weapon or its typology than in the circumstances of discovery. Thus many of the swords in the Wilde collection and others in the National Museum are simply labelled ‘Ireland’ and it is unlikely that any amount of diligent research will shed light on their recovery. The decrease in finds from such activities since the start of the twentieth century must largely be due to increased mechanisation. In recent years metal detectorists have been responsible for finds such as the Class 3 sword and associated decorated bronze bowl

from Tamlaght, Co. Antrim (Warner 2006). The find spot was excavated following discovery, a rare example of the context of a sword find being recorded. SWORD USE Few swords have been found in complete, pristine condition. Most have some degree of damage or breaks to the hilt or blade or both. On many swords such as the weapon from Inishmore, Co. Cavan (fig. 4) the hilt has broken off above the shoulders across the rivet holes or the blade has snapped at its narrowest point. On others, such as the swords from ‘Ireland’ (fig. 1) and Knockadoo, Co. Roscommon, (fig. 2), the blade has snapped at the narrowest width. Large numbers of swords also show damage to the extremities, either the point (the sword from Meelick, Co. Roscommon, fig. 3), the shoulders or the terminal. There are a number of potential reasons why this breakage should occur. These can usefully be summarised as through use and wear in prehistory, by deliberate damage or breakage prior to deposition, by corrosion, or by accidental damage after the sword had been placed in the ground, for example by the plough or by the actions of the discoverer. The concern here is with what happened to these swords in prehistory. From ‘birth’ to ‘death’ of a weapon, the possible causes for damage can be summarised as production error, use in conflict, use in practice, use in ritual, deliberate damage prior to deposition, and breaking up for recycling. Production error The successful manufacture of a bronze sword involved a huge amount of skill and experience. Modern casters of replica weapons are acutely aware that it is not until the clay mould is broken open that it becomes possible to know whether or not the whole casting procedure has been successful. There are variables such as the quality of the copper and tin, the quality of the mould materials, and so forth. However, it seems that not all swords that appear imperfect were destined for recycling. Casting flaws such as small holes where air has entered the alloy during cooling are often visible on swords which have all the appearances of being used. This suggests that imperfect weapons were acceptable, presumably depending on the status of

111

Ian Colquhoun

the intended owner and, perhaps, whether the flaws could be hidden under the details of the hilt plates, scabbard or chape. Use in Conflict A bronze sword was designed as a weapon intended to kill and maim. A warrior using a sword in combat would have been pitted against one or more adversaries equipped with swords, shields and spearheads. It seems unlikely that a warrior would deliberately wield his sword against the sword of his enemy in the manner of modern sword fighters, as he would be well aware that hitting these weapons would easily damage or weaken his own sword and thus endanger his own life. It is more likely that his aim would be to injure or kill his opponent, and that damage was caused when the opponent’s bronze weapons got in the way. Experimental work with replica swords (Bridgford 1997; Molloy 2004; 2007; this volume) has shown that hitting one weapon against another produces the type of edge damage visible on the majority of Irish swords examined to date. This generally takes the form of nicks of varying sizes along the bevelled blade edges, especially at the widest part of the blade. Small nicks are visible on the blade edges of the sword from Keelogue Ford, Co. Galway (fig. 5). Early medieval iron swords, with steel edges, were capable, if written sources are to be believed, of slicing through human bone from shoulder to waist (Thompson 2004, 38). B. Molloy’s experiments with replica bronze swords and pig carcasses in a Dublin Food Research Centre (Molloy 2007) could not manage a comparable feat, but the blows did cut well into the bone. In doing so the blade edges were buckled. Damaged bronze spearheads have been found embedded in human skeletons. At West Littleton Down, Tormarton, Gloucestershire, two skeletons were found and excavated in 1968 in advance of a gas pipeline. One had the fragments of two spearheads embedded in the spinal cord and pelvis, while the other had a hole in the pelvis apparently caused by a spearhead. These two young men, aged around nineteen, had been placed without apparent ceremony in a ditch or pit; two further bodies were found in a subsequent excavation in 1999 (Knight, Brown and Grinsell 1972; Osgood 2000, 21 f.). An earlier find from Dorchester on Thames is that of a human skeleton where the pelvis had been pierced by a triangular bladed spearhead which had bro112

ken off in situ. Such was the force of the blow that tip had bent (Osgood 1998, 21). Given the amount of damage to these spearheads as the result of human contact, it is likely that at least some of the damage seen on Irish swords is as a result of such contact. Many swords show evidence of repair, especially where the hilt has broken across the rivet holes. New hilts were added, sometimes with strengthening ribs, presumably to help in the casting process and strengthen the repair. As this part of the sword would not be visible, it would not matter that the resulting repair was unsightly, as many were. Such a repair can be seen on the sword from Keelogue Ford, Co. Galway (fig. 5). Use in Practice Whether we view Late Bronze Age swordsmen as aristocratic warriors with time on their hands or as farmers and part time fighters, they would have needed to practise and improve their skills with their swords, spears and shields. To be effective modern combat training is designed to be as realistic as possible, using real weapons, which frequently get damaged, and soldiers do get injured. It follows that at least some of the damage and wear now visible on prehistoric swords is due to repetitive use of the weapons for practice. It is also likely that they would be handled for such training far more frequently than they were used in anger against an adversary. While wooden swords such as the example from Cappagh, Co. Kerry (Waddell 2000, fig. 13. 5) may have been used for training as suggested by K. Kristiansen (2002, 325) these replicas would never have been able to replicate the weight and feel of solid bronze. I believe, then, it is feasible to suggest that some training involved the use of real weaponry and it follows that some damage is due to use for practice. Use in ritual and deliberate damage prior to deposition It has already been suggested that swords served many purposes other than their use as killing or maiming weapons. They were, perhaps above all, weapons of prestige and as such would have played a role within the posturing and appearance of the warrior class or stratum of society. The physical use of such weapons would have played a role in ritual, perhaps equivalent to the modern practice of shooting into the air to cele-

Irish swords: Use and abuse

brate or merely displaying skills in front of an audience, perhaps composed of potential allies. A special role may have been played by Class 5 / Gündlingen swords. By comparison with Class 4 swords, these give the impression of being far more decorative and valuable. When they appeared in Ireland they would perhaps have been the most prized of all swords, a type of weapon carried only by those at the top of society and an aspiration to the rest. It is likely that it was due to their prestige that they had such an influence on the design of Class 4 swords. The very possession of such a weapon must have been a symbol of status within society, as it was in early historic times. How valuable a sword would be in the Late Bronze Age and what level of status it suggested to society at large is difficult to gauge as we do not know how many swords were in circulation at any one time and what the population could have been. F. Pryor (2003) looked at the numerous bronze weapons and fragments found during his own excavations at Flag Fen in Cambridgeshire and at another nearby site, Bradley Fen. Taking into account the finds at these two sites over twenty years or so he changed his opinion over the two decades and relegated the modern comparison from being the equivalent of an anti-aircraft missile to that of the essential equipment of every soldier. This comparison is backed up by the observation that Class 4 / Ewart Park swords give the impression, as a group, of being almost utilitarian, without decoration or embellishment. Perhaps they are best described as the equivalent of a modern rifle. Within the Wilde Collection in the National Museum in Dublin there are two Class 4 swords from a hoard found at Knockadoo in Co. Roscommon. The hoard, with two associated spearheads, has been published (Eogan 1983, hoard no. 126). The circumstances of discovery are unknown and it is likely that what remains was part of a larger hoard. One sword is in relatively good condition, with just a few nicks along the edges. The other has been bent and broken, and one edge in particular has suffered quite spectacular damage, with chunks almost gouged out of the edge (fig. 2). In short, it appears that the sword has been deliberately bent and damaged to put it out of use. Another weapon, unfortunately unprovenanced, (Eogan 1965, no. 496) has similar damage along both edges (fig. 1). A sword discovered in the 1980s from Meelick, also Co. Roscommon, has edges which are turned in and appear to have been hit with a blunt object (fig. 3).

Most Irish swords have not suffered this treatment. The same is true of most British swords, though examples of bent and damaged swords are not hard to find – for instance in the hoard from Blackmoor in Hampshire (Colquhoun 1979). The fact that swords with significant edge damage also seem to have been bent and broken suggests that this is part of a deliberate procedure. While the bending could be viewed as part of the procedure for breaking the sword for recycling, the excessive, almost frenzied damage to the top of the blade edge, below the ricasso, suggests a procedure planned in advance and carried for a purpose other than the utilitarian procedure of preparation for casting. In short, it suggests that the damage was part of the procedure for ending the life of this particular sword, whether because of its connection with an individual or because of some other role the sword played within the Late Bronze Age psyche. It is rare for organic hilt plates and pommels to survive. A notable exception was a bog find, a Class 2 sword from Lisletrim, Co. Monaghan (Evans 1881, fig. 358; Eogan 1965, no. 29). While the decomposition of organic hilt plates is understandable under many conditions, such as river deposition, it would not seem to be unreasonable to suggest that more should have survived as bog finds. Many swords are found without rivets. It is possible that the hilt plates and pommels were normally removed before deposition, as the hilt plates may have had a part to play in ritual or as an heirloom to be used on another weapon. Hilts would have been a conspicuous and possibly personalised part of the sword. This was certainly the case in Anglo Saxon England; the recently discovered Staffordshire Hoard contained dozens of ornamented hilt and pommel pieces which had been removed from the swords (Leahy 2009). Breaking up for recycling The majority of the Irish swords are, as outlined above, from rivers or bogs. Most are single finds, or there is no recorded evidence to indicate that they were found in association with any other bronzes or artefacts. Few hoards from Ireland appear to have been destined for the melting pot. Hoards composed of fragments of broken bronzes, common in the south east of England, are rare in Ireland, the one major exception being the Co. Roscommon hoard with its lozenge sectioned sword blade fragment (Waddell 2000, 201

113

Ian Colquhoun

fig. 81. 1). Because of the circumstances of acquisition, through the activities of a Victorian dealer, its status as an Irish hoard has been questioned (Eogan 1983, 49).

CONCLUSIONS The purpose of this paper has been to summarise the recent research into bronze swords in Ireland and to examine some aspects of their discovery and use. Whatever role luck played in the discovery and survival of any particular weapon, the sheer number of swords found, especially those of Class 4, suggests that there must have been plenty of bronze weapons around in the first centuries of the first millennium BC. There are swords yet to be found, and many which have been found and either melted down or disappeared from the public eye. Most swords appear to have suffered edge damage as the result of being used to hit another hard object. These weapons were designed for conflict and their primary use was in a combat role. Each sword may have served many other roles: as an object of desire, as a weapon intimately connected with a memory or ceremony; as a gift; as the booty of war. They were valuable but not that valuable; and when they had served their purpose many, along with spearheads, shields, and other bronzes, were given up to the earth. ACKNOWLEDGEMENTS Thanks are due to staff at the National Museum of Ireland, especially M. Lannin, and to T. Cowie at the National Museum of Scotland. While opinions expressed are my own, thanks are also due to my tutors at the University of Durham, M. Diaz-Andreu and T. Moore, for helping to put some structure into my thoughts. BIBLIOGRAPHY Becker, K. 2008. Left but not Lost. Archaeology Ireland 22 (1), 12–15. Bourke, L. 2001. Crossing the Rubicon. Bronze Age Metalwork from Irish Rivers. Galway, National University of Ireland. Brewis, W. P. 1923. The Bronze Sword in Great Britain. Archaeologia 73, 253–265. Bridgford, S. 1997. Mightier than the pen? (An edgewise look at Irish bronze swords). In J. Carman (ed.) Material Harm: Archaeological 114

Studies of War and Violence, 95–115. Glasgow, Cruithne. Colquhoun, I. A. 1979. The Late Bronze Age Hoard from Blackmoor, Hampshire. In C. Burgess and D. Coombs (eds), Bronze Age Hoards: Some Finds Old and New. British Archaeological Reports, British series 67, 99–106. Oxford, BAR. Colquhoun, I. A. and Burgess, C. B. 1988. The Swords of Britain. Prähistorische Bronzefunde IV, 5. München, C.H. Beck. Cooney, G. and Grogan, E. 1999. Irish Prehistory: A Social Perspective. Bray, Wordwell. Cowen, J. D. 1933. Two bronze swords from Ewart Park, Wooler. Archaeologia Aeliana 4th series, 10, 185–198. Cowen, J. D. 1951. The Earliest Bronze Swords in Britain and their origins on the Continent of Europe. Proceedings of the Prehistoric Society 17, 195–213. Cowen, J. D. 1967. The Hallstatt sword of bronze: on the continent and in Britain. Proceedings of the Prehistoric Society 33, 377–454. Eogan, G. 1965. Catalogue of Irish Bronze Swords. Dublin, The Stationery Office. Eogan, G. 1983. The hoards of the Irish Later Bronze Age. Dublin, University College. Evans, J. 1881. The Ancient Bronze Implements, Weapons, and Ornaments of Great Britain and Ireland. London, Longmans, Green and Co. Harding, A. F. 2000. European societies in the Bronze Age. Cambridge, Cambridge University Press. Hodges, H. W. M. 1956. Studies in the Late Bronze Age in Ireland: 2. The Typology and Distribution of Bronze Implements. Ulster Journal of Archaeology 19, 29–56. Knight, R. W., Browne, C. and Grinsell, L.V. 1972. Prehistoric Skeletons from Tormarton. Transactions of the Bristol and Gloucestershire Archaeological Society 91, 14–17. Kristiansen, K. 2002. The tale of the sword – swords and swordfighters in Bronze Age Europe. Oxford Journal of Archaeology 21 (4), 319–332. Leahy, K. 2009. Trophies of Kings. Current Archaeology 236, 12–21. Mallory, J. P. and McNeill, T. E. 1991. The Archaeology of Ulster. Belfast, The Institute of Irish Studies, Queen’s University. Molloy, B. 2004. Experimental Combat with Bronze Age weapons. Archaeology Ireland 17/4 (66), 32–34. Molloy, B. P. C. 2007. What’s the bloody point? Bronze Age swordsmanship in Britain and

Irish swords: Use and abuse

Ireland. In B. P. C. Molloy (ed.), The Cutting Edge: Studies in Ancient and Mediaeval Combat, 90–111. Stroud, Tempus. Needham, S. 2001. When expediency breaches ritual intention: the flow of metal between systemic and buried domains. Journal of the Royal Anthropological Institute 7 (2), 275– 298. Osgood, R. 1998. Warfare in the Late Bronze Age of Northern Europe. British Archaeological

Reports, International series 694. Oxford, BAR Publishing. Pryor, F. 2003. Britain BC. London, Harper Collins. Thompson, L. 2004. Ancient Weapons in Britain. Barnsley, Pen and Sword Military. Waddell, J. 2000. The Prehistoric Archaeology of Ireland. Bray, Wordwell. Warner, R. 2006. The Tamlaght Hoard and the Creeveroe Axe. Emania 20, 20–28. Ian Colquhoun Department of Archaeology Durham University South Road, Durham, DH1 3LE Great Britain

115

116

INTEGRATING FORM, FUNCTION AND TECHNOLOGY IN ANCIENT SWORDS. THE CONCEPT OF QUALITY Marc Gener ABSTRACT

FORM, FUNCTION AND TECHNOLOGY

Weapons are, down to the bare basics, specific tools (form) designed to facilitate a specific task (function) and built accordingly to the available possibilities (technology). Beyond this indispensable but clinical analysis, the integration of these objects within their historical context generates a generally high number of more or less complex questions. Using some samples of Iberian carp’stongue swords from the Ría de Huelva hoard as a case study, we will present a discussion about how a relationship between these concepts can help us to approach such an apparently subjective idea as ‘quality’: when an object can be considered ‘good’ or ‘bad’. Also, given the frequency to which an opinion on that regard is passed on in descriptions and analysis of elements of the material culture, the question to what degree this perception can be evaluated from an archaeological point of view is presented and discussed. ‘

When approaching the study of a particular weapon, we start with the object itself. With the idea that, like a natural object is the product of a natural environment, a manufactured object is the product of a human and social environment, and that this environment leaves a ‘fingerprint’ on the object. A fingerprint that is not only traceable, as a consequence of its use, but also technological, as a consequence of the use that was allegedly intended for it, and of the subsequent technical decisions that have been taken, in each case, in its process of manufacture (Matthews 2005). Going down to a simplified version of the situation, we can say we start our study with an object. With luck, it comes with an archaeological context, and even maybe with a chronological and geographical attribution. Then, we proceed to examine the object with all the techniques at hand. We study its form (shape, decoration, etc.) via taxonomy, and try to compare it with other similar objects, or with iconographical evidence, if available. If the right technical equipment and knowledge are available, we try to obtain information, via archaeometry, about the material or materials it is made of, and the technological processes applied to it. With all of this, we might end up with a reasonably good idea about HOW the object was made. But speaking from a very general point of view, an object is a tool. A mean to accomplish an end. This end is what defines its function. And its function is what is going to help us tell WHY the object was made the way it was made. In short, from the study of the form we can answer the ‘hows’, which can be verified through experimentation with replicas. Leading on from the study of its possible function, we want to answer the ‘whys’. And both are necessary to better understand, in the wider sense of the word, our object. Furthermore, looking at the question from the point of view of the human being who produced the object, and, again, simplifying, we can say that they faced a problem, a task that had to be fulfilled. And we can say that the technology that was available at a certain geographical and chronological point represented the means the creators of the object had at their disposal in order to actually fulfil the intended task. In a broad sense it includes as much material potential to do

KEYWORDS Swords – quality – functionality – technology – metallurgy – Late Bronze Age – Ría de Huelva INTRODUCTION The objective of the talk given at the ‘Warfare in Bronze Age Europe: Manufacture and Use of Weaponry’ meeting was to offer a brief personal perspective about weapons as objects of the material culture, and how our research can be affected by how they were perceived by those who designed, made and used them. It does not pretend to bring to light anything really new, but to summarize, what many people are already doing, more or less consciously. Hopefully, along the way we will be introducing one or two concepts that may have some use, even if it is only to stir some discussion. Many simplifications and generalizations are consciously introduced in order to provide a framework to convey the argument. Swords are used here as a paradigmatic example, but a lot of what is presented in this work can be applied to any object of the material culture, with the logical adaptations.

117

Marc Gener

something as well as the knowledge about how to do it. These three parameters (form, function and means) are inextricably linked, and one cannot be properly understood without the others, not without missing a big part of the picture. Frequently, when the answer to what the function was is not satisfactorily reached, there is a certain tendency to consider the object as ‘ceremonial’, out of an exclusion process. This phenomenon has had a special incidence in the field of weaponry, where in the early stages of these studies, speculation about functionality was barely ever supported by any kind of experimentation or real familiarity1 with weapons. The subsequent conclusions became quickly dogmatic. Of course, this does not mean, in any way, that the very important ceremonial and symbolic aspects of weaponry should be disregarded. They must not. It is just an historical question of degree, oversimplification and of abuse of the term. At the end, the object we are dealing with becomes, in a certain sense, the material part of the solution that was devised in order to solve the problem, given certain circumstances and certain means. But we do not want just a part, we want to know the whole of the solution. Also, we want to know the problem. And, also, we want to know as much as possible from the context that actually generated the problem and that influenced the decisions that led to that solution. If we go back to our case study, the sword, it would seem at first that the problem that must be solved is just the taking of a fellow human being’s life; and that the sword is the answer to this specific problem. But that is not the case, of course. The contextual factors that influence both the problem and the solution are multiple, and must be considered. For example, the actual performance is done in a context: we have shields, body protection, etc. There are also different types of combat (collective, individual, ritual, etc.) with different sets of needs. The performance of the tool is also product of factors like the level of its contemporary technology, the degree of availability of materials or the skill of the smith not to mention the skill of the actual performer. In addition to that, there are even more intangible features to be considered: social aspects of the object, the context of the action, symbolism, etc.

The problem, in short, has a global nature. The solution, inevitably, will, too. In consequence, the approach must also be global.

1 Even some potential exceptions, as could be the case of Richard Burton’s ‘The Book of The Sword’, are more an evidence of the prejudices of the author regarding the use of weapons than an example of an epistemologically coherent application of this knowledge to the functional

analysis of those weapons the author is not directly familiar with. 2 See, just as a small sample of different approaches and results: Bridgford 2000; Kristiansen 2002; Mödlinger and Ntaflos, 2007; Molloy 2008.

THE SWORD It is in this context that we approach the study of our particular subject of choice: the sword. Following what has been introduced in the previous section, we consider this object as being the result of its functional, and technological aspects, as well as of its contextual (we can call them ‘social’) aspects, of which, of course, neither the functional nor the technological aspects can be considered independent. We apply our knowledge to it in order to extract as much information as we can of these aspects with all the tools available: taxonomy, materials analysis, technological studies, usewear analysis, experimental archaeology, etc. With all of this, hopefully, we will get our ‘How’, or at least a reasonable approximation. Then we will need to find our ‘Why’. Now, the concept of ‘Function’ has, in fact, a pretty wide meaning. The sword has, unquestionably, a strong symbolic aspect, but we will start by concentrating, in first instance, on the martial aspect. The initial consideration is that the sword is a tool, designed and built for the application of violence onto another human being, in order to kill or seriously disable them. And it is from this function that its value as a symbol derives. This is the basic hypothesis behind all the works on the functional analysis of Bronze Age swords, and use-wear studies of the so called ‘combat damage’. This is also the hypothesis that questioned one of the still ongoing currents of thought: that these swords were not really meant for battle. An idea which fundamental arguments have always boiled down to variations or elaborations of the concept: ‘these swords are functionally poor’ (see for example Pearce 1998; Carrión et al. 2002; González García 2009). Well, studies, some better founded than others, have been carried out, especially in recent years, about how the Bronze Age swords might have been used2. Typology, in first instance, and then materials and technological analyses allowed for experimental archaeology to build reproductions

118

Integrating form, function and technology in ancient swords. The concept of quality with a mechanical behaviour close enough to the originals. From there, use-wear, ergonomics and biomechanics have been the basic tools to develop an approximation to their use. Forensic Archaeology has also been incorporated into this argumentation, with some results that would support the use of swords in the contemporary period remains (Campillo 1977; Cloquell and Aguilar 1996). Recent researchers are now on the path of proving that these swords were perfectly capable to withstand the rigours of serious battle, and that they actually did so (Molloy 2007). But this discussion, among many other things, brought to attention a concept that maybe deserves some additional thought.

every technological context. This would allow us, the modern researchers, to work with scales of degrees of quality, which, if limited to this, could perhaps be useful for descriptive purposes, but hardly for anything else. But the limits given by technology and how they are able to meet the demand of the task at hand also configure the expectations of the actual users in regard to the performance of the tool created. They configure the knowledge of what the tool is supposed to do. All of this leads us to acknowledge that, in the context of where the object, the sword, belongs, quality is in fact a perception. And trying to understand this perceived quality, from the point of view of the subject of this perception, can give us a great insight into the relationship between form, function and technology of the object we are studying, about why things were done the way they were done.

QUALITY As already mentioned, there is still nowadays a current of thought that states that bronze swords were functionally questionable as actual weapons, meaning from a functional point of view, that bronze swords, as mechanical tools, are bad. Well, if bronze swords are bad, what would be then considered a good bronze sword? This question has many ramifications, and here is where it comes handy: the concept of ‘quality’. Quality is better defined as a relative scale. Thus, we would speak of degree of quality, which would be defined as how well a tool can perform its intended function. This concept involves only the functional quality and does not refer for instance to the quality of decoration. Typically the scale goes from the item that is absolutely unable to perform its intended function, up to the ideal: the item that performs the task perfectly. Of course, Platonism aside, this is just an ideal. But, then, if we stop to think about it, is exactly this ideal that lingers behind the arguments that say that the bronze swords are ‘bad’, that is, unable to be used as swords. In our world of ISO standards, we like to think we have a pretty good idea of what is ‘good’ and what is not, but of course this cannot be applied to any other historical context. As already mentioned, the tool is adapted to the job through technology, in its wider sense. Technology helps us to determine a framework, an upper limit for our scale to get an idea of up to which point the object can be adjusted to the task. In this sense, we should be able to create a theoretical kind of absolute scale for

THE TOLEDO SWORDS To illustrate this concept an example will be given. For this purpose the famous Toledo Swords are taken, which we studied more deeply as part of a wider problem elsewhere (Gener 2007). We have some 17th century AD sources regarding the price of sword blades in some cities in Spain. Due to conflicts with dealers inflating the prices, especially in times of higher demand occasionally the ruling bodies of the city had to fix the prices at what certain goods could be sold. In this particular case, the subjects of the regulation were sword blades, and the documents offer a list of different types, classified by their place of origin, and their corresponding price. In the two lists used for the study, Toledo is always marked prominently as the place where the most expensive blades, that we suppose of rapier type, given the context, came from. Period literature – not only Spanish but also European, Shakespeare even on a few occasions3 – bears witness of the high regard in which Toledo blades were held. It also documents, although not always so explicitly, how such a ‘good’ blade was supposed to behave. It had to be flexible and elastic enough to bend to a near impossible degree and return to true form without any deformation. It had to keep a keen edge and a sharp point, able to cut and pierce through materials that ranged from the difficult to the impossible, and had to be shi-

3 Cf. Lope de Vega in ‘Las flores de Don Juan’ (Act 1, Scene 1; after Menéndez Pelayo 1968), or William Shakespeare in ‘Otello’ (Act 5, Scene 2), ‘The Merry Wives

of Windsor’ (Act 3, Scene 5) or ‘Romeo and Juliet’ (Act 1, Scene 4; after Craig 1914).

119

Marc Gener

Figure 1: Location of the Ría de Huelva Hoard. Key: [a] Paleographic Reconstruction of the Ría de Huelva, after Garrido y Orta 1989 and Ruiz Mata 1990. The filled circle marks the actual location

Huelva hoard were closely studied for this purpose. The Ría de Huelva hoard, dating around the 10th/9th century BC, is of great importance for the quantity and type of material included in it (especially weapons: swords, arrowheads, daggers and spearheads and butt caps, but also objects like fibulae, rivets, buttons, etc.) and for its location at the bottom of a river’s bed.4 The hoard has been the subject of many studies, not only concerning its actual content but also about the creation of the hoard itself, having been mainly related to the ritual practices of the time (RuizGálvez 1995; Brandherm 2007). The typology of the objects is associated with the Atlantic Late Bronze Age world, and is the largest hoard in the Iberian Peninsula from this cultural context. It must be mentioned that this finding was not as much excavated as ‘extracted’ from the bottom of the mouth of the river Oriel in 1923 during the dredging works of the Huelva Harbour. The dredge operators observed that some bronze objects were being lifted together with the mud of the river bed. In the following weeks, the dredging works brought up to 400 objects and fragments to the surface, configuring the hoard we known today (fig. 1) (Ruiz-Gálvez 1995, 15). The whole of it was taken to the Museo Arqueológico Nacional, in Madrid, where the objects were restored and some of them exhibited. Among the studies concerning the objects of the hoard, are some on the metallurgical aspects, which have produced significant results about their manufacture, placing them in the wider context of the Late Bronze Age peninsular metallurgical technology (Rovira 1995; 2007). Other archaeometrical studies have contributed additional data, like Lead Isotope analyses which revealed that the origin of the metal used

ny and stay this way for a long time. What can be concluded from this example is, that all the virtues for a blade to be perceived as ‘good’ were related to its mechanical properties, which allowed it to perform its intended task with a varied degree of excellence. And this is what made them ‘good’, as well as valuable, not only in gold but also in prestige. The swords belonged to people who were important and wanted to show it. In addition to ostentatious decoration they wanted also quality blades, and there was a thriving market all over Europe of blades made in different places, but marked as having been made in Toledo by a prestigious maker. It has still to be proven empirically that the Toledo blades had better mechanical properties than those from other places. Metallographical analysis and material testing should be performed exhaustively on original Toledo blades, something that can be difficult to accomplish due to the shortage of genuine material of this kind and the obvious limits to the invasiveness of the techniques that can be used. However it would give us a great insight into the degree of control over a given technology, its reproducibility, implementation and transmission. A wide study on this type of technology would give us a good idea about the expectations a user had about a tool (a weapon or armour), which in turn gives us information about why they faced combat in a particular way, and how. THE RÍA DE HUELVA SWORDS All of this, of course, can be applied diachronically. This can be illustrated with another example from much earlier in time and some fragments of the Late Bronze Age swords from the Ría de 4

Albelda 1923; Díaz 1923; Almagro Bosch 1940; Terrero 1944.

120

Integrating form, function and technology in ancient swords. The concept of quality

2

3

4

5

Figure 2: Fragments Mus. no. 24-60-69 (above) and 24-60-47 (below) of a sword from the Ría de Huelva Hoard. – Figure 3: Detail of one fragment (Mus. no. 24-60-47) showing the point of breakage with the other fragment (Mus. no. 24-60-69). – Figure 4: Detail of the surface of the fragment Mus. no. 24-60-47. – Figure 5: Detail of the surface of the fragment Mus. no. 24-60-47

in different objects is much more varied than originally thought (Montero Ruiz et al. 2007). We had the privilege of examining some fragments of the swords from the hoard, as well as sampling some for metallurgical analysis. This work is still in progress, and there have been some problems with the study. These arise from the particular circumstances of the find, and how much of the damage to the swords occurred prior to deposition, and how much is due to the extraction procedures. This situation is aggravated by the heavy restoration some of the swords underwent when deposited in the museum, which makes it now impossible to differentiate between ‘ancient’ and ‘recent’ damage (Gener et al. 2009). For the previously mentioned illustrative purposes, two fragments of a tin bronze sword blade of the carp-tongue type (Mus. no. 24-60-47 and 24-60-69, see fig. 2), will be used. We will concentrate in the materials properties of our subject of study. As a general rule and with all other circumstances being equal, when swords break, they do it always preferably where a structural defect

in the casting is present (shrinkage porosity, inclusions, bubbles, etc.). Due to its morphology and its intended use as a sword of this type, it is bound to be indeed subjected to a rather extreme mechanical stress (Molloy 2007; Verhoeven 2007). At the point of breakage between the two fragments presented here is a big bubble in the bronze (fig. 3). When bubbles occur in the surface, they can be found almost certainly in the interior structure as well (Stefanescu 2002), and in our particular case the surface is riddled with irregularities and cavities, which cover the full spectrum of sizes and depth, from small pores to true craters, going in some places even all the way through the blade (fig. 4. 5). In short this sword would have all the visual signs of being prone to break under stress, making it less than ideal for performing its intended task. This sword would look to us like a ‘bad’ sword, and probably did look equally bad to its original prospective user. Nonetheless, the microstructure tells us that this sword had its blade cold worked, annealed and extensively cold worked again. With an alloy 121

Marc Gener

of close to 9 % tin, the first phase of cold working and annealing resulted in a small size of grain, and after the last phase of cold working with an average hardness of 187 HV. It can be safely stated that a lot of work and care was involved in making it as functional as possible. A sword that given its visual appearance would be, at first thought, only good for the scrap pile, but that is finished like a perfectly functional one, brings up a question: why, when it was so bad, was so much work invested after its casting? If it was made only for display or ceremonial use, why, such a work-intensive finishing? Well, maybe it was not that bad. Maybe a structurally unpredictable sword was better than no sword. Maybe the user knew fairly well the limits of the material, and was aware of what to expect of this sword. Maybe his concept of quality, his concept of how able that particular sword was of performing its intended task was conditioned by the fact that he knew pretty well the characteristics of the task at hand, and how the tool intended to perform. And, armed with this knowledge, the user decided that, maybe, this ‘bad’ sword was, at the end, ‘good enough’. These kinds of questions are not just an exercise of the mind. They configure a particular landscape regarding the relationship between an object, its maker and its user. They relate to the actual technological knowledge of certain chronological and geographical contexts, and give a different perspective on the possibility of use of certain kind of objects. Which, in this particular case, happen to be swords: paradigmatic instruments of combat, potential recipient of the best that metals technology was able to offer in a given moment and symbols of a privileged class. At some point, someone used them to defend and to conquer, wore them to show others that they were ready to kill and die on behalf of their community, and that they deserved privileges for that. How it was perceived, not only as a symbol, but in its aspect as a tool for war, is certainly important. CONCLUSIONS When undertaking a study of an object of the material culture we study its form, in order to find out how it was made, but determining its function will tell us why it was made in a particular way. Technology, in a wider sense, is the framework that defines the means available through which this function can be fulfilled, and configures the expectations about the performance possibilities of the object. How well a tool can perform its in-

tended function is what defines the degree of its quality. Modern standards cannot be applied to quality, as it would then become just a descriptive tool. Instead, quality must be studied in the framework of the original object, trying to understand how a tool was expected to perform by the actual users. In this sense, we talk about a perceived quality. This approach will give us insight into the relationships of an item with its context, helping us to expand the information we are able to extract from our studies in such fields as antique technology or functional analysis of ancient artefacts. BIBLIOGRAPHY Albelda, J. 1923. Bronces de Huelva. Révue Archéologique XVIII, Vè serie, 222–226. Almagro Bosch, M. 1940. El Hallazgo de la Ría de Huelva y el final de la Edad de la Edad de Bronce en el Occidente de Europa. Ampurias II, 85–143. Burton, R. F. 1884. The Book of The Sword. London, Chatto and Windus. Brandherm D. 2007. Las espadas del Bronce Final en la Península Ibérica y Baleares. Prähistorische Bronzefunde IV, 16. Stuttgart, Steiner. Bridgford, S. 2000. Weapons, warfare and society in Britain 1250–750 BC. Unpublished PhD thesis, Sheffield University. Campillo Valero, D. 1977. Paleopatología del cráneo en Cataluña, Valencia y Baleares. Barcelona, Montblanch-Marín. Cloquell, B. and Aguilar, M. 1996. Herida por espada en un niño argárico. Revista de Arqueología, 17 (184), 10–15. Craig, W. J. (ed.) 1914. The Oxford Shakespeare: The Complete Works of William Shakespeare. London, Oxford University Press. Díaz, J. 1923. Objetos de bronce en la Ría de Huelva. Sociedad Española de Antropología 2, 37–40. Garrido, J. P. and Orta. E. M. 1989. La necrópolis y el hábitat orientalizante de Huelva, 5–93. Dirección Provincial de la Consejería de Cultura, Huelva. Gener, M. 2007. Metallographic study of some 17th and 18th c, European sword (rapier) blades. Proceedings of the 2nd International Conference ‘Archaeometallurgy in Europe’, Aquileia, Italy, 17–21 June, 2007 [electronic format, ISBN 88-85298-61-3]. Gener, M., Montero, I. and Rovira, S. 2009. Hasta aquí puedo leer. Discusión sobre un estudio de fragmentos de espadas del bronce final

122

Integrating form, function and technology in ancient swords. The concept of quality del depósito de la Ría de Huelva. In J. Barrio Martín and E. Cano Díaz (eds), Proceedings of ‘MetalEspaña’08’, Universidad Autónoma de Madrid – Cantoblanco, Madrid, Spain, 10– 12 April, 2008, 234–241. San Sebastián de los Reyes, Gráficas 82. González García, F. J. (2009), Between Warriors and Champions: Warfare and Social Change in the Later Prehistory of the North-Western Iberian Peninsula. Oxford Journal of Archaeology 28 (1), 59–76. Kristiansen, K. 2002. The tale of the sword – swords and swordfighters in Bronze Age Europe. Oxford Journal of Archaeology 21(4), 319–332. Matthews, S. 2005. The materiality of gesture: Intimacy, emotion and technique in the archaeological study of bodily communication, article presented at the round table session ‘The archaeology of gesture: Reconstructing Prehistoric Technical and Symbolical Behaviour’ at the 11th annual meeting of the European Association of Archaeologists, 5–11 September, 2005, Cork (Ireland). [Publ. online with Semioticon: the Open Semiotics Resource Centre. Available at: http://www. semioticon.com/virtuals/archaeology/materiality.pdf ] Menéndez Pelayo, M. (ed.) 1968. Obras de Lope de Vega. Madrid, Atlas. Mödlinger, M. and Ntaflos, T. 2007. Manufacture and use of Bronze Age swords. Multidisciplinary investigation of Austrian Full Hilted and Griffzungen-type swords. Proceedings of the 2nd International Conference ‘Archaeometallurgy in Europe’, Grado and Aquileia, Italy, 17–21 June, 2007 [electronic format ISBN 88-85298-61-3]. Molloy, B. P. C. 2007. What’s the bloody point?: Bronze age swordsmanship in Ireland and Britain. In B. P. C. Molloy (ed.), The Cutting Edge: Studies in Ancient and Medieval combat, 90–111. Stroud, Tempus. Molloy, B. P. C. 2008. Martial arts and materiality: a combat archaeology perspective on Aegean swords of the fifteenth and fourteenth

centuries BC. World Archaeology 40 (1), 116–134. Montero Ruiz, I., Hunt Ortiz, M. A. and Santos Zalduegui, J. F. 2007. El depósito de la Ría de Huelva: procedencia del metal a través de los resultados de los análisis de Isótopos de Plomo. In J. Cielis, G. Delibes de Castro, J. Fernández Manzano and L. Grau Lobo (eds), El hallazgo leonés de Valdevimbre y los depósitos del Bronce Final Atlántico en la Península Ibérica, 194–209. Estudios y Catálogos 17, Museo de León. Pearce, M. 1998. Reconstructing prehistoric metallurgical knowledge: the Northern Italian Copper and Bronze Ages. European Journal of Archaeology 1 (1), 51–70. Rovira S. 1995. Estudio Arqueometalúrgico del depósito de la ría de Huelva. In M. Ruiz Gálvez, Ritos de paso y puntos de paso. La Ría de Huelva en el mundo del Bronce Final europeo. Complutum, Extra 5, 33–58. Rovira S. 2007. Estudio arqueometalúrgico. In D. Brandherm, Las espadas del Bronce Final en la Península Ibérica y Baleares. Prähistorische Bronzefunde IV, 16, 155–175. Stuttgart, Steiner. Ruiz Mata, D. 1990. La Ría de Huelva: un foco clave de la protohistoria peninsular. Appendix to J. Terrero, Armas y objetos de bronce extraídos de los dragados del puerto de Huelva. Reimpresión de Clásicos de la Arqueología de Huelva 3, 57–70. Ruiz-Gálvez, M. 1995. Ritos de paso y puntos de paso. La Ría de Huelva en el mundo del Bronce Final europeo. Complutum, Extra 5. Stefanescu, D. M. 2002. Science and engineering of casting solidification. New York, Springer. Terrero, J. 1944. Armas y objetos de bronce extraídos en los dragados del puerto de Huelva. Reimpresión de la Excma. Dip. Prov. de Huelva, Clásicos de la Arqueología de Huelva 3. Verhoeven, J. D. 2007. Steel Metallurgy for the Non-Metallurgist. Metals Park (OH), ASM International. Marc Gener Instituto de historia Centro de Ciencias Humanas y Sociales (IH/CCHS - CSIC) C/Albasanz, 26–28 Madrid 28037 Spain 123

124

SLASH AND STAB: SOME REMARKS ON THE MANUFACTURE, FUNCTION AND DEPOSITION OF CARP’S-TONGUE SWORDS (SUMMARY) Dirk Brandherm and Magdalena Moskal-del Hoyo The following is a brief summary of our attempt to bring together results from two earlier separate studies on the subject of carp’s-tongue swords which arrived at quite similar conclusions regarding their development and chronology (Moskal 2003, 2007; Brandherm 2007; Brandherm and Burgess 2008). A more comprehensive revision of these earlier studies by the present authors, which is also to take into account results from recent work by B. Quilliec (2007), is underway and will feature an expanded and fully revised catalogue of carp’s-tongue swords known to date. Constraints of time on the authors’ side and of available space in this volume do not permit to present a full revision at this stage. The present contribution then is only to give a concise overview of the main points we were trying to make in our paper presented at the Vienna workshop. First of all, the criteria previously established for a morphological distinction between Type Nantes and Type Huelva carp’s-tongue swords – specifically the outline of the grip-shoulder transition – were cross-checked against the occurrence of different ricasso shapes in both groups. While rectangular and sub-rectangular ricassi clearly predominate among Type Nantes swords and curved ricassi are much more common in the Huelva than in the Nantes type, it emerges that hilt outline as a typological criterion in this particular case produces groups that are morphologically more homogeneous overall and more coherent in terms of their geographical distribution than any groups defined according to the shape of the ricassi. Huelva and Nantes type swords are not only morphologically distinct and show different geographical distribution areas (fig. 1), with Type Huelva mainly distributed in the Iberian Peninsula and Type Nantes chiefly north of the Pyrenees, but also differ significantly in terms of alloy composition. In no case does the lead content in Huelva type swords exceed 0.5 %, whereas Nantes type swords typically are characterized by a higher and much more variable lead content. On the other hand, the percentage of tin in Nantes type swords shows considerably less variation than in Huelva type pieces, where it ranges from less than 8 % to more than 13 %. In general, Type Huelva carp’s-tongue swords are characterized by binary bronzes while Type Nantes represents

a use of leaded bronze alloys (cf. Rovira 1995, 34–39; Rovira and Gomez 1998, 82). Both types also differ with regard to deposition practice. While Type Huelva swords are known mostly from wet contexts, frequently as single finds, Type Nantes swords are mainly found in hoards of the so-called ‘carp’s-tongue complex’ on both sides of the Channel, characterized by a rather homogeneous composition, including specific types of weapons and tools (fig. 3a. 3b). As these hoards are composed chiefly of scrap metal, Type Nantes swords are usually encountered in much more heavily fragmented state than Huelva type swords. On the other hand, Type Huelva swords are totally absent from ‘carp’s-tongue complex’ hoards. Where Huelva type swords do appear in hoards, the composition of the respective assemblages has much more in common with the large weapon hoards of the Wilburton phase than with the typical ‘bric-a-brac’ carp’s-tongue assemblages, indicating a considerably earlier date of Type Huelva swords than hitherto accepted. Unlike most other ‘Atlantic’ sword types, Huelva swords might have originated in the Iberian Peninsula or in the Western Mediterranean, resulting from a hybridisation between early Urnfield leaf-shaped swords, such as Types Hemigkofen or Erbenheim, and late rapiers, as e.g. the Sardinian ‘votive’ pieces from the Arzachena hoard (cf. Begemann et al. 2001, 46 fig. 2) which might also provide the prototype model for the diagnostic midrib featured by most carp’s-tongue swords. This would help to explain the distinct impression of a rapier-like point having been quite unorganically attached to an early Atlantic leaf-shaped sword given by some of the Iberian Huelva type swords (e.g. Brandherm 2007, no. 57). The third main type of carp’s-tongue sword, Type Monte Sa Idda, shows a peculiar distribution that differs from the previous types, restricted almost entirely to the southern half of the Iberian Peninsula and to Sardinia – with a few notable exceptions. Chronologically this type apparently succeeds Huelva type swords and would be at least partially contemporaneous with Type Nantes. This is also supported by metal analyses from Monte Sa Idda type swords, indicating a comparatively high lead content,

125

Dirk Brandherm and Magdalena Moskal-del Hoyo

Figure 1: Distribution of Huelva ( ● ) and Nantes ( ▲) type carp’s-tongue swords 126

Slash and stab: Some remarks on the manufacture, function and deposition of carp’s-tongue swords (summary)

Figure 2: Distribution of undiagnostic carp’s-tongue blade fragments 127

x

x

x

x

Spearhead, incised x ornament on socket Double-edge socketed knife/puñal x

Hog‘s back knife

x

Socketed knife

x

Socketed gouge

x

Socketed hammer

x

x

x

x

x

x

x

x

x x

x

x

x

x

x

x

x

Socketed axe with winged ornament

x

x

x

x

x

x

Socketed axe with vertical ribs

x

x

Winged axe

x

x

x

x

x

x x

x

x

x

x

x

Bugle-shaped object Bracelet with incised ornament Ingot

x x

x

x

x

x

x x

x

x x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x

x

Nantes, ‘Prairie-de-Mauves’ x

x

x

x

x

x x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x

x x

x

x x

x

x

x

x x

Nantes ‘Jardin des Plantes’

x

x

x x

Locmariaquer, ‘Pont-er-Vil’

Lessay

x

x

x

x

x

x

x

x

x

x

Tanged sickle

x

x

x

Tanged chisel

Juvincourt-et-Damary

Île de Groix ‘Men-Stang-Roh’

Île de Batz ‘Île Verte“

Hérouville-Saint-Clair

x

x

x x

x

x

x

Socketed axe

Axe mould

x

x x

Double-edge tanged knife/puñal

x

Guingamp o área de Bourbriac

x

x

Graville-Sainte-Honorine

x

Gouesnach ‘Menez -Tosta’

x

Gausson ‘La Cour’

x

Fouesnant ‘Le Douric’

x

x

Festigny ‘Bois-des-Roches’

Spearhead

x

Cubzac-les-Ponts

x

Chédigny

x

Challans ‘La Villate’

Azay-le-Rideau

x

Bringolo

Auvers

Ewart Park sword

OBJECTS/ HOARDS

Bangor ‘La Celastrène’

Amiens ‘Le Plainseau’

Dirk Brandherm and Magdalena Moskal-del Hoyo

x

x

x

x

x

x

x

x

x

x x x

x

x x

x

x

x

x

Figure 3a: Occurrence of common types associated in hoards with carp’s-tongue swords from the Channel core area and adjacent regions

much better in line with Type Nantes than with Huelva type weapons. The lack of any significant overlap in the distribution areas of Type Nantes and Type Monte Sa Idda might be explained by the realignment of trade networks in the Western Mediterranean at this stage, when southern Iberia

seems to have lost some of its prior focus on trade along the Atlantic seaboard and became increasingly involved in the east-west trade run by the Phoenicians. Thus, not only are different types of carp’stongue swords defined by quite distinctive mor-

128

x

x

x

x

x

x

x

x x

x

Socketed knife

x

x

x

x x

x

x

x

x

Socketed hammer

x

x

x

x

x

x

x

x

x

x

Socketed axe

x

x

x

Socketed axe with winged ornament

x

x

x

Socketed axe with vertical ribs

x

Winged axe

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x

x

x

x

x

Bracelet with incised ornament

x

Ingot

x x

x

x

x

x

x

x

x

x

x

x

x

x x

x

x

x

x

x

x

x

x

x

x x x

x

x

x x

x

x

Tanged sickle Bugle-shaped object

x

x

Tanged chisel

Axe mould

x

x

x

x

Socketed gouge

x

Saarlouis ‘In der Bins’

x

x

Vange

x

x

x

Stormouth

x

Minster-in-Thanet, ‘Ebbsfleet I’

x

Levington

x

Leigh (II)

Hoaden

x

Eastbourne ‘Beachy Head’

x

x x

x

Grays Thurrock (I)

x

Egham ‘Petters Sports Field’

x

Cumberlow Green

x

Addington

x

Double-edge tanged knife/puñal Hog‘s back knife

Saint Mary‘s

x

Spearhead, incised x ornament on socket Double-edge socketed knife/puñal

Longy

x

Saint-Yrieix-sur-Charente, ‘Vénat’

x

Saint-Martin-de-la-Place

Spearhead

Saint-Marc-le-Blanc ‘Le Plessix’

x

Questembert Parc-aux-Bœufs’

x

Plouescat ‘Parc-Bodenez’

Notre-Dame-d’Or ‘La Grimaudiere’

Ewart Park sword

Plouénan ‘Le Bourg’

OBJECTS/ HOARDS

Neuvy-sur-Barangeon ‘Petit-Villatte’

Slash and stab: Some remarks on the manufacture, function and deposition of carp’s-tongue swords (summary)

x x

x

x

x

x

Figure 3b: Occurrence of common types associated in hoards with carp’s-tongue swords from the Channel core area and adjacent regions (continued)

phologies, but at least in part they also differ in alloy composition and in their typical find contexts. However, it has to be admitted that only a minority of carp’s-tongue swords are sufficiently well preserved to allow attribution to any specific type, the vast majority consisting in undiagnos-

tic blade fragments. Most of these are found in northern France and south-east England (fig. 2), mainly in hoards of the ‘carp’s-tongue complex’. It does not come as a complete surprise then that a preliminary analysis of the composition of these hoards shows many similarities with those hoard

129

Dirk Brandherm and Magdalena Moskal-del Hoyo

100 N = 103 hoards

90 80 70 60 50 40 30 20 10

Axe mould

Ingot

Bracelet with incised ornament

Bugle-shaped object

Tanged sickle

Tanged chisel

Winged axe

Socketed axe with vertical ribs

Socketed axe with winged ornament

Socketed axe

Socketed hammer

Socketed gouge

Socketed knife

Hog's back knife

Double-edge tanged knife

Double-edge socketed knife

Spearhead, incised ornament on socket

Spearhead

Ewart Park sword

Nantes sword

0

%

Figure 4: Frequency of common types in hoards containing carp’s-tongue swords from the Channel core area and adjacent regions

assemblages that contain hilt fragments clearly attributable to Type Nantes swords (fig. 4). Correspondingly, in 26 of 103 analysed hoards are undiagnostic blade fragments associated with Type Nantes hilts. Other than carp’s-tongue sword fragments, with regard to weapons found in ‘carp’s-tongue complex’ hoards, fragments from Ewart Park swords form the most numerous group, and among the tools different types of socketed axes and socketed gouges are dominant. Also hog’sback knives and bugle-shaped objects are very common. Our future work on hoard composition will look further into regional variation. The metal analyses available from carp’stongue sword fragments of undetermined type included in those hoards indicate a dominance of leaded bronzes. All of the available evidence thus seems to indicate that any blade fragments from hoards found north of the Pyrenees most likely are from Type Nantes swords. On the other hand, in the Iberian Peninsula not only is the percentage of undiagnostic carp’s-tongue fragments recovered from settlements or as single finds much higher than the corresponding figure from sites north of the Pyrenees, but the available metal analyses also show mainly binary bronzes, similar to Type Huelva swords (cf. Rovira 2007, 156–158). Thus, based on ge-

ographical distribution, archaeological context and alloy composition we would conclude that most undiagnostic carp’s-tongue sword fragments from Iberia very likely are from Huelva type swords (fig. 2). The data obtained from morphologically undiagnostic fragments thus is totally in line with what was stated above concerning the distinctive character of different types of carp’s-tongue swords, not only with regard to morphology, but also to alloy composition and find contexts. This in particular regards types Huelva and Nantes, which are characterized by quite different main distribution areas, different prevailing find contexts and a different metallurgical make-up. While chronology certainly plays an important part in these differences, a full explanation of the observed discrepancies will need to take a broader view of the complex cultural background of these items. BIBLIOGRAPHY Begemann, F., Schmitt-Strecker, S., Pernicka, E. and Lo Schiavo, F. 2001. Chemical composition and lead isotopy of copper and bronze from Nuragic Sardinia. European Journal of Archaeology 4, 43–85.

130

Slash and stab: Some remarks on the manufacture, function and deposition of carp’s-tongue swords (summary) Brandherm, D. 2007. Las espadas del Bronce Final en la Península Ibérica y Baleares. Prähistorische Bronzefunde IV, 16. Stuttgart, Steiner. Brandherm, D. and Burgess, C. 2008. Carp’stongue problems. In F. Verse, B. Knoche, J. Graefe, M. Hohlbein, K. Schierhold, S. Siemann, M. Uckelmann and G. Woltermann (eds), Durch die Zeiten... Festschrift für Albrecht Jockenhövel zum 65. Geburtstag. Internationale Archäologie, Studia honoraria 28, 133–168. Rahden, Marie Leidorf. Moskal, M. 2003. Schyłkowobrązowe miecze z językiem karpia w Europie. Unpublished M.A. thesis, Jagiellonian University, Kraków. Moskal, M. 2007. Późnobrązowe miecze z językiem karpia na terenie Europy. In J. Chochorowski (ed.), Studia nad epoką brązu i wczesną epoką żelaza. Księga poświęcona Profesoriwi Markowi Gedlowi na pięćdziesięciolecie pracy w Uniwersytecie Jagiellońskim, 465–531. Kraków, Wydawnictwo Uniwersy-tetu Jagiellońskiego.

Quilliec, B. 2007. L’Épée atlantique: échanges et prestige au Bronze final. Mémoires de la Société Préhistorique Française 42. Paris, Société Préhistorique Française. Rovira Llorens, S. 1995. Estudio arqueometalúrgico del depósito de la Ría de Huelva. In M. Ruiz-Gálvez Priego (ed.), Ritos de paso y puntos de paso. La Ría de Huelva en el mundo del Bronce Final europeo. Complutum Extra 5, 33–57. Madrid, Servicio de Publicaciones Universidad Complutense. Rovira Llorens, S. 2007. Las espadas del Bronce final de la Península Ibérica: estudio arqueometalúrgico. In D. Brandherm, Las espadas del Bronce Final en la Península Ibérica y Baleares. Prähistorische Bronzefunde IV, 16, 155–175. Stuttgart, Steiner. Rovira Llorens, S. and Gómez-Ramos, P. 1998. The Ría de Huelva hoard and the Late Bronze Age metalwork: a statistical approach. In C. Mordant, M. Pernot and V. Rychner (eds), L’atelier du bronzier en Europe du XX-e au VIII-e siècle avant notre ère. Actes du colloque international “Bronze 96” de Neuchâtel et Dijon 1, 81–90. Paris, CTHS. Dirk Brandherm School of Geography, Archaeology and Palaeocology (GAP) Queen‘s University Belfast Belfast, BT7 1NN Northern Ireland, Great Britain Magdalena Moskal-del Hoyo Departament de Prehistòria i Arqueologia Universitat de València Avgda. Blasco Ibáñez, 28 46010 – València Spain

131

132

THE DISCOVERY OF A NEW TACHLOVICE SWORD AT PONT-SUR-SEINE, AUBE, FRANCE Sylvie Boulud-Gazo ABSTRACT A complete sword was recently found during an archaeological survey at Pont-sur-Seine (Aube), in the north-eastern part of France. The sword was found in a very large pit, without any associated finds. This particular context indicates a probable one piece hoard. The sword is in a very good state of preservation. Its characteristic hilt, with its organic pieces as well as the shape of its blade, both indicate that it belongs to the Tachlovice type. This type of sword is quite rare: around thirty examples are known in Europe amongst which eight have been found in France. Technical analyses that have been carried out on the sword have lead to a better comprehension of its manufacture. This paper presents the main results of these analyses. This new discovery is also a good opportunity to reconsider typological aspects and to update our knowledge of the geographical distribution and the archaeological contexts of Tachlovice swords in Europe. KEYWORDS Tachlovice type sword – Bronze final IIIb (Ha B2/3) – Paris Basin – manufacture – tools – decorative punch INTRODUCTION

Late Bronze Age swords are only rarely unearthed during developer funded excavations except in funerary contexts where preservation conditions are not always optimal. Taking this fact into consideration, the recent discovery of a sword with a solid hilt (and tang) dating from the Bronze final III b (Ha B2/3) at Pont-sur-Seine (Aube) in North-Eastern France can be considered to be a unique find. Discovered during an excavation directed by INRAP (Institut National de Recherches Archéologiques Préventives), the sword is in excellent condition enabling a certain number of observations and analyses to be carried out on the materials used, as well as manufacture and decoration techniques. This discovery also provides an opportunity to revise

the catalogue of French Tachlovice swords and to place their Late Bronze Age production characteristics in an European context. THE PONT-SUR-SEINE SITE AND THE AREA AROUND THE TOWN OF NOGENT-SUR-SEINE Pont-sur-Seine is situated on the north-western limits of the Aube department, which lies southeast of the Paris Basin, approximately 10 km from the town of Nogent-sur-Seine (fig. 1). It is very near to the neighbouring administrative departments of the Marne, the Seine-et-Marne and the Yonne. The area has many known Late Bronze Age archaeological sites that are located along the Seine valley in the area around Nogent-surSeine and more to the West in ‘the Bassée’ and at the confluence of the rivers Seine and Yonne. Funerary sites are particularly well represented for the first two stages of the Late Bronze Age (Bz D–Ha B1) as well as settlement sites and metal hoards in wet and dry contexts. Famous cemeteries dating from the first and second stages of the Final Bronze Age are to be found near Pontsur-Seine (at Barbuise-Courtavant, La Saulsotte, or La Villeneuve-au-Châtelot) and contemporary settlements are also located on the same sites. For the final stage of the Late Bronze Age, there seem to be a smaller number of sites, for example at Barbuise-Courtavant only funerary contexts date from this period. During the Late Bronze Age, this area of the Paris Basin can be seen as a contact zone between two techno-complexes that covered the North and the Centre of France: the Atlantic Complex to the West and the NorthAlpine complex to the East. At the ten hectare site at Pont-sur-Seine domestic and funerary structures dating from the Neolithic, the Bronze Age and medieval times were excavated from February to September 2007 by five INRAP teams. The excavation of the Bronze Age settlement and cemetery, that included inhumations and cremations, was undertaken by S. Fournand. Work on finds to determine precise dating is still in progress. One of the burials at Pont-sur-Seine is an inhumation in a sitting position, identical to mortuary practices that were brought to light in the

133

Sylvie Boulud-Gazo

100

10 0

BarbuiseCourtavant La Villeneuve-au-Châtelot La Saulsotte

MELUN

A

L'

u be

100

Marollessur-Seine

NOGENT

Pont-sur-Seine 0 20

eine La S

100

0 10

Barbey Cannes-Ecluse Villeneuve-la-Guyard

200

20 0

10 0

200

10 0

NEMOURS

nne L ' Yo

Villethierry

TROYES

e nn La Va

SENS

200 100

Migennes 100

m

L'Ar

on

anc

200

Le Loing

0 20

40 0

AUXERRE 200

200

La

400

0 50

400

e Lo ir

200

400

200

500

ein Le Se r

Sermizelles

40 0

La Cur e

0 40

0 40

200

400

500

200

0

25 km

500

500

500

N 40 0

500

40 0

DAO fond de carte S. Rottier

Figure 1: Localisation of the Pont-sur-Seine site (Aube). Several important Late Bronze Age sites are located nearby (infography: S. Rottier, S. Boulud)

Barbuise-Courtavant ‘Les Grèves de Frécul’ and Barbey ‘Les Cents Arpents’ cemeteries (Piette et al. 2005; Rottier 2005). A number of finds have been dated to the first phase of the Late Bronze Age (Bz D–Ha A1): cannelated pottery and grave goods that include copper alloy pins with crosiershaped heads, Villethierry type pins, an amber bead, two copper alloy bracelets and one lignite bracelet. The second phase of the Late Bronze Age (Ha A2–B1), called Rhin-Suisse-France orientale or ‘Urnfield Culture’, is only represented by settlement structures and the final phase of the Late Bronze Age (Ha B2/3) is represented by a cremation burial (pers. comm. T. Nicolas and S. Fournand). The sword itself was discovered in the filling of a large irregular shaped pit measuring 10x6 m, it has a flat profile and lies about 30 cm below the ancient occupation level (fig. 2, A). The sword was slightly bent and one of its sides damaged

during the mechanical excavation of the pit (fig. 2, B. C). Few other finds apart from pottery sherds were found in the pit, whose function remains to be determined. It does not seem to have been a refuse pit, the more likely hypothesis being that of an extraction pit that was rapidly filled in. The datable pottery sherds (fig. 3) indicate the final phase of the Late Bronze Age (Ha B2/3) or even the Early Iron Age (Ha C1). The context of this large pit does however leave unanswered questions. This type of pit is frequently found in domestic contexts always in association with other settlement structures such as storage pits and post-holes that reveal different wooden constructions. This unfortunately is not the case here, but the pit’s location at the edge of the excavation suggests that a contemporary site could be lying outside of the excavation limits. The circumstances of the sword’s discovery and indeed its original context are quite out of the

134

The discovery of a new Tachlovice sword at Pont-sur-Seine, Aube, France

Fosse 2305 A

B

Sword

AB = 10,38m 0

3m

2A

2B

1

2

6

5

4 0

10 cm

3

2C

Figure 2: A. Schematic drawing of the irregular shaped pit 2305 with the sword positioned within (drawing and infography S. Fournand/Inrap GEN, S. Boulud). – B. The sword at discovery in the middle of the pit 2305 (photography S. Fournand/Inrap GEN). – C. Close up of the sword in the fill (photography S. Fournand/Inrap GEN) Figure 3: Pottery finds from pit 2305 where the sword was discovered (drawing and infography M. van Eck/ Inrap GEN) 135

Sylvie Boulud-Gazo

100

10 0

Pont-sur-Seine

MELUN

A

L'

La Rochette

u be

100

NOGENT

Brienne

ne

0 20

100

ei La S

0 10

TROYES

200

20 0

10 0

200

nne L ' Yo

10 0

NEMOURS

Saint-Thibault

e nn La Va

SENS 200

100 100

m

L'Ar

Aubepierresur-Aube

on

anc

200

Le Loing

0 20

40 0

AUXERRE 200

200

La

400

0 50

400

e Lo ir

200

400

200

500

ein Le Se r 40 0

La Cur e

0 40

0 40

200

400

500

200

0

500

25 km

500

500

N 40 0

500

40 0

DAO fond de carte S. Rottier

4

2 4

1 3

5 Figure 4: Map illustrating Bronze final III (Hallstatt B) sword finds in the South-West of the Paris Basin near to Pont-sur-Seine (infography: S. Rottier, S. Boulud) Figure 5: Bronze final III (Hallstatt B) swords in the South-West of the Paris Basin. –1. Brienne (Aube; after Gaucher and Mohen 1972, 541-2). – 2. La Rochette, près Melun (Seine-et-Marne ; after Gaucher and Mohen 1972, 533-1b). – 3. Saint-Thibault (Aube; after Villes 1995, 103). – 4. Aubepierre-sur-Aube (Aube; after Déchelette 1913, 61). Scale varies 136

The discovery of a new Tachlovice sword at Pont-sur-Seine, Aube, France

ordinary, as many Late Bronze Age swords in this area come from votive deposits mainly in watery contexts and less often from burials. To our knowledge no other complete Late Bronze Age sword has been found in a settlement context in this part of the Paris Basin. It seems certain that this sword has not just been thrown away, as it was complete, undamaged and deposited carefully in the pit fill. The manual excavation of the pit subsequent to the sword’s discovery also confirms that it was not a burial as no bones were found near to the sword and this absence is not due to acidic soils as several well preserved inhumations were found on the same site (and the preserved bone on the sword’s hilt itself). The sword has thus been intentionally deposited, isolated in this large pit. Even if the Seine flows less than a hundred metres from this site, the original context of the sword’s deposit was on dry land. Several other swords dating from the Late Bronze Age have been found in the southeastern part of the Paris Basin. Most were discovered accidentally during the dredging of different rivers. Votive river deposits were obviously made here during the Bronze Age in the same way as elsewhere in the Paris Basin and along the Seine valley (Mohen 1977; Brun 2000). Examples of swords dating from the Bronze final III (Ha B) were discovered in this area (fig. 4. 5), such as the Klein-Auheim type sword at Brienne (Aube), the Mayence type sword at La Rochette (Seine-et-Marne), the Ewart ParkChallans type sword at Saint Thibault (Aube) or even the Mâcon type sword found long ago at Aubepierre-sur-Aube (Aube).1 Funerary contexts are rarer and to the best of our knowledge are limited to the first stage of the Late Bronze Age (Bz D–Ha A1). Two Rixheim swords were discovered at Barbuise-Courtavant (‘tombe Morel’, Aube) and at Migennes (Yonne; Morel 1875; Muller et al. 2007). A NEW TACHLOVICE SWORD The Pont-sur-Seine sword is exceptionally well preserved. It is complete, measuring 84 cm in length and weighs about one kilo (fig. 6. 7). It has a straight sided blade with a lenticular section. A decorative plate made out of an organic material that covers the pommel has also been preserved, which is extremely rare. However, the two 1

Dechelette 1913, 61; Gaucher and Mohen 1972, 5331, 541-1; Villes 1995, 103.

137

Figure 6: The Pont-sur-Seine sword before cleaning, length 84 cm (photography J.-G. Aubert, Arc’antique)

Sylvie Boulud-Gazo

BONE

0

0

FACE B

FACE A

1 cm

20 cm 0

5 cm

Figure 7: The Pont-sur-Seine sword (drawing and infography S. Boulud-Gazo) 138

The discovery of a new Tachlovice sword at Pont-sur-Seine, Aube, France

A

B

Figure 8: A. The sword’s hilt before cleaning (photography J.-G. Aubert, Arc’antique). – B. X-ray image of the sword’s hilt (x-ray J.-G. Aubert, Arc’antique)

small plates that were maintained on each side of the grip by three rivet heads have not survived, probably because they were made out of a different material from that used for the pommel plates. This ornamentation is complemented by a geometric patterning at the top of the blade. The hilt and the blade of the sword are typical of Tachlovice swords (fig. 8, A). This type of sword was first defined by H. Müller-Karpe (1961), as E. Sprockhoff in his 1934 paper originally classed the Tachlovice type sword in the Auvernier sword category (Sprockhoff 1934). The hilt of the Tachlovice sword is indeed very similar to that of an Auvernier sword, as only the shape of the pommel differs. For Tachlovice swords, the pommel is navicular shaped and circular or calotte shaped for the Auvernier swords. The grip and the cross-guard are however very similar to the point that it is impossible to distinguish between the two sword types when the pommel is missing. The name Tachlovice co139

mes from a hoard with two swords of this type discovered in Bohemia (Kytlicová 2007, pl. 146 A). In her inventory of solid hilted swords of southern Germany, I. von Quillfeldt uses the term ‘Hostomice’ to describe these swords (v. Quillfeldt 1995, 223–225). This modification has created some confusion and in this paper the traditional name type will be used, giving as much detail as possible as to the shape of the pommel. A synthesis of European Tachlovice swords has been recently published (Sicherl 2008), in which the author proposes a typological classification based on the different construction principles of the hilt. Of the swords that have been classified as Tachlovice, it appears that different techniques where used for the fabrication of the hilt with or without detachable end piece at the top of the pommel and for attaching the hilt to the blade either by a second casting on the tang or by a system of rivets at the cross-guard. Using this typology,

Sylvie Boulud-Gazo

9 Figure 9: View from top of the sword’s pommel: detail of the end detachable piece attached by two rivets and the bone plate (photography J.-G. Aubert, Arc’antique)

the Pont-sur-Seine sword corresponds to the Hostomice construction type 2 (Typ Hostomice, Konstruktionsprinzip 2; Sicherl 2008, 242 f.) which includes the fixing of the detachable end piece at the top of the pommel using two rivets and by a hilt that was cast onto the blade. The swords that belong to this group mainly come from France, the Pont-sur-Seine sword thus confirming the development of a particular technique for swords found on the western limits of the Tachlovice sword diffusion zone. 10

LABORATORY ANALYSES Figure 10: View of the upper part of the blade during cleaning (photography J.-G. Aubert, Arc’antique)

The x-ray radiography With the exceptional preservation of the Pontsur-Seine sword, we grasped the opportunity to complete a series of analyses pertaining to the materials used, how the hilt was made, how it was attached to the blade and the decorative techniques used. X-rays and photographs of the sword before cleaning were provided by the Restoration and Research laboratory Arc’antique in Nantes (by J.-G. Aubert). The x-ray images confirm that the hilt was in fact made by a second casting onto the blade, as gases trapped between the tang and the cross-guard had formed characteristic bubbles that were visible on the tang (fig. 8, B). It is also very clear how the small plate was fixed onto the top of the pommel by two rivets. It seems however that originally four rivets were intended, as two other rivet holes were started under the pommel but never finished. Furthermore, the x-ray images reveal several faults in the metal probably due to bubbles formed during the casting. They can

be seen in different areas on the hilt’s surface, especially near the cross-guard. Analysis of the organic material For the identification of the pommel’s organic plate, the sword was subsequently taken to the Centre de Recherche et de Restauration des Musées de France (C2RMF) in Paris (by I. Reiche, K. Müller and B. Moignard). The objective was to be able to differentiate between antler, bone and ivory, as simple observation of the surface was not sufficient. The analyses carried out used non-intrusive methods: observation using a binocular microscope, x-rays and gamma-rays produced by proton beams (PIXE/PIGE) using the AGLAE accelerator. An elementary analysis of calcium and magnesium levels enables us to differentiate between antler, bone and ivory, however, it is important 140

The discovery of a new Tachlovice sword at Pont-sur-Seine, Aube, France

A

B Figure 11: View of the decor of the blades sides A and B (photography J.-G. Aubert, Arc’antique)

to bear in mind that as we are dealing with an archaeological object the chemical composition has probably been modified over time. Elements such as aluminium, silicon, iron, and especially copper all clearly indicate an alteration of the sword’s surface whilst in the ground. Calcium and magnesium levels were finally compared to a data base of old and modern bone, antler and ivory and the plate attached to the pommel was found to be bone (fig. 9). Elementary analysis Analysis of the metal’s elementary composition was carried out in December 2009 using the same AGLAE accelerator at the C2RMF laboratory (by B. Mille), however we do not yet have the results. Readings were taken on each of the sword’s parts in order to compare the composition of the copper alloy. As with the analysis of the bone plate, non-intrusive methods were used. As soon as the results of the analysis will be known, readings taken along the blade, the 141

hilt, the rivets and their heads will be compared with each other and then compared with other objects of the very Late Bronze Age from both the Atlantic and Continental spheres. The cleaning process Having been buried in the ground for a long period of time, the surface of the metal was covered with a layer of soil mixed with the corroded copper alloy (fig. 10). The sword’s manual and mechanical cleaning by Arc’antique took four months (by M. Vieau). The manual cleaning aimed to preserve as much as possible of the metal’s smooth surface that would probably have been damaged if chemical cleaning had been used. Chemical cleaning does alter the metal’s surface and this method would have wiped away any scratches, decorative details or use marks. An ultrasound scalpel, a manual scalpel (used most) and an engraving burin were used in the mechanical cleaning process. The burin was particularly useful in separating the

Sylvie Boulud-Gazo

0

12

13 A

13 B

Figure 12: Detail of the ricasso edge with its herringbone motif (photography J.-G. Aubert, Arc’antique) Figure 13: A. Detail of the motif using a millimetre scale. The motif is made up of three concentric semicircles and a central conic point. – B. Detail of the motif located directly under the cross-guard, the exterior line has been partly erased (photography J.-G. Aubert, Arc’antique) Figure 14: Proposed restitution of the decorative punch used to decorate the sword’s blade (infography: S. Boulud) 142

14

10 mm

Figure 15: Tachlovice type swords discovered in France. – 1. Pont-sur-Seine (Aube; photography J.-G. Aubert, Arc’antique). – 2. Aliès at Menet (Cantal; photography © Trustees of the British Museum). – 3. Lyon Area (Rhône; Mohen 1971, fig. 8). – 4. L’Epineuse at Vénarey-les-Laumes (Côte-d’Or; Mohen 1971, fig. 8). – 5. Chauny (Aisne; after Vuaillat 1969, fig. 2, 3). – 6. Grigny (Rhône; after Chapotat 1973, fig. 3, 2). – 7. Simandre (Saône-et-Loire, Bonnamour 1990, fig. 26, 35). – 8. Humes (Haute-Marne, drawing S. Boulud-Gazo). Scale varies 143

Sylvie Boulud-Gazo

corrosion from the metal’s surface, thus ensuring the sword’s original surface remained intact. Each stage of the cleaning process was photographed (by J.-G. Aubert).

intervention after the tracing of the motif, which of course would not have been possible if the motif hade been directly etched onto the metal (fig. 13, B).

THE BLADE’S DECORATION

The tools used for tracing the semicircles

The sword is decorated with a geometric pattern at the top of the blade at the ricasso and below. Before cleaning, the pattern was only partially visible on one of the faces, the other side being almost completely covered by a layer of corrosion (fig. 7. 11, A. B). In spite of this a regular patterned décor was observed on both sides of the blade and the cleaning revealed just how regular and how finely executed the motifs really are. The cleaning revealed as well that both of the ricassi edges are decorated with a very fine herringbone pattern (fig. 12). The well preserved surface of the metal and the meticulous cleaning have enabled detailed examination of the sword’s decoration in view of understanding how the motifs were created and which tools were used. Similarly to the decoration of the two ricassi, the blade is also decorated on both sides, with an almost symmetrical pattern. The one side (A) is decorated with two lines comprising of eight motifs of concentric semicircles with a central concave point (fig. 11, A). The two lines mirror each other on either side of the blade’s central fuller. Below this pattern, the same motif of three concentric semicircles with a central point is repeated from which three very regularly etched lines parallel to the blade edge run to the blade tip. The fuller is underlined by two parallel lines, one on each side. On the other side (B), the motif is almost identical except that there are only seven semicircles ordered into two lines instead of eight (fig. 11, B). Except for this minor difference, the general organisation of the motifs is the same on both sides of the blade. The decoration is particularly fine. The semicircular motifs measure 5.7 mm wide and 3 mm high and they are so regularly drawn that the same instrument must have been used to create them (fig. 13, A). At first it was thought a compass with three points was used to trace each motif onto the wax or clay model before casting. Cleaning and further examination under the binocular microscope, confirms that the decoration was indeed created before casting on a wax model (observations by B. Armbruster). Evidence of tracing before casting is visible on one of the motifs as the outer lines have almost been rubbed out, while the other lines are intact, indicating

In contrary to what was thought before the sword was cleaned, the semicircular motifs were probably created by punching the décor onto the wax model using a punch and not by tracing the lines with a pivoting instrument such as a compass. The motifs are identical to each other within the tenth of a millimetre and such small differences can be explained by a slight movement of the hand (more or less pressure being applied and variation of the punch’s angle) or by corrections made to the motif after punching. The motifs were compared using macroscopic photography that confirmed that only a punch can produce such a regular tracing. A punch similar to the decorative punches used to create concentric circular motifs found in the Génelard hoard (Saône-et-Loire) or in the Larnaud hoard (Jura) could have been used (Armbruster 2000, 115). The smallest punches from both hoards measure less than 1 cm in diameter and were probably used to create particularly fine motifs. In the case of the Pont-sur-Seine sword, a similar hemispherical shaped tool can be imagined made out of metal, bone, antler or even wood. The punch would have had a relief décor of three semicircular arcs with a high-relief central cone-shaped protuberance (fig. 14). TACHLOVICE SWORDS IN FRANCE Bringing the inventory up to date The last synthesis of Tachlovice swords published in France dates from 1969 (Vuaillat 1969). D. Vuaillat’s inventory mentioned six swords to which three new finds can be added: the Simandre sword (Saône-et-Loire) discovered in the Saône river (Bonnamour 1990, fig. 26, 35), the Grigny sword (Rhône) from the Rhône (Chapotat 1973, fig 3, 2) and the Pontsur-Seine sword (Aube) (fig. 15, 7. 6. 1.). Three other swords with broken grips from Grigny and Lyon (Rhône) can be classed as either Tachlovice or Auvernier type swords (Vuaillat 1969, fig. 2, 8; Chapotat 1973, fig. 3, 1. 3), as without the pommels the two types cannot be distinguished from one another (see above).

144

The discovery of a new Tachlovice sword at Pont-sur-Seine, Aube, France

2

1 3

Figure 16: Tachlovice type sword from the Aliès hoard (Cantal). – 1. Aliès hoard at Menet as illustrated in the ‘Revue Archéologique’ 1872 (after Anonyme 1872, pl. XXV). – 2. Sword from the Aliès hoard at Menet at the British Museum (photography © Trustees of the British Museum). – 3. Sword of ‘unknown provenance’ at the MAN and published by D. Vuaillat in 1969 which is in reality a copy of the Tachlovice sword from the Aliès hoard (after Vuaillat 1969, fig. 2, 7). Scale varies 145

Sylvie Boulud-Gazo

Combining a fresh look at archive material with new research in museum collections, it seems that two examples mentioned by D. Vuaillat do in fact belong to the same sword (fig. 16). The sword classed as being of ‘unknown provenance’ at the Musée des Antiquités Nationales (MAN) in Saint-Germain-en-Laye (Acc.no. 24.748; Vuaillat 1969, fig. 2, 7) is in fact a copy of the sword from the Aliès hoard found at Menet (Cantal), the original being in the British Museum (Acc.no. 1890-0519-1; photograph communicated by B. Roberts). The possible identity of these two examples was first suggested by P. Abauzit whilst publishing the Aliès hoard. However, Abauzit thought that the original sword was housed at the MAN and the copy in the British Museum (Abauzit 1973, 281). The Tachlovice sword from the Aliès hoard, belonging to the Gréau collection, was exhibited at the Universal Exhibition in 1878, having been lent to the museum’s director by a collector in order to make a moulding at the MAN workshop. The exhibition was being used as an excuse to make copies of the more interesting archaeological finds. The museum’s archives mention that the sword was later sold in Paris in 1880 without mentioning the buyer or the sword’s future home. It is only ten years later that the sword finally arrives at the British Museum (1890). However, two other swords from the Aliès hoard, a Mörigen type sword and a Weltenburg type sword are missing and are not mentioned in the Musée des Antiquités Nationales archives in Saint-Germain-en-Laye and are not in the British Museum’s collections either (fig. 16, 1). Their whereabouts remain unknown. So this up to date inventory brings the number of Tachlovice swords found on France soil to eight certified swords and three other possible ones (fragmented) (fig. 15). A sword with an extremely complex hilt, found in an open-cast mine at Oswald (Bas-Rhin), also reminds us of the Tachlovice type sword (Lambot 1981, 284–288). Its pommel has a navicular shaped detachable element which resembles those on Tachlovice swords, however, the tang without its solid hilt is more comparable to the Klein-Auheim sword type. The same pommel mounting and rectangular end piece are also found on two Tachlovice swords discovered in Germany at Kirschgartshausen and Ladenburg. These two swords were classed by B. Sicherl as Kirschgartshausen swords (Sicherl 2008, 243 f.). To conclude, even with its characteristic pommel, the Oswald sword does not have all the elements of a Tachlovice sword, and seems to be more of a hybrid or prototype, perhaps created

as the answer to a technical problem encountered in the process of copying a Tachlovice sword. DISCOVERY CONTEXTS AND TYPO-CHRONOLOGY All of the Tachlovice swords discovered in France are from hoards and none are from funerary contexts, in contrast to Eastern Europe where examples of this type are found in burials. Three are from terrestrial contexts: the Aliès hoard (Menet, Cantal), the Epineuse hoard at Vénarey-les-Laumes (Côte-d’Or; Nicolardot and Verger 1998) and the Pont-sur-Seine sword (Aube). Five swords were found in watery locations: a sword found upriver from Chauny in the Oise river (Aisne; Pilloy 1904), the Humes sword (Haute-Marne; Royer 1912) discovered in an old river bed of the Marne, a sword from the Lyon area (Rhône; Mohen 1971, 43–45) found in the Rhône or the Saône, the Grigny sword found whilst dredging the Rhône and the Simandre sword (Saône-et-Loire) found in the Saône. From a typological point of view, the French swords mainly belong to the Typ Hostomice, Konstruktionsprinzip 2 defined by Sicherl (2008, 242 f.). Their pommels have a detachable end piece, fixed to the pommel by double rivets. The organic element probably covered the section between the end of the sword and the navicular pommel, as can still be seen on the Pont-surSeine sword. With this type of fitting, the hilt is attached to the blade’s tang by a second casting. Of the eight known examples, five belong to this construction type, namely the swords from Aliès, Chauny, Epineuse, Hûmes and Pont-sur-Seine. Two other swords from Grigny and the Lyon area are clearly imitations of this construction type but are of lesser quality and do not correspond exactly to the defined type. The Simandre sword belongs to a slightly different type, without a detachable end piece on the pommel and with a hilt that is attached to the sword using rivets (Type Bothenheiligen, Sicherl 2008, 246 f.). Votive deposits from watery locations do not allow a precise dating of the objects discovered. The land hoards do however give good chronological indications. The Aliès hoard comprises the Tachlovice sword and two other swords dating to the last stage of the Late Bronze Age (BF IIIb / Ha B2/3). The hoard from Epineuse at Vénareyles-Laumes contains a number of objects most of which are characteristic of the Bronze final IIIb (Nicolardot and Verger 1998, 28). The dating of the Tachlovice swords and sub-variants

146

The discovery of a new Tachlovice sword at Pont-sur-Seine, Aube, France

1B

1A

0

2A

5 cm

2B

Figure 17: The decoration of two Bronze final III (Ha B) swords in comparison. – 1. Sword from Brienne (Aube); A. Tang and upper part of the blade (Gaucher and Mohen 1972, 541-2); B. Detail of the decoration (photography S. Boulud-Gazo). – 2. Sword from Humes (Haute-Marne); A. Hilt and upper part of the blade (drawing S. Boulud-Gazo); B. Detail of the decoration (photography S. Boulud-Gazo) 147

Sylvie Boulud-Gazo

to the very end of the Late Bronze Age is confirmed elsewhere by hoards that contain multiple objects such as the Stendal hoard found in Saxony-Anhalt or the Bothenheilingen hoard from Thuringia (Wüstemann 2004). In Bohemia, Tachlovice swords, belong to the Trteno-Hostomice phase dating to Ha B2/3 (Kytlicová 2007, 172 fig. 2, 213. 214). Two later variants of the Tachlovice swords, the two swords from Gornja Radgona (Radkersburg) and Gornje Vrhpolje (Stična) found under tumuli in Slovenia, appear in Hallstatt contexts in the most eastern part of the distribution zone (Gornja Radgona variant; Sicherl 2008, 249 f.). Tachlovice swords, though produced during a short period at the end of the Late Bronze Age (BF IIIb / Ha B2/3), seemed to have influenced Hallstatt sword production even though swords from this later period, of which only two Slovene examples are known, are quite different from the originals. In the same light, the Goldes sword (Austria) also discovered in a funerary context could be a chronological link in the genesis of this most recent variant (Krämer 1985, 46–164). The Pont-sur-Seine sword does not bring new elements to the dating of Tachlovice swords, the few pottery sherds found in the pit with the sword can not be precisely dated, even if the proposed date of Ha B2/3–Early Iron Age is according to the traditional dates given for the Tachlovice swords. DECORATION OF LATE BRONZE AGE LONG SWORDS (BRONZE FINAL III / HA B) The semi-circular motifs on the Pont-sur-Seine sword can be deemed as typical of the Late Bronze Age. They are fairly frequently found, with variations in shape, organisation and association, on long swords (J. D. Cohen’s Langschwerter; Cowen 1955, 97–108), which appear during the Ha B1 period and are still produced during the whole of the Ha B 2/3. Semi circles and concentric circles are often found on Klein-Auheim, Port-Nidau, Briest and Mainz type swords, as well as other Tachlovice swords. J. D. Cowen talked of ‘late-styled decor’(Schwertgriffe mit späterer Dekoration; Cowen 1955, 102 f.) insisting on the fact that these motifs were used for a relatively short length of time and were characteristic of the very end of the Late Bronze Age (Bronze final III / Ha B). It would be useful to write a synthesis on the decorative motifs of Ha B long swords considering all sword types and decorations and not only those of the Tachlovice swords, as swords dating

to the same period were probably produced by the same regional workshops and the same tools were probably used to decorate them. Identical motifs that were used to decorate two distinct swords suggest that these objects were probably produced in the same workshop. By comparing all of the swords decorated with this motif, it rapidly appears that every example is slightly different and it is impossible to find two swords that are decorated in exactly the same way, as the motifs and patterns are similar, but there are always slight variations that make the decoration unique. So it seems that each sword has been individually decorated using standardised motifs. This could be due to swords being made to specific orders, each sword being individualised by using a distinct decorative design that aims to reflect the personality of the recipient. Tracing the tools used to decorate several swords, even if each pattern is unique, could bring a solution to the problem of not being able to identify two swords that are strictly the same, as the ability to pinpoint identical decors would confirm hypotheses about production regions (Middle Rhine, East Paris Basin, Middle Danube, Carpathian Basin, etc.) and would further the question of the circulation and exchange of prestige goods in Europe. Only one of the other French Tachlovice swords, an old find in Humes (Haute-Marne), has a decorated blade. The pattern comprises of circles and concentric semicircles that form a line on the ricasso and underneath at the top of the blade (fig. 17, 2, A. B). Unfortunately, the sword was the victim of a rather aggressive cleaning that has damaged the original surface and rubbed off some of the décor. Fabrication marks that could have given some indication as to the operational sequence in the patterning of the décor have also been erased. In spite of this the regularity of the motifs and their identical dimensions suggest that the same operation sequence was used as for the Pont-sur-Seine sword. The concentric semicircles on the sword discovered at Humes measure a maximum of 5 mm long and are much smaller than the motifs on the Pont-sur-Seine sword. This is probably due to the use of a similar tool but of a different size. Another sword from the Aube department at Brienne also has an interesting concentric semicircular décor associated with a herringbone pattern on the blade’s ricasso (fig. 17, 1, A. B). The discovery context of this KleinAuheim type sword is not known, however, its patina indicates that it could have come from a watery context. Even though the decoration is very similar to that of the Pont-sur-Seine sword, close observation reveals that the motif was not

148

Pont-sur-Seine

Tachlovice and variants Tachlovice or Auvernier Auvernier

^ ^

The discovery of a new Tachlovice sword at Pont-sur-Seine, Aube, France

Figure 18: Distribution map of Type Tachlovice and Auvernier swords in Europe, with localisation of the Pontsur-Seine site (Aube). Map updated after Müller-Karpe 1961, Stockhammer 2004 and Sicherl 2008 149

Sylvie Boulud-Gazo

worked in the same way. A punch does not seem to have been used, as the patterns are too irregular and different. They have clearly been etched by hand, line after line probably using a sharp stiletto tool. Only under a binocular microscope it is eventually possible to distinguish between motifs worked before casting using a wax or clay model or if they were directly etched onto the metal. TACHLOVICE SWORDS IN OTHER AREAS OF EUROPE Sicherl includes 35 complete swords and ten swords with incomplete hilts that belong to the Tachlovice or the Auvernier group in his recent synthesis. Eight French swords are mentioned in the inventory, but as discussed here, the sword of ‘unknown provenance’ can be excluded and the Pont-sur-Seine sword that was not known when the paper was published needs to be added. Overall, the number of French swords does not change. Thirty swords were discovered in known contexts: seven (including the two Slovenian swords) come from burials, thirteen from land hoards, seven from rivers and three were isolated discoveries. From a geographical perspective, the funerary contexts are exclusively found in the eastern part of the distribution zone and the watery deposits are well to the west. The land deposits either hoards or single finds are recorded throughout the zone. These differences give the impression that the swords were not ‘consumed’ in the same way depending on which part of the distribution zone they were found in. The distribution of Tachlovice swords and their variants illustrates two preferential use zones: the western zone including France, Switzerland, West Germany, and Holland; and to the east: East Germany, the Czech Republic and Poland. In between these two zones Tachlovice swords are rare, but if we compare the distribution of Tachlovice swords with that of Auvernier swords, we can see that the Auvernier swords are mainly found in the area where the Tachlovice swords are rare, as if the two types exclude each other (fig. 18). This observation has also been mentioned by Sicherl and can be seen as particularly significant. It seems that these different production zones have chosen to adopt one type or the other, but rarely both types as if the two models were in competition with each other. Exceptions do however exist, the two sword types being found in the same region or even in one hoard as is the case for the Bothenheilingen hoard in Thuringia (Germany; Wüstemann 2004, 192 f. pl. 80, 505; 84, 518).

CONCLUSION The discovery of a new Tachlovice sword at Pont-sur-Seine has given us the opportunity to reconsider the French corpus of these weapons by replacing them in a larger European context. The different studies and analyses have shed new light on what materials were used and how these weapons were manufactured. Observations of the blade’s decoration, the hypotheses formulated as to the operational sequence and the tools used open up new perspectives in relation to the manufacture and circulation of Late Bronze Age swords. The next step is to study Tachlovice swords in the light of long bladed Bronze final III (Ha B) swords in different regions with particular emphasis on the study of the decorative motifs. This aspect will perhaps confirm the existence of distinct production zones and possible workshops in the case for Tachlovice swords, where very different technical solutions for the fabrication of the hilt were adopted within the same distribution zone. Translation: Rebecca Peake ACKNOWLEDGMENTS The present paper is the result of close collaboration with my restoration-curator colleagues of the Arc’antique lab in Nantes and the C2RMF in Paris. I would particularly like to thank S. Lemoine (Arc’antique) for overseeing the interventions of the two labs and M. Vieau (Arc’antique) and J.G. Aubert (Arc’antique) for their patience and the high quality of their work. B. Armbruster, S. Fournand and T. Nicolas have also contributed to the existence and the progress of this work, and I would like to thank them sincerely. Last but not least, my thanks go to B. O’Connor who kindly accepted to present this paper at the Vienna Workshop and to R. Peake for the precise translation of the text. BIBLIOGRAPHY Abauzit, P. 1973. Le déconcertant dépôt d’épées d’Aliès (Menet, Cantal). Bulletin de la Société Préhistorique Française 70, 9, 279–281. Anonyme, 1872. Découverte d’épées de bronze à Aliès (Cantal), Revue Archéologique, 1872, 2nd semestre, 337 f., pl. XXV. Armbruster, B. R. 2000. Goldschmiedekunst und Bronzetechnik. Studien zum Metallhandwerk

150

der Atlantischen Bronzezeit auf der Iberischen Halbinsel. Monographies instrumentum 15. Montagnac, Ed. Monique Mergoil. Bonnamour, L. 1990. Les armes de l’Âge du Bronze. In L. Bonnamour (ed.), Du silex à la poudre… 4000 ans d’armement en val de Saône. Catalogue d’exposition (1990–1991), 19–51. Montagnac, Ed. Monique Mergoil. Brun, P. 2000. D’étonnants sacrifices de richesses métalliques. Les abondantes découvertes de l’âge du Bronze dans la Seine. In L. Bonnamour (ed.), Archéologie des fleuves et rivières, 150–154. Paris, Ed. Errance. Chapotat, G. 1973. Les dragages de Grigny (Rhône): inventaire des trouvailles et étude des épées de l’âge du Bronze. Revue archéologique de l’Est et du Centre, 24, 343–359. Cowen, J. D. 1955. Eine Einführung in die Geschichte der bronzenen Griffzungenschwerter in Süddeutschland und den angrenzenden Gebieten. Bericht der RömischGermanischen Kommission, 36, 52–155. Déchelette J., 1913. La collection Millon, Paris. Gaucher, G. and Mohen J.-P. 1972. Typologie des objets de l’âge du Bronze en France, Fascicule I: Epées. Paris, Société préhistorique française. Kytlicová, O. 2007. Jungbronzezeitliche Hortfunde in Böhmen. Prähistorische Bronzefunde XX, 12. Stuttgart, Steiner. Krämer, W. 1985. Die Vollgriffschwerter in Österreich und der Schweiz. Prähistorische Bronzefunde IV, 10. München, C. H. Beck. Lambot, B. 1981. Quatre armes de l’âge du Bronze final découvertes anciennement en Alsace. Bulletin de la Société préhistorique française, 78, 9, 281–288. Mohen, J.-P. 1971. Quelques épées à poignée métallique de l’âge du Bronze conservées au Musée des Antiquités Nationales. Antiquités Nationales, 3, 29–46. Mohen, J.-P. 1977. L’Âge du Bronze dans la région de Paris. Catalogue synthétique des collections conservées au Musée des Antiquités Nationales. Paris, Ed. des Musées Nationaux. Morel, L. 1875. Découverte d’une sépulture renfermant une épée de bronze à Courtavant (Aube). Bulletin Monumental, 5th series, 3, 41, 250–259. Müller, F., Staniazeck, L. and Roscio, M. 2007. Une nécropole de l’âge du Bronze à Migennes (Yonne). Archéologie en Bourgogne, 8. Dijon, Publication de la DRAC Bourgogne - SRA. Müller-Karpe H. 1961. Die Vollgriffschwerter der Urnenfelderzeit aus Bayern. Münchner 151

Beiträge zur Vor- und Frühgeschichte, Bd. 6. München, C. H. Beck. Nicolardot, J.-P. and Verger, S. 1998. Le dépôt des Granges-sous-Grignon (Commune de Grignon, Côte-d’Or). In C. Mordant, M. Pernot and V. Rychner (eds), L’Atelier du bronzier en Europe du XX° au VIII° siècle avant notre ère. Actes du colloque international Bronze ’96, Neuchâtel et Dijon, T. III : Production, circulation et consommation du bronze, 9–32. Paris, CTHS. Piette, J., Rottier, S. and Depierre, G. 2005. Les pratiques funéraires du début du Bronze final dans les nécropoles de Barbuise-Courtavant et de La Saulsotte (Aube). In C. Mordant and G. Depierre (eds), Les pratiques funéraires à l’âge du Bronze en France. Actes de la table ronde de Sens (1998) 433–475. Paris, CTHS. Pilloy, J. 1904. Une épée en bronze découverte dans la rivière Oise. Bulletin Archéologique du Comité des travaux historiques et scientifiques, 272–277. von Quillfeldt, I. 1995. Die Vollgriffschwerter in Süddeutschland. Prähistorische Bronzefunde, IV, 11. Stuttgart, Steiner. Rottier, S. 2005. Pratiques funéraires originales de la phase ancienne du Bronze final de Barbey « Les Cent Arpents » (Seine-et-Marne). In C. Mordant and G. Depierre (eds), Les pratiques funéraires à l’âge du Bronze en France. Actes de la table ronde de Sens (1998), 459–474. Paris, CTHS. Royer, C. 1912. Choses de l’âge du Bronze. Bulletin de la Société archéologique et historique de Langres, 87, 1st trimestre, 356 fig. 3. Sicherl, B. 2008. Gedanken zur Schwertproduktion und –distribution in der späten Urnenfelderzeit am Beispiel der Tachloviceschwerter. In F. Verse, B. Knoche, J. Graefe, M. Hohlbein, K. Schierhold, C. Siemann, M. Uckelmann, G. Woltermann (eds), Durch die Zeiten…, Festschrift für A. Jockenhövel zum 65. Geburtstag, Internationale Archäologie, Studia honoraria, Bd. 28, 243–257. Rahden, Marie Leidorf. Sprockhoff, E. 1934. Die germanischen Vollgriffschwerter der jüngeren Bronzezeit. Römisch-Germanische Forschungen 9. Berlin/ Leipzig, de Gruyter. Stockhammer, P. 2004. Zur Chronologie, Verbreitung und Interpretation urnenfelderzeitlicher Vollgriffschwerter. Tübinger Texte, Bd. 5. Rahden, Marie Leidorf. Villes, A. 1995. Saint-Thibault (Aube): Epée hallstattienne en bronze. In Fastes des celtes anciens, Catalogue de l’exposition de Troyes

Sylvie Boulud-Gazo

et Nogent-sur-Seine, Sainte-Savine. Ed. des Musées de Troyes et de Nogent-sur-Seine, 103. Vuaillat, D. 1969. Les épées d’Auvernier et de Tachlovice, leur répartition en France. Etat

de la question en 1968. Bulletin de la Société Préhistorique Française, 66, 283–288. Wüstemann, H. 2004. Die Schwerter in Ostdeutschland. Prähistorische Bronzefunde, IV, 15. Stuttgart, Steiner. Sylvie Boulud-Gazo Université de Nantes - UMR 6566 CReAAH UFR d’Histoire, Histoire de l’Art et Archéologie Chemin de la Censive du Tertre BP 81227 44 312 Nantes Cedex 3 France

152

RITUAL OBJECT OR POWERFUL WEAPON – THE USAGE OF CENTRAL EUROPE BRONZE AGE SWORDS Marianne Mödlinger ABSTRACT During the period directly after the end of the Second World War, our past and especially the Bronze Age was interpreted as a rather peaceful period. Archaeological evidence of warfare and combat was ignored or, in the case of arms and armour, interpreted as symbolic weapons, not intended for actual use. During the last 15 years this point of view has changed – at least regarding weapons. Traces of use-wear on swords and spearheads, impacts of weapons on bones and of course fortified settlements were interpreted as what they are – evidence of combat and fighting. However, questions remain as to how these arms and armour were produced and by whom, as well as to why different types of e.g. swords were produced. In this work, the manufacture, development and usage of Central European Bronze Age swords will be discussed, showing the close interaction between craftsman and warrior. As a basis, 80 swords, mainly from Austria, but also from Hungary, Italy, Slovenia and the Ukraine, were analysed. To discover more information about the manufacture and usage of these swords, the alloy composition, the post-casting treatment and the casting quality were analysed using various methods such as SEM, EMPA, XRF, NRCA, x-ray, ToF-ND, 3D-Ct, metallographic analyses and visual examinations of the swords’ surfaces. KEYWORDS Bronze Age – swords – metallography – alloy composition – x-ray – manufacture – usage MANUFACTURE The design of a weapon, its shape and decoration, depend not only on its further usage but are also influenced by various factors. The craftsman or the craftsmen as well as the future owner or the buyer of arms and armour are part of a society, family, village, region and cultural group with its associated values, religious conceptions and working traditions. Therefore, the design of a weapon is not an arbitrary creation but rather the result of several cultural and technical influences

such as fighting techniques or religious aspects or rules, usually only allowing small variations – until the moment these influences demand changes or further development of the weapon. As P. Stockhammer (2004) pointed out in detail, many local decoration elements, which might also have symbolic character, may not have been understood in other regions. Furthermore, in some regions there was not only a preference for certain types of swords but also for particular lengths and shapes (see Molloy this volume, fig. 1–3). Compared with the (metal) hilt, which is usually decorated, the blade, as the active part of the weapon during a fight, is more open to technological changes. For this reason, this article will focus more on manufacturing techniques used to create the blade than on the decoration of the hilt. In Central and Eastern Europe, relatively few moulds have been discovered that were used to cast sword blades. In contrast with Northern Europe, Ireland and Great Britain, they are not made of clay but usually of various types of stone, e.g. sandstone. Known finds include HeilbronnNeckargartach (Germany), Font, Kanton Fribourg (Switzerland), Majaki (Ukraine), Pobit Kamăk (Bulgaria), Gura Idrici (Rumania), Velem (Hungary) and Piverone, Coriano and Castione dei Marchesi (Italy) (Mödlinger 2011). According to the primary usage of the swords, the gating is located at the upper end or the tip of the blade (see below). Only one mould, located in Erlingshofen (Germany), is known for the production of the metal hilt. This four part bronze mould is almost complete, with just the upper part missing. To demonstrate its usability, several casting experiments and simulations were carried out to prove it could be used directly for the casting of the bronze hilt (Wirth 2003). Nevertheless, the production of wax models for lost-wax-casting has to be considered as well. Regarding the small number of casting moulds for blades and even more for metal hilts in Central and Eastern Europe (and the later cast-on of metal hilts on the blade), it seems reasonable to assume the usage of clay moulds, where blade and hilt were cast using the lost-wax technique. Analysing the alloy composition of more then 50 swords – all following catalogue numbers refer to the ones in Mödlinger 2011 –, dating from Bz C to Ha B1, it is not surprising that the 153

Marianne Mödlinger

A B C Figure 1: A: Drawing of an x-ray of a metal-hilted sword. – B: Hilt of a metal-hilted sword of Type Liptau (cat. no. 24), containing the upper end of the blade and showing nearly no shrink holes there. – C: Blade tip of the same sword; a concentration of shrink holes, giving information about the location of the gate at the lower end of the blade

composition ranges from 7 to 13 %, with the majority from around 9–11 %, which is, due to its good workability and casting ability, the most common alloy in Bronze Age Europe (for the results of all swords, see Mödlinger 2011, appendix). Interestingly, the Sn-content does not depend on the type or dating of the sword. However, it is not surprising that the hardness as well as the amount of deformation increases with tin content, as does the appearance of (α+δ) Eutectoid. With EPMA, the CuSn-inclusions were also analysed and found to contain 16.9– 23.8 % sulfur, 0–6.6 % iron and 0.006–0.63 % tin. Trace elements in the bronze – nickel, iron, antimony, arsenic, zinc and lead – are usually below 0.5 %, with lead in just two cases up to 1.5 % (results SEM; flange sword type Tenja, metal-hilted sword type Q). One sword (unknown findspot, in the Natural History Museum in Vienna since 1899, Invno. 35.011) does not show much similarity with the others: The Sncontent is below 1.7 %, the Sb-content around 1 % (Mödlinger 2011, cat.no. 38). The alloy and 154

shape of the sword as well as the fact that it was cast in one piece, without separate bronze hilt or rivets, indicate an unusual, presumably non-violent use. The casting quality is rather poor with lots of blow holes across the blade and the solid hilt. Furthermore, the blade is broken or cut intentionally. X-ray analyses of the swords (with the exception of the swords from Styria, which at that time were not able to be removed from the museum) have provided deep insights into casting technology, the quality and usage of the swords as primarily stabbing or slashing weapons. In particular, the blades of metal-hilted swords demonstrate a different method of production, changing in Bz D. The older or earlier swords, namely the Achtkant type swords, intended to be used primary as stabbing weapons, were cast with the gating located at the upper end of the blade, the tang. This lead to a larger number of even bigger blow holes in the haft area, guaranteeing a better casting quality at the tip of the sword, making it more stable during the usage

Ritual object or powerful weapon – The usage of Central Europe Bronze Age swords

2A

2B

3

Figure 2: Traces of use-wear on the edges. – A: Cat.no. 50, a flange-hilted sword of Type Asenkofen (Bz C2), which is usually interpreted as only a stabbing weapon, which obviously was not the case. – B: a metal-hilted sword of Type Liptau (Ha A2; cat.no. 25). Figure 3: 3D-CT of sword cat.no. 40. Casting defects are clearly visible, especially on the blade (above) as well as on the hilt (below)

and during various impacts on the tip. The blow holes in the haft area – which is the weakest point of a sword – did not seem to be a big problem, as the stress of the material is flowing in one line straight to the hilt, not causing any side stress. This argues against the primary usage of these swords as slashing weapons. Using the sword as a slashing weapon usually places a large amount of stress on the haft area due to the side impact. If the blade is not fixed properly into the hilt, the rivet holes might break or, even worse, the blade can break in the haft area as well. To avoid this, the founders changed the location of the pouring gate at the end of Bz D and started to cast the blades of the metal-hilted sword with the tip upwards (fig. 1), accepting a worse casting quality at the tip of the sword, which might break more easily in this area. In such cases, the sword could still be used as a stabbing weapon, albeit with a higher risk of breaking the blade tip. Riegsee type swords demonstrate this change not only by revealing both ways of casting blades, but

also by the fact that the early blades of this type of sword primarily have the attributes of stabbing weapons used while the later Riegsee type swords have attributes for slashing and stabbing arms (see below). When we consider organic-hilted swords, no changes in the casting method or in location of the gating due to type or dating of the swords seem to occur (fig. 7). This might be caused by the fact that organic-hilted swords were already used as cut and thrust weapons much earlier and more often than metal-hilted swords, as suggested by the injuries documented on at least ten male skeletons of the cemetery of Olmo di Nogara, Verona (Italy) or the injuries from Elche and Alicante (Spain) (see below). Furthermore, swords of type Annenheim, Tenja or Traun show several traces of usage as a slashing weapon, e.g. worn or torn out rivet holes, resharpening or nicks and cuts on the lower third part of the blade (fig. 2. 4). Nevertheless, not all younger metal-hilted blades were cast with the tip upwards and 155

Marianne Mödlinger

A

B

C

D

Figure 4: A: Cat.no. 26; organic wrapping of the hilt (see Mödlinger and Grömer 2005). – B: Cat.no. 15; flashing, residues from the multipart-mould. This sword is the second known with residues of organic wrappings of the hilt. – C: Cat.no. 45; resharpening of the blade of a flange-hilted sword. – D: Cat.no. 40; sharpening of the edge

therefore guarantying a better casting quality, as was demonstrated by a 3D-Ct on a Ha B sword from lower Austria. Carrying out a 3D-Ct on this Schalenknauf type sword whose blade was broken in the haft area, the reason for the damage was easily identified. Hitting the target with the blade, the energy of the blow concentrated in the sword’s weakest point, the haft area, and, in combination with the bad casting quality, caused the breakage of the blade in this region (fig. 3). In summary, we are dealing with a perfect looking but low-quality sword, which was used as weapon until it broke (Mödlinger 2008). From the writing of Homer, we also know that swords seem to break from time to time during usage, e.g. when Menelaos´ sword breaks into five pieces while he was attacking Paris (Iliad 3, 361–363). In addition to the manufacturing traces found inside the sword using x-ray, 3D-Ct, metallography, etc., other hints about manufacturing techniques may be extracted from a sword’s surface. Unfortunately, many of these traces have been destroyed in the restoration process or treatment after finding: broken blades were soldered together, hammered straight, 156

the corrosion was eliminated using etchings or wire brushes or the swords were completely sand blasted. Such treatments tend to eliminate all traces of the manufacturing process or ornamentation. Nevertheless it was possible to document the usage of damaged punches on several swords, but no traces of the same punch were found on different swords. Traces of hammering are common, but unspecific. Usually they are found in the haft area of the hilt from attempts to improve the fit of hilt and blade (the shoulders of the hilt were hammered on the blade). Further visible traces of manufacture or the casting process include polishing, sharpening or relicts of the surface of the wax model of the sword hilt. Alternatively, the duct on the upper end of the hilt may still visible as a hole in the pommel or residues of the flashing of the multi-valve mould for the production of the hilt (fig. 4). Pores and blow holes can sometimes be seen on the surface of the blade (and hilt!) as well. Due to traces of use, such as worn and torn out rivet holes on the same swords, we can consider that even with the visible casting defects, the sword was used as a weapon nevertheless (fig. 7).

cat.no.

method

% tin

max. HV

δ-phase

δ-phase

eutectoid

eutectoid

slip lines

slip lines

205

20 - 25

x

x

-

35

XRF

10.4

120

2

XRF

10.9

126

15

x

-

-

37

SEM

10.3

3

XRF

7.2

132

20

x

-

-

38

XRF

1.7

4

EDX 10.1

135

20

x

x

-

39

XRF

5

EDX

11

167

40

x

x

-

40

EDX

6

XRF

10.6

213

20 - 30

x

-

-

44

8

XRF

9.7

237

30 - 50

-

-

-

46

EDX

13

9

EDX

13

205

20 - 30

x

x

-

47

SEM

8.3

10

SEM

9.8

113

15

x

x

-

48

EDX

9.2

199

12

SEM 11.7

120

15

-

-

x

50

EDX

9.2

14

EDX

12

193

15 - 20

x

x

-

51

EDX

15

XRF

11.5

200

20 - 25

x

x

-

54

16

EDX

9.6

220

30 - 40

x

-

-

18

EDX

9.5

100

50 - 60

-

-

-

21

XRF

10.3

143

20

-

-

22

SEM 10.3

199

20 - 25

x

23

XRF

8.4

108

15

-

24

XRF

9.2

186

15 - 20

-

-

-

67

EDX

25

XRF

7.8

97

10

x

-

-

68

EDX

26

XRF

7.9

144

20

-

-

-

69

EDX

27

XRF

9.1

257

40 - 60

x

-

-

70

28

XRF

11.5 corr.

65 - 70

?

-

-

30

SEM

8.2

100

10

-

-

-

31

SEM 11.4

207

20 - 30

x

x

32

XRF

10.7

172

80

x

-

33

SEM 12.7

196

15 - 25

x

34

SEM 10.2

222

30 - 40

x

% tin

% deformation

% deformation

SEM 12.7

method

1

cat.no.

max. HV

Ritual object or powerful weapon – The usage of Central Europe Bronze Age swords

15

x

x

-

244 40 - 50

-

-

x

106

10

-

-

-

7.8

163

20

-

x

-

10.4

229 30 - 40

x

-

x

141

x

-

-

106 60 - 70

-

x

-

176 15 - 20

x

-

x

20

x

x

-

269

5 - 10

x

-

x

7.5

97

10

-

-

-

SEM

11.5

155

20

x

x

x

55

SEM

8.7

163

25

-

-

x

56

EDX

11

210 20 - 30

x

x

-

-

60

EDX

9.8

130

20

x

-

-

x

-

61

EDX

10

88

10

-

-

-

-

-

66

EDX

11

240 35 - 40

x

x

-

13

137

x

x

-

12

196 15 - 20

-

x

-

10.5

146

x

-

-

EDX

10.5

252 30 - 40

x

x

-

71

SEM

8.8

180

20

x

-

-

74

SEM

8.7

94

10 - 15

-

-

-

x

75

EDX

9

-

77

x

-

79

-

-

EDX

9.5

20

20 40

87

15

-

-

-

214

10

x

x

-

x

x

x

135 15 - 20

Figure 5: An overview of the most important results of the metallographic analyses. The amount of deformation and the hardness of each edge was measured. The amount of tin is rounded

METALLOGRAPHY Concerning metallographic analyses, we have to consider that the results are only significant for the area sampled and not for the whole object – or, in the case of a sword – for the whole blade or the edges. Interpretations surrounding the manufacture of the whole objects are, strictly speaking, just assumptions. Supposedly, the aim of the smith was to create a blade which was sharp and

hard rather than with brittle edges and a tough core that was not too soft. In most cases, this goal was achieved in the area sampled, as the results of the metallographic analyses demonstrate. To obtain more information about the manufacture and post casting treatment of the blades, 53 sword blades were sampled for metallographic analyses. The samples were etched with FeCl3, ammonium persulfate and Klemm II solution to reveal the microstructure. The samples were ta-

157

Marianne Mödlinger

A

B

C

D

E

F

Figure 6: Metallography. – A: Cat.no. 21; etched with Klemm II, 100x, 5 mm from the edge inside the blade. – B: Cat.no. 8; etched with ammonium persulfate, 100x, edge. – C: Cat.no. 30; etched with ammonium persulfate, 200x; 5 mm from the edge inside the blade. – D: Cat.no. 33; etched with ammonium persulfate, 100x; deformed twins and slip lines are visible. The extreme elongation of the CuS-inclusions testifies that this blade must have received a total reduction close to 80%. – E: Cat.no. 75; corrosion on the blade; unetched, polarized light, 200x. – F: For the eye visible dendrites (casting structure) in the corrosion on the blade of sword cat.no. 24 158

dating

finding place

cat.no. (see also Mödlinger 2011)

Donau bei Persenbeug, NÖ

Bad Wimsb.-Neydharting, OÖ Bz C

Rovereto, Trento, Italien

Gratwein, Stmk.

Unbekannt

Albrechtsberg/Pielach, NÖ

Baierdorf, NÖ

Nöfing, OÖ

Nöfing, OÖ

Peggau, Stmk.

Gallneukirchen, OÖ

Lorch, OÖ

Oberburgau, Salzburg

Mitterndorf, Stmk.

Loretto, Bgld.

Lansach, K

Tultschnig, K

Bovec, Slowenien

4

5

6

7

8

9

10

159

11

12

13

14

15

16

17

18

19

20

21

metal-hilted Type Illertissen

metal-hilted Type Schwaig

metal-hilted Type Schwaig

tip

tongue

tip

tip

no x-ray

?

tongue

?

X

tip

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

int.

x

fun.

x

x

x

fun.

fun.

x

int.

x

x

int.

x

resharpening

Ha A1

Ha A1

Ha A1

metal-hilted Type Schwaig

metal-hilted Type Schwaig

Riegsee type sword

Riegsee type sword

Riegsee type sword

Riegsee type sword

Riegsee type sword

X

tongue

?

X

no x-ray

?

x

x

hole in disc

Ha A1

Ha A1

Bz D

Bz D

Bz D

Bz D

Bz D

Riegsee type sword

Riegsee type sword

Riegsee type sword

Riegsee type sword

Achtkant type sword

Achtkant type sword

tongue

tongue

tongue

tongue

ricasso

Bz D

Bz D

Bz D

Bz D

Bz C

Bz C

Achtkant type sword

Achtkant type sword

Achtkant type sword

Spatzenhausen

int.

x

x

x

nicks/cuts on the edge

Bz C

Bz C

Bz B1

Unbekannt

Greiner Strudel, OÖ

2

x

traces of use-wear

x

x

no x-ray

x

x

x

x

x

x

x

x

no x-ray

x

?

x

x

x

x

x

?

cracks at rivet holes; metal hilt

3

sword type X

location of gating

Early metal hilted sword

hole in pommel

Bz B1

tip missing

Au am Leithagebirge, NÖ

traces of manufacture worn-/torn-out rivet holes

1

general information

Ritual object or powerful weapon – The usage of Central Europe Bronze Age swords

160

Unterradlberg, NÖ

Wildon, Stmk.

Graz, Stmk.

41

42

Kleinmünchen, OÖ

36

40

Schlögen, OÖ

35

Vulchovica, Ukraine

Deinham, OÖ

34

39

Desselbrunn/ Rüsdorf, OÖ

33

Budapest, Ungarn

Greiner Strudel, OÖ

32

Unbekannt

Unterschauersberg, OÖ

31

37

Unterradl, NÖ

30

Bz B1

Bz B1

Ha B2

Ha B1

Ha A

Ha A

Ha A

Ha A

Ha A

Ha A

Ha A

Ha A

Ha A(1)

Ha A1

Ha A1

Ha A1

Ha A1

Griffplatten Type ‚Gambrin‘

Griffplatten Type ‚Wildon‘

metal-hilted Type Wörschach

metal-hilted Type Wörschach

metal-hilted Type D, variante 4

metal-hilted Type Q, variante 1

metal-hilted, conic pommel

metal-hilted, conic pommel

metal-hilted Type Erlach

metal-hilted Type Erlach

metal-hilted Type Erlach

metal-hilted Type Erlach

metal-hilted Type Unterradl

metal-hilted with ‚Pilzknauf‘

metal-hilted with ‚Pilzknauf‘

metal-hilted Type Aldrans

metal-hilted Type Aldrans

metal-hilted Type Liptau

metal-hilted Type Liptau

no x-ray

no x-ray

tongue

tip

X

?

tongue

tip

?

tip

tip

tip

tip

Knv

?

tip

?

tongue

tip

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

int.

x

x

x

x

x

x

x

x

int.

int.

x

x

x

x

x

ev.

resharpening

38

Greiner Strudel, OÖ

St. Pölten, NÖ

28

29

Tscheraditz, Tschechien

Unbekannt

27

Ha A1

Ha A1

tip

hole in pommel on x-ray

metal-hilted Type Liptau

x

x

x

nicks/cuts on the edge

Ha A1

x

traces of use-wear

x

x

x

no r.holes

x

x

x

recent

?

x

?

x

x

cracks at rivet holes; metal hilt

26

Năsăud, Rumänien

25

cat.no. (see also Mödlinger 2011)

Velká, Slowakei

finding place

24

dating

Vulchovica, Ukraine

sword type

23

location of gating tip

ricasso

metal-hilted Type Liptau

hole in disc

Ha A1

tip missing

Oberravelsbach, NÖ

traces of manufacture worn-/torn-out rivet holes

22

general information

Marianne Mödlinger

161

Stein im Jauntal, K

Freudenberg, K

Mauthausen, OÖ

Krajn, Slowenien

Bad Wimsbach-Neydh., OÖ Bz C2-D

Siegendorf, Bgld.

Annenheim, K

Annenheim, K

Unbekannt

Traun, OÖ

Graz, Stmk.

Graz, Stmk.

54

55

56

57

58

59

60

61

62

63

Bz C2-D

Bz C2-D

Bz C2-D

Bz C2-D

Bz C2-D

Bz C2-D

Bz C2-D

Bz C2-D

Bz C1

Bz C1

Bz C1

Bz C1

53

Bz C1

Bz C1

52

Grubbeg, Stmk.

49

Bz C1

Bz B

Linz-Fischdorf, OÖ

Gusen, OÖ

48

51

Prägarten, NÖ

47

Bz B2

Persenbeug, OÖ

Wels-Rosenau, OÖ

46

flange-hilted Type Traun

flange-hilted Type Traun

flange-hilted Type Traun

flange-hilted Type Traun

flange-hilted Type Annenheim

flange-hilted Type Annenheim

flange-hilted Type Annenheim

flange-hilted Type Annenheim

flange-hilted Type Tenja

Griffplatten Type

flange-hilted Type Nitzing

flange-hilted Type Nitzing

flange -ilted Type Mining

flange-hilted Type Asenkofen

flange-hilted Type Asenkofen

flange-hilted (?)

GriffplattenType ‚Gambrin‘

flange hilted Type B. Keszhely

flange hilted

Griffplatten Type ‚Sauerbrunn‘

tongue

no x-ray

tongue

tongue

?

tip

tip

tongue

?

?

?

tongue

tongue

?

tongue

Z

?

tongue

no x-ray

X

x

x

fun.

x

x

x

x

x

x

x

missing

missing

x

x

ev.

x

x

resharpening

Bz B2

Bz B1

tongue

traces of use-wear

x

x

x

nicks/cuts on the edge

50

Graz, Stmk.

cat.no. (see also Mödlinger 2011)

45

finding place

Mölltal, K

dating

44

sword type Griffplatten Type ‚Sauerbrunn‘

location of gating

Bz B1

tip missing

Pichlern, K

traces of manufacture

x

x

x

x

x

x

x

int.?

x

x

x

x

x

worn-/torn-out rivet holes

43

general information

Ritual object or powerful weapon – The usage of Central Europe Bronze Age swords

cracks at rivet holes; metal hilt

hole in disc

hole in pommel on x-ray

ricasso

Augsdorf, K

Baierdorf, NÖ

Langmannersdorf, NÖ

Millesi Teich, Völkendorf, K

Ardeschitza, Rosenbach, K

Schwabegg, K

Grein, OÖ

Velden, K

73

74

75

76

77

78

79

80

Oed bei Traun, OÖ

70

Riegersburg, Stmk.

Wimpassing, OÖ

69

72

Unbekannt

68

Wels-Waidhausen, OÖ

Linz, OÖ

67

flange-hilted Type Rixheim

Rod-tanged Type Terentola

flange-hilted Type Traun

162 blade fragment

flange-hilted (?)

flange-hilted

Rod-tanged Type Arco

Rod-tanged

flange-hilted Type Reutlingen

?

?

tongue

X

tongue

?

tongue

X

no x-ray

tongue

?

?

tip

tongue

tongue

tongue

missing

missing

missing

int.

x

int.

x

x

x

x

x

x

ev.

x

x

x

nicks/cuts on the edge

Ha A-B

?

Ha A

Ha A1

Ha A1

Ha A1

BzD-HaA1 flange-hilted Type Reutlingen

BzD-HaA1 flange-hilted Type Reutlingen

BzD-HaA1 flange-hilted Type Reutlingen

BzD-HaA1 flange-hilted Type Reutlingen

BzD-HaA1 flange-hilted Type Reutlingen

BzD-HaA1 flange-hilted Type Reutlingen

BzD-HaA1 flange-hilted Type Reutlingen

Bz D

Bz D

Bz C2-D

no x-ray

tip missing

71

Baierdorf, NÖ

cat.no. (see also Mödlinger 2011)

66

finding place

Priel, K

dating

65

sword type flange-hilted Type Traun

location of gating

Bz C2-D

traces of use-wear resharpening

Graz-Puntigam, Stmk.

traces of manufacture

no r.holes

x

x

x

no r.holes

no r.holes

x

x

x

x

x

x

x

wor-/torn-out rivet holes

64

general information

Marianne Mödlinger

cracks at rivet holes; metal hilt

hole in disc

hole in pommel on x-ray

ricasso

Ritual object or powerful weapon – The usage of Central Europe Bronze Age swords

Figure 7, previous pages: The most important traces of manufacture and usage for all the swords analysed. The point of balance was not measured on the organic hilted swords due to the fact that we do not know the material and the different weight of possible materials (bone, antler, hard and soft wood, etc.) the hilt was made of. – ‘X’ at ‘location of gating’ stands for an incomplete blade. The ‘int.’ at ‘tip missing’ stands for an intentionally broken off blade; ken where the blades were supposedly used the most, e.g. on blades with parallel edges, massive midrib and narrow blades closer to the tip of the blade and on swords with a leaf shaped blade with a broad, flat cross-section on the broadest part of the blade. Some swords were sampled vice versa to document if the treatment was uniform. The post-casting treatment such as hammering/hardening and annealing of the blade or the edges reveals information about the hardness, brittleness, flexibility and cutting potential of a weapon. Although a greater tin content and cold working increase the hardness of the edges, they also decrease toughness, potentially making the bronze brittle. Up to 14 % tin the alloy is usually cold worked, as we can see on the blades. Over 14 % (not the case for any sword blade analysed) bronze needs different treatment (e.g. quenching) if one wants to flatten or harden it. For all swords analysed, only the edges were cold worked but never the midrib or the inner parts of the edge. In these areas, changes in the microstructure appear to be the result of the annealing process. To increase deformation, the edge needs to be reheated and re-crystallisation can take place. While annealing the edge, the rest of the blade was affected by the heat treatment as well. All of the 53 swords analysed were cold worked on the sampled area of the edge after casting. Six swords did not receive any further treatment (fig. 6, B); the other blades continued to be worked. After the first circle of cold working, the edges were annealed. Seven swords did not receive any further treatment (fig. 6, A). The other 40 swords were cold worked and annealed in several circles again with cold working as the final step of production (fig. 6, C). Three out of these 40 swords were exposed to the fire of a cremation, so the intensity of cold working can only be observed in terms of the degree of deformation of the CuS-inclusions and we do not know the final step of post-casting treatment. On the edges, shrink holes and porosity were nearly eliminated during cold working and hardness increased up to 270 HV and the deformation on the edge reached up to 80 %, as indicated by the extreme elongation of the CuS-inclusion (fig. 6, D). Interestingly, there is no difference in the post-casting treatment of metal-hilted and or163

ganic-hilted swordblades. Nevertheless, the intensity of after-cast working decreased over the course of time, while the quality of the casting remained the same (i.e. no tendency indicates that swords were casted greater or lesser skill in different periods). USAGE – STABBING OR CUT AND THRUST WEAPONS? The sword could intentionally or unintentionally hit objects such as swords, shields, spears, javelins, daggers, organic weapons or bone during a fight, but also any other kind of hard object in the fray. These impacts lead to different damage marks on the blade or even no visible damage at all. Nevertheless, use-wear that indicates the analogous use of the swords can be extracted from nearly every sword. Although few swords show significant use-marks such as nicks and cuts, bent or broken tips, there are some indications that imply the usage of the swords as effective weapons: torn out rivet holes, re-sharpened blades, worn down ricassi, organic wrappings of the hilt, etc. (fig. 7). The background surrounding the (mis)-interpretation of the usage of swords as stabbing or slashing weapons is given by B. Molloy in this volume. In general, swords used (primarily) as a stabbing weapon usually have a narrow blade, a compact cross-section with small edges and a massive midrib as well as parallel edges. A hole in the disc pommel as well as ricasso on the upper end of the edges is unknown, despite the youngest swords Type Riegsee (as below). The point of balance on a sword, which is critical for the handling and the movement of the sword, becomes more important as the blade’s length and weight increases. On early swords – metalhilted swords usually used primary as stabbing weapons – the point of balance is usually close to the hilt so the blade can be moved with less force and more precise (Mödlinger 2007, 110 fig. 13). Early Riegsee type swords show exactly these attributes, however later ones are quite different: the blade is typically broader, the midrib flat and the parallel edged blade itself is often instead a leaf shaped blade. Furthermore, some

Marianne Mödlinger

of the Riegsee type swords show a ricasso, a typical attribute of swords used in a slashing mode: namely one out of 23 Austrian and ten out of 36 German Riegsee type swords (v. Quillfeldt 1995; Mödlinger 2009). The first ricasso usually dates to Ha A, but obviously can be dated back to the end of Bz D. Riegsee type swords with ricassi are unknown in other countries (according to the PBF-publications). We also have to take into account that some swords could have been produced without ricassi, which probably was added later, if the sword was still in use. Of the 119 Riegsee type swords known so far, only the four swords from the deposits at Uioara de Sus und Piricse I are part of a Ha A context, but without ricasso (Stockhammer 2004, 60; footnote 173). Analysing the point of balance of these Riegsee type swords, we see that it is in the middle of the Achtkant type swords, used primarily as stabbing weapons, and the point of balance of the younger Dreiwulst type swords, used primarily for slashing. With the development of the shape and cross-section of the blade, the point of balance of the Riegsee type swords migrated towards the tip. Later swords, which were used as cut and thrust weapons, put more weight on the main area of impact on the edge to cause even more harm to the opponent or his armour. Torn or worn out rivet holes on the blades and the shoulders of the hilt indicate heavy use of the sword as a slashing weapon. Alternatively, this could have also resulted from contact with other weapons or a hard object. However, we find these manufacturing traces not only on the later Riegsee type swords and Ha A swords but also on older metal-hilted swords and of course on organic-hilted swords, which proves at least a temporary usage of the swords as a slashing weapon. For example, on the radiographs of D. Ankner we see that seven out of the ten Riegsee type swords have ricasso work and torn out rivet holes. Nine of the 26 swords without ricassi also show these traces of usage (Ankner 1977). This leads to the conclusion that the Riegsee type swords are the missing link or an intermediate type between primary stabbing and slashing type metal-hilted swords (Mödlinger 2009). In addition, even the earliest swords analysed, dating to Bz B (e.g. cat. no. 1, 40–47), and later swords (see fig. 5. 7) show worn or torn out rivet holes, indicating the usage as slashing weapons as well. This phenomenon is found in Great Britain too – nine out of ten usually as stabbing swords interpreted weapons show use-wear typical for the application as a slashing weapon (Oakeshott 1960). In contrast, blade-to-blade impact marks caused by sword-to-sword contact appear only very 164

occasionally. Blade-to-blade impacts were normally avoided in order not to damage the edge. The small number of blade-to-blade impact marks on the sword edge can be explained by the fact that a blow is usually parried with the flat side of the blade, a shield or another weapon. Some of the swords, however, show impressive nicks and cuts on the edge. This use-wear does not seem to depend on the type or dating, the number of nicks and notches is too small. Nevertheless more usewear is typically present on the later metal-hilted swords. Concerning organic-hilted swords, there do not seem to be any connections with period or type. Surprisingly, a lot of these swords were rated as stabbing weapons, despite not being used accordingly. For a detailed description of the various types of nicks, cuts and notches on sword edges see Molloy in this volume. The two metal-hilted swords show remains of organic wrapping of the hilt (cat. no. 15. 26; see Moedlinger and Grömer 2005). No other metalhilted swords with organic wrappings are known. The wrapping was made of thin S-twill, which provided a firmer grip and lessened the danger of concussions of the wrist. Bent or broken points are documented on both slashing and stabbing swords. 19 out of 31 complete metal-hilted swords and 16 out of 25 complete swords with organic hilt show bent or broken points, which indicate the usage as a stabbing weapon. Also, traces on shields indicate that these swords were used to stab armour (e.g. the shields from the river Thames or the Type Yetholm shield from the Shannon at Athlone, M. Uckelmann this volume). Some of the swords have intentionally broken points or were deliberately destroyed: they were bent until unusable or broken into two or more pieces. This damage cannot be classified as use-wear. One of the main arguments that swords have been used as weapons is the presence of re-sharpened blades (see fig. 4, C). A notched blade had to be fixed (cold worked or ground) before further use. Re-sharpening of the edge might indicate either long-term utilisation with only a few resultant notches or a short usage period with more nicks on the edge (and maybe a poorly trained warrior). At least three of the metal hilted and nine of the organic hilted swords have re-sharpened blades. Although 40 swords with organic hilts were analysed, there does not seem to be any pattern concerning the re-sharpening of specific types of swords or certain periods. Even so, all the re-sharpened blades of full hilted swords belong to one type of swords; the Dreiwulstschwert type, a Ha A weapon used primarily for slashing.

Ritual object or powerful weapon – The usage of Central Europe Bronze Age swords

Though more organic hilted then metal hilted swords were re-sharpened, we cannot assume the latter were used less because they have the big advantage that it is possible to re-use the hilt and attach a new blade on an older hilt: the basic rivet type connection allows badly damaged blades to be easily exchanged. Reused blades can be easily identified, since they usually have surplus unused or dislocated rivet holes. Secondary used hilts can be identified in cases where they were connected with much younger blades. At least four / 10 % of the metal hilted swords analysed had a re-used hilt or blade (cat. no. 1, 13, 16, 32 and maybe even 31). Re-used blades are easy to recognise using x-ray – they have more rivet holes then necessary. Re-used hilts can be only recognized when the blade belongs to another type of sword or is obviously much younger then the hilt. Nevertheless, at least one sword has a re-used hilt: a Riegsee type hilt with a broad, flat, late Ha A blade (cat. no. 16). INJURIES Information about the number of injuries caused by swords is scarce in Bronze Age Europe – mostly due to the lack of bone preservation and anthropological analyses as well as the practice of cremation and the fact that many injuries did not leave any traces on the bone. The following provides a short overview of the known injuries caused or most likely caused by using a sword (Mödlinger 2011, 88–92; also see there for further references to the following finds discussed). Most sword-related injuries are located on the head and the upper part of the body; obviously the defensive armoury was not always sufficient (a project on analysing the Bronze Age defensive armour in Eastern Europe is already in progress by the author). A section of the parietal bone of the man from Wiligrad, Germany, was hacked off (15th century BC). The injury was caused by a sword used as a slashing weapon. The man from Alicante, Spain had a 98 mm long fracture of the right parietal bone caused by a sword. Evidence suggests that he survived the injuries and lived for some years in Bronce I (Bz C), most likely without any mental damage. In contrast, the approx. 18 month old baby from Elche, Spain, did not survive a deadly blow to the head caused by a sword c. 1700– 1500 BC. Further male victims of sword injuries are known from Sund, Norway (c. 1500–1050) with injuries e.g. on the front side (!) of the thoracic vertebrae or on the femur (Fovea Capitis)

(Fyllingen 2003). The approx. 40-year old man from Kråkerøy, dating from Ha A2/B1 suffered a blow to the right side of the cranium and an almost cut through thoracic vertebra. The 5060 year old man from Kehlheim, Germany, was hurt by a sword from behind on the lower area of the left parietal (Ha B). The highest number of Bronze Age people showing injuries caused by swords is known from the cemetery of Olmo di Nogara, close to Verona (Italy) (not published yet; pers. comm. M. Letizia Pulcini). The graves are dated to Bz B2–Bz C2, the man of grave 475 was buried with a sword Type ‘Sauerbrunn’ (Bz B2/C1) and a ceramic vessel (Bz C2). His left femur shows a strong impact, caused by a sword. A piece of the bone was cut off respectively broken off. The side behind the blade is polished and has a semicircular shape; the side in front of the blade has an uneven edge. In grave 168, the man was buried with a long dagger (Bz B2/C2). His skeleton is showing injuries caused by a strike on the left scapula from the top to the bottom. Possible victims who died from the usage of swords are known from: Antofts Windypit: possibly a woman, who was killed with a sharp blow to the head around 1350 BC; Mykene, Greece: a man in Grave Z with a sword-cut over the left eye and on the left temple, dating 1630– 1500 BC; Athenian Agora: individual with three traumas to its head, most likely caused by the blow of a blade, dating to LH IIIA and Stillfried, Austria: young girl with four impressions on the right side of the skull, caused by a blunt weapon and an injury above the right forehead caused by a sharp weapon, maybe a sword or an axe (late Urnfield Culture). SUMMARY Analysing a sword’s shape, surface and its material characteristics leads to a better understanding of the usage of a sword as either a stabbing and/or slashing weapon. While we cannot simplify the usage to purely stabbing or slashing, one was typically the dominant function. Indeed, most of the swords were used in both ways as indicated by their attributes: the casting process, the shape of the blade, the point of balance as well as various traces of use-wear on the blades and the hilt itself indicate the according usage of the weapon. Traces of use – worn- or torn-out rivet holes, worn down ricassi, nicks and cuts on the blade, bent or broken sword tips, re-sharpened edges, organic wrapping of the hilt, secondary use of blade and hilt as well as injuries on human

165

Marianne Mödlinger

bones – clearly demonstrate that these swords were used primary as weapons. But the exclusive use as symbolic weapons, status symbols or ritual objects does not require any functional optimisation such as a cold working of the blade to increase hardness or an optimum point of balance – ‘there can be no rituals or symbols without the reality of what they signify’ (Kristiansen 1999, 188). Furthermore, considering Bronze Age swords – which were in fact part of the first weapons invented purely to kill humans – as unused or un-usable weapons would demonstrate the absurdity of a three thousand-year-old improvement process. Nevertheless, a secondary function as a symbol of status or ritual object is possible and cannot be excluded. BIBLIOGRAPHY Ankner, D. 1977. Röntgenuntersuchungen an Riegseeschwertern. Ein Beitrag zur Typologie. Archäologie und Naturwissenschaften 1, 269–459. Bridgford, S. D. 2000. Weapons, Warfare and Society in Britain, 1250 – 750BC. Unpublished PhD thesis, University of Sheffield. Fyllingen, H. 2003, Society and Violence in the Early Bronze Age: An Analysis of Human Skeletons from Nord-Trondelag, Norway. Norwegian Archaeological Review 36/1, 27– 43. Kristiansen, K. 1999. The Emergence of Warrior Aristocracies in Later European Prehistory and Their Long-Term History. In J. Carman and A. Harding (eds), Ancient Warfare, 175– 189. Stroud, Sutton Hampe, R. 1986. Homer. Ilias. Dithingen, Reclam. Mödlinger, M. 2007. Herstellung und Verwendung mittel- und spätbronzezeitlicher Schwerter aus Österreich. Das Altertum 52, 101–130.

Mödlinger, M. 2008. Micro-X-ray Computer Tomography in Archaeology: Analyses of a Bronze Age Sword. Insight – Non-Destructive Testing and Condition Monitoring vol. 50, Issue 5, 323–326. Mödlinger, M. and Trnka, G. 2009. Untersuchungen an Riegseeschwertern aus Ostösterreich. In T. L. Kienlin and B. W. Roberts (eds), Metals and Societies. Studies in honour of Barbara S. Ottaway. Universitätsforschungen zur prähistorischen Archäologie 169, 350–357. Bonn, Habelt. Mödlinger, M. and Grömer, K. 2005. Metallographische und textilkundliche Untersuchungen an einem urnenfelderzeitlichen Schwert aus Nordböhmen. Archäologie Österreichs 16/2, 51–55. Mödlinger, M. 2011. Herstellung und Verwendung bronzezeitlicher Schwerter Mitteleuropas. Eine vertiefende Studie zur mittelbronze- und urnenfelderzeitlichen Bewaffnung und Sozialstruktur. Universitätsforschungen zur prähistorischen Archäologie 193. Bonn, Habelt. Molloy, B. P. C. 2009. For Gods or Men? A reappraisal of the function of European Bronze Age shields. Antiquity 83, 1052–1064. Oakeshott, E. 1960. The archaeology of weapons. London, Boydell and Brewer. Osgood, R. 2006, The Dead of Tormarton: Bronze Age Combat Victims? In T. Otto, H. Thrane and H. Vandkilde (eds), Warfare and Society. Archaeological and Social Anthropological Perspectives, 331–340. Arhus, Aarhus University Press. Stockhammer, P. 2004. Zur Chronologie, Verbreitung und Interpretation urnenfelderzeitlicher Vollgriffschwerter, Tübinger Texte 5. Rahden, Marie Leidorf. Wirth, M. 2003. Rekonstruktion bronzezeitlicher Gießereitechniken mittels numerischer Simulation, gießtechnologischer Experimente und werkstofftechnischer Untersuchungen an Nachguss und Original. Aachen, Shaker. Marianne Mödlinger Landesmuseum Kärnten Abteilung Urgeschichte Museumgasse 2 9021 Klagenfurt Austria

166

NEUTRON RESONANCE CAPTURE ANALYSIS AND THE BRONZE AGE Hans Postma, Peter Schillebeeckx and Marianne Mödlinger

ABSTRACT In this paper the application of neutron-resonance capture analysis (NRCA) to bronze artifacts from the Bronze Age period is discussed on the basis of some examples. The experiments were carried out at the GELINA facility in Geel, Belgium. With NRCA the overall elemental compositions of objects can be determined, while the remaining activation is negligible after a short waiting time. The relation with respect to other neutron-based methods is briefly explain-ed. This non-destructive analytical method is of particular interest to studying well-preserved and precious artifacts for which sampling is not permitted as is necessary with other methods to determine elemental compositions. KEYWORDS Epithermal neutrons – bulk mass analysis – time-of-flight spectroscopy – bulk mass analysis – elemental compositions – Bronze Age artefacts INTRODUCTION A source of information that is useful in the study of ancient bronze objects concerns their elemental compositions, which can vary considerably even if the study is restricted to the Bronze Age period. There are several different kinds of methods available to derive elemental compositions of objects, methods ranging from purely chemical to neutron-based, all with certain limitations either in the sensitivity for detecting certain elements, or the necessity to take samples, or lthe imited penetration of radiation into objects. In the case of the X-ray fluorescence method, the generated soft Xrays can only emerge from a thin surface layer. Three methods have been developed, which are based on the capture of neutrons by nuclei. The oldest method, known as Neutron Activation Analysis (NAA), was developed since the time nuclear research reactors became available. In its modern version high-resolution Ge-detectors are used to recognize gamma transitions of radioactive isotopes produced by the capture of

thermal neutrons, and thus can identify elements in objects. It cannot be used for complete objects because they will become strongly radioactive, and therefore cannot be returned to the owners for a long period. The second method is based on detecting gamma radiation emitted during the capture of neutrons. Also in this method Gedetectors are used to detect gamma rays with a high-energy resolution necessary to recognize elements. This so-called Prompt Gamma Neutron Activation Analysis (PGNAA) method uses thermal or sub-thermal neutron beams mainly from research reactors. The third method based on neutron capture has been developed rather recently. It is based on the occurrence of peaks (due to socalled resonances) in the probability of capturing neutrons plotted as a function of neutron energy. This method, called Neutron Resonance Capture Analysis (NRCA), requires a ‘white’ spectrum of neutrons over a sufficiently wide range of energies. Energies of these so-called epithermal neutrons can be determined with the time-of-flight (TOF) technique. That is, the time (t) needed for a neutron to travel a certain well-known distance (L) from its source to the sample where it is captured, provides the neutron velocity v = L/t and thus its energy using the equation E = ½m(L/t)2, in which m is the neutron mass. For measuring the time-of-flight it is necessary to have a start pulse and a stop pulse. The start pulse can be obtained from the reaction, which creates the neutron. The stop pulse can be derived from the detection of the prompt gamma radiation emitted during the capture process. The detection of this gamma radiation can be done with large detectors, which do not need to have good energy resolution, but good timing properties and low neutron sensitivities are essential. The latter is important since neutrons are also scattered from the object placed in the beam. The detection efficiency per captured neutron is considerably larger compared to NAA and PGAA. NRCA EXPERIMENTS AT THE GELINA FACILITY The NRCA measurements discussed in this paper are carried out at the GELINA facility of the EC-JRC Institute for reference Materials and

167

Hans Postma, Peter Schillebeeckx and Marianne Mödlinger

Figure 1: A typical resonance spectrum obtained with a Geistingen socketed axe on loan from the Limburg Museum in Venlo (NL). Some resonance peaks of Cu, Sn, Sb, As and Ag are indicated with their resonance energies in eV

Measurements (IRMM) in Geel, Belgium. The basic unit of the GELINA facility is a 150-MV accelerator, which is used to produce fast electrons in bunches as short as 1 ns with a maximum repetition rate of 800 Hz. Bremsstrahlung is produced by stopping these electrons in a disk of uranium. Bremsstrahlung in turn generates neutrons of high energies in this disk. These neutrons are slowed down in two small watercontaining vessels just above and below the Udisk. Neutrons exiting these containers constitute an energy spectrum with a flux roughly inversely proportional with neutron energy above about 1 eV. For NRCA experiments reported in this paper, epithermal neutrons from 1 to about 3000 eV are used. See W. Mondelaers and P. Schillebeeckx (2006) for more details about GELINA. A large number of evacuated and collimated tubes viewing these containers allow neutrons to reach experimental stations. For the reported NRCA experiments flight paths of about 13 and 29 m lengths have been used. Two kinds of filters are inserted into the beam. i) An overlap filter, which removes low energy neutrons from the beam in order to avoid overlap of consecutive bursts of neutrons. For this purpose a ¾ mm thick sheet of cadmium

metal or a disk containing 10B are used. The additional advantage of removing thermal neutrons from the beam is that activation of objects placed in the beam is reduced by a very large factor. ii) A filter, which reduces the strong Bremsstrahlung flash at the gamma detector used to signal capture events. Otherwise this detector will not function for several µs after each burst. Figure 1 shows an example of a resonance spectrum obtained at the GELINA facility, translating the TOF to neutron energy in electronvolts (eV). The example concerns a socketed axe of the Geistingen type. In this spectrum several resonances of various elements used for the analysis are indicated. Elements can be recognized on the basis of the energies of one or more resonances. Sometimes the width of a resonance can be helpful. Amounts of elements can be determined from the numbers of capture counts in resonances. An aspect in neutron physics, which plays an important role in the analysis of neutron resonance spectra concerns neutron scattering at nuclei. There are two kinds of scattering, i) at the nuclear potential well when the neutron approaches a nucleus, and ii) scattering inside the nu-

168

Neutron resonance capture analysis and the Bronze Age

Figure 2: This figure shows two examples of resonance lines (at left the 230-eV resonance of a thick piece (14 mm) of copper and at right the 21.4 eV resonance of antimony) with their scattering-capture structure at their high-energy sides

cleus. The first mode is called potential scattering and is independent of neutron energy. The second scattering mode shows the same resonance structure as capture and is known as resonance scattering. Due to scattering neutrons lose energy. This is the recoil energy, which depends on the scattering angle θ and is given by:

As a consequence resonance peaks will show a structure at its high-energy side due to single, double and multiple scattering processes followed by capture. It depends on the size and shape of the object and its significance increases with the thickness of the object. For the higher energy resonances this scatter-capture (SC) structure is largely separated from the capture peak, that is, only a small part of it is under the capture peak. However, at low-energy resonances the capture peak and SC-structure are more difficult to separate. An accurate determination of the number of counts in the capture peak requires correction for this phenomenon. Two examples of capture peaks and SC-structures are shown in figure 2.

are known and for an object with a simple shape. But for irregular shapes of most archaeological and art objects this is difficult if not impossible. In addition resonance parameters are not always known accurately. Therefore it is so far preferred in the study of archaeological and art artifacts to use calibration samples and to derive ratios of elemental amounts, which can later be transferred into absolute values if it can reasonably be assumed that all elements have been detected, with maybe the exception of some minor/trace elements. Say we consider two elements, X and Y, each with a resonance, µ and λ. The count rates Nλ and Nµ in these resonances are determined for an object as well as for a calibration sample of a known weight ratio. The weight ratio of two elements can be determined with the aid of the following expression:

The F-factors will be discussed later. The calibration factor Kµλcal is given by:

ANALYSIS OF RESONANCE SPECTRA Resonances observed in a spectrum as in figure 1 can be used to identify and to quantify elements in an object. Elements are mainly recognized on the basis of resonance energies; sometimes resonance widths are useful too. Elemental amounts can be obtained from the areas of resonance peaks. In principle it is possible to do this in an absolute way if the neutron flux and resonance parameters 169

In fact the determination of the weight ratio of two elements by this method depends on a double ratio of four count rates and therefore detection efficiencies for capture events at the two resonances and also the energy dependence of the flux cancel to a large degree. The F-factors in the above expressions are correction factors for the count rates Nµ and Nλ related to

Hans Postma, Peter Schillebeeckx and Marianne Mödlinger

Figure 3: Self-shielding factors (F-factors) plotted for some copper and tin resonances as a function of their areal densities in gram/cm2

the fact that the neutron flux diminishes during traversing the object, notably at resonances by capture and scattering. The F-factor is known as absorption or self-shielding factor; it can be calculated with the aid of the total cross-section taking Doppler broadening into account. For thin samples or weak resonances the self-shielding factors are close to one. They are known factors for the calibration sample, but they should be considered as variables for the object depending on the traversed thickness. Figure 3 shows Ffactors for some Cu- and Sn-resonances. For certain elements it is often possible to choose several resonances to perform the analysis. For instance three resonances for both copper and tin can often be chosen for the analysis of a tin-bronze. In that case weight ratios for nine pairs of resonances are obtained. Without correction for self-shielding these weight ratios are different if we are dealing with a thick sample and/or strong resonances. Taking the self-shielding effect into account and by varying the ‘effective’ areal densities of Cu and Sn the Sn/Cu weight-ratios can be made equal. With the minimum square method this leads to a mean Sn/Cu weight ratio plus a related effective areal density in g.Cu/cm2. Figure 4 shows an example of Sn/Cu weight ratios for the pairs of resonances indicated in this figure, uncorrected and corrected for self-shielding. NRCA can be applied to almost all elements except light elements. Elemental detection efficiencies can be quite different depending on properties of resonances. Elements with low-energy resonances can be well detected.

Broadly speaking, elements with suitably resonances below about 10 eV can be detected in the low ppm range, those with resonances between about 10 to 500 eV in the order of 10-3 to 10-4, and elements with resonances in the few keV region in the order of 10-2. Notably in the keV range overlap of resonances can be a serious problem requiring peak-shape analysis. In figure 5 the periodic table is shown with the elements marked with their first resonances below 10 eV, above 10 but below 100 eV etc. At

Figure 4: The Sn/Cu weight ratios of the blade of the Buggenum sword determined from nine pairs of resonances with the left side uncorrected and the right side corrected on the basis of a variance procedure, leading to an optimum areal density of 4.5 g.Cu/cm2. See Postma et al. (2009) for a detailed discussion about this method

170

Neutron resonance capture analysis and the Bronze Age

Figure 5: Periodic table showing the elements with different energy ranges in which the first suitable resonances occur. This indicates the applicability of NRCA. Elements with low energy resonances can be better detected compared to elements with only high-energy resonances

this point it should be noted that elements with suitably resonances only in the higher energy range (say several keV or higher) might be better studied with PGNAA. In this sense NRCA and PGNAA are complementary methods. BRONZE AGE OBJECTS STUDIED BY NRCA NRCA experiments have been carried out at the GELINA facility with several Bronze Age objects. These objects were made available for scientific studies under the strict condition that only non-destructive, non-invasive methods were going to be used. Some of these studies based on NRCA will be briefly reviewed in this paper.

Meuse. It is a very beautifully decorated sword originating presumably from the Danube region. With NRCA the tin content is determined at 10 locations along the length of the blade. The averaged tin content of the blade is 11.1 % while the hilt is made from a 13.1 % tin-bronze. The tin content of the blade increases slightly from the tip to the hilt. Together with the determination of the Sn/Cu weight ratio, the areal density of Cu is determined. The two parts have the minor elements Sb, As, Ag, In and some Co and Zn. Together with their ratios to Cu, the effective total density along the length of the blade has been determined. Details of the analyses combined with neutron diffraction results are published in Postma et al. (2009). The Juthphaas dirk

The Buggenum sword This sword, named after a village close to Roermond in the Netherlands, was found during the dredging of a lateral canal near the river

This is a curious and rare object found during dredging in the Rhine River near the town of Jutphaas, Netherlands. It is a non-typical blade of a dirk, unsuitable to be used as weapon, and therefore considered to have been a ceremoni171

Hans Postma, Peter Schillebeeckx and Marianne Mödlinger

Figure 6: A bar plot of six very identical socketed axes of the Geistingen type showing the strong variation in compositions. Two of them were on loan from the Limburg Museum in Venlo (NL), two from the Gallo-Roman Museum in Tongeren (B) and two from De Valkhof Museum in Nijmegen (NL).

al object. There are no indications for fixing it to a hilt. There are four other objects known of exactly the same shape and decorations, all with no hafting, but of different sizes. The elemental composition of the Jutphaas dirk is determined by NRCA and found to be very similar to three other objects of the same shape found at Oxborough, Beaune and Kimberley (Needham 1990). Socketed axes Compositions of several bronze axes have been determined by NRCA (Postma et al. 2007). One group is known as the Geistingen socketed axes, which are all very thin-walled and coming from the Meuse and Neder-Rhine area. They cannot be used as functional axes; they are considered to be ceremonial objects of the late Bronze Age (Butler and Streegstra 2002/2003). In the bar plot of figure 6 the results of NRCA analyses of six

Geistingen axes, which are very similar in shape and appearance, are plotted. Their compositions are remarkably different. Austrian swords Two all-metal swords from Museum Joanneum in Graz and two other swords from the Landesmuseum in Kärnten in Austria were brought to the GELINA facility for an NRCA study. They turned out to be tin-bronzes with Sn/ Cu ratios ranging from 0.10 to 0.15. These four swords show considerable variation of the minor elements Sb, As, Ag, In and Co. Remarkably the ‘Riegsee’ sword from the Graz museum has a hilt made from a very pure copper with very small amounts of Sb, Ag and In while As and Co are not detected in this hilt, that is, their ratios to copper are below 50 and 40 ppm. Its blade is quite a normal tin-bronze

172

Neutron resonance capture analysis and the Bronze Age

object location Sword 1

Sn/Cu weight ratio

Sb/Cu

As/Cu

Ag/Cu

Co/Cu

In/Cu

„effective thickness” gCu/cm2

Achtkantschwert, Museum Joanneum Graz No. 8977

A

0.1020 ± 0.0015

1.72E-4

7.82E-4

6.57E-5

2.90E-4

< 2e-6

2.9

B

0.1015 ± 0.0015

1.76E-4

7.42E-4

6.97E-5

2.90E-4

< 2e-6

3.4

D

0.0955 ± 0.0020

2.13E-4

6.72E-4

7.74E-5

2.60E-4

< 2e-6

5.7

Sword 2

Riegseeschwert, Museum Joanneum Graz No. 16910

A

0.1076±0.0015

3.10E-3

3.14E-3

2.40E-4

3.40E-4

3.90E-5

2.5

B

0.1035±0.0015

2.60E-3

3.40E-3

2.40E-4

3.40E-4

4.70E-5

2.5

D

0.1140±0.0015

4.80E-5

< 2e-5

2.70E-5

< 4e-5

6.00E-6

9.5

3.60E-5

2.7

2.30E-5

3.3

Sword 3 A

0.1245±0.0015

Sword 4 A

Dreiwulstschwert, Landesmuseum Kärnten No. A/194 1.59E-3

3.73E-3

2.95E-4

8.60E-4

Dreiwulstschwert, Landesmuseum Kärnten, No. 9644 0.1552±0.0020

1.39E-4

3.33E-4

1.02E-4

1.70E-4

Table 1: The weight ratios of elements in the four Austrian swords with respect to copper. A = lower part blade, B = upper part blade, D = top hilt (pommel)

with considerably larger amounts of these minor elements. Preliminary results of the elemental ratios are quoted in table 1. The Escharen double axe. This is a rare find from the southern part of the Netherlands. NRCA showed that it is mainly copper with the following additional elements: Sn 0.003 wt%, Sb 0.121 wt%, As 0.740 wt%, Ag 0.043 wt% and traces of Au and Co. Hence it is not made from a tin-bronze, but rather it is arsenic bronze. Butler (1995/1996) considers this to be an axe of the type Zabitz, variant Westregeln, and attributes this axe to the Bell Beaker period, or maybe earlier.

CONCLUDING REMARKS In this paper it is shown that neutron resonance capture analysis (NRCA) can be used to determine elemental compositions of artifacts in a nondestructive manner; it is not necessary to prepare the object before the measurement or to take samples. Whole objects can be studied within certain size limits. Compositions of bronze objects may vary considerably even within the series of artifacts from the Bronze Age, which look very similar. Activation of objects, directly after an NRCA run at GELINA, is normally low and after a short waiting period negligible. A drawback is the necessity to bring objects to dedicated accelerator facilities of which there exist only two in Europe, the GELINA and ISIS (Harwell, Great Britain) facilities.

ACTIVATION ACKNOWLEDGEMENTS Thermal neutrons especially activate objects. As mentioned earlier in this paper, filters are inserted in the beam at GELINA, which remove thermal neutrons to a very large factor. In the resonance region the total number of neutrons captured during NRCA experiments is too small at the GELINA facility for the significant production of long-living isotopes. In general it can be stated that the remaining activation of objects by neutron capture at GELINA is small and after a short waiting time fully negligible; that is, the activity is below internationally agreed levels below which an object is not considered to be a radioactive object.

All experiments reported in this paper are carried out at the GELINA facility, very skillfully operated by the dedicated staff of W. Mondelaers. We very much appreciate the loan of the Buggenum sword, Jutphaas dirk, and Escharen double axe from the National Museum of Antiquities in Leiden (Netherlands), the four Austrian swords on loan from Museum Joanneum in Graz and the Landesmuseum Kärnten, and the Geistingen socketed axes on loan from the Limburg Museum in Venlo (Netherlands), the Gallo-Roman Museum in Tongeren (Belgium) and the De Valkhof

173

Hans Postma, Peter Schillebeeckx and Marianne Mödlinger

Museum in Nijmegen (Netherlands). Part of the research mentioned in this paper has been carried out under the EU FP6 Ancient Charm project, funded by the European Commission under the contract No. 15311. BIBLIOGRAPHY Butler, J. J., 1995/1996, Bronze Age Metal and Amber in the Netherlands (II:1). Catalogue of the Flat Axes, Flanged Axes and Stopridge Axes, Palaeohistoria 37/38, 159–243. Butler, J. J., Steegstra, J., 2002/2003, Bronze Age Metal and Amber in the Netherlands (III:2), Catalogue of the Socketed axes. Part A. Palaeohistoria, 43/44, 263–319. Mondelaers, W., Schillebeeckx, P., 2006, GELINA, a neutron time-of-flight facility for

high-resolution neutron data measurements, Notiziario Neutroni e Luce di Sincrotrone 11/2, 19–25. Needham, S., 1990, Middle Bronze Age ceremonial weapons: new finds from Oxborough, Norfolk and Essex/Kent, The Antiquaries Journal 70, 239–252. Postma, H., Butler, J. J., Schillebeeckx, P., van Eijk, C. W. E., 2007, Neutron resonance capture applied to some prehistoric bronze axes, Il Nuovo Cimento 30C/1, 105–112. Postma, H., Amkreutz, L., Borella, A., Clarijs, M., Kamermans, H., Kockelmann, W., Paradowska, A., Schillebeeckx, P., Visser, D., 2009, Non-destructive bulk analysis of the Buggenum sword by neutron resonance capture analysis and neutron diffraction, Journal of Radioanalytical and Nuclear Chemistry DOI 10.1007/s10967-009-0405-z. Hans Postma RD&M, Faculty of Applied Sciences Delft University of Technology Mekelweg 15 2629 JB, Delft The Netherlands Peter Schillebeeckx EC-JRC IRMM Retieseweg 111 B2440 Geel Belgium Marianne Mödlinger Landesmuseum Kärnten Abteilung Urgeschichte Museumgasse 2 9021 Klagenfurt Austria

174

THE SINTASHTA BOW OF THE BRONZE AGE OF THE SOUTH TRANS-URALS, RUSSIA Andrey Bersenev, Andrey Epimakhov and Dmitry Zdanovich

ABSTRACT The Sintashta culture of the South Urals (20th– 18th centuries cal BC) stands out from the other Bronze Age cultures of Northern Eurasia due to substantial evidences for advanced war technologies. The category of distance weaponry is the most prominent among the finds from a military sphere of life. This work analyses artefacts which are the parts of bow made from horn. The authors propose different variants of reconstructions for strengthening the bow. KEY WORDS Bronze Age – Sintashta culture – South TransUrals – distance weaponry – strengthened bow INTRODUCTION The Sintashta materials have been the subject of intense and often heated debate. This cannot be explained exclusively by their relatively recent discovery and extreme originality; it is rather the fact that their significance goes far beyond regional importance. The Sintashta culture played a very important role in the formation of the Srubnaya and Andronovo cultural families of the Late Bronze Age, and the Sintashta people were unique in their attempt to create an early complex society in the Eurasian steppe (Koryakova and Epimakhov 2007). The discovery and investigation of the eponymous site of Sintashta came about in the 1970s (Gening et al. 1992), but the awareness of its value occurred only later. After subsequent large-scale excavations of Sintashta settlements and cemeteries, which illuminated the extraordinary characteristics of these sites, scholars began to change their views regarding many questions relating to the archaeology of the Bronze Age. A complete review of the history of the Sintashta culture is given in the introductory article by D. Zdanovich (2002). The sites of the Sintashta culture are dated to the 20th–18th centuries cal BC (Hanks et al. 2007). Due to the wide use of aerial photography and better site recognition, the area of distribution, number, size, and settlement pattern of sites are

now well established. In the Trans-Urals steppe, the systems of closed fortifications are only connected to the final period of the Middle Bronze Age (20th–18th centuries BC). The settlements of Sintashta type are concentrated in the northern steppe of the southern Trans-Urals (fig. 1). As a general rule, Trans-Uralian settlements are accompanied by cemeteries. These compactly located sites were united under the name of ‘Country of Towns’ given by their principal investigator, G. Zdanovich (Zdanovich and Batanina 2002; Zdanovich and Zdanovich 2002). Some archaeologists accepted this term as a sort of metaphor, others commented on it rather vociferously. The burial grounds are known from wider areas, including Kazakhstan (Logvin 2002) and Cis-Urals (Tkachev 2007). However, settlements have not yet been found in these regions. Twenty-three fortified settlements have been discovered during the past three decades. Most of them contained between one and four building horizons (Zdanovich and Batanina 2002). Out of the twenty three settlements, seven have been excavated to varying degrees. The best studied are Sintashta, Arkaim, Ustye and Kamennyi Ambar (Ol’gino). Twelve cemeteries are known in the Trans-Urals, but this number cannot be regarded as final, because cemeteries are not known for several settlements. Conversely, a few cemeteries have no visible connection to settlements (Epimakhov 2009). In the Trans-Urals, nine cemeteries have been excavated (200 burials), in the Pre-Urals six cemeteries (46 burials), and in northern Kazakhstan only one cemetery (28 burials). In the Pre-Urals, the Sintashta burials and kurgans are included in multi-stage cemeteries. Overall, the database of the Sintashta culture is rather rich, but the full extent of its potential has still not been realized. ARCHAEOLOGICAL REPRESENTATION Settlements The settlements are usually located in the central parts of the large river valleys, near a stream or ravine or river mouth. They occupy the spacious flat, dry ground of the first fluvial terrace, yet naturally fortified areas were rarely used. A wa-

175

Andrey Bersenev, Andrey Epimakhov and Dmitry Zdanovich

Figure 1: Map of the Sintashta culture. – 1 Stepnoe-M cemetery. – 2 Solntze II cemetery. – 3 Kamennyi Ambar-5 cemetery

part wall could reach 3 to 4 m in height and 4 m in width. In two fortresses, stones were also used on the outside. The internal space of the settlements is almost entirely occupied by buildings, organised into sectional blocks. Usually settlements consist of two circles of houses of standard dwellings, separated by a street and a central square. Generally, the centre of the settlement was free of any buildings. The number of houses, which are usually rectangular or trapezoidal, correlates with the size of a settlement. The house sizes are very similar, usually between 100 and 250 m2. The construction principles are standard with a frame-pillar construction exclusively. The longer longitudinal walls were adjacent to the next house.

ter barrier usually separated the cemetery from the settlement. The architecture of the Sintashta settlements is characterized by a closed system of fortifications. Their monumental character and complexity in comparison with other sites are very impressive. The fortification systems are represented by an oval or round shaped plan with radial grouped houses. In some settlements, the traces of reconstructions are recorded both archaeologically and by aerial photography. Despite the monumental character of fortification, real traces of military activity have still not been discovered. The closed fortifications consisted of ramparts and ditches, surrounded by a fence or wall. Fortified grounds enclose between 6000 and 20000 m2. The wood and earth building technique was commonly used. The ram176

The Sintashta bow of the Bronze Age of the South-Trans-Urals, Russia

All houses had a standard floor plan: an economic area, a living area, and a small antechamber or porch going out to the centre of the settlement. The heating system of such a large house was fairly efficient: on three sides, it was isolated from the cold by adjacent houses and the defensive wall, and the fireplace in front of the entrance created a form of heat air lock. Every house contained one or several wells next to the remains of a cupola shaped furnace where traces of metallurgy have been recorded. The living space occupied an area of 35–65 m2, which might accommodate 20–30 people (Epimakhov 1996; Grigory‘ev 2002). The excavated sites yielded mainly everyday material such as animal bones, pottery fragments, spindle whorls and tools for leather working. Yet a fairly large number of finds connected to metal production and metalworking (pestles, abrasives, nozzles, slag, metal drops) is also typical for the Sintashta sites. These traces are usually evenly distributed on the surface of a dwelling. The metal inventory from the settlements is represented by blade objects demonstrating the decline of traditions of the Circumpontic metallurgical complex and the forming of Eurasian metallurgical prototypes. These are slightly curved sickles, knife-daggers, fishhooks, awls, and chisels. Compared to the large number of finds connected to metal production, the number of stone moulds is not large. Special objects or markers of social status are almost completely absent in the settlement collections. Therefore, the settlement finds do not reveal socially determined areas nor specialised buildings and blocks. Cemeteries The Sintashta cemeteries occupy the flat portion of the first and second river terraces. The number of funeral complexes within one cemetery can vary from five pits to several dozen. Burial grounds are generally found close to the fortified settlements. The only difference between the burial grounds is the degree of their density. At present, funerary sites are visible as kurgans. According to some scholars (Grigoriev 2002), the Sintashta burial ground did not have a common mound; each grave was marked by a special construction that took the form of a small hill before being destroyed. The structures above the grave were made of wood and soil; there are no reliable examples of the use of stone, except the ones from the Pre-Urals area. Sacrificial deposits consisting of parts of animals and pots

were placed into the grave and the floor under the barrow, which constituted the major elements of the funeral area. The ways of organising the space under the kurgan – number, composition, orientation, and locality of graves with regard to the central grave – vary. About 85 % of multi-grave barrows have a quite clear planning structure. The funeral area was enclosed by a circular ditch (sometimes with a small bank) with entrances or causeways. The largest (3–4 m in length) or double graves usually have the central location. In the Sintashta period, the graves were oriented North/South. The other structures were situated around the central complex with different degrees of regularity. Therefore, within some barrows a clear contrast can be recognized between the ‘centre’ – more significant – and the ‘periphery’ – of lower status, but undoubtedly connected with the central grave. Stratigraphic observations confirm the multi-stage character of the mortuary complex formation. It is obvious that some burials were secondary to the primary complex. Alongside such an established pattern some divergence can be observed. The pit graves differ in size and complexity. They are more or less similar in form: rectangular with a stable length:width ratio (3:2). All grave structures were of wood and earth; stone is very rare. The mortuary chamber was furnished with a wooden frame covered by one, two, or even three ceilings made of logs and straw. The large graves yielded individuals of various gender and age compositions; all ages are represented. No clear markers of social inequality have been discovered. However, 5 % of all graves contained couples: a man and woman lying in a position facing each other. Some of the larger graves contained up to eight individuals. Nevertheless, 52 % of the burials were single individuals, mostly children. The deceased were placed on organic bedding on their left (rarely on the right) side in a contracted position with hands near the face. One of the most outstanding characteristics of the Sintashta funeral ritual is the abundance and variability of animal sacrifices, chiefly domestic animals: horse, cattle, sheep, and dog. The variability is also expressed in the numerous combinations of different animal body parts and where they were located. Whole skeletons obviously cannot be interpreted as funeral food; this is especially true in relation with horse depositions in combination with chariot traces or finds of cheekpieces. Notable regularities are: a horse usually accompanied a man; children and young women were usually given small horned animals, which

177

Andrey Bersenev, Andrey Epimakhov and Dmitry Zdanovich

Figure 2: Armament of the Sintashta culture. 1 mace. – 2 battle-axe. – 3 knife?? Chisel?. – 5 spearhead. – 6 cheek-piece. – 4. 7–16. 18–23 – arrowheads. – 17 armour parts. – Provenance: 1. 3. 5–23: Kamennyi Ambar-5 cemetery. – 2. 4: Sintashta cemetery. – Material: 1. 9–16 stone. – 2–5 bronze. – 6 horn. – 7. 8. 17 bone 178

The Sintashta bow of the Bronze Age of the South-Trans-Urals, Russia

being numerous occupied the lowest position in the hierarchy of species of sacrificed animals. The horse was on the top of this hierarchy. Grave goods are numerous, and the most frequent is pottery. Metal goods split into the following categories: weaponry – spearheads, axes, knife-daggers, arrowheads, darts; implements – knives, sickles, needles, awls, gouges; ornaments – pendants for braids, beads, pendants in one and half twists. There are as well many objects made from bone and horn. Above all, the cheek pieces should be noted. The objects made from bone also include arrowheads, small ‘spades’, knife handles, and spindle whorls. The set of stone objects consists of numerous arrowheads of two types (tanged or with a truncated base), pestles, ‘anvils’, and abrasive stones. There are no obvious luxury goods. In some cases, the remains of two-wheeled chariots were found, which were furnished with wheels with 10–12 spokes and an imitation of harnessing for a pair of horses. In general, the wheel traces are recorded in form of two parallel oval-elongated imprints with lens-like sections. They are usually situated 130–160 cm apart from each other. This distance can be regarded as the gauge of a chariot. We would not insist on that the ritual imperatively required placing the whole chariot into the grave (yet, undoubtedly, such examples are well known), but rather similar gauge size supports this hypothesis. These ‘wheel imprints’ also correlated with a post hole, which is usually located in or near the centre of the grave floor. Such ‘charioteer’s graves’ (14 %) often contained adult males accompanied by numerous weapons. Most of these are collective graves and they were located either in the centre or in the periphery (Epimachov and Korjakova 2004). The Gender and age of the deceased is reflected in the grave good composition. Weaponry except for knife-daggers (which may have been considered a tool rather than a weapon) are accessories of male burials. Ornaments, awls, and needles are considered female attributes. This differentiation appears to have taken place from about three years of age. Objects interpreted as the markers of social status and social power – maces, spearheads, chariots – are encountered both in central and in peripheral graves. Hence, the Sintashta funeral ritual is distinguished by its high variability in comparison with other Bronze Age cultures in the Ural. However, it is not easy to interpret these characteristics. Brilliance and diversity of the Sintashta funeral rite was probably partly conditioned by the necessity to assign a new system of cultural stereo-

types to a society at an early stage of its consolidation. The subsequent decrease in the variability of the burial rites demonstrates a standardisation of the ritual framework. Certain ritual elements passed to other forms (for instance from material to verbal ones), or they could remain implicit, and consequently they would not have had special attention and material realization. This process can be observed in the Petrovka culture, the ‘daughter-branch’ of the Sintashta culture. DISTANCE WEAPONRY This brief review demonstrates that the Sintashta sites possess extensive evidence of advanced war technologies: an enclosed system of settlement fortification, chariots and a full set of armaments in burials (fig. 2). But evidence of actual military activities at the settlements or even injuries on bones on the individuals in the cemeteries are absent. For instance, only around a dozen of stone and bone arrowheads were found at the Kamennyi Ambar (Ol’gino) settlement. However, the military sphere of life was clearly symbolised in mortuary rites, but the arrowheads have been most frequently placed in burials. The long-distance weaponry is represented by bone or horn parts of the bow. This long-distance weaponry is found in one fifth of all burials (46), but these are mainly disturbed (28). These graves contain both individual and collective burials. In total, 237 arrowheads and eight parts of bows are known. The typology of arrowheads is based on their material and construction. We divided the arrowheads into the following groups: stone (tanged or with flat base), metal (bronze) and bone. The difference in the quantity of arrowheads is also high. Morphological groups correlate with the quantity. The first two groups (stone arrowheads) numerically predominate: 48 % the tanged ones, and 38 % the ones with a flat base. In general, the tanged arrowheads are considered as a ‘visiting card’ of the Sintashta warfare, although they were chiefly found in cemeteries. The number of arrowheads varies from one to 20 in a closed Sintashta complex. Single arrowheads are known from undisturbed graves; i.e. that these items were deliberately placed here in such small quantity. Among the most striking examples is a male burial with a chariot (Krivoe Ozero cemetery) and a female burial from the Sintashta burial ground. Here so-called ‘quiver’ sets of arrowheads represent the special interest. Unfortunately, we have no confirmed evidences

179

Andrey Bersenev, Andrey Epimakhov and Dmitry Zdanovich

Figure 3: Parts of the Sintashta bow. 1 arrow rest. – 2 clasp or arrow rest? – 4–6 top-ends of the bow limbs. – Provenance: 1–4 Kamennyi Ambar-5 cemetery. – 5. 6 Stepnoe-M cemetery

for a reconstruction of the quiver assembly. These rare cases allow us, to suggest an arrow of about 60–70 cm in length. Nevertheless, at least 15 ‘quiver sets’ are potentially known, but the diversity in the quantity, shape and material of arrowheads within the same ‘quiver’, as in more than half of all cases, challenges any early assumptions. 180

THE SINTASHTA BOW: MATERIAL PROOF AND CONTEXT Research on the problem of the steppe zone bow during the Bronze Age is extremely scarce (Shishlina 1990; Nelin 1996). D. V. Nelin assumed the knowledge of the strengthened or

The Sintashta bow of the Bronze Age of the South-Trans-Urals, Russia

composite bow in the Sintashta population (Nelin 1996, 61). N. I. Shishlina came to the conclusion that two variants existed in the Srubnaya culture – the self bow and composite bow (Shishlina 1990, 32–34). Due to the paucity of reliably attributed finds, at the time Shishlina’s proposal could not be evaluated against the archaeological material. Until recently, research on the design of the Sintashta bow has been lacking. This can be explained firstly by the rarity of finds. This situation changed thanks to the investigation of the Stepnoe cemetery in 2007 (D. Zdanovich and E. Kupriyanova’s excavations) and by the exploration of the cemeteries of Solntze II (Epimakhov 1996) and Kamennyi Ambar-5 (Epimakhov 2005). ARCHAEOLOGICAL DATA Burial ground Solntze II, kurgan 11, grave pit 1. The pit is of a large size, and has wooden board and siding in the floor (ground) part. In ancient times the grave was ‘robbed’. It appears that such ‘robberies’ in many cases were of ritual character. Notwithstanding, total ‘robbery’ (in the course of which the anthropological remains had been removed before they finally lost the soft tissues) the burial preserved numerous animal sacrifices and a number of artefacts at the bottom of the grave: a vessel, a pair of bronze hooks and an object of elaborate shape suggesting for a bow limb (top-end). This last object is made from horn and the maximum dimensions are 72 x 32 x 27 mm (fig. 4, 3). The rectangular (in the plan view) base has a massive S-shaped asymmetrical hook at its reverse side. This hook has a longitudinal grooved channel apparently to achieve a maximal tight adhesive joint where it is attached to the wooden base. A tenon was used for the same aim. The round profile of the recessed hole was made in the out-shot of the hook. It is very likely that in addition to the discussed methods they were also used to wrap the base and arc of the bow from the outside. The smoothly sharpened end of the hook is lost. Stepnoe сemetery, kurgan 4, grave pit 13. Tomb 13 is one of the largest on the plateau of the kurgan. The dimensions of the burial chamber are 3.15 x 2 m, and the depth is 2.5 m. The ceiling of the chamber was held up by a supporting pole. On the ceiling (1–1.3 m over the pit floor) along its southern wall numerous remains of sacrificial animals were found, including a

whole skeleton of a dog. The pit was ‘robbed’ in antiquity. At the time of the ‘robbery’ the burial chamber was still hollow. At the bottom of the pit, fragments of two skeletons were discovered in situ (bones of lower limb with adjacent fragments of hip bones). The buried children, both about nine years old, possibly a girl(?) and a boy, were positioned parallel to each other on their left sides, with their heads directed to the northern part of the chamber. Notwithstanding that the grave was ‘robbed’, numerous grave goods were preserved: pottery (four vessels with ornaments), a cheek piece with pins, wooden goods with metal clamps, stone implements, 13 arrows with stone, bone and horn heads, animal astragals, silver adornment and also objects of horn, which can be interpreted as parts of bows. Two items were found in the south-eastern part of the grave. At the feet of the male skeleton near the southern wall of the pit lay a S-shaped, elaborately worked, object with a hook (no. 1; fig. 4, 1). The other artefact, apparently also related to the bow (no. 2; fig. 4, 2) was found at the eastern end of the wall of the pit. It can be attributed as a part on the end of the limb or as a ‘ledge’ like an original ‘backsight’, which was attached to the handle riser section and used as an arrow rest. The distance between these finds is 1.3 m. Both items lay sideways. The main features and dimensions (70 x 35 x 25–27 mm) of the S-shaped piece from the described burial bear close resemblance to the find from the cemetery of Solntze II. The main distinctions concern the length of the base, arrangement of the hook and several other minor characteristics. For example, the rubbing strip is complemented with a small ledge, the end of the hook has a flat cut, and the inner blind hole has rectangular cross-section. The horn ‘pan’ (40 x 19 x 16 mm) has a rectangular base with three transversal arranged salient edges, which form two ‘cavities’. Depending on the method of reconstruction it was used as the arrow rest or as the bow-string notch. The base of the artefact has a longitudinal groove (channel). Two tips of a bow limb were found in the western half of the same burial. They are made from the endings of elk prongs (fig. 3, 5. 6). These finds lay on the bottom, parallel and close towards one another, with sharp ends (points) directed to the pit wall. In the proximity of these items there were ten arrowheads: seven made of bone and horn, three made of stone. The arrowheads were scattered randomly on the floor of the pit.

181

Andrey Bersenev, Andrey Epimakhov and Dmitry Zdanovich

Figure 4. Parts of the Sintashta bow. – 1. 3. S-shaped hooks, for the upper end of the bow limb. – 2 ledge, which served as arrow rest (Variant 1) or as string notch (Variant 2). – Provenance: 1. 2 Stepnoe-M cemetery. – 3 Solntze II cemetery 182

The Sintashta bow of the Bronze Age of the South-Trans-Urals, Russia

Characteristics and preservation of the tips of the bow limbs diverge notably. The tips of bow limbs are 118 and 69 mm in length and have a practically identical external diameter of 19–20 mm. The inside of both objects has a dead taper sleeve, however, whereas in the short (no. 2, fig. 3, 5) the bush occupies almost all internal room (55 mm), then in the long one (no. 1, fig. 3, 6) – it hardly exceeds half of the internal room (69 mm). Both artefacts have a full-length perforation (6 mm) in the reduced neck portion (thinner part). The larger specimen is complemented with a small side cutting (this part of the second one was ruined in antiquity). Kamennyi Ambar-5 cemetery, kurgan 2. The excavation of this cemetery contributed information about the bow construction: in grave pit 2 a ‘pan’ was found, and in grave pit 15 – half-finished tips of bow limbs and a central lath. The first find comes from the central tomb, which is very big, but also a completely robbed burial and human bones, in contrast to numerous traces of animal sacrifices, were not found at all. The small (29 x 19 x 21mm) horn item found there, has a rectangular base with a longitudinal groove and two vertical step-shaped arises from the outside, which form the central channel for the arrow (fig. 3, 1). The set of objects from pit 15 (the same kurgan) is much bigger and included a quiver with different-types of arrowheads (fig. 2, 7–16. 18–23), parts of a bow, a knife and four vessels. The grave pit contained an individual burial of a young man and had a number of extraordinary characteristics unlike other burials (Epimakhov 2005): such as many animal sacrifices; an ancient weathered shoulder of a large animal put on the shoulder of the buried person; a fragment of gold-foil near the hand; and an infant burial on top of the ceiling. For the production of the half-finished tips of bow limbs horn was used, namely the antler tips of elk and saiga, of 136 and 115 mm length (fig. 3, 3. 4). In spite of the heavy polishing of the surface the objects have no traces of use-wear (Usachuk 2005, 187), but this is not surprising in the context of the incompleteness of the overall manufacturing process. Since such examples are well known, one may suggest that imitations of originally used objects were used in funeral rites: similar horn segments were found in other cemeteries (Solntze II, kurgan 11, burial 2; Kamennyi Ambar-5, kurgan 4, burial 8), as well as with weaponry. Nevertheless, these artefacts can be interpreted in the military sphere only with caution.

The interpretation of a bone plate with two opposite cuttings (79 x 29 x 6 mm) (fig. 3, 2) is less definite. Its inner surface has a weakly expressed longitudinal groove that suggests an attachment to a cylindrical surface (the handle riser section of a bow?), however, traces of usage (polishing, folding), according to A. N. Usachuk’s conclusion (2005), were visible on all surfaces including the cuttings. This fact has allowed Usachuk to offer another interpretation of the item: as a clasp. The context of the find does not exclude such a theory. In our proposal we have introduced a detailed record of the artefacts, which were found in the large tombs (both in individual and collective ones). These graves contained burials of adults of both genders and children. In some cases stone, horn and bone arrowheads were found alongside bow parts and other grave goods. Authentic wooden bow parts have not yet been discovered and the location of the horn parts does not give clear answers. RESULTS OF USE-WEAR ANALYSIS The studied S-shaped objects, notwithstanding their formal resemblance, essentially vary in their manufacturing technology. Thus, workmen used carving to make the artefact from Stepnoe (fig. 4, 1) as main technique of the further shaping of the item. This can be seen from differently layered tracks in the form of linear recessed and protruding grooves. This regularity in the tracks’ combination suggests that the tool used for shaping had a small working part (a blade with two sharpened edges) and apparently could have been made of stone, which is confirmed through the ‘wave-like’ bottom of the piece. The main technique of the after-treatment of the corresponding part from the Solntze cemetery (fig. 4, 3) was grinding and probably even polishing, because the surface of the studied item is very smooth. ‘Polishing out’ the surfaces during maintenance can be excluded, since, as far as surface configuration is concerned, it is rather elaborate to undergo uniform friction in the entire area. The second distinctive technological method is the manufacture of a recessed hole for a fixing pin. The square (in plan) hole on the topend piece from the Stepnoe cemetery was made by slotting. We failed to determine if pre-drilling had taken place, as in the bottom of the hole such traces were not detected. Another top-end had a round hole, made with the help of a mounted and flat perforator (borer, drill). Both skilled

183

Andrey Bersenev, Andrey Epimakhov and Dmitry Zdanovich

workmen used all accessible manufacturing methods of their time, beginning with the material selection, creating it and finishing with polishing the item. Compared with these elaborated and even artistically manufactured top-ends, the tips of the bow limbs (Stepnoe, fig. 3, 5. 6) look plain, if not ragged. Tool marks were not detected on the outside surfaces of the two studied items and the inner surface of the longitudinal hole for fixing the piece onto the ends of the bow limb also did not show any tool traces. The ledge (?) from Stepnoe (fig. 4, 2) was made by chiselling with the help of a tool with a metal tip. The traces on one of the guiding grooves give evidence of this – they were not completely polished while in use. The hollow on the base is made by scraping. The long longitudinal tracks preserved at the surface of the base very probably did not undergo any finishing touches (shaping or polishing), probably for a tighter connection with the bow arc when glued. Shock-absorbing traces are most clearly presented on the two top-ends (fig. 4, 1. 3), since here a wearing down (and loss) of parts of the items material is visible. The abrasion of some sections is considerable and sums up to 3–5 mm (!) from a supposed newly-worked piece. Taking into consideration a rather high hardness of the used material, it is possible to presume that functional stress on this structural bow part was enormous. The question of use depends not only on the bow strength, meaning the power of the string on the top-ends, but also on the number of shots made. In the introduced cases, we can talk about thousands of shots. A further interesting observation of the use-wear traces on the hollow of the winding hook of the top-end is a clearly defined direction of abrasion: the first one described – right, the second one – left. This appears to show evidences for different bow modifications: for a left-hander and for a right-hander. The fact that the top-ends, as structural parts of the bow, were used until ceasing to function, does not contradict this interpretation. The surfaces of the horn tips of the bow limbs from the cemetery at Stepnoe show no traces of tool working and use-wear. Only insignificant traces, under 1 mm, of wear on the hole rim for the bow string’s end (fig. 3, 6), which drastically differs from the worn down hollow of the hook, for the bow-string’s end, and the choppy rim of the socket let us suppose a moderate lifetime of this bow part. No definite conclusions can be stated about the second (smaller) ending, due to its poor preservation and some structural 184

features. The socket of this ending (fig. 3, 5) is so deep that it is cross-cutting the hole for the bow-string’s end and this probably caused damage to the ending’s wall. It is highly likely that this tip is not a piece that was used and damaged while in service, but rather a failed attempt to make such a horn part of the bow. Traces of usage on the ‘ledge’ from Stepnoe (fig. 4, 2) can be clearly seen on the middle and inner parts of the low edge of this piece. A wear-down of up to 0.5–0.8 mm of the walls can be observed, this ostensibly proves a theory of usage of this bow element as an arrow rest during the shooting. VARIANTS OF RECONSTRUCTIONS The find-context and traces of use-wear allow reconstructions of two variants of the Sintashta composite bow to be proposed. Variant 1 (after A. Bersenev and A. Epimakhov) A conical piece was used as a bottom bushing and string notch (fig. 3, 3–6). It was placed on the end of the lower bow limb and ensured a firm fastening of the bow string and also guaranteed the fast substitution of the string in case of breakage. The duality of finds and their location side by side are indirect evidences of this version. The parts where not attached to the limb, when positioned in the grave. The upper limb end was furnished by a S-shaped top-end (fig. 4, 1. 3) which was fastened firmly to the wood base. The load on this part was considerable. One more horn element, with a rectangular concave base and cross-located lugs (fig. 3, 1; 4, 2) was positioned on the middle of the handle riser section over the grip. It could be used as arrow rest (or plate) and has the marks of thorough fixation. Variant 2 (after D. Zdanovich) The second version has been developed due to the finds from the Stepnoe cemetery. The conditions of discovery suggest that fragments of two polytypic bows were found. One was discovered in the western part of the grave and the other on in the eastern part. Type 1. The distinctive feature is the mounting of tips at both limb ends (fig. 3, 5. 6). Since they were found parallel to each other in the grave, they can be interpreted as the result of deliberate

The Sintashta bow of the Bronze Age of the South-Trans-Urals, Russia

destroying of the bow in half. Such a breakage with this kind of weapon is founded on the combination of flexible limbs and rigid ends. Type 2. This reconstruction is characterised by the presence of short laths on the ends of limbs. Thus the parts with cross-located lugs (fig. 4, 2) did not serve as arrow rests as interpreted in Variant 1, but are seen as string notches. The S-shaped element (fig. 4, 1) was placed on the end of the opposite limb. Both artefacts were manufactured in a very similar manner and identical material was used. Despite these arguments, the archaeological context of the finds does not suggest they should be interpreted as parts of the same bow. It is well-known that most of the ancient bow strings were fitted with loops on both ends to ensure better string safety. However, another method of attaching is possible, where one end of the limb is bent and the string is fastened by using a self-tightening knot (Medvedev 1966, 19). The Sintashta bow (type 2) is probably characterized by this method. In this case the lath (fig. 4, 2) served to fasten one end of the string tightly and the S-shape top-end (fig. 4, 1) was used to put the bow in a run position with a detached string. Another assumption on the bow morphology of type 2 can be made: the limbs were asymmetric and of different thickness (1.6–1.8 and 2.7 cm). Variant 2 proposes the existence of two well developed strengthened bow types in the Sintashta culture. CONCLUSION Summing up our review we can propose that the Sintashta population used a technology to produce and wield a strengthened bow. To the wooden bow limb parts of horn (or bone) were added. Short lengthed arrows (50–65 cm) point indirectly to smaller sized bows, because it implies a small brace height. The proposed versions of reconstructions of the strengthened bow do not deny simple variants of bow design as the ‘self bow’. Considerable variation of materials, morphology and the large quantity of arrowheads are indirect arguments for this. The tradition of including a composite bow or parts of it into the grave declined in the post-Sintashta period by degrees. Uralian and Siberian bows of later periods differ from the Sintashta types significantly (Solov’ev 2003, 34. 42). It is possible to say, that several independent traditions of bow making in North Eurasia can be identified.

BIBLIOGRAPHY Epimakhov, A. V. 1996. Kurgannyi mogil’nik Solntze II – nekropol’ ukreplennogo poseleniya srednei bronzy Ust’e. In N. O. Ivanova (ed.), Materialy po arheologii i etnografii Uzhnogo Urala, 22–42. Chelyabinsk, Kamennyi poyas. Epimakhov, A. V. 2005. Rannie komplexnye obschestva Severa Tzentral’noi Evraszii (po materialam mogil’nika Kamennyi Ambar-5). Vol. 1. Chelyabinsk, Chelyabinskii dom pechati. Epimakhov, A. V. 2009. Settlements and necropolises of the Bronze Age of the Urals: opportunities of reconstruction of social dynamics. In B. K. Hanks and K. M. Linduff (eds), Social Complexity in Prehistoric Eurasia: Monuments, Metals and Mobility, 74–90. Cambridge, Cambridge University Press. Epimakhov, A. V. and Koryakova, L. N. 2004. Streitwagen der eurasischen Steppe in der Bronzezeit: Das Wolga-Uralgebirge und Kasachstan. In M. Fansa and S. Burmeister (eds), Rad und Wagen. Der Ursprung einer Innovation. Wagen im Vorderen Orient und Europa, 221–236. Mainz, von Zabern. Gening, V. F., Zdanovich, G. B. and Gening, V. V. 1992. Sintashta: arheologicheskie pamyatniki ariyskikh plemen Uralo-Kazahstanskikh stepei. Vol. 1. Chelyabinsk, Uzhno-Ural’skoe knizhnoe izdatel’stvo. Grigoriev, S. A. 2002. Ancient Indo-Europeans. Chelyabinsk, Rifei. Hanks, B. K., Epimakhov, A. V. and Renfrew, A. C. 2007. Towards a Refined Chronology for the Bronze Age of the Southern Urals, Russia. Antiquity 81, 353–367. Koryakova, L. and Epimakhov, A. V. 2007. The Urals and Western Siberia in the Bronze and Iron Age. Cambridge, Cambridge University Press. Logvin, V. N. 2002. The Cemetery of Bestamak and the Structure of the Community. In K. Jones-Bley and D. Zdanovich (eds), Complex Societies of Central Eurasia from the 3rd to the 1st Millennium BC: Regional Specifics in Light of Global Models, 189–201. Journal of Indo-European Studies, Monograph Series 45. Washington, Institute for Study of Man. Medvedev, A. F. 1966. Ruchnoe metatel’noe oruzhie 9luk i strely, samostrel. VIII-XIV vv. Svod arheologicheskih istochnikov. Vyp. Е136. Moscow, Nauka. Nelin, D. V. 1996. Luk i strely naseleniya Uzhnogo Zaural’ya i Severnogo Kazahstana epohi bronzy. In V.A. Ivanov (ed.), XIII Ural’skoe

185

Andrey Bersenev, Andrey Epimakhov and Dmitry Zdanovich

arheologicheskoe soveschanie, 60–62 Chast’ I. Ufa, Vostochnyi universitet. Shishlina, N. I. 1990. O slozhnom luke srubnoi kul’tury. In S. V. Studzitzkaya (ed.), Problemy arheologii Evrazii. Trudy Gosudarstvennogo istoricheskogo muzeya. Vyp. 74. 23–37. Moscow, Gosudarstvennyi istoricheskii muzei. Solov’ev, A. I. 2003. Oruzhie I dospekhi. Sibirskoe vooruzhenie: ot kamennogo veka do srednevekov’ya. Novosibirsk, Infolio-press. Tkachev, V. V. 2007. Stepi Uzhnogo Priural’ya i Zapadnogo Kazahstana na rubezhe epoh srednei i pozdnei bronzy. Aktobe, Akyubinskii oblasnoi tzentr istorii, etnografii i arheologii. Usachuk, A. N. 2005. Kamennoambarskie psalii (trassologicheskii analis). In A. V. Epimakhov (ed.), Rannie komplexnye obschestva Severa Tzentral’noi Evraszii (po materialam mogil’nika Kamennyi Ambar-5). Vol. 1, 179– 189. Chelyabinsk, Chelyabinskii dom pechati. Zdanovich, D. G. 2002. Introduction. In K. Jones-Bley and D. Zdanovich (eds), Complex

Societies of Central Eurasia from the 3rd to the 1st Millennium BC: Regional Specifics in Light of Global Models. Journal of IndoEuropean Studies, Monograph Series 45, ix–xxxviii. Washington, Institute for Study of Man. Zdanovich, G. B. and Batanina, I. M. 2002. Paleography of the Fortified Centers of the Middle Bronze Age in the southern TransUrals according to Aerial Photography Data. In K. Jones-Bley and D. Zdanovich (eds), Complex Societies of central Eurasia from the 3rd to the 1st Millennium BC: Regional Specifics in Light of Global Models. Journal of Indo-European Studies Monograph Series 45, 120–138. Washington, Institute for Study of Man. Zdanovich, G. B. and Zdanovich, D. G. 2002. The ‘Country of Towns’ of Southern Trans-Urals. In M. G. Levin (ed.), Ancient Interactions: east and west of Eurasia, 249–263. Cambridge, McDonald Institute for Archaeological Research. Andrey Bersenev Archaeology Scientific Centre Communy st. 68 454000 Chelyabinsk Russia Andrey Epimakhov Institute of History and Archaeology, South Ural branch Communy st. 68 454000 Chelyabinsk Russia Dmitry Zdanovich Chelyabinsk State University Brat’ev Kashirinykh st. 129 454001 Chelyabinsk Russia

186

THE FUNCTION OF BRONZE AGE SHIELDS Marion Uckelmann

ABSTRACT With the end of the Middle Bronze Age and the beginning of the Late Bronze Age shields made of a single piece of bronze sheet come into use. All over Europe about 86 of these metal shields are recorded, as well as two wooden and one leather shields and two wooden shield formers from Irish bogs. The main focus of distribution is in the British Isles, followed by a larger group in southern Scandinavia and more scattered finds from Germany, Poland, Czech Republic and the Carpathian basin. This paper will take a closer look at the possible function of the shields based on their technological characteristics, for example metal thickness, weight, strengthened rim and riveted on parts. New research on these features, together with metallographic and material analyses will be considered and compared with the visible signs of use wear on the shields, in order to evaluate a possible function as protective armour in combat. Old and newer experiments on the use of shields will be taken into account as well. KEYWORDS Shields – sheet metal work – combat – votive deposits INTRODUCTION Shields are an important part of the warrior panoply in the Bronze Age. The main part of the shields would have been made out of wood or leather, but due to their organic nature very few have come down to us. But widespread images of shields indicate that they were known in many parts of Europe and their use can be assumed from the late Middle Bronze Age and the Late Bronze Age (from the 13th century BC) onwards. Two wooden and one leather shield as well as two wooden shield-formers were found in Irish bogs. A C14 dating of one the organic shieldformers suggests the use of round shields in the British Isles already in Early/Middle Bronze Age times. The majority of the shields were found in ‘wet’ undatable contexts such as bogs and

rivers and only very few finds from hoards in the Carpathian Basin and in one case from Denmark can be dated through their associations. With the end of the Middle Bronze Age and the beginning of the Late Bronze Age, organic shields were adapted to metal forms, made from one piece of bronze sheet. Today about 86 of those metal ones are known from Europe, some survived only in fragments and some are known only through literature (Uckelmann, in prep.) DISTRIBUTION An astonishingly high number of around 50 shields were found in the British Isles, comprising more than half of all known pieces. A third belong to the Nordic Bronze Age of Southern Scandinavia and Northern Germany, and smaller groups come from Southern Germany and the Carpathian Basin, as well as single shields from Poland and the Czech Republic (fig. 1). This distribution map must be complimented by evidence of possible other shields, for example represented through nails, probably used as fittings on wooden shields. A smaller number of these are present in rich graves of the Middle Bronze Age Tumulus culture of Southern Germany (Stary 1980). But also other metal fittings that might have been used in some way for uncommon types of shield must be considered (Uckelmann, in prep.). It is important to add to this distribution of where shields were known and used to where images of shields are present. Many round shields are engraved on the so-called warrior steleas in Iberia, which show a very close resemblance to actual shields. The earliest encravings bear the shield as central symbol (Harrison 2004). Shields are also represented relatively frequently in many Scandinavian rock carvings, but there they are more roughly executed (Coles 2005). Some of the bronze statues from Sardinia known as bronzetti carry round shields, but in general they show no resemblance to the metal ones, and probably relate to organic forms (Thimme 1980). Many shield images and models are known from the Eastern Mediterranean, but almost no originals have survived. Different shapes are common in this region, such as the figure of eight or tower

187

Marion Uckelmann

Figure 1: Distribution of the different types of Bronze Age shields. – Multiple finds of shields of the same type are indicated with a letter in or next to the symbol: B = Beith (5–6); Ba = Bamberg (2); C = Church Wilne (2); F = Fröslunda (16); H = Herzsprung (2); Y = Yetholm (3). – Multiple finds of shields of different types are indicated with symbols in or close to the other symbol. – Shields of unclear provenance are positioned with an arrow next to the country 188

The function of Bronze Age shields

Type

Number

Diameter

Thickness

Weight

Lommelev-Nyírtura

6

67 cm

0,9–1 mm

2,2 kg

Nipperwiese

8

38–44 cm

1–1,3 mm

1,5–2,2 kg

Harlech

6

50–68,9 cm

0,1–1 mm

1–2,75 kg

Coveney

2

45,7 / 52,5 cm

0,3–0,5 mm

0,9/1,2 kg

Athenry-Eynsham

6

23–35 cm

0,3–1,2 mm

0,9–1,2 kg

Yetholm

25

55–70 cm

0,4–0,7 mm

1,2–2 kg / 2,6 kg

Herzsprung

22

c. 71x67

0,4–0,5 mm

1,4–1,5 kg

Group Plzeň

3

c. 51x48 / 68x61 cm

1–1,3 mm

2,4–3,4 kg

Table 1: Table showing the number, diameter, sheet thickness and all over weight of the different types of bronze shields

shields. Round shields are rare and a rather late form (Snodgrass 1999). MANUFACTURE The shields were hammered out from a tin-bronze blank. No blanks or moulds for such a purpose have been detected yet, but it is assumed that the blank, probably cast as a round form in a ceramic mould which in experiments,1 the cast discs would be as large as 19–20 cm. The metallographic structure of some shields suggests, due to the grain size in the cold-worked sheet, a starting diameter of the blank of about 15 cm (Goodway/ Chen 1996). Such a disc had to be flattened out into a sheet of, in most cases, 50–70 cm and an average weight of around 1.5 kg. Hammering to produce a thin bronze sheet involved a process which took many rounds of annealing in order to keep the metal from turning brittle. This was a rather time consuming process; in experiments a large disc could be expanded only 2 mm per hammering phase. To expand a blank disc of 20 cm to 60 cm would therefore take approximately 200 rounds of hammering and annealing. The technique of rolling or waltzing sheet metal can only be verified through the distinctive tools known only from the 15th century AD onwards (Born 1997). On some of the shields the hammer-marks are still visible, but most of them were polished before the decoration was embossed. The back of some shields especially bears very clear punch marks of the hammering process where the hammer was worked radially over the surface and then crossing these radial marks to expand the bronze. On the front of certain shields, namely the Type Herzsprung of the Nordic Bronze Age, very 1

The experiments were carried out in the workshop of the bronze smith N. Burridge, Cornwall, United

fine and thin long marks are still visible. They reach from the centre along the whole width and were made before the decoration with ribs and bosses was punched in. Comparable marks are also common on other large sheet metal objects, such as cauldrons. It is still unclear what kind of a technique and/or tool produced such traces. With the exception of the very small shields, all have a rolled over rim, which in some cases comprises a bronze wire for strengthening. Through this the rim becomes the strongest part of the shield being with the wire and rolled rim of sheet metal the thickest section of the shield and therefore the most suitable to withstand sword blows. When the body of the shield was finished, a grip of rolled sheet metal and thickened ends for the rivets or of either a massive bronze strap with flattened ends for the rivets were riveted onto the shield. In most cases loops or tabs were also attached to fasten a string to carry the shield over the shoulder. On a few shields, mainly the ones of Type Herzsprung, painted or incised lines and calliper points show how carefully and precise the ornamentation was planned (Thrane 1977; Uckelmann 2005). Shield features The shields have a round or slightly oval form, the diameter differs between 20 and 85 cm, but they measure mostly around 50 to 70 cm. The thickness of the sheet varies between 0.3 and 1.4 mm, and can vary quite a bit on one shield. But there is no evidence that the rim part represents a thinning out. Sometimes the area of the central boss is thinner. The weight of each shield lies mostly between 1 and 2 kg, but the heaviest one Kingdom. See: www.bronze-age-craft.com.

189

Marion Uckelmann

Figure 2: Bronze shield from Lommelev Mose, Falster, Denmark. Drawing from the back, and section on the side, Dm. 69 cm (Nationalmuseet København, M 9855; Uckelmann in prep., no. 3)

is 3.4 kg (tab. 1). Due to decorative and technical aspects, a number of types can be distinguished. Seven types and one group and two single shields can be differed. Type Lommelev-Nyírtura The Type Lommelev-Nyírtura includes only one complete shield from the eponymous find site of Lommelev Mose on Falster, Denmark (fig. 2) and fragments of shields from five hoards located in the Carpathian Basin. These finds have a central role in the dating of the shields, because they can be dated through their associated finds to the earlier Urnfield period (ältere Urnenfelderzeit, Bz D / Ha A1) or the 13th century BC, which makes them the oldest bronze shields so far in 190

Europe. The shield from Lommelev has a round shape and a diameter of 68.7 cm. This size can also be estimated for the fragments. The ornamentation consists of concentric ribs and bosses. The bosses are arranged in double rows which are not continuous. The gaps become smaller at the outer rim of the shields and form triangular groups creating a star-like decoration. The decoration on the fragments cannot always be reconstructed in the same way, the beams of the star seem to have the same number of bosses and are not getting wider. On the reverse side of the Lommelev shield a grip has been riveted on, which is made of a tube of rolled bronze sheet. There were probably tabs attached as well, perhaps to hold a strap or wire to carry the shield over the shoulder, but only the rivets have survived.

The function of Bronze Age shields

Type Nipperwiese This type was found in northern and central Germany, with two pieces in Britain. Eight shields are known with a diameter between 38 and 44 cm and a thickness of the sheet body between 1 and 1.3 mm. These shields were the sturdiest since the sheet of most of the others are hammered thinner. The weight of the Nipperwiese shields ranges between 1.5 and 2.2 kg and all shields have a grip and tabs riveted on. The ornamentation of all eight shields consists of only two concentric ribs making this group very homogenous in appearance. However, they do differ slightly in the attachment of grip and tabs, and no two shields are fitted exactly in the same way. Type Harlech Only four shields and two more from the close variant Trent form this type and all are found in England and Wales. The diameters are between 50 and 68.5 cm. The metal thickness varies greatly, between 0.1 and 1 mm. Two of the shields weigh around 1 kg and the one from the river Leah 2.75 kg, two more are large fragments of around 0.5 kg and the last one could not be weighed. As on the Type Nipperwiese shields, the ornamentation consists of concentric ribs only, in the case of Type Harlech with more ribs, between six and ten, and on the variant Trent with 21 and 63 ribs. The reverse side of the shields are fitted with a grip of rolled bronze sheet, fastened on with rivets as well as tabs. These are triangular in shape, sometimes perforated, and typical for shields from the British Isles. Type Coveney The Type Coveney is formed by only two shields, but the extraordinary design justifies a type on its own. The diameters are 47.5 and 52.5 cm with a metal thickness of 0.3–0.5 mm and a weight of c. 1 kg. Both have two extra rivets next to the tabs, in one case with large conical heads. The function of these rivets is unknown. The unique ornamentation is formed by meandering ribs which wound around the shield face, but also following a general concentric line around the central shield boss. The ends of these ribs end in snakeheads, with punched-in eyes. If the ornamentation on both shields are compared closely it becomes obvious

that they are very similar. Although the shield from Coveney is larger and has one rib more, the rib-width and the bows of the meander are almost exactly overlapping. Even though they were found far away from each other, one in Aberdeenshire and the other in Cambridgeshire, they must have been made either in the same place, or one was crafted using the other as a model. Type Athenry-Eynsham The shields of Type Athenry-Eynsham are the smallest sized shields and the six pieces were found only in the British Isles. The diameter varies between 23 and 35 cm, with a metal thickness of 0.3 and 1.2 mm and a weight of 0.9 and 1.2 kg. All show one concentric rib as decoration, and some show one or two rows of bosses. This type is less homogenous than other types. All shields have grip and tabs on the back. The grips can be made out of a massive bronze strip as well as of sheet tube, and the tabs are quite large, with big perforations. In contrast to the other shields, this type shows no rolled over rim (fig. 3). Type Yetholm This type is found only in the British Isles with the exception of one find from Denmark. With 26 known shields it is the most numerous type of shields. The diameter is quite large lying between 55 and 70 cm with a metal thickness of 0.4–0.7 mm. The Yetholm shields weigh around 1.2–2 kg, and in one case even 2.5 kg. The handle and tabs on these shields are all very similar and comparable to the ones on the other shields from the British Isles, while the tabs are generally of small size. The decoration is very regular and consists of alternating concentric rows of bosses and ribs. Most shields have 20 to 30 alternating ribs and small bosses, but a few shields have less rows, four to eleven, with larger bosses (fig. 4). Type Herzsprung The distribution of these shields is restricted to the Nordic Bronze Age, mainly southern Scandinavia. 16 of the 22 shields come from one find spot alone: a dried-up bay of Lake Vänern, Sweden near Fröslunda. The diameter of these

191

Marion Uckelmann

Figure 3: Bronze shield from Athenry, Ireland. Drawing from the back, and section on the side, Dm. 33.5 cm (British Museum, London, 1888.7-19,1; Uckelmann in prep., no. 24)

slightly oval shields is around 71 x 67 cm, with a metal thickness of 0.4–0.5 mm and a weight of 1.4–1.5 kg. The special ornamentation in the centre of these shields is the main feature for the type description: three wide and plain ribs running around the oval shield boss, the inner one has a gap and the two outer ones bear a U-shaped notch. The outer zone on the shield body is decorated differently with alternating ribs and rows of bosses. The handle on these shields is formed differently to other shields. Here a rolled tube of bronze sheet is riveted on the body with three rivets over a smaller rectangular piece of sheet, only the middle rivet connects the shield, handle and sheet, the two outer ones secure the handle from being ripped. The central boss is much shallower than on other shields, but the handle is heavily bent, and gives enough room for the hand of the bearer. The five shields and formers of organic material, are due to their decoration, close to Type Herzsprung (Uckelmann 2008; in prep.). Group Plzeñ This is not an actual type but due to similar decorative motifs the three shields form a group. One was found in the Czech Republic and the other two come probably from Denmark, but are unprovenanced. The shields are of oval form, and through decoration are related to the Type Herzsprung. The diameters are between

51 x 48 cm and 68 x 61 cm. The metal thickness is 1–1.3 mm which explains the relatively heavy weight of 2.4–3.4 kg for the shields. The integrating element in the decoration is the circular notch in the central shield boss, the rest of the shield body is rather plain, and adorned only with ribs or boss rows. The handle and tabs are all fitted in different ways. The shield from Plzeñ-Jíkalka shows at least in this aspect some resemblance to the Type Nipperwiese shields. CHRONOLOGY The dating of the shields is still problematic, since most of them are single finds or associated only with other shields. The fragments of shields found in the Carpathian hoards are well dated through their associations and belong to the 13th century (Bz D / Ha A1/2; Patay 1968). A hoard in Skydebjerg, Denmark dating to Period V (c. 925–800 BC) includes a fragment of a Type Herzsprung shield and gives a terminus ante quem for this type (Albrectsen 1957). The close resemblance with some of the shield images on the Iberian stelae and the early dating of the organic shields (see below) make it possible that the Type Herzsprung origins are as early as the late 13th century BC (Uckelmann 2008). The C14 result of the shield former from Kilmahamogue shows an Early to Middle Bronze Age date and that of the wooden shield from Cloonlara a Middle to Late Bronze Age 192

The function of Bronze Age shields

date.2 Three more organic shields were recently C14 dated and a 13/12th century BC date can be securely stated. In addition, a further metal shield is being AMS-dated through small pieces from remaining wood in the bronze sheet grip. Therefore, the long discussion about the date of the shield from Plzeñ-Jíkalka can finally be ended: it belongs to the 13th century BC.3 For the British types, the recent find of a Type Yetholm shield at South Cadbury suggests a date based on the find circumstances and the stratigraphical situation, as well as the metal alloy, of the Penard metalwork phase (c. 1300/1250–1125 BC) and at the latest the Wilburton metalwork phase (c. 1125–975BC) (Coles et al. 1999; Needham et al., in prep.). The Types Harlech and Coveney are seen to be contemporary, since they are associated with Yetholm shields. Type AthenryEnsham and Nipperwiese are extremely difficult, since they are all single finds, but the latter can be seen as contemporary with the shield from PlzeñJíkalka, whilst the former is similar to the larger British Types in form and technical features. The available dates for all shields suggest that the round shields were an invention that occurred in Britain and Ireland.

rather thin (0.3–1.4 mm), but when held in the hand they appear as quite sturdy objects, which are specially reinforced through the rolled rim and the fluting of the shield body through ribs and boss rows. The grip and tabs are well connected to the shield and fully functional. The alloy and hardness of the surface has been analysed in some cases. The examined shields show a percentage of copper between 85 and 90 % and tin between 9 and 13.4 %. This makes a rather hard bronze, comparable to a modern cast alloy. It is surprising that this alloy was used – and deliberately chosen – to make bronze sheet, since from a modern point of view it seems too hard and too brittle for hammering. The reason for such a high tin proportion could be the lower melting point, or the special colour. It is also possible that this alloy was chosen to make shields because of its heightened hardness and tensile strength. The measured hardness varies between shields as well as on a shield itself. This is due to the manufacture process as areas which have been worked on after the final annealing are far harder. From the few analyses it becomes obvious that even shields from the same types have varying hardness levels (Needham et al., in prep.). More testing has to be done to understand the differing results of these hardness tests on the shields surfaces.

FUNCTION OF THE SHIELDS Whether the metal shields of the Bronze Age were actually used and if so what for remain ongoing debates. What answers can be gleaned rely on the technical attributes and traces of use wear on the shields as well as on the find circumstances and experimental approaches.

Traces of use wear Traces of wear or abrasion are seldom found on the shields, which is mostly due to corrosion and patina. Repairs can be counted as traces of use; they appear on some shields, but are not common. Some shields have little sheet-straps riveted over small fractures and in one case a fracture is sewn together with a bronze wire (Klockhoff 1995). The smoothed over rim of a sharp-edged hole is probably also an ancient repair, the hole is likely to have been inflicted by a spearhead. Two more signs of damage can be found on the same shield, which was found in the river Thames near Long Wittenham.5 There is

Technical attributes The size of the shields is the same as that of used shields in contemporary and later times. Even the very small shields cannot be labelled as miniatures because the grip is a normal size. It is known from more recent times that they are quite useful in combat.4 The metal thickness of the shields is 2

Kilmahamogue (Uckelmann, in prep., no. 84), dating after Hedges et al. 1991, 128 f. OxA-2429: BP 3445 ± 70 (c. 1950–1600 cal BC). Cloonlara (Uckelmann, in prep., no. 83), dating after Hedges et al. 1993, 316, Oxa-3228: BP 3150 ± 90 (c. 1630–1190 cal BC). 3 The dating for these shields, the leather shield from Cloonbrin (National Museum Dublin, 1908:156), the wooden shield from Annadale (National Museum Dublin, 1863:1754), the wooden shield-former from Churchfield (National Museum Dublin, 1942:1844) and the wooden

fragments from the bronze shield from Plzeň-Jíkalka (Západočeské muzeum Plzeň, 8432) was undertaken in the laboratory of Groningen University by J. Lanting and will be fully published in Uckelmann, in prep., no. 85. 81. 82. 86. 4 A 13th century AD fencing book describes the use of a buckler. The so-called Walpurgis-Codex or Tower Fechtbuch, Royal Armouries, MS I. 33; Forgeng 2003. 5 Ashmolean Museum, Oxford, 1980:212; Uckelmann, in prep. no. 9.

193

Marion Uckelmann

A

E

B C D Figure 4: Bronze shield from Yetholm (National Museums of Scotland, Edinburgh, X. DN1; Uckelmann in prep., no. 49). – A: Photograph of the front, Dm. 60.2 cm. – B: Detail of damage on the front of the shield boss. – C: Detail of damage on the back of the shield boss. – D: Detail of shield tab, length: 2.4 cm (scale c. 1:1). – E: Detail of handle, side view, length: 14.8 cm

one larger hole of trapezoidal shape which could also have been derived from a spearhead, and smaller round one perforation. Another shield also from the Thames near London6 bears a cut in the upper half, maybe from a sword tip, and a trapezoid hole, probably from a spearhead. One of the three shields from Yetholm, Scotland, also shows as well some damage: the central boss is pierced by a sharp pointed object, most likely a 6

sword (fig. 4 A–C). Other shields show weapons marks, but they are not common and not bound to one type of shields. DEFENSE Some shields show weapon marks and the technical characteristics that suggest that some of

British Museum, London, 1856,7-1,1350; Uckelmann, in prep. no. 42.

194

The function of Bronze Age shields

them could have been used in combat. J. M. Coles undertook an important study on shields in the 1960s and conducted the first experiments on the usage of shields in combat. Unfortunately, he took a rather thin shield as a model and used, due to availability, only 0.3 mm thin copper, rather than copper-alloy, sheet to reproduce a shield. In an adventurous testing with one individual holding a shield against repeated sword blows, the shield could not withstand any sword blows and was cut through to the rolled rim (Coles 1962). Newer experimental approaches carried out by B. Molloy showed that some metal shields, especially the thicker ones, are quite capable of protecting an individual in combat. He used copper sheet as well, but of 0.9 mm thickness, to reproduce three shields of different sizes, two of them with a rolled over rim. These shields were tested with swords and spears thrown at them: they survived the test without being damaged to the extent that they were unusable, with only minor dents visible (Molloy 2009). Coles produced very valuable research on shields, but due to his experiments came to the conclusion that metal shields could have been only used for display and were non-functional in combat. This opinion was adopted by most later researchers. Yet, the closer examination of the shields revealed more evidence for weapon damage, and newer experiments with copper-sheet shields show that the interpretation of the function of metal shields has to be changed, although clearly not for all of them. To prove this, further testing, with a replica that is much closer to the original, ideally hammered out sheet from a blank of tin-bronze, has to be carried out. A project of such nature is currently being undertaken, including metal analyses of the shields kept in the British Museum and practical work carried out together with the bronze smith N. Burridge. The project is funded by the British Museum. The surviving shields of organic material certainly could have been used in battle. Coles also reproduced and tested a hardened and waxed leather shield which was found quite suitable fending off sword blows (Coles 1962). In the experiments by Molloy a leather shield replica was also tested and was quite effective in warding off sword blows and spear throws, but dented and buckled after about 40 blows. However, the shield withstood all testing of cuts and thrusts of bronze swords. The marks of the attacking weapons were comparable to those on the original shield from Cloonbrin 7

Society of Antiquaries London, LDSAL 80; Uckelmann, in prep. no. 34.

(Molloy 2009). Traces or damages had been noted before, but could not be related to weapon marks until testing and comparative analysis. It has often been suggested, that the thin sheet metal shields were strengthened with a layer of leather on the back, but no archaeological evidence on the actual shields can be found to support this. Since many shields come from bogs, remnants of organic material for the backing would have been recorded in some cases. DISPLAY For Bronze Age people these shields were precious objects, this becomes obvious through their limited numbers, the high amount of material used, and the time intensive and high skilled production as well as their final use as votive offerings (see below). Some of the shields bear an extensive ornamentation, for example, the shield from Beith with over 9000 singly punched bosses.7 The shields could have been used to mark a social position or as a badge of rank. Since they are never found in graves, they seem not to represent a personal item, but rather a communal property, where the shield was at certain times assigned to a person according to their position (that could be a chieftain or a leader in battle). In their own time the shields glistening golden glare in the hands of their bearer must have bestowed an impressive vision in the eye of the beholder and maybe an enemy in battle or a participant in a ceremonial rite. On the shield from Sørup on Falster,8 Denmark, four rings are riveted on to the rim. The rings show abrasion and the shield was probably hung up. Where it was placed is of course speculative, but it is easy to envision the shield hanging in a community hall or in the house of the leading man, or as well in a ritual building, the latter one is found with shields in temples in antiquity. DEDICATION There is some evidence for a ritual function of the shields. The few shields with figurative décor are bearers of ideological and/or religious content: one shield of Type Herzsprung shows a row of water-birds formed from punched in bosses and points; another shield of unique character bears a rare design of the common bird-sun-boat 8

Nationalmuseet København, B. 10988b; Uckelmann, in prep. no. 90.

195

Marion Uckelmann

Figure 5: Distribution of Bronze Age sheet metal armour 196

The function of Bronze Age shields

motif; and the two shields of Type Coveney are decorated with winding snakes. The symbol with U-notches on the Type Herzsprung shields might also have a religious meaning (Uckelmann 2008; in print). It is the find context of the shields that especially suggests a ritual interpretation, at least at the time of their deposition after their time of use. 75 of the 90 shields come from a wet context, rivers, bogs and lakes. The few finds in a landscape context come mainly from hoards in the Carpathian Basin. Almost half of the shields come from multiple depositions, but with only shields as further associations: on seven sites two shields were discovered together; and on one site each three, five to six and 16 shields. These multiple finds of shields are potentially comparable to the well known ritual of dances with arms in antiquity (Hagberg 1989, 41 f.; Uckelmann in prep.). Many of the shields received special treatment prior to the deposition. Some are recorded to have been placed standing on edge, the five or six from Beith even in a circle. A few shields show damage, and less frequently white sand as bedding or burnt remains were observed (Uckelmann, in prep.). Detailed studies on regions with water finds, such as the river valleys of the Thames in England and the Shannon in Ireland, as well as the Fenlands in East Anglia and the southern German area have revealed that the shields fit in the general deposition patterns, with the exception that they are rather rare objects.9 In these studies most of the finds from watery contexts are interpreted as votive gifts, due to special choice and treatment of the objects, as well as the fact that most them could not be retrieved again. All these aspects suggest that the shields were votive gifts. The motivations behind these offerings remain unclear, but one can imagine them, for example, as offerings for a victorious battle. Alternatively, rivers can be interpreted as borders as well as transport routes, and the offerings could have wished for a safe crossing or to strengthen the borders. For the Fen sites maybe the offerings were placed to wish for the waters to stop rising (Evans 2002). OTHER DEFENSIVE ARMOUR Compared with the distribution of the other metal defensive weapons such as helmets, cuirasses and shin guards, a clear distinction becomes obvious (fig. 5). Only in the Carpathian area are all types found, but these have survived mainly in frag9

York 2002; Bourke 2001; Evans 2002; Falkenstein 2005.

197

ments. A complete set of defensive armour – although in fragments – is known from only one site: Nadap, Hungary in a huge hoard of metal objects (Petres 1982). In the British Isles shields are the only form of metal armour. Shields and cuirasses are never found in the same area; helmets and shields are known in small numbers in Germany and Denmark and appear together on the Iberian stelae and in Nordic Rock Art, but on the later not on the body. In France and the south Alps region helmets and shin-guards are common and in the west Alps region cuirasses appear as well, but there are no metal shields in the rest of Western Europe. The numbers of the other objects of armour show that they are as rare as the shields: ca. 90 helmets; 51 shin guards and 30 cuirasses (cf. Uckelmann in prep.). The meaning behind this distribution is not yet clear. The images of the shields show that their use was probably widespread over Europe, but in organic material. Helmets, shin-guards and cuirasses are, like the shields, manufactured from thin bronze sheet and had to be skilfully hammered out and formed into shape. They also have riveted on parts as well. The decorative elements bear the same punched in bosses, points and ribs, and very few have bird motifs. Some of the pieces show traces of use wear and repairs, indicating evidence for longer use-time. Unlike the shields, helmets, cuirasses and shin-guards were worn over some kind of bedding, most likely leather. In some cases the edges of helmets and cuirasses show punched in holes and the shin-guards have wire loops, where the bedding could be sewn on. The other armour also shows a similar deposition pattern as the shields: the majority of finds in Western Europe were deposited in wet contexts whereas in the Carpathian Basin the finds are mainly in hoards. Only the shin-guards seem to have a different meaning, since they appear very often in graves, a context where almost none of the other pieces of defensive armour are found. CONCLUSION Organic shields made of leather or wood were almost certainly the main material for shields throughout much of Europe in the Bronze Age, even though they rarely survive in the archaeological records. Shields were used in combat and their development and spread should probably be seen in relation to the use of swords. Different sized shields probably indicate different

Marion Uckelmann

combat styles (Molloy 2009). The form of the round shield appears, according to the recent dating of organic examples first, in the British Isles, where most of the metal ones were produced in the following centuries. This manufacture of shields of bronze sheet, as well as the other sheet armour, that occurs at the end of the Middle Bronze Age and the beginning of the Late Bronze Age in Central, Western and Northern Europe shows a pan-European desire to elevate the meaning of these objects and maybe their bearers. This can be related to broader changes in Bronze Age societies, and especially the role of the warrior. In different regions, different parts of armour were preferred, but treated in the same way when deposited. Together with the images of warriors and weapons this provides evidence that the metal defensive armour was possibly used to mark exceptional warriors or leaders. The bronze shields were valuable and elaborately worked objects and had their own meaning in the martial environment of the Bronze Age people. They were clearly used but not simply for a single purpose. During the ‘lifetime’ of the shield it went through different stages in its meaning and function. At the beginning as the precious product of a well skilled and trained craftsman. In its time of active use, some of the metal shields very probably protected their bearer in combat and were also used as markers of a social position and/or as a device in ritual ceremonies. After their time of active use, their function changed and they were transferred – most likely by the community – to another sphere. BIBLIOGRAPHY Albrectsen, E. 1957. Skydebjergfundet. Fynske Minder 1957, 75–78. Born, H. 1997. Zur Herstellungstechniken der Bronzen des Gefäßdepots aus dem Saalegebiet. Mit einem Beitrag von T. Schmidt-Lehmann. Acta Praehistorica et Archaeologica 29, 1997, 69–96. Bourke, L. 2001. Crossing the Rubicon. Bronze Age Metalwork from Irish Rivers. Bronze Age Studies 5. Galway, National University of Ireland. Coles, J. M. 1962. European Bronze Age Shields. Proceedings of the Prehistoric Society 28, 1962, 156–190. Coles, J. M., Leach, P., Minnitt, S. C., Tabor, R. and Wilson, A. S. 1999. A Late Bronze 198

Age shield from South Cadbury, Somerset, England. Antiquity 73, 1999, 33–48. Coles, J. M. 2005. Shadows of a northern past. Rock carvings of Boshuslän and Østfold. Oxford, Oxbow. Evans, C. 2002. Metalwork and ‘Cold Claylands’: Pre-Iron Age Occupation on the Isle of Ely. In T. Lane and J. Coles (eds), Through Wet and Dry. Essays in Honour of David Hall, 33–53. Heckington, Heritage Trust of Lincolnshire. Falkenstein, F. 2005. Zu den Gewässerfunden der älteren Urnenfelderzeit in Süddeutschland. In B. Horejs, R. Jung, E. Kaiser and B. Teržan (eds), Interpretationsraum Bronzezeit. B. Hänsel von seinen Schülern gewidmet, 491– 504. Bonn, Habelt. Forgeng, J. L. 2003. The Medieval Art of Swordsmanship. A Facsimile and Translation of Europe’s Oldest Personal Combat Treatise, Royal Armouries MS I.33. Leeds, Royal Armouries. Goodway, M. and Chen, Y. 1996. The Fröslunda shields: Cymbals or Symbols? Laborativ Arkeologi 9, 1996, 49–52. Hagberg, U. E. 1989. Die Bronzeschilde von Fröslunda – ein neu entdeckter Hortfund von Herzsprungschilden aus Schweden. Vorläufige Mitteilung. Jahresschrift für Mitteldeutsche Vorgeschichte 72, 1989, 39– 42. Harrison, R. J. 2004. Symbols and warriors. Bristol, Western Academic and Specialist Press. Hedges R. E. M., Housley, R. A., Bronk Ramsey C. and van Klinken G. J. 1991. Radiocarbon dates from the Oxford AMS system: datelist 12. Archaeometry 33, part 1 (Febr. 1991) 128–129. Hedges R. E. M., Housley, R. A., Bronk Ramsey C. and van Klinken G. J. 1993. Radiocarbon dates from the Oxford AMS system: datelist 17. Archaeometry 35, part 2 (Aug. 1993) 316. Klockhoff, M. 1995, Sköldarna ur konservatorns synvinkel. In P. Jankvas (ed.), Långt Borta och Nära. Skrifetr från Skaraborgs Länsmuseum 21, 33–57. Nossebro, Rydin. Molloy, B. P. C. 2009. For Gods or men? A reappraisal of the function of Bronze Age shields. Antiquity 83, 2009, 1052–1064. Needham, S., Northover, P. and Tabor, R. in prep. South Cadbury: the last of the bronze shields. Patay, P. 1968. Urnenfelderzeitliche Bronzeschilde im Karpatenbecken. Germania 46, 1968, 241–248.

The function of Bronze Age shields

Petres, É. F. 1982. Neue Angaben über die Verbreitung der spätbronzezeitlichen Schutzwaffen. Savaria 16, 1982, 57–80. Snodgrass, M. A.1999. Arms and armour of the Greek. Baltimore, John Hopkins University Press. Stary, P. F. 1980. Das spätbronzezeitliche Häuptlingsgrab von Hagenau, Kr. Regensburg. In K. Spindler (ed.), Vorzeit zwischen Main und Donau, 46–97. Erlangen, UniversitätsBund Erlangen-Nürnberg. Uckelmann, M. 2005. Die Schilde von Herzsprung – Bemerkungen zu Herstellung, Funktion und Deutung. Jahresschrift für Mitteldeutsche Vorgeschichte 89, 2005, 159–188. Uckelmann, M. 2008. Irland oder Iberien – Überlegungen zum Ursprung einer Ornamentform der Bronzezeit. In F. Verse, B. Knoche, J. Graefe, M. Hohlbein, K. Schierhold, S. Siemann, M. Uckelmann and G. Woltermann (eds), Durch die Zeiten ... Festschrift für Albrecht Jockenhövel zum

65. Geburtstag, 259–268. Rahden, Marie Leidorf. Uckelmann, M. 2010. Zur Ornamentik jungbronzezeitlicher Schilde. In F. Bertemes and H. Meller (eds), Der Griff nach den Sternen. Wie Europas Eliten zu Macht und Reichtum kamen. Tagung Halle 2005, Bd. 5, 553–562. Halle (Saale). Uckelmann, M., in prep. Die Schilde der Bronzezeit in Nord-, West- und Zentraleuropa. Prähistorische Bronzefunde III, 4. Stuttgart, Steiner. [PhD thesis, University of Münster 2008] Thimme, J. 1980. Kunst und Kultur Sardiniens. Ausstellungskatalog. Karlsruhe, Badisches Landesmuseum. Thrane, H. 1977. Bagsiden – et kig bag den fine facade af broncealderns metalhandvaerk. Fynske Minder 1977, 37–46. York, J. 2002. The life cycle of Bronze Age metalwork from the Thames. Oxford Journal of Archaeology 21 (1) 2002, 77–92. Marion Uckelmann Department of Archaeology University of Exeter Laver Building, North Park Road Exeter EX4 4QE United Kingdom

199

200

METALLURGICAL AND ARCHAEOLOGICAL INVESTIGATIONS ON THE MACEDONIAN SHIELDS OF STARO BONČE (3RD CENTURY BC) Paolo Piccardo and Dusko Temelkoski ABSTRACT Recent discoveries (in 2006) in the area of Staro Bonče, close to the city of Prilep (Republic of Macedonia) revealed the presence of three shields belonging to the army of King Demetrius I Poliorcetes. The shields, fragmented in more than 150 pieces, were submitted to a careful archaeometric investigation on samples obtained before restoration, that allowed to rebuild the shields nearly completely. This paper reports the typological and metallurgical characteristics leading to find out information concerning the king (and therefore the period of usage), the nature of the base materials (tin bronze) and the most probable support used under the thin metallic sheet. The three shields were distinct by their decoration (the shape of the little stars in the frieze) and size (diameters of decorations and of the whole shield), and by the alloy composition (i.e. the tin and lead relative amounts were differing in content and homogeneity of distribution). The unexpected presence of elements such as chromium or zinc in the external edge of the alteration patina found on the inner face of the shields allowed to formulate sound hypotheses about the use of leather as support (or as main layer in interface with the metal) under the decorated bronze sheet. KEYWORDS Bronze shields – Macedonia – microstructure analyses – SEM-EDXS-BSE-SE – Raman – leather INTRODUCTION In the second half of 2006, after the casual finding of two fragments of thin bronze sheets in the village Podmol, further excavations started in the area of the Hellenistic–Roman cemetery of Staro Bonče, located in the north-eastern part of the Pelagonia valley about 20 km southeast of Prilep (fig. 1). The shape and decoration found on those fragments show without doubts that they were parts of exclusive items: antique Macedonian shields (Grozdanova 2007; Piccardo et al. 2008). In the surroundings of the village of

Bonče, traces of the ancient location called Staro Bonče had been already found. Numismatic finds conserved at the museum’s collection suggest a two-phased settlement: the first one from the middle of the 4th century BC up to the first quarter of the 3rd century BC and the second one on the pass of the 3rd century AD to the 4th century AD. Among other things, the excavations carried out in 2006 lead to the discovery of a pit of irregular circular shape with, among other finds, numerous fragments from such as ceramic vessels, one iron knife, several weights of kidney or pyramid shape and a total amount of over 150 fragments belonging to Macedonian shields (Grozdanova 2007). THE SHIELDS Only a few shields of this type have been found until now (Grozdanova 2007). The shields were already mentioned by classical authors such as Plutarch (Scott-Kilvert and Griffith 1973) and others (Anglim et al. 2002; Anderson 1976; Markle 1977;1999) from where it results that the diameters of these plate convex circular shields were ranging from around 60–77 cm. Usually, the Phalanx soldiers were carrying a small shield (i.e. pelte), while the elite members of the infantry of the King’s personal guard were wearing larger shields, the aspis. The reconstructed size of the fragments from Staro Bonče is pointing to shields used by members of the King’s personal guard. The discovered shields are richly decorated (fig. 2): a band of three circular ribs, parallel to the outer edge, is found on the brim of the shields. They bear towards the centre a frieze, composed by seven fields, each of them consisting of three sickle-shaped relief lines, surrounding a star with eight rays. In the centre of the shield episema a star (i.e. Macedonian sun) is presented with twelve rays. The central star is surrounded by a circular band in which the name of the Macedonian king Demetrius is inscribed in genitive: ΒΑΣΙΛΕΩΣ ΔΗΜΗΤΡΙΟΥ. The shield’s external edge is bent inside, probably to fasten the metallic sheet on a support made of organic materials such as wood or leather or both of them (Markle 1977; van Wees 1994a; Hanson 1993). With the telamon, a special handling system made of leather (Markle

201

Paolo Piccardo and Dusko Temelkoski

Figure 1: Geographical position of Staro Bonče in the modern Republic of Macedonia and in the ancient Macedonian kingdom

1999), the warrior was able to have both hands free to hold more than one weapon. It was obvious that the 150 fragments freshly excavated were belonging to more than one shield. This was confirmed on one hand by typological observations by the shape of the fragments, the decorations (as it will be discussed hereafter) and the inscriptions which gave the most important contribution, and on the other hand by elemental analyses and technical investigations. All of these observations were leading to the first assumption, that at least three shields were found in the excavation in Staro Bonče in 2006. The central star and the surrounding seven little stars have two different types of rays: one with needle-shaped form and one with dropshaped form (i.e. with fattened and well-rounded root) as shown in figure 3 (A. B). Carrying out systematic observations on all fragments containing complete and partly preserved smaller stars, it became clear that we were dealing with two shields with needle-rays stars, and at least one more shield with smaller stars showing rays of the second type described above. The typological

distinctions resulted coherent with the patina characteristics on the surface of the fragments. However, one of the most important data supporting the existence of at least three shields lies in the identification of the preserved letters of the inscription ΒΑΣΙΛΕΩΣ ΔΗΜΗΤΡΙΟΥ which surrounds the central star. Three letters are appearing one time only in the inscription: Λ, M and Τ; they are preserved on three different fragments. The letters Σ and Η should appear two times per each inscription; five of them have been found. By this simple information (table 1) it is possible to suggest that the number of shields which were discovered on the locality of Staro Bonče should be three. INSCRIPTIONS Following this, the shields are marked as shield A, shield B and shield C. The typological investigation carried out on all fragments showed that larger parts are missing, especially of shield C. Only one well preserved fragment of the edge

202

Metallurgical and archaeological investigations on the Macedonian shields of Staro Bonče (3rd century BC) Letter

B

A

Σ

I

Λ

E

Ω

Δ

H

M

T

P

O

Y

Freq.

2

2

5

3

3

2

1

1

5

3

3

2

1

2

1 Shield

External diameter

Diameter of inscription

Thickness of inscription

Rays‘ length of the little stars

A

74

27.70

3.20

2.0

B

≈72

25.80

2.50

2.5 – 3.0

C

≈66

20.50

2.50

2.0 – 2.5

2 Table 1: Counting the preserved letters of the central inscription Table 2: Dimensional characteristics of the shields. All measures are reported in cm

of shield A allowed to measure the original diameter of 74 cm. The rim of the shield is bent inwards (not outside, as it was common with Bronze Age shields), probably to better fix a possible organic support as leather, linen or wood. Inside the bent edge, no residues of a metallic wire (i.e. the ‘soul’ made to strengthen the edge of the shield) were found, but a black, fatty powder, mixed with soil. From the same shield, a number of fragments were recovered such as the central sun, which has rays in form of a lengthy drop, and is almost fully preserved, and the inscription: ΒΑΣΙΛ_ _ Σ ΔΗMΗΤΡΙΟΥ. The field band containing the inscription is 3.4 cm wide and the diameter of the middle part, including the band, is 27.7 cm wide. The rays of the smaller stars are as well needleshaped and measure 2 cm in length. We recorded one fragment of the mid star of shield B, showing rays in the shape of lengthy drops, and the following letters of inscription: BΑΣΙΛEΩΣ _ _ _ΗΤΡ_ _ Υ. The field-band with the inscription is 3.2 cm wide and the diameter of the middle part, including that band, is 25.8 cm. The rays of the smaller stars are radial disposed around the central star. They are similar to shield A, but shorter and measure 2.5–3 cm of length. It is possible to suggest that the most probable diameter of shield B is at least 72 cm. The smallest number of fragments survived for the reconstruction of shield C. Only the following letters of the inscription are preserved: ΙΛ _ Τ. The field–band were the inscription is placed is 3.2 cm wide and the total diameter of the central part is 20.5 cm. The rays of the small stars have the shape of a lengthy drop and measure from 2–2.5 cm in length. The rays of the small stars have the shape of a lengthy drop and measure from 2–2.5 cm of length. The smaller convex central area suggests that the total diameter of the shield might be smaller than those of the previous

ones. Therefore and taking into account that the three ribs of the edge measure almost the same width of the other shields, it is possible to assume that Shield C had a total diameter of about 66 cm. Thanks to the preserved inscription, it is possible to date the shields. The genitive is referring to the King as the owner of the shields, that they were ‘belonging to’ king Demetrius. The size of the shields points to the infantrymen as their user, in detail the infantry being part of the Macedonian corps d‘élite of the army and probably to the personal guard of the king himself. In the known line of Macedonian kings two have the same name. Demetrius I Poliorcetes, who ruled from 294 till 285 BC, and Demetrius II, who ruled between 239 and 229 BC. The epigraphical analysis confirmed that other shields belonging to the same type (Karametrou-Menteside 1987; Adam-Veleni 1989; Dakaris 1993), that the Staro Bonče shields can be assigned to Demetrius I Poliorcetes. This is proven on one hand by the associated pottery, dating to the end of the 4th and the beginning of the 3rd century BC, and on the other hand due to coins found in the same area, showing Demetrius I. ANALYTICAL APPROACH BY ARCHAEOMETRY In order to complete the study of these shields and to gain information about their manufacture, a few fragments were sampled. The results discussed in the following are meant to highlight: - the nature of the alloys used to produce the shields, - the manufacturing technique (of basic importance to better understand technological skills and solutions), - the nature of the organic support, and

203

Paolo Piccardo and Dusko Temelkoski

2

3A

3B

Figure 2: A schematic drawing of a Macedonian shield, inspired by the shields found in Dion and in Staro Bonče Figure 3: Details of the central star of shields. – A: Shield C with lengthy drop rays. – B: Shield A with needle like rays. Shield B has the same typology of stars as shield A. 204

Metallurgical and archaeological investigations on the Macedonian shields of Staro Bonče (3rd century BC) Shield

Cu

Sn

Pb

S

Metallurgical features

A

92–93

5–6

1–2

0.1–0.3

Large sized recrystallized grains, residual primary heterogeneity of the solid solution

B

91

9

0.1–0.2

Recrystallized, homogeneous solid solution

C

90–91

7–8

0.1–0.3

Recrystallized, slightly heterogeneous solid solution

0.5–1

Table 3: SEM – EDXS analyses of the alloys (wt. % average or range) metallographic observations

- the degradation suffered during the long disposition in the soil. Taking into account that copper is the base component of the alloy, the research approach is following a modus operandi of investigation specifically adapted to archaeological finds and inspired by the characterization of metallic materials (Pinasco et al. 2007). One sample (2–4 mm2) per shield was mounted in epoxy resin and polished in order to show cross sections corresponding to the radial and to the tangential direction from the centre to the edge of the shield. All analyses were performed on samples without using metallographic etchants in order to avoid any possible pollution of the inclusions and, mainly, of the corrosion products. The common use of etchants such as ethanol solution of iron chloride and hydrochloric acid or Klemm’s reagents is helpful to highlight microstructural feature but has the negative effect to modify the composition of corrosion products and non metallic parts. In order to rescue the original composition of these constituents of the finds a new session of grinding and polishing is needed that is often not possible on archaeological samples due to their limited size. Nevertheless, the corrosion process is often allowing to obtain information from micro-structural features of the original alloy (even in nearly complete mineralization of the metal, Pinasco et al. 2007) due to main corrosion mechanism starting from grain boundaries then entering mechanical and thermal twins and, finally, the core of the crystals (Piccardo et al. 2007). The equipment used has been selected depending on the information requested: optical microscopy with Bright Field (BF) and Dark Field (DF) contrast for basic microstructural information and for the characterization of oxidised compounds in the alteration patina; Scanning Electron Microscopy equipped with Secondary and Backscattered Electrons detectors (SEMSE, SEM-BSE) for topographic and compositional contrast in order to analyse the metallurgical features of the alloy. Furthermore, an Energy Dispersive X-Ray Spectroscopy (EDXS) for quantitative elemental analyses; Raman micro-

Spectroscopy (RS) was used to study the corrosion in a more detailed way. RESULTS AND DISCUSSIONS In table 3 a synthesis of the results achieved are collected. The alloys are ranging between 5–9 wt.% of tin and 0–2 wt.% of lead. Each shield has a different alloy composition. A few more fragments selected for further analyses confirmed the differences and made it possible to assign the fragments to one of the three shields. The main inclusions are copper sulphides, which is quite characteristic for prehistoric bronzes. The larger range of composition found in shields A and C corresponds to a residual heterogeneity of the α-bronze solid solution and the distribution of tin in bands, following the direction of deformation, which is coherent with a mechanical work (hammering) performed at room temperature and coupled with recrystallisation annealing at a temperature in the range of 500–600°C and for a time not allowing a complete homogenization of the alloy. Figures 4 to 6 are showing a representative SEM-BSE image per each sample where the metallurgical features are visible. The intergranular corrosion (fig. 7) is similar in all three shields and is clearly showing a microstructure typical for slightly deformed recrystallised bronze. The corrosion is following the grain boundaries of regular polygonal grains and the straight lines of mechanical twins edges inside the grains. The latter are visible due to a light cold deformation after annealing that was not affecting the grains geometry. As a matter of fact, this kind of deformation could result from a last finishing treatment as well as from the usage of the shields. It is nevertheless logic to imagine that the usual microstructure of the shields should be able to absorb shocks without breaking or damaging the soldier (e.g. by too intense vibration), as it is the case for recrystallized bronzes. The estimated thickness of the metal sheets of all three shields is the same and ranges between 0.3–0.4 mm. This is coherent with the use 205

Paolo Piccardo and Dusko Temelkoski

4

5

6 7 Figure 4: SEM–BSE micrograph on the cross-section sample corresponding to shield A. The alloy is rich in lead (white inclusions) and showing linear bands of differing gray levels in accordance with the local amount of tin (the lighter the gray level the higher the amount of tin) Figure 5: SEM–BSE micrograph of the cross section sample corresponding to shield B. The alloy is completely homogeneous and no lead inclusions are visible Figure 6: SEM–BSE micrograph of the cross section sample corresponding to shield C. The solid solution is showing light but still present residual heterogeneity in bands and small inclusions of lead Figure 7: SEM–BSE micrograph of the cross section of shield C and representative of the microstructural aspects of the three objects. Examples of intergranular corrosion are clearly showing the grain boundaries and a few mechanical twins

of this kind of shields applied on a thicker support made of organic materials. The composition of the alloy affects more the production process than the efficacy of the shield as a defensive weapon. However, differences in the chemical composition of the alloy (relative amount of lead and tin mainly) affect some characteristics such as castability (i.e. the ability to fill thin moulds), workability (i.e. the amount of stress an alloy can absorb during a deformation process without breaking), and annealing parameters (i.e. temperature and time of thermal treatment between two cold working steps). It is therefore clear that differing characteristics will result in more or less important differences in the manufacturing work flow.

Previous studies in castability (Piccardo et al. 2009; Piccardo et al. 2010) have shown that lead and tin have a synergic effect in increasing this parameter. Therefore, with the alloy of shield A, it should be possible to cast thinner ingots than with the alloy of shield B. The behaviour of the alloy used to produce shield C is intermediate. At the same time a tin-amount ranging between 8–11 wt.% is increasing the elongation (i.e. plastic deformation) under stress, thus improving the workability of the alloy (Ammanati et al. 2004). Concerning the annealing process, one can consider that temperature and time of the thermal treatment could be quite similar for all shield bronzes due to the balance between a lack of precise temperature

206

Metallurgical and archaeological investigations on the Macedonian shields of Staro Bonče (3rd century BC)

Figure 8: OM-DF example of the corrosion patina with compounds recognisable by their colours; under the blue carbonates (midline upwards) and phosphates a dark brownish layer of soil mixed with small particles, probably rich of residuals organic compounds, is visible (below midline; small band with light grey dots)

control of the furnace and an improved empirical knowledge of this process. In fact, a temperature ranging between 550–600°C is suitable for nearly all α-bronzes and the time is in the magnitude order of minutes. These annealing parameters are perfectly working to recrystallise an alloy and to allow further cold deformation, but are not totally efficient for homogenising the solid solution. This can be obtained only if the cumulative time of the thermal treatment (of recrystallisation) is lasting long enough to allow the complete diffusion of tin into the matrix. MANUFACTURING TECHNIQUES To outline all information gained so far, we can suggest that the manufacturing technique was roughly the same for all alloys taken: - casting a thin ingot probably in the shape of a disk (thickness between 3 and 5 mm, esti-

mated diameter 40 to 50 cm, taking into account that manufacturing implies a certain amount of loss on the edges), - the as-cast ingot was water quenched immediately after solidifying, in order to obtain an alloy ready for the following steps, - cold hammering alternating with recrystallisation annealing, followed by water quenching (Pinasco et al. 2007; Ammanati et al. 2004) in order to avoid a (α+δ)-eutectoid precipitation with consequent embrittlement of the alloy until final thickness and size are achieved, - grinding and polishing of the surface, - decoration with friezes by cold deformation and finishing, and - fastening with the supporting materials (e.g. leather and/or wood on which handle parts where also mounted). Taking into account the two more differing alloys (the ones of shield A and B) it is possible

207

Paolo Piccardo and Dusko Temelkoski

9

10

Figure 9: Quantitative X-Ray map of the distribution of major elements in the corrosion patina. The central image is corresponding to the SEM-BSE vision of a portion of the blue compounds of figure 8 Figure 10: Quantitative X-Ray map of the distribution of minor elements in the corrosion patina (see fig. 9). Chromium, iron and nickel are concentrating on the edge at the interface with the soil rich layer (see fig. 8) 208

Metallurgical and archaeological investigations on the Macedonian shields of Staro Bonče (3rd century BC) to discuss main differences in the manufacturing process. The alloy of shield A is characterised by a higher castability, therefore the thickness of the ingot should be thinner and the diameter larger, thus the craftsman needed a smaller number of thermomechanical steps to achieve the final shape than were necessary for the other two alloys/shields. The alloy of shield B was cast in a slightly thicker and smaller disk ingot due to its reduced castability characteristics and then deeply wrought to the final shape. From the metallurgical point of view, despite the similarities of the final thickness, the alloy of shield B is logically characterised by a more homogeneous solid solution compared to the one of shield A. Image analyses performed on sulphides confirmed the smaller amount of deformation applied to this shield. This discussion opens several hypotheses, of which the first two are more speculative: 1. The alloy composition corresponds to the typical production of a certain period and therefore the three shields were manufactured in different moments. At the same time, the alloy of shield A can be considered as the least expensive, due to the lower amount of tin, which is an expensive element due to is rarity among the three main elements of the alloy, and the higher amount of cheaper lead. It might mean that the evolution in the techniques led to an improvement in casting properties of alloys by changing lead content without major modifications on the mechanical resistance of the shield. 2. The alloy composition corresponds to a specific workshop and is a sort of ‘signature of the producer’ and no hypotheses can be made about the time of production of the shields; they can even be made at the same time. The very small differences in colour (Ammanati et al. 2004) could correspond to the ‘taste’ of the customer and the differences in workability and resistance in battle as well as the expenses for the alloy could be reflected in the shield price. The mechanical resistance and efficacy of the shield is influenced as well by the support and thus is difficult to argue about without having further information of the organic support. 3. The alloy composition is depending only on raw materials and/or recycling and the craftsman was adapting the manufacturing process to the properties of the alloy. Following this, one can consider casting in an open (eventually edgeless) mould, taking care to cover immediately the molten alloy with charcoal powder in order to limit oxidation, that will be anyway removed by water quenching of the as-cast ingot once solidi-

fied. The thickness and diameter of the ingot are depending on the castability and the higher this parameter, the thinner and larger the ingot. The craftsman just had to apply a smaller amount of deformation unaware of the composition. What we see from our studies (e.g. homogeneity of solid solution) is, in this case, limited to our researches and totally unknown by the craftsman who was anyway obtaining his final product with characteristics suitable for a defensive armour as a shield. Although supporting one hypothesis more than another might look too drastic; the authors consider the latter the more consistent one. Recycling was a common practice and usually more or less similar alloys were cast together, making the selection only on empirical tests like response to hammering and comparison of colours. The kingdom of Demetrius I Poliorcetes was lasting less than a decade, thus, if the attribution to this king is correct, it is difficult to imagine any change in techniques in such a short time space. Moreover, the attitude of Demetrius I Poliorcetes to expand his kingdom by attacking neighbouring countries as well as the need to defend his kingdom against Pyrrhus of Epirus in 287 BC, is in accordance with the need of a large production of arms and armour (and thus shields as well) within a short period. THE NATURE OF SUPPORT AND THE DEGREE OF DEGRADATION With DF optical microscopy (fig. 8) it is possible to distinguish the oxidised compounds by their specific colours (e.g., green and blue for differently hydrated carbonates, red for cuprous oxide and orange for mixed cuprous and stannic oxides). Raman spectroscopy (RS) confirmed the presence of copper corrosion products as: copper carbonates (azurite and malachite), cuprous oxide, mixed oxides of copper (I) and tin (IV), but also found out the probable presence of phosphates in the edge of the sample, as confirmed by X-Ray maps (fig. 9). Large parts of the external patina are amorphous and not giving specific results by RS due to the high level of fluorescence. Some mixture of soil products, corrosion products and eventual organic materials can be hypothesised but not demonstrated. However, it was evident in all samples that the internal side (i.e. the one in close contact with the support) was characterised by two main aspects: the thinner layers of corrosion compared to the external side and the presence of ele-

209

Paolo Piccardo and Dusko Temelkoski

2003; van Driel 2002; Budrugeac and Miu 2008) and chromium in leather is not coming from the soil (Strzelczyk et al. 1997). The study of P. T. Nicholson and I. Shaw (2000, 305) on Egyptian materials and technology is supporting our suggestion: ancient tanning and dying techniques are described and the usage of oak gall as a source of tannins is clearly demonstrated. It is known that in ancient practice mineral pigments (e.g. iron sulfates or chromium oxides) were added to concentrated solutions of Oak gal tannins in order to give them a specific tint (Senvaitienë et al. 2005). Such salts were known in antiquity and easily recognised by their intense colours. Being natural products, they could contain impurities (like nickel or zinc) while not really changing their properties. Taking into account all the previous data and information, it is plausible to suggest that the bronze shields were mounted and fastened on support made of wood and leather or simply leather, which was tanned and dyed with an organic base mixture of chromium and iron rich salts. These elements are permanently fixed in the leather and then participating to the formation of the external patina during the alteration process.

ments as chromium, nickel, zinc and iron. The presence of a support was thus slowing down the alteration degree on one side of the shield as long as it resisted before complete degradation. While decomposing itself, the organic shield support was leaving its traces in form of chromium, nickel and iron. The presence of these metallic elements differing from those typical of the alloy were analysed and quantitative XRay maps were realised in order to better understand the reason of their presence (fig. 10). As being major elements found in nearly all soils, iron, calcium, silicon and phosphorus are quite typical in the corrosion patina. However, nickel, zinc and, mainly, chromium are quite rare and thus their presence can originate from something different than the soil. According to H. van Wees (1994b), V. D. Hansen (1993) and L. A. Tritle (2007) leather and wood, often applied in several layers, were quite common materials for the production of weaponry as stand alone objects (Prag et al. 2008) or as support of metallic shields. The absence of organic materials now is not helping much to define if leather or wood and leather were used, but nevertheless it is possible to suggest that the material in direct contact with the inner face of the shield was leather. A few fibres were recognised during the DF observations in the soil rich patina over the edge of the cross sections on all samples. Their reduced size and very weak resistance to laser and electron beam made any kind of analysis impossible. Nevertheless the shape was quite consistent with some examples found in literature (Strzelczyk et al. 1997; Lindner and Neuber 1990). Moreover, well tanned leather is known to show a good resistance to moisture and to mechanical stress that are basic characteristics for materials to be used in special conditions (such as a combat or a long trip) and meant to last as long as possible with limited risk of failure. Tanned leather has an increased resistance to deterioration (Lindner and Neuber 1990; Covington et al. 2001). During the long permanence of the shields in the soil, this can result in a longer lasting protection of the bronze surface in close contact with the support to the corrosion agents. This would be consistent with the different thickness of the patina found in the same sample; the one edge of the cross section which is corresponding to the external face of the shield is much more corroded than the inner side. Chromium is actually one of the base elements for tanning and dying salts used in modern leather industry (Mazzocchin et al.

CONCLUDING REMARKS The discovery of at least three shields in Staro Bonče represents an important and still enigmatic step beyond the history and culture of Macedonia. The real meaning of the pit where the shields as well as ceramics and an iron knife were found is still unclear. For the moment it is only possible to presume that those finds were deposited in a pit presumably of ritual purpose in a broader sense. However the material culture being at the origin of the shields represents a ‘state of the art’ of technological advances in metalworking and is contributing to the continuous evolution of the history of techniques. As a matter of fact the archaeometric investigations performed on the three shields allowed the authors to reconstruct their manufacture techniques, the alloys used – each shield was cast with a different alloy, these ranging from 5–9 % of tin – and the thermomechanical working flow applied by the ancient craftsmen. Moreover, the characterisation of the corrosion patina gave information concerning at least one of the materials used to build the support to which the metallic decorated surface was fastened: tanned and dyed leather. 210

Metallurgical and archaeological investigations on the Macedonian shields of Staro Bonče (3rd century BC) ACKNOWLEDGEMENTS

Hanson, V. D. 1993. Hoplite technology in phalanx battle. In V. D. Hansen (ed.), Hoplites: The authors would like to acknowledge proThe classical Greek Battle Experience, 63–85. fessors B. Minčeva-Šukarova and O. Grupče London, Routledge. from the Institute of Chemistry, Faculty of Karamitrou-Menteside, G. 1987. O Mακεδουικοσ Natural Sciences and Mathematics, ‘Sts Cyril ταφοσ Σπηλιασ Εορδαιασ. To Αρχηεολοικο and Methodius’ University – Skopje, Republic Eργο στε Mακεδσυια και Θρακε I, 23–36. of Macedonia, for the help and the support they Lindner, W. and Neuber, H.-U, 1990. Preservation gave in order to realise the cooperation. Also, we in the Tannery. International Biodeterioration would like to thank Dr. V. Tanevska for being the 26, 195–203. first who believed in this work and supported us Langhorne J. and Langhorne, W. 1836. Plutarch. with Raman microspectroscopy. Furthermore, Lives. Translated from the original Greek, we want to thank A. Ervas, conservator and outwith notes, critical and historical, and a life standing blacksmith, Trento, and R. Amendola, of Plutarch. Baltimore, W. and J. Neal. University of Genoa, for sharing the studies on Markle, M. 1977. The Macedonian Sarissa, Copper base alloys fluidity and castability. Spear, and Related Armor. American Journal of Archaeology 81/3, 323–339. Markle, M. 1999. A Shield Monument from Veria BIBLIOGRAPHY and the Chronology of Macedonian Shield Types. Hesperia 68/2, 219–254. Adam-Veleni, P. 1989. Χ ασ  Mazzocchin, G. A., Agnoli, F. and Mazzocchin,    . Ancient S. 2003. Investigation of a Roman age ‘bulk Macedonia: Vth International Symposium, pigment’ found in Vicenza. Analytica Chimica vol. 1, 17–28. Thessalonike, Institute for Acta 475, 181–190. Balkan Studies. Nicholson, P. T. and Shaw, I. 2000. Ancient Ammanati, N., Martellucci, E., Ienco, M. G., Egyptian materials and technology. Piccardo, P. and Natali, S. 2004. Valutazione Cambridge, Cambridge University Press. dei parametri cromatici dei bronzi in funzio- Piccardo, P., Mille, B. and Robiola, L. 2007. Tin ne del tenore di stagno. Proc. 30° Convegno and copper oxides in corroded archaeologiAssociazione Italiana di Metallurgia, Vicenza cal bronzes. In P. Dillmann, G. Beranger, P. 17-19 nov. 2004, paper 10 [CD Rom]. Milan, Piccardo and H. Matthiesen (eds), Corrosion Associazione Italiana di Metallurgia. of metallic heritage artefacts. European Anderson, J. K. 1976. Shields of Eight Palms‘ Federation of Corrosion 48, 239–262. Width. California Studies in Classical Cambridge, Woodhead Publishing. Antiquity 9, 1–6. Berkeley and London, Piccardo, P., Amendola, R., Genocchio, P., University of California Press. Temelkoski, D., Minčeva-Šukarova, B. and Anglim, S., Jestice, P. G., Rice, R. S., Rusch, S. Grupče, O. 2008. Studies on 3rd century BC M. and Serrati, J. 2002. Fighting techniques of shields found in the Republic of Macedonia. the ancient world 3000 BC – 500 AD: equipProceedings of the Balkan Symposium on ment, combat skills, and tactics. New York, Archaeometry ’Science Meets Archaeology Thomas Dunne Books. and Art History’. Ohrid (Rep. of Macedonia). Budrugeac, P. and Miu, L. 2008. The suitability [http://archaeometry.ukim.edu.mk/events/ of DSC method for damage assessment and Announcement%20Ohrid.pdf] certification of historical leathers and parch- Piccardo, P., Amendola, R., Adobati, A. and ments. Journal of Cultural Heritage 9, 146– Faletti, C. 2009. Study of the fluidity of cop153. per-base alloys. La Metallurgia Italiana 5, Covington, A. D., Lampard, G. S., Menderes, O., 31–38. Chadwick, A.V., Rafeletos, G. and O’Brien, P. Piccardo, P., Amendola, R., Mille, B. and Adobati, 2001. Extended X-ray absorption fine structuA. 2010. Valutazione dell’influenza di stagno re studies of the role of chromium in leather e piombo sulla fluidità di leghe binarie a base tanning. Polyhedron 20, 461–464. rame. 33° Convegno Associazione Italiana di Dakaris, S. 1994. Dodona. Congleton, Old Metallurgia, 10–12 November 2010, Brescia Vicarage. (I), lecture 27. Milan, Associazione Italiana di Metallurgia. Grozdanova, V. B. 2007. Macedonian shield from Bonče. Situla 44 [Scripta Praehistorica Pinasco, M. R., Ienco, M. G., Piccardo, P., Pellati, G. and Stagno, E. 2007. Metallographic apin Honorem Biba Teržan], 863–874. 211

Paolo Piccardo and Dusko Temelkoski

proach to the investigation of metallic archaeological objects. Annali di Chimica 97, 553–574. Prag, A. J. N. W., Garner, R., Pantos, E., Bennett, S. L., Mosselmans, J. F. W., Tobin, M. J., Kockelmann, W., Chapon, L. C., Salvado N. and Pradell T. 2008. How the Greeks got ahead: Technological Aspects of Manufacture of a Corinthian Type Hoplite Bronze Helmet from Olympia. In S. A. Paipetis (ed.), Science and Technology in Homeric Epics, 205–220. Dordrecht, Springer. Scott, I., Griffith, G. T., 1973. The Age of Alexander. Nine Greek Lives by Plutarch. London, Penguin books. Senvaitienë, J., Beganskienë, A., Tautkus, S., Padarauskas, A. and Kareiva, A. 2005 Characterization of historical writing inks by different analytical techniques. Chemija, Lietuvos mokslo akademija T. 16/3–4, 34–38.

Strzelczyk, A. B., Bannach, L. and Kurowska, A. 1997. Biodeterioration of Archaeological Leather. International Biodeterioration and Biodegradation 39/4, 301–309. van Driel, M. C. 2002. Practical Evaluation of a Field Test for the Identification of Ancient Vegetable Tanned Leathers. Journal of Archaeological Science 29, 17–21. van Wees, H. 1994a. The Homeric way of war: the Iliad and the hoplite phalanx (I), Greece and Rome, Vol. XLI, No. 1, 1–19. London, Routledge. van Wees, H. 1994b. The Homeric way of war: the Iliad and the hoplite phalanx (II), Greece and Rome, Vol. XLI, No. 2, 2–26. London, Routledge. Tritle, L. A. 2007. War and peace among the Greeks. In K. A. Raaflaub (ed.), War and Peace in the Ancient World, 172–190. Oxford, Blackwell. Paolo Piccardo Dipartimento di Chimica e Chimica Industriale University of Genoa Via Dodecaneso 31 16146 Genoa Italy Dusko Temelkoski NI Institute and Museum-Prilep www.muzejprilep.org.mk Republic of Macedonia

212

Abstracts workshop Vienna 2009

THE RESULTS OF THE X-RAYS EXAMINATIONS OF SWORDS FROM POLAND Michał Bugaj Warszawa, Poland The paper concisely discusses the results of the X-ray examinations of the bronze swords with bronze hilts found in Poland. The results are presented on a wider, European background. The Late Bronze Age specimens have been given special attention, due to their highest occurrence

ratio in this period. Their analysis enabled the pinpointing of Late Bronze Age manufacturing sites of bronze swords in Poland, and shed light on their local charcteristics. Moreover, additional information on bronze sword manufacturing and construction methods has been gained.

‘BY WEAPONS MADE WORTHY’? PRODUCTION, USE AND DEPOSITION OF BRONZE AGE WEAPONRY David Fontijn Leiden, The Netherlands The above quote comes from Beowulf, and may be seen as exemplifying the well-rehearsed notion that pre- and protohistoric weaponry held high symbolical value. If there is one period to which this seems to apply, this would be the European Bronze Age. Weapons, and swords in particular, are associated with prestigious warrior equipment, signalling supra-regional identities. Ceremonial weapons are among the most outstanding pieces of craftsmanship known for this period. Nevertheless, weapons are made, used and discarded by people, and current research shows that their life-cycles may be various. The presentation was discussing the results of research on the life-cycles of weaponry,

concentrating on their production ‘birth’ and deposition ‘death’ of weaponry. It includes the results of recent experimental production of bronze swords by Jeroen Zuiderwijk in the Netherlands as part of our research project. Swords appeared to be made quite easily in some numbers, which begs the question, how such products could become so prominent in the adornment of elite warriors? This paper will throw into doubt whether swords as a category had ‘prime value’ and argue that we should understand the history of swords from their entire ‘cultural biography’. I will argue that, for archaeology, the crucial thing is to try to understand the selection of weapons in acts of permanent, deliberate deposition.

METAL ALLOY AND TEXTURE ANALYSIS OF BRONZE AGE SWORDS BY NEUTRON DIFFRACTION Evelyne Godfrey1, Winfried Kockelmann1 and Marianne Mödlinger2 1 ISIS, United Kingdom; 2Klagenfurt, Austria Metal microstructures and minerals can be observed in two-dimensions i.e. by sampling, preparation, and examination under the microscope. Metals and minerals can of course also be measured in three-dimensions, by diffraction analysis, e.g. X-ray diffraction. The major drawback of microscopy is that the metal artefact has to be cut; usually just one small sample

is permitted, and this may not be representative of the whole object, especially in the case of a functional and technically complex item such as a sword, where various portions of the blade may have had different heat treatment and intensity of mechanical working. By applying a non-destructive diffraction method, it is possible determine the properties of numerous

213

Abstracts workshop Vienna 2009

regions of the object. Conventional laboratory X-ray diffraction is useful for surface analysis of small artefacts and for powdered samples but in order to penetrate intact large objects, diffraction analysis using neutrons rather than X-rays is necessary. Eight full-metal hilted bronze swords of the Late Bronze Age from Austria were recently analysed at the ISIS Neutron Facility by the authors. Time-of-flight neutron diffraction (ToFND), a completely non-destructive method, was used to measure the bulk alloy composition and microstructure (crystallographic ‘texture’) across several sections along the blade of each sword. A total of around 150 data points – each covering a volume of 5 mm x 15 mm x the whole thickness of the object – were collected on the eight swords, in lines of overlapping areas from the edge to the mid-rib, providing the equivalent

of four or five non-invasively acquired ‘cross sections’ per sword. The data is being processed using Reitveld analysis, in order to quantify the copper, tin, and lead contents of the metal. The amount of tin segregation in the bronze alloy will indicate how close the metal is to either the as-cast condition on the one hand or the fully annealed state on the other. The intensity of mechanical working will be reflected in the crystallographic texture data. The results from the analysis from one of the swords have been given as an example. The aim of experiment was to non-destructively ‘map’ the microstructure of each sword across more areas than could be possible previously with microscopy, in order to demonstrate whether the use of the sword – e.g. as a stabbing or a slashing weapon – is indeed reflected by differences in working at particular parts of the blade.

LATE BRONZE AGE CRESTED HELMETS IN THE EASTERN ALPS Andreas Lippert Vienna, Austria In Böckstein, at the Northern foot of the Hohe Tauern, a high mountain range of the Eastern Alps forming the border of the counties of Salzburg and Carinthia, a fragment of a bronze ornamented crested helmet consisting of two different parts was recently found. It came from the sandy bank of a river along an old path up to the mountain pass of the Korntauern. In 1838 a complete crested helmet of the same type was discovered near the mountain pass of Pass Lueg, which connects the Northern flat lands of Salzburg with the Alpine area. The site lies high above the gorges of the Salzach river. Besides the helmet mining picks, an axe and copper ingots were found there. A half of another crested helmet appeared 2002 as a part of a big hoard in Fließ in Northern Tyrol and again in the vicinity of an important mountain pass. All three helmets and the fragments of helmets have been investigated archaeologically and metallurgically in a research project carried out by the Department of Prehistoric Archaeology at Vienna University (Andreas Lippert), the Interdisciplinary Archaeology Forum in Vienna (VIAS, Mathias Mehofer) and by the Curt-Engelhorn-Zentrum in Mannheim (Ernst Pernicka and Joachim Lutz). The helmets can be interpreted as offerings on significant

214

mountain pass routes. They form a late group of Bronze Age crested helmets in the middle of the Eastern Alps, while related groups of helmets are known in the areas north and west of the Alps. Dating to the Early Iron Age, however, is the Italian group of crested helmets of the Villanova and Etruscan Cultures. This group clearly has evolved from the earlier groups of crested helmets. The helmets of the type of Pass Lueg provide evidence of far reaching trade connections across the Alps in the 12th and 11th century BC. Questions about the origin of the copper used for the alloy of the bronze helmets can now be answered to a certain extent. The analyses of the techniques of the production and shaping of the helmets as well as of the ornamentation have shown that there were different, yet closely related workshops. PUBLISHED AS Lippert, A. 2010. Spätbronzezeitliche Kammhelme an transalpinen Fernhandelsrouten (mit einem Beitrag von M. Mehofer). Archäologisches Korrespondenzblatt 40, 49–65.

Abstracts workshop Vienna 2009

WEAPONS GRADE BRONZE Peter Northover Oxford, United Kingdom During the Bronze Age in Britain the manufacture of weapons became an increasingly specialised task driven by the requirements of the technology adopted. Spearheads were the first objects to be regularly cast hollow, while the progression from dagger to rapier to sword meant casting slender objects of ever increasing length. This paper will address the questions of whether or not specific alloys were ever developed and adopted for these tasks, and of how the alloys actually used were adapted to the needs of the weapons. Is there such a thing as weapons grade bronze? A simple answer is that there was not: in most cases the alloys in regular use for other products were equally well suited to swords and spearheads. There are some instances where a specific selection was made: In Chalcolithic Europe arsenical copper might be used for dag-

gers and halberds while unalloyed copper was used for axes, and in Middle Bronze Age Wales somespearheads had raised tin contents, possibly to give a harder cutting edge which could be sharpened by grinding rather than by re-forging a very narrow blade. The adoption of leaded bronze in the later Bronze Age in parts of Europe could have given more scope for alloy selection but generally it did not. The mechanical properties required of a sword blade might have prompted some selection but the bronzes used were sufficiently robust in their properties not to need this In the end the way to a good sword or spearhead was quality control at all stages of manufacture and this was more usually done for swords than spearheads. Finally some thought will be given to the question of why iron replaced bronze for weapons.

COMPUTED TOMOGRAPHY FOR THE 3D ANALYSIS OF METALLIC ARCHAEOLOGICAL FINDINGS Dietmar Salaberger and Johann Kastner Wels, Austria X-ray computed Tomography (CT) was established during the last few decades for the nondestructive testing and geometrical measurement of industrial components. The well established medical diagnostic method was modified to fulfil the special needs of industrial applications. Industrial CT scanning devices facilitate very high resolutions and x-ray sources that provide high energetic x-rays. For archaeological finds, CT is a very suitable method since the specimen is not destroyed or even affected by the measurement and valuable information about the inner structures can be acquired. Specimens made of metal provide a special challenge regarding CT measuring technique. The physical laws that determine the absorption of x-rays lead especially in dense materials (e.g. iron, copper) to measurement artefacts, meaning deviations between CT data and reality. The proper choice of measurement parameters and

application of artefact reduction software enable a noticeable improvement in image quality. Using examples, we discussed the possibilities, advantages and limitation borders of x-ray CT. On the pictures a belt buckle with mineralised remains of wool is shown. Metallic parts are highly visible in the CT data since the x-ray absorption is very high. This leads to higher gray values corresponding to the density of the material. In the slice images the damascening, that cannot be seen from outside, is clearly visible. Due to the mineralization of the wool it can be visualized by CT too. As additional examples CT results of a gynaecological instrument from Dion, Greece, damascene swords and several block excavations were shown. Acknowledgement: This work was supported by the Austrian Research Promotion Agency (FFG).

215

Abstracts workshop Vienna 2009

EXAMPLES OF IMAGERY IN ORNAMENTATION ON LATE BRONZE AGE WEAPONS Stefan Wirth Dijon, France Ornamentation on Late Bronze Age metalwork is generally purely geometric and usually avoids any figural depictions. Exceptions can be found mainly on metal vessels and certain weapon types. These possess rich ornamental designs, which should be understood as a meaningful system of signs that corresponds to the highly symbolic dimension of the objects. At the centre of these complex systems of decoration is the image of the ‘bird-sun-boat’ (barque solaire aux oiseaux, J. Déchelette 1909). The histories of this symbol and of the Urnfield culture of Central Europe are closely connected. The image brings the observation of the ever cyclical movement of the sun to life and was a symbol of certain cosmological ideas which were apparently valid for centuries. The configuration of the image is always applied strictly symmetrically and can be shortened and abstracted. In a close concurrence with the object that carries the image, three-dimensional, and even palindromic, structures are created. These are of mnemonic quality: the essential message is not

configured, but can still be identified by insiders. The individual creation of the image through the hand of the craftsman possesses therefore an identity-creating dimension. Exceedingly individualised versions of this symbol are present on the hilts of certain full-hilted swords (Vollgriffschwerter), corresponding to a personalised, i.e. developed for the sword owner, manufacture of the weapon. E. Sprockhoff (1954) already recognised in the hilt decoration on certain swords an encoded version of the ‘bird-sun-boat’ symbol. Through the Prähistorische Bronzefunde series a large part of the relevant pieces are now available for a closer observation and on a European scale. However, it was found that until now many details did not receive enough attention in the research. Only through a large-scale study, including the other objects bearing this symbol such as sheet vessels will the broader connections become visible. Translation: Marion Uckelmann

Figure 1: Sword from Pliening, Kr. Ebersberg, Germany (after v. Quillfeldt 1995). – Figure 2: Sword from Hatting, Land Tirol, Austria (after Wilhelm 2003, drawing A. Blaickner, Innsbruck). Scale 1:2 216

Abstracts workshop Vienna 2009

BIBLIOGRAPHY Déchelette, J. 1909. Le culte du soleil aux temps préhistoriques. Paris. Sprockhoff, E. 1954. Nordische Bronzezeit und frühes Griechentum. Jahrbuch des Römisch-

Germanischen Zentralmuseums 1, 28–110, pl. 7–9. Wirth, S. 2006. Vogel-Sonnen-Barke. Reallexikon der Germanischen Altertumskunde (2nd edition), vol. 32, 552–563, pl. 27.

STUDIES ON THE PRODUCTION AND FUNCTION OF LATE BRONZE AGE SWORDS Michael Siedlaczek Berlin, Germany Investigations of a large number of swords, including analyses of production- and stressmarks, as well as usage traces and damages reveal important information on production techniques and function of this special weapon in Bronze Age Europe. More than 150 Late Bronze Age swords from Central Europe were examined in a study macroand microscopically. Investigations of sword

morphology, casting quality, mounting manner including observations of surface traces, like abrasion, fracture damages, notches, resharpening traces, etc. were carried out and allow statements on a more empirical basis. Beside the theoretical examination of manufacture and usage, also experimental castings as well as usage of the replicated swords additionally offered new perspectives on these weapons.

217

Abstracts EAA, Riva del Garda 2009

EXPERIMENTAL COMPARISON OF SPEARHEADS USE-WEAR ANALYSIS Kate Anderson Edinburgh, Scotland A practical understanding of how weapons were used in the past can provide illumination on the types and scale of conflict taking place, in addition to the quantity and consistency of training required of participants. However, the majority of research into the functionality and modes of use of spearheads is dominated by binary thinking, forcing considerations of use to fall into discussions on throwing or thrusting. Examination of the Late Bronze Age spearheads of northern Britain show that 40 % are edge hardened, 32 % exhibit diagnostic combat damage patterns on their edges (identical to those seen on contemporary swords) and less than 19 % show any form of point damage, as might be expected if primary usage involved throwing or thrusting. Therefore, this paper argues that traditional approaches attempting to divide spears into these two groups on the basis of morphology is misleading. Research into functionality focusing

on edge wear analyses, rather than morphology, may produce more nuanced, realistic results. The aim of this paper was to present the results gained through testing of the hypothesis that the spearheads were hafted with a very short shaft and wielded in a similar manner to the swords. A number of spearheads, cast in 1 % leaded bronze, were commissioned, in addition to swords and shields, and were used in a range of experiments designed to determine diagnostic spear point damage as a result of combat, and the relationship between damage types and strikes. PUBLISHED AS Anderson, K. forthcoming. Slashing and thrusting with Late Bronze Age spears: analyses and experiment. Antiquity (2011).

HALBERDS AND RED HERRINGS. WEAPONRY AND WARFARE IN THE EARLIER BRONZE AGE OF SOUTHEAST SPAIN Dirk Brandherm1, Sandra Montón2, Margarita Sánchez3, Gonzalo Aranda3 1 Belfast, Northern Ireland; 2Barcelona, Spain; 3Granada, Spain; The introduction during the Earlier Bronze Age of halberds and swords as the first specialised combat weapons in southeast Spain, as well as the interpretation of other implements, such as daggers and axes from El Argar burials also as combat weapons has given rise to the notion that warfare was endemic to the structure of Argaric society. Concepts such as ‘armed conflict’, ‘war’, ‘violence’, ‘warrior’, and ‘militarism’ are commonly encountered in studies on the El Argar culture. Only on rare occasions, however, is the use of these concepts critically examined, and very little attention so far has been paid to what empirical evidence there actually is for a widespread use of such weapons in interpersonal combat. In this paper we follow two different lines of enquiry to address these questions. On the one hand we are examining halberds, swords and daggers from El Argar contexts for traces of combat-induced use-wear, on the other hand we are presenting the results of a survey carried out on

skeletal material from Argaric burials to identify combat-induced injuries. With regard to the first line of enquiry, only some of the halberds display use-wear patterns indicative of their probably use in interpersonal combat. The very limited number of halberds in the archaeological record, though, would not necessarily seem to support the notion of endemic warfare. These results are consistent with those from our second line of inquiry, where our survey of skeletal material shows a lack of injuries from edged weapons. PUBLISHED AS Aranda-Jiménez, G., Montón-Subías, S. and Jiménez-Brobeil, S. (2009). Conflicting evidence? Weapons and skeletons in the Bronze Age of south-east Iberia. Antiquity 83, 1038– 1051.

218

Abstracts EAA, Riva del Garda 2009

WEAPONS OR TOOLS? INSIGHTS INTO THE USAGE OF METALWORK IN COPPER AGE ITALY Andrea Dolfini Newcastle, United Kingdom Interpretations concerning the function of prehistoric metal artefacts has been neglected until recently. The reason seems to lie in a plurality of factors ranging from the frequent lack of contextual data to a scholarly interest centred on either production or deposition, at the expense of the middle phase of the life-cycle of objects. This problem has been further aggravated by the lack of a specific analytical technique to investigate whether, what for and to what extent metal artefacts were used in antiquity. Over the last decade, however, the principles and methods of usewear analysis have been applied to prehistoric copper and bronze objects from around Europe with encouraging results. It emerged that, although still pioneering, this technique can give

valuable insights into the ancient usage of metalwork. This paper discusses the results of usewear analysis carried out on a sample of copper and arsenical-copper axes, daggers and halberds from pre-Bronze Age sites in central Italy. It explores in particular whether wear marks can be interpreted as resulting from craft activities or combat. PUBLISHED AS Dolfini, A. 2011. The function of Chalcolithic metalwork in Italy: an assessment based on use-wear analysis. Journal of Archaeological Science 38, 1037–1049.

THE SPIRIT OF THE SWORD AND SPEAR Mark Pearce Nottingham, United Kingdom From its appearance in the Norse sagas or the Arthurian cycles, we are used to the concept that the warrior’s weapon has an identity, a name. In this paper I shall pick up an idea put forward by J. D. Cowen (1966, 294) who argued that in the twin spiral motif on Boiu type swords ‘…its makers intended … a pair of eyes and maybe a brow over them as well. These would then be the eyes of the sword itself, or rather of the spirit within it …’. I shall argue that the same explanation can be put forward for the decoration on a number of early Iron Age Italian spearheads (Bietti Sestieri and Macnamara 2007, no. 387. 388. 458) and that it might even explain the loops in the basal-looped spearheads of the British Isles, which date to the local middle Bronze Age (1300–31000 BC) (Davis 2006). If a later Bronze Age / early Iron Age belief in the spirit or identity of a weapon as documented by Boiu swords or Italian spearheads is accepted this has implications for their biographies. I will argue that this suggests that they may have

been conserved as heirlooms or exchanged as prestige gifts for much longer than is generally assumed and that it has implications for our understanding of the deposition of weapons in tombs, where they may have a ‘guardian spirit’ function. BIBLIOGRAPHY Bietti Sestieri, A. M. and Macnamara, E. 2007. Prehistoric metal artefacts from Italy (3500– 720 BC) in the British Museum. BM Research Publication 159. London, British Museum. Cowen, J. D. 1966. The origins of the flange-hilted sword of bronze in continental Europe. Proceedings of the Prehistoric Society 32, 262–312. Davis, R. 2006. Basal-looped Spearheads. Typology, chronology, context and use. British Archaeological Reports, International series 1497. Oxford, BAR Publishing.

219