Monday 14 December 2020

Deciphering ceramics manufacturing techniques from the medieval city of Qalhât, Oman.

At the beginning of the first millennium AD, the Indian Ocean became a region of long-distance trade between the Middle East, India, South(east)-Asia and Africa, thanks to monsoon winds. The Qalhât site (Sultanate of Oman) located near the Ra’s al-Hadd was a key harbour city for short- to long-distance trade from the 13th to the 15th century. The city was founded around 1100 AD, sacked by the Portuguese in 1508 and becametotally abandoned during the 16th century. More than two hundred thousand ceramic shards have been  excavated from this site, and preliminary visual classifications based on shape and decors, as well Raman and X-ray fluorescence studies have been carried out. Many questions are still open for pottery characterised by a simple monochrome glaze and a porous body intermediate between that of terra cotta and faience. There are two well defined groups by visual examination, the so-called ‘Blue speckled’ and ‘Brown speckled’ wares; these are commonly found in many sites of the Arabic Peninsula and their place of production is highly disputed.

In a paper published in the journal Boletín de la Sociedad Española de Cerámica y Vidrio on 20 July 2020, Liliana Gianni of MONARIS (Molécule aux Nano-Objets : Réactivité, Intéractions et Spectroscopies) at Sorbonne Université, Hélène Renel of the Laboratoire Orient & Méditerranée at Sorbonne Université, Aleksandar Kremenović at the Faculty of Mining and Geology at the University of Belgrade, and Philippe Colomban, also of MONARIS, present the results of a studt which sought to obtain objective information concerning theirplaces of production, local, imported from Iran or elsewhere.

‘Blue speckled’ ware refers to pieces with a medium porous body, generally reddish and a glaze showing a wide variety ofshades of blue or green, sometimes pinkish or purple, or greyish and even nearly black. This glaze is characterised by millimetre dark speckles and is often slightly bumpy. Shapes of pottery are very typical, mostly ranging from wide shallow bowls with a straight lip or, most often, a flange, to jars in part. This ware has been found on the coastal sites of Iran, south Arabia, India and East Africa dating to the 14–16th/17th centuries. It is variously known as ‘Persian Blue speckled’, ‘Monochrome’or ‘Blue monochrome’, and is most commonly given a south Iranian origin, although a south Arabian production is also sometimes considered.

‘Blue speckled’ samples and corresponding cross sections for (a) 1; (b) 3; (c) 5; (d) 6 and (e) and (f) 8 shards. Gianni et al. (2020).

‘Brown speckled’ wares include a variety of pieces with arather fine grained, dense body, greyish or light red, sometimes with both colours. The glaze is brown to green, with millimetre dark speckles, showing orange splashes in a greenish matrix when the body has two colours. It is also common on the late sites of the Persian Gulf area and south Arabia, from the 16th century onwards, and is known as ‘Bahlâ’ or ‘Khunj’ wares. Its origin is under debate, whether from the Oman peninsula where similar pieces are still produced Its origin is under debate, whether fromthe Oman peninsula where similar pieces are still produced today, or from south Iran, in the district of Khunj in the Hormuz straights where wasters were seemingly identified in the 1970s. Shards of this type are even sometimes (mis)interpreted as of South Asian origin due to the high quality of the body and are often difficult to distinguish by eye sight but two different groups of shapes may nevertheless be individualised, one including flanged plates similar to the ones of ‘Blue speck-led’ ware, the other one with rather deep bowls with straight pointed lips as in Qalhât, each group possibly testifying a different origin or dating.

‘Brown speckled’/green samples and corresponding cross sections for Qa12, Ba21a, Mo20, Sa18.1 and Sa18.2 shards. Gianni et al. (2020).

The technology of these glazed ceramics seems to be among the simplest one: only one type of (clay-rich) earth or a simple mixture of common clay and sand was sufficient. The elemental and Raman analysis allowed the identification of the glaze type. Some phases, stable in contact with the molten glaze and rare from the geological point of view, as well as some minor or trace elements were used as provenance markers. The green body can then be coated with a glaze-precursor slip and put directly in the kiln. This simple technology limits the chance to find specific chrono-technological markers. Firing of terra cotta- and faience-like pottery is generally conducted at a relatively low temperature, between 850°C (apparition of liquid phases activating the sin-tering) and 1150°C (highest temperature achieved by many kilns), sufficient to form a transient liquid phase that will weld quartz and feldspar grains decreasing the porosity but without developing a significant volume of new phases as observed for stoneware and porcelain. Working techniques of traditional potters may have continued using raw materials coming from the same quarries without much change for centuries after the beginning of the exploitation. In this way, comparison with modern and contemporary traditional ceramics can be very informative. Thus, visually similar modern ceramics (19–21st centuries) were collected from Oman sites around Bahlâ which is a well-known pottery production centre, located in inland Oman, about 200 km to the northwest of Qalhât and close to the archaeological site of Salut (around 200 km from Qalhât and 40 km from Bahlâ). The Late Islamic period (i.e. 17th century onwards) Bahlâ wares excavated from Qattara Oasis, close to al Ain city, about 200 km northwest of Bahlâ were recently studied.

In Gianni et al.'s study, a multistep analysis protocol was pursued to obtain comparative information: (i) macrostructure examination of pottery sections and Raman microspectroscopy of glazes and minerals of the body (using the mapping modalities) in association with X-Ray Powder Diffraction, (ii) glaze composition analysed by scanning electron microscope/energy-dispersive X-ray spectroscopy, and (iii) thermal expansion measurements to determine the firing temperature.

Five representative ‘Blue speckled’ and five ‘Brown speckled’ (one from 13th century and 4 post-medieval) shards were selected on the basis of previous work in which different groups were identified: hundreds of shards have been classified according visual criteria, 65 of them being analysed by Raman microspectrometry (including 31 ‘Blue and Brown speckled’ shards), plus respectively 58 of them at the laboratory and 41 other at Qalhat site by portable X-ray fluorescence. Blue shards represent allthe variety of enamels. The good homogeneity of the ‘Brown speckled’ wares observed by archaeologists and previous work led to the selection of a single medieval sample. The selected shards were cut with a saw and the cross sections were observed with an optical microscope. However polishing liquids generate a strong fluorescence, detrimenta lto Raman analysis due to the penetration of chemicals into the porous body. Porosity was evaluated in an area of 6 mm² per sample using ImageJ software. Pore types were classified as a function of their dimension with small (5–50 μm), medium (50–100 μm) and large (over 100 μm) size. The number of pores and percentage of three dimension classes (small, medium and large pores) were calculated to obtain the most characteristic porosity dimension for each sample. The bubble number was calculated in the same area of 6 mm² while the mean thickness was extrapolated along the entire sample section.

identificationFor all samples of the two groups, X-Ray Powder Diffraction patterns showthe presence of three abundant mineral phases: quartz, augite (a pyroxene), and albite (a plagioclase feldspar). Only sample Qa12 contains quartz as the abundant mineral with augiteand albite as moderately to minorly present minerals. Differentiation of the old Qa12 sample is thus obvious. Minor amounts of paragonite (a mica) werre observed in sample 6. Amorphous X-Ray Powder Diffraction patterns are usually obtained for glazes coloured ingreen or blue. However, for sample Ba21, two mineral phases were noticed such as diopside (previously reported for Late Islamic Arabic period productions) and calcite. Mineral phases with minor amounts which were detected by Raman analysis were not detected by X-Ray Powder Diffraction. Furthermore, from X-Ray Powder Diffraction patterns, the content of each mineralphase could be estimated to be less than 5%.

X-Ray Powder Diffraction patterns for samples Ba21 (top) and 6 (bottom). Indexed phases of quartz (PDF# 01-089-1961), augite (PDF#01-088-0856), albite (PDF# 01-089-6428) and paragonite 2M1 (PDF# 01-076-5968) are presented in the form of straightintensity lines at the bottom. Gianni et al. (2020).

Previous methodological investigations have demonstrated that Raman mapping could be more efficient than X-ray powder diffraction to identify the minor phases. However, X-Ray Powder Diffraction gives a more realistic view of the major crystalline phases.

waresBody: Samples of ‘Blue speckled’ group look rather similar withtheir red bodies. Examination of the cross sections showed that samples 5 and 8 appear darker (more red) inthe core of the section and lighter towards the two sides, suggesting different firing conditions than the others. The number of pores and their dimensions were calculated by theobservations under the optical microscope and the images taken at low magnification. Considering the number of pores present in the same area, bodies of the samples turn out to be similar with a medium level of porosity (24–58 pores per unit area) with the exception of sample 3 which is more compact; 24 pores were calculated for samples 1 and 8 while samples 5 and 6 Exhibit 58 and 40 pores per surface unit, respectively. The percentage of the pore size in all samples is comparable: around 60% large pores, around 25% medium pores andaround 10% small pores. Sample 5 has a lower percentage (2%) of small pores and sample 3 differs from the others withthe total absence of pores in the area considered. The dimension of the smallest pores is similar whereas some differences in the large pores were found. However, distribution of the pore size is analogous. Thus, macrostructures are very similar, except for sample 3 where the pore size is much smaller indicating a higher firing temperature or the use of finer particles or a composition with more flux. Regarding the body, no specificity can be identified for sample 5 which is assigned to a potential production of Iran on the basis of visual examination.

The most frequent Raman spectrum collected in mapping analysis is assigned to quartz. The presence of dolomite is also confirmedby Raman scattering. The unusual presence of phosphate in almost all the samples is worth mentioning (except for sample 8), with the characteristic phosphorus-oxygen mode peak at about 960 cm⁻¹. This phase may arise from minerals or more likely from ashes of animal residues (dried animal dung) or contamination (soil, ingredients kept in the pottery). Gianni et al. note that fish head and bone residues, likely used as fuel (due to oil-rich content) have been found close to Qahlât kilns. Thus, pottery exhibiting a high level of phosphate such as samples 3 and also Qa12, Ba21, Sa18.1 may have been produced using similar raw materials and/or technology or may have had similar use. Only sample 8 does not exhibit the phosphate signature.Also, in all bodies of ‘Blue speckled’ ware, frequent titaniumoxide polymorphs such as anatase (140 cm⁻¹) and rutile (444 and 605 cm⁻¹) are present, suggesting firing temper-atures below about 1200°C (i.e. under the temperature required for the complete transformation of anatase into rutile from about 900°C for nanometric grains to about 1200°C for coarse grains). Anatase is abundantly present in samples 1, 3, 5 and 6. The rare presence of rutile was detected in samples 3, 5 and 8 (also in Mo20 and Sa18.2; totally absent in Ba21 and Sa18.1.

Hematite and magnetite were found in all the samples (abundant in samples 1, 3 and 8), confirming the use of oxidizing firing atmosphere (as proposed by the observations of different colours between the core and sides of cross sections) while maghemite was detected in samples 3, 5 and 6. Maghemite transforms into hematite by heating above 800°C. Augite was only clearly found in sample 1 (like sample Sa18.1). Albite and microcline arecommonly observed. Calcite and dolomite present in some of the samples propose lower firing temperatures (below 1200°C). Traces of organic molecules (oleic acid) probably resulted from food storage in the vessels were found insample 5.

Glazes ofsamples 3 and 8 show a heterogeneous dispersion of the blue colour. Composition analyses show that copper oxide content ranges between 0.3 (sample 3) and 1.62 percent by weight (sample 5). The tin oxide content is relatively high (1.2–1.9%), in agreement with the possible use of bronze residues (cassiterite was not detected by Raman scattering, thus tin was dissolved within the glaze network and did not opacify). Thiscould be due to the mixing of colourless and coloured glazes (anima/corpo technique) which could be consistent with the importation of the raw materials used for preparation of the glaze (the precursor powder) and the speckle distribution. Some bubbles were found in the glazes of samples 1, 3 and 5, which indicate the reaction with the body, suggesting a single step firing process of an enamelled green body. The number of bubbles for unit area is similar as well as dimensions of the pores for samples 1 and 5 while larger bubble diameters were measured for sample 3 (richer in alumina). No bubbles were observed in the glazes of samples 6 and 8 with ×5 magnification. The thicknesses of glazes are comparable for the top and bottom sides as well as between all the glazes. These technological differences (presence of bubbles, colour dispersion and mono- or multistep firing) may correspond to different productions (place or time?) or even different locations in the kiln, the gradient of temperature being rather big in an ancient kiln. The highest copper oxide contentof the glaze of sample 5 explains the darker colour but no other significant differences were found regarding the body. The composition of all glazes is very similar, except the alumina concentration of sample 3 which is a little higher. Also, the variable content of potassium, which Gianni et al. note is an important criterion, (potassium oxide forms 2.65% by weight of the total oxide content in sample 3, for example, 5.18% in sample 5, 3.5% in sample 6, and 5.75% in sample 8). However, this isnot associated to a high level of magnesium (magnesium forms less than 0.6% by weight of the total oxides in samples 1, 5, 6, and 8, and between 2.2 and 2.3 in sample 3), which could suggest the use of ashes as flux. The equal con-centration of titanium (titanium oxide forms between 0.3 and 0.6% by weight of total oxides in all samples) causes Gianni et al. to suppose the same type of sand as well as the comparable concentrations of barium (barium oxide makes up 0.3% by weight of the total oxides in samples 1, 5, and 8, and between 0.5 and 0.7% by weight in samples 3 and 6). The glazes also containa small amount of lead oxide (which forms 1-2% by weight of the total oxides in samples 1, 3, and 8, and 2.5% in samples 5 and 6), tin oxide (roughly 1.8% by weight in all samples), copper oxide (less than 0.4% by weight in all samples except sample 5, where it comprised 1.6% by weight). Gianni et al. also mention that cobalt was not detected in the glazes, the turquoise colour being only due to copper ions. There are no arguments to discriminate sample 5 from the others and it can be concluded that all glazes are identical with a single origin of production.

Comparison of the glaze compositions of ‘Brown speckled’ (a), (b) and ‘Blue speckled’ wares (c), (d); (a) and (c): refractory oxide content vs. flux, (b) and (d): flux composition. Gianni et al. (2020).

All samples havethe same Raman broad spectrum with the vibrational mode at about 510 cm−1and the stretching mode at 1090 cm⁻¹, a signature assigned to alkali glass (sodium oxide + potassium oxide (+ calcium oxide) flux). The intensity of the roughly 960 cm⁻¹ broad component, characteristic of poorly polymerised silicate network is consistent with a moderate firing temperature. In sample 5, the characteristic resonance Raman feature of the so-called amber iron sulphite chromophore at about 420 cm⁻¹ explains the strong darkening of the blue hue. The glaze structure is also different. The lack ofthe 960 cm⁻¹component and the shift to 1100 cm⁻¹ of the silcon oxide stretching mode indicate a different composition, as also confirmed by the highest silica and alumina contents measured by Scanning Electron Microscopy with Energy Dispersive X-Ray Analysis. There are no differences between blue and grey glazes because only the glass signature was detected in the Raman signature implying that the colouring copper ions were dissolved in the glaze network.

Representative Raman spectra recorded on sample glazes: (a) ‘Brown speckled’, (b) ‘Blue speckled’ wares. Gianni et al. (2020).

In the cross sections of the ‘Brown speckled’ ware samples, a similarity was observed between Qa12 and Sa18.1 as well as Ba21 and Mo20.The grey core of the body develops a darker red layer towards the two opposite sides of the shards, suggesting oxidising firing atmosphere.

The red body of sample Sa18.2 shows a typical black core indicating the use of a clay rich in humic acids (such clays exhibit a better plasticity). Pore dimensions, number and percentage of small, medium and large pores were calculated as for ‘Blue speckled’ samples. The higher porosity levels measured for the modern samples Ba21 and Mo20 are consistent with a two-step firing process (glaze deposited on an already fired put porous body) while the matrix of the ancientsample Qa12 appears compact without any pores under the used magnification in the area analysed. Usually, the heating rates used in ancient times are much lower than those presently used, which could lead to better sintering for a similar firing temperature and smaller bubble size due to the higher viscosity of the glaze. Samples Ba21 and Mo20 display about 150 pores for the area analysed of comparable dimensions (Ba21: 51–128 μm; Mo20: 35–250 μm) taking into accountthat the range includes the smallest and the biggest pore found in the area. Intermediate porosity levels were calculated for the two shards of Salut samples (Sa18.1 and Sa18.2) respectively with 125 and 80 pores for the area analysed. Samples Sa18.1 and Sa18.2 appear similar with the presence of small, medium or large pores while Ba21 is exclusively characterised by large pores.

Feldspars such as albite and microcline were detected in all the samplesby Raman analysis, except samples Qa12 and Mo20. Microcline was also found in samples Ba21 and Sa18.1, Sa18.2 while albite signal is absent only in sample Qa12. Consequently, thesand used for samples Ba21, Sa18.1 and Sa18.2 is differentfrom that used for samples Qa12 and Mo20. It is probable that no sand was specifically added for Qalhât shard Qa12. Thus, different raw materials have been used for these shards. Calcium phosphate was abundantly found inall the samples while it is scarcely present in samples Mo20 and Sa18.2. In fact, phosphate is rarely found in pottery body and this appears as a common characteristic common with ‘Blue speckled’ wares. Rutile and anatase were also abundantlyfound in samples Qa12 and Sa18.2 and scarcely in sample Mo20.

Iron oxide is present in the form of hematite, magnetite and maghemite. Magnetite and maghemite are absent in sample Sa18.1 and the latter was also not found in sample Ba21. The presence of hematite in all samples is consistent with an oxidising atmosphere during firing. A major presence of augite was clearly found only in sample Sa18.1 by Raman analysis, although augite was detected by X-Ray Powder Diffraction in all samples. This demonstrated that one technique may not be sufficient to detect all the phases present. Augite is a type of pyroxene usually found in specific rocks such as ophiolites, lavas and gabbro. Ophiolites are common in Oman Mountains and detection of augite could be a reliable marker of local production. Iranian geology also incorporates ophiolite and augite, but far away from Hormuz Detroit. Calcite is absent in samples Qa12 and Sa18.2 while dolomite was only detected in sample Sa18.1. 

All the samples selected are covered by a glaze layer with colours of brown (Qa12 and Sa18.1), green (Ba21 and Mo20) and yellow-brown (Sa18.2) and similar thickness (about 150 μm). Samples Qa12, Ba21 and Sa18.1 have glazes on both sides. Observation of the cross sections showed the similarity of samples Qa12 and Sa18.1 (no significant inter-phase between the body and the glaze) and also some similarity with sample Sa18.2. An intermediate layer of aboout 80 μm between the upper glaze layer and the paste can be distinguished for samples Ba21, Mo20 and Sa18.2. Such an interlayer could result from a deposit (engobe?) or from a reaction zone due to a deposit of the enamel slurry on a green body and a single step cofiring process that promotes the extent of the reaction zone. No bubbles were found in any glazes, indicating a low viscosity at the firing temperature and a deposition on already prefired body.

The elemental analysis shows two groups. On one hand, samples Ba21, Mo20 and Sa18.1 (very close in the values of refractory oxide content vs. flux) are coated with analkali-containing glaze and on the other hand, samples Qa12 and Sa18.2 are glazed with a lead-based composition (more widespread in the graph than the others). Each value is the mean of data measured on three spots. The concentration of lead in the latter two samples is variable and higher than 14% by weight (the others below 4% by weight). This indicates that the glazing technology of sample Sa18.2 is the most comparable tot hat of the mediaeval Qa12 shard. This is consistent with a technological change from lead-rich glaze for medieval and post-medieval periods to an alkali-rich glaze for the modern period. As usual, lead-based copper-containing glaze is green. The high level of iron oxide measured for Qa12, Sa18.1 and Sa18.2 glazes explains the brown colour.

Raman sig-natures show the characteristic silicone oxide bending and stretching broad modes of poorly polymerized amorphous silicates, i.e. glassy silicates with relatively low temperature of melting. Due to the incomplete reaction between the raw materials used to prepare the glaze, signatures of quartz and feldspars were recorded in many places. For instance, in sample Mo20, the contribution of unreacted quartz and feldspar grains is shown, hindering the observation of the glaze signature (broad components). It should be noted that the size of the probed volume by the Raman microspectrometer (roughly 5 μm × 5 μm × 15 μm) is much smaller than the glaze thickness, which guarantees the lack of contamination by the signature of body phases and the possibility to identify the glaze heterogeneity. Raman spectra of Qa12 and Sa18.2glazes are similar, without features at about 750 cm⁻¹, according to a significant lead content, in agreement with the glaze composition.

Additional iron oxide signature with spinel structure, likely magnetite was found in Ba21 and Sa18.1 glazes, which indicates firing under reducing atmosphere or the use of a clay rich in humic acids. The green colour of samples Ba21 and Mo20 is however not only related to a high level of copper. The colouration is likely due to the combination of iron⁺² and iron⁺³ ions, but clear observation of the Raman signature of spinel could indicate the presence of ions promoting the spinel structure such as chromium. The very surprising high level of barium oxide (about 12% by weight) detected in Ba21 glaze is interesting. The narrow 667 cm⁻¹ peak indicates traces of calcium-antimony oxide opacifier. 

The analysed samples obviously belong to different productions. Samples Sa18.2 and Qa12 are very different from the others, both regarding the body and especially the lead-rich glaze. On the other hand, the glaze of other shards is alkali-rich and this can be interpreted as a technological change between the medieval/post medieval and modern period.

Thermal expansion analysis was pursued to make the precise estimation of the firing temperatures. Thermal expansion measurement has been one of the favourite techniquesto control ceramic production and qualify raw materials and glazes. The onset of the thermal expansion jump is considered at the firing temperature.

Thermal expansion curves measured on sample bodies: top, ‘Blue speckled’wares, bottom, ‘Brown speckled’. Gianni et al. (2020).

The curves obtained indicate that the ‘Blue speckled’ ware ceramics were fired around 1140°C for samples 6 and 8 and at a little higher temperature, 1150°C for the others. The closer values may clearly indicate the same type of kiln. In accordance with the identification of glaze compositions and phases, thermal expansion curves confirm the homogeneity of the group. Sample 5, with dark glaze coloured both by copper ions and iron sulphite amber chromophore shows small events at about 300–500°C which were not detected in the other samples (traces of cristobalite?)

Thermal expansion analysis shows the firing temperature ranges for ‘Brown speckled’ ware from about 1080°C for sample Sa18.2 to 1120–1150°C for other samples. Ancient sample (Qa12) was fired at 1120°C. The firing temperatures are fully consistent with the number of pores measured per unit area, the lowest number for the lowest firing temperature and the highest number for the highest firing temperature, except for sample Qa12. Thermal expansion curve of sample Qa12 exhibits a strong event at about 600°C which is characteristic of the α to β-quartz transition, according to the large amount of quartz in the body, i.e. consistent with the use of coarse sand and/or raw earth, without specific preparation. This finding also demonstrates that more recent pottery samples are different from the 13th century ones.

Representative archaeological shards excavated at Qalhât and collected at the important active neighbouring site of Bahlâ were analysed and compared to modern pottery produced in a regional pottery production centre. Augite was detected in allthe shards by Energy Dispersive X-Ray Analysis and in some of them by Raman. Local production of pottery with body having a high content of augiteis thus very likely but more information on the geology of the Oman Sultanate and Iran is needed. Identification of phosphate in almost all the samples (except sample 8) suggests the same procedure of firing, the use of phosphate-containing raw materials or water, or the same type of contamination in relation with the use as vessel or the conservation in the soil. Recording of the oil signature by Raman scattering in someshards is consistent with contamination and a further chromatographic study of the soluble food residues to be extractedfrom the pottery body would be interesting to identify the organic matter and document the function of the utensils.

Elemental compositions of the ‘Blue speckled’ glazes are more or less similar with some differences only found in sample 3 (higher level of alumina) and sample 5 (the amber iron sulphite chromophore darkens the turquoise colour obtained by copper ions according the higher copper oxide and iron oxide content).Considering the different characteristics, the shards appear very similar and may have the same local origin of production. Many other common minerals were found such as anatase (the firing temperature below 1200°C) and rutile (high temperature polymorph, consistent with a firing temperature close to 1150°C) in samples 1, 3, 5, 6 and 8. Whatever the very homogeneity of the bodies at visual examination, specific characteristics can also be identified such as the presence of unreacted feldspar and other particular minerals (e.g.dolomite). The main differentiation can be made by considering the glaze. Importation of the glaze (or glaze precursor(s) only) is possible, the high similarity of the glazes contrasting with the variation of characteristics regarding the body.

Considering the ‘Brown speckled’ glazes, cross sections of samples Qa12 (medieval) and Sa18.1 (absence of an interface, consistent with single step firing) and Ba21, Mo20 and Sa18.2 (presence of an interface, prefiring of body before enamelling?) have different characteristics while the glaze thickness is comparable. Elemental and Raman analysesid entified samples Mo20 and Ba21 as alkali glazed (Ba21, barium-based glaze), as in the case of ‘Blue speckled’ glazes. However, samples Qa12 and Sa18.2 are covered by mixed lead-rich glazes. Feldspar-containing sand appears typical of Bahlâ productions. On the contrary, rather high levels of rutile and anatase are characteristics of Qalhât samples. Important similarity was remarked in compositional and technological aspects between the medieval Qalhât shard (Qa12) and samples from the ruined village near Salut (in particular Sa18.2), with the use of similar technological procedures (lead-rich glaze, single firing in similar kilns) which fit with a slightly lower temperature of firing. The other modern samples (Mo20 and Ba21) belong to another technological procedure (multi-firing?); this can be interpreted as a technological change between the medieval/post medieval and modern period. Sample Ba21 is characterised with its very special specialbarium-based glaze. Finally, Qalhât sample Qa12 and Salut sample Sa 18.2 are glazed with lead-rich glazes whereas alkaline glazes are used for the others. In conclusion, modern Bahlâ pottery cannot be considered as a continuation of Qalhât production (stopped around 1508 with the destruction of the city).

See also...

Follow Sciency Thoughts on Facebook.

Follow Sciency Thoughts on Twitter.