Exploring Ancient Textiles: Pushing the Boundaries of Established Methodologies

By Oxbow Books

Over the past 30 years, research on archaeological textiles has developed into an important field of scientific study. It has greatly benefited from interdisciplinary approaches, which combine the application of advanced technological knowledge to ethnographic, textual and experimental investigations. In exploring textiles and textile processing (such as production and exchange) in ancient societies, archaeologists with different types and quality of data have shared their knowledge, thus contributing to well-established methodology. In this book, the papers highlight how researchers have been challenged to adapt or modify these traditional and more recently developed analytical methods to enable extraction of comparable data from often recalcitrant assemblages. Furthermore, they have applied new perspectives and approaches to extend the focus on less investigated aspects and artefacts.

The chapters embrace a broad geographical and chronological area, ranging from South America and Europe to Africa, and from the 11th millennium BC to the 1st millennium AD. Methodological considerations are explored through the medium of three different themes focusing on tools, textiles and fibres, and culture and identity. This volume constitutes a reflection on the status of current methodology and its applicability within the wider textile field. Moreover, it drives forward the methodological debates around textile research to generate new and stimulating conversations about the future of textile archaeology.

• Language: English
• Publisher: Oxbow Books
• Release date: Jul 20, 2022
• ISBN: 9781789257267

Book preview

Preface

Ian Shaw

As a participant in the conference that formed the basis for this book, I was privileged not only to hear the initial versions of many of the papers published here but also to engage in discussion with the many textile researchers gathered in Liverpool in October 2018. It is now an equal pleasure to contribute this brief preface to a book that will, I am sure, make a lasting contribution to global studies of ancient textiles, particularly in the area of scientific methodology. The geographical and chronological range of the book is impressive, stretching from Cornwall to Peru and Russia, and from early prehistory to the dawn of the Middle Ages. The cutting-edge scientific methods employed include strontium isotope analysis, energy dispersive x-ray spectroscopy, and use-wear analysis, while the more theoretically informed papers take the study of textiles in the direction of such prominent issues as self-representation, neuroarchaeology, socio-economic networking, and mobility of craftworkers. As this book demonstrates, textile production, like other forms of ancient technology, is often deeply embedded in idiosyncratic social and cultural structures.

Since my own research over the last few decades has focused primarily on ancient technology, specifically in Egypt and the Eastern Mediterranean, I am hugely aware of the great strides that have been made not only in terms of archaeological methodology across a wide range of materials and crafts, but also with regard to increased recognition of our need to analyse and interpret this new data within relevant theoretical frameworks. Crucially, therefore, many of the papers in this volume not only focus on the application of innovative analytical techniques but also on the profound ways in which society shapes technology, and, conversely, the technological shaping of society itself. How are we to unpick these complex cultural interconnections, when, as Bill Gates has observed: ‘The advance of technology is based on making it fit in so that you don’t really even notice it, so it’s part of everyday life’?

As the editors indicate in their introduction to this book, studies of textiles (and also of the tools used to produce them) have been very much a part of the inexorable rise of combinations of scientific and theoretical approaches to early technology. Our AHRC-funded project at the University of Liverpool (‘Contextualizing textiles: using the Bolton Museum collection to explore social and international contexts of Egyptian Bronze Age-Islamic cloth’), pursued jointly with Bolton Museum, and incorporating the PhD research of two of the editors of this volume, has hopefully played a small role in this process of change, at least in the relatively neglected area of ancient Egyptian and Nubian textile studies.

The subtitle of this volume focuses on ‘pushing boundaries’, and many of the papers in these pages clearly indicate that progress in textile research methodology in recent years has often been radical and paradigm-changing rather than simply incremental. Two good examples of such leaps forward in recent years that are highlighted in the introduction to this volume are, firstly, the increasing recognition that splicing was an important aspect of textile production in most ‘cloth cultures’, and, secondly, the increasing ability of traceological analysis of textile tools to elicit subtle distinctions between actual use wear and the effects of manufacturers’ marks and post-depositional actions. Such specialist studies of both textiles and tools have been particularly enhanced by the work of the Centre for Textile Research at the University of Copenhagen, which has clearly been the single biggest influence on modern archaeological approaches to textiles, and its multifarious contributions can be seen in most of the papers gathered here.

Introduction: Ancient tools and textiles – Thinking outside the box

Gabriella Longhitano, Sarah Hitchens, Alistair Dickey, and Margarita Gleba

Advances in methodological practices and analytical techniques continue to push forward the rigour of archaeological textile research. For example, in 2021, Neolithic textile material from Çatalhöyük, studied on multiple occasions in the past (Burnham 1965; Ryder 1965; Vogelsang-Eastwood 1987; Rast-Eicher and Jørgensen 2018), was re-identified as tree bast using Scanning Electron Microscopy (Rast-Eicher, Karg, and Bender Jørgensen 2021). Similar re-analysis of material from Wadi Mubarra’at (Israel), 9500 cal BC (Shamir and Rast-Eicher 2020), has also led to a re-identification of the fibre, again, tree bast. Dye analysis of Iron Age textiles recovered from Danish bogs using High Performance Liquid Chromatography demonstrated that 80% of them were dyed and brightly coloured rather than decorated using naturally pigmented wool only (Vanden Berghe et al. 2009). This shows that there are always new details to (re)discover in textile research, even when it comes to the old and well-known finds, which have the potential to drastically change our understanding of past activities relating to textile production. Methodological development is constantly ongoing, as it seeks to meet the needs and aspirations of textile and broader archaeological and historical investigations.

The case studies presented in this book aim to reflect upon the status of the viability and applicability of certain current methodological approaches within the wider archaeological textile research field. This includes a wide array of indirect and direct evidence of textile activity and its products. Over the past two decades, numerous important developments in excavation, documentation, and conservation methodologies have allowed new evidence to be recovered, ranging from fragments of extant garments to mineralised and imprinted textiles to a range of textile implements. With the advancement of the research, a plethora of scientific studies have been established to let these new sources of evidence unravel their stories. In turn, this has been bringing into motion a wider discussion on how to combine in rigorous ways analytical and more traditional approaches for extracting comparable data from (often recalcitrant) assemblages.

The preservation of textiles

One of the main problems within archaeological textile research is the unevenly distributed evidence in the archaeological record. As organic material, textiles are inherently liable to decomposition, unless specific conditions alter this natural process (Wild 1988; Peacock 2005; Jones et al. 2007). Hyper-arid climates like Peru (Beresford-Jones et al. 2018) or Egypt (Barber 1991; Kemp and Vogelsang-Eastwood 2001), wet or waterlogged environments (Felding 2016; Knight et al. 2019), permafrost conditions as found in Greenland (Østergaard 2004; Hayeur Smith 2014), the presence of salt (Bichler et al. 2005; Grömer et al. 2013; Grömer and Rösel-Mautendorfer 2014), or exposure to fire leading to charring (Rast-Eicher and Dietrich 2015; Rast-Eicher, Karg, and Bender Jørgensen 2021) will often afford unique preservation conditions for textiles and fibres. The likelihood of a textile surviving in the archaeological record also depends on the type of fibre the textile is made from. Animal fibres survive better in acidic environments whilst plant fibres preserve better in more alkaline environments (Cybulska and Maik 2007; Mannering and Skals 2014). Furthermore, textiles survive in archaeological contexts much more frequently than is commonly believed, in the form of imprints (Jansone 2017; Rammo 2019; Ulanowska 2021) or mineralised fragments (Chen et al. 1996, 1998; Price and Gleba 2012; Margariti and Papadimitriou 2014; Gleba 2017; Meo in this volume). Even when textiles themselves do not survive, their presence can be inferred from the position of skeletal remains in a burial (James in this volume). Indeed, the quantity of archaeological material is constantly increasing, particularly thanks to improved excavation and conservation procedures (Jones et al. in Gillis and Nosch 2007), as in the case of the amazing Late Bronze Age textile and fibre finds from Must Farm in the UK (Knight et al. 2019).

The preservation of textile tools

Unlike the textiles themselves, textile tools are frequently discovered at archaeological sites and constitute the single most important and plentiful type of evidence for the assessment of the technology, techniques, scale of production, and thus the economic aspects of past textile industries (Andersson 2003; Gleba 2008; Andersson Strand and Nosch 2015). Indeed, on many archaeological sites, the only evidence for the presence of textiles comes in the form of tools used in spinning and weaving. This is often in the form of spindle whorls and loom weights as they were often made from inorganic material such as clay, metal, or stone and thus survive well in the archaeological record. However, other textile implements such as distaffs (Rahmstorf 2015; Langgut et al. 2016), needles (Adams 1996; Walton Rogers 1997; Adams and Adams 1998), spindles (Walton Rogers 1997; Adams 2013; Langgut et al. 2016), spinning bowls (Dothan 1963; Kemp and Vogelsang-Eastwood 2001; Ruiz de Haro 2018; Spinazzi-Lucchesi 2020), shears (Walton Rogers 1997; Notis and Sugar 2003; Rosell Garrido and Spagiari in this volume), shuttles (Adams and Adams 1998; Adams 2013), weaving combs (Adams 2010, 2013), and weaving tablets (Adams 1996; Gleba 2008) have also been excavated. Implements such as distaffs, spindles, shuttles, and weaving combs are usually made from organic materials such as wood or bone and often do not survive in the archaeological record.

Analytical techniques applied to the study of textiles

In recent decades, various scientific methods have been co-opted from other disciplines to aid in fibre and dye identification, fibre analysis, and determination of whole textile structures. The combination of scientific knowledge and hardware using e.g. digital portable microscopes (such as Dino-Lite), Scanning Electron Microscope-Energy Dispersive X-ray Spectroscopy (SEM-EDS; see Dickey in this volume), Portable X-Ray Fluorescence (pXRF), and Attenuated Total Reflection-Fourier Transform Infrared Reflectance (ATR-AFTIR), have allowed textiles to be measured and described in new ways and often in great detail.

The methods used in the analysis of archaeological textiles range from traditional structural analysis using a hand lens to the advanced methods mentioned previously. When possible, technical analysis of textiles should be nondestructive and based upon visual examination. The basic analysis of any textile should include the following: textile dimensions, weave type, thread diameter, spin direction, and twist angle, thread count, colour description, and a record of selvedges, fringes, or borders, decorative elements, structural sewing, wear and tear, as well as a general description of the textile (Walton and Eastwood 1988). This information can be collected with just the use of a hand lens or a stereomicroscope, depending on the quality and preservation of the textile and the size of the fragment.

One of the recent advances in textile structural analysis is the wider recognition of splicing, one of the earliest yarn-making technologies used, in prehistoric Europe and western Asia (Gleba and Harris 2019), as well as in South America (Beresford-Jones et al. 2018). It is well attested in prehistoric European contexts such as the Neolithic pile-dwelling settlements in eastern Switzerland (Leuzinger and Rast-Eicher 2011). Splicing was also a common technique used in linen production during the Chalcolithic period in the southern Levant (Shamir 2015) and was also commonly used in ancient Egypt (Cooke et al. 1991; Granger-Taylor 1998, 2003; Kemp and Vogelsang-Eastwood 2001; Dickey 2019). Its recognition as a ubiquitous technique in the past forces scholars to re-evaluate the textile chaîne opératoire and the use of certain tools, such as spindle whorls in the Neolithic and Early Bronze Age.

Fibre identification and analysis

Some of the most significant advances in textile archaeology over the last decade have been in fibre identification and analysis. Identification of fibres is important as it provides valuable information about the use of wild and cultivated textile resources.

Archaeological textile fibres can be divided into three groups: animal, plant, and mineral (such as asbestos). The most common fibres encountered in archaeological contexts, however, are those derived from plants and animals. They are distinguishable by their molecular composition as plant fibres are cellulose based and animal fibres are made of proteins.

One of the oldest methods used in the identification of fibres is the burn test. By design, it is destructive and provides little information beyond the fact that animal or plant material was used. Textile samples that are highly degraded can be identified using chemical tests such as solubility measurements. However, chemical tests such as zinc chloride iodide can only provide information about the biological source of the fibre (Gleba 2008, 64). Chemical and solubility tests are unable to distinguish between different species of animals and plants (Landi 1998; Gleba 2008).

Currently, the fastest, most affordable, and easiest means by which to identify archaeological textiles fibres is microscopy since it allows for much closer observation of fibre structures. This includes techniques such as interference microscopy, Transmitted Light Microscopy (TLM), Polarised Light Microscopy (PLM), Scanning Electron Microscopy (SEM), including the Environmental Scanning Electron Microscopy (ESEM) and Variable Pressure SEM, and Transmission Electron Microscopy (TEM). Unless the fibres are heavily degraded, distinguishing between plant and animal fibres is generally straightforward. The principal morphological features of animal fibres are cuticular scales, the shape and position of which on the fibre’s surface can be diagnostic for different species (Appleyard 1960; Textile Institute 1985, 5; Rast-Eicher 2016, 11–13). Plant bast fibres have characteristic nodes or dislocations, while cotton, a seed fibre, has a ribbon-like structure with convolutions (Rast-Eicher 2016, 14, 73).

Microscopic methods, however, are less reliable in distinguishing between fibres of different species of plants or animals that have similar characteristics. Thus, many bast fibres share common features such as a lumen, dislocations (nodes), and cross markings (Haugan and Holst 2014, 952). These shared features make it difficult to distinguish between different bast fibres (Lukesova 2017). Recently, numerous studies have attempted to develop additional protocols that might help resolve bast fibre identification problems, focusing on features such as fibre and lumen cross-sections (Lukesova and Holst 2021), fibre microfibrillar orientation (Bergfjord and Holst 2010), and other morphological traits. To date, most of these studies have primarily focused on European bast fibres (flax, hemp, nettle), but these approaches are beginning to be expanded to species used in the Pacific (Patterson, Lowe, and Smith 2017) and South America (Alday in this volume).

Plant fibres, particularly tree-bast and bast fibres like flax, were the first raw material used to make mats, strings, cords, etc. (Rast-Eicher 2005; Breniquet 2008, 85–90; Karg 2011; Banck-Burgess 2018; Karg et al. 2018; Médard 2018), and they have remained in use even after the introduction of animal fibres.

Among these, sheep wool has been the major fibre for textile making. Yet, although the sheep domestication process commenced in the Fertile Crescent approximately 9000 years ago, the direct evidence for the use of wool fibre in textile production can be dated back no earlier than the fourth millennium BC (Good 1999). In the following millennia, wool became an important and, in some areas, the primary textile material (Sabatini and Bergerbrant 2020; Shishlina et al. in this volume). The reason for this relatively late adoption of wool as a textile material is the fact that the early domesticated sheep did not look anything like the modern animals and produced very little usable fibre. The modern wool fleeces were achieved through selective breeding of sheep (Ryder 1969, 1983). The ever-increasing number of analysed samples demonstrates that Bronze Age sheep had fleeces with extremely fine underwool and coarse hair and kemp, but by the Iron Age they were more homogeneous (Rast-Eicher 2008; Gleba 2012; Rast-Eicher 2013). These developments happened earlier in some areas than others, so that when large sets of textiles are analysed, distribution patterns permit the identification of fleeces that do not fit the general pattern and may therefore be identified as of nonlocal origin (Rast-Eicher and Bender Jørgensen 2013, 1231).

Microscopy is useful for identifying relatively well-preserved plant and animal fibres, however, the most reliable way to distinguish between different species of animal fibres is by ancient DNA analysis or palaeoproteomics. These biomolecular methods can also answer deeper questions such as the origins of the raw materials used, species, or even breed identification, and provide more precise dating. Studies on ancient DNA may be used in the identification of animal species as well as specific gene development studies such as those observed in the evolution of wool and hair or animal migration and domestication (Ørsted Brandt and Allentoft 2020). Polymerase Chain Reaction Sequencing and Mass Spectrometry-Based Peptide Sequencing has also been used to study ancient rope fibres from the Christmas Cave in Israel and to distinguish between flax and hemp fibres (Murphy et al. 2011).

Another method to identify animals to species is proteomics. For example, analysis of relict proteins has shown that the Salish of west coast North America made their blankets out of dog hair, interweaving it with goat, and that the woolly dog was superseded by sheep by the late nineteenth century (Solazzo et al. 2011).

Dye identification

Addition of colour has been an integral part of textile making, contributing to pattern design (Cardon 2007). Yet, archaeological textiles often survive as discoloured rags or mineralised formations, making it difficult to visualise what colour they originally had. This can in some cases be reconstructed through dye and mordant analysis using High or Ultra-Performance Liquid Chromatography, HPLC/UPLC (Vanden Berghe et al. 2009). The technique allows identification of chemical dye components and their combination, which can then be matched to the database of known plant and animal dye sources.

Dating

Until recently, textiles were dated primarily using methods of relative dating, which are based on context or comparison of the object to similar items found on dated archaeological sites, yet the results are simply terminus ante quem as the textile may have been old when it was buried. Furthermore, the traditions of continuity meant that certain technical or artistic features were reproduced virtually unchanged for centuries, thus one item seemingly identical to another could have been made hundreds of years later or indeed earlier than the ‘dated’ example. The most consistent way to accurately date fibres has been by using radiocarbon analysis. It is a radiometric dating method that uses the naturally occurring radioisotope carbon-14 (¹⁴C) to estimate the age of carbon-bearing (i.e. organic) materials. The Accelerator Mass Spectrometry method (AMS), developed in the last 20 years is proving a particularly important dating tool. Textiles made over 500 years ago are especially suitable for ¹⁴C dating and may even give more precise dates than other material (Mannering et al. 2010). Dating a large number of textiles, particularly of a specific type, can help in tracing chronological developments in the spread of textile technologies (e.g. in the case of wool, see Shishlina et al. in this volume) and styles.¹ The technique is also important for the identification of fakes and forgeries, as well as textiles without well-known provenance. For example, recent re-dating of a textile from the Hallstatt salt mine presumed to date to the Bronze Age has shown that they in fact are from the sixteenth century AD (Grömer et al. 2020). Another old textile find from the Vesuvian area assumed to be of Roman date, also turned out to be much later (Galli et al. 2018, 276).

Provenance

Identifying the provenance of archaeological textiles in absolute terms is often difficult, if not impossible. Developments in strontium isotope tracing have seen its application in provenancing organic materials, thus aiding in ancient human and animal mobility studies. Strontium isotope ratios have recently been shown to be a useful indicator for wool fibre provenance in some cases. Strontium analyses of some Danish Iron Age textiles found in bogs demonstrate that not all of them had a local origin (Frei et al. 2009a, 2009b). This method was used in the analysis of the Egtved Girl and her garments, buried in an oak coffin in Denmark dated to 1500–1100 BC (Frei et al. 2015). The use of Sr to provenance textiles is somewhat controversial, however, since studies indicate that ⁸⁷Sr/⁸⁶Sr ratios of archaeological wool textiles recovered from wet burial environments do not accurately reflect wool provenance even after cleaning (von Holstein et al. 2016) and the isoscapes may have been affected by recent anthropogenic factors (Thomsen and Andreasen 2019).

In addition to strontium, combined carbon, nitrogen, and hydrogen isotopes can also be used in archaeology to establish geographic origin. Gradients in stable isotopes identified in modern studies of European sheep meat and wool have now been successfully applied to medieval archaeological wool samples from Iceland, the United Kingdom, Germany, and Sweden. Analysis has shown the isotopic composition of wool and bone collagen samples clustered strongly by settlement, demonstrating the feasibility of provenancing keratin preserved in anoxic waterlogged contexts (van Holstein et al. 2016).

Analytical techniques applied to the study of textile tools

Until recently, textile tool studies involved the cataloguing of finds and development of their typologies based on shapes (Crewe 1998; Rahmstorf 2015). A major impetus to textile studies in Mediterranean Europe and more broadly has been without doubt the establishment of the Centre for Textile Research (CTR) at the University of Copenhagen in 2005, funded by the Danish National Research Foundation. The Tools and Textiles Texts and Contexts project has revolutionised the way we study textile tools not only in the Mediterranean but across past cultures.² The new methods developed at CTR allow researchers to calculate the range of textile qualities obtainable using specific tools – an indispensable tool in the absence of actual textiles on many sites.

More recently, however, this methodology has been found difficult to apply to other geographical areas and periods. Several papers in this volume highlight the necessity of experimenting with CTR methodology to adapt it to other kinds of assemblages (Ferrero in this volume; Meo in this volume), less well-studied techniques such as splicing and braiding, and to use it in conjunction with other core methodologies such as GIS (Grzybalska 2010). Reflection upon methodology as a whole has prompted some authors to extend the focus on less studied artefacts such as shears (Rosell Garrido and Spagiari in this volume).

Spindle whorl analysis

The weight, diameter, and height of a spindle whorl affects how the whorl spins and influences the type of yarn produced. They are also the three most important measurements needed for the analysis of spindle whorls (Andersson Strand and Nosch 2015, 146). The material from which the whorl was made, its shape, perforation diameter, and decoration are also important to note (Liu 1978; Barber 1991, 39–78; Crewe 1998; Andersson and Nosch 2015). These attributes are part of the whorl’s functional characteristics.

The two most important measurements are weight and diameter (Smith and Tzachili 2012, 144). The weight of a spindle whorl influences how the tool spins and the thickness of yarns produced. It is generally accepted that heavier whorls were used to spin thicker yarns and lighter whorls were used to spin thinner yarns (Olofsson 2015, 32). Thinner yarns also contain fewer fibres than thicker yarns which contain more fibres. It has been suggested that heavier whorls are better for spinning thicker or longer fibres such as flax, while lighter whorls are best for spinning shorter fibres such as wool (Crewe 1998, 5–7; Smith and Tzachili 2012, 144).

Whorl diameter is the second most important measurement needed for the analysis of spindle whorls. The whorl’s diameter determines how fast it will rotate on a spindle and, as a result, how it will influence the tightness of the fibre spun (Smith and Tzachili 2012, 144).

However, some spinning experiments showed that the skill of the individual spinner influenced the quality of the spun yarn more than the whorl’s mass, moment of inertia, or the fibre (Kania 2013). The nature of fibre used also has an effect on the type of tool used (Andersson 1999, 2003, 25). In the end, it is the relationship between all these parameters that influences and determines the functionality of the spindle and its whorl as well as the types of yarns produced (Hudson 2014).

Loom weight analysis

The primary function of loom weights is to keep the warp of a warp-weighted loom taut by tying them to the warp threads. The warp-weighted loom was used from the Neolithic period onward over a large geographical area, comprising most of Europe and parts of western Asia (Barber 1991). Weights are often recovered on archaeological sites as they were most often made of fired or unfired clay, although stone examples are also known (e.g. Hoffmann 1964; Poursat 2012). They can thus be used as a proxy for textile production, particularly in places where organic preservation of textiles is poor. For example, the adoption of Cretan-style discoid loom weights at a number of Bronze Age settlements across the southern Aegean has been used to trace the spread of the warp-weighted loom across the region (Cutler 2021). Loom weights can also be used as cultural indicators and also allow us to track craftspeople’s mobility (Cutler 2021; Longhitano in this volume).

Theoretically, anything of sufficient weight can be used as loom weight, even simple stones or pebbles, as long as there is a hole or some other feature allowing the warp threads to be attached to it securely. However, the type of weight affects the weaving process.

Ethnographic studies have recorded a wide variety of shapes and sizes. The biggest loom weights recorded weigh more than 3 kg (Hoffmann 1964, 21). Hoffmann (1964, 42) also reported in Scandinavian communities the use of loom weights of remarkably different sizes in the same loom set up. Women simply tied proportionately more warp threads to heavier weights, and fewer to the lighter ones for producing very coarse fabrics.

Experimental tests with warp-weighted looms have been carried out since the 1940s with the main focus on the textile to be reconstructed (e.g. Broholm and Hald 1940; Andersson Strand and Nosch 2015). Overall, these tests have shown that weight and thickness are the primary functional parameters of a loom weight. The weight of the loom weight plays a crucial role as, in combination with the thread diameter, it determines the thread tension, namely the required weight to keep the warp in place during weaving, and the number of threads that can be attached to a single loom weight. However, it is worth noting that the warp tension depends also on how tightly the yarns are spun, on the fibre quality, and on how well fibres have been cleaned and sorted (see e.g. Andersson Strand 2015, 39–44).

Thickness is another important factor that has been investigated by experimental studies. Thickness determines how closely the loom weights hang side by side, and this has a major effect on the density and on the quality of the weave. In general, it has been demonstrated that the thicker the loom weight the looser the weave, the thinner the loom weight the denser the fabric. Also, tests have shown that it should be preferable to use loom weights with a total width, when hanging in a row, which is identical or slightly larger than the width of the fabric to be produced. Indeed, warp threads hanging outwards, or inwards would not be optimal as the threads would not be evenly distributed and the weaving would not be either even or regular (Olofsson et al. 2015, 91–2).

Experiments were carried out to test whether a find should be interpreted as loom weight or not, as in the case of the so-called spools (e.g. Ræder Knudsen 2012; Olofsson et al. 2015). Their function has been widely debated, as no ancient representations of spools as we understand them (for holding thread) exist.

Use wear analysis

Another important development in textile tool studies has been the use wear analysis. Studying the use wear marks on the surface of textile tools can help identify the possible use of an artefact as not all round perforated objects are spindle whorls nor were all tools found in graves used. This approach has its starting point in the assumption that different stages of tool production, use, and discard leave differing marks on the tool’s surface. This type of analysis is not new and has been applied to ceramic vessels (Skibo 2015). However, traceological analysis has only recently been carried out on ceramic spindle whorls and spools (Forte and Lemorini 2017). Recent traceological analysis of Italian ceramic textile tools dated to the first millennium BC as well as of experimental ceramic textile tools has allowed researchers to distinguish between manufacturing marks (modelling, surface treatments, and firing techniques), use wear marks, and damage caused by post-depositional processes (Forte and Lemorini 2017).

It is not always easy to demonstrate that weights were used as loom weights, unless they are found in sets (Wild 1970, 62; Barber 1991, 92–3). Other objects, such as thatch weights, spit supports over fires, pot supports, fishnet weights, or supports for spindle to facilitate the unwinding of the spun thread onto a spool or into a ball, might be confused with loom weights (Wace and Thompson 1912, 43, fig. 19; Mingazzini 1974, 209–11, 215; Barber 1991, 97, n. 11; Buchner and Rittmann 1948, 40, fig. 9 in Gleba 2008, 127). Loom weights could also be (re)used for some of these other functions. In some cases, the analysis of use wear marks can help to distinguish these