INTRODUCTION
In Etruria, as in most ancient societies, textiles were necessary for a wide range of purposes, including clothing (Bonfante, Reference Bonfante2003), armour (Gleba, Reference Gleba2012), sails, furnishings — coverings for floors and furniture, wall hangings and awnings — as well as bags, sacks, wrappings and even books (van der Meer, Reference van der Meer2007). Despite their great economic and social importance, textile production and consumption have only recently entered into the discourse on Etruscan crafts, technology, economy and the socio-religious sphere (see Gleba, Reference Gleba2008; Pitzalis, Reference Pitzalis2011; Lipkin, Reference Lipkin2012).Footnote 2 This is partially due to a limited number of extant Etruscan textile fragments, which are constantly increasing (Gleba, Reference Gleba2017a, Reference Gleba2017b). A more important lacuna has been the lack of systematic studies on textile tools, although items such as spindle whorls and loom weights are ubiquitous finds and constitute the single most important and plentiful type of evidence for understanding the technology and assessing the scale of textile manufacturing in Etruria (Gleba, Reference Gleba2008; Lipkin, Reference Lipkin2012).
The tools are found in a variety of contexts. Spinning and weaving tools found in tombs (Pitzalis, Reference Pitzalis2011; Lipkin, Reference Lipkin2012; Laurito, Reference Laurito2015) and votive deposits (Gleba, Reference Gleba, Herring and Lomas2009) indicate the social significance of textile production. Finds from settlement areas, on the other hand, are crucial because their context is directly related to their functional use. Although virtually no complete ancient looms from Etruria survive at present, the position and size of a loom may be reconstructed where loom weights are found in a primary destruction context.Footnote 3 The scale of textile production may be estimated from the distribution of spindle whorls, loom weights and spools. Additionally, recent advances in textile experimental archaeology, pioneered at the Danish National Research Foundation's Centre for Textile Research at Copenhagen University (CTR), now make it possible to analyse spindle whorls and loom weights in order to gain insight into the types of thread and fabric they would have been suitable to make (Andersson Strand and Nosch, Reference Andersson Strand and Nosch2015).
Most of the information about Etruscan life comes from rich cemeteries, because almost all of the major settlements have been continuously occupied to this day, making them archaeologically inaccessible. One exception is Poggio Civitate, because of its accidental seventh-century BC destruction that was followed by a deliberate sixth-century BC destruction and the lack of subsequent occupation. Poggio Civitate is unique in preserving a ‘freeze-frame’ picture of an Etruscan settlement during several phases of occupation. This picture probably provides a more typical reflection of Etruscan life rather than a unique representation of its time. This is particularly important with regard to the production activities attested on the site, due to both the quantity and quality of data pertaining to such activities retrieved and recorded. Textile tools have been found during almost every season since the excavations began at Poggio Civitate in 1966. Judged solely by the number of spinning and weaving implements, Poggio Civitate appears to have been a significant textile-producing centre.Footnote 4 The site thus furnishes excellent material for a case study of Etruscan textile production, as the information that can be retrieved from textile implements found throughout the site can be compared with the technical data gained from the analysis of textile fragments found in the contemporary burials at the nearby Poggio Aguzzo.
POGGIO CIVITATE AND POGGIO AGUZZO DI MURLO
The site of Poggio Civitate is strategically situated roughly 20 km southwest of Siena in Tuscany near the Crevole river, a tributary of the Ombrone, which would have linked it with the major Etruscan coastal centres. Its location near the rich copper and iron deposits of the Colline Metallifere likely contributed to the site's emergence and prosperity. Two summits, known as Poggio Civitate and Poggio Aguzzo, form the hill. Monumental buildings are located on the plateau of Poggio Civitate, called locally Piano del Tesoro (Fig. 1).
Fig. 1. The view of Poggio Civitate (left summit) and Poggio Aguzzo (right summit) from the east. (Image: M. Gleba.)
Over 50 years of excavation and study have identified two distinct chronological phases at Poggio Civitate: the Orientalizing, spanning 675–600 BC, and Archaic, dated to 600–535 BC.Footnote 5 The Orientalizing phase is currently represented by three structures: the Orientalizing Complex 1 (OC1, previously known as the Lower Building) identified as the residence of a prominent local family; the Orientalizing Complex 2 (OC2, previously the Southeast Building), believed to have been a workshop structure for manufacturing activities including bronze casting, bone and antler carving, ceramic production, glass production, food processing and textile manufacture; and the recently excavated Orientalizing Complex 3 (OC3, previously known as the Tripartite Building), which may have been a building of religious nature. After the seventh century BC, the buildings were destroyed by a conflagration, and the site was levelled and a monumental four-winged structure was built. This square building (previously known as the Upper Building), measuring 61 m on each side, is one of the largest structures known during the Archaic period from the Mediterranean (Fig. 2). Various theories have been proposed for its function including a political meeting hall, a religious sanctuary, an aristocratic residence and even an Etruscan version of an agora (Tuck, Reference Tuck, Carpino and Bell2016: 110). The excavators currently believe that it combined the functions of the various Orientalizing buildings into a single edifice, dating to the early sixth century BC. The Archaic structures of the site were dismantled between 550 and 535 BC for reasons yet unknown.Footnote 6 In 2015, evidence of an earlier phase of monumental building immediately west of Piano del Tesoro came to light, consisting of the remains of a rectangular structure (EPOC4), which according to the excavators was built within the first quarter of the seventh century BC (Tuck et al., Reference Tuck, Kansa, Kreindler and O'Donoghue2017).
Fig. 2. Simplified plan of Poggio Civitate building with Orientalizing structures in black. (Drawing: V. Herring after Poggio Civitate plan.)
The nearby hill of Poggio Aguzzo is believed to be the necropolis of the settlement excavated at Poggio Civitate and is located north of the settlement site. In 1972, nine tombs were excavated, which represent only a portion of the cemetery as further excavations were not possible. All of the tombs date to the middle–late seventh century BC and thus cover only part of the period during which the settlement of Poggio Civitate was active (Tuck, Reference Tuck2009). In contrast to the well-known rich Etruscan Orientalizing tombs elsewhere (e.g. Riva, Reference Riva2010), these are rather modest fossa burials with locally produced grave goods consisting of eating and drinking pottery, iron weapons and tools and bronze objects, with one imported Greek aryballos (Tuck, Reference Tuck2009).
FUNCTIONAL TEXTILE TOOL ANALYSIS
Specific tools were associated with various stages of textile manufacture: fibre processing, spinning/splicing yarn, weaving fabric, and secondary processes such as sewing fabric together to create garments. Not all stages of the textile chaîne opératoire are documented in the archaeological record since many tools, especially those used in the preparation of fibres, could be made of perishable materials such as wood or bone, and have not survived time and the elements.Footnote 7 Surviving implements are mostly associated with spinning, weaving and finishing. These include fired clay spindle whorls, loom weights and spools and metal tools, such as needles and, sometimes, distaffs and spindles. Functional analysis of textile tools, which prioritizes their physical attributes insofar as they can affect tool function, can be used, in combination with other sources of evidence, to examine the nature and scale of textile production in particular sites and regions during particular periods.
A spindle whorl is positioned on a spindle shaft and acts as a flywheel, providing weight and tension for spinning fibres into thread (Ryder, Reference Ryder1968; Crewe, Reference Crewe1998: 5–8; Gleba, Reference Gleba2008: 103–9). The weight of a spindle whorl provides information on what types of thread one can make with the given whorl.Footnote 8 In general, lighter whorls are better suited to produce thinner thread, and heavier whorls are best suited to produce thicker thread. A very light whorl is unlikely to be used for spinning thick thread, but an experienced spinner can make fine thread using a heavy whorl, especially if the spindle is supported on the ground or in a bowl, rather than suspended (Grömer, Reference Grömer, Bichler, Grömer, Hofmann de Keijzer, Kernand and Reschreiter2005: 110–11). The diameter of a spindle whorl affects how tightly the thread is spun (Barber, Reference Barber1991: 53). Whorls with narrower diameter spin faster and are good for making more tightly twisted thread.Footnote 9 The tight twist makes the thread stronger, and able to withstand more tension.
Skill, personal habits and the properties of the fibre also play an important role in shaping the thickness and tensile strength of the resulting thread, but these factors are difficult to incorporate in the calculations at present (Andersson Strand and Nosch, Reference Andersson Strand and Nosch2015: 46). Notwithstanding these variables, the experimental work published to date allows us to link spindle whorls of a certain weight with a range of yarn thickness and output per hour (Mårtensson et al., Reference Mårtensson, Nosch and Andersson Strand2009: 378, 393). The calculations below are based on experiments performed by two spinners at CTR, spinning wool and linen with 4 g, 8 g, 18 g and 44 g whorls. This corresponds to the range of spindle whorl weights found at Poggio Civitate, and to the predominant raw materials attested across pre-Roman Italy (Gleba, Reference Gleba2017a, Reference Gleba2017b), although we cannot be sure about the exact varieties of sheep wool used in antiquity (Gleba, Reference Gleba2012).
The presence of loom weights typically indicates weaving on a warp-weighted loom, an upright type of loom, which was used in Italy at least from the Middle Neolithic period onwards (Gleba, Reference Gleba2008: 122). The loom weights’ function is to apply tension to the vertically hanging warp threads. If there is not enough tension, it is very hard to change the shed, which creates problems during weaving. If there is too much tension, it is more difficult to insert weft threads, and in extreme cases the warp is likely to wear out and break. Different types of thread need different amounts of tension on the loom (Olofsson et al., Reference Olofsson, Andersson Strand, Nosh, Andersson Strand and Nosch2015: 88). In general, a thin thread needs less tension than a thick thread. This is important, because even a small difference in the thread diameter makes a big impact on the appearance of the textile, as can be seen in fabrics which have approximately the same number of warp and weft threads but are made with yarns of different thicknesses. The type of fibre (e.g. flax, different varieties of wool), the way it is prepared and incorporated into a yarn (e.g. in the case of flax whether it is draft-spun or spliced),Footnote 10 also affect how much tension the yarn needs. However, at this stage these variables in the calculations cannot be accounted for, because all of the experiments on which this method is based have been conducted with draft-spun wool of one sheep breed, draft-spun flax and machine-spun wool (Olofsson et al., Reference Olofsson, Andersson Strand, Nosh, Andersson Strand and Nosch2015: 78–9, 94).
The weight of a loom weight dictates how many threads needing a particular tension can be attached to it (Gleba, Reference Gleba2008: 134; Olofsson et al., Reference Olofsson, Andersson Strand, Nosh, Andersson Strand and Nosch2015: 87–97). For example, if the loom weight weighs 500 g, one can attach ten warp threads that require 50 g tension, or 25 warp threads that require 20 g tension. On the other hand, if the loom weight weighs 200 g, one can attach only four warp threads that require 50 g tension, or ten warp threads that require 20 g tension. The weavers consulted by CTR considered that attaching at least five and no more than 30 warp threads to one loom weight are the limits of what is practical (Mårtensson et al., Reference Mårtensson, Nosch and Andersson Strand2009: 389–90). However, weaver Lena Hammarlund reports that in her experience, weaving with four threads per loom weight and 50 threads per loom weight can also be practical, depending on the set-up and the yarn (Hammarlund, pers. comm., 2017). These limits determine which set-ups are evaluated as ‘possible’ or ‘unlikely’ (see Table 1).
Table 1. Loom weights from Poggio Civitate: tension and warp thread density distribution.
In the most basic weave structure, tabby or plain weave, the horizontal weft runs over one warp thread, and under the next. To weave a tabby textile on the warp-weighted loom, two rows of loom weights are used. To weave twills, where the weft thread passes over two warp threads in a regular sequence, two, three or four rows of loom weights are needed.
The functional tool analysis method developed by CTR assumes that in an optimal loom set-up, the weights hang close together and the width of the row of loom weights is the same as the width of the textile to be produced (Olofsson et al. Reference Olofsson, Andersson Strand, Nosh, Andersson Strand and Nosch2015: 92). The thickness of a loom weight, the number of threads attached to the weight and the number of rows of loom weights used control how closely the warp threads will be spaced in the finished fabric. A fabric can be balanced, with approximately the same number and type of warp and weft threads per unit of length. Alternatively, it can be either warp-dominant or weft-dominant, when one of the thread systems has a higher number of, and/or thicker, threads than the other thread system per unit of length. Since a loom weight's shape affects the nature of the finished textile, different types of loom weights are optimal for making different types of cloth. For example, light, thin loom weights would be best suited for producing dense fabrics made with fine thread, whereas heavy, thick loom weights would be optimal for weaving more open or weft-dominant textiles made with thicker thread.
TEXTILE TOOLS FROM POGGIO CIVITATE
Textile tools catalogued to date at Poggio Civitate include 581 spindle whorls, 99 loom weights, 875 spools, one bone and fourteen bronze needles (Fig. 3).Footnote 11 This is a very large assemblage compared with contemporary sites in Italy. Loom weights at other sites sometimes number in the hundreds (Gleba, Reference Gleba2008: 135), but it is rare to find more than several dozen spindle whorls and spools in a settlement excavation (see Gleba, Reference Gleba2008: 163–71 for an overview; Quercia Reference Quercia2017; Gleba et al. Reference Gleba, Harris, Cutler, Marín-Aguilera and Dimovaforthcoming). We cannot calibrate our comparison, although we know the number of tools at each site depends partly on its history of abandonment and reuse and the scale of excavation. Nonetheless the contrast remains striking.
Fig. 3. A selection of (a) spindle whorls, (b) loom weights and (c) spools from Poggio Civitate. (Images courtesy of Poggio Civitate project.)
SPINDLE WHORLS
The great number of spindle whorls at Poggio Civitate indicates that the scale of yarn production likely exceeded the requirements of even a large household. The whorls weigh between 2 g and 48 g, excluding one incomplete example with preserved weight of 58 g,Footnote 12 but most weigh less than 30 g. Two-thirds (over 300 whorls) are very light, weighing 10 g or less (Fig. 4), which makes them optimal for spinning thin thread. This thread would need less than 20 g tension on the loom. According to the experiments, whorls weighing under 4 g would be suitable for spinning wool thread needing c. 13 g tension per thread on the loom (Olofsson et al., Reference Olofsson, Andersson Strand, Nosh, Andersson Strand and Nosch2015: 85). One hundred and eleven of the whorls weigh between 11 and 20 g, and would be suitable for spinning slightly thicker thread, while a relatively small number of heavier whorls would be good for spinning thick thread, needing up to around 50 g tension on the loom. Therefore, while the spindle whorl assemblage indicates that a range of thread types could have been spun at the site, most of the whorls were best suited for making fine yarn. The overall weight range is similar to spindle whorls recorded in contemporaneous settlement and funerary contexts for which the information has been published (Gleba, Reference Gleba2008: 108; Laurito, Reference Laurito2015: 13–14) or recorded by the authors (Fig. 5); however, the mean weight is lower at Poggio Civitate (10 g) due to the prevalence of very light spindle whorls.
Fig. 4. Spindle whorl weight groups (N = 479).
Fig. 5. Spindle whorls from Poggio Civitate, compared to other settlements in Etruria and Latium (PROCON database; Ficana – Lipkin, Reference Lipkin2012: 31; Poggio Colla – Meyers, Reference Meyers and Banck-Burgess2015b: 254).
Almost 90 per cent of the whorls are biconical truncated in shape, suitable for spinning yarn of medium twist, with lower numbers of standard biconical, conical and other types (Fig. 6). The distribution plot of spindle whorl weights and diameters demonstrates that there is no clear link between shape and size of the whorls, i.e. whorls of the same shape come in different sizes and vice versa (Fig. 7). Biconical and conical spindle whorls also predominate at other contemporaneous Etruscan and Italic sites (Gleba, Reference Gleba2008: 107; Laurito, Reference Laurito2015: 15).
Fig. 6. Spindle whorl shapes.
Fig. 7. Spindle whorl diameter and weight by type (N = 433).
LOOM WEIGHTS
The assemblage from Poggio Civitate consists of 99 loom weights, which seems small in comparison with the hundreds of spindle whorls and spools. The loom weights are all truncated pyramidal in shape, and cover a wide weight range, from 70 g to 795 g, although most are under 500 g (Fig. 8). This range is broad and consistent with ranges at several other contemporary Etruscan sites (Gleba, Reference Gleba2008: 135), which suggests that textiles of various qualities were being produced at Poggio Civitate (Fig. 9).Footnote 13 The loom weights are suitable for use with warp threads needing a range of tensions, from 5 g to 75 g per thread (Table 1). Weaving with higher tension (>75 g) is possible using the Poggio Civitate loom weights, but this would yield textiles with only two to three warp threads per cm, which have not been archaeologically documented in Italy to date. Such set-ups are therefore unlikely. At the other end of the scale, weaving with very low tension (<10 g/thread) is a theoretical possibility that has not yet been sufficiently tested on the warp-weighted loom.
Fig. 8. Loom weight thickness and weight distribution (N = 60).
Fig. 9. Loom weights from Poggio Civitate, compared to other settlements in Etruria and Latium (published data and PROCON database; Monte Savino – Macellari, Reference Macellari and Vitali1983: 123–6; Forcello di Bagnolo di San Vito – Vay, Reference Vay and de Marinis1988: 171–2; Acquarossa – Wikander, Reference Wikander1986: 112, 125 n. 282; Accesa – Camporeale, Reference Camporeale1997: 199; Ficana – Lipkin, Reference Lipkin2012).
The loom weights could be used to make both tabbies and twills, depending on how the loom was arranged. Tabby can be woven by arranging the loom weights in two rows. Twill can also be woven with two rows of loom weights, as demonstrated by ethnographic examples from Scandinavia (Hoffman, Reference Hoffman1964). However, it is much easier to work with four rows: the weight is distributed better, which stops the warp threads from becoming loose or tangled when one changes the shed (Haynes, Reference Haynes1975). A three-row set-up is also possible, for weaving 2/1 twill, although very few 2/1 twills have been documented in Etruria to date. The loom weights’ productive possibilities are assessed in relation to these possible set-ups. Table 1 provides an overview of the Poggio Civitate assemblage.
The theoretical range of possibilities should also be compared to the parameters of archaeological textiles from Italy, in order to identify the most likely way in which the tools were used at Poggio Civitate. Archaeologically documented twills from the Italic peninsula typically have fifteen to 30 threads per cm, although in rare cases the thread count can be as low as five or as high as 40 threads per cm (Gleba, Reference Gleba2017b: 20, 22, fig. 7). Tabbies have four to 30 threads per cm, but thread counts below ten are rare in archaeological textiles from Etruria (Gleba, Reference Gleba2017b: 16–20).
The loom weights from Poggio Civitate could have been used for making cloth with low thread counts (Table 1), but the textile data suggest that this is not very likely. In other words, ancient weavers probably used only part of what we calculate as the tools’ range of possibilities. The loom weights from Poggio Civitate could be used to weave twills with typical archaeologically attested thread counts in a variety of yarns. For example, the textiles from Poggio Aguzzo, discussed below, have an average quality of eighteen to twenty warp threads per cm, and could have been made using the loom weights from Poggio Civitate (Table 1). The heaviest loom weight, measuring 8.4 cm in thickness and weighing 795 g (Artifact ID PC19720317), can be used to make textiles with such thread counts using yarn requiring c. 15–20 g tension per thread in twill, with four rows of loom weights. With two rows of loom weights, one would need to attach 84 threads per loom weight, which is far from the limits of what is practical. Therefore, the production of tabbies with this loom weight is unlikely.
Another example, the loom weight with dimensions closest to the average (Artifact ID PC 20040172), can be used to make such textiles with yarn that requires less tension on the loom (5–11 g per thread). It could theoretically be used to make tabbies, working with two rows of loom weights, but with very low tension (5 g). The lightest loom weight from Poggio Civitate, Artifact ID PC 19690354, would provide very low thread tension in the same set-up (<5 g per thread). As mentioned above, experiments with such low tension have not been conducted. At the current state of our knowledge, it is likely that this loom weight was used for weaving textiles with very few warp threads per cm.
As Table 2 shows, the tools from Poggio Civitate are better suited for weaving twills rather than tabbies.Footnote 14
Table 2. Productive possibilities of the Poggio Civitate loom weights: assessment of set-ups for weaving twill and tabby with 20 warp threads/cm (th is thread(s), lw is loom weight). Low tension: theoretically possible set-up but has not been tested experimentally.
SPOOLS
Fewer than 200 of the spools are intact; however, their size range is representative of the overall assemblage. They are all light, with more than two-thirds weighing 50 g or less (Fig. 10). Spools from other Etruscan settlement and funerary contexts generally do not exceed 100 g and mostly weigh under 50 g (Fig. 11). Their function has been debated (see Gleba, Reference Gleba2008: 140–3). The larger spools could theoretically be used on the loom, as experiments have shown (Olofsson et al., Reference Olofsson, Andersson Strand, Nosh, Andersson Strand and Nosch2015: 92–7), and considering that small loom weights of about 40 g are known at Incoronata (Meo, Reference Meo, Gleba, Busana, Meo and Tricomi2018) and Torre di Satriano (Quercia, Reference Quercia2017) in southern Italy, Selinunte in Sicily (Quercia, Reference Quercia2017), as well as in Greece (Spantidaki, Reference Spantidaki2016: 180–213). However, the spools from Poggio Civitate would work better as weights for tablet weaving: this technique involves passing threads through holes in the corners of (usually) square tablets, which, when rotated forward or back, force the threads to form different sheds (Collingwood, Reference Collingwood1996). By rotating cards in different combinations, it is possible to achieve numerous and complex patterns. This method is suitable for weaving narrow bands, such as belts, heading bands for the warp of a warp-weighted loom, or decorative borders for the ground textile weave. Tablet weaving in Etruria is well attested by the presence of such borders on textiles (Gleba, Reference Gleba2017a, Reference Gleba2017b), as well as by the finds of tablets, metal clasps and bone spacers with pegs (Gleba, Reference Gleba2008: 139, 150–3).
Fig. 10. Spool thickness and weight distribution (N = 178).
Fig. 11. Spools from Poggio Civitate and other sites in Etruria: weight range (PROCON database).
NEEDLES
Finally, one bone and fourteen bronze needles with a single eye have been found at Poggio Civitate. The complete bronze items range in length between 5.1 and 8.5 cm, while the single bone needle is thicker and measures 4.5 cm in length. Not only do they indicate that sewing was practised on the site, they also suggest that several workers could have been sewing simultaneously, because in normal domestic contexts needles are quite rare. Although it is considered that Etruscan and ancient Mediterranean garments were generally made of textiles taken off the loom and shaped on the body with belts and pins, thus requiring minimal sewing, burial evidence suggests that aristocratic garments were frequently decorated with hundreds and even thousands of buttons, appliqués and other types of decorations, which would have required the use of a needle (Gleba, Reference Gleba2017b).
ORGANIZATION OF PRODUCTION
Mapping the distribution of the tools shows that each category clusters in certain locations in the north, south and western parts of the site. The precise dating of many textile tools is problematic, because the depth of many objects recovered in the early excavations was not recorded, and material from the two phases was often mixed. Therefore, the majority of the tools broadly date to a c. 100-year period between the mid-seventh and mid-sixth century BC. However, the find spots, in some cases, suggest a more specific Orientalizing or Archaic date (Fig. 12). Furthermore, about a third of the spools and some spindle whorls bear stamped, incised or impressed decoration, which allows stylistic attribution to the Orientalizing period.
Fig. 12. Heat map of (a) spindle whorl and (b) loom weight distribution at Poggio Civitate, showing concentrations in specific areas. (Image: Adam Green.)
Clusters of loom weights, spindle whorls and spools were identified near OC2/Workshop. This location is well outside the area of the Archaic building, and thus the tools almost certainly derive from the multi-craft workshop of the Orientalizing period. This indicates that textile production was practised alongside the other manufacturing activities that took place within OC2, documented by the deposits of the debris of ivory, glass and bronze working, as well as the presence of unfired tiles (Nielsen, Reference Nielsen1998: 98–9).
Thirteen loom weights were recovered from OC2. These items have a more restricted weight range than the overall assemblage of loom weights from the site, weighing between 120 and 300 g, suggesting that there may have been a focus on the production of certain types of textiles in this building. They would be suitable for use with thin to medium thread types, needing no more than c. 30 g tension on the loom. If they were used to make tabby textiles, the fabrics would have been quite open if balanced, or weft-dominant. The spindle whorls from OC2 also lie at the lighter end of the spindle whorl weight range, all being under 30 g, and most weighing under 20 g.
Another cluster of spindle whorls appears to be associated with OC1/Residence (Tesoro Rectangles 5–9), which yielded 123 spindle whorls. Of these, 80 are noted in trench books as coming from Stratum D, the layer associated with the Orientalizing phase. All but two of these spindle whorls are truncated conical, and, of 105 weighed objects, 99 weigh under 18 g, with the average weight of the group being 6.5 g, i.e. significantly lower than the overall average of 10 g for the site as a whole. Such a large concentration of rather uniform spinning tools in the northern part of OC1 suggests that they may have been kept there in storage, like the large banquet service recovered from the same building (Tuck, Reference Tuck, Gleba, Marín-Aguilera and Dimovaforthcoming).
Although not all parts of the site have been excavated, overall the numbers of spindle whorls with respect to loom weights suggests that a large amount of yarn was being spun at the site. The uniform proportions and great quantity of the spindle whorls suggest that they were designed to equip a number of spinners making standardized yarn. The small size of the whorls points to the production of fine thread, which requires considerable skill and concentration (Grömer, Reference Grömer, Bichler, Grömer, Hofmann de Keijzer, Kernand and Reschreiter2005: 110–11). The recent portable X-ray fluorescence (pXRF) analysis of the archaeological ceramic finds and extant clay sources at the site (Bauer et al., Reference Bauer, Tuck, Duncan and Moore2017) indicates that this vast quantity of standardized tools was produced at the site. The concentration of large quantities of very uniform implements in OC2 suggests that this area was a workshop and that spinning was likely one of the major crafts at Poggio Civitate in terms of occupied labour.
In contrast, the number of loom weights at the site is relatively low, especially considering that a single loom would require many loom weights. It is possible that weaving was carried out in other areas of the site, not yet excavated, or that the loom weights were recovered after the destruction of the site. This inconsistency may also be explained by the extraordinary quantity of spools on the site. The spools (over 875) are even more numerous than the spindle whorls and, if the hypothesis that they were used in tablet weaving and other band weaving is correct (Ræder Knudsen, Reference Ræder Knudsen, Gleba and Mannering2012), it may indicate that the weaving workshop at Poggio Civitate specialized in the production of tablet-woven textiles. These textiles could have been used as headbands, belts or to decorate garments and other textiles. This interpretation is suggested by the popularity of patterned borders on garments depicted in Etruscan figurative representations (Bonfante, Reference Bonfante2003) and examples of such borders found at Poggio Aguzzo (examined below).
The great quantity of standardized tools found together suggests that many of these items were used simultaneously, and hence raises the question of why a workshop was set up at Poggio Civitate and how it operated.
Deducing the scale of production from archaeological data is an exercise fraught with complexities, nuances and unknowable variables (see Costin, Reference Costin1991, Reference Costin and Maschner2005). It would be unwise to assume a straightforward link between the number of tools found at a site and a certain mode of production, without taking into account spatial distribution, social and political organization. In any case, the criteria one uses must be defined.
In the case of Poggio Civitate, we can address two criteria archaeologically: the number of people involved and the spatial concentration. Hence, by workshop we mean a production unit which is concentrated in one area or building and which involves more people than a family or household. Starting with the number of spindle whorls (581) we can estimate in broad terms how many spinners could work simultaneously. In pre-Roman Italian graves, women are often buried with one or several whorls, making up a set for spinning and plying different yarns (Gleba, Reference Gleba2008: 175–8). The size of these sets varies widely and may not reflect what the deceased used in life in every case. Even a conservative ratio of five tools for each spinner at Poggio Civitate would amount to 113 workers. If even half of them — 57 spinners — worked simultaneously, this would have been a very large workshop. Whether it operated on a full-time or seasonal basis, whether the same people did the spinning and tablet weaving, what their circumstances were and how they related to other activities at Poggio Civitate remain open questions.
TEXTILES FROM POGGIO AGUZZO
More tangible are the answers to the questions regarding what the textiles produced at the site may have looked like, what material they were made of and, in some cases, how they were used. Fragments of mineralized textiles have been identified on three objects from two tombs dating to the seventh century BC (Fig. 13), located at the nearby site of Poggio Aguzzo (Tuck, Reference Tuck2009; Gleba, Reference Gleba, Carpino and Bell2015b), and hence contemporary with the majority of textile tools at Poggio Civitate. A summary of the technical textile data is provided in Table 3.
Fig. 13. Micrographs of mineralized textiles from Poggio Aguzzo at 50×: (a) twill from Tomb 1, Inv. 72-04; (b) tablet weave from Tomb 1, Inv. 72-05; (c) twill from Tomb 1, Inv. 72-05; (d) tabby from Tomb 4, Inv. 72-13. Inv. is inventory number. (Images: M. Gleba.)
Table 3. Technical data of textiles from Poggio Aguzzo (thread counts are in threads per cm; * tablets per cm; diameters measured in mm).
One of these textile fragments is preserved on an iron spear counterweight from Tomb 1, and is a 2/2 balanced twill made of wool as indicated by the scanning electron microscopy (SEM), with approximately twenty threads per centimetre (Fig. 13a). The threads are single z- or clockwise-twisted, around 0.4 mm in diameter. This quality of yarn corresponds well with the mean weight of the spindle whorls at Poggio Civitate, as light whorls (4–25 g) are optimal for spinning yarn with diameters 0.3–0.6 mm (Grömer, Reference Grömer, Bichler, Grömer, Hofmann de Keijzer, Kernand and Reschreiter2005: 110–11; Möller-Wiering, Reference Möller-Wiering, Andersson Strand and Nosch2015: 107). Some of the loom weights recovered from the settlement would be suitable for making twills with a similar thread count, as outlined above. The twill has a very matted surface, obscuring the weave structure, suggesting that it was either quite worn or intentionally fulled.
The second textile fragment, also from Tomb 1, is a wool tablet weave, preserving at least seventeen tablets alternating the direction of the tablet rotation in a pattern 3Z3S (Fig. 13b), and appears to be connected to the 2/2 twill (Fig. 13c), i.e. it was probably a tablet-woven border. Tablet borders are usually found on outer wear garments of special and/or ceremonial nature such as mantles (Stauffer, Reference Stauffer, Gleba and Mannering2012; Gleba, Reference Gleba2017b). This item fits well with the presence of large numbers of spools at the settlement, which indicates that tablet weaves were also being produced at the site.
The third textile fragment, preserved on an iron knife from Tomb 4, is a dense balanced tabby weave, with approximately twenty threads per centimetre (Fig. 13d). The threads are approximately 0.4–0.5 mm in diameter. While some of the spindle whorls would have been suitable for spinning thread of this quality, none of the loom weights recovered at the site so far would have been suitable for making balanced tabbies with so many threads per centimetre. Tabby textiles produced with the extant weights would have been open or weft-faced. Much thinner loom weights would be better suited for producing dense, balanced tabbies. So if the fragment from Tomb 4 was made on the warp-weighted loom, it may have been brought to the site from elsewhere.
In addition to the technical data, SEM of samples from two of the textiles provides information about the quality of wool and its preparation. The former is determined on the basis of fibre diameter measurements (Gleba, Reference Gleba2012), while the latter is indicated by the scale direction of individual fibres within a thread. The wool in the Poggio Aguzzo textiles from Tomb 1 has a rather wide distribution curve, generally skewed to fine, with maximum diameter not exceeding 50 µm and with mean diameter and mode above 20 µm (Fig. 14). This type of wool developed in Europe at least by the seventh century BC. Similar distributions are found in wool samples from other contemporaneous and later Italian and European sites (Rast-Eicher, Reference Rast-Eicher2008; Gleba, Reference Gleba2012; Rast-Eicher and Bender Jørgensen, Reference Rast-Eicher and Bender Jørgensen2013).
Fig. 14. Histogram of wool fibre diameter distribution in textile from Poggio Aguzzo Tomb 1, Inv. 72-04.
Microscopy also allows assessing whether wool was processed before spinning. The combination within the same thread of fibres with scales pointing in opposing directions indicates that the wool used for making the thread was combed (Fig. 15). Absence of coarse hair and kemp also suggests that the wool was processed. Combing may realign the fibres from their original position on a staple and remove very fine and very coarse fibres.
Fig. 15. Scanning electron micrograph of negative fibre casts of textile from Tomb 1 (Inv. 72-74), showing scales pointing in opposite directions. (Image: M. Gleba.)
CONCLUSIONS
The data provided by the textile tools indicate that textiles were manufactured at Poggio Civitate on a large scale, in a centralized manner. The great quantity of standardized, light spindle whorls clustering in specific areas of the site suggests that at least several dozen workers were employed in a spinning workshop, oriented towards making very fine yarn. Tablet weaving was likely another major craft activity at the site, as indicated by the presence of over 870 spools. While the smaller number of loom weights may have served for weaving textiles for use on site, the scale of spinning and tablet weaving is indicative of a larger-scale production.
Who was this production for? Textiles, being highly portable and valuable, are excellent candidates for exchange goods and it may be easy to assume that a workshop was oriented towards long-distance exchange. However, other lines of evidence suggest that we should challenge this initial assumption and think more carefully about the scale of local demand. Imports constitute only a very small fraction of the material found at Poggio Civitate; conversely, exceptionally few items manufactured at the site have been found outside Poggio Civitate or Poggio Aguzzo (Tuck, Reference Tuck, Carpino and Bell2016: 108). Despite the monumental size of OC2/Workshop (the building was at least 52 m long), and the variety of materials produced there (pottery, metalwork, bone and antler carving, glass manufacture, textile production and food processing), it may have catered primarily to internal consumption. Raw materials such as bone and horn, metal ores and clay, as revealed by the pXRF analysis, were sourced locally. A recent pilot study of sheep husbandry and mobility using stable and strontium isotopes to analyse sheep teeth from Poggio Civitate showed that all the animals were kept in the same geological environment, which corresponds to the local signature (Trentacoste et al., Reference Trentacoste, Lightfoot, Le Roux, Buckley, Esposito and Gleba2019). The latest interpretation of the site by the excavators is that OC2/Workshop was likely engaged in satisfying the needs of the community in the immediate vicinity under the aristocratic control of the site, which was largely locally oriented (Tuck, Reference Tuck2014, Reference Tuck, Carpino and Bell2016). The elite of Poggio Civitate controlled the manufacturing processes, some of which were seasonal, such as food production, while other activities were episodic (e.g. communal banqueting).
It is difficult to say whether the textile evidence conforms to this picture or challenges it, for two reasons. First, we cannot identify textile exports from Poggio Civitate because in contrast to ceramic, terracotta and metal finds, Etruscan textile evidence to date does not provide the resolution that would allow for assigning provenance of individual fabrics to a specific site. All of the fabrics recovered at Poggio Aguzzo and their technical characteristics are common in Etruria and, more widely, in pre-Roman Italy (Gleba, Reference Gleba2017a, Reference Gleba2017b). Nevertheless, considering the internally oriented economy of the site, it is likely that the majority of textiles produced at Poggio Civitate were intended for local consumption.
Second, in order to understand whether the workshop produced predominantly for local demand or for exchange, we need to assess the ‘internal’ or ‘local’ needs of Poggio Civitate, and that is not a straightforward task since demographic estimates are not yet possible. In addition to utilitarian purposes, textiles would have been made as a means of generating wealth that could be stored and exchanged in various contexts, sometimes at a later date (Harris, Reference Harris2018). Textiles would have been needed also for occasions of conspicuous consumption, such as banqueting. At Poggio Civitate, banqueting has been documented by a large dining service recovered from OC1/Residence, which would have accommodated over 100 individuals (Tuck, Reference Tuck, Gleba, Marín-Aguilera and Dimovaforthcoming). High concentrations of cut and butchered animal bone clustered within and around OC2 and non-elite residential areas point to a robust practice of slaughter and food preparation (Kansa and MacKinnon, Reference Kansa and MacKinnon2014). Textiles played a central role in such communal commensal contexts in the form of garments, tents, furnishings etc., as Etruscan iconography of banqueting demonstrates. In fact, one of the Archaic architectural terracotta plaques from Poggio Civitate illustrates a banquet (Small, Reference Small1971), and the couches are clearly draped with large cloth pieces (Fig. 16). The large concentration of spindle whorls in OC1/Residence may have been intended for an episodic textile production occasioned by a large banquet.
Fig. 16. Banquet terracotta plaque from Poggio Civitate. (Drawing: V. Herring after Small, Reference Small1971.)
Yet while production strategies at Poggio Civitate were mainly inward-oriented during the seventh and sixth centuries BC (Tuck, Reference Tuck, Carpino and Bell2016), the textiles recovered at Poggio Aguzzo indicate that even smaller, inland, Etruscan centres consumed textiles similar to those worn by the elite of the larger Etruscan cities of Cerveteri, Vulci or Tarquinia. The high specialization and large-scale production of textiles at the Poggio Civitate workshop proves the role that textiles played in the development and reinforcement of a community among the elites even outside the urban centres (Riva, Reference Riva2010). The unique window into the past that was inadvertently created by the several destruction events and the abandonment of the settlement at Poggio Civitate provides us with a glimpse of the economic and social investment required for textile production by analysing tools for textiles.
Acknowledgements
The research leading to these results has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP/2007-2013-312603) and was carried out within the scope of the project Production and Consumption: Textile Economy and Urbanisation in Mediterranean Europe 1000–500 BCE (PROCON). Our thanks go to Anthony Tuck, the current director of the Poggio Civitate excavation project, for access to the material and his continuous support of our research. We also thank Beatriz Marín-Aguilera and Jean MacIntosh Turfa for reading and commenting on the early draft, and the anonymous PBSR reviewers whose thoughtful critique helped to improve the paper. Any remaining errors are our own.
