2.4.1 Glaze Types and Compositions

Table 2 provides a summary of representative compositions of transparent and opacified Islamic glazes from the eighth to sixteenth centuries CE. The high lead glazes (60–67 wt% PbO, < 2 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) were used for all the yellow glazes opacified with lead stannate, as well as for some white glazes opacified with tin oxide and some transparent glazes. Lead-alkali glazes (15–45 wt% PbO, 4–14 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) were used extensively for the white glazes opacified with tin oxide, as well for the transparent glazes. A small number of lead-alkali glazes with the lowest lead oxide contents are, perhaps, better termed alkali-lead glazes (4–5 wt% PbO, 10–12 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ). Finally, soda-lime glazes (10–22 wt% Na₂O, 5–12 wt% CaO+MgO) were also extensively used for the transparent glazes.

Following on from Roman and Byzantine potters (Armstrong et al. Reference Armstrong, Hatcher, Tite and d’archimbaud1997; Walton and Tite Reference Walton and Tite2010), Islamic potters again produced the transparent high lead glazes either by applying lead oxide (or some other lead compound) by itself to the surface of the pottery body, or by applying a mixture of lead oxide-plus-quartz. With both methods, a small amount of clay can be included in the glaze slurry, and the application can be to either unfired or biscuit fired bodies. A further variation is that the lead oxide-plus-quartz mixture can be applied either in the raw state or after pre-fritting. As shown by Hurst and Freestone (Reference Hurst and Freestone1996), the two primary glazing methods can be distinguished by subtracting the percentages of lead oxide and any intentionally added colorant (e.g., copper oxide) from the glaze composition, and renormalising the resulting composition to 100%, as shown in the following equation:

$C_i{}^{*}=\frac{C_{i}x100}{{\text{Σ}}_{\text{i+1}}{}^{\text{N}}{\text{C}}_{\text{i}}\text{–}\left({\text{C}}_{\text{PbO}}{\text{+ C}}_{\text{CuO}}\right)}$(2.4)

where C_i is the concentration of oxide i in the glaze and C_i* is the adjusted concentration of oxide i. This adjusted glaze composition is then compared with the composition of the body. When glazing was by the application of lead oxide by itself, the adjusted glaze and the body compositions should be the same. In contrast, when glazing was by the application of a lead oxide-plus-quartz mixture, the silica content of the adjusted glaze should be higher than that of the body, and the alumina and other oxide contents should be lower.

Similarly, following on from the Seleucid (c.310–250 BCE), Parthian (c.250 BCE–226 CE) and Sassanian (226–637 CE) ceramics produced in Mesopotamia (McCarthy Reference McCarthy1996), Islamic potters used soda-lime glazes produced from a soda-rich plant ash. However, soda-lime glazes appear to have been used much less frequently than high lead glazes. Instead, Islamic potters made extensive use of a hybrid lead-alkali glaze containing a very wide ranges of both lead oxide and alkali contents. One reason for extensive use by Islamic potters of glazes with a significant lead content is that a lead-tin calx is a key ingredient in the production of opaque white and yellow glazes, opacified respectively with tin oxide and lead stannate, which represents an important innovation introduced by Islamic potters (Section 2.1). Also, as a result, there are no tin-opacified soda-lime Islamic glazes.

In addition, high lead glazes have a number of advantages as compared to soda-lime glazes (Tite et al. Reference Tite, Freestone and Mason1998). First, because of the solubility of the alkali in water, the preparation and application of alkali glazes are more difficult than in the case of lead glazes. Second, a key advantage of lead glazes is that they provide a broad melting range. This results in a larger temperature window for firing the glaze, which is particularly significant in traditional ceramic kilns where temperature variations of up to 100°C could occur across different areas of the kiln. Further, the broad melting range also enables a more gradual transition from solid to liquid state during firing which allows the glaze to fully react and mature, resulting in smoother surfaces and better glaze bonding. In addition, lead glazes allow for better matching of thermal expansion with that of the body again reducing the risk of glaze ‘crazing’ during cooling; and their lower surface tension, as compared to that of alkali glazes, results in better ‘wetting’ properties which reduces the risk of glaze ‘crawling’. Finally, lead glazes exhibit greater optical brilliance which increases with increasing lead content.

One advantage of lead-alkali glazes is that, as compared to high lead glazes, they have a higher viscosity in the molten state, resulting in a reduced mobility of the colorant ions and therefore, less running and merging of the associated colours. In addition, one potential problem with all lead glazes is the risk of reduction of lead oxide to lead metal which results in blackening and blistering of the glaze, and this risk is reduced as the lead content of the glaze is reduced. A further possible influence on the choice between high lead and lead-alkali glazes is the relative availability and cost of lead and soda-rich plant ash, depending on prevailing political and economic circumstances.

2.4.2 Glaze Colorants

In the production of glazed ceramics from their beginnings in fourteenth century BCE in Mesopotamia (Paynter Reference Paynter, Shortland, Rehren and Freestone2008), through the Seleucid (c.310-250 BCE), Parthian (c.250 BCE–226 CE) and Sassanian (226–637 CE) periods again in Mesopotamia (McCarthy et al. Reference McCarthy1996), to the Roman and Byzantine periods (Armstrong et al. Reference Armstrong, Hatcher, Tite and d’archimbaud1997; Walton and Tite Reference Walton and Tite2010), only a copper blue-green colorant was used until the emergence of Islamic glazed ceramics in the eighth century CE. By adapting the techniques employed by Byzantine glass workers, Islamic potters progressively extended their range of glaze colorants, both through the use of additional colorant ions dissolved in the glaze, and by the introduction of particulate colorants.

The principal ion-based colorants used by Islamic potters were copper (Cu²⁺) for a range of blue-green colours, cobalt (Co²⁺) for an intense blue colour, and manganese (Mn³⁺) for a purple colour. In some cases, with the cobalt and manganese colorants, either the original mineral phases used as colorant are not entirely dissolved or new mineral phases precipitate. As a result, the colorant can have a particulate component as well providing ions in solution in the glaze.

The colour associated with copper ions depends first on the composition of the glaze in which it is dissolved. Thus, according to Weyl (Reference Weyl1951: 163–167), copper ions produce a blue colour in an alkali-lime glaze, a turquoise colour in lead-alkali glaze, and a green colour in a high lead glaze. According to Ligand theory, these changes are the result of the stretching of electron orbitals which results in a change of wavelength. In addition, the nature of green colour of the glaze changes when other colorant ions, such as cobalt, iron, and chromium, are dissolved in the same area as the copper ions.

The particulate colorants first used by Islamic potters were their newly introduced opacifiers (Section 3.1), lead stannate yellow and tin oxide white. In addition, lead antimonate reappears as a yellow opaque colorant in Egypt in tenth century (Section 4.1.1) from where it spread to Tunisia and Sicily (Section 5.1.2).

The two other particulate colorants used by Islamic potters are bole red, which consists of a glassy frit containing a very fine mixture of quartz and haematite particles, and chromite black. Bole red appears to have been occasionally used as a colorant in the production of Raqqa underglaze decorated ware in the thirteenth century CE (Section 4.3.3). However, bole red was extensively used in the decoration of the later Iznik ‘Rhodos’ ware from about 1560 CE onwards.

Chromite black is an oxide mineral of the spinel group which can be represented by a chemical formula of the type FeCr₂O₄, where varying amounts of magnesium can substitute for the iron, ultimately forming MgCr₂O₄; and varying amounts of aluminium can substitute for the chromium, ultimately becoming FeAl₂O₄ (i.e., hercynite). In addition, as first observed in Iznik ware glazes (Tite Reference Tite1989), chromite particles sometimes contain relatively significant amounts of copper, manganese, cobalt and nickel, which have diffused from the adjacent decorative colorants. Crushed, angular particles of chromite, which do not dissolve and diffuse within the glaze, were used to produce fine, black to dark-green linear underglaze decoration. As a refractory pigment, chromite was used to draw black to dark-green lines which prevented other coloured glazes from blending into each other.

Chromite was used as a pigment through much of the Islamic world. Its early application appears to be associated with slip-painted wares of Eastern Iran (Section 5.1.1), later extending to underglaze decorated (Section 5.2) and mina’i (Section 5.2.2) wares of Iran, as well as being used for Raqqa-type underglaze decorated wares of Syria (Section 4.3.3). Subsequently, its use is noted in Anatolia, first, by Seljuq potters (Section 6.1) and later, by Ottoman potters in the decoration of Miletus ware (Section 6.4) and Iznik ware (Section 6.5).

Two other decorative techniques used by Islamic potters were lustre decoration, as discussed in Section 2.2, and sgraffito decoration produced in Egypt (Section 4.1.2) in which incisions or carvings are made through an applied slip to reveal the contrasting colour of the underlying body.

In addition, the technique of cuerda seca, in which chromite was used to draw black to dark-green lines in order to prevent adjacent coloured glazes from blending into each other, began to be used in Iran in the Timurid period from the mid fourteenth century CE onwards, principally in the production of tiles rather than ceramic vessels (O’Kane Reference O’Kane, Bloom and Blair2011). However, being towards the end of the period, during which the majority of the ceramics considered in this paper date, the cuerda seca technique is not discussed further.

2.4.3 Inglaze, Overglaze, and Underglaze Methods of Decoration

The three methods used to decorate glazed ceramics are inglaze, overglaze (or onglaze), and underglaze decoration. The inglaze technique involves applying decoration onto the surface of an unfired glaze, and allowing it to mature simultaneously with the glaze during firing. Second, the overglaze technique entails adding decorations to pottery after it has been glazed and fired. Once the decoration is applied, the pottery undergoes a further firing. Lustreware stands as the earliest example of overglaze decoration in the Islamic world. Another type of Islamic overglaze decorated pottery is represented by the Persian enamelled mina’i wares (e.g., see Figure 7). Finally, the underglaze technique involves painting directly onto the ceramic body, which is subsequently covered by a transparent glaze and fired (e.g., see Figure 8). This technique represents the final significant technological innovation in medieval Islamic pottery, and was initially extensively practiced in Iran, Egypt, and Syria in the twelfth and thirteenth centuries CE, later spreading to Anatolia and China.

Figure 7 Backscattered SEM photomicrograph of a section through white glaze into the stonepaste body of a Persian mina’i sherd. The opaque white glaze is opacified by tin oxide particles, with a layer of iron-rich red enamel applied on top of the white glaze. The enamel shows as the darker part on the surface of the glaze. The stonepaste body appears to be highly vitrified (Victoria and Albert Museum, acc. no. 625.1868(6)).

Figure 8 Backscattered SEM photomicrograph of a section through transparent turquoise glaze into the stonepaste body of an underglaze painted sherd from Moshkin Tepe, although it does not appear to be a local production. The black chromite pigment is applied on the surface of the stonepaste body and covered with a transparent turquoise glaze.

3.1.1 Coptic Glazed Wares, Cream Wares, and Hijazi Wares

The term Coptic Glazed Ware (CGW) was first used by Rodziewicz (Reference Rodziewicz1976: 63–64) to designate an assemblage of bowls excavated in Kom el-Dikka, Alexandria, Egypt, that occurred immediately after late Roman levels. Further excavations have identified CGW in various sites across Egypt, Jordan, Palestine, and Syria (for a review see Matin et al. Reference Matin, Tite and Watson2018: 43). Petrographic studies have shown that the light orange-coloured bodies of CGW were made of kaolinitic clay from the Aswan region, Egypt (Matin et al. Reference Matin, Tite and Watson2018; Salinas et al. Reference Salinas, Pradell, Matin and Tite2019; Ting and Taxel Reference Ting and Taxel2020), which confirms earlier propositions of Egyptian manufacture by Rodziewicz (Reference Rodziewicz1976, Reference Rodziewicz1978, Reference Rodziewicz1983) and Whitcomb (Reference Whitcomb1989) based on stylistic forms. CGW are characterised by open dish forms with flat bases, copying styles of Late Roman fine wares (Figure 9a). The bodies are covered with a thin white/greyish slip (Figure 9b; see also, Ting and Taxel Reference Ting and Taxel2020) and either painted and covered with an overall high-lead translucent amber glaze, or decorated in discrete bands with high-lead opaque yellow, green, and/or brown glazes, with some areas left unglazed. The latter type of CGW, decorated with bands of opaque glazes, are significant in that they represent the first examples of the invention and use of tin-based opacification technology of ceramic glazes. Chemical analyses by Ting and Taxel (Reference Ting and Taxel2020: Table 4) and Matin et al. (Reference Matin, Tite and Watson2018: Table 7) have shown 55–75 wt% PbO and up to 2.5 wt% SnO₂ and negligible alkalis (< 0.3 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) in the opaque glazes. The glazes were opacified by the use of lead stannate type II (Pb(Sn,Si)O₃) in the glaze with the Pb/Sn ratios usually greater than 35, representing the calx composition (for the opacification technology see Section 2.1). In the yellow glazes, the lead stannate also acted as the yellow colorant, but in the green and brown glazes, copper oxide and manganese oxide were added respectively as the colorants. For the translucent amber glazes, iron oxide was used as the colorant. The subsequent developments of tin-based opacification of glazes are seen in the so-called Yellow Glazed Family wares (Section 3.1.2) and later in Samarra-type wares (Section 3.2.2).

(a) An example of a Coptic Glazed Ware

(Ashmolean Museum, acc. no. EA1974.48);

(b) a typical backscattered SEM photomicrograph of a section through opaque yellow glaze into the slip and body of a Coptic Glazed Ware.

Cream wares (also known as Mefjar ware after their discovery at Khirbat Mafjar near Jericho) were found at various archaeological sites in Syria, Palestine, and Jordan dated to the late eighth to ninth century and were suggested to have been produced in the Syria-Palestine region (see Walmsley Reference Walmsley, Villeneuve and Watson2001). They are characterised by pale cream bodies, often in the form of delicate, thin-walled jars, jugs, and water flasks with incised, appliqué, and moulded decoration. A variety labelled ‘Cream Splash ware’ is suggested by Whitcomb (Reference Whitcomb1991: 53) to have been a close descendant of CGW. Cream Splash wares are similar to CGW both in their open dish forms as well as in the styles of paintings applied under high-lead transparent glazes. Another derivative of CGW were the Hijazi ware, after excavation of assemblages and wasters of glazed wares throughout the Hijaz (Hamed Reference Hamed1988). Hijazi wares exhibit similar painting motifs to CGW, although also often show cross-hatching decoration applied on dark red-orange fabric which Whitcomb (Reference Whitcomb1991: 53) proposed may suggest later developments. Comprehensive compositional and mineralogical analysis of the bodies and glazes of Cream wares and Hijazi wares are still largely lacking which hinders further interpretations on provenance and use of raw materials.

3.1.2 Yellow Glazed Family

Further development and use of CGW decorative techniques is characterised in ceramics excavated throughout Syria and Anatolia, including for instance at Tell Aswad/Raqqa (Watson Reference Watson1999b), Al-Mina (Vorderstrasse Reference Vorderstrasse2005: 75–78), Qinnasrin (Whitcomb Reference Whitcomb1999: 81–83), Antioch (Waagé Reference Waagé1948), and Tarsus (Bagci Reference Bagci2016). The wares were labelled ‘Yellow Glaze Family’ (YGF) by Watson (Reference Watson1999b: 81) after his classification of the sherds from Tell Aswad/Raqqa, and are characterised by their broad-based bowl shapes and the reddish to pale yellow colour of their bodies (see Figure 4). Despite its name, the YGF encompasses not only opaque yellow glazed wares but the first examples of opaque white glazed wares, with green and/or brown decorations. Of all the ceramics found in the above archaeological sites, YGF wares make up one of the main types in the corpus; three-quarters of the excavated glazed sherds in Tell Aswad, for instance (Watson Reference Watson1999b).

YGF wares were made from calcareous clays containing 10–20 wt% CaO. The yellow glazes are of the very high lead type containing greater than 60 wt% PbO and negligible alkali (<2 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) and the white glazes of the high lead-alkali type containing 35–50 wt% PbO and 3–6 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ (Matin et al. Reference Matin, Tite and Watson2018). The yellow glazes were opacified by lead stannate (Pb(Sn,Si)O₃) crystals with the Pb/Sn ratios greater than 16, representing the calx composition, and the white glazes by tin oxide crystals with the Pb/Sn ratios varying between 2.5 and 9 (see Section 2.1). In the green and brown glazes, copper oxide and manganese oxide were used as colorants.

3.2.1 Monochrome Green- and Turquoise-Glazed Jars

Being one of the most long-lasting pottery traditions, the manufacture of monochrome green- and turquoise-glazed jars in southern Mesopotamia during the Parthian and Sasanian periods continued into the Umayyad and early Abbasid periods. Over the twelve centuries of their production (from third century BCE to tenth century CE), the jars were made using calcareous alluvial clays from Mesopotamia and covered by alkali glazes (see, e.g., Hill Reference Hill2006; Pace et al. Reference Pace, Bianco Prevot, Mirti and Venco Ricciardi2008). The glazes were coloured by dissolved copper but the difference in green and turquoise colours are not currently fully understood. Further subdivisions based on the appearance of glazes, body fabrics, and forms in different assemblages have also been suggested by archaeologists (see, e.g., Priestman Reference Priestman2005: 234–240; Kennet Reference Kennet2004; Boucharlat Reference Boucharlat and Finkbeiner1993). The jars show a considerable variation in sizes, usually from about 20 to 80 cm height, and are decorated by different techniques including appliqué, thumb impressed ridges, and incision. After the eighth century, the range of designs (Priestman Reference Priestman2016: 2–3) and the scale of production expanded, and the jars became a significant object of trade.

The main purpose of these jars was the storage and transport of foodstuffs – especially date and date-syrup, but also oils and cane sugar (Wright Reference Wright1984: 44). They were distributed widely for their contents throughout upper Mesopotamia, East Africa, the Persian Gulf, Southeast Asia, and the Far East (see, e.g., Priestman Reference Priestman2021). In China, archaeological remains of green glazed jars were found as early as the late Eastern Han period (25–220 CE) in a tomb in Hepu. In general, they were mostly associated with the ports of Yangzhou and Guangzhou, where Middle Eastern merchants and communities resided, or at Buddhist monasteries where the contents of the jars would have been valued for their medicinal or ritual purposes (Cheng Reference Cheng2016: 58–63 cited in Wen Reference Wen2018: 324).

By the tenth century, the manufacture of monochrome green- and turquoise-glazed jars went through a decline. The political and social turmoil of the tenth century and the weakening of the Abbasid empire may have been partially responsible. Harvey (Reference Harvey2021) argued, based on the study of early Islamic textual documents, that the prophet’s hadith/saying in which the use of green-glazed jars for the storage or fermentation of date-wine was prohibited may have contributed to such decline. The invention of white tin-opacified glazing technology and the development of the so-called Samarra-type wares may also have been a contributing reason, although green-glazed jars still continued for about a century after the proliferation of Samarra wares.

3.2.2 Samarra-Type Pottery

The Abbasid palatial city of Samarra, Iraq, was founded in 836 CE and abandoned in 884 CE. The excavation of the site before the First World War and the subsequent publication of ceramics by Sarre (Reference Sarre1925) marked a major turning point in the understanding of Islamic ceramics. The so-called ‘Samarra-type’ or ‘Samarra-horizon’ pottery, dated to the middle of ninth century, was characterised by exceptionally fine open bowl shapes, covered by opaque white glaze with a range of inglaze or overglaze decorations (see Northedge and Kennet Reference Northedge, Kennet and Grube1994: 23–25; Whitehouse Reference Whitehouse1979). In addition, the Samarra excavations unearthed an unprecedented number of imported Chinese stonewares of the Tang dynasty. The most typical shapes of Samarra-type wares, typically with shallow curved profiles and a rounded flared rim, were almost exact copies of Chinese wares. The vessels were entirely covered by a background opaque white glaze, imitating the smooth, white surfaces of imported Chinese wares. The white surfaces were then either left plain, or decorated with a range of new innovative techniques executed by local potters, including blue (and green)-on-white, green-splashed, or lustre overglaze decorations (Northedge and Kennet Reference Northedge, Kennet and Grube1994).

The bodies of Samarra-type wares are made of buff-coloured fine-grained Southern Mesopotamian calcareous clays (20–25 wt% CaO), hitherto widely designated as ‘Basra fabric’ (see, e.g., Mason Reference Mason2004; Hallett Reference Hallett2000). It appears, however, that the alluvial geology of the region is very homogenous for the purpose of petrofabric studies, and therefore similar clay types could have been used in various production centres over this vast geographic region. The petrographic fabrics and compositions of the bodies do not appear sufficiently distinctive to make secure suggestions concerning provenance (Matin et al. Reference Matin, Tite and Watson2018: 47; see also, Priestman Reference Priestman2011). In order to avoid any misinterpretation, the term ‘Southern Mesopotamian fabric’ is preferred here which refers to the alluvial plains of southern Iraq and Iran, rather than a specific site of production.

White glazes of Samarra-type pottery are of three compositional types and represent different compositions and microstructures: (1) Alkali glazes containing 12–15 wt% ${\text{Na}}_{\text{2}}{\text{O + K}}_{\text{2}}\text{O}$ and less than 1 wt% PbO. These glazes show a combination of air bubbles, partially reacted quartz and other mineral particles dispersed throughout the thickness but are not opacified by tin oxide particles. This glaze type was widely practiced in southern Mesopotamia during the pre-Islamic period (see, e.g., Paynter Reference Paynter2001). (2) Alkali-lead glazes containing 10–12 wt% ${\text{Na}}_{\text{2}}{\text{O+ K}}_{\text{2}}\text{O}$ and 5–15 wt% PbO. The glazes are opacified by tin oxide particles, with Pb/Sn ratios of approximately 1 (see Section 2.1), as well as containing air bubbles and quartz grains. (3) High lead-alkali type containing 35–50 wt% PbO and 3–6 wt% Na₂O+ K₂O. The glazes are again opacified by tin oxide with calx Pb/Sn ratios between 2.5 and 9.

3.2.3 Susa Assemblage

Another key archaeological assemblage of Samarra-type pottery was found in Susa, Iran (Kervran Reference Kervran1977, Reference Kevran1979; Guillermina et al. Reference Guillermina, Peli and Makariou2005). The pottery assemblage includes all the typical wares discussed above (i.e., blue-on-white, green splashed, and lusterwares) as well as a wide range of opaque yellow and white glazed wares. The significance of these latter wares is in that they indicate striking similarities to the earlier Coptic Glazed Wares and Yellow Glazed Family wares of the Eastern Mediterranean, particularly in their shape, design, and glazing techniques. Essentially in the form of shallow bowls with flat bases or low footrings and steep walls, the wares were entirely covered by opaque yellow and white glazes decorated with geometrical designs or with runs of green and/or brown from the rim. The composition of opaque yellow glazes of these Samarra-type wares are similar to those of Coptic and Yellow Glazed Family wares of Eastern Mediterranean, containing very high lead oxide contents (> 60 wt% PbO) and opacified by the proliferation of lead-stannate crystals. The green and brown glazes were again achieved using respectively copper and manganese oxides which have blended into underlying tin-opacified white glaze (Matin et al. Reference Matin, Tite and Watson2018).

The evidence from the Susa assemblage further reinforces the connections with the earlier ceramic inventions that had taken place in the Eastern Mediterranean. The eastwards transfer of technologies and styles from the Eastern Mediterranean to Mesopotamia may have been related to the movement of the Islamic capital from Damascus, Syria, to Baghdad, Iraq, after the overthrow of the Umayyads by the Abbasids in 762 CE, which may have encouraged the migration of craftspeople and technologies.

4.1.1 Polychrome Glazed Ceramic Production in Egypt

Following on from Coptic Glazed Ware (Section 3.1.1), polychrome glazed ceramics continued to be produced into the second half of the ninth century. These ceramics were previously occasionally referred to as the ‘Fayyumi wares’ after a substantial number of them were discovered in the Italian excavations in the Fayyum region, although no evidence of production or wasters have so far been found. Similar to the Coptic Glazed Wares, the polychrome glazed wares were made of pink bodies using the kaolinitic clay from Aswan and with comparable very high lead glazes (about 60 wt% PbO, < 1 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) (Salinas et al. Reference Salinas, Pradell, Matin and Tite2019). However, although these glazes contained some 1 wt% SnO₂, no particulate lead stannate or tin oxide were now detected (Salinas et al. Reference Salinas, Pradell, Matin and Tite2019).

Next, through into the tenth century, polychrome opaque glazed ceramics, with calcareous clay bodies (16–25 wt% CaO) and high lead glazes (40–60 wt% PbO, 2–6 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ), were produced in the Fustat region near to modern Cairo from the second half of 9^th century through into 10^th century (Salinas et al. Reference Salinas, Pradell, Matin and Tite2019). These ceramics may include as many as six different glaze colours (yellow, amber, green, turquoise, white, and brown) which are applied side by side forming colour bands occupying the whole surface of the ceramic. However, instead of lead stannate, the yellow and amber glazes are coloured and opacified by particulate lead antimonate together with higher amounts of dissolved iron as compared to the other glaze colours. The antimony oxide contents are in the range 1.0–1.7 wt% Sb₂O₃ for the amber glazes and 0.2–0.5 wt% Sb₂O₃ for the yellow. The iron oxide contents vary from 2.4 to 2.8 wt% FeO for the amber glazes, and 1.3–2.3 wt% FeO for the yellow glazes. The green glazes are coloured by dissolved copper (0.9–1.3 wt% CuO) and opacified with lead antimonate (0.2–0.9 wt% Sb₂O₃). Turquoise glazes are again coloured by dissolved copper (0.2–2.6 wt% CuO) but opacified with tin oxide, and the brown are coloured by dissolved manganese (1.5–4.1 wt% MnO).

Next, with the Fatimid conquest of Egypt in 969 CE and the foundation of Fustat as the capital of the Fatimid dynasty, calcareous clay bodies, high lead glazes, and lead antimonate and tin oxide opacifiers continued to be used. Initially, the ceramics were decorated with transparent yellow and green glazes, coloured respectively with dissolved iron and copper oxides, over a white tin-opacified background. These ceramics were similar in appearance to the earlier polychrome ceramics produced in Tunisia except that, in Tunisia, a tin-opacified white glaze was not used for the white background, as discussed in Section 5.1.

Subsequently, in Egypt, the ceramics were decorated with areas of lead antimonate opacified yellow glaze, together with areas of transparent green glaze, which spread over a substantial area of the white tin-opacified background.

4.1.2 Sgraffito

Sgraffito (or sgraffiato) wares are characterised by slip-painted earthenwares with incisions or carvings through the slip into the body, and covered with a transparent lead-based glaze. In Egypt, sgraffito wares appear to have developed from the local polychrome glazed wares (Bongianino Reference Bongianino2014: 33), characterised by porous pink, red, or buff clay bodies which are typically composed of Nile alluvium clay or mixtures with calcareous clay. The ceramics bodies are covered with a quartz-based slip (Mason and Keall Reference Mason and Keall1990), and subsequently glazed. A parallel-hatched incision is commonly seen on these wares.

Syrian sgraffito wares represent light orange to pink bodies depending on the amount of lime present. Based on analysis of a small number of samples by Mason Reference Mason2004, the high lead glaze appears to contain 60–65 wt% PbO and is usually coloured amber/yellow by iron oxide or green by copper oxide. The glaze is sometimes splashed with copper-oxide green, iron-oxide amber/brown, and manganese-oxide purple/dark brown.

4.1.3 Lustrewares

The early phase of Egyptian lustreware production is represented by earthenwares with tin-opacified glazes and showed close stylistic links with the Abbasid monochrome lustrewares of the Samarra-type pottery (Philon Reference Philon1980). Based on the earliest pieces with datable inscriptions, it appears that production had begun by the year 1000 CE.

The earthenware body was typically composed of Nile alluvium clay mixed with calcareous clay (approx. 20 wt% CaO, 10 wt% Al₂O₃). Through experimentation, it appears that significant amounts of quartz were added to the otherwise clay body, resulting in a transition towards early stonepaste bodies that contained only around 30% clay (discussed in Section 2.3). The glazes were of the lead-alkali compositional type (25–35 wt% PbO, 5–9 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) and opacified by typically 2.5–12 wt% SnO₂ (Matin and Ownby Reference Matin and Ownby2023). Based on microprobe analysis of lustre layers in six samples, Pradell et al. (Reference Pradell, Molera, Smith and Tite2008: 2652) showed that the lustre layers on some of the samples are silver-rich with minimal copper (2–2.9 wt% Ag, 0.2 wt% Cu) others are made of a combination of copper and silver (1 – 1.9 wt% Cu, 1.8–5.4 wt % Ag).

4.1.4 Turquoise-Glazed Wares

Found at various sites across the Euphrates from Syria down to Northern Iraq, including Qal’at Ja’bar, Qasr al-Hayr al-Sharqi, and ‘Ana. Turquoise-glazed wares appear to have been produced between the twelfth and fourteenth centuries and were most probably introduced contemporaneously to the Syrian stonepaste Tell-Minis wares (see Section 4.3). The buff to pink coloured body was made of calcareous clay with mixed sand, which may have been naturally present or deliberately added as a temper. They are noticeably distinct in form from the pre-Islamic and early Islamic turquoise glazed jars of southern Mesopotamia, discussed in Section 3.2.1, and open-form tablewares are often discovered, although closed forms are also present.

4.2.1 Lustrewares

The lustre-decorated stonepaste wares of Egypt represent striking designs, typically featuring central figural motifs either drawn in lustre on a plain background or in reverse on a lustre background. A wide range of lustre colours were used, most commonly yellow, green, and brown, derived from different copper and silver amounts (Pradell et al. Reference Pradell, Molera, Smith and Tite2008; see also Section 2). The stonepaste bodies are composed of around 85 wt% SiO₂, 4–4.5 wt% Al₂O₃ and 2.5–5 wt% CaO and as discussed above exhibit a range of petrofabrics with minimal to extensive vitrification (approximately 2–4 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ). The glazes are of the lead-alkali compositional types containing 20–35 wt% PbO and 6–11 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ which were opacified with 5.5–19 wt% SnO₂, sometimes coloured green/turquoise by copper oxide, or blue by cobalt oxide (Mason Reference Mason2004; Matin and Ownby Reference Matin and Ownby2023). The wide range of tin oxide content observed in the ceramics serves as evidence of the diverse calx recipes used, further reinforcing the notion of a dynamic and experimental phase in ceramic production during this period in Egypt (see Section 2.1; see also, Matin Reference Matin2019: 1160). In contrast, glazes associated with the whiter, finer, and harder stonepaste bodies are colourless and transparent containing approximately 28–35 wt% PbO and 6–8 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ (Matin and Ownby Reference Matin and Ownby2023).

4.2.2 Fustat Fatimid Sgraffito (FFS)

Although a confusing label, the term FFS is used in the literature to refer to incised and carved stonepaste wares representing use of monochrome glazes (Bongianino Reference Bongianino2014). Until the comprehensive work by Bongianino (Reference Bongianino2014, Reference Bongianino2015, Reference Bongianino2017), there has been limited research on this common type of pottery in Egypt. It appears that FFS were produced in response to the high demand for imported Chinese celadons and porcelains in Fatimid Egypt. The potters developed white stonepaste bodies with visible incisions and carvings, covered by a thin transparent glaze. These ceramics provided affordable alternatives in a similar style, incorporating Chinese elements while also offering a fresh interpretation within the FFS tradition (Bongianino Reference Bongianino2014: 40–45). The range of stonepaste bodies appear similar to those of lustrewares although sherds of FFS have been analysed only sporadically and a more comprehensive study is necessary to comprehend the full range of variations. The composition of the glazes is similar to transparent glazes of lustrewares and the most common colours are honey/brown by iron oxide, green by copper oxide, and blue by cobalt oxide (Matin and Ownby Reference Matin and Ownby2023).

4.3.1 Tell Minis Ware

‘Tell Minis’ ware represents the earliest production of stonepaste in Syria and is known in western literature after a village in north-west Syria where an assemblage of pottery was claimed to have been recovered and offered to the British Museum (acquired by the Victoria and Albert Museum and the David Collection; see Porter and Watson Reference Watson1987). It is characterised by its remarkable thinness and fineness and its white, compact stonepaste body. It appears to be the product of the eleventh and first half of the twelfth century and has been found in excavations at several sites across the Euphrates (for a review see Tonghini Reference Tonghini1998: 42–43) as well as various public and private collections.

Analysis of Tell Minis wares from Qal’at Ja’bar indicates that the stonepaste bodies are as expected made mainly of silica at around 85 wt% SiO₂. The chemical composition includes 5–7 wt% Al₂O₃, 2–5 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ , and only minor (0.9–3.5 wt%) CaO which represents the nature of the clay used. The wares are decorated with one of the three techniques each with or without incised and carved motifs: monochrome glazed, in-glaze painted, or lustre decorated. The glazes are of the lead-alkali type including 18–30 wt% PbO and 10–15 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ (Tonghini Reference Tonghini1998: 91). Most glazes are transparent, but Porter and Watson (Reference Watson1987: 181) reported that some lustre glazes were opacified by tin oxide and noted one non-lustred turquoise glaze that was tin-opacified. Monochrome glazes and the glazes associated with lustrewares can be either colourless or coloured. Turquoise colour is achieved by adding 1–3 wt% CuO, blue by 0.4–1.5 wt% CoO, and purple by adding as high as 20 wt% MnO. The polychrome in-glaze painted wares are similarly blue, turquoise, or purple, and always applied to a colourless glaze. For the lustre-decorated wares, the lustre survives only in traces and is typically of a pale gold-green colour with green tint (Tonghini Reference Tonghini1998: 39, 91).

4.3.2 Intermediate Ware

While chemically and petrographically similar to Tell-Minis ware, the Intermediate wares represent the more friable body and coarser quality of Raqqa wares to the extent that the two types are visually indistinguishable and can be differentiated only based on laboratory examination (see Tonghini Reference Tonghini1998: 42; ‘Related wares’ in Porter and Watson Reference Watson1987: 186; ‘Groupe intemediare’ in Poulsen Reference Poulsen, Rus, Poulsen and Hammershaimb1957). Chronologically, it appears that the Intermediate ware overlapped Tell-Minis wares (or were introduced shortly after) but were produced for a longer period and ceased only shortly after the production of Raqqa wares. Tonghini (Reference Tonghini1998: 46) suggested that potters may have turned to producing Intermediate wares as a response to the marked rising-demand for stonepaste pottery. Their focus may have been more on increasing production rates or minimising fuel costs by firing wares at lower temperatures, rather than prioritising quality.

4.3.3 Raqqa Ware

The most widely discussed type of stonepaste pottery from Syria, the so-called ‘Raqqa wares’ reached the European Art market in the late nineteenth century and were labelled after their most probable provenance. Later archaeological excavations showed that Raqqa was indeed a ceramic production centre but that the so-called Raqqa wares were also produced at a number of other sites across the upper Euphrates and probably in other areas of Syria, such as Damascus (Tonghini Reference Tonghini1998: 50; Jenkins-Madina Reference Jenkins-Madina2006). On the basis of historical and stratigraphic evidence, the production of Raqqa wares began in the second half of the twelfth century and had fully developed by the thirteenth century. Tonghini (Reference Tonghini1998: 49) suggested that while certain kilns seem to have been destroyed during the Mongol invasions of the thirteenth century, other production centres continued to function but eventually by the fourteenth century Raqqa wares had completely disappeared. Compared to Tell-Minis and Intermediate wares, Raqqa wares were more widely distributed across and beyond the Levant region. This was likely due in part to the prosperous economy of the Ayyubid period and the expansion of trade routes.

Raqqa wares are characterised by their rather rough, friable and porous bodies which stands in contrast with the dense and hard bodies of the preceding Tell Minis wares. Visual examination and chemical analysis of Raqqa wares from Qal’at Ja’bar indicate that the stonepaste bodies are made predominantly of quartz at around 82–90 wt% SiO₂ with very low amounts of clay indicated by 3–5 wt% CaO, 1.5–4 wt% MgO, and 2–4 wt% Al₂O₃. The ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ contents range between 2 and 5 wt% (Tonghini Reference Tonghini1998: 90). Chemical analyses of a set of sixty-eight samples of Raqqa stonepaste from museum collections by Smith (Reference Smith and Jenkins-Madina2006) suggested a typical composition of 75 wt% SiO₂, 2–5 wt% Al₂O₃, and just over 3 wt% Na₂O. Only a sub-group of seven objects represented higher silica contents at approximately 80 wt% SiO₂.

The range of decorations varies from plain monochrome glazed to lusterwares. Raqqa wares also encompass a range of underglaze decoration techniques characterised by direct painting on the body, followed by the application of a colourless or coloured (turquoise, blue, or purple) transparent glaze (as discussed in Section 2.4.3). These include black-painted decoration under a transparent turquoise glaze, as well as black or black-and-blue painted designs under a transparent glaze. Polychrome underglaze-painted wares, often referred to as ‘Rusafa wares’, derive their name from the city located southwest of Raqqa. However, there is currently no evidence to designate Rusafa as the production centre for Syrian polychrome underglaze wares. These wares feature a variety of pigment colours including black, blue, turquoise, red, or purple, and exhibit significant figural and motif similarities with Persian polychrome overglaze enamel mina’i wares (Tonghini Reference Tonghini and Grube1994: 254; Watson Reference Watson2004: 294; see Section 5.2.5). They appear to have played a crucial role in the transmission of figural arts as well as coloured pigments between Iran and Anatolia during the twelfth/thirteenth centuries (see also Section 5).

The transparent glazes associated with the plain monochrome, lustre, and underglaze-painted wares are of the alkali-lime compositional type (14–22 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ). The plain monochrome glazes are either colourless, coloured turquoise by approximately 1–3 wt% CuO, blue by 0.7–0.9 wt% CoO, or purple by 4–6 wt% MnO. The most common underglaze-painted decoration was black achieved from chromium-based spinels (Franchi et al. Reference Franchi, Tonghini, Paloschi, Soldi and Vincenzini1995: 203). The black paint was sometimes paired with copper-based turquoise or cobalt-based blue underglaze painting. Red underglaze painting from bole pigments (a mixture of haematite and quartz; comparable to the one used on later Iznik wares discussed in Section 6.5) or dark red/purple from manganese oxide was added to this palette (Soustiel Reference Soustiel1985: 118). Raqqa lustrewares may also show underglaze painting in blue or turquoise, as also seen on Persian and occasionally Egyptian lustrewares. The lustre decoration represents a characteristic chocolate brown colour achieved my mixed copper and silver salts. Again, the glaze-body interaction layer, predominantly quartz and cristobalite, as well as some pyroxenes were found (Pradell et al. Reference Pradell, Molera, Smith and Tite2008: 2655).

4.3.4 Mamluk Ware

This type of stonepaste pottery represents the Mamluk period and was manufactured at least from the late thirteenth to fifteenth centuries. Mamluk wares are sometimes referred to as ‘Syrio-Egyptian’ as they were produced both in Syria and Egypt. Production workshops were found in Damascus and Fustat, and it is possible that other centres existed. Mamluk wares were widely distributed across Syria, Jordan, Palestine, although rarely found in Northern Syria. Tonghini (Reference Tonghini1998: 54) suggested that this scarcity may be attributed to the region’s low occupation after the Mongol invasion. They were also widely distributed in Europe perhaps as a result of the extensive network of trade facilitated by the Portuguese, Spanish, and Italian ships that sailed between North Africa and Europe.

Mamluk wares are characterised by their fine-grained, gritty, and porous stonepaste bodies. It appears that uncrushed rounded quartz was used together with crushed quartz (Tonghini Reference Tonghini1998: 51) but petrographic analysis distinguishing different type of body fabrics have not been conducted. Based on analysis of Mamluk wares from Qal’at Ja’bar the bodies consist of 84–89 wt% SiO₂, and a predominantly calcareous clay represented by 6–8 wt% CaO. Compared to the preceding Syrian stonepaste wares, the Mamluk wares contain lower amounts of alkalis at around 2–3 wt % ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ (Tonghini Reference Tonghini1998: 90).

The colourless, transparent glaze is often thickly applied and occasionally exhibits a green tint. The wares are either underglaze-painted with black and sometimes black and blue pigments from chromium spinels and cobalt oxides, respectively. Chemical analysis of glazes from Qal’at Ja’bar indicates that they are of the alkali-lime compositional type with approximately 17–19 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ (Tonghini Reference Tonghini1998: 90).

5.1.1 Slip-Painted Wares

The peak of creative innovation, encompassing both designs and pottery-making techniques, emerges in the slip-painted wares of the tenth century. These slip-painted wares, found in Eastern Iran and Central Asia, are covered with a white slip (or occasionally black or red slip) and feature thick, solidly painted designs applied over this underlying slip, all covered with a transparent lead glaze. This method of decoration distinguishes them from splashed slip-painted ceramics (Section 5.1.2), where white slips cover the entire body, decorated with colour splashes and sometimes incisions carved through the slip into the clay-based body.

Slip-painted pottery is found in a range of designs and qualities, with higher-quality examples originating from Nishapur, Samarqand, and Tashkent. Recent archaeometric analyses of slip-painted wares from hitherto little-known sites for instance in Bust and Lashkar-i Bazar in Afghanistan by Gulmini et al. (Reference Gulmini, Giannini, Lega, Manna and Mirti2013), southern Kazakhstan by Klesner et al. (Reference Klesner, MacDonald, Dussubieux, Akymbek and Vandiver2019), and Termez (southern Uzbekistan) by Molera et al. (Reference Molera, Martínez Ferreras and Fusaro2020) and Martínez Ferreras et al. (Reference Martínez Ferreras, Fusaro and Gurt Esparraguera2019) have expanded our views of their production in this region.

Slip-painted pottery presents several variations in its decoration. A type that has been particularly praised by collectors and museums were the white slip-painted wares, typically featuring epigraphic, animal, or abstract decorations primarily in black, although red is also occasionally used. These are often referred to as black on white wares. Occasionally, the colours are reversed, with the ware covered in a black slip and the decorations applied in white slip under a transparent glaze. Epigraphic wares with Kufic inscriptions, in particular, have garnered significant attention due to their elegant and modern designs. Other decorative styles involve the use of a variety of polychrome slips, yellow-staining black painted wares, and slip-painted imitations of Samarra-type pottery, especially lustre-imitation wares. These typically feature black, red, and/or olive-green decorations.

The production of slip-painted wares persisted, albeit with diminishing quality, during the eleventh and twelfth centuries. Concurrently, a series of sgraffito wares with incised patterns through a slip emerged across Iran and neighbouring regions in Northwestern Iran, extending eastwards into Afghanistan (see Section 4.1.2. for Egyptian and Syrian sgraffito wares). Common varieties include plain types featuring incised decoration through a white slip, occasionally decorated with copper oxide green decoration under a transparent lead glaze. Other types, displaying a more controlled use of colour, have been identified and named by dealers based on the towns where many were claimed to have been found, such as ‘Garrus’, ‘Amol’, and ‘Aghkhand’ wares.

5.1.2 Splashed Slip-Painted Wares

Splashed slip-painted wares, along with buff wares (Section 5.1.3), constitute some of the largest pottery groups discovered in Nishapur and have also been unearthed in Samarqand and other locations in the Eastern Iranian and Central Asian region. These ceramics are distinguished by their white slip coating decorated with splashes of copper green, iron brown, and manganese black under a transparent lead-based glaze. A variant of this ware is further decorated with incised designs.

The general compositions of the high lead glaze and pigments are similar to those found on slip-painted wares. Holakoeei et al. (Reference Holakooei, de Laprouse and Car.2019: 768) conducted an in-depth analysis of the manganese black pigment used in splashed wares from Nishapur and Samarqand. Based on the proportion of iron/manganese in the pigment, their findings suggest that the black pigment in the splashed wares was achieved using a manganese-rich mineral, with minimal iron content.

5.1.3 Buff Wares with Opaque Lead Stannate Yellow Decoration

Buff wares, named after the buff-coloured ceramic body, are a type of polychrome painted wares under a transparent lead glaze that became known following excavations conducted by the Metropolitan Museum of Art in Nishapur between 1935 and 1948 (Sardar Reference Sardar2015). They were the largest group of glazed pottery found in Nishapur (Wilkinson Reference Wilkinson1973: 3–53), but sherds of this type were also unearthed in Merv. A distinctive feature of buff wares is the prevalent use of lead-stannate opaque yellow pigment complemented by designs in black, green, and occasionally white slip, featuring geometric patterns or dense animated motifs. Iconographers have suggested that these animate motifs might represent local Persian festivities that had survived from the pre-Islamic Sasanian period. However, a more recent perspective by Siméon (Reference Siméon2017) proposes that they may instead depict rituals conducted by the army of the Buyids, which was a Persian Shia dynasty that gained influence in the tenth century, primarily in the western part of the Samanid Empire, which ruled over a predominantly Sunni population.

The production technology of buff wares also appears to exhibit influences from the western regions. Their prevalent use of lead-stannate yellow pigment, deep bowl shapes, and certain geometric and animal figures bear a resemblance to the Yellow Glazed Family wares of Syria (Section 3.1.2). An analysis of the yellow decoration on a single shard of Nishapur buff ware by Mason (Reference Mason2004: 135, sherd ROM.20) revealed a composition of 59.0 wt% PbO, 34.8 wt% SiO₂, and 1.1 wt% SnO₂. Non-invasive analyses of three Nishapur buff ware sherds by Holakoeei et al. (Reference Holakooei, de Laprouse and Car.2019: 766) confirmed the presence of lead-stannate (PbSnO₃) in the yellow as well as green pigments. Copper and iron were responsible for the green and red colours, respectively. A mixture of iron and manganese was used for black decorations (Holakoeei et al. Reference Holakooei, de Laprouse and Car.2019: 764).

5.1.4 Opaque Yellow Glazed Wares

The use of opaque yellow glaze was not unique to buffwares. Known as Opaque yellow glazed wares, these ceramics featured an opaque yellow glaze with green decoration. Wilkinson (Reference Wilkinson1973: 205) documents their production and use in Nishapur during the ninth and tenth centuries, though production seems to have ceased thereafter. Chemical analysis of three sherds from Nishapur, four sherds from Takht-i Sulaiman, and one sherd from Merv indicate that the yellow glazes were of the high lead type with approximately 65 wt% PbO, 26 wt% SiO₂, and 4 wt% SnO₂, with Pb/Sn calx ratios ranging between 10.5–28.3. The microstructure shows a proliferation of lead stannate crystals throughout the glaze. However, the decorations on the Takht-i Sulaiman and Merv sherds may not strictly adhere to those of the Nishapur opaque yellow glazed wares. The Merv sherds feature brown decorations, while the Takht-i Sulaiman sherds exhibit brown and white decorations alongside yellow and green ones.

The red clay body, which appears similar to that used for slip-painted and splashed slip-painted wares of Nishapur, contain 7–16 wt% CaO and 2.5–5 wt% MgO and is of a highly immature fabric (Matin Reference Matin2016: 69–76; Matin et al. Reference Matin, Tite and Watson2018: 61–64).

5.1.5 Opaque White Glazed Wares

Little known and rarely discussed, the opaque glazed ceramics of Samarqand feature entirely distinct decorative styles when compared to the earlier Samarra-type ceramics. Covered with opaque white glaze, they are decorated with designs predominantly in green and occasionally in purple/brown. The opaque white glazes employed belong to the lead-alkali category and contain approximately 8–10 wt% Na₂O+ K₂O and 15–19 wt% PbO (Matin et al. Reference Matin, Tite and Watson2018). These glazes achieve their opacity through the inclusion of 3.4–6.1 wt% SnO₂ (Pb/Sn ratios ranging from 3.7 to 5), along with the presence of quartz particles and air bubbles (Matin et al. Reference Matin, Tite and Watson2018: 61–64).

5.2.1 Lustre Ware

Three styles of Kashan lustre ware have been identified by Watson (Reference Watson1985a, Reference Watson1976) which appear to indicate the stages of its development. The earliest stage appears to be the ‘monumental style’ where the main figures were depicted in reverse, that is, instead of painting the main motifs in lustre, they were left white against a lustre background. Rather distinct from this style is the ‘miniature style’ with designs which as the name suggests were closely related to the manuscript illumination of the period, which is missing, but is seen to a large extent in the later manuscripts of fourteenth and fifteenth centuries. Watson (Reference Watson1976) has suggested that the ‘miniature style’, which is painted in lustre on a white background, was most probably developed primarily for the enamelled mina’i wares (Section 5.2.2). Finally, the ‘Kashan’ style, with designs in reverse on a lustre ground appeared to have been an amalgamation of the two styles (see, e.g., Figure 10a).

Figure 10

(a) An example of a Kashan Lustre Ware

(Ashmolean Museum, acc. no. EA1956.33),

(b) a typical backscattered SEM photomicrograph of a section through opaque white glaze (lead-alkali compositional type) into stonepaste body of a Kashan lustre ware.

The glazes were of two compositional types: lead-alkali type (12–24 wt% PbO, 5–13 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) opacified by typically 3.5–8 wt% SnO₂ (see, e.g., Figure 10b), and a soda-lime transparent type (17–20 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) (Kingery and Vandiver Reference Kingery and Vandiver1986: 116; Mason Reference Mason2004; Pradell et al. Reference Pradell, Molera, Smith and Tite2008). It appears that the soda-lime glazes were often used to glaze the reverse of dishes and interior of the closed vessels. The lustre layers appear to be a mixture of copper (2–5 wt% Cu) and silver (<5 wt% Ag), with copper contents significantly higher than those of Egyptian Fatimid lustres (Pradell et al. Reference Pradell, Molera, Smith and Tite2008: 2661; see also Section 5.1.3).

5.2.2 Mina’i Ware

The Persian term mina’i, adopted by dealers and collectors since the twentieth century, refers to overglaze enamelled ceramics from medieval Iran. In this technique, once the glazed vessel was fired it was painted with various enamels and fired for the second time. The second firing was performed at lower temperatures than the first, preventing the underlying glaze from softening, while still high enough to allow the enamels to fuse. Mina’i wares feature figural representations closely tied to contemporary Persian manuscript painting, even though the manuscripts themselves have not survived. They also feature geometric, floral, and epigraphic designs, coming in various sizes and shapes of vessel, with bowls being the most common. Due to the popularity of mina’i wares in the art market during the 1930s, unscrupulous dealers went to great lengths to assemble new complete vessels from unrelated sherds. They concealed these efforts through intensive overpainting and restoration. Therefore, caution is needed when dealing with museum complete pieces (see e.g., Norris and Watson Reference Norris and Watson2021; Watson Reference Watson1985b).

There seem to have been close connections between mina’i and lustre production, and pieces with both lustre and enamel decorations represent complex production processes (McClary Reference McClary2022). Kashan stands out as the most convincing centre for its production. Based on inscriptions on known pieces, mina’i wares was produced between 1180 and 1219 (Watson Reference Watson and Hillenbrand1994: 171). However, there are suggestions that the chronology needs to be reconsidered based on evidence of mina’i tiles in Konya dated as early as 1174 (McClary Reference McClary2016). Abu’l Qasim Kashani, in his 1301 CE treatise, asserts that this technique, referred to as haft-rang [seven-coloured], had ceased to be practiced. Recent evidence of mina’i-decorated sherds, identified as the Fatimid Fustat Sgraffito type based on the foot shapes and glazing manner, as presented by Watson (Reference Watson, Blair, Bloom and Williams2024), suggests that the mina’i technique may have travelled westward to Fustat. Additionally, it may have served as inspiration for figural images and motifs on Syrian Rusafa wares (see Section 4.3.3).

The base glazes are of the lead-alkali compositional type (14–22 wt% PbO and 7–12 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) and opacified by 6–10 wt% SnO₂ (see Figure 7). They are either white or coloured turquoise by addition of 1–2 wt% CuO as colorant (Mason et al. Reference Mason, Tite, Paynter and Salter2001; Mason Reference Mason2004: Table 6.5; Wen and Pollard Reference Wen and Pollard2014).

In some cases, applied relief with a composition similar to the stonepaste body was added over the base glaze. Cut gold leaf gilding was also sometimes applied using an organic binder after the first firing, before any of the enamels were applied. SEM microstructures often reveal bubbles under the gold leaf, likely caused by the burning of organic binders during the second firing (Holakooei et al. Reference Holakooei, Mishmastnehi, Moloodi Arani, Röhrs and Franke2023).

A wide range of colour decorations are used on mina’i wares. Koss et al. (Reference Koss, McCarthy, Chase, Smith, McCarthy, Chase, Cort, Douglas and Jett2009), in their chemical and microstructural analysis of a set of mina’i sherds, highlighted the presence of a range, rather than a singular mina’i technology. Notable variations in enamel compositions, the sequence of application, and firing regimes are evidence of this range, often correlating with stylistic distinctions.

Based on their microstructural studies, most enamel layers exhibit an even linear interface, with minimal interaction with the underlaying base glaze. This suggests that these enamels were likely applied after the initial firing. However, turquoise glaze, as well as certain varieties of blue and purple, appear to have been applied as an inglaze over the raw base glaze before firing.

Turquoise glaze was achieved by 1–2 wt% copper oxide. A range of blue were obtained by 0.1–6.5 wt% CoO, accompanied by 0.3–2.1 wt% FeO and 0.6–5.3 wt% As₂O₃. These blues vary in microstructure, chemistry, appearance, intensity of colours, and ways of application and have been categorised to different types by Koss et al. (Reference Koss, McCarthy, Chase, Smith, McCarthy, Chase, Cort, Douglas and Jett2009) and Wen and Pollard (Reference Wen and Pollard2014). Based on their analyses it appears that in some cases, the cobalt pigment was mixed with a mixture similar to that of the base glaze powder and applied, while in others, the cobalt pigment was processed into a frit (see also, Mason et al. Reference Mason, Tite, Paynter and Salter2001).

The main black pigment used on mina’i wares was chromite containing 1.2–4 wt% Cr₂O₃, accompanied by 0.6–3.6 wt% FeO and 1.2–3 wt% MgO. This pigment was used for drawings under the base glaze, as well as opaque, dense accent areas (such as facial features and hair), although the chromite used is more finely ground in the former case which allows for finer and thinner application of underdrawings. Another type of black was achieved by a mixture of manganese and iron oxides (0.5–6.5 wt% MnO and 3–8.1 wt% FeO).

The red pigment consisted of iron oxide (2.5–8.5 wt% FeO), sourced from hematite, and the pink appears to be a mixture of hematite and tin oxide (1.2 wt% FeO and 4.7 wt% SnO₂). The purple was achieved by manganese oxide (1–3 wt% MnO), and olive-green type with mottled yellow was found to consist of chromium, iron, and copper, although the specific chemical composition was not provided (see Koss et al. Reference Koss, McCarthy, Chase, Smith, McCarthy, Chase, Cort, Douglas and Jett2009: 43).

The archaeometric examination of pigments indicates a strikingly similar use of chromium-based pigments on mina’i wares as seen in earlier slip-painted wares from Eastern Iran. The use of chromite pigment for black and the identification of chromium in an olive-green pigment with mottled yellow on mina’i wares closely parallel the practices developed by earlier potters in Eastern Iran. Abu’l Qasim Kashani, in his treatise, mentioned that the chromite, known as muzarrad in medieval Persian, was procured from the mountains of Jajarm in Khurasan, Northeastern Iran. While it remains uncertain whether this was also the source for the chromium-based pigments in slip-painted wares, the fact that the mine(s) were situated in a similar region raises an intriguing possibility linked to the initial use of these pigments in Eastern Iranian wares.

The connections between mina’i and slip-painted wares extend beyond the use of chromite pigments; stylistic parallels are also apparent. Both feature epigraphic decorations around the rim and showcase large sizes, reaching up to approximately 50 cm in diameter. These shared characteristics suggest a continuity and further development of technologies and styles initially employed in Samanid wares, evolving into the twelfth-century Seljuq wares.

5.2.3 Lajvardina Ware

Lajvardina wares, adapted from mina’i ceramics, were produced during the thirteenth and fourteenth centuries within the Ilkhanid Period in Iran. The designs are non-figural, typically featuring geometric patterns. Reflecting the influence of the Mongol period in Iran, Chinese motifs such as phoenixes, lotuses, and peonies are commonly seen particularly on lajvardina tiles. These tiles were extensively used at Takht-i Sulaiman in northwestern Iran, serving as the summer palace for Abaqa Khan, the grandson of Ghengis Khan and the second Ilkhanid ruler. It appears that potters from Kashan were brought to the site to produce these tiles (Masuya Reference Masuya1997: 226).

Chemical analyses of lajvardina sherds from Takht-i Sulaiman, Sultaniya, Rayy, Yazd, and Kashan indicated that base glazes come in blue, achieved with 0.3–0.8 wt% CoO, accompanied by 0.3–1.2 wt% As₂O₃ and 1–2.3 wt% FeO), or turquoise, achieved with 1.4–3.2 wt% CuO. White base glazes, which from the data currently available, appear to be exclusive to lajvardina sherds from Takht-i Sulaiman, are achieved with 1.2–8.9 wt% SnO₂ (Holakooei et al. Reference Holakooei, Mishmastnehi, Moloodi Arani, Röhrs and Franke2023).

The glazes fall into three main compositional types: lead-alkali (8–38 wt% PbO and 10–14 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ) opacified by 5–11 wt% SnO₂, or alkali-lime (4–26 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ ; 2–8 wt% CaO) without opacification (Osete-Cortina et al. Reference Osete-Cortina, Doménech-Carbó, Doménech, Yusá-Marco and Ahmadi2010; Röhrs et al. Reference Röhrs, Dumazet, Kuntz and Franke2022; Holakooei et al. Reference Holakooei, Mishmastnehi, Moloodi Arani, Röhrs and Franke2023). Additionally, one sample appear to belong to the low lead-alkali compositional type (4 wt% PbO; 3.5 wt% K₂O, and unspecified levels of Na₂O), with 1.2 wt% SnO₂. The main distinguishing factor between glaze compositions is the presence or absence of lead-tin calx for opacification, categorising the final glaze composition as either (low) lead-alkali or alkali-lime. Based on analysed sherds, all turquoise base glazes fall under the lead-alkali type, blue glazes fall under both lead-alkali and alkali-lime, and white base glazes exhibit all three compositional types.

The gilding, along with the black, white, red enamels, and blue and turquoise inglazes, are similar to those observed on mina’i wares. Black enamels employed either chromite or manganese oxide. Analysis of chromite black pigments on sherds from Takht-i Sulaiman identified chromium and iron, but no manganese, indicating the sole use of iron-chromite. Tin oxide and iron oxide (from hematite) were used for white and red enamels, respectively. Similar to the mina’i wares, chromium-based olive green/yellow were also identified on lajvardina sherds from Takht-i Sulaiman and PbCrO₄ was also identified. Cobalt and copper oxides were introduced for blue and turquoise inglaze decorations respectively.

5.2.4 Interactions with China – Underglaze Blue-Painted Wares

At the core of the development of blue-and-white wares were two pivotal technologies that found their way from Iran to China. The first was the underglaze painting technology, and the second involved the use of cobalt blue pigment, specifically sourced from the cobalt mine in the village of Qamsar, situated southwest of Kashan (Matin and Pollard Reference Matin and Pollard2015, Reference Matin and Pollard2017). These two technologies converged in Jingdezhen, in eastern China, during the second quarter of the fourteenth century, giving rise to the production of exquisite porcelain ware with underglaze cobalt blue painting. Watson (Reference Watson2020: 347) draws attention to discarded pieces of early Yuan Chinese Blue-and-white porcelain bearing Persian verses, supposedly inscribed by a native Persian, as evidence supporting the idea that these techniques were transferred through direct personal contact.

The earliest Persian Chinoiserie underglaze blue-painted copies seem to have originated in Samarqand, potentially produced by potters forced to relocate from Iran and Syria to Samarqand by Timur (see, e.g., Watson Reference Watson1987). Mason (Reference Mason, Golombek, Mason and Bailey1996: 35, 112, 114) argues that the execution and motifs of these wares appears to follow imitations of blue-and-white Yuan porcelain excavated at Hama and Fustat. It appears, however, that the Syrian potters working in Samarqand would have needed to experiment with locally available materials, resulting in discernibly different stonepaste body fabrics from those in Syria. Nevertheless, they did share the use of quartz sand, as opposed to crushed quartz (Mason Reference Mason, Golombek, Mason and Bailey1996: 35).

It appears that potters subsequently began to migrate from Samarqand westwards and established workshops in cities like Nishapur, Mashhad, Tabriz, and others. Mason (Reference Mason, Golombek, Mason and Bailey1996) proposed various petrofabric groups based on petrographic analysis of samples tentatively or securely associated with these cities. Subsequently, Tite et al. (Reference Tite, Wolf and Mason2011) conducted chemical analyses to determine the composition of their stonepaste bodies. For instance, the ‘Nishapur petrofabric’ group, as identified by Mason (Reference Mason, Golombek, Mason and Bailey1996: 38, 42), included wasters and pottery from the kiln site at Nishapur, along with a dish inscribed with its place of manufacture as Nishapur. Chemical analysis by Tite et al. (Reference Tite, Wolf and Mason2011, 572) indicated that the stonepaste body consisted of 92 wt% SiO₂, 3 wt% Al₂O₃, 1.5 wt% CaO, and 2 wt% ${\text{Na}}_{\text{2}}{\text{O+K}}_{\text{2}}\text{O}$ .

A crucial aspect in the production of both Chinese Blue-and-white porcelain and their Persian imitations was the acquisition and use of cobalt blue pigment. The cobalt mine in Kashan is well-documented in Persian and Chinese historical manuscripts (for a review of Persian sources see Matin and Pollard Reference Matin and Pollard2015; for Chinese sources see Watt Reference Watt1979 and Wen Reference Wen2012: 70–130), and surveys, along with chemical and mineralogical analyses of its ore (Matin and Pollard Reference Matin and Pollard2017). It has been established that the cobalt ores from Kashan were of the high arsenic type, often accompanied by other high-iron minerals. The presence of arsenic and iron alongside cobalt in the analysis of cobalt pigments serves as a distinctive fingerprint indicating the source as the Kashan mine. Extensive analyses of Chinese blue-and-white porcelain, as well as Persian and other Islamic imitations, have revealed that the Kashan mine was the sole source of cobalt used in Chinese blue-and-white porcelain during the Yuan and early Ming dynasties. The fate of Kashan cobalt in China beyond this period remains unclear, and it is uncertain whether it was combined with other sources or entirely replaced by Chinese local sources. In Iran, while cobalt blue was used in the pre-Mongol period, its prominence increased significantly after the Mongol invasions. It became a dominant secondary colour on lustrewares, the primary glaze colour in lajvardina wares and subsequently featured in the underglaze blue-painted wares of the Timurid and Safavid periods. Chemical analyses conducted by Wen (Reference Wen2012: 305–336) on a wide range of Islamic glazed wares indicate that cobalt from the Kashan mine continued to be used for much longer, persisting until the nineteenth century, not only in Iran but also across the Islamic world. One notable exception seems to be the ceramics of the Iznik type (Section 6.5), which, based on its chemical composition, appear to have employed cobalt from a different source.