Approaching archaeological evidence for workshops and production in ancient Greece

Workshops are the physical space for manufacturing activities. In the Greek world, workshops were defined with the word ἐργαστήριον or ἐργαστηρίδιον, denoting the physical space where work (ἔργον) was carried out. The term had broad meaning and was used to describe a range of workshops (e.g. carpenter’s workshop, coppersmith, silversmith, dyer’s house, weaving workshops, pottery workshop, sculptor’s workshop) as well as shops (Martelli Reference Martelli2011). Archaeologists often expand the concept of a workshop to include specific technological processes and standardized specialized productions (Costin Reference Costin, Hodgkinson and Tvetmarken2020).

Identifying workshops in the archaeological record can be a challenging task, especially due to the occurrence of multiple craft activities in the same location, the variability of possible production processes, the perishability of certain materials, excavation techniques, site-formation processes and the difficulty of identifying and interpreting tools and facilities (Costin Reference Costin, Hodgkinson and Tvetmarken2020, 179). In addition, the material remains of production activities vary depending on the specific requirements of each production process, increasing the challenge of identifying and understanding the spatial and temporal organization of production. Certain crafts, such as pottery production for example, require permanent installations and leave noticeable evidence (kilns, waste, standardization of production demonstrated through typology and/or materials analysis), while others, such as textile production, may leave only limited material traces (e.g. loom weights), or no remains at all. The analysis of ancient crafts and manufacturing activities has increasingly adopted contextual and holistic approaches to the study of technology, and scholars recognize the need to approach questions concerning the organization of production through interdisciplinary methodologies (see, for example, the recent publications by Hodgkinson & Tvetmarken Reference Hodgkinson and Tvetmarken2020; Partida & Graml Reference Partida and Graml2025).

Since individual workshops can be the locations for multicraft production, it is important to understand the material choices that brought the different activities together and to explore the ways these may have interacted. Based on both archaeological material and ethnographic evidence, several studies demonstrate that a narrow understanding of the constituent elements of workshops risks impeding our understanding of their nature and their associated activities (some examples include Brysbaert Reference Brysbaert2011; Hodgkinson & Tvetmarken Reference Hodgkinson and Tvetmarken2020; Rebay-Salisbury et al. Reference Rebay-Salisbury, Brysbaert and Foxhall2014). These studies highlight the need to move beyond the question of the presence or absences of single categories of artefacts. To complicate the issue further, recent analysis of archaeological evidence of pottery production demonstrates that a standardized outcome of a production process could be organized in numerous ways and under different workshop configurations (Murphy Reference Murphy, Verboven and Laes2017). In such cases, careful analysis of the spatial distribution of archaeological finds can help reconstruct the relationship between different craft activities and raw materials (on this point, see Murphy & Poblome Reference Murphy and Poblome2021). Narrow definitions of workshops may privilege certain types of production and periods over others and may overlook ‘the power and vitality of alternative modes of production organisation’ (Costin Reference Costin, Hodgkinson and Tvetmarken2020, 181). Moreover, when written sources about the organization of production exist, such as Vitruvius’ well-known description of the production of Egyptian blue in Vestorius’ workshop (De arch. 7.11.1), they might be wrongly perceived as the one and only production process, while the archaeological evidence, archaeometric analysis and experimental reconstructions may point to alternative production methods (Jackson & Smedley Reference Jackson, Smedley, Martinón-Torres and Rehren2008; Martinón-Torres Reference Martinón-Torres, Martinón-Torres and Rehren2008, 27).

When it comes to colourants (pigments and dyes), our current understanding of the materials and techniques used in antiquity is weighted towards the materials less prone to deterioration (Brøns Reference Brøns, Thavapalan and Warburton2019) and is highly influenced by the interpretation of the rich textual sources (especially Theophr., Lap.; Vitr., De arch.; Plin., HN). To understand the possible remains of pigment production, it is important to consider the nature of the technology, the chaînes opératoires and the variability of the production, in order to identify specific indicators for production spaces and processes (Costin Reference Costin, Hodgkinson and Tvetmarken2020, 186). Since pigments can derive from various source materials and production techniques, we argue that specific criteria should be defined for each pigment or pigment category by comparing the archaeological evidence with written sources, while also taking into account post-depositional degradation and alteration.

In our attempt to characterize craft activities performed in the workshop in the Koan agora through the archaeological finds, we found the systematic work of Iphigenia Tournavitou (Reference Tournavitou, French and Wardle1988) to provide a solid classificatory foundation upon which we could structure our analysis. According to this approach, the following criteria are evaluated to confirm the presence of a workshop: a) architecture; b) pottery; c) built-in facilities; d) tools; e) worked material, including raw materials, half-worked or unfinished pieces, waste, and finished objects; and f) the location of the workshop in relation to the administrative centre and resources.

However, as Ann Brysbaert points out, this model presupposes a single craft was carried out in the workshop space, without recognizing the possibility of multicrafting (Brysbaert Reference Brysbaert, Rebay-Salisbury, Brysbaert and Foxhall2014, 41). In a recent article on cross-craft relationships in late antique Sagalassos, Elizabeth A. Murphy and Jeroen Poblome assert that the modern conventions in archaeological practice of pragmatically dividing archaeological materials into clearly defined artefact-based categories ‘reflect and promote a system of parochialization that privileges objects, materials, and technologies over archaeological and social context. Disconnected from the full assemblage of materials, these artefacts are only subsequently interpreted as collective corpora derived from archaeological contexts’ (Murphy & Poblome Reference Murphy and Poblome2021, 101). This, in turn, they argue has reinforced single-industry analyses rather than acknowledging multiple crafts operating synchronously within the same spatial context. By drawing on, and combining, the already established theoretical concepts of chaîne opératoire and cross-craft interactions, Brysbaert has offered a useful theoretical framework for thinking about the interconnectedness of various craft technologies and social agents (Brysbaert Reference Brysbaert2011, esp. 2–5; Reference Brysbaert, Rebay-Salisbury, Brysbaert and Foxhall2014). Combining these two concepts allows us to study what she refers to as ‘horizontal’ or multiple chaînes opératoires in the production process and how cross-craft interactions might take place through the sharing of spaces, raw materials, tools and craft expertise and knowledge.

Such cases of co-existence of multiple crafts within a single workshop are increasingly being recognized in the archaeological record and have come to be referred to as ‘multicrafting’ and ‘multicraft production’ (Shimada Reference Shimada2007). We therefore employ the term multicrafting as defined by Shimada (Reference Shimada2007, 5): ‘the same or closely related artisans performing a major part or the entirety of two or more distinct crafts involving different media or raw materials in close proximity to each other to produce items primarily for suprahousehold consumption’. Multicraft production, on the other hand, describes the practice of different individuals that are specialized in one or more crafts and practise their specialized activities in the same space (Shimada Reference Shimada2007, 5).

Evidence confirming that the same people carried out distinct crafts (multicrafting) as opposed to distinct individuals or groups performing different crafts in the same space (multicraft production) may not be easily traceable in the material remains. However, we believe these concepts can help us think more carefully about the relationship between two crafts. Below we argue that the archaeological record of the Koan workshop provides an instructive case of multicrafting or multicraft production, where two or more crafts were carried out in conjunction within a shared workshop space. Starting from the presentation of the material evidence for pigment manufacture and metallurgical activity in the workshop, we trace the work processes that must have taken place within this space. In this analysis, we pay particularly close attention to the material aspects of cross-craft interaction within the workshop space to illustrate areas of overlapping knowledge and skills. Since cross-craft interactions were, however, by no means confined to the workshop space but extended to other areas of the socio-cultural systems, we extend the analysis to include the larger network of workshops in the Koan agora.

The Koan workshop: primary excavation evidence for pigment production and lead metallurgy

Excavations carried out by Charis Kantzia of the 22nd Ephorate of Prehistoric and Classical Antiquities in the modern city of Kos from 1984 to 1989 brought to light a pigment workshop where earth pigments and Egyptian blue pellets were found among fire-structures and remains related to lead metallurgy (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991).Footnote ¹ The workshop is located in the stoa complex that lines the eastern border of the agora immediately north of an E-W passage that provided access to the ancient agora square (Fig. 1). To the East, the workshop is delimited by the central N-S running commercial street that leads to the ancient port, while the full extent of the workshop to the north inside the stoa complex has not been revealed (Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019).

Figure 1.

Topographical plan of the eastern part of the excavations in the Damtsa plot. (1) workshop area; (2) east-west passage; (3) commercial street. (Modified after Kantzia Reference Kantzia1992, 636.)

According to C. Kantzia, the workshop dates to the late Hellenistic period, when it was established on top of a destroyed floor of square terracotta tiles (0.42 m in length), which belonged to an earlier Hellenistic phase of the agora (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991). Recent ¹⁴C dating confirms this hypothesis and places the workshop within the second century and the first half of the first century bce (Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019). The exact character and function of the earlier phase of the room remains uncertain, but Kantzia interpreted the presence of metallurgical slag as an indication of earlier manufacturing activities at the site (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 217).

A precariously built east–west wall is fundamental for the organization of the workshop space, which extended on either side of this wall (Fig. 2.1). Kantzia described the wall as ‘intensively burnt’ on both faces, indicating that pyrotechnological processes were carried out in its vicinity (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 218). The east–west wall features four V-shaped apertures, which possibly functioned as ventilation for fire-structuresFootnote ² (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 222–3; Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019, 178). A circular wellhead was installed at the eastern end of the east–west wall (Fig. 2.2). According to Kantzia, this is a later construction that disturbed the already abandoned workshop and destroyed at least one fire-structure (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 216–17). Excavations carried out from 2009 to 2011 by E. Papanikolaou within the framework of restoration activities at the agora by the Ephorate of Antiquities of Dodecanese revealed traces of a second wall to the north of and running parallel to the east–west wall described by Kantzia (Fig. 2.3). To the north of this wall, Papanikolaou found traces of another poorly preserved fire-structure (pers. comm. 2017; cf. Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019, 179). This later discovery indicates that the workshop space continued further to the north, following a compartmentalized layout with firing structures arranged along east–west traversing walls (Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2021, 222–3). If we allow for the same amount of space in the unexcavated northern part (i.e. north of the northern wall) as the area between the east–west wall and the southern boundary wall of the workshop, the minimum dimensions of the workshop would have been approximately 12×9 m, i.e. 108 m².

Figure 2.

1Topographical plan of the workshop. (1) east–west wall; (2) well head; (3) northern wall; (4) eastern fire-structure; (5) western fire-structure; (6) litharge rods; (7) lead lumps. (The plan is a composite of two plans: Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 214–15 and Kantzia Reference Kantzia1992, 636.)

Two fire-structures are associated with the southern east–west wall. The fire-structure in the eastern part that was destroyed by the later wellhead (Fig. 2.4) seems itself to have partially destroyed the earlier terracotta tile floor (Kantzia Reference Kantzia1992, 638). This fire-structure, measuring 0.53×0.48×0.11 m, was fashioned out of tiles containing hay and twig inclusions, and the associated debris covered an area more than 3 m in length (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 218). The second fire-structure (Fig. 2.5) measures 3×1 m and was identified 0.66 m to the south of the opening located in the western part of the east–west wall. There, 98 Egyptian blue rounded pellets with approximate dimensions ranging from 2 to 4 cm in diameter were found among large quantities of burnt pinecones and twigs. Ten more Egyptian blue pellets were found in other locations of the workshop and its surroundings. The shape of the pellets is considered a characteristic ingot shape of a late production, as opposed to the larger size of earlier Egyptian blue ingots, i.e. flat rounded or rectangular cakes, bowl-shaped cakes and sack-shaped pieces (Panagiotaki et al. Reference Panagiotaki, Tite, Maniatis, Kousoulis and Lazaridis2015; Weatherhead & Buckley Reference Weatherhead, Buckley and Kemp1989). The semi-circular stone mound is understood to mark the southern limit of the fire-structure. Among the remains of this fire-structure, which include vitrified wall-fragments, tiles, burnt pinecones and corroded copper objects, were also several pieces of a white substance, identified as quartz (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 227). A large yellow lump was characterized as a mixture of silica, goethite and beudantite (PbFe₃(AsO₄)(SO₄)(OH)₆), but no further details are provided in the publication. Layers of sand, charcoal, pumice, and dolomite lumps, as well as pigments of red, brown, green, yellow, violet purple and blue colour, were found in a 20–27 cm thick layer at the bottom of the fire-structure (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 235–6).

Lumps of red, yellow and brown earth were also found outside and among the remains of the fire-structures (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 223). The lumps are brittle, roughly spherical, and occasionally have a flattened surface (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 223–4). The publication states that chemical analysis revealed that the red lumps are iron-based (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 224), but the results of the analysis were not presented in the publication. Selected samples of earth pigments from the workshop were recently characterized, and the findings revealed the presence of hematite-based red earths, red lumps containing both hematite and lead tetroxide, goethite- and jarosite-based yellow earths, and green earth, characterized as celadonite (K(Mg,Fe^II)(Fe^III,Al)[Si₄O₁₀](OH)₂) with quartz and muscovite impurities (Kostomitsopoulou Marketou et al. Reference Kostomitsopoulou Marketou, Kouzeli and Facorellis2019).

A deposit of approximately 50 litharge (lead oxide, PbO) rods and numerous fragments of litharge were discovered to the south of the wellhead (Fig. 2.6). The rods measure up to 11.0 cm in length and 1.5–2.0 cm in diameter (Fig. 3). To the west of the litharge rods, large quantities of amorphous lead lumps and lead drops were found in an area covering 0.8×0.8 m. (Figs 2.7 & 4). The lead lumps were found at an opening in the east west wall, suggesting that the slag solidified as it flowed out of a fire-structure’s aperture. According to Kantzia, these finds indicate the in situ reduction of litharge to metallic lead (PbO to Pb) (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 221). She suggests that the litharge rods were imported for this purpose, while the lead produced could have been used to manufacture pigments. The excavations also recovered a cupel, similar to those found in the metallurgical workshops of Laurion (Conophagos Reference Conophagos1980, 315–16).Footnote ³ Even though Kantzia questioned the in situ silver extraction by cupellation (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 222), recent analysis of the surface of an Egyptian blue pellet from the western fire-structure demonstrated the presence of a corroded silver droplet (Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019). This may suggest the extraction of silver in the workshop; however, further research is needed to confirm the occurrence of silver metallurgy at the site.

Figure 3.

Two fragments of hollow, tubular litharge rods from the deposit south of the well head. Inv. no. ΑΕ 535. (Photograph: A. Marketou, courtesy of the Hellenic Ministry of Culture.)

Figure 4.

Amorphous lead lumps and lead drops from the area of the eastern fire-structure. Inv. no. ΑΕ 878. (Photograph: A. Marketou, courtesy of the Hellenic Ministry of Culture.)

To the east of the lead slag deposit, a finely ground white material was found, identified through chemical analysis as calcite (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 221). White lumps found in the area of the workshop were identified by chemical analysis as dolomite (CaMg(CO₃)₂) and quartz (SiO₂) (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 221). Kantzia also lists pumice, copper and iron slag (scoriae), copper and iron nails, a corroded copper sheet, tools and animal bones among the finds, all of which support the presence of manufacturing activities that exploited a range of materials (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991, 220–21).

To summarize, the excavation of one of the rear rooms of the agora’s eastern stoa complex, which is centrally located in the agora and near the area of the ancient city’s commercial quarter (Livadiotti Reference Livadiotti2018), uncovered at least two identifiable fire-structures (built-in installations). The excavations recovered light tools, a range of possible raw materials, waste, and finished products. The analytical resolution of the excavation documentation data, which originated from a rescue excavation carried out more than 30 years ago, makes interpreting the evidence challenging. The range of materials found at the site (litharge, amorphous metallic lead, various tonalities of Egyptian blue, earth pigments, metal scrap, pumice, etc.) indicates that at least two crafts were carried out in the same space: pigment production and lead metallurgy.

The presence of both successfully produced Egyptian blue pellets and pellets that can be characterized as waste products (unsuccessfully produced) has been considered evidence of local production (Kantzia & Kouzeli Reference Kantzia and Kouzeli1991; Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2022). However, based on a comparison of the Koan evidence with other possible production sites, in a recent article Cavassa (Reference Cavassa and Jockey2018) excludes the possibility of local production, predominantly due to the absence of firing vessels (saggars) among the archaeological finds. These differing interpretations led us to revisit the evidence of pigment production from Kos and to examine the ways material remains are interpreted and craft production is understood.

Defining the Koan workshop and pyrotechnology: space, architecture and location

The architectural plan and the basic construction of a building do not in themselves constitute definitive criteria for the identification of workshops (Tournavitou Reference Tournavitou, French and Wardle1988, 450). On the contrary, buildings can be modified and repurposed to house workshop spaces. On Kos, however, the workshop area was established on top of the tile-floor of a previous Hellenistic phase, beneath which the recovery of metal slag indicates that metallurgical activities were carried out before the establishment of the late Hellenistic workshop, indicating a tradition of craft production at the workshop’s location (Kantzia Reference Kantzia1992, 638; Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019).

Permanent, built-in installations can also be indicators of workshop space and are constructed when needed for specific production processes. Several types of pigments, including Egyptian blue and green, lead white, calcined red earth, verdigris and sandyx Footnote ⁴ are the products of pyrotechnological processes requiring built-in installations (fire-structures). Others, such as earth pigments, require (in principle) minimum treatment, such as grinding and mixing, and these are less likely to leave easily identifiable material traces of production.

For the first category, depending on the temperatures required for their manufacture, different fire-structures can be used, including open hearths, kilns or furnaces. The production of Egyptian blue, in particular, requires high-performance fire-structures that can maintain temperatures ranging from 850–1050°C for several hours (Pradell et al. Reference Pradell, Salvado, Hatton and Tite2006). Therefore, the identification of Egyptian blue workshop spaces should be supported by the presence of material remains suggesting the existence and operation of fire-structures, which is the case of the workshop area identified in the Koan agora. The second category requires relatively low temperatures (approx. 250°C for c. 45 minutes) for the pyrotechnological production process. For example, calcined red earth (ochra exusta: Plin., HN. 35.35) production involves the altering by heat of goethite (FeO[OH]) to hematite (Fe₂O₃) (Siddall Reference Siddall2018, 201), and because of the low temperatures needed, the remains of ochre firing may be impossible to identify archaeologically.

The Koan workshop also revealed ‘secondary’ architectural elements, namely the precariously built east–west walls that supported the fire-structures and divided the workshop space into compartments. The small openings in the wall can be interpreted as ventilation facilities, allowing air to circulate in the chamber of the fire-structure (see Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019, fig. 11). Pyrotechnological facilities can be hosted in either an open-air setting or inside a roofed building. Dioscorides’ description (De Materia Medica, 5.74) of the metallurgical process to produce pompholyx (zinc oxide or zinc copper alloy), which is produced from the smelting of cadmeia (impure zinc oxide), falls into the latter category. The organization of the workshop space described by Dioscorides, as reconstructed by Wilsdorf (Reference Wilsdorf1977, 17), closely resembles the Koan workshop, where the fire-structures are supported by similar walls, featuring small apertures. This compartmentalization of the workshop space also resembles the building identified as related to the production of Egyptian blue in Puteoli, found by the excavations of Giuseppe De Criscio at the end of the nineteenth century (see Cavassa Reference Cavassa and Jockey2018, fig. 4, 26, 27–8). The building is described as composed of a series of independent large rooms on the ground floor, with a second floor of similarly constructed rooms of smaller dimensions (Cavassa Reference Cavassa and Jockey2018, 28, citing De Criscio Reference De Criscio1881, 15–16).

The other pigments (red, brown, green and yellow), which can be grouped under the umbrella term ‘earths’, identified at the Koan location may have required levigation to increase the concentration of the chromophore and produce more saturated pigments (Siddall Reference Siddall2018, 201). Therefore, depending again on the scale of production, other types of built-in installations, such as cisterns or vats, may be required. Such levigation cisterns have not been identified in the excavated part of the Koan workshop. However, in his discussion of cinnabar procurement, Theophrastus (Lap. 58) writes that the delicate levigation process for the purification of cinnabar was carried out in copper vessels to allow for better control and thereby minimize the risk of losing the precious pigment. Theophrastus’ account, therefore, provides an alternative to such built-in cisterns.

Finally, an important criterion for a workshop’s location is its access to bulk and potentially heavy supplies (raw materials, water and fuel). Therefore, ground floors with street access are generally preferred. The Koan workshop is located at ground level and faces one of the most important thoroughfares of the ancient city, connecting directly with Kos’s commercial port (Livadiotti Reference Livadiotti2018, 59; see also Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019). The port of Kos was a significant trade hub of the Hellenistic world. The island is situated on one of the most important trade routes of the ancient Mediterranean, linking the Black Sea region to the southeast with the coast of Asia Minor (Höghammar Reference Höghammar, Höghammar, Alroth and Lindhagen2016, 97) and the gulf of Gökova and Halicarnassus (today’s Bodrum) to the east. After the synoecism of 366 bce, the foundation of the city of Kos and the establishment of the port, Kos emerged as a port town with wide and stable connections across the eastern Mediterranean and the Black Sea, including the port cities of Sinope, Tyros and Alexandria (Höghammar Reference Höghammar, Höghammar, Alroth and Lindhagen2016), which were important hubs for the trade of colourants (red earth, murex purple and Egyptian blue, respectively). Moreover, the fertile Island of Kos was well known for its exports, which, according to literary sources, included high-quality wine, silk textiles and perfumes (cf. Sherwin-White Reference Sherwin-White1978, 224–43). Thus, Kos was an attractive emporion for traders throughout the Hellenistic period (Höghammar Reference Höghammar, Höghammar, Alroth and Lindhagen2016, 133). Hence, the location of the workshop in the centre of the ancient city and in direct connection to the port would have been beneficial for the import of raw materials and fuel and for the export of local products. This suggests that, theoretically, workshops that rely on a range of raw materials would preferentially be located in port contexts.

Tracing Egyptian blue production at the Koan workshop: literary evidence and small finds

For pyrotechnological processes associated with the production of Egyptian blue, ceramic equipment, such as stilts and saggars, is to be expected among the material remains of the workshop. Vitruvius, in his graphic description of the Egyptian blue production process, informs us about the use of saggars: ‘When dry they [the Egyptian blue pellets] are collected in an earthenware jar, and the jars are put in a furnace’ (De arch. 7.11.1). Such vessels (saggars) have been found in Puteoli, Cumae, Liternum and Memphis, representing the primary evidence for Egyptian blue production in these locations (Cavassa Reference Cavassa and Jockey2018). However, only the workshop at Memphis has been definitively identified, and the rest, apart from Puteoli for which literary evidence also exists (Vitr., De arch. 7.11; Plin., HN 33.161), are only suggested as locations for the production of Egyptian blue based on the presence of saggars (Cavassa Reference Cavassa and Jockey2018), without other forms of evidence (fire-structures, workshop remains). As mentioned above, the absence of saggars among the archaeological finds from the Koan workshop was considered crucial evidence to exclude the Koan workshop from the known Egyptian blue production sites (Cavassa Reference Cavassa and Jockey2018). However, the discovery of specialized pottery or tools largely depends on abandonment and post-abandonment factors at the workshop (was the workshop looted, were tools removed for use elsewhere, etc.) (Tournavitou Reference Tournavitou, French and Wardle1988, 451). Thus, conditions of abandonment and the rate of disturbance and overbuilding should be taken into consideration. Therefore, we argue that the absence of saggars from the workshop on Kos cannot in itself be considered a sufficient reason to exclude the in situ production of Egyptian blue. Equally, the discovery of saggars in Puteoli, Cumae and Liternum indicates that Egyptian blue workshops were situated nearby, but cannot alone positively identify the exact location of a workshop space.

While keeping in mind our limited understanding of the Koan workshop space (Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019), the absence of saggars might indicate an alternative mode of production (Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2022). Whereas the Egyptian blue pellets found in saggars often adhere to each other due to their close contact during firing (see, for example, Rehren Reference Rehren and Carnall2013), the pellets from Kos are separated and feature one flattened surface, indicating that they were spread out on a flat surface without contacting each other during firing (Fig. 5). That the production of Egyptian blue in the Koan workshop represents an alternative mode of production is also supported by the size of the pellets, which range from 2 to 4 cm in diameter. The (published) dimensions of the pellets from other sites are smaller, ranging from 1.5 to 2 cm in diameter (Tite & Hatton Reference Tite, Hatton, Gosden, Hamerow, De Jersey and Lock2007, 75–6).

Figure 5.

Composite photograph showing Egyptian blue pellets from fire-structure 1 at various stages of successful firing. (Photograph: A. Marketou, courtesy of the Hellenic Ministry of Culture.)

Among the small finds, the most distinctive remains of pigment production are the raw materials, the waste and the final products. Among Egyptian blue pellets, the proportions of successfully produced pellets with the characteristic blue hue throughout the material (47%), partially successful pellets with blue and green, grey or brown hues (23%) and waste or unsuccessful pellets with hues ranging from brown to grey and purple (30%) provide strong evidence for local production (Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2021). Recent research linking the archaeological and archaeometric evidence for the local Egyptian blue production (Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2022; Kostomitsopoulou Marketou et al. Reference Kostomitsopoulou Marketou, Andriulo, Steindal and Handberg2020), coupled with experimental reconstructions of the suggested production process, could provide crucial information about the technology of production and the expected archaeological traces of such activities (Becker Reference Becker2022; pers. comm.). Finally, in his account of the manufacturing process of Egyptian blue, Vitruvius also describes how copper filings were used as a copper source, and it is possible that the filings from the bronze and copper found in the workshop could have served this purpose.

Tools, such as mortars, which are expected finds in a pigment-production workshop, were recovered within the workshop area (Fig. 6). Even though their exact use cannot be established at present, they may have been used in the production of purer forms of certain mineral pigments, including for example celadonite (identified at the workshop), azurite, or malachite, which require only minimal mechanical processing that can be achieved simply by grinding with mortars.

Figure 6.

(a) Fragment of a burnt marble basin. Inv. no. AE 372; (b) fragment of a marble mortar and finger-shaped pestle. Inv. no. SK 186. Found in the workshop and possibly used in production processes. (Photograph: A. Marketou, courtesy of the Hellenic Ministry of Culture.)

Tracing metallurgy and pigment production at the Koan workshop: literary evidence and small finds

Beyond the evidence supporting Egyptian blue production and pigment treatment, excavations of the workshop on Kos brought to light amorphous lead lumps and tubular litharge rods, as described above. Limited evidence from other sites suggests an interconnection between lead metallurgy and Egyptian blue production. For example, one Egyptian blue pellet was found in the context of a Hellenistic lead metallurgical workshop at Pherae in Thessaly (Asderaki-Tzoumerkioti & Rehren Reference Asderaki-Tzoumerkioti, Rehren and Karampelopoulos2006). Moreover, the shipwreck known as Planier 3, with a cargo of mainly Campanian wine amphorae, brought to light 20 pellets of realgar (As₄S₄), 41 kg of litharge and a small number of Egyptian blue pellets (Tchernia 1968, 302; Reference Tchernia2016, 80). The presence of both litharge and Egyptian blue among the cargo reinforces the statement of Pliny that Puteoli was an important litharge production site (Plin., HN 33.106–107). It should be noted, however, that none of these sites are located close to silver/lead mines; rather, they seem to exploit sea trade routes and access to multiple raw materials and fuel resources, as well as knowledge repositories of travelling craftspeople. These finds, corroborated by literary evidence, illustrate a relationship between Egyptian blue and the by-products of silver and lead metallurgy that remains to be explored further.

Lead compounds have frequently been detected in Hellenistic and Roman Egyptian blue samples of various provenances. The presence of lead in Egyptian blue pellets dating to Roman times from, for instance, Pompeii and several sites in Gaul has been interpreted as indicative of the presence of metallic lead or leaded copper alloy in the starting mixture (Blet et al. Reference Blet, Guineau and Gratuze1997, 126; Tite & Hatton Reference Tite, Hatton, Gosden, Hamerow, De Jersey and Lock2007, 86). The addition of lead oxide in the production of Egyptian blue would facilitate the pyrotechnological process, working as a fluxing agent (Zhang et al. Reference Zhang, Ma and Berke2019). Moreover, lead in the form of cerussite has been identified in Pompeiian Egyptian blue samples that have been further treated to produce cosmetics and pigments designated to a secco painting (Delamare Reference Delamare2013, 29–30). The evidence from the Koan workshop, where lead was also found in traces among the analysed Egyptian blue samples (Kostomitsopoulou Marketou et al. Reference Kostomitsopoulou Marketou, Andriulo, Steindal and Handberg2020), presents yet another possible explanation for the presence of lead in Egyptian blue: potential contamination of the material evidence due to shared workshop space. Since lead metallurgy and pigment production were both carried out in close proximity to each other, it is possible that lead contamination occurred as a result of working conditions in the workshop.

Litharge and lead could also be transformed into pigments, in what Brøns (Reference Brøns, Thavapalan and Warburton2019, 323–8) describes as ‘overlaps between craft traditions’. Lead white, a pigment known as psimithium in Greek and cerussa in Latin, which are lead carbonate and lead acetate respectively, was produced by heating metallic lead in the presence of vinegar in vessels intended to contain the acetic acid vapours. According to both Pliny and Vitruvius, the neighbouring Island of Rhodes was a producer of such lead-white pigment of a premium quality (Theophr., Lap. 56; Vitr., De arch. 7.7.1; Plin., HN 34.175). Firing of litharge (PbO) in an oxidizing atmosphere would also result in the production of lead tetroxide (Pb₃O₄), a pigment known in antiquity as cerussa usta or sandarach (Plin., HN 34.176) (Becker Reference Becker2022). Such material crossovers, or ‘overlaps’, in craft productions might also be visible in the Koan workshop in an iron and lead-containing red pellet, which may correspond to what Pliny the Elder describes as sandyx (Plin., HN 35.30, 40) (Kostomitsopoulou Marketou et al. Reference Kostomitsopoulou Marketou, Kouzeli and Facorellis2019).

The earth pigments identified in the Koan workshop, including hematite-based red earths, goethite, jarosite and celadonite (Kostomitsopoulou Marketou et al. Reference Kostomitsopoulou Marketou, Kouzeli and Facorellis2019), highlight the connection between earth pigments and mining known from ancient texts. Theophrastus, for example, comments that the best red earth comes from the small iron mines of Kea (Theophr., Lap. 52). According to Vitruvius (De arch. 7.7.1), the best quality of ochre came from the mines of Laurion in Attica, where a vein of ochre was followed as though a vein of silver. Likewise, among the sources for the different pigments, Pliny the Elder includes gold, silver and iron mines (HN 33.158; 35.35). The fact that these materials are found in a workshop context with evidence for both pigment and metallurgical activity emphasizes the material relationships between pigments and metals, which might extend to shared trade networks between metals, metal ores and pigments.

A multicrafting workshop: integrating pigment production and lead metallurgy

The juxtaposition of raw materials found in the Koan workshop seems to represent a clear example of multicrafting or multicraft production, where at least two craft technologies are carried out in the same space (Shimada Reference Shimada2007, 5). By introducing the notion of multicrafting and expanding from the mere sharing of space to the possibility of shared knowledge by a group of craftspeople, we aim to avoid the compartmentalization of the workshop’s manufacturing processes according to modern distinctions between end products like metallic lead and pigments. Rather, the limited space of the workshop highlights the interconnectedness of pyrotechnological processes that shared knowledge, skills, raw materials, facilities, tools and a common understanding of technological and physico-chemical principles. This aligns with what Brysbaert (Reference Brysbaert, Antoniadou and Pace2007, 336) classifies as the ‘composite objects/features’ that constitute the material manifestation of cross-craft interactions. In this regard, the Koan workshop recalls the close interrelationship between glass and metallurgical technologies (Fenn Reference Fenn2015) and the possible co-existence of potters and metalworkers in the same workshops (MacGillivray Reference MacGillivray, Hägg and Marinatos1987).

The archaeological context of the Koan workshop also offers a glimpse into the identities of the craftspeople (τεχνίται), even if the occupational identities of craftspeople are difficult to comprehend (Flohr Reference Flohr, Verboven and Laes2017, 147). The size and spatial layout of the workshop must have entailed close communication between the craftspeople on a daily basis, what Flohr (Reference Flohr, Verboven and Laes2017, 151–3) calls the ‘first social layer’. Even though the specific organization of production of Egyptian blue remains largely obscure due to the lack of pertinent evidence, epigraphical evidence from Ptolemaic Egypt offers some insights into the organization of pigment production. Nicola Reggiani (Reference Reggiani2012) has suggested that the kal(l)ainopoioi recorded on ostraca and a papyrus dating to the third century bce should be understood as a group of specialized professional producers of a turquoise frit (possibly also Egyptian blue, even though the evidence is indirect). Notably, the epigraphic record indicates that these kal(l)ainopoioi in Egypt were organized in an association (called ethnē) and that a tax was imposed on their consumption of natron, which was probably used as a flux, the so-called nitrikēs kal(l)ainopoiōn (Reggiani Reference Reggiani2012, 34, n. 30). This evidence implies well-organized and regulated large-scale production in Ptolemaic Egypt.

This glimpse into the possible organizational structure seems difficult to reconcile with the picture that emerges from the Koan workshop. Indeed, as already mentioned, the absence of saggars and the different firing techniques used for Egyptian blue production in the workshop provide an alternative mode of production to the one described by Vitruvius and to what is implied by the evidence for the kallainopoioi. Instead, the occupational identities of the craftspeople working in the Koan workshop should perhaps not be understood as persons specialized solely in pigment production or metallurgy, but rather as craftspeople skilled in a variety of activities. Such manufacturing environments are known to have existed in the Hellenistic and Roman periods, as reflected in early alchemical texts that describe a wide range of elaborate treatments applied to many different materials and that would have required proficiency in working with diverse substances, tools and firing techniques (Martelli Reference Martelli2011).

We may ask what the purpose of such a multicrafting workshop in the Koan agora might have served. One way would be to envisage its primary function as an auxiliary or support workshop for other industries located in the general vicinity. Artisan workshops historically tended to cluster together in geographical zones, often associated with other supporting or related crafts (Goodman Reference Goodman, Wilson and Flohr2016). The concept of such a supporting workshop is exemplified by Building F on the ‘Street of the Marble Workers’ in the Athenian Agora (Hochscheid Reference Hochscheid and Rodríguez Pérez2017). Based on lumps of pigments and vats that were associated with the building’s two construction phases, the refinement of pigments is suggested as one of the activities carried out in that workshop space. In this way, the workshop can be understood as a support workshop that produced pigments intended for the painting of the marble sculptures produced in nearby marble workshops.

The presence of other workshops in the general area of the Koan agora might support the idea of a network, or cluster, of workshops (for a brief overview, see Kostomitsopoulou Marketou Reference Kostomitsopoulou Marketou2019, 64–6, with nn. 21–25). Large quantities of corroded copper and bronze artefacts including statues and statuettes were found just south of the east–west passage in another room of the eastern stoa complex that was interpreted as a metallurgical workshop dated to the first century ce (the context has not been fully published, but see Giannikouri et al. Reference Giannikouri, Skerlou, Papanikolaou and Giannikouri2011, 368, with further refs; Kantzia Reference Kantzia1992, 640). Three moulds for the manufacture of terracotta figurines have been found in the eastern part of the agora, which strongly indicates the existence of a coroplastic workshop. Two of the moulds were found on the street that defines the eastern border of the agora, i.e., behind the eastern stoa complex, whereas the exact findspot of the third remains uncertain (Giannikouri et al. Reference Giannikouri, Skerlou, Papanikolaou and Giannikouri2011, 369; Skerlou & Grigoriadou Reference Skerlou, Grigoriadou and Giannikouri2014). Finally, an early Hellenistic pottery workshop (Kantzia Reference Kantzia, Drougou, Zervoudake, Maragkou, Papapostolou and Touratsoglou1994) and a late Roman glass workshop (Skerlou Reference Skerlou2013) have also been identified to the east. Although these workshops span a broader chronological range than the late Hellenistic pigment and metallurgical workshop in the stoa, they can be seen as part of a wider workshop district, serving an auxiliary role for neighbouring workshops. We might imagine, for instance, that the various pigments identified in the workshop could have been produced for the coroplastic workshop, where they could have been used in the painting of terracotta figurines and other terracotta products, just as Hochscheid (Reference Hochscheid and Rodríguez Pérez2017) proposed for the marble workshops in Athens. However, at least for now, we have no direct evidence for this practice. Metal scrap from the neighbouring bronze workshop (if the two ever co-existed) might have been utilized in the pigment productions.

Finally, the pigment and metallurgical workshop in the eastern stoa complex of the agora was closely connected to the main commercial zone in the city near the port (Livadiotti Reference Livadiotti2018, 58–9). This location within the eastern stoa complex fits the pattern of metallurgical workshops established in agoras for lead production from litharge as early as the fourth century bce (Kakavogiannis Reference Kakavogiannis2005, 284). Kakavogiannis argues that it would have been preferable to import litharge instead of metallic lead since it was significantly cheaper and could be worked into lead relatively easily. Information presenting tax laws for the Asian Roman provinces that is preserved in a first-century ce inscription from Ephesus, the lex portorii Asiae, points to the exemption of unworked earth (γῆ) from taxation (Cottier et al. Reference Cottier, Crawford, Crowther, Corbier, Mitchell and Nijf2008). The taxation of the raw mineral, which had an uncertain value, should be lower than the final processed product (Carrelli Reference Carrelli and Vergilio1996, 186). While the specific taxation laws regulating materials trade on Kos are uncertain, by importing unprocessed ores and processing them locally, workshops established in port-cities, such as the one on Kos, would be highly profitable. The establishment of workshops close to trade hubs provided easy access to all the materials available through established maritime trade links. Returning to the case of Egyptian blue, literary sources support this argument. Vitruvius (De arch. 7.11) names Alexandria as the place where Egyptian blue material was first produced, even if archaeological evidence in support of Alexandrian workshops has yet to be found. Similarly, both Puteoli and Kos were important commercial and manufacturing hubs of the Graeco-Roman trade network, with strong relationships with Egypt (Sherwin-White Reference Sherwin-White1978, 26, 82–131). The relationship of Kos with Ptolemaic Egypt, and especially Alexandria, is particularly close, with Koan doctors, workers, and soldiers travelling to Egypt to serve the Ptolemies (Sherwin-White Reference Sherwin-White1978, 97, 106, 299). In this context, the exchange of technology, materials, and even craftspeople could easily come about. Pigment workshops organized in port cities would indeed benefit from the direct access to imported raw materials and fuel, and even technological knowledge.