The systematic techno-stylistic and chemical study of glass beads from post-15th century West African sites
PMCID: PMC11809889
PMID: 39928606
Abstract
The systematic chemical analysis of large collections of archaeological glass beads is essential to better understand trade patterns at different times around the world. Glass beads’ trade towards and within sub-Saharan West Africa grew exponentially over time to culminate with the establishment of the Atlantic Trade. Although these artefacts are very commonly found in archaeological contexts dating after the 15 th century CE, the assemblages are generally poorly studied from a chemical point of view. We present here the study of 916 glass beads found in five archaeological sites in Ghana, Mali, and Senegal, in contexts dated between the 15 th and the mid-20 th century CE. Besides the techno-stylistic classification of the whole assemblage, the compositional study of a sub-group of 578 monochrome and polychrome glass beads was performed. The 798 glass samples composing the selected beads were therefore classified based on their main chemical composition. Moreover, major, minor, and trace elements analysis by Laser Ablation-Inductively Coupled Plasma-Mass Spectrometry (LA-ICP-MS) and the statistical analysis of the results by Principal Component Analysis (PCA) led to the identification of the probable origin of the glass. Different suppliers were distinguished for the Ghanaian earlier beads and the Senegalese and Malian later ones, in relation to the different European trade partners at different times.
Full Text
We present here the techno-stylistic and chemical characterisation of 916 glass beads found in five archaeological sites in Senegal, Mali, and Ghana in the framework of the research project “Human Population and Palaeoenvironment in Africa (HPPA)” coordinated by the laboratory Archaeology and Peopling of Africa (APA), today ARChaeology of Africa & ANthropology (ARCAN), at the University of Geneva in collaboration with several research groups in Switzerland, France, Germany, Mali, Senegal, and Ghana [1].
The first glass beads reached sub-Saharan West Africa in the first millennium BCE through sporadic exchanges across the Sahara Desert [2, 3]. From the 8th century CE, trans-Saharan trade routes progressively developed to peak between the 10th and the 15th century CE with the expansion of the Sahelian states. Among other commodities, glass beads were exchanged at different spatial scales, from the large trade centres to the rural areas [4]. Trans-Saharan trade then gave way to the Atlantic trade with the Europeans and the Americans from the 15th century onward, although it never completely disappeared. Outposts were built along the coasts of West Africa first by the Portuguese between the 16th and the 17th century, then by the Dutch, the French, and the British between the 17th and the 19th century [5, 6]. The main good sought by the Europeans at the very beginning was gold, but then ivory, pepper, precious wood, leather, and slaves became valuable trade goods, exchanged for textiles, guns, metals, alcohol, and glass produced in Europe. Historical documents such as travel accounts, invoices, and detailed trade reports, as well as archaeological finds, like in Elmina, Ghana [7–9], in Garumele, Niger [10], and in the Falémé Valley and other sites in Senegal [11–13], show how glass beads were especially important commodities, in large demand in West Africa [7, 12, 14–17].
Glass beads traded by Europeans in West Africa had specific typologies depending on the good they were exchanged for, and sample cards and catalogues were created by the beadmakers to categorize the items [18–21]. Distinctive qualifying terminology was hence used based on the origin and the type of the beads, as well as depending on the recipients of the goods [16: 22].
Local glass beads production has also been attested in various West African sites where imported glass items were reprocessed to create beads according to local taste [22–25].
The ingredients to produce man-made glass have changed at different historical times depending on raw material availability, technological knowledge, and market demand. The chemical characterization of archaeological glass artefacts and production waste showed that the main ingredients used for glass manufacture in the Mediterranean basin and Middle East between the 2nd millennium and the 9th century BCE were silica and halophyte plant ash [26–28], to switch to quartz limestone sand and mineral flux (mainly natron) between the 10th century BCE and the 9th century CE [29–32], and to go back again to vegetal flux (plant ash) from the 8th century CE onwards [33–35], although in the Middle East, east of the Euphrates River, the use of plant ash probably never ceased and resumed on a larger scale around the 3rd century CE [36–38]. Around the same time, in continental Europe glass artisans started to use wood ash as flux to lower the temperature of quartz sand and lime as glass stabilizer, leading to the production of potassic and, later, mixed akali glass. This transition was progressive, and many different recipes were ventured from the 8th to the 16th century CE, showing a gradual increase in CaO and P2O5 content and a progressive decrease in K2O content [27, 39–41]. Besides, the use of quartz sand, beech ash, and potash led to the production of the Bohemian glass in the regions of Bohemia and Silesia [40]. Another important production in Medieval Europe was lead glass. The use of lead to lower the glass working temperature was sporadic in the British Isles around the 10th–11th century CE and in Eastern Europe in the 12th–13th century CE, and more intense in Central and Northern Europe from the 16th century CE onwards [27, 42]. Finally, one of the most important glass productions in medieval and post-medieval Europe was the Italian glass manufacture (mainly in Liguria, Tuscany, and Veneto [43–46]). The production of soda-lime glass in Northern Italy started in the 6th–7th century CE with the importation and reworking of raw glass produced in the Middle East, making the distinction between the two types very difficult. It is not before the 13th century CE that primary production of glass in Murano and Venice started, using plant ash imported first from the Middle East and from Spain, and Southern France later, as well as the silica source imported from Ticino, in Switzerland [46]. The Venetian glass industry flourished and led the glass market between the 13th and the 17th century CE, showing a constantly developing technology, an increasing glass quality (from common glass to Vitrum Blanchum to Cristallo), and a vast variety of artefact types produced. Around the 16th century the raw materials supply became scarce, and glass artisans began to move towards Central and Northwestern Europe, taking with them the secrets of the Venetian glassmaking [46, 47]. Flawless imitations of the prestigious Venetian glass, both from a typological and a chemical point of view (the so-called Façon de Venise), started to sprout especially in the Netherlands, in British islands, and in France, effectively putting an end to the Venetian glassmaking monopoly at the end of the 17th century. Venice remained, however, an active and dynamic player in the glass beads market until the 19th century [17]. In the 19th century new industrial materials were introduced in the European glassmaking process, such as synthetic soda and potash for the base glass, fluorine as an opacifier, as well as various chromophores like nickel for grey glass, uranium for yellow glass, and cadmium and selenium for red glass [48, 49].
Old Buipe site is located 10 km north of the Black Volta River, close to its confluence with the White Volta River, in north-central Ghana (Fig 1).
According to written sources and oral traditions, Old Buipe was one of the main centres of power of the Gonja state between the 15th and the 18th–19th centuries CE. Moreover, it was an important commercial hub on the sub-Saharan trade routes between the Niger and the tropical forest [50]. Surveys and archaeological excavations carried out between 2014 and 2022 revealed a very extensive archaeological site consisting of settlement mounds, some of them with impressive dimensions and containing imposing architectural remains [50–52]. The main occupational sequence of the site took place between the 15th and the 18th century CE. The 19th century CE witnessed a first shrinkage in size and a shift of the settlement towards the north, where the colonial-period town flourished. By the mid-20th century CE, most of the town was abandoned in favour of a new site a dozen km to the east, and only a small village remained in Old Buipe. Archaeological finds include artifacts of local and regional production (ceramic, clay smoking pipes, iron objects, etc.), as well as importation items like European clay smoking pipes and ceramics, cowries, and glass beads; with the notable exception of the glass beads, most of the importation items originates from 19th and 20th century CE contexts. A total of 18 glass beads found between 2015 and 2017 and coming from the earlier occupational phase were selected for techno-stylistic and chemical analysis [53]. Most of the beads are composed by multiple layers of glass for a total amount of 45 different glass samples analysed (Table 1).
Alinguel is a large settlement site located in the Falémé River valley in Eastern Senegal (Fig 1). This site is mentioned in the 19th century travelogues of various travellers and trade agents as a hinge between the regions where the Boundou and Bambouk kingdoms were established [54]. The site was excavated in 2012 and 2013, leading to the identification of three settlement areas occupied during several centuries. Radiocarbon dating allowed to determine three occupational phases, the earlier one between the 1st and the 10th century CE, the middle one between the 11th and the 13th century CE, and the latter one between the 17th and the 19th century CE [54]. The extensive excavation of the site delivered a considerable number of potsherds, as well as metallic objects, lithic grinding tools, botanical macro-remains, and 227 beads made of glass (N = 208), ceramic (N = 10), stone (N = 4), copper (N = 2), iron (N = 1), bone (N = 1), and indeterminate material (N = 1). Glass beads were recovered from various chronological levels of the central area of the site: 7 were found in the layers dating from the middle phase, 178 from the recent occupational levels, whereas 23 come from the surface [53]. The beads from the 11th–13th century context resulted to be displaced from the more recent layers and were therefore included in the 17th–19th century assemblage [53]. A total of 207 beads were included in the techno-stylistic classification, one of the beads being too degraded to be identified. Moreover, 184 glass samples composing 129 beads were selected for chemical characterization; only the beads coming from the surface and the beads presenting an advanced chemical degradation were excluded from the analysis (Table 1).
Toumbounto is a settlement site located on the left bank of the Falémé River in Eastern Senegal (Fig 1). As for Alinguel, this site is mentioned in several written accounts of travellers, describing it as a remarkable village linked to the Boundou kingdom in the 19th century CE [54]. The site, excavated between 2012 and 2015, consists of 76 structures of various functions and one grave. Eight of these structures, namely three storage structures, two habitations, one structure linked to metallurgical activities, one grave, and one structure of unknown function, were extensively excavated [54]. Oral accounts and radiocarbon dating of several charcoal fragments and one calcinated seed led to the identification of one main occupational phase between the 17th and the 19th century CE linked to the architectural structures, as well as scant traces of a much older occupation around the 1st century CE [54]. The archaeological finds of various nature coming from the main occupational phase include 704 beads, 690 of which made of glass, 8 made of stone, 5 made of undetermined material, and 1 made of metal [53]. A total of 687 glass beads were studied and included in the techno-stylistic classification, 426 of which were selected to be chemically analysed. These beads are composed from one to four layers of glass, for a total amount of 561 glass samples analysed (Table 1).
The archaeological site of Farabana is located on the Malian side of the Falémé River Valley, 10 km east of the border between Senegal and Mali (Fig 1). The written accounts of the Compagnie du Sénégal, the French chartered company administering part of Senegal, mentioned the construction in 1724 of a fort near the village where the Bambouk king had his residence, precisely Farabana [55]. The excavation of the site in 2014 and 2016 led to the discovery of the ruins of this fort, which was used between 1724 and 1758 according to the texts and confirmed by the radiocarbon dating of six charcoal fragments [56: 151]. The archaeological finds include ceramics and smoking pipes of local production, cast-iron cannons, as well as surprisingly sporadic importation items, such as tableware, glass bottle fragments, and clay smoking pipes [56, 57]. Only four glass beads were recovered during excavation, one of which was found in surface and therefore excluded from the analysis. A total of 4 glass samples were chemically characterised (Table 1).
Tyi is a large settlement site composed by several districts located on the top of the Bandiagara Escarpment in Mali (Fig 1). Three structures from two of the districts were excavated between 2008 and 2010, specifically a habitation in Tyi-jo and two open-air pottery firing structures in Tyi-jo and Tyi-kun. The study of the archaeological finds and the oral surveys among the people living in the present village of Tyi in the plain led to the conclusion that Tyi-jo was occupied between the end of 16th and the end of 17th century CE, whereas Tyi-kun between the 18th and mid-20th century CE [58]. A total of 13 beads were recovered during the surface survey and the excavation of Tyi-kun, 8 of which made of glass. Of these, only two were found in stratigraphic position and included in the techno-stylistic and chemical study [53]. A total of 4 glass samples were chemically analysed (Table 1).
Shape: shape of the cross and longitudinal sections of the bead based on Horace Beck classification system [59].
Colour: Munsell colour code of each glass composing the bead [60].
The 798 glass samples composing 578 monochrome and polychrome beads selected for characterisation were analysed by Laser Ablation–Inductively Coupled Plasma–Mass Spectrometry (LA–ICP–MS) at the Institut de Recherche sur les Archéomatériaux (CEB-IRAMAT) of Orleans in France. A Thermo Fisher Scientific ELEMENT XR mass spectrometer was used for the analysis and the glass was sampled with a Resonetics M50E excimer laser working at 193 nm and operating at 5 mJ energy and 10 Hz pulse frequency. A 20 s pre-ablation was applied to reach the unaltered bulk glass, followed by 30 s of signal acquisition, corresponding to 10 mass scans. Setting the beam diameter between 50 and 100 μm led to the creation of craters 150–250 μm deep, invisible to the naked eye. The argon/helium flow carrying the sampled aerosol to the injector inlet of the plasma torch works at a rate of 1 l/min for Ar and 0.65 l/min for He. The signal in count-per-second was measured in low-resolution mode for 58 isotopes, from lithium to uranium. Standard Reference Materials from the National Institute of Standards and Technology (NIST SRM) 610 and the Corning reference glasses B, C, and D were used for external calibration. Reference materials Corning A and NIST SRM612 were analysed with the beads as unknown samples. The in-house archaeological sample APL1 was used for chlorine quantification and 28Si was used as an internal standard. The protocol detailed in [61] and [62] was followed to calculate the concentration of each element, with detection limits from 0.01% to 0.1% for major elements, and from 20 to 500 ppb for minor and trace elements.
The macroscopic and microscopic analysis of 916 glass beads led to their classification in 63 different techno-stylistic types based on their morphology, optical properties, and manufacturing techniques (Figs 2 and 3). The beads were separated into 15 categories, namely monochrome beads of different colour (white, colourless, pink, red, green, blue, orange, yellow, and black), Cornaline d’Aleppo, Green Hearts, striped beads, eye beads, feather beads, and chevron beads. S1 Table lists the beads types’ morphological, optical, and manufacturing characteristics previously described.
The full chemical composition of the 798 glass samples composing the selected 578 monochrome and polychrome beads included in the techno-stylistic classification, as well as three stylistically unidentifiable fragments of beads, can be found in S2 Table.
Principal Component Analysis (PCA) was applied to the collection taking into account the oxides linked to the silica source (SiO2, Al2O3, CaO, Fe2O3), to the flux (K2O, MgO, CaO, P2O5, Cl, B2O3, PbO), and to the colouring agents (PbO, Fe2O3, MnO, CuO, As2O3, SnO2, Sb2O3) in order to identify the main chemical groups. Keeping in mind that some of the oxides can be associated to multiple ingredients, it was possible to differentiate between four main compositional groups, i.e. soda/soda-potash-lime silica glass (S/SPL), potash-lime silica glass (PL), lead and lead-alkali silica glass (divided into three clusters depending on the alkali content: lead silica–LS; lead-soda-potash–LSP; lead-potash–LP), and lead-soda silica glass of intermediate composition (LSS) (Fig 4). The main oxide average composition of the main chemical groups can be found in S3 Table.
Table 2 lists the sub-groups of each main chemical groups indicated by the colour of the glass (blk: black, b: blue and turquoise, cls: colourless, g: green, o: orange; p: pink, r: red, w: white, y: yellow) and the chemical element characterising the colourant. The number of analysed glass samples for each sub-group, as well as the correlation between these groups and the techno-stylistic types of the corresponding beads are also indicated (cf. Figs 2 and 3, and S1 Table).
The average composition of each chemical sub-group is reported in S4 Table.
The only monochrome red bead of the collection was found in Toumbounto and it is composed by alumina-silica glass fluxed with soda, coloured by cadmium sulfoselenide, and containing 5 wt% of boron (group S/SPL r2, type 47). It is interesting to notice that a fragment of glass with the same composition was found in Dourou-Boro, an archaeological site not far from Tyi-kun in Mali, in a context dating from the 7th–9th century CE [4]; given that this chromophore was used only after the 19th century CE [63, 64], this fragment was interpreted as the result of looting of the funerary site.
The ruby glass composing all Cornaline d’Aleppo analysed is lead-potash silica glass containing 8 to 15 wt% of potash and 3 to 7 wt% of soda. The colouring agent of this glass is colloidal gold in concentrations between 53 and 435 ppm (N = 50, group LP p, types 12 to 17). The technology for creating ruby red glass through metallic gold is ancient and involves the precipitation of Au(0) nanoparticles though addition of tin salts, iron oxides or, more recently, arsenic salts as reducing agents [49, 65–67]. The high amount of arsenic present in this glass suggests the use of an arsenic compound together with gold to obtain the ruby colour, as described in Venetian recipes dating from the end of the 17th century [68]. It is possible to see a pattern in the amount of gold and lead oxide in relation with the typology of the beads. Type 12 shows in fact (with one exception) lower amount of PbO (11±2 wt%) and higher amount of Au (275±52 ppm), whereas types 13 to 17 show higher concentration of PbO (24±3 wt%) and lower concentration of Au (105±32 ppm) (Fig 5).
The so-called Green Hearts are generally composed by two layers of glass, one opaque carmine red on the exterior and one transparent to translucent, colourless to green one on the interior. Few beads also show a transparent colourless glass layer on the external surface. The 88 Green Hearts found in Alinguel, Toumbounto, and Farabana can be divided into three types, from 18 to 20, based on their size and finish (S1 Table). A total of 163 glass samples were chemically analysed, 79 of which are green, 82 are red, and 7 are colourless. All glass samples, except the ones composing one bead from type 19, show the same base composition, i.e., soda-lime silica glass with around 13 wt% of soda, 3 wt% of potash, and 10 wt% of lime. The only factor changing is the amount of copper and iron oxides giving colour to the glass (Fig 6).
The colourless layer of glass on the surface of the beads has negligible amount of copper and iron oxides (N = 7, group S/SPL cls, types 19, 20). On the other hand, the 82 red glass samples from group S/SPL r1 show higher amount of CuO (2.5±0.8 wt%) and Fe2O3 (2.7±0.8 wt%), whereas the 79 green glass samples from group S/SPL g contain 0.9±0.8 wt% of CuO, 1.6±0.4 wt% of Fe2O3, and varying amount of antimony oxide (between 0.1 and 3.2 wt%), in relation with the opacity of the glass. It is difficult to understand the nature of the glass chromophore without a molecular analysis of the glass. However, the different amount of copper and iron oxides could explain mechanisms of red glass colouration. As the green colour is generally given by CuO, the red colour could be either given by Cu2O or metallic Cu particles in suspension [49, 67, 69]. It has been proven that in red glass containing copper oxide in concentration higher than 5 wt% and high amount of lead oxide, the formation of Cu2O crystals is favoured [70, 71], which is not the case for the analysed samples. Besides, the high concentration of iron oxide in our samples can explain the precipitation of the metallic copper particles in the glass, as it could work as a reducing agent for the copper oxide [67, 71].
The feather beads decoration is made by winding rods of different glass colour around the bead core and then creating a pattern by dragging the trails in different directions [72].
Chevron beads were created by blowing a mass of glass into a star-shaped mould, repeating the process layer after layer with different coloured glass. After drawing a glass cane and cutting the singular pieces, the layers were revealed by mechanical abrasion of the extremities of each bead [20, 72, 73].
The four chevron beads found in Old Buipe show different number of layers and colours (types 60 to 63), like the very degraded red and white chevron bead not included in the techno-stylistic classification (n/a) (Table 2). Besides, the composition of the glass of each colour is the same of the ones composing the other beads from this archaeological site. The white layers of types 60, 62, 63, and “n/a” are made of lead-soda silica glass opacified by tin oxide (N = 4, group LSS w), whereas the white glass from type 61 is a soda-lime glass opacified by tin oxide (N = 1, group S/SPL w1); the red, blue, and colourless glass samples of all chevron beads have a soda/soda-potash-lime-silica base composition and they are coloured respectively with metallic copper (N = 6, group S/SPL r1), cobalt (N = 4, group S/SPL b1), and various amount of tin oxide depending on the opacity (N = 3, group S/SPL cls). All glasses show a positive linear correlation between Sn2O and PbO, indicating the probable use of lead-tin calx to opacify the matrix as for most of the other beads found in Old Buipe (see infra) (Fig 7) [74].
Many of the analysed beads coming from the excavated sites in Ghana, Senegal, and Mali with a specific typology can be associated to well-known European productions thanks to historical documents, travel accounts, and trade reports describing and classifying these items. The Cornaline d’Aleppo, also known as white hearts or Hudson Bay beads, for example, were produced from the beginning of the 19th century in Venice and can be found, among others, in the catalogues of the Levin Collection (1851–1863), displaying the beads typologies produced specifically for the trade with West Africa [18, 19]. In these catalogues, it is also possible to find the Green Hearts, or pre-white hearts, exchanged since the 17th century [19, 21], which are also displayed in the sample cards of the Venetian society Società Veneziana per l’industria delle conterie and of the Amsterdam’s factory J.F. Sick & Co. dated from the end of the 19th and the beginning of the 20th century [19]. The samples cards from these two societies exhibited in addition a great variety of monochrome beads, also known as seed beads or rocailles, largely present in our collection. Chevron beads, also known as rosetta beads, were originally produced in Venice in the 15th and 16th century and later imitated by other production centres in the Netherlands, Germany, and in North America. All chevron beads patterns from our collection can be found in the catalogues of chevron beads for the West African trade created by John and Ruth Picard [20, 21], as well as in the private collection of Cesare Moretti [75]. In the Picard catalogues, it is also possible to find a good match for our blue multifaceted beads known as Russian Blues [76], produced originally in Bohemia after the 18th century CE [47, 72, 77, 78]. The eye beads and the feather bead from our collection can also be found in these catalogues (respectively on page 21 of [76] and page 3 and 32 of [20].
Many types of beads from our collection can also be matched to glass beads found in several archaeological sites in sub-Saharan West Africa [7–13]. However, few of these assemblages were chemically analysed. The typological analysis of the beads alone is in fact not enough to pinpoint the production sites of the glass and the artifacts. If Venice and Murano were the main European production sites between the 13th and the 16th century CE, many expert artisans started to migrate towards northern Europe to create new production centres, while Venice still remained very productive [17, 46, 47, 79, 80]. The chemical analysis of the glass composing the beads might help distinguish between different productions dated 17th to 20th century CE, although it still remains a complicated task, as the glass recipes and raw materials used in different glasshouses were often extremely similar, if not the same. As of today, no systematic chemical analysis of the glass beads exhibited in the commercial catalogues has been done, so we can rely for the provenance study solely on the characterisation of glass artefacts and production waste found in archaeological sites in Europe [81–86], North America [87–95], and sub-Saharan Africa [10, 96–100]. Furthermore, the approximate chemical composition of various 15th–20th century glass types calculated from multiple Venetian recipe books can be used for comparison [48, 74, 75, 101–103].
In order to evaluate the possible provenance of the glass samples, the four main compositional groups discussed above, namely soda/soda-potash-lime silica glass, potash-lime silica glass, lead and lead-alkali silica glass, and lead-soda silica glass of intermediate composition (Fig 4, Table 2, and S3 Table), were taken into account. To simplify the provenance study, some of the glass samples of intermediate composition were included in the discussion either of the soda/soda-potash-lime silica glass or the lead and lead-alkali silica glass.
The white bead from type 2, the yellow bead from type 41, and the red bead from type 47 found in Toumbounto show very high levels of alumina (10.5–13.7 wt%), suggesting that they are faïence, sintered glass or Prosser-molded glass or ceramic beads probably produced after the 19th century [63]. This hypothesis is supported by the modern colourants detected, i.e. the Imperial Red (cadmium sulfoselenide), and the chromium yellow, used as glass chromophores in Venice starting from the second half of the 19th century [48, 63, 92, 104]. Since no classical opacifying agents (e.g. Sb, Sn, or As compounds) were detected, the opacity of these beads is probably due to the presence of unmelted minerals in the glass matrix, although this should be confirmed by the molecular analysis of the glass.
The other alkali-lime silica glass samples can be further divided in two subgroups in relation to the tin and lead oxide contents (Fig 8).
The majority of the glass samples composing Old Buipe beads shows a positive correlation between lead and tin, suggesting a coherence in the glass supply and indicating the use of the so-called calce di stagno e piombo (lead-tin calx) for opacification, a method used by Venetian glassmakers until the 19th century [74]. Moreover, the high Sn/Pb content in the glass due to the use of tin as opacifier suggests a production predating 1650, after which the use of antimony as an opacifier became systematic [84]. This is coherent with the results of the study of North American archaeological bead assemblages of probable Dutch origin, for which the tin content is used as a chronological marker in opposition to antimony and arsenic [89, 92, 93]. The composition of Old Buipe glass is comparable to the chemical group 2 found in the North American sites of Petun and Seneca, although with lower levels of Na2O and CaO [93].
Old Buipe glass samples have the typical composition of tin-opacified lattimo glass produced in Venice between the 15th and 17th century [75]. The comparison with the composition of glass assemblages from European archaeological sites and museum collections shows that they are comparable with the glass beads found in the late 17th century Hammersmith Embankment site in London (chemical group 1) and the glasswork waste from the contemporary Kg10 production site in Amsterdam [82, 105] (Fig 9).
Comparison between Old Buipe glass composition and the composition of glass from Amsterdam and London [82, 105], Old Broad Street, London [106], Louvre [103], and Venice [75].
Moreover, compositional and typological similarities can be found between Old Buipe chevron beads and the chevron beads found in the 16th–19th century glasswork site of Neularten, Germany, probably imported from Venice or Amsterdam [83, 84].
Concerning the soda and soda-potash-lime silica glass beads from Alinguel, Toumbounto, and Farabana, the composition of contemporary European and African archaeological glass and glass beads assemblages can be used to understand the possible origin of these beads (Fig 10).
Comparison between mixed alkali and soda-lime glass composition and the composition of glass from Amsterdam and London [82], Antwerp, Mechelen [86], Garumele [10], Louvre [103], Middleburg [85], Rouen [81], Venice [101], and Kulumbimbi and Lumbu, Angola [98, 99]. For Venetian glass: C = Cristallo, VB = Vitrum Blanchum.
The base glass of our assemblage can be compared to most archaeological European glass assemblages for which a Venetian production is suggested, as well as to the Cristallo and Vitrum Blanchum glass produced in Venice between the 15th and the 17th century [101]. Some of our beads, namely few turquoise beads from type 21 and the red and green glass composing some Green Hearts from type 18 and 19, have lower potash, magnesia, lime, and alumina content, much like Cristallo glass produced with purer sources of silica. However, unlike Cristallo, they contain higher amount of phosphorus, chlorine, and iron, especially the Green Hearts.
The other glass samples, more similar to the Vitrum Blanchum references, show higher amount of potash, much like the glass samples from 17th century Mechelen and Antwerp, in Belgium [86]. Here, to explain this difference in K2O content, the authors suggest a production in the Low Countries instead of Venice, which is supposedly supported by the trace element content. Unfortunately, the trace element concentrations were not published, so it is not possible to compare them to the composition of our samples. However, a higher potash content is typical of Venetian glass produced between the 16th and 18th century adding to the glass matrix the so-called grepola, the deposit of the bottom of the wine barrel (potassium bitartrate). The higher potassium oxide content could be also explained by the addition of tartar (K-bearer) to the glass matrix, practice more and more common from the 17th century onward [46, 107].
It is interesting to notice that techno-morphologically consistent soda-lime glass beads from Garumele, the only contemporary assemblage from a West African site to be chemically analysed, show lower concentrations of soda and lime, as well as higher concentration of potash and alumina compared to our beads, so a different production site is probable (Fig 10).
Moreover, comparing our assemblages to northern American assemblages, a parallel can be found with the soda-lime beads from 18th–19th century Sullivand’s Island attributed to the Venetian production, although the latter show in average lower alumina content [94]. Even though beads from other North American sites [87, 89–93] show lower soda and lime content and higher potash and alumina content, like Garumele beads, the opacifier content, especially for white beads, can be used as an indicator of possible production. As already mentioned, tin, antimony, and arsenic were used as chronological markers for North American bead assemblages of possible Venetian or Dutch production, dating antimony opacified beads between the 17th and the 19th century [89, 92]. Considering that Amsterdam stopped to be an important glass beads production centre after the beginning of the 18th century [94], we can suggest that the mixed alkali and soda-lime beads from our assemblage are of Venetian manufacture, probably dating between the late 17th and the 19th century.
This group includes the glass composing the colourless monochrome beads (type 11), the blue monochrome beads from types 31 and 33, and one blue glass sample not included in the techno-stylistic analysis (TO-29-80-3b). All beads were found in Alinguel and Toumbounto, Senegal. Their high content in potash and phosphorus corresponds to the use of forest wood ash as a flux, typical of the 18th–19th century Bohemian glass production [64, 68, 94]. Typologically and chemically consistent beads were found in North American [88, 94], European, and West African [21, 64] sites dating after the 18th century.
The colour of the blue beads is given by cobalt (233–624 ppm), which can be correlated to the presence of arsenic (928–3875 ppm), nickel (60–508 ppm), bismuth (50–576 ppm), and silver (0.3–3.6 ppm), as well as uranium in type 33 (S1 Fig). This is in line with the composition of the minerals extracted from the Erzgebirge mountain range, in Germany, one of the main sources of cobalt pigments for European glass and glaze production from the 14th century onwards, and especially between the 17th and 19th century [99, 108, 109]. Other Europeans ores showing similar mineral associations (e.g. France, Sweden, and Bohemia) can’t however be excluded [110–112].
The study of the potential provenance of the glass samples from this group was done by comparing the chemical composition to the Venetian glass recipe books, as well as the lead and mixed-alkali lead glass found in various European, African, and North American archaeological sites (Fig 11).
Comparison between lead and mixed-alkali lead glass composition and the composition of glass from Amsterdam [82], Garumele [10], Louvre [103], MET [102], Murano [75], Rouen [81], and Southern Africa [100].
Orange and green glass show a composition similar to the glass of the same colours found in Garumele, Niger, of probable Venetian manufacture and suggested to be dated between the late 17th and the 19th century [10]. Moreover, the comparable Nb, Zr, and Sr content indicates that a similar source of raw materials was probably used for manufacture. Similarities can be found also with the composition of the yellow and green enamels of Venetian glass from Louvre [103], showing however higher amount of tin and antimony oxides, and slighter higher soda content. Finally, the composition of the yellow glass can be compared to the theorical composition of the Venetian anime that were mixed with transparent glass to create yellow opaque glass in the 19th and 20th century [75].
Arsenic opacified white glass is mentioned in the Venetian recipe books starting from the 18th century CE. The arsenic white monochrome beads from our collection can be compared to the opaque Girasole and Smalto white glass from Murano mentioned by Moretti and Hrelich [75], as well as to the white wound beads found in Garumele [10], and to the white beads from late 18th–19th century Sullivan’s Island site, in northern America, of probable Venetian origin [94]. On the other hand, the arsenic white glass composing our Cornaline d’Aleppo (as well as the ruby red glass from the same beads) has a very different composition from the red-on-white beads from Sullivan’s Island, suggested to be of Bohemian origin [94]. Correlations can be found with the composition of the red-on-white beads found in Angola, although the gold content of the ruby red glass is much lower than in our beads [99]. Although no systematic chemical analysis of red-on-white glass beads has been found in literature, the typical composition calculated from the Venetian recipe books [68, 94] is in fact comparable to our Cornaline d’Aleppo’s composition.
This group comprises one Green Heart (TO-2-6), the three blue beads from type 24, and one of the green monochrome beads from type 34. The other samples included in this group as a result of PCA (Table 2) were included in the provenance discussion of the other groups.
Both glass composing sample TO-2-6 contain higher soda and alumina concentration compared to the other Green Hearts and the green layer show traces of chromium oxide. This suggests that the bead was produced after the 19th century, when chromium green glass was produced in Venice [48]. It is also the case for the green bead from type 34 having similar base composition as the other green beads from the same type but showing a higher amount of soda and alumina and lower amount of lead, as well as a very high content of chromium oxide (0.3 wt%).
Morphological, optical, and manufacturing characteristics of each type of beads showed in Figs 2 and 3. Abbreviations: Category: Mon.: Monochrome, C.A.: Cornaline d’Aleppo, G.H.: Green Heart. Site: AL: Alinguel, TO: Toumbounto, OB: Old Buipe, FA: Farabana, TK: Tyi-kun. Shape: IR: irregular. Colour: w: white, cls: colourless, p: pink; r: red, g: green, b: blue, o: orange; y: yellow, blk: black. Diaphaneity: OP: opaque, TL: translucid, TR: transparent. The + sign in the colour and diaphaneity sections refers to different glasses in the same layer of the bead. The number of layers, colour, and diaphaneity of glasses in brackets refer to some of the beads of the same type. Technique: D: drawing, W: wounding, M: moulding. n.d.: indeterminate.
Sections
"[{\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref001\"], \"section\": \"Scope and objectives\", \"text\": \"We present here the techno-stylistic and chemical characterisation of 916 glass beads found in five archaeological sites in Senegal, Mali, and Ghana in the framework of the research project \\u201cHuman Population and Palaeoenvironment in Africa (HPPA)\\u201d coordinated by the laboratory Archaeology and Peopling of Africa (APA), today ARChaeology of Africa & ANthropology (ARCAN), at the University of Geneva in collaboration with several research groups in Switzerland, France, Germany, Mali, Senegal, and Ghana [1].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref002\", \"pone.0318588.ref003\", \"pone.0318588.ref004\", \"pone.0318588.ref005\", \"pone.0318588.ref006\", \"pone.0318588.ref007\", \"pone.0318588.ref009\", \"pone.0318588.ref010\", \"pone.0318588.ref011\", \"pone.0318588.ref013\", \"pone.0318588.ref007\", \"pone.0318588.ref012\", \"pone.0318588.ref014\", \"pone.0318588.ref017\"], \"section\": \"Glass beads trade in West Africa\", \"text\": \"The first glass beads reached sub-Saharan West Africa in the first millennium BCE through sporadic exchanges across the Sahara Desert [2, 3]. From the 8th century CE, trans-Saharan trade routes progressively developed to peak between the 10th and the 15th century CE with the expansion of the Sahelian states. Among other commodities, glass beads were exchanged at different spatial scales, from the large trade centres to the rural areas [4]. Trans-Saharan trade then gave way to the Atlantic trade with the Europeans and the Americans from the 15th century onward, although it never completely disappeared. Outposts were built along the coasts of West Africa first by the Portuguese between the 16th and the 17th century, then by the Dutch, the French, and the British between the 17th and the 19th century [5, 6]. The main good sought by the Europeans at the very beginning was gold, but then ivory, pepper, precious wood, leather, and slaves became valuable trade goods, exchanged for textiles, guns, metals, alcohol, and glass produced in Europe. Historical documents such as travel accounts, invoices, and detailed trade reports, as well as archaeological finds, like in Elmina, Ghana [7\\u20139], in Garumele, Niger [10], and in the Fal\\u00e9m\\u00e9 Valley and other sites in Senegal [11\\u201313], show how glass beads were especially important commodities, in large demand in West Africa [7, 12, 14\\u201317].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref018\", \"pone.0318588.ref021\", \"pone.0318588.ref016\"], \"section\": \"Glass beads trade in West Africa\", \"text\": \"Glass beads traded by Europeans in West Africa had specific typologies depending on the good they were exchanged for, and sample cards and catalogues were created by the beadmakers to categorize the items [18\\u201321]. Distinctive qualifying terminology was hence used based on the origin and the type of the beads, as well as depending on the recipients of the goods [16: 22].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref022\", \"pone.0318588.ref025\"], \"section\": \"Glass beads trade in West Africa\", \"text\": \"Local glass beads production has also been attested in various West African sites where imported glass items were reprocessed to create beads according to local taste [22\\u201325].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref026\", \"pone.0318588.ref028\", \"pone.0318588.ref029\", \"pone.0318588.ref032\", \"pone.0318588.ref033\", \"pone.0318588.ref035\", \"pone.0318588.ref036\", \"pone.0318588.ref038\", \"pone.0318588.ref027\", \"pone.0318588.ref039\", \"pone.0318588.ref041\", \"pone.0318588.ref040\", \"pone.0318588.ref027\", \"pone.0318588.ref042\", \"pone.0318588.ref043\", \"pone.0318588.ref046\", \"pone.0318588.ref046\", \"pone.0318588.ref046\", \"pone.0318588.ref047\", \"pone.0318588.ref017\", \"pone.0318588.ref048\", \"pone.0318588.ref049\"], \"section\": \"Glass production in the Mediterranean basin and Europe\", \"text\": \"The ingredients to produce man-made glass have changed at different historical times depending on raw material availability, technological knowledge, and market demand. The chemical characterization of archaeological glass artefacts and production waste showed that the main ingredients used for glass manufacture in the Mediterranean basin and Middle East between the 2nd millennium and the 9th century BCE were silica and halophyte plant ash [26\\u201328], to switch to quartz limestone sand and mineral flux (mainly natron) between the 10th century BCE and the 9th century CE [29\\u201332], and to go back again to vegetal flux (plant ash) from the 8th century CE onwards [33\\u201335], although in the Middle East, east of the Euphrates River, the use of plant ash probably never ceased and resumed on a larger scale around the 3rd century CE [36\\u201338]. Around the same time, in continental Europe glass artisans started to use wood ash as flux to lower the temperature of quartz sand and lime as glass stabilizer, leading to the production of potassic and, later, mixed akali glass. This transition was progressive, and many different recipes were ventured from the 8th to the 16th century CE, showing a gradual increase in CaO and P2O5 content and a progressive decrease in K2O content [27, 39\\u201341]. Besides, the use of quartz sand, beech ash, and potash led to the production of the Bohemian glass in the regions of Bohemia and Silesia [40]. Another important production in Medieval Europe was lead glass. The use of lead to lower the glass working temperature was sporadic in the British Isles around the 10th\\u201311th century CE and in Eastern Europe in the 12th\\u201313th century CE, and more intense in Central and Northern Europe from the 16th century CE onwards [27, 42]. Finally, one of the most important glass productions in medieval and post-medieval Europe was the Italian glass manufacture (mainly in Liguria, Tuscany, and Veneto [43\\u201346]). The production of soda-lime glass in Northern Italy started in the 6th\\u20137th century CE with the importation and reworking of raw glass produced in the Middle East, making the distinction between the two types very difficult. It is not before the 13th century CE that primary production of glass in Murano and Venice started, using plant ash imported first from the Middle East and from Spain, and Southern France later, as well as the silica source imported from Ticino, in Switzerland [46]. The Venetian glass industry flourished and led the glass market between the 13th and the 17th century CE, showing a constantly developing technology, an increasing glass quality (from common glass to Vitrum Blanchum to Cristallo), and a vast variety of artefact types produced. Around the 16th century the raw materials supply became scarce, and glass artisans began to move towards Central and Northwestern Europe, taking with them the secrets of the Venetian glassmaking [46, 47]. Flawless imitations of the prestigious Venetian glass, both from a typological and a chemical point of view (the so-called Fa\\u00e7on de Venise), started to sprout especially in the Netherlands, in British islands, and in France, effectively putting an end to the Venetian glassmaking monopoly at the end of the 17th century. Venice remained, however, an active and dynamic player in the glass beads market until the 19th century [17]. In the 19th century new industrial materials were introduced in the European glassmaking process, such as synthetic soda and potash for the base glass, fluorine as an opacifier, as well as various chromophores like nickel for grey glass, uranium for yellow glass, and cadmium and selenium for red glass [48, 49].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g001\"], \"section\": \"Old Buipe, Ghana\", \"text\": \"Old Buipe site is located 10 km north of the Black Volta River, close to its confluence with the White Volta River, in north-central Ghana (Fig 1).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref050\", \"pone.0318588.ref050\", \"pone.0318588.ref052\", \"pone.0318588.ref053\", \"pone.0318588.t001\"], \"section\": \"Old Buipe, Ghana\", \"text\": \"According to written sources and oral traditions, Old Buipe was one of the main centres of power of the Gonja state between the 15th and the 18th\\u201319th centuries CE. Moreover, it was an important commercial hub on the sub-Saharan trade routes between the Niger and the tropical forest [50]. Surveys and archaeological excavations carried out between 2014 and 2022 revealed a very extensive archaeological site consisting of settlement mounds, some of them with impressive dimensions and containing imposing architectural remains [50\\u201352]. The main occupational sequence of the site took place between the 15th and the 18th century CE. The 19th century CE witnessed a first shrinkage in size and a shift of the settlement towards the north, where the colonial-period town flourished. By the mid-20th century CE, most of the town was abandoned in favour of a new site a dozen km to the east, and only a small village remained in Old Buipe. Archaeological finds include artifacts of local and regional production (ceramic, clay smoking pipes, iron objects, etc.), as well as importation items like European clay smoking pipes and ceramics, cowries, and glass beads; with the notable exception of the glass beads, most of the importation items originates from 19th and 20th century CE contexts. A total of 18 glass beads found between 2015 and 2017 and coming from the earlier occupational phase were selected for techno-stylistic and chemical analysis [53]. Most of the beads are composed by multiple layers of glass for a total amount of 45 different glass samples analysed (Table 1).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g001\", \"pone.0318588.ref054\", \"pone.0318588.ref054\", \"pone.0318588.ref053\", \"pone.0318588.ref053\", \"pone.0318588.t001\"], \"section\": \"Alinguel, Senegal\", \"text\": \"Alinguel is a large settlement site located in the Fal\\u00e9m\\u00e9 River valley in Eastern Senegal (Fig 1). This site is mentioned in the 19th century travelogues of various travellers and trade agents as a hinge between the regions where the Boundou and Bambouk kingdoms were established [54]. The site was excavated in 2012 and 2013, leading to the identification of three settlement areas occupied during several centuries. Radiocarbon dating allowed to determine three occupational phases, the earlier one between the 1st and the 10th century CE, the middle one between the 11th and the 13th century CE, and the latter one between the 17th and the 19th century CE [54]. The extensive excavation of the site delivered a considerable number of potsherds, as well as metallic objects, lithic grinding tools, botanical macro-remains, and 227 beads made of glass (N = 208), ceramic (N = 10), stone (N = 4), copper (N = 2), iron (N = 1), bone (N = 1), and indeterminate material (N = 1). Glass beads were recovered from various chronological levels of the central area of the site: 7 were found in the layers dating from the middle phase, 178 from the recent occupational levels, whereas 23 come from the surface [53]. The beads from the 11th\\u201313th century context resulted to be displaced from the more recent layers and were therefore included in the 17th\\u201319th century assemblage [53]. A total of 207 beads were included in the techno-stylistic classification, one of the beads being too degraded to be identified. Moreover, 184 glass samples composing 129 beads were selected for chemical characterization; only the beads coming from the surface and the beads presenting an advanced chemical degradation were excluded from the analysis (Table 1).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g001\", \"pone.0318588.ref054\", \"pone.0318588.ref054\", \"pone.0318588.ref054\", \"pone.0318588.ref053\", \"pone.0318588.t001\"], \"section\": \"Toumbounto, Senegal\", \"text\": \"Toumbounto is a settlement site located on the left bank of the Fal\\u00e9m\\u00e9 River in Eastern Senegal (Fig 1). As for Alinguel, this site is mentioned in several written accounts of travellers, describing it as a remarkable village linked to the Boundou kingdom in the 19th century CE [54]. The site, excavated between 2012 and 2015, consists of 76 structures of various functions and one grave. Eight of these structures, namely three storage structures, two habitations, one structure linked to metallurgical activities, one grave, and one structure of unknown function, were extensively excavated [54]. Oral accounts and radiocarbon dating of several charcoal fragments and one calcinated seed led to the identification of one main occupational phase between the 17th and the 19th century CE linked to the architectural structures, as well as scant traces of a much older occupation around the 1st century CE [54]. The archaeological finds of various nature coming from the main occupational phase include 704 beads, 690 of which made of glass, 8 made of stone, 5 made of undetermined material, and 1 made of metal [53]. A total of 687 glass beads were studied and included in the techno-stylistic classification, 426 of which were selected to be chemically analysed. These beads are composed from one to four layers of glass, for a total amount of 561 glass samples analysed (Table 1).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g001\", \"pone.0318588.ref055\", \"pone.0318588.ref056\", \"pone.0318588.ref056\", \"pone.0318588.ref057\", \"pone.0318588.t001\"], \"section\": \"Farabana, Mali\", \"text\": \"The archaeological site of Farabana is located on the Malian side of the Fal\\u00e9m\\u00e9 River Valley, 10 km east of the border between Senegal and Mali (Fig 1). The written accounts of the Compagnie du S\\u00e9n\\u00e9gal, the French chartered company administering part of Senegal, mentioned the construction in 1724 of a fort near the village where the Bambouk king had his residence, precisely Farabana [55]. The excavation of the site in 2014 and 2016 led to the discovery of the ruins of this fort, which was used between 1724 and 1758 according to the texts and confirmed by the radiocarbon dating of six charcoal fragments [56: 151]. The archaeological finds include ceramics and smoking pipes of local production, cast-iron cannons, as well as surprisingly sporadic importation items, such as tableware, glass bottle fragments, and clay smoking pipes [56, 57]. Only four glass beads were recovered during excavation, one of which was found in surface and therefore excluded from the analysis. A total of 4 glass samples were chemically characterised (Table 1).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g001\", \"pone.0318588.ref058\", \"pone.0318588.ref053\", \"pone.0318588.t001\"], \"section\": \"Tyi-kun, Mali\", \"text\": \"Tyi is a large settlement site composed by several districts located on the top of the Bandiagara Escarpment in Mali (Fig 1). Three structures from two of the districts were excavated between 2008 and 2010, specifically a habitation in Tyi-jo and two open-air pottery firing structures in Tyi-jo and Tyi-kun. The study of the archaeological finds and the oral surveys among the people living in the present village of Tyi in the plain led to the conclusion that Tyi-jo was occupied between the end of 16th and the end of 17th century CE, whereas Tyi-kun between the 18th and mid-20th century CE [58]. A total of 13 beads were recovered during the surface survey and the excavation of Tyi-kun, 8 of which made of glass. Of these, only two were found in stratigraphic position and included in the techno-stylistic and chemical study [53]. A total of 4 glass samples were chemically analysed (Table 1).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref059\"], \"section\": \"\", \"text\": \"Shape: shape of the cross and longitudinal sections of the bead based on Horace Beck classification system [59].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref060\"], \"section\": \"\", \"text\": \"Colour: Munsell colour code of each glass composing the bead [60].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref061\", \"pone.0318588.ref062\"], \"section\": \"Chemical characterisation\", \"text\": \"The 798 glass samples composing 578 monochrome and polychrome beads selected for characterisation were analysed by Laser Ablation\\u2013Inductively Coupled Plasma\\u2013Mass Spectrometry (LA\\u2013ICP\\u2013MS) at the Institut de Recherche sur les Arch\\u00e9omat\\u00e9riaux (CEB-IRAMAT) of Orleans in France. A Thermo Fisher Scientific ELEMENT XR mass spectrometer was used for the analysis and the glass was sampled with a Resonetics M50E excimer laser working at 193 nm and operating at 5 mJ energy and 10 Hz pulse frequency. A 20 s pre-ablation was applied to reach the unaltered bulk glass, followed by 30 s of signal acquisition, corresponding to 10 mass scans. Setting the beam diameter between 50 and 100 \\u03bcm led to the creation of craters 150\\u2013250 \\u03bcm deep, invisible to the naked eye. The argon/helium flow carrying the sampled aerosol to the injector inlet of the plasma torch works at a rate of 1 l/min for Ar and 0.65 l/min for He. The signal in count-per-second was measured in low-resolution mode for 58 isotopes, from lithium to uranium. Standard Reference Materials from the National Institute of Standards and Technology (NIST SRM) 610 and the Corning reference glasses B, C, and D were used for external calibration. Reference materials Corning A and NIST SRM612 were analysed with the beads as unknown samples. The in-house archaeological sample APL1 was used for chlorine quantification and 28Si was used as an internal standard. The protocol detailed in [61] and [62] was followed to calculate the concentration of each element, with detection limits from 0.01% to 0.1% for major elements, and from 20 to 500 ppb for minor and trace elements.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g002\", \"pone.0318588.g003\", \"pone.0318588.s001\"], \"section\": \"Techno-stylistic classification\", \"text\": \"The macroscopic and microscopic analysis of 916 glass beads led to their classification in 63 different techno-stylistic types based on their morphology, optical properties, and manufacturing techniques (Figs 2 and 3). The beads were separated into 15 categories, namely monochrome beads of different colour (white, colourless, pink, red, green, blue, orange, yellow, and black), Cornaline d\\u2019Aleppo, Green Hearts, striped beads, eye beads, feather beads, and chevron beads. S1 Table lists the beads types\\u2019 morphological, optical, and manufacturing characteristics previously described.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.s002\"], \"section\": \"Chemical classification\", \"text\": \"The full chemical composition of the 798 glass samples composing the selected 578 monochrome and polychrome beads included in the techno-stylistic classification, as well as three stylistically unidentifiable fragments of beads, can be found in S2 Table.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g004\", \"pone.0318588.s003\"], \"section\": \"Chemical classification\", \"text\": \"Principal Component Analysis (PCA) was applied to the collection taking into account the oxides linked to the silica source (SiO2, Al2O3, CaO, Fe2O3), to the flux (K2O, MgO, CaO, P2O5, Cl, B2O3, PbO), and to the colouring agents (PbO, Fe2O3, MnO, CuO, As2O3, SnO2, Sb2O3) in order to identify the main chemical groups. Keeping in mind that some of the oxides can be associated to multiple ingredients, it was possible to differentiate between four main compositional groups, i.e. soda/soda-potash-lime silica glass (S/SPL), potash-lime silica glass (PL), lead and lead-alkali silica glass (divided into three clusters depending on the alkali content: lead silica\\u2013LS; lead-soda-potash\\u2013LSP; lead-potash\\u2013LP), and lead-soda silica glass of intermediate composition (LSS) (Fig 4). The main oxide average composition of the main chemical groups can be found in S3 Table.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.t002\", \"pone.0318588.g002\", \"pone.0318588.g003\", \"pone.0318588.s001\"], \"section\": \"Chemical classification\", \"text\": \"Table 2 lists the sub-groups of each main chemical groups indicated by the colour of the glass (blk: black, b: blue and turquoise, cls: colourless, g: green, o: orange; p: pink, r: red, w: white, y: yellow) and the chemical element characterising the colourant. The number of analysed glass samples for each sub-group, as well as the correlation between these groups and the techno-stylistic types of the corresponding beads are also indicated (cf. Figs 2 and 3, and S1 Table).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.s004\"], \"section\": \"Chemical classification\", \"text\": \"The average composition of each chemical sub-group is reported in S4 Table.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref004\", \"pone.0318588.ref063\", \"pone.0318588.ref064\"], \"section\": \"Monochrome beads\", \"text\": \"The only monochrome red bead of the collection was found in Toumbounto and it is composed by alumina-silica glass fluxed with soda, coloured by cadmium sulfoselenide, and containing 5 wt% of boron (group S/SPL r2, type 47). It is interesting to notice that a fragment of glass with the same composition was found in Dourou-Boro, an archaeological site not far from Tyi-kun in Mali, in a context dating from the 7th\\u20139th century CE [4]; given that this chromophore was used only after the 19th century CE [63, 64], this fragment was interpreted as the result of looting of the funerary site.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref049\", \"pone.0318588.ref065\", \"pone.0318588.ref067\", \"pone.0318588.ref068\", \"pone.0318588.g005\"], \"section\": \"Cornaline d\\u2019Aleppo\", \"text\": \"The ruby glass composing all Cornaline d\\u2019Aleppo analysed is lead-potash silica glass containing 8 to 15 wt% of potash and 3 to 7 wt% of soda. The colouring agent of this glass is colloidal gold in concentrations between 53 and 435 ppm (N = 50, group LP p, types 12 to 17). The technology for creating ruby red glass through metallic gold is ancient and involves the precipitation of Au(0) nanoparticles though addition of tin salts, iron oxides or, more recently, arsenic salts as reducing agents [49, 65\\u201367]. The high amount of arsenic present in this glass suggests the use of an arsenic compound together with gold to obtain the ruby colour, as described in Venetian recipes dating from the end of the 17th century [68]. It is possible to see a pattern in the amount of gold and lead oxide in relation with the typology of the beads. Type 12 shows in fact (with one exception) lower amount of PbO (11\\u00b12 wt%) and higher amount of Au (275\\u00b152 ppm), whereas types 13 to 17 show higher concentration of PbO (24\\u00b13 wt%) and lower concentration of Au (105\\u00b132 ppm) (Fig 5).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.s001\", \"pone.0318588.g006\"], \"section\": \"Green Hearts\", \"text\": \"The so-called Green Hearts are generally composed by two layers of glass, one opaque carmine red on the exterior and one transparent to translucent, colourless to green one on the interior. Few beads also show a transparent colourless glass layer on the external surface. The 88 Green Hearts found in Alinguel, Toumbounto, and Farabana can be divided into three types, from 18 to 20, based on their size and finish (S1 Table). A total of 163 glass samples were chemically analysed, 79 of which are green, 82 are red, and 7 are colourless. All glass samples, except the ones composing one bead from type 19, show the same base composition, i.e., soda-lime silica glass with around 13 wt% of soda, 3 wt% of potash, and 10 wt% of lime. The only factor changing is the amount of copper and iron oxides giving colour to the glass (Fig 6).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref049\", \"pone.0318588.ref067\", \"pone.0318588.ref069\", \"pone.0318588.ref070\", \"pone.0318588.ref071\", \"pone.0318588.ref067\", \"pone.0318588.ref071\"], \"section\": \"Green Hearts\", \"text\": \"The colourless layer of glass on the surface of the beads has negligible amount of copper and iron oxides (N = 7, group S/SPL cls, types 19, 20). On the other hand, the 82 red glass samples from group S/SPL r1 show higher amount of CuO (2.5\\u00b10.8 wt%) and Fe2O3 (2.7\\u00b10.8 wt%), whereas the 79 green glass samples from group S/SPL g contain 0.9\\u00b10.8 wt% of CuO, 1.6\\u00b10.4 wt% of Fe2O3, and varying amount of antimony oxide (between 0.1 and 3.2 wt%), in relation with the opacity of the glass. It is difficult to understand the nature of the glass chromophore without a molecular analysis of the glass. However, the different amount of copper and iron oxides could explain mechanisms of red glass colouration. As the green colour is generally given by CuO, the red colour could be either given by Cu2O or metallic Cu particles in suspension [49, 67, 69]. It has been proven that in red glass containing copper oxide in concentration higher than 5 wt% and high amount of lead oxide, the formation of Cu2O crystals is favoured [70, 71], which is not the case for the analysed samples. Besides, the high concentration of iron oxide in our samples can explain the precipitation of the metallic copper particles in the glass, as it could work as a reducing agent for the copper oxide [67, 71].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref072\"], \"section\": \"Feather beads\", \"text\": \"The feather beads decoration is made by winding rods of different glass colour around the bead core and then creating a pattern by dragging the trails in different directions [72].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref020\", \"pone.0318588.ref072\", \"pone.0318588.ref073\"], \"section\": \"Chevron beads\", \"text\": \"Chevron beads were created by blowing a mass of glass into a star-shaped mould, repeating the process layer after layer with different coloured glass. After drawing a glass cane and cutting the singular pieces, the layers were revealed by mechanical abrasion of the extremities of each bead [20, 72, 73].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.t002\", \"pone.0318588.g007\", \"pone.0318588.ref074\"], \"section\": \"Chevron beads\", \"text\": \"The four chevron beads found in Old Buipe show different number of layers and colours (types 60 to 63), like the very degraded red and white chevron bead not included in the techno-stylistic classification (n/a) (Table 2). Besides, the composition of the glass of each colour is the same of the ones composing the other beads from this archaeological site. The white layers of types 60, 62, 63, and \\u201cn/a\\u201d are made of lead-soda silica glass opacified by tin oxide (N = 4, group LSS w), whereas the white glass from type 61 is a soda-lime glass opacified by tin oxide (N = 1, group S/SPL w1); the red, blue, and colourless glass samples of all chevron beads have a soda/soda-potash-lime-silica base composition and they are coloured respectively with metallic copper (N = 6, group S/SPL r1), cobalt (N = 4, group S/SPL b1), and various amount of tin oxide depending on the opacity (N = 3, group S/SPL cls). All glasses show a positive linear correlation between Sn2O and PbO, indicating the probable use of lead-tin calx to opacify the matrix as for most of the other beads found in Old Buipe (see infra) (Fig 7) [74].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref018\", \"pone.0318588.ref019\", \"pone.0318588.ref019\", \"pone.0318588.ref021\", \"pone.0318588.ref019\", \"pone.0318588.ref020\", \"pone.0318588.ref021\", \"pone.0318588.ref075\", \"pone.0318588.ref076\", \"pone.0318588.ref047\", \"pone.0318588.ref072\", \"pone.0318588.ref077\", \"pone.0318588.ref078\", \"pone.0318588.ref076\", \"pone.0318588.ref020\"], \"section\": \"Hypothesis on the provenance of the beads\", \"text\": \"Many of the analysed beads coming from the excavated sites in Ghana, Senegal, and Mali with a specific typology can be associated to well-known European productions thanks to historical documents, travel accounts, and trade reports describing and classifying these items. The Cornaline d\\u2019Aleppo, also known as white hearts or Hudson Bay beads, for example, were produced from the beginning of the 19th century in Venice and can be found, among others, in the catalogues of the Levin Collection (1851\\u20131863), displaying the beads typologies produced specifically for the trade with West Africa [18, 19]. In these catalogues, it is also possible to find the Green Hearts, or pre-white hearts, exchanged since the 17th century [19, 21], which are also displayed in the sample cards of the Venetian society Societ\\u00e0 Veneziana per l\\u2019industria delle conterie and of the Amsterdam\\u2019s factory J.F. Sick & Co. dated from the end of the 19th and the beginning of the 20th century [19]. The samples cards from these two societies exhibited in addition a great variety of monochrome beads, also known as seed beads or rocailles, largely present in our collection. Chevron beads, also known as rosetta beads, were originally produced in Venice in the 15th and 16th century and later imitated by other production centres in the Netherlands, Germany, and in North America. All chevron beads patterns from our collection can be found in the catalogues of chevron beads for the West African trade created by John and Ruth Picard [20, 21], as well as in the private collection of Cesare Moretti [75]. In the Picard catalogues, it is also possible to find a good match for our blue multifaceted beads known as Russian Blues [76], produced originally in Bohemia after the 18th century CE [47, 72, 77, 78]. The eye beads and the feather bead from our collection can also be found in these catalogues (respectively on page 21 of [76] and page 3 and 32 of [20].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref007\", \"pone.0318588.ref013\", \"pone.0318588.ref017\", \"pone.0318588.ref046\", \"pone.0318588.ref047\", \"pone.0318588.ref079\", \"pone.0318588.ref080\", \"pone.0318588.ref081\", \"pone.0318588.ref086\", \"pone.0318588.ref087\", \"pone.0318588.ref095\", \"pone.0318588.ref010\", \"pone.0318588.ref096\", \"pone.0318588.ref100\", \"pone.0318588.ref048\", \"pone.0318588.ref074\", \"pone.0318588.ref075\", \"pone.0318588.ref101\", \"pone.0318588.ref103\"], \"section\": \"Hypothesis on the provenance of the beads\", \"text\": \"Many types of beads from our collection can also be matched to glass beads found in several archaeological sites in sub-Saharan West Africa [7\\u201313]. However, few of these assemblages were chemically analysed. The typological analysis of the beads alone is in fact not enough to pinpoint the production sites of the glass and the artifacts. If Venice and Murano were the main European production sites between the 13th and the 16th century CE, many expert artisans started to migrate towards northern Europe to create new production centres, while Venice still remained very productive [17, 46, 47, 79, 80]. The chemical analysis of the glass composing the beads might help distinguish between different productions dated 17th to 20th century CE, although it still remains a complicated task, as the glass recipes and raw materials used in different glasshouses were often extremely similar, if not the same. As of today, no systematic chemical analysis of the glass beads exhibited in the commercial catalogues has been done, so we can rely for the provenance study solely on the characterisation of glass artefacts and production waste found in archaeological sites in Europe [81\\u201386], North America [87\\u201395], and sub-Saharan Africa [10, 96\\u2013100]. Furthermore, the approximate chemical composition of various 15th\\u201320th century glass types calculated from multiple Venetian recipe books can be used for comparison [48, 74, 75, 101\\u2013103].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g004\", \"pone.0318588.t002\", \"pone.0318588.s003\"], \"section\": \"Hypothesis on the provenance of the beads\", \"text\": \"In order to evaluate the possible provenance of the glass samples, the four main compositional groups discussed above, namely soda/soda-potash-lime silica glass, potash-lime silica glass, lead and lead-alkali silica glass, and lead-soda silica glass of intermediate composition (Fig 4, Table 2, and S3 Table), were taken into account. To simplify the provenance study, some of the glass samples of intermediate composition were included in the discussion either of the soda/soda-potash-lime silica glass or the lead and lead-alkali silica glass.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref063\", \"pone.0318588.ref048\", \"pone.0318588.ref063\", \"pone.0318588.ref092\", \"pone.0318588.ref104\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"The white bead from type 2, the yellow bead from type 41, and the red bead from type 47 found in Toumbounto show very high levels of alumina (10.5\\u201313.7 wt%), suggesting that they are fa\\u00efence, sintered glass or Prosser-molded glass or ceramic beads probably produced after the 19th century [63]. This hypothesis is supported by the modern colourants detected, i.e. the Imperial Red (cadmium sulfoselenide), and the chromium yellow, used as glass chromophores in Venice starting from the second half of the 19th century [48, 63, 92, 104]. Since no classical opacifying agents (e.g. Sb, Sn, or As compounds) were detected, the opacity of these beads is probably due to the presence of unmelted minerals in the glass matrix, although this should be confirmed by the molecular analysis of the glass.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g008\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"The other alkali-lime silica glass samples can be further divided in two subgroups in relation to the tin and lead oxide contents (Fig 8).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref074\", \"pone.0318588.ref084\", \"pone.0318588.ref089\", \"pone.0318588.ref092\", \"pone.0318588.ref093\", \"pone.0318588.ref093\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"The majority of the glass samples composing Old Buipe beads shows a positive correlation between lead and tin, suggesting a coherence in the glass supply and indicating the use of the so-called calce di stagno e piombo (lead-tin calx) for opacification, a method used by Venetian glassmakers until the 19th century [74]. Moreover, the high Sn/Pb content in the glass due to the use of tin as opacifier suggests a production predating 1650, after which the use of antimony as an opacifier became systematic [84]. This is coherent with the results of the study of North American archaeological bead assemblages of probable Dutch origin, for which the tin content is used as a chronological marker in opposition to antimony and arsenic [89, 92, 93]. The composition of Old Buipe glass is comparable to the chemical group 2 found in the North American sites of Petun and Seneca, although with lower levels of Na2O and CaO [93].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref075\", \"pone.0318588.ref082\", \"pone.0318588.ref105\", \"pone.0318588.g009\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"Old Buipe glass samples have the typical composition of tin-opacified lattimo glass produced in Venice between the 15th and 17th century [75]. The comparison with the composition of glass assemblages from European archaeological sites and museum collections shows that they are comparable with the glass beads found in the late 17th century Hammersmith Embankment site in London (chemical group 1) and the glasswork waste from the contemporary Kg10 production site in Amsterdam [82, 105] (Fig 9).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref082\", \"pone.0318588.ref105\", \"pone.0318588.ref106\", \"pone.0318588.ref103\", \"pone.0318588.ref075\"], \"section\": \"Old Buipe glass.\", \"text\": \"Comparison between Old Buipe glass composition and the composition of glass from Amsterdam and London [82, 105], Old Broad Street, London [106], Louvre [103], and Venice [75].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref083\", \"pone.0318588.ref084\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"Moreover, compositional and typological similarities can be found between Old Buipe chevron beads and the chevron beads found in the 16th\\u201319th century glasswork site of Neularten, Germany, probably imported from Venice or Amsterdam [83, 84].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g010\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"Concerning the soda and soda-potash-lime silica glass beads from Alinguel, Toumbounto, and Farabana, the composition of contemporary European and African archaeological glass and glass beads assemblages can be used to understand the possible origin of these beads (Fig 10).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref082\", \"pone.0318588.ref086\", \"pone.0318588.ref010\", \"pone.0318588.ref103\", \"pone.0318588.ref085\", \"pone.0318588.ref081\", \"pone.0318588.ref101\", \"pone.0318588.ref098\", \"pone.0318588.ref099\"], \"section\": \"Soda and soda-potash-lime silica glass beads from Alinguel, Toumbounto, and Farabana.\", \"text\": \"Comparison between mixed alkali and soda-lime glass composition and the composition of glass from Amsterdam and London [82], Antwerp, Mechelen [86], Garumele [10], Louvre [103], Middleburg [85], Rouen [81], Venice [101], and Kulumbimbi and Lumbu, Angola [98, 99]. For Venetian glass: C = Cristallo, VB = Vitrum Blanchum.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref101\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"The base glass of our assemblage can be compared to most archaeological European glass assemblages for which a Venetian production is suggested, as well as to the Cristallo and Vitrum Blanchum glass produced in Venice between the 15th and the 17th century [101]. Some of our beads, namely few turquoise beads from type 21 and the red and green glass composing some Green Hearts from type 18 and 19, have lower potash, magnesia, lime, and alumina content, much like Cristallo glass produced with purer sources of silica. However, unlike Cristallo, they contain higher amount of phosphorus, chlorine, and iron, especially the Green Hearts.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref086\", \"pone.0318588.ref046\", \"pone.0318588.ref107\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"The other glass samples, more similar to the Vitrum Blanchum references, show higher amount of potash, much like the glass samples from 17th century Mechelen and Antwerp, in Belgium [86]. Here, to explain this difference in K2O content, the authors suggest a production in the Low Countries instead of Venice, which is supposedly supported by the trace element content. Unfortunately, the trace element concentrations were not published, so it is not possible to compare them to the composition of our samples. However, a higher potash content is typical of Venetian glass produced between the 16th and 18th century adding to the glass matrix the so-called grepola, the deposit of the bottom of the wine barrel (potassium bitartrate). The higher potassium oxide content could be also explained by the addition of tartar (K-bearer) to the glass matrix, practice more and more common from the 17th century onward [46, 107].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g010\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"It is interesting to notice that techno-morphologically consistent soda-lime glass beads from Garumele, the only contemporary assemblage from a West African site to be chemically analysed, show lower concentrations of soda and lime, as well as higher concentration of potash and alumina compared to our beads, so a different production site is probable (Fig 10).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref094\", \"pone.0318588.ref087\", \"pone.0318588.ref089\", \"pone.0318588.ref093\", \"pone.0318588.ref089\", \"pone.0318588.ref092\", \"pone.0318588.ref094\"], \"section\": \"Soda/soda-potash-lime silica glass\", \"text\": \"Moreover, comparing our assemblages to northern American assemblages, a parallel can be found with the soda-lime beads from 18th\\u201319th century Sullivand\\u2019s Island attributed to the Venetian production, although the latter show in average lower alumina content [94]. Even though beads from other North American sites [87, 89\\u201393] show lower soda and lime content and higher potash and alumina content, like Garumele beads, the opacifier content, especially for white beads, can be used as an indicator of possible production. As already mentioned, tin, antimony, and arsenic were used as chronological markers for North American bead assemblages of possible Venetian or Dutch production, dating antimony opacified beads between the 17th and the 19th century [89, 92]. Considering that Amsterdam stopped to be an important glass beads production centre after the beginning of the 18th century [94], we can suggest that the mixed alkali and soda-lime beads from our assemblage are of Venetian manufacture, probably dating between the late 17th and the 19th century.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref064\", \"pone.0318588.ref068\", \"pone.0318588.ref094\", \"pone.0318588.ref088\", \"pone.0318588.ref094\", \"pone.0318588.ref021\", \"pone.0318588.ref064\"], \"section\": \"Potash-lime silica glass\", \"text\": \"This group includes the glass composing the colourless monochrome beads (type 11), the blue monochrome beads from types 31 and 33, and one blue glass sample not included in the techno-stylistic analysis (TO-29-80-3b). All beads were found in Alinguel and Toumbounto, Senegal. Their high content in potash and phosphorus corresponds to the use of forest wood ash as a flux, typical of the 18th\\u201319th century Bohemian glass production [64, 68, 94]. Typologically and chemically consistent beads were found in North American [88, 94], European, and West African [21, 64] sites dating after the 18th century.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.s005\", \"pone.0318588.ref099\", \"pone.0318588.ref108\", \"pone.0318588.ref109\", \"pone.0318588.ref110\", \"pone.0318588.ref112\"], \"section\": \"Potash-lime silica glass\", \"text\": \"The colour of the blue beads is given by cobalt (233\\u2013624 ppm), which can be correlated to the presence of arsenic (928\\u20133875 ppm), nickel (60\\u2013508 ppm), bismuth (50\\u2013576 ppm), and silver (0.3\\u20133.6 ppm), as well as uranium in type 33 (S1 Fig). This is in line with the composition of the minerals extracted from the Erzgebirge mountain range, in Germany, one of the main sources of cobalt pigments for European glass and glaze production from the 14th century onwards, and especially between the 17th and 19th century [99, 108, 109]. Other Europeans ores showing similar mineral associations (e.g. France, Sweden, and Bohemia) can\\u2019t however be excluded [110\\u2013112].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g011\"], \"section\": \"Lead and lead-alkali glass\", \"text\": \"The study of the potential provenance of the glass samples from this group was done by comparing the chemical composition to the Venetian glass recipe books, as well as the lead and mixed-alkali lead glass found in various European, African, and North American archaeological sites (Fig 11).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref082\", \"pone.0318588.ref010\", \"pone.0318588.ref103\", \"pone.0318588.ref102\", \"pone.0318588.ref075\", \"pone.0318588.ref081\", \"pone.0318588.ref100\"], \"section\": \"Lead and mixed-alkali lead glass composition.\", \"text\": \"Comparison between lead and mixed-alkali lead glass composition and the composition of glass from Amsterdam [82], Garumele [10], Louvre [103], MET [102], Murano [75], Rouen [81], and Southern Africa [100].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref010\", \"pone.0318588.ref103\", \"pone.0318588.ref075\"], \"section\": \"Lead and lead-alkali glass\", \"text\": \"Orange and green glass show a composition similar to the glass of the same colours found in Garumele, Niger, of probable Venetian manufacture and suggested to be dated between the late 17th and the 19th century [10]. Moreover, the comparable Nb, Zr, and Sr content indicates that a similar source of raw materials was probably used for manufacture. Similarities can be found also with the composition of the yellow and green enamels of Venetian glass from Louvre [103], showing however higher amount of tin and antimony oxides, and slighter higher soda content. Finally, the composition of the yellow glass can be compared to the theorical composition of the Venetian anime that were mixed with transparent glass to create yellow opaque glass in the 19th and 20th century [75].\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref075\", \"pone.0318588.ref010\", \"pone.0318588.ref094\", \"pone.0318588.ref094\", \"pone.0318588.ref099\", \"pone.0318588.ref068\", \"pone.0318588.ref094\"], \"section\": \"Lead and lead-alkali glass\", \"text\": \"Arsenic opacified white glass is mentioned in the Venetian recipe books starting from the 18th century CE. The arsenic white monochrome beads from our collection can be compared to the opaque Girasole and Smalto white glass from Murano mentioned by Moretti and Hrelich [75], as well as to the white wound beads found in Garumele [10], and to the white beads from late 18th\\u201319th century Sullivan\\u2019s Island site, in northern America, of probable Venetian origin [94]. On the other hand, the arsenic white glass composing our Cornaline d\\u2019Aleppo (as well as the ruby red glass from the same beads) has a very different composition from the red-on-white beads from Sullivan\\u2019s Island, suggested to be of Bohemian origin [94]. Correlations can be found with the composition of the red-on-white beads found in Angola, although the gold content of the ruby red glass is much lower than in our beads [99]. Although no systematic chemical analysis of red-on-white glass beads has been found in literature, the typical composition calculated from the Venetian recipe books [68, 94] is in fact comparable to our Cornaline d\\u2019Aleppo\\u2019s composition.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.t002\"], \"section\": \"Lead-soda silica glass of intermediate composition\", \"text\": \"This group comprises one Green Heart (TO-2-6), the three blue beads from type 24, and one of the green monochrome beads from type 34. The other samples included in this group as a result of PCA (Table 2) were included in the provenance discussion of the other groups.\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.ref048\"], \"section\": \"Lead-soda silica glass of intermediate composition\", \"text\": \"Both glass composing sample TO-2-6 contain higher soda and alumina concentration compared to the other Green Hearts and the green layer show traces of chromium oxide. This suggests that the bead was produced after the 19th century, when chromium green glass was produced in Venice [48]. It is also the case for the green bead from type 34 having similar base composition as the other green beads from the same type but showing a higher amount of soda and alumina and lower amount of lead, as well as a very high content of chromium oxide (0.3 wt%).\"}, {\"pmc\": \"PMC11809889\", \"pmid\": \"39928606\", \"reference_ids\": [\"pone.0318588.g002\", \"pone.0318588.g003\"], \"section\": \"Techno-stylistic classification of the beads.\", \"text\": \"Morphological, optical, and manufacturing characteristics of each type of beads showed in Figs 2 and 3. Abbreviations: Category: Mon.: Monochrome, C.A.: Cornaline d\\u2019Aleppo, G.H.: Green Heart. Site: AL: Alinguel, TO: Toumbounto, OB: Old Buipe, FA: Farabana, TK: Tyi-kun. Shape: IR: irregular. Colour: w: white, cls: colourless, p: pink; r: red, g: green, b: blue, o: orange; y: yellow, blk: black. Diaphaneity: OP: opaque, TL: translucid, TR: transparent. The + sign in the colour and diaphaneity sections refers to different glasses in the same layer of the bead. The number of layers, colour, and diaphaneity of glasses in brackets refer to some of the beads of the same type. Technique: D: drawing, W: wounding, M: moulding. n.d.: indeterminate.\"}]"
Metadata
"{}"