The Social Implication of a “mixed” Ceramic Assemblage: Understanding the Predynastic Community of Nag el-Qarmila (First Nile Cataract, Egypt) Through Pottery Technology

Abstract

This article investigates the social landscape of the First Nile Cataract region in the 4th mill. BCE through the application of petrographic and technological approaches to pottery analysis. The study focuses on the ceramic assemblages from the settlement WK15 and the necropolis WK14 at Nag el-Qarmila and identifies for the first time three technical traditions along with their synchronic and diachronic interaction. It provides new data on the more ancient phases of the so-called Naqadan productions and their relationship with the Shale Ware and the Nubian Black-Mouthed Ware. The results allow us to better frame the mixed assemblage that characterizes a site far from the elite centers and where relations between mobile and sedentary groups were, therefore, more fluid. The data obtained also constitute a comparative nucleus for the growth of technological and petrographic studies at other sites and periods, and the numerous insights that emerged will be useful for directing future research.

Résumé

Cet article étudie le paysage social de la région de la Première Cataracte du Nil au 4e millénaire avant notre ère en appliquant les approches pétrographiques et technologiques à l'analyze des poteries. L'étude se concentre sur les assemblages céramiques du village WK15 et de la nécropole WK14 à Nag el-Qarmila et identifie pour la première fois trois traditions techniques ainsi que leur interaction synchronique et diachronique. Elle fournit de nouvelles données sur les phases les plus anciennes des productions dites Nagadiennes et sur leurs relations avec la “Shale Ware” et avec la “Black-Mouthed Ware” nubienne. Les résultats nous permettent de mieux cerner l'assemblage mixte qui caractérise un site éloigné des centres élitaires et où les relations entre groupes mobiles et sédentaires étaient, par conséquent, plus fluides. Les données obtenues constituent également un noyau comparatif pour le développement d'études technologiques et pétrographiques sur d'autres sites et d'autres périodes, et les nombreuses aperçues qui en découlent seront utiles pour orienter les recherches futures.

Introduction

This paper explores the social implications of a mixed ceramic assemblage from a small Predynastic community that of Nag el-Qarmila in southern Egypt. The assemblage is made of ceramics commonly attributed on a typological basis to the so-called Egyptian Naqada culture, the Nubian A-Group culture, a Shale Ware with parallels in the Egyptian Western Desert, Upper Egypt, and the Sheikh Muftah culture, plus a few hybrid items. What is the reason for such a diversified array of pottery vessels and their use in combination by a small and peripheral community? What can this combination tell us about pottery modes of production? What can we learn about the people living in and making up the community, and their regional and supra-regional network of communities? To address these questions, a technological and petrographic analysis was performed for the first time on the pottery from the village and its adjoining cemetery to highlight the various chaînes opératoires that characterize the assemblage and understand whether the variability is functional or sociological. In particular, the technological approach allowed us to reconstruct the various manufacturing procedures present at the site, from clay collection to firing, while the petrographic analysis was mainly aimed at describing the fabrics in detail, both from a technological and mineralogical point of view. It also obtained information on the possible local or exogenous origin of the raw material and consequently better frame the various productions we have in Nag el-Qarmila. Furthermore, as a secondary objective, the two analytical approaches have made it possible to continue “mapping” the fabrics and the chaînes opératoires widespread in predynastic Egypt for future research.

Background

The First Cataract is located almost mid-way along the Nile, and it has been a crossroads since prehistory. The granitic outcrops and the rocky landscape have made the region a natural boundary, which subsequently became the political border between Egypt and Nubia (modern-day Sudan). The Nag el-Qarmila settlement and necropolis were both in use during the 4th millennium BCE, and their importance lies in being the southernmost Predynastic sites recently investigated (apart from the settlement of Elephantine, which is, however, slightly later and larger in size and scope). Another fundamental element is that we have both a settlement and a cemetery, a non-negligible factor in a research field in which most of the data comes from funerary contexts where the range of material is limited, and domestic productions are poorly represented. The preponderance of data from necropolises has pushed previous research to focus heavily on elites. In this case, however, the settlement material and adopted methodology have allowed us to have a bottom-up approach looking at the Predynastic society from the point of view of a small community established in a peripherical area of the soon-to-be ancient Egyptian state. The sites, located approximately 17 km north of Aswan on the west bank of the Nile, in a small valley facing the river, were partially excavated between 2007 and 2012 by the Aswan–Kom Ombo Archaeological Project—AKAP directed by M. C. Gatto (Polish Academy of Sciences) and A. Curci (University of Bologna), (Fig. 1; for a detailed description of the sites see: Gatto, 2014, 2016; Gatto & Siegel, forthcoming; Gatto et al., 2009a, 2009b). At that time, their preservation was good enough to provide valuable archaeological information; now they have been severely damaged by new housing. A series of areas were excavated in the village (WK15) and cemetery (WK14), as well as a storage facility located on the first terrace of the slope overlooking the valley from the south (WK22), (Fig. 1).

Fig. 1

Map with the location of the sites WK14 and WK15 in Nag el-Qarmila and details of the areas excavated by the Aswan-Kom Ombo Archaeological Project. Elaborated with QGIS 3.22 using a satellite image from Bing as a basis by S. Nicolini

The cemetery WK14, found on the northern wadi terrace, was poorly preserved because of the sandy sediment in combination with wind erosion, ancient plundering, and modern disturbance that resulted in major post-depositional damage to the tombs. The area of the cemetery was defined by a concentration of artifacts; its maximum extension was estimated to be around 65 × 50 m, of which approximately half was excavated. The tombs found in the north-western portion of the site were placed on a Middle Pleistocene dune, the rest were dug into a Late Pleistocene alluvial deposit. Out of 60 identified features, 32 were excavated and turned out to be tombs, of which only two were intact (Tombs 14 and 59).

The settlement WK15 is located near the cemetery but closer to the wadi bottom on a Late Pleistocene alluvial deposit and was indicated by a concentration of material, and the maximum extension of which was estimated to be approximately 60 × 60 m. The analyzed ceramics come from excavation Area B, a 5 × 8 m unit that was investigated to evaluate the nature and state of preservation of the site. The remains of the settlement primarily consist of fireplaces, postholes, and in situ material. No areas dedicated to specific activities have been identified. Three radiocarbon determinations were made on samples from the first occupational phase (composed of two main occupation layers: Layer 16 is the earliest in the sequence, superimposed by Layer 11), providing a date for the village’s main occupation to the first half of the 4th mill. BCE. According to the pottery assemblage retrieved from those early layers, and from a second, younger occupational phase, closer to the surface and with a poorer state of preservation (Layer 3), the first phase of the village was dated by S. Hendrickx to Naqada IC-beginning of IIA of the Predynastic relative chronology, and the second phase to Naqada IIC-IIIA2, following the understanding of the time (Hendrickx, 2006). In light of an ongoing revision of the Predynastic ceramic chronology, mainly based on the work by R. Hartmann at Abydos (Hartmann, 2016) and N. Buchez at Adaima (Buchez, 2021), we also revised the chronology of the beginning of the occupation phase of the site. According to these studies, some ceramic forms until now attributed to the Naqada IC phase have been backdated, widening the chronological range to Naqada IA-IC. Since we found similar material in Nag el-Qarmila, we also decided to broaden the chronological range of the first phase of the settlement to Naqada IA-beginning of IIA, and we aim to publish the details of this revision in a specific article. This wider chronological range remains compatible with the radiometric determinations of 3800–3300 BCE (Dee et al., 2013). Concerning the second phase of the settlement, we confirm the chronology proposed by Hendrickx (Naqada IIC-IIIA2). The assemblage from the cemetery is characterized by ceramics ranging from Naqada IA to Naqada IIIA2, but its heavy disturbance does not allow a clear distinction for the stratigraphic distribution of the material.

The first classification of the ceramic assemblage from the village and cemetery of Nag el-Qarmila was performed by the late S. Hendrickx, who created a chrono-typology mainly based on diagnostic fragments from the 2007-2008 excavations, and by the late H.-Å. Nordström, who analyzed the Nubian fabrics (Gatto et al., 2009a, 2009b). Now we have resumed the study of the ceramic assemblage by incorporating the chrono-typological analysis into a technological and petrographic approach to be able to address more in-depth inquiries that go beyond the mere dating and cultural affiliation of the ceramic vessels, as recently formulated by Gatto (2021). The aim was to better understand the social world in which such a multi-tradition assemblage came into being.

The chaînes opératoires approach is particularly suitable for obtaining an anthropological reading of archaeological materials and consequently addresses our questions. Research in experimental psychology and ethnography has indeed shown how the passing down of technical traditions takes place within the same social group, with a tutor educating the apprentice about the location of supply sources, properties of the materials, tools, techniques, and gestures (Reed & Bril, 1996; Bril, 2002). At the end of this process, the trainees will have developed specific cognitive and motor skills to make objects according to the models shared by their social group, and changing the manufacturing routine becomes very difficult (Dobres, 2000; Gosselain, 2000). As a result, the learning processes constitute the basis of the creation and transmission of technical traditions (meaning with this term a recurrent way of doing which is passed down from generation to generation within a determined social group), (Gosselain, 2000; Roux, 2019). Furthermore, the same pot-shape can be obtained with different techniques (molding, coiling, percussion, wheel throwing, etc.); therefore, the chosen ones also represent a cultural choice. These dynamics ensure that the chaînes opératoires are stable over time. Their association with evidence from archaeological contexts allows us to address broader questions about social boundaries, interactions between groups and their subsequent evolution, and to formulate hypothesis on the nature of the identified social groups to which the artisans belonged (professional community, itinerant artisans, communities with different subsistence patterns, etc.) (Stark, 1998; Roux, 2019; Roux et al., 2017). The chaînes opératoires adopted to make pots are the result of the transmission of technical traditions that are closely linked to the distinctiveness of the groups to which the potters belong and are consequently an invaluable source of information for investigating the social sphere of ancient societies (Gosselain, 2000; Stark et al., 2008; Roux et al., 2017; Roux, 2019).

Materials and Methods

The adoption of the technological approach (also including petrography) was motivated by its effectiveness in addressing issues such as those raised for the Nag el-Qarmila sites, as demonstrated by its application on contemporary assemblages from the Nile Delta (Bajeot & Roux, 2019; Bajeot & Buchez, 2021, 2022) and Upper Egyptian material kept in museum collections (the latter as part of the TECHNOPREGYPT Project, funded by the National Science Centre of Poland and the European Union through the POLONEZ BIS program, reg. No. 2021/43/P/HS3/03262). Moreover, this methodology is particularly suitable for Egyptian archaeological material as it can be easily implemented during fieldwork, without the need for expensive equipment, since exporting any type of sample for analysis is prohibited. Even in the case of thin sections, in fact, the samples were made and analyzed in the field.

Petrographic Analysis

Fourteen thin sections were prepared and analyzed at the Kom Ombo storehouse in the frame of a larger sampling campaign that included 50 samples from excavations carried out by the AKAP mission in the Aswan area. Twelve samples come from the necropolis and two from the settlement (Table 1). The analysis of the thin sections with a petrographic microscope at 100× magnification followed standard procedures (see Reedy, 2008). For each section, its color in plane (PPL) and cross-polarized (XPL) light was noted, an estimate was made for the frequency of inclusions relative to clay matrix, and the sorting of the inclusions was specified. The minerals identified in the thin section were listed as those that represent the main inclusions and those that are less common. For the inclusions, both their general shape and size range were noted. Information on raw material source comes from the Geologic Map of Egypt (1981) and from extensive previous petrographic work on Egyptian pottery (see Ownby & Brand, 2019).

Table 1 Thin sections sample list

Technological Analysis

The ceramic assemblage, kept in the storehouse in Kom Ombo, was viewed in its entirety and almost all of it was studied. Concerning the settlement WK15, a total of 2499 samples including diagnostic pieces and bodysherds coming from Area B (from now on WK15B) were analyzed. Material from phases 1 (187 diagnostic samples and bodysherds) and 2 (1164 diagnostic samples and bodysherds) was prioritized, while material from the surface was sampled (284 diagnostic and bodysherds, the largest and most suitable for technological analysis). Part of the sherds previously studied by S. Hendrickx (ca. 150 diagnostic samples) were excluded to give precedence to those that had never been seen. The material from the necropolis WK14 has been analyzed in its entirety (1204 diagnostic samples and bodysherds). For details on the number of bodysherds and diagnostic samples see Tables 2, 3, and 4.

The analysis of the sherds was carried out with the naked eye and employed a binocular microscope (×10–×40 magnification) and a digital microscope (Dino Lite AM4113ZT). The samples were examined following Roux’s classification criteria (Roux, 2016, pp. 257-274, Roux, 2019 pp. 217-232) based on three successive rounds of sorting: the first sorting is by technical groups (identification of the manufacturing process), the second is by techno-petrographic groups (identification of the petrographic groups within each technical group), and the third is by techno-morphological and stylistic groups (classification of the morphological and stylistic types within each techno-petrographic group).

The identification of the manufacturing techniques had as a prerequisite the combination of at least two or more diagnostic traces (preferential fractures, characteristics of the radial sections, manufacturing traces visible on the internal and external surfaces). The diagnostic traces have been identified and interpreted by referring to experimental and ethnographic reference data (e.g., Forte, 2020; Livingstone-Smith, 2007; van Doosselaere, 2011). The manufacturing traces have been described following the terminology proposed by Roux (2019).

The ceramic material is generally quite well preserved even if light alterations of the surfaces are diffused and, less frequently, the surfaces are heavily eroded due to exposure to wind. As expected, the material from the necropolis is better preserved and has yielded the greatest number of more complete forms.

Results

Three techno-petrographic groups have been identified and called NAQ (technical tradition referring to the so-called Naqada culture), SHALE (technical tradition characterized by shale fabrics, possibly linked to Western Desert nomads of the Sheikh Muftah culture), and MNR (Middle Nile Region, technical tradition linked to the Nubian A-Group culture). The names of the groups were chosen to recall those known in the literature and therefore make it easier to follow the analysis.

In Tables 2, 3, and 4 and in the description of the groups, we give the breakdown of the analyzed samples for WK15B and WK14 but we must underline that while the assemblage from WK15B is divided into two chronological phases (albeit large), based on stratigraphy, the lack of a clear stratigraphy for the assemblage from the cemetery WK14 forced us to treat the material as a single assemblage dating from Naqada I to Naqada IIIA. For this reason, we do not give the relative percentages of the techno-petrographic groups for WK14 in the tables, as they would give a distorted picture of the relative presence of the various groups. For a disturbed site such as that of WK14 (very common condition for Upper Egyptian sites) in the future, it will be necessary to attempt a seriation work of the Naqadan diagnostic sherds to try to identify the chronological development of the necropolis and also to date the associated A-Group pottery, and possibly the Shale ware, which cannot currently be dated on a typological basis.

The NAQ Group

The NAQ is represented by the sub-groups NAQ-SILT (WK14 = 175 diagnostic samples and bodysherds; WK15B phase 1 = 74 diagnostic samples and bodysherds, phase 2 = 299), NAQ-MIXED (WK14 = 122 diagnostic samples and bodysherds; WK15B phase 1 = 5 diagnostic samples and bodysherds, phase 2 = 5), and NAQ-VEG (WK14 = 232 diagnostic samples and bodysherds; WK15B phase 1 = 4 diagnostic samples and bodysherds, phase 2 = 150) (Table 2) that are characterized by the same manufacturing procedure described below except for the fabrics and surface finish that differ.

Table 2 Classification of the diagnostic samples and bodysherds from the cemetery WK14 and the settlement WK15B. The percentages relating to WK14 should be considered only as suggestive since the chronological span is extremely broad. Elaborated by integrating our data with that collected by S. Hendrickx in 2009

The NAQ-SILT Sub-Group

During the first occupation phase of the village (Naqada IA-beginning of IIA), it constitutes 40% of the settlement assemblage, and 25% in the second phase (Naqada IIC-IIIA) (Table 2).

The fabric characterizing this sub-group is made of alluvial clay with 40% frequency of mineral inclusions according to the petrographic analysis of thin sections (corresponding to the estimate of 5–10% of minerals obtained at low magnification of ×10 based on the quantification chart published in Courty et al., 1989 and that we include here alongside the petrographic estimation to facilitate comparison with our previous works where petrographic analysis was not done) (Fig. 2). The petrographic data suggested all inclusions originate from within the clay deposit. The homogeneous gradient of minerals and the absence of plant remains points towards the acquisition of a naturally fine clay (raw material sources chosen specifically for these characteristics) or processing of the clay through sieving or decantation. All five samples subjected to petrographic analysis are silty in thin section with a similar assemblage of minerals such as quartz, feldspars, mica, pyroxene, and amphibole (Table 1, Fig. 3a). One Red-Polished bowl (WK14.6) has a siltier clay than the other sherds, which comprise two Black-Topped beakers, a Black-Topped vessel, and a Red-Polished bowl. The paste is quite compact, with scattered vacuoles, suggesting moderate wedging.

Fig. 2

Techno-petrographic groups mentioned in the text. Photos taken with a digital microscope by J. Bajeot

Fig. 3

Thin section images in cross-polarized light. a Sample WK14.5, NAQ-SILT. b Sample WK14.10, NAQ-MIXED. c Sample WK14.8, SHALE. d Sample WK15B.2, MNR. Photos by M. Ownby

All types of NAQ-SILT vessels (for the typology see Tables 3 and 4 and the end of this section) were roughed out through the coiling technique, with small coils (height between 1.5 and 2 cm) placed from the inside. This technique results in radial sections to oblique coil junctions, oblique porosity, and concentration of vacuoles in correspondence with the joints (Fig. 4a–c). On the external and internal surfaces, the coils are slightly perceptible to the touch and are sometimes visible on the internal surface (Fig. 4b). The coiling technique is also sometimes witnessed by preferential horizontal fractures. The base is made from a spiral coil sometimes evidenced by the presence of a bulge in the center of the base and, in radial section, by elongated oblique junctions and oblique porosity on either side of the center of the base (Fig. 4b). The peripheral coil of the base was then stretched obliquely to attach to it the first coil of the wall.

Table 3 Typology chart based on diagnostic sherds from cemetery WK14. The percentages are not provided as they would be misleading due to the impossibility of identifying chronological phases for the necropolis. Elaborated by integrating our data with that collected by S. Hendrickx in 2009

Table 4 Typology chart based on diagnostic sherds from settlement WK15B. Elaborated by integrating our data with that collected by S. Hendrickx in 2009

Fig. 4

Chaîne opératoire of the NAQ group. a–c Spiraled base and oblique coils visible on a NAQ-SILT vessel (Black-Topped). d Spiraled base and oblique coils visible on a NAQ-MIXED section (Wavy-Handle). e Oblique junction on a NAQ-VEG section (jar). f Shaving marks on a NAQ-SILT vessel kept in the Egyptian Museum, Turin (Black-Topped); shaving marks on a NAQ-MIXED bowl. Photos by J. Bajeot

On the medium-low part of the external surfaces, traces of shaving are sometimes still visible (Fig. 4f). This manufacturing step consisted of shaving the leather hard clay with a sharp tool designed to remove clay chips to thin the walls and obtain the final shape, especially in the case of bases. This technique differs from scraping that is instead applied on wet clay with discontinuous pressures exerted with a rigid tool to profile the walls. The diagnostic traces of shaving are flattened areas and deep striations which start from the base and go up the walls with a slightly radial orientation. The shaving is hardly visible on the sherds belonging to this group since the traces are erased by subsequent surface treatments. However, the possibility of studying complete samples from museum collections allowed the first author to extensively document this technique.

The inside of the vessels was smoothed with the hands and with a hard tool, as testified by a combination of reticulated ribbed striations left by fingers and arrest marks of the dragged tool (Fig. 5b). This smoothing, which makes the surface very compact, tends to be horizontal or multidirectional. The inner rim is characterized by horizontal smoothing with more evident threaded striations. With the exception of the bowls which are also burnished on the inside, there does not seem to be any particular importance given to the aesthetic aspect of the interior of beakers and jars with wide openings. In fact, the inner surface is often characterized by evident undulations and smoothing traces. On the contrary, the exterior was carefully smoothed to obtain a surface as flat and even as possible to then perform a perfect burnishing.

Fig. 5

Chaîne opératoire of the NAQ group. a Slipped and burnished surface of a NAQ-SILT vessel (Black-Topped). b Smoothed internal surface characterized by arrest marks and finger imprints. c External smoothed surface of a NAQ-VEG jar. d Smoothed internal surface characterized by arrest marks and finger imprints. Photos by J. Bajeot

Slipped and Burnished (Black-Topped and Red-Polished)

This fabric and chaîne opératoire are used to produce red vessels with a black band around the rim (Black-Topped, Petrie, 1921) and red vessels (Red-Polished, Petrie, 1921). The outer surface of all the NAQ-SILT pots was slipped with an iron-rich slip (sample WK14.6, Ownby, 2022). The slip is homogeneous and covers the vessel completely, being found on the external surfaces, including the base, and stopping at the inner rim (Fig. 5a). Then, the external surfaces of closed forms and the external and internal surfaces of open vessels were burnished when the clay was dry, as evidenced by the gloss and by very fine and narrow streaks with clear edges only visible under the binocular microscope. The term Polished in the name “Red-Polished” used here does not refer to a different surface treatment; all the vases were burnished but the name of this category in the literature has been retained.

The color of the surfaces is rather homogeneous as is that of the breaks, with completely oxidized fractures for the walls of most of the body and completely reduced and black for the walls of the rim, suggesting a good control over firing (Fig. 4b). The firing temperature was estimated petrographically between 800 and 850 °C based on moderate optical activity of the paste as at higher temperatures the clay minerals are sintered reducing optical activity (see Rice, 1987, p. 431).

The different characteristics between Black-Topped and Red-Polished vessels were obtained through a diversified firing procedure: completely oxidizing for the Red-Polished and a combination of oxidizing atmosphere and reduction of the rim band in the case of Black-Topped.

While the Black-Topped vessels have external surfaces characterized by a deep red and black color, the inside is often gray and beige (Fig. 4b, observation also made on samples from museum collections). Small vases tend to have a darker interior, and medium and large ones tend more towards beige. No chronological or spatial variations of this phenomenon were observed. This element contrasts with the deep black internal surfaces obtained during firing experiments (Hendrickx et al., 2000; Baba & Saito, 2004, Smythe, 2005), highlighting that further research on this subject would be needed to better assess the firing process.

Within our assemblage, the red burnished finish mainly characterizes bowls (23 samples from WK14, 3 from WK15B phase 1 and 5 from phase 2) and more rarely jars (3 samples from WK14 and 2 from WK15B phase 2) (Fig. 6m–o), and one carinated bowl which we highlight here for its importance in dating the most ancient phase of the site (Fig. 6p, similar to Hartmann, 2016, Taf.129, U-412.1365) (Tables 3, 4, and 5).

Fig. 6

Typology of the NAQ group. Drawings by S. Hendrickx and H. Joris, elaborated by J. Bajeot

Table 5 Chronological distribution of the techno-petrographic groups and of the diagnostic forms based on the stratigraphy of WK15B and on the general chronology for Naqadan ceramics (here NAQ technical tradition) and parallels with contemporary sites. Question marks indicate that a determined group is not datable, or that we are not sure about the presence of a determined group in a specific chronological phase due to the stratigraphic problems related to the more superficial layers of WK15

The black-topping mainly marks beakers with straight or slightly flaring rims (25 samples from WK14, 3 from WK15B phase 1 and 9 from phase 2) (Fig. 6c–g), jars with wide opening and direct rim (21 samples from WK14, 2 from WK15B phase 1 and 1 from phase 2) (Fig. 6h–l), small cups (2 from WK14) (Fig. 6a–b), (similar to Hartmann, 2016, Taf. 150, U-531.1576, U-532.1577), and bowls (11 samples from WK14, 2 from WK15B phase 1 and 3 from phase 2) (Tables 3 and 4).

According to typology, both Red Polished and Black-Topped vessels can be dated from Naqada IA to Naqada IIA, (Table 5) (Hartmann, 2016, Abb. 151; Buchez, 2021).

The NAQ-MIXED Sub-Group

This sub-group is made of a total of 122 samples from WK14, 5 from WK15B phase 1 (2.5%) and 5 from phase 2 (0.5%) (Table 2).

In this group, there are possibly various fabrics (marls and mixed clays) (Fig. 2). The field analysis mainly focused on the preparation of the clay paste and on manufacturing processes, relying on previous descriptions that included this material in the “marls” group of the Vienna System, which includes a wide diversity of calcareous and shale clay pastes (see Nördstrom & Bourriau, 1993). The petrographic analysis indicated that there may be more types of fabrics. From the point of view of clay preparation, these clay pastes have a rather homogeneous appearance (fine clay, inclusion size, and compactness) and are difficult to distinguish with the naked eye. So only the fabric of the two samples analyzed in thin section are described (Table 1 and Fig. 3b), and we aim to do a more in-depth analysis of this category of fabrics in the future.

The two petrographically analyzed samples (a Wavy-Handled jar and a large bowl), featured a now well-known paste of shale clay with common micritic limestone and microfossils (see Ownby and Köhler, 2021, Mixed (shale) clay fabric). It has been suggested this raw material comes from around the Qena bend, and the clay is from a naturally mixed deposit where shale is eroding adjacent to limestone deposits. The fabric is characterized by 10–20% of sand-sized mineral inclusions of homogeneous size (percentage established macroscopically according to the quantification chart by Courty et al., 1989). Petrographically, the amount of inclusions was recorded at 40% and are mostly quartz and feldspar grains, limestone pieces, and shale fragments. Most are fine-sand sized (see Wentworth, 1922). Organic inclusions are generally absent. The paste is quite compact, characterized by low porosity and a quite dense texture indicating thorough kneading.

The manufacturing chaîne opératoire for the production of all the pots included in this sub-group is the same as described for the NAQ-SILT group, coiling with internal apposition, and spiraled bases with the peripheral coil stretched to attach to it the first coil of the wall (Fig. 4d). The height of the coils is rather homogenous and comprised between 1.5 and 2 cm. The only slight variation in the chaîne opératoire has been observed in the case of a very distinctive type of globular painted vessel (D-ware) with a large, rounded base (1 sample recorded in WK14 but well known in museum collections). The lower part of the body was obtained by coiling into a concave mold inside which the coils were assembled and joined. This technique is testified by the very regular and homogenous profile of the outer surface contrasting with the evident undulations visible on the internal surface and by the shaving marks visible at the largest diameter, where the part built in the mold is joined to the upper part obtained by free coiling.

Shaving is a step that characterizes all NAQ productions and is particularly visible in this sub-group (Fig. 6g). Except for the globular shapes mentioned above, all types of open and closed forms show evident shaving traces on the outside in the medium-low part. It is done from bottom to top, starting at the base and with a slightly radial movement that goes around the vessel.

The homogeneous pinkish, orange, and beige color (or red in the case of slipped vessels) of the internal and external surfaces suggests a controlled firing process. Also, the fractures are homogeneously oxidized with the very rare presence of a slightly darker, but always pink, inner core. Petrographically, they are suggested to have been fired between 800 and 850 °C based on the optical activity of the paste.

The finishing treatments are as follows: smoothed, softened, slipped and burnished, smoothed/softened, and painted.

The rarity of this group in phase 1 of the settlement and their later dating compared to the dating of phase 1 (Naqada IA-beginning of IIA) suggests that they may be intrusive (Table 2).

Smoothed

Only one bowl (WK15B phase 2) is characterized by a shaved and smoothed surface, without other finishes.

Softened

After smoothing and shaving, the leather-hard paste was re-humidified and rubbed with a rigid tool. This step results in a compact microtopography, inserted grains, and fine thickenings with scalloped edges linked to the displacement of leather-hard clay.

This finish has been observed on the Wavy-Handled jars (3 samples from WK14 and dating to Naqada IID, similar in shape to W 47-49, Petrie, 1921) 2 fragments of jars with rolled rim and 1 bowl fragment from WK14 (Naqada IID-IIIA) (Table 5, Fig. 6s–t).

Slipped and Smoothed

In this sub-group, we only find bowls with the inside and the outer rim band red slipped and smoothed. The slip is very compact and thickly covers the inside, while it is thinner on the external rim band. Three samples are recorded in this group: 1 from WK15B phase 1, and 2 samples from WK14 (Naqada IID-IIIA, Petrie’s L 3, 7b, 7g) (Table 5).

Slipped and Burnished

In this sub-group, we only find bowls with the inside and the outer rim band red slipped and burnished, dated to Naqada IID-IIIA (Table 5). The nature of the slip coverage is identical to that of the previous group. The burnishing is not perfectly uniform and is mostly in horizontal panels with overlapping edges on the inside, while on the outer rim, it is horizontal. To this group belong 2 fragments from WK14 and 2 from WK15B phase 1.

Smoothed/Softened and Painted

This group features fragments of oval and globular pots decorated with spirals and “fish flakes” painted decorations (D-ware). The red color was applied on the dry surface after smoothing of leather-hard clay or softening by rubbing the re-humidified surface with a hard tool (very often these two surface treatments are not clearly distinguishable and both types of traces can be found on the same sherd). We have two fragments of D-ware from WK14, one with spiraled and one “fish flakes” decoration (Petrie’s types D 63a and 67) (Fig. 6q–r). According to typology they can be dated from the end of Naqada IIB to Naqada IID (Hartmann, 2016, Abb.151; Buchez, 2008, figs. 3, 5–7) (Table 5).

The NAQ-VEG Sub-Group

For this sub-group, no samples were examined petrographically as it has been well-characterized through previous studies and is easily distinguished at low magnification. Therefore we present the macroscopic description here:

the fabric is made of alluvial clay with 5% of mineral inclusions and 20–25% of plant temper (estimated at ×10 magnification based on the quantification chart by Courty et al., 1989) (Fig. 2). The dough was not worked long as demonstrated by the presence of vacuoles, very often triangular in shape, left by the air trapped in the dough and not completely evacuated during wedging. The plant fragments reach a maximum length of 6 mm and are very often still intact.

The NAQ-VEG assemblage is composed of 232 diagnostic samples and bodysherds from WK14 and by 4 diagnostic samples and bodysherds from WK15B phase 1 (2%) and 150 from phase 2 (13%).

The manufacturing procedure for obtaining all the shapes made with this fabric is the same as described for the other NAQ sub-groups as witnessed by the same diagnostic traces. For this reason, we recall here only the main points: the base was made with a spiraled coil and the walls with coils fixed with internal apposition. Then the leather-hard clay was preformed by shaving (Fig. 5e). The height of the coils is rather homogeneous and between 1.5 and 2 cm.

The only recorded finish is smoothing made on a wet paste that left fluidified striations and reticulated ribbed striations relating to fingerprints (Fig. 5c).

The typology of the vases made with this fabric is wider than the other groups (Tables 3 and 4) and can be dated from Naqada IIC to Naqada IIIA (Table 5). It consists of large jars with everted rolled rims and flat bases (1 sample from WK15B phase 1 and 1 from phase 2, and 21 samples from WK14 dating from Naqada IIC to Naqada IIIA, of which 5 Petrie’s R84), Petrie’s R 81 tapered jars with wide opening and small rounded base (4 samples from WK14, Naqada IIC) (Fig. 7u), one distinctive bottle of Petrie’s type R 91a-c from WK14 and dated to Naqada IIB-C (Buchez, 2008, fig. 3), vats (3 samples from WK14), tronco-conical bowls (Naqada IIC-IIIA, Petrie’s R 26, 5 from WK14) and one sherd belonging to a concave bowl with simple rim and bumpy external surface from WK14 and dated to Naqada IIIA (Buchez, 2008, fig. 5; Bajeot & Buchez, 2021, fig. 6) (Table 5).

Fig. 7

Chaîne opératoire of the SHALE group. a, b oblique coils and spiraled base. c Complete pot characterized by horizontal preferential fractures, finger impression on the median part, and scraping marks toward the bottom. d, e Slipped (?) and burnished surfaces. f, g Smoothed surfaces. Photos by J. Bajeot

The SHALE Group

This group predominates in the settlement with 56% in phase 1 (98 diagnostic samples and bodysherds) and 59% in phase 2 (683 diagnostic samples and bodysherds), together with the NAQ-SILT which stands at 40% in phase 1 and 25% in phase 2 (Table 2). It is also well represented in the necropolis with 206 diagnostic samples and bodysherds out of a total of 1204 (Table 2).

This assemblage is marked by a fabric made from shale clay naturally characterized by very distinctive laminar shale inclusions of variable color and angular and sub-angular fragments of minerals (Figs. 2 and 7a). Of note, in none of the samples did the shale inclusions appear as temper. The quantity and size of the inclusions are variable and, in some samples, sand is also present. The frequency of inclusions has been estimated at around 30% according to petrographic analysis (20% through macroscopic observation at ×10 magnification based on the quantification chart by Courty et al., 1989). The four samples analyzed in thin section all had slightly different features, possibly indicating varying areas of production or utilization of different raw material deposits within the same region (Table 1 and Fig. 3c). The shale fragments are similar in appearance to the clay paste, suggesting a shale deposit weathering in place with natural shale pieces that were not fully rehydrated back to clay. Sample WK14.8 (hemispherical bowl with burnished inner and outer surfaces) has some sand and silty quartz possibly suggesting a Nile clay component that could be natural to the clay deposit. Rare micritic limestone is present, along with rare fine-grained sandstone rock fragments. Sample WK14.11 (sherd belonging to an open form with smoothed surfaces) has frequent coarse-sized shale inclusions with some shale fragments also containing silty quartz, which is common in the paste. Very rare are micritic limestone fragments. Sample WK14.12 (sherd belonging to an open form with burnished surfaces) is quite similar to both, but the shale fragments lack silty quartz and micritic limestone is slightly more common. Somewhat different is Sample WK15B.1 (sherd belonging to an open form with smoothed surfaces) with coarse shale fragments and some micritic limestone. Common quartz and feldspars are either sand natural to the secondary shale deposit or some Nile clay that is also likely inherent to the raw materials. The presence of sandstone and siltstone fragments suggests a source near to these deposits that are known in southern Egypt and the southern oases. A petrographic analysis of Neolithic pottery from the Western Desert noted some with sandstone parent material, but the presence of shale was not mentioned (Zedeño, 2002). As already observed by Hendrickx and Nordström (Gatto et al., 2009a), the variation in fabric coarseness seems random as it is not connected with specific pottery shapes, sizes, and wall thickness. This element possibly suggests the use of different local clay sources and the lack of attempts to homogenize the general aspect of the fabric. The paste is quite well worked as evidenced by the small and sparse pores, but there are also large vacuoles.

The pots belonging to this group are represented by two form categories: bowls with rounded bases and simple rims, and conical and hemispherical deep bowls of various dimensions with rounded bases and simple rims. Both categories are characterized by a marked variability (Fig. 8a–d). Because of their simple typology that does not seem to change over time, they cannot be dated typologically but only on the basis of stratigraphy. Concerning WK15B the SHALE characterizes the two phases of the village with important quantities. However, based on the data at our disposal, it is not possible to say when this production declines (Table 5).

Fig. 8

a–d Typology of the SHALE group. e–i Typology of the NMR group. Drawings by S. Hendrickx and H. Joris, elaborated by J. Bajeot

These pots are entirely built with coils witnessed on the outer surfaces by undulations and fissures of the coil junctions and by very common preferential fractures (Fig. 7c). In radial section ca. 2–2.5 cm high coils with oblique junctions are clearly visible (Fig. 7a, b). Inclusions and pores are in fact organized in a radial pattern in correspondence to the center of the coil (due to the rolling movement used to prepare the coils) and by an oblique orientation in correspondence to the junctions. About half of the samples have coils positioned with internal apposition (therefore with junctions having the same orientation), and the other half shows alternating coils. No specific patterns related to shape or surface finish were noted for this variation of the coiling technique. The presence of the two varieties in phases 1 and 2 of the settlement suggests that they are contemporary, but the extreme rarity of in situ material, and problems linked to the stratigraphy and consequently to precise chronological phases do not allow us to be certain. It is certainly an element that will need to be explored more thoroughly in the case of analysis of Shale ceramics from other sites in the future. The bases are made with a spiraled coil possibly laid in a concave support in the case of larger shapes as witnessed by the frequent presence of a horizontal depression that runs around the lower part of the vessels (Figs. 7b and 8a, b, Buchez, 2004, Fig. 3.15-17). The outer surface is characterized by a very distinctive bumpy aspect that is due to the overlapping of vertical finger impressions left during the assembly of coils with discontinuous pressures, and by oblique and horizontal finger impressions left by the hands in support of the wall during the regularization and smoothing of the inside (Fig. 7c, d, and f). Sometimes some deep striations indicate a scraping stage on wet clay whose traces are in most cases erased by subsequent smoothing (Fig. 7c). The mouth of the vessels was left slightly wavy (Fig. 7c–g).

According to the petrographic analysis, this pottery was possibly fired around 750–800 °C due to the higher optical activity of the paste, and in an oxidizing atmosphere. Surfaces and sections are always completely oxidized but the surface color is not homogenous and ranges from beige to pink to reddish. We Recorded two Finishes: Smoothing and Burnishing.

Smoothing

This finish characterizes both deep bowls (24 samples from WK14, 23 from WK15B phase 1 and 19 from phase 2) and shallow bowls (27 samples from WK14, 2 from WK15B phase 1 and 6 from phase 2). Although it seems to be more frequent on deep forms (Tables 3 and 4).

The finishing is more precise on the internal surface which is better regularized and smoothed than the outside (Fig. 7f, g). The more or less compact surfaces point towards smoothing performed at different stages of drying, from wet to leather-hard clay. On the external surface smoothing is testified by a fluidified surface and protruding grains, and it was possibly obtained with the hands (reticulated ribbed striations) and/or with a soft tool (more regular threaded striations and distinctive traces due to the detachment of the tool from the surface) and possibly also the use of hard tools (clay over-thicknesses related to the displacement of clay with the instrument). The smoothing of the interior also shows the use of different types of tools. In all cases, the smoothing is horizontal below the rim and multidirectional on the rest of the surface. The variability of smoothing is very high and forms a continuum. It does not therefore seem to be related to the shape or size of the containers but to a non-systematic approach to this processing step.

Most of the sherds belonging to this group bear soot traces possibly caused by their use as cooking pots (Buchez, 2004; Friedman, 1994, p. 633, 669). These carbon traces have a dusty appearance and are quite superficial, they never affect the section, therefore they are not traces due to the firing of the vases themselves (Fig. 7c). Furthermore, the fragments belonging to this group are the only ones to have these traces, so we can exclude that the soot is due to post-depositional processes (Forte et al., 2018). In the majority of cases, the soot is localized only on the outer surfaces.

Burnishing

Besides the smoothed forms, 18 fragments of shallow bowls and 1 sample of deep bowl from WK14 are marked by a more intense pinkish-red burnished surface that seems to suggest the presence of a covering slip obtained from earthy hematite with added water, of which some traces can be seen in thin section. A deeper analysis would be needed to confirm it, as no clear traces such as dripping have been observed and the burnishing tends to darken the surface color and create a thin layer that may imitate the effect of a slip.

Both the internal and external surfaces are burnished (Fig. 7d, e). Inside it tends to be more covered and compact, while outside the covering is uneven, due to the bumps and striations that characterize the surface and which do not allow for a good finish. The burnishing can be more or less shiny, indicating that it could be carried out at various drying stages of the clay. As with the smoothed forms, no pattern was observed for this variability, which again seems to be attributable to a non-systematic approach to the execution of this finish. This surface finish is mainly found on small and medium-sized bowls, whose form and function probably overlap with that of the NAQ-SILT Red-Polished bowls, and is rarely applied on cooking pots (Table 3).

The MNR Group

This group constitutes a minority of the assemblage from the village (3% in the first phase and 3% in the second) but is well represented in the necropolis (346 samples out of a total of 1204) (Table 2).

The group MNR mainly consists of the so-called Black-Mouthed Ware, shallow or deep open forms with direct rims and rounded or slightly pointed bases characterized by a black mouth and internal surface, and a reddish-brown external surface (2 open deep forms from WK15B phase 1, 2 bowls and 1 deep open form from WK15B phase 2, 116 deep open forms and 8 bowls from WK14) (Figs. 8e–i and 9c–g); and by two fragments of red burnished bowls with rounded base from WK14 (Tables 3 and 4).

Fig. 9

Chaîne opératoire of the MNR group. a, b Oblique coil junctions. c Black-Mouthed burnished jar with non-covering slip. d Black-Mouthed jar with slightly rippled surface. e Burnished internal surface. f, g Sherd with external rippled surface (slip and burnishing are only partially preserved due to poor conservation state) and burnished inner surface. Photos by J. Bajeot

The fabric is made of Nile clay with mineral and rare plant inclusions whose frequency has been petrographically estimated at around 40-50% (Table 1 and Fig. 3d), (low magnification estimation based on the quantification chart by Courty et al. (1989) is 10–15%), (Figs. 2 and 9a, b). Sample WK14.1 has Nile clay with some silt-sized but also some medium-sized and rare coarse-sized grains. Calcrete fragments (recrystallized calcium carbonate on soil deposits) are frequent. Sample WK14.2 has the more typical silty Nile clay. Calcrete is absent. Sample WK15B.2 is also a silty Nile clay lacking calcrete. The variation in Nile clay can reflect natural variation within the source material. However, calcrete has been noted within the Nile clay used for Nubian-style vessels from the Aswan area (Ownby, 2022), so it could suggest at least one vessel was made from similar raw materials. This may be further supported by the other two vessels being very similar to the NAQ-SILT sherds analyzed petrographically. Possibly these were made in a different area more commonly used to make NAQ ceramics. The paste overall has a compact aspect, with small and sparse vacuoles, and was therefore quite well wedged.

All the pots belonging to this technical tradition are made according to the following chaîne opératoire: the base was formed from a spiraled coil as evidenced by the very common preferential fractures caused by the detachment of the base; by the presence of a bulge in the center of the base; and, in radial section, by elongated oblique junctions and oblique porosity on either side of the center of the base. The walls were made of coils assembled with internal apposition as shown by oblique coil junctions and by the oblique porosity visible on radial sections (Fig. 9a, b). The coils vary in size from one vase to another (between 1.2 and 3.5 cm), and this variability seems random, as it is not linked to the shape or size of the vase. The very homogeneous thickness of the rather thin walls, the regular profile, the lack of perceptible undulations, the surfaces with inserted grains and compact microtopography, the lack of striations, and the rounded bases, suggest preforming by percussion on leather hard paste to give the vessels their final shape. Possibly after the percussion stage, two coils were added to form the rim. The two coils that constitute the rims are, in fact, more perceptible to the touch as they were not subject to compression, and they are more visible in the radial section. The percussion stage also left a distinctive, slight rippled texture more commonly found on the outer surface and rarely on the outer and inner surfaces (Fig. 9d, f). Sometimes this texture is visible only with oblique light and can cover the entire external surface or be limited only to the lower part. The ripple was probably caused by contact with a mat laid in a hole, or in a concave support, on which the vessel was hammered from the inside (comparable traces have been observed on Dogon pottery by A. Gallay). The pot was rotated and repositioned during beating as witnessed by the organization of the imprints into partially overlapping panels. The slightly different textures of the imprints probably correspond to different types of matting. When more evident, this ripple originating from the manufacturing process could also have a decorative purpose. Regarding the occasional presence of a very slight ripple inside the vessels, further research will be necessary to determine the cause. All the vessels belonging to this group are as follows: slipped and burnished.

Slipped and Burnished

After shaping, the pots were slipped with the help of a soft tool as suggested by the streaked and non-covering color of the surface (Fig. 9c). Then, the internal and external surfaces were completely burnished when the clay was dry as evidenced by the gloss and by fine and narrow streaks with clear edges sometimes visible with the naked eye. Occasionally, however, burnishing bands are visible due to the finishing of a slightly less dry paste, but currently, no specific pattern related to shape/size and stratigraphy was observed for these small variations in burnishing, which are also normal in the case of handmade productions. Often, under the burnishing, raised ribbed striations left by smoothing and/or slipping are visible. Generally, the burnishing is carefully carried out although the slightly rippled surfaces do not help, as they make the surface sinuous and less easily reachable by the instrument. There is the possibility of burnishing made with a soft tool or with a double step, with a rigid and then soft instrument, as suggested in some cases by an overlapping of multidirectional striations (for comparisons see Forte, 2019). Experiments would be necessary to better highlight this finishing procedure (also for the NAQ-SILT), as it is very difficult to identify the tools and methods of burnishing since the possibilities are many (various types of hard and soft tools, such as wool, leather, horn, bone, pebble, wood, and the possibility of adding fatty substances which disappear during firing).

Occasionally the rim is decorated with oblique impressed notches described in the literature as “milled rim” (5 samples from WK14, Fig. 8e, g–h), (the first to use the term was G. Reisner in 1910, in his Archaeological Survey of Nubia).

The firing technique created a perfectly black rim and internal surface, and quite homogenous brown-reddish external surfaces, showing good control over firing. The firing temperature has been estimated at around 800 °C for most vessels as seen in thin section.

The completely black radial sections and the homogenous black coloring of the inside of the vases contrast with the red and black color of the sections and the gray and beige nuances of the inner surface of the NAQ-SILT Black-Topped pottery (Fig. 9e, g). This diversity points to the adoption of different firing methods for the MNR tradition to obtain a relatively similar result, a red vase with a black band.

As mentioned above, this finish characterizes our entire assemblage consisting of two pottery categories: bowls and deep open forms with simple rims and rounded bases, except for two fragments of red burnished bowls fired in a completely oxidizing atmosphere.

Currently, the A-Group pottery can only be dated on the basis of the Naqadan pottery associated with it. Therefore, based on our data we can say that we certainly have it during Naqada I and II, while the disturbed stratigraphy of the necropolis and of the superficial layers of the settlement does not allow us to understand how long it lasts (Table 5).

Hybrid Sherds

Three sherds with unusual features were found on the surface of the necropolis area WK14. They belong to concave bowls with milled rims and red burnished surfaces and have been identified as hybrid forms mixing Nubian (milled rim) and Naqadan (Red-Polished treatment and shape) characteristics (Gatto, 2021, p. 374 and fig. 5). Further, the paste is not the one used for the NAQ-SILT, as it is distinguished by mineral inclusions of variable size, by small, black, and round plant fragments, an incompletely covering slip and burnish, and a slightly bumpy external surface that recalls the manufacturing of the SHALE vessels. Unfortunately, the size of the fragments did not allow us to conduct an in-depth analysis of the radial sections, but based on the data at our disposal, these bowls would appear to be imitations of Red-Polished vessels probably made by potters belonging to the SHALE technical tradition (an element already pinpointed by the presence of red burnished bowls in the SHALE group), and on which aspects of all three traditions present in the region are reflected. None were examined petrographically. Due to the discovery of these fragments on the surface and the lack of parallels that we know of, it is not possible to date these fragments.

Undetermined

Two fragments of N-ware (Petrie, 1921) bowls found in the necropolis were analyzed (Fig. 2). Unfortunately, the small size of the sherds did not allow for a complete analysis. It was possible to note only that the fabrics are made with Nile clay and are characterized by small mineral and vegetable inclusions. The walls seem made with coils. One sherd is characterized by a dark gray burnished internal surface and by a dark gray external one decorated with a vertical and an oblique burnished band with the space in between the bands filled with impressed dots (Gatto, 2021 fig. 6). At present, specific parallels for this sample have not been found, and it is tentatively dated to Naqada IIA-B (Hartmann, 2016, Abb. 151). The second sherd has dark gray surfaces. The external surface is burnished and characterized by a horizontal band of impressed dots below the rim; the external one is decorated with partly preserved triangles filled with impressed dots (possibly similar to Hartmann, 2016, Taf. 61 U-234.648, 124 U-384a.1305, 126 U-393.1324; Petrie, 1921, Pl. XXVI. 2 and 6) and dating to Naqada IIA-B.

Additionally, in the necropolis, two complete vases of a unique type were found in Tomb 59: a small concave bowl with a rounded base and simple rim, and a deep open form, also with rounded base and simple rim (Gatto, 2021 fig. 7). The presence in the tomb of an amulet representing the head of a bearded man (Gatto, 2016) and a NAQ-VEG vessel of Petrie’s type R 91a-c, suggests it dates to the early Naqada II phase (Hartmann, 2016, Abb. 151; Buchez, 2008, planche typologique 17). The fabric is light brown and seems made of clay and mineral inclusions along with shale fragments. The vessels are coiled, based on the undulations perceptible on the smoothed surface. On the external surface of the larger sample, vertical scraping striations are clearly visible on the median part. The completeness of these two pots prevented the analysis of the fabric and the rough-out technique from going beyond these general observations. The shape recalls the bowls belonging to the SHALE group but the manufacturing traces visible on the external surfaces are not quite the same. They have been put in parallel with Reisner’s Type EDII Mud wares based on drawings and descriptions (Gatto, 2021, p. 376; Reisner, 1910, fig. 293), but further research is needed to confirm the nature and the provenance of these pots and the N-ware, which due to their very low occurrence are to be considered anecdotal at Nag el-Qarmila. None were examined petrographically.

Discussion

The first extensive application of petrographic and technological analyses to an Upper Egyptian Predynastic pottery assemblage revealed three different ceramic traditions: the NAQ and the SHALE, both present in the settlement and the necropolis, and the MNR mainly represented in the cemetery.

The NAQ refers to a specific chaîne opératoire that sees the construction of the pots with a spiraled coil for the base and internal apposition of coils for the wall and rim. Then the vessels were shaved and smoothed on leather hard clay to obtain the final shape and receive the finishing. Within this group, we have three sub-groups that, with the same chaîne opératoire, use different fabrics and surface finish. The NAQ-SILT sub-group is distinguished by the use of a very fine and well-worked silty Nile clay whose origin is difficult to establish due to the ubiquity of this material in the Nile Valley, and it is mainly used for making slipped and burnished bowls, beakers, and jars of Red-Polished or Black-Topped type. In Nag el-Qarmila, this sub-group is the oldest, dating to Naqada I to IIA (Table 5). With the end of Naqada IIB, we have the appearance of the NAQ-MIXED group, initially to make small and medium painted jars (D-Ware) and then to produce Wavy-Handled jars, undecorated bowls, and storage jars with smoothed/softened surfaces (Table 5). This sub-group is characterized by a fabric made of very well-wedged and compact mixed (shale) clays probably from the Qena area (see above the problem related to the classification of these fabrics), marked by a light beige or pinkish surface color. At the same time, we have the appearance of the NAQ-VEG, with a coarser fabric made of Nile silt mixed with abundant plant temper. This fabric was used to make a large variety of vessels such as bowls, vats, storage jars, and cooking pots and possibly the first examples of bread molds. All these pots have smoothed surfaces. The NAQ tradition therefore shows a diversification over time of fabrics and finishes that correspond to specific shapes and which consequently have a functional value that increases over time.

The SHALE is a technical tradition based on the use of shale clay possibly originating in the broad Aswan region (although the Western Desert cannot be ruled out) and characterized by abundant laminar shale inclusions which give this fabric its characteristic appearance. The quite well-worked dough was used to produce coils to build the base (spiraled coil) and the walls (internal apposition or alternated coils). Then the surfaces were scraped and the forms were finished by smoothing or burnishing. Two categories of containers are recorded: bowls (smoothed or burnished) and cooking pots (mostly smoothed). This group appears in Naq el-Qarmila at the very beginning of the stratigraphy and seems to be present throughout the occupation of the site.

The MNR group consists of a chaîne opératoire which uses a silty Nile clay rich in mineral inclusions and tiny plant fragments, whose characteristics are different from those of the NAQ-SILT, suggesting different sources of supply within the broad Aswan region. The pots (bowls and deep open forms with rounded bases and direct rims of variable size and shape) were built with a spiraled coil for the base and internal apposition of coils for the walls, and the form was finalized by percussion. Then the pots were slipped and burnished. Also, this group characterizes the earliest stages of the sites of Nag el-Qarmila and appears to persist throughout the occupation of the site.

The three chaînes opératoires have in common the rough-out based on the coiling technique with internal apposition for making base and walls (Fig. 10). We know from ethnography and experimental archaeology that the rough-out techniques and gestures are the most enduring and resistant to change since they do not affect the final shape of the pots (Gelbert, 2003, Roux, 2016). The fact that these three productions share the same roughing-out is therefore particularly significant and could indicate a common origin for these three ceramic traditions to be found in the 5th mill. BCE or earlier. The first author of this paper was able to analyze sherds belonging to 5 open globular shapes with rippled surfaces (one of which almost complete) dated to the 5th mill. BCE, retrieved in the Aswan region by the AKAP team. It emerged that these ceramics were manufactured with the same procedure described for the MNR group (apart from some differences in the fabric and texture of the rippled surface), confirming the presence of this rough-out technique also for the previous millennium. It would now be necessary to expand the study to Badarian and Tasian ceramics to better understand the genesis and relationships of the three 4th mill. BCE chaînes opératoires.

Fig. 10

Summary diagram of the rough-out and preform stages of the NAQ, SHALE, and MNR technical traditions

These three technical traditions were subsequently developed separately, each with a different fabric and a different characteristic preform stage (Fig. 10). The final shape of the NAQ vessels was obtained by pressure, shaving the leather hard clay with a sharp tool. Shaving is a marker specific to NAQ that becomes increasingly evident over time. The SHALE vessels were preformed by scraping the wet clay, and the MNR pots by percussion on leather hard clay.

During Naqada I, the NAQ production was characterized by only one fabric (NILE-SILT) used to produce several pottery types (mainly bowls and beakers), all slipped and burnished with red or Black-Topped final effects. Then, towards the end of Naqada I production diversified, with different types of fabrics and finishes depending on the shape and, therefore, on the function of the vessels (NAQ-SILT, NAQ-MIXED, NAQ-VEG, see Tables 3 and 4). According to the preliminary data of the project TECHNOPREGYPT (which will be the subject of an article dedicated to NAQ-SILT production), we can hypothesize that the NAQ was a specialized production (as already proposed by Friedman, 1994, p. 648; Baba, 2021 and Hendrickx, 2011) that developed a specific manufacturing routine aimed at producing vessels with a precise aesthetic aspect. This was obtained with a rather long and careful procedure including the preparation of a very levigated clay (naturally present in specifically chosen sources or sieved or decanted) and of a well-worked paste. Along with a smooth surface obtained through shaving and smoothing, which was necessary to carry out a perfect burnishing, and careful firing to attain homogenous red surfaces and a specific firing technique for the Black-Topped pottery, this chaîne opératoire was transmitted within a determined group of potters who developed specific knowledge and motor skills. This production possibly flourished in the domestic context (household specialization, Forte, 2019), as suggested by the fact that the most represented shapes (Black-Topped beakers and Red-Polished bowls) show great variability in terms of shape and size (see Fig. 6; Buchez, 2008, fig. 3; and the catalog of Hartmann, 2016), which points towards a low number of pieces produced by each potter, and probably by a relatively high number of potters who made them (for the development of specialized domestic productions see Forte, 2019). Later, the forms became more homogeneous in size, shape, and finish, and the fabrics indicate different possible production centers as evidenced by the presence of mixed (shale) sherds (Ownby and Köhler,, 2021) from the Qena bend in Nag el-Qarmila or, for example, the use of silty clay to create the same types of vessels in the Delta during Naqada III (Bajeot & Buchez, 2021, figs. 9 and 15). For Nag el-Qarmila, it is difficult to establish whether the NAQ-SILT from the earliest phase was produced locally or whether it came from nearby sites. Indeed, the ubiquity of Nile clay, which is widespread throughout Egypt, does not allow for the identification of specific sources of supply and therefore a study of Nile clay pottery circulation channels.

Concerning the SHALE group, variability of raw material sources, general appearance of the fabric, shape, and dimensions within the same pottery class, the adoption of different tools for the execution of the surface treatments, and the not well-controlled firing, suggest a production managed by many potters (domestic production?) aimed at producing cooking pots and bowls. This domestic pottery has been identified by R.F. Friedman (1994) as typical of the region of Hierakonpolis up to Gebelein, and the same observations were later confirmed for the site of Adaima (Buchez, 2004). The technological analysis showed how this ceramic was made according to an original chaîne opératoire that differs from the NAQ, which contextualizes the Shale Ware in the frame of Predynastic ceramics. These results indicate that although the SHALE could be considered a local production since it makes up half of the ceramic assemblage of the oldest phase of WK15B (Naqada IA-beginning of IIA), it was not produced by the same potters who made the NAQ-SILT. The two chaînes opératoires imply in fact different gestures and techniques which are not linked to the obtaining of a specific shape (which can easily be obtained with both procedures) but to the knowledge acquired by the potter since he/she was a child within his/her community. This interpretation of the chaînes opératoires is based on a quantity of studies of experimental psychology and ethnography which have demonstrated how potters (but it also applies to the producers of any other artifact) have the tendency to always use the manufacturing routine and gestures learned within their own community (Dobres, 2000; Gosselain, 2000).

We therefore have two ceramic traditions that have two different channels of transmission within the same community at the regional level. One is aimed at producing simple forms such as cooking pots and bowls with low investment in preform and surface treatment, and the second is a production involving more advanced level skills for the manufacture of time-consuming tableware and storage vessels. This phenomenon is also confirmed by the presence of NAQ-SILT Red Polished bowls, SHALE red burnished bowls, and hybrid red burnished bowls that are made differently although the shape, and possibly function, is more or less the same.

At the present state of our knowledge, however, the archaeological contexts do not allow us to determine who was responsible for the implementation and transmission of these two technical traditions. The possibilities are, in fact, innumerable such as gender, social class, workshop/domestic production, itinerant potters (as suggested for the production of Red-Polished and Black-Topped pottery by Friedman, 1994, p. 648), etc.

It is interesting to note that the amounts of SHALE are greater at Nag el-Qarmila WK15B (59%) and Adaima (40% in the layers dated to Naqada IC-IIB, Buchez, 2008, p. 18) than at Hierakonpolis (25% at HK14, Naqada IC-IIA, Friedman, 1994, p. 630; 18% at HK11 in the layers dated to Naqada IC-IIA, Friedman et al., 2002, p. 59). Moreover, at the first two sites, the SHALE seems to disappear much later than at Hierakonpolis. In the latter site, in fact, after the end of Naqada I, it is strongly restricted to specific activities related to brewing at locality HK24A (Friedman, 1994, p. 669-671), and it is rarely present at other localities (Friedman et al., 2002), while instead at Adaima, despite a constant decrease, it is still recorded until Naqada IIIB-C1 (Buchez, 2004, 2008). This may be because the first two sites are provincial while at Hierakonpolis an elite emerged very early on, stimulating the development of more standardized specialized productions such as the NAQ-VEG pottery (R-Ware in Petrie’s classification), (Friedman, 1994; Di Pietro, 2019) that supplanted the SHALE domestic pottery more quickly than at smaller sites (Buchez, 2004).

Another notable element is the presence of important quantities of Shale ware (19% of the total analyzed samples) in the necropolis WK14 that does not find parallels in contemporary sites such as Adaima and Hierakonpolis, except for the Predynastic cemetery of Sheikh Mohamed, a few km south of Nag el-Qarmila, where Shale Ware is mentioned among the findings (Junker, 1919, but we do not know the quantity). This lack of evidence in funerary contexts of the First Cataract region may be the result of biased publications. The funerary assemblage of WK14 features more bowls than cooking pots compared to the settlement (see Tables 3 and 4). This functional class disparity between the village and cemetery suggests that almost all of the ceramics pertain to the necropolis and were not discarded there as waste. Of these bowls, several are red burnished and could recall the NAQ-SILT Red-Polished ones (Table 3). But it is not clear if the SHALE red burnished bowls, and in general the number of bowls of the necropolis, were meant to meet a greater demand for bowls that could not be covered with the NAQ-SILT Red-Polished ones or if it was a ritual necessity. Further research on burial customs would be necessary to better frame this theme.

The currently known sources of shale clay are located in the Aswan region (for instance, in the Qubbet el-Hawa area within the sandstone layers, Bloxam et al., 2007, p. 143) and towards the desert. The large quantities of Shale pottery found in Nag el-Qarmila (and more generally at Hierakonpolis and Adaima) suggest the possible exploitation of outcrops closer to the Nile but still unknown. However, it should be underlined that Shale Ware has a rather wide distribution in the Egyptian Western Desert and in the Dakhla and Kharga Oases, where the first traces date back to the 6th mill. BCE. It continues throughout the 4th and the 3rd mill. BCE, when it is associated with mobile pastoral communities known as the Sheikh Muftah culture (for an in-depth overview of the Shale Wares see Gatto, 2013). In the future, it will therefore be essential to carry out a broader study on Shale Ware to verify the technical relationship between the samples from the Nile Valley and those from the Western Desert. Further, to understand if this domestic tradition emerged in the Nile Valley in the 4^th mill. BCE independently from the desert one, or whether it originated in the Western Desert and was later brought to the Nile Valley by groups from this region (Friedman, 1994, p. 892-893; Gatto, 2013).

The characterization of the manufacturing procedure for the pottery made with Nile clay and plant temper (NAQ-VEG in this paper and Nile C in the Vienna System, Nordström & Bourriau,, 1993) and its consequent attribution to the NAQ group, enabled a better understanding of the emergence of this pottery category. In fact, it is evident that this ceramic paste with which utilitarian vessels were produced (such as, for example, the so-called beer jars) did not develop from the SHALE domestic tradition but rather from the increasingly specialized NAQ. This phenomenon shows how potters working in the NAQ tradition at some point also began to produce utilitarian pottery, which led, with different timing depending on the site, to the disappearance of the SHALE domestic production. This phenomenon, albeit with some differences, is also observed in the Delta, where the NAQ tradition, at a certain point, replaces the local LEC tradition (Bajeot & Buchez, 2021).

Finally, the MNR group is characterized by a preform made by percussion and by a careful finish that also points towards a specific tradition aimed at producing Black-Mouthed jars. The characteristics of the fabric (coarser than the fabric used to produce comparable pots in the Second Cataract Region, Nordström, 1972) and its composition (in particular the sample with calcrete) point towards an origin in the Aswan region. However, this does not exclude the possible presence of some importation from Nubia, a possibility that should be explored by broadening the petrographic analysis. In the future, it would be necessary to expand the analysis of the 4th mill. BCE Nubian pottery to a larger geographical area to see if this is the only chaîne opératoire adopted for all the productions, or if it is specific to the Black-Mouthed jars, and if there are geographical varieties. Currently, the identification of this ceramic tradition has made it possible to establish its relationship to and differences with the NAQ and SHALE groups. It was also determined through the study of fragments from the 5th mill. BCE and from the Second Intermediate Period (Bajeot, 2022) that it is an extremely enduring chaîne opératoire that does not appear to undergo substantial changes over time, contrary to what happens for the NAQ.

But what does the presence of these three groups in the small site of Nag el-Qarmila indicate? (Fig. 11).

Fig. 11

Map showing the approximate diffusion of the three technical traditions NAQ, SHALE, and MNR during the period Naqada I-II, and the sites cited in the text. Elaborated by J. Bajeot

The possibility that the three different groups of potters were all active at the site can be ruled out due to the small size of the settlement, where few families probably lived. A more considered hypothesis, formulated by M. C. Gatto, suggests that the Shale Ware and the Black-Mouthed Ware are the result of a fusion that, in the Aswan region, constituted the local production characterized by a strong Nubian substratum; and in this framework, the NAQ pottery would have been imported from nearby sites, such as Hierakonpolis, witnessing a strong Naqadan influence (Gatto, 2021). In light of the results of the technological analysis, it was demonstrated that the substantial differences between the SHALE and the MNR technical traditions show that these two types of pottery were not made by the same potters. The presence of important quantities of Shale Ware at Adaima and Hierakonpolis in association with NAQ productions and the absence of Black-Mouthed vessels suggest instead that the SHALE was a production linked in some way to the NAQ rather than to the MNR. This is also confirmed by the fact that in Nag el-Qarmila 90% and 70% of the ceramics found respectively in the settlement and the necropolis are made up of SHALE and NAQ productions. These elements make this second hypothesis unlikely and push us to formulate a third, although still preliminary and to be explored further as research progresses:

The groups NAQ and SHALE are the primary ones and are the expression of two groups of potters who operate within the same society, the Naqadan one. One takes care of the production of cooking pots and bowls, and the second is the bearer of particular know-how aimed at preparing containers for presentation and storage, with precise aesthetic characteristics that require longer manufacturing times. This second group will develop over time, moving from specialized domestic production to workshop production, diversifying fabrics, shapes, and finishes with a strongly functional purpose. And it will slowly go to supplant the SHALE tradition, which will decline sharply. As this pottery is common, production close to the village in the broader area certainly seems likely.

It is worth noting the spatial distribution of the NAQ tradition and of the Shale Ware during the 4th mill. BCE (Fig. 11): Shale Ware is found in the lower portion of Upper Egypt (south of Armant/Luxor), with a chance of being present also in the area south of the First Cataract, and the oases of the Western Desert. The NAQ pottery is, instead, present along the Nile, from Middle Egypt to the Dakka Plain, in Lower Nubia. The latter has a larger extent but is confined to an N-S axis, mostly along the river; the former (if it is really made according to the same chaîne opératoire identified at Nag el-Qarmila) has a smaller extent that crosses the valley and the desert following a W-E axis.

Alongside these predominant groups, we find the MNR group that mainly consists of Black-Mouthed jars, which overlap in use and general appearance with the NAQ productions (Black-Topped and Red-Polished). MNR is a minority in the village (3%) but it is well represented in the necropolis (32% of the complete assemblage). Unfortunately, the poor state of conservation of the necropolis does not allow a true understanding of the role of this ceramic in funerary rituals and its evolution over time. Nevertheless, it is clear that it had a certain importance in the mortuary context, which was not reflected in the settlement. Such a consistent presence of Black-Mouthed vessels (absent at Hierakonpolis and Adaima), the mixed character of the assemblage, and the likely origin of the raw material in the Aswan region are probably the only archaeologically visible traces of an integrated system of contacts between sedentary and mobile communities living in the same area. Nag el-Qarmila could be a community which precisely because of its position on the fringes of areas influenced by prominent sites such as Hierakonpolis has more fluid identity traits, incorporating more varied influences. These close contacts are less visible in residential contexts but seem to find greater expression in funerary contexts. Based on the available data, it is not possible to understand what types of relationships existed between these groups but, likely, they were not merely commercial but part of an integrated system based on different lifestyles (mobile/sedentary). The heterogeneity of these communities at the regional level could explain the composition of the funerary assemblage and the presence in the necropolis of the two unique vases found in Tomb 59 (apparently associated with two NAQ vessels and an amulet representing a bearded head) described in the “undetermined group,” and the rare presence of N-Ware (also in the “undetermined group”) that are anecdotal in the cemetery and without comparisons in the settlement. These communities should not be considered in contrasting terms but rather as parts of a fluid system of interactions, away from the “pressures” that probably developed in the centers where the elites possibly pushed towards greater coherence and identity unity. A similar phenomenon has been observed at Adaima, where alongside the predominant ceramic groups, there are probably different minority technical traditions (still to be investigated). For example, the one represented by particular cooking pots (called Nubian by N. Buchez) used as containers to bury newborns (Buchez, 2008, p. 78) and by small bowls, cups, and mats in association with Naqadan material found in a cluster of tombs and also attributed by Buchez to Nubian groups. This material has been interpreted by Buchez as the results of contacts and circulation of people (possibly based on marital practices), which makes Adaima a more composite community than previously thought (Buchez, 2008, p. 78), and finds a parallel in Nag el-Qarmila, even if with different materials and proportions, and consequently different modalities.

Conclusions

With the present study, for the first time, three different technical traditions were identified at the site of Nag el-Qarmila, and they are suggested to represent reciprocal relationships. This enabled revisions to previous hypotheses and new ones to be established on a more solid scientific basis and brought out new research insights and directions to set future investigations.

The very promising results achieved allowed us to propose a new interpretative model for the “mixed” assemblage recovered at Nag el-Qarmila, and to reach a better understanding of the social dynamics of the First Nile Cataract region during the Predynastic period. The richness of technical traditions in such a small site possibly witnesses the fluidity of the region, far from the large centers where the development of the elite will lead to the birth of the state. This will also result in a change in the area under study which will see the establishment of the political border between Egypt and Nubia.

Beyond the interpretations given to the data, the detailed description of the technical groups and of the processes involved will form the basis for expanding the study to other sites and other periods, functioning as a basis of comparison. The technological approach, indeed, works best on a large geographical scale and long chronological periods, therefore, the more sites and periods investigated with this method, the better the results will be. In combination with petrographic data, a full understanding of the chaîne opératoire can be made along with insights into local production and non-local pottery acquisition. The ability of this methodology to obtain an anthropological reading of the archaeological material, and consequently to address more in-depth inquiries that go beyond the mere dating and cultural affiliation of the ceramic vessels, shows how important its inclusion is in the basic analysis of ceramic assemblages along with typological analysis.