Microscopic fabric analyses of Aswani and Merotic pottery

January 17th, 2018

Compound microscopic analysis

The 15 samples of R30 wares from Qasr Ibrim that were selected for this study were examined under a x30 compound microscope to visually identify any inclusions and verify their previous fabric categorisation as Family A wares. More specifically, fresh breaks on the sherds show that this family is characterised by a hard and dense paste, pinkish buff to pinkish tan in colour. Adams (1986b, 526) notes that the Aswani fabric is made out of a fine clay which contains fine to medium quartz sand grains, black opaque particles, dark red particles and sometimes large white opaque particles. The present visual analysis shows that all 15 sherds fall under the same category of Aswan wares but their fabric can be subdivided into two groups: a finer and a coarser version. The finer version is found in samples R30/1-4, 6, 8-9 and 15. The fabric is pink (5YR 7/3 or 7/4) and it contains fine quartz grains, small black and white opaque particles of up to 5% frequency, and clay pellets (up to 2%). The coarser version is represented by samples R30/5, 7 and 10-14. The fabric colour is exactly the same; however, the paste contains larger quartz particles and clay pellets (up to 5%). All sherds were tested with 10% dilute hydrochloric acid and showed no reaction. This verifies that the fabric of this family is made out of non-calcareous clay, which matches the granitic geology of the local region (Schuelter, 2006, 88-89; Furon, 1963, 152-165). Furthermore, the Aswani clays are likely to come from a kaolinitic source that fires to a pink colour.

The compound microscope examination of the 15 samples of W26 wares suggested three different fabric variations:

  1. Samples W26/1, 4, 6 and 9 appear slightly coarser and softer than the other samples. They contain small to medium-sized quartz sand, some small grey opaque particles and clay pellets (up to 2%). The fracture of sherds 1 and 6 is light reddish brown (5YR 6/4 or 5/4) fading to pink or pinkish white (5YR 8/3 or 8/2) towards the centre. The fracture of sherds 4 and 9 is reddish yellow (5YR 6/6). These four sherds are likely to come from alluvial clays that fire to a reddish colour. Furthermore, they cannot be characterised as ‘eggshell’ wares because their thickness varies between 4mm and 5mm. According to Adams’ fabric descriptions they seem to be a better match with the N Family of Nubian wheel-made coarse wares, and more specifically the N.I subgroup. This family is postulated to be of Nubian origin, even though the production centres have not been verified. Their texture is described as medium to coarse and their paste contains fine to medium quartz sand. Adams (1986b, 454) describes them as “made from ordinary Nile mud”. The appearance of these four sherds in the M Family group may indicate that Adams’ categorisations of Meroitic and Nubian wheel-made wares should be carefully revised. It is likely that Adams’ fabric categorisations within ware W26 were misled by his stylistic observations. In other cases, it is also likely that Adams’ N Family contains mixed fabrics (alluvial and kaolinitic), which can be mistakenly identified because they fire to similar colours to the M Family fabrics. Thin section analysis is a useful technique to prove this point.

  2. The second variation of the M Family wares is represented by sherds W26/2-3, 5, 7 and 11-15. These sherds are both thin-walled ‘eggshell’ wares with an average thickness of between 2mm and 2.5mm (sherds 3, 14, 15), and slightly thicker wares (sherds 2, 5, 7, 11-13) that exhibit an average thickness of between 3mm and 4mm. Their fabric colour is either a white to pinkish white core between pink margins (5YR 8/1 or 8/2) (sherds 2, 6, 11, 13) or entirely pink (5YR 7/3 or 7/4) (sherds 3, 5, 7, 12, 14 and 15). These sherds are likely to come from kaolinitic clays, similar to the Aswani clays, which fire to similar pinkish colours. Their paste contains small to medium sand grains and clay pellets (up to 2%), and sherds 7, 11 and 15 contain small rounded red and black particles. The inclusions of the three last sherds exhibit great similarity with the Aswani fabrics, yet they do not match the description of ware W27, which according to Adams (1986, 439) is the only undecorated Meroitic pale pink ware made in imitation of Aswani ware R30. This observation is interesting. It is possible that the clays of other wares that belong to the M Family (and not only W27) were selected from areas that matched the Aswani fabrics. Thin section analysis could again be used to prove this point.

  3. Finally, sherds W26/8 and 10 seem to be completely different to Adams’ general fabric description of ware W26, even though they exhibit the same pinkish white colour as the previous category (5YR 8/2), and they come from eggshell wares with 2mm average wall thickness. Their texture is very fine and the visual identification of inclusions is difficult. Both sherds seem to contain sparse mica. Again, thin section analysis appears to be the most appropriate technique to investigate the provenance of these fabrics.

Tests with 10% dilute hydrochloric acid showed no reaction in all 15 sherds. This test indicates that all W26 samples are made out of non-calcareous clays.

Thin section analysis

The thin section analysis of the Aswani R30 sherds verifies that all samples were made out fine clay that originates from the same area. More specifically, the samples exhibit a rather homogeneous fabric, the clay of which derives from sedimentary rock formations, widely distributed in Egypt and particularly in the Aswan region.

Generally, the majority of the rocks in Egypt are sedimentary and the most characteristic ones are the Nubian sandstone and the shelly limestone (Furon, 1963, 152-165). The main geological bed of Egypt is composed of marine and non-marine deposits, formed during the Cenozoic period. These are clastics, carbonates, gypsum and local basalt-dolerite dykes, which spread from the northern part of Egypt to a large zone west of Aswan (Schuelter, 2006, 88-89). However, the geology of the area around Aswan is composed of Cretaceous sediments that formed during the Mesozoic period. According to the geological map of Egypt (Schuelter, 2006, 89), these rocks are mainly clastics, phosphates and carbonates. The same geological bed extends south along the Nile, forming the same geological composition round the area of Qasr Ibrim. Close to the modern city of Aswan, another geological formation contains granitoids, slightly metamorphosed clastic and volcanic formations, and local gabbros (felsites and porphyries).

The main characteristic of all R30 samples in thin section is the presence of large concentrations of pyroxenes in their fabric. The existence of pyroxenes is quite common along the Nile, especially in the areas around Aswan where the two geological formations described meet. Of course, pyroxenes can also travel from other places along the river and become naturally mixed with sediments of local origin. The second characteristic of the R30 samples is that their fabric can be subdivided into two different versions: a fine version that is found mainly in finewares (samples R30/1-4, 6-9 and 15) and a coarse version that is found in thicker amphorae sherds and exceptionally in one fineware sherd (samples R30/5 and 10-14). This observation matches almost perfectly the results produced by the compound microscope analysis and only one sample (R30/7) was mistakenly identified under the normal compound lens. Two fabric group subdivisions can now be defined, as follows:

  1. Fabric ASW1 is the finer fabric, containing abundant, well-sorted, angular, small to medium quartz grains, 5% pyroxenes, 3% large clay pellets, 1% plagioclase feldspar, less than 1% iron oxide, occasional large pieces of siltstone and sparse mica (Plates 1 and 2).

Plate 1: The fabric of sherd R30/1 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

Plate 2: The fabric of sherd R30/3 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

  1. Fabric ASW2 is the coarser fabric, containing abundant, poorly sorted, angular quartz grains that come in a variety of sizes, fragments of siltstone (approximately 5%), large clay pellets (up to 5%), 3-5% iron oxide, 2% pyroxenes and sparse plagioclase feldspar (Plates 3 and 4).

Plate3: The fabric of sherd R30/10 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

Plate 4: The fabric of sherd R30/13 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

The existence of a coarser fabric used mainly for the manufacture of thicker amphorae and other table-wares indicates that the Aswani potters performed deliberate selection and levigation of their clay according to the type of vessel they aimed to build.

The thin section analysis of the W26 sherds proves the existence of four different fabric groups in the Qasr Ibrim assemblage. The results match several of the conclusions of the previous compound microscope analysis, and they also revealed a larger and more complex variety of fabrics. Under the compound microscope, the main differences among the W26 sherds could be identified in terms of clay colour, visible inclusions and coarseness of the fabrics. However, under the petrological microscope four different fabric groups were identified based on their mineral inclusions, as follows:

  1. Fabric MR1 is a fairly coarse fabric, containing approximately 20% poorly sorted angular quartz grains and sparse plagioclase feldspar. This fabric was identified in samples W26/1, 2 and 5-7 (Plates 5 and 6). Only one sample (W26/1) contained occasional limestone, but not enough to form a separate fabric group.

Plate 5: The fabric of sherd W26/1 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

Plate 6: The fabric of sherd W26/5 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

  1. Fabric MR2 is a coarse fabric with poorly sorted inclusions; it contains 20-25% small rounded quartz grains accompanied by some larger angular quartz grains, 2-3% pyroxenes and sparse plagioclase feldspar. This fabric is found in samples W25/3 and 4 (Plate 7) and it generally resembles the Aswani coarse fabric ASW2.

Plate 7: The fabric of sherd W26/3 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

  1. Fabric MR3 is a completely different fabric with a variety of mineral inclusions. It has a medium texture, containing 15-20% very small, well-rounded quartz grains combined with sparse – but larger – well-rounded quartz grains, some mica (between 5% and 10%), up to 5% iron oxide, and hornblende (between 2% and 3%). This fabric was detected in sample W26/9 (Plate 8).

Plate 8: The fabric of sherd W26/9 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

  1. Fabric MR4 has a very fine texture, containing very small, rounded to sub-rounded quartz grains, mica (up to 10%), medium-sized pieces of iron oxide (less than 5%) and no visible hornblende. This fabric was identified in samples W26/8, 10-15 (Plate 9, 10 and 11).

Plate 9: The fabric of sherd W26/8 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

Plate 10: The fabric of sherd W26/10 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

Plate 11: The fabric of sherd W26/14 in thin section under Polaroid light, viewed through a 10/0.25/160 lens with the analyser in

As mentioned above, the geology of the area around Qasr Ibrim contains mainly clastic, phosphate and carbonate rock formations. However, Meroitic pottery may come from areas further south of Qasr Ibrim, from other production centres in the Kingdom of Kush, now located in the modern Republic of Sudan. According to the geological map of Sudan (Schuelter, 2006, 219), the variability in the fabrics of Meroitic pottery could be attributed to a large variety of rock formations. The area directly south of Qasr Ibrim is composed of Mesozoic formations, mainly clastic sediments, evaporates, limestones and Nubian sandstones. Further south the geological bed is followed by a Precambrian and Undifferentiated basement complex. The geological formations contain some intrusive rocks (granites, syenites, pegmatites, granodiorites, basic and ultrabasic rocks). Other geological formations in Sudan include a combination of gneisses and schists (granitic gneisses, migmatites, chamockitic granites, amphibolites, pyroxene granulites and metasediments), along with basalts, rhyolites and trachytes. These formations are detected in two large zones: one in a circular area between Khartoum and the Egyptian border, and another one next to the coast of the Red Sea and the area around Port Sudan (Schuelter, 2006, 219).

A discussion of the results of thin section analysis

Previous thin section analysis of Aswani samples performed by Laurence Smith (1999) indicated two main fabric divisions that perfectly match the results of the current analysis of ware R30. More specifically, Smith’s Group 1 contained moderate size quartz grains with different degrees of angularity, small quantities of iron oxide, opaque minerals, muscovite and biotite mica, rutile, epidote, zircon and pyroxenes (probably augite) (Smith, 1999, 13). This fabric matches fabric ASW1 of the present analysis. Smith’s Group 2 had similar mineralogy to his Group 1. However, the main difference in this fabric was the high frequency of iron oxide grains; the fabric of Group 2 also contained fine crystalline quartz inclusions, which were likely to be from mudstone or siltstone fragments (Smith, 1999, 14). This group matches fabric ASW2 of the present analysis.

According to previous fabric analysis on Meroitic finewares recovered at Musawwarat es Sufra, it has been demonstrated that they were made out of fine white or cream kaolinitic clay, while several other samples exhibited pink fabric. The two main identified fabrics were recorded as: a) pure fine white ‘kaolin’ clay, and b) a mixture of ‘kaolin’ clay with residual alluvial clays (Edwards, 1999, 26-27). Such mixing of clays has been also suggested by Smith (1996) after thin section analysis of other Meroitic finewares from the south regions of Kush. A number of potential clay sources have been suggested in the Meroe area by Robertson (1975, 26-27) and Smith (1996, 23-25). Edwards (1999, 27) noticed that similar clay sources were available in Musawwarat es Sufra, lying in the stone quarries to the west of the Great Enclosure. Similarly, Anne Seiler (1998) characterised both of these fabrics as ‘kaolinitic’: C1 was described as a kaolinitic fabric with fine black and white inclusions, with a generally white (5YR 8/1) fracture but with some variation towards very pale brown. C2 was described as a kaolinitic fabric with fine black and red inclusions with a generally pale red (2.5YR 7/2-7/3) fracture. These fabric types correspond to Adams’ M Family (W26 and W35), while they also match the Type F of Shinnie and Bradley (1980, 154-155).

The present thin section analysis of the Meroitic W26 samples indicates a greater variability in fabrics than the analysis conducted by Seiler (1996) and many fabric differences to the ones proposed by Smith (1999). More specifically, previous work on samples from Qasr Ibrim by Smith (1999) distinguished three main fabric groups:

Group 1 fabric contained about 30% quartz, ranging from small to medium grains, frequent iron oxide, rare mica (both muscovite and biotite) and 1-2% of other mineral inclusions, such as hornblende, pyroxenes and plagioclase feldspar. This fabric was postulated to be a mixture between a residual clay and Nile silt, since there were some similarities in mineralogy with Nile silt wares, particularly in terms of the presence of heavy minerals and plagioclase (Smith, 1999, 25-26). It is interesting that in the present assemblage of ware W26 no sherds matched this specific fabric description.
Smith’s fabric Group 2 contained 10-15% of moderate-sized rounded to sub-rounded quartz grains, 10% of which were silt sized. Other inclusions were 5-10% iron oxide, sparse mica, and sparse quantities of very small grains of pyroxenes, hornblende and plagioclase feldspar (1-2% all together) (Smith, 1999, 26). In the present analysis this specific fabric resembles fabric MR2; however, the absence of hornblende and iron oxide in fabric MR2 may indicate that this is in fact another type of mixed clay.
Fabric Group 3 was quite distinctive, containing relatively sparse grains of which the majority were silt sized. The inclusions were virtually all quartz, with some feldspar and sparse iron oxide (Smith, 1999, 26). This specific group matches perfectly with the present fabric MR4.

Smith (1997, 81-83) suggests that the clays from the northern and southern regions of Nubia exhibit a fair degree of variability and they can be distinguished by the presence, absence and abundance of the following minerals: muscovite and biotite group micas, plagioclase feldspar, hornblende and pyroxenes. Smith divides his northern Nubian clays into two main groups: firstly the Meinarti/Aswan group of the northern Meroitic clays, which is characterised by high muscovite, moderate plagioclase and absence of hornblende (Smith, 1997, 87). Such fabrics have been recorded as far as Musawwarat es Sufra, which lies a long distance south of Qasr Ibrim (Edwards, 1999, 46). Secondly, the Kalabsa group of the northern Meroitic clays, which is characterised by moderate muscovite, low plagioclase and low hornblende (Smith, 1997, 83). In the present W26 assemblage from Qasr Ibrim, no wares were found to match the description of the Kalabsa group. However, fabrics MR1 and MR2 seem to match the general description of the Meinarti/Aswan group, even though other inclusions, such as pyroxenes, indicate that MR2 could be a clay mixture.

Under the compound lens, some of the wares that were identified as MR1 and MR2 in thin section appear to match Adams’ N Family of Nubian silt wares. However, in Smith’s analysis, samples made from pure Nile silt were found to contain frequent quartz, very frequent iron oxide, frequent biotite mica, rare hornblende and rare basaltic fragments (Smith, 1999, 16). Finewares (such as W26) are not generally expected to be made from pure alluvial silt (such as the N.Is) (Smith, 1999, 16), even though they can be mixtures of alluvial and kaolinitic clays; therefore, it is likely that MR1 and MR2 were made of mixtures of fine clays and alluvial silt that contains other various inclusions. Fabric MR2 could possibly be made of a mixture of fine northern clays from the Aswan region and Nile silt.

According to Smith, clays of southern Nubian origin contain moderate muscovite, plagioclase and moderate hornblende. More specifically the Meroe area clays tend to be characterised by frequent to very frequent quartz (30-40%), fairly frequent iron oxide grains, micas and sparse opaque minerals (Smith, 1997, 83). This petrological description strongly suggests that fabric MR4 is derived from the southern clays of the Meroe region.

In Smith’s previous work on other samples from Qasr Ibrim (1997, 86), a fabric described as QI1 was found to contain low muscovite, plagioclase and hornblende. This fabric matches almost perfectly the present fabric MR3. Smith suggests that this QI1 fabric is not compatible with any northern clay group (Meinarti/Aswan or Kalabsa) and the muscovite content is too low to match any of the southern clay groups (Meroe or Musawwarat es Sufra). Therefore, the source of this kind of clay cannot be determined on the basis of mineralogical study alone (Smith, 1997, 86). A similar suggestion is possible for fabric MR3.

In conjunction with previous thin section analyses performed by Smith, the conclusions of the present analysis indicate that the four fabrics from the W26 assemblage of Qasr Ibrim can be tied to different clay sources. Fabric MR1 appears to be a mixed clay (kaolinitic and alluvial) that matches Smith’s Aswan/Meinarti clay group; therefore, it can be postulated that the clay was extracted in the northern region of Nubia, within the vicinity of Qasr Ibrim, Meinarti and Aswan. Fabric MR2 shows great similarity with the Aswani coarse fabric ASW2. It is also nearly identical to fabric MR1; however MR2 contains more quartz grains and pyroxenes. MR2 may therefore be made out of another form of northern mixed clay from the same area, yet selected from very close to the River Nile, where pyroxenes are found more abundantly in the clay. Fabric MR3 matches another fabric from Qasr Ibrim described by Smith (QI3), which is completely impossible to characterise on the basis of mineralogical analysis. Finally, fabric MR4 belongs to Smith’s southern clay group, possibly originating from Meroe. The existence of this fabric proves that pottery from the south of Kush could have travelled as far as the Qasr Ibrim region.