Made from Mud: Functional Categorization and Analyses of Bronze Age Earthen Materials from Western Turkey
Abstract
30
Full text
STUDIA HERCYNIA XXIV/1, 30–65 Made from Mud: Functional Categorization and Analyses of Bronze Age Earthen Materials from Western Turkey Jana Mokrišová– Christopher H. Roosevelt– Christina Luke – Caitilin R. O’Grady ABSTRACT This contribution presents the results of apilot study of earthen materials excavated at the Middle to Late Bronze Age site of Kaymakçı, located in western Anatolia. It argues that systematic collection and analysis of fragmentary and difficult ‑to ‑identify earthen materials is challenging, yet crucial. These materials inform on activities of which traces are preserved in the archaeological record but which have been largely under‑ ‑researched. Flourishing studies on earthen findings foreground architectural materials, such as mudbrick, and well ‑preserved features and objects. However, earthen objects and architectural features were utilized more widely than in building architecture and only asmall portion of excavated sites has good preservation. We, therefore, present the different categories of earthen materials discovered at Kaymakçı, specifically ar‑ chitecture, installations, and portable items. Our work demonstrates that by incorporating new knowledge of archaeological remains at the site and re ‑studying the earthen assemblage it is possible to gain abetter understanding of the morphological, functional, and social aspects of this dataset. KEYWORDS Late Bronze Age; western Anatolia; Kaymakçı; architecture; mudbrick; earthen materials. INTRODUCTION Studies of earthen architecture and earthen features of the built environment in the Bronze Age and later periods in the eastern Mediterranean have multiplied since the 1980s and ac‑ celerated in recent years (e.g., Jerome 1991; Shell 1997; Nodarou– Frederick– Hein 2008; Sauvage 2008; Mielke 2009; Homsher 2012; Costi de Castrillo– Philokyprou– Ioannou 2017; Devolder– Lorenzon 2019). Afocus on extant features, such as mudbricks and wattle and daub construction from the Middle East, Egypt, Anatolia, and the Aegean allowed for development and improvement of existing analytical methods and insights into the technical, environmental, and cultural milieu of ancient communities (e.g., Guest ‑Papamanoli 1978; French 1984; Rosen 1986; Emery– Morgenstein 2007; Nodarou– Frederick– Hein 2008; Love 2012; 2013; Rosenstock 2009; Lorenzon– Iacovou 2019). Studies of the processes of formation and degradation of Mediterranean mudbrick architecture have also increased in the past few decades (e.g., Shaffer 1993; Stevanovic 1997; Friesemetal. 2011; Friesem– Kar‑ kanas– Tsartisidou 2014; Forgetetal. 2015; Peinettietal. 2017; Cammas 2018). Moreover, new work on fragmentary building materials and features have now supplemented the dataset (Avrami– Guillaud– Hardy eds. 2008; Claasz Coockson 2010; Jazwa forthcoming). To‑ gether, this research has shed unprecedented light on earthen architecture and placed it at the center of holistic research combining archaeological, architectural, and scientific approaches. Despite such gains, at least two significant gaps in research remain. First, published contributions tend to discuss well ‑preserved examples and almost complete architectural features. Only asmall portion of excavated sites can boast conditions ideal for the preserva‑
31JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY tion of features and objects made from earth, and such asituation is uncommon to most ar‑ chaeological sites. Second, publications focus predominantly on mudbrick from architectural contexts. This trend is understandable because excavations of mudbrick architecture have highlighted its important role in vernacular traditions of the eastern Mediterranean and the Levant since the Neolithic period (e.g., Aurenche 1981). Moreover, this is in part due to the long ‑acknowledged difficulty of identifying earthen materials in their secondary and tertiary contexts (Goldberg– Macphail 2006, 227, 279, 283; Cammas 2018). Yet, features and objects made from amixture of sand, clay, silt, and organic and inorganic aggregates were utilized more widely than in building architecture. Hearths, ovens, linings, containers, trays, and other items were also made in similar materials. Such features and objects propose challenges for archaeologists, because their remains are often preserved in only limited traces, and their identification is ever more difficult owing to alack of published comparative examples, in contrast to the growing body of literature on mudbrick architecture. This article tackles this combined problem and addresses the most common preservation scenarios– that of very fragmented archaeological remains with only afew well ‑preserved diagnostics– by presenting arange of earthen features and objects. These include, but are not limited to, mudbrick from the Middle to Late Bronze Age site of Kaymakçı in western Turkey. We argue that systematic collection and analysis of fragmentary and difficult ‑to ‑identify earthen materials is challenging yet very rewarding because it informs on activities of which traces are preserved in the archaeological record but which have been largely understudied, if not wholly ignored. In short, we stress the importance of thorough study of all types of excavated materials collected in asystematic manner. Excavations at Kaymakçı started in 2014. Since then, our team has documented avaried range of earthen features. In excavations between 2014 and 2018, 364 samples of earthen materials were recovered, demonstrating that the mixture of sand, silt, clay, and organic and inorganic temper created malleable utilitarian materials abundant across the site. Such mixtures were used not only in the construction of buildings, but also in the manufacture of arange of features of both utilitarian and aesthetic value. Indeed, while extant mudbrick architecture at Kaymakçı is limited, the abundance of earthen features such as fragmentary hearths, ovens, and trays attest to the application of the material in broader contexts. In what follows, we present an overview of the different categories of structural and utilitarian earthen features and objects found at the site. KAYMAKÇI: THE ARCHAEOLOGICAL CONTEXT The archaeological site of Kaymakçı is located in the middle Gediz River valley in western Anatolia, modern Turkey, on aridge above the western edge of Lake Marmara (the ancient Gygaean Lake or Lake Coloe) (Fig. 1). The site was discovered during aregional survey in 2001 and studied intensively in subsequent surveys between 2006 and 2013; excavations began in 2014 (Rooseveltetal. 2018; see also Roosevelt– Luke 2008; 2009; 2010; 2011; 2012; 2013; Luke– Roosevelt 2009; Rooseveltetal. 2014; Roosevelt– Luke– Sekedat 2016).1 Earthen architectural remains from ancient and modern times have been investigated within the scope 1 The Kaymakçı Archaeological Project (KAP) is apart of Gygaia Projects, aresearch collaboration directed by Christina Luke and Christopher H. Roosevelt that promotes archaeological research, sustainable management of cultural and natural heritage resources, and the engagement of inter‑ ested communities in the Marmara Lake basin.
32 STUDIA HERCYNIA XXIV/1 of the project over the past ten years through pedestrian survey, excavations, and ethnographic work (Luke– Cobb 2013; Luke– Roosevelt– Scott 2017; O’Gradyetal. 2018). This contribu‑ tion focuses exclusively on archaeological evidence for the use of earthen materials through excavated remains from the ancient citadel of Kaymakçı, and thus complements previously and concurrently conducted ethnographic, heritage, and conservation studies. Kaymakçı is positioned on the lower promontory of the Gür Dağ ridge, in an area rich in natural resources such as wood, earth, and water. Fortifications enclose an area of 8.6 ha, Fig. 1: Map of the middle Gediz River valley in western Turkey, showing the location of Kaymakçı (©Gygaia Projects).
33JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY making it one of the largest 2nd millennium BCE citadels in western Anatolia. The settlement is an ideal candidate for one of the major centers of the so ‑called Arzawa Lands, specifically the Seha River Land (Roosevelt– Luke 2017). Material culture at Kaymakçı is mostly local inland western Anatolian, with only few imports identified to date (Rooseveltetal. 2018, 664–665). The site was occupied as early as the Middle Bronze Age (2000–1700/1650 BCE), but the main Fig. 2: Plan of excavation areas at Kaymakçı (©Gygaia Projects).
34 STUDIA HERCYNIA XXIV/1 phase of activity belongs to the Late Bronze Age (1700/1650–1200 BCE), when the fortifications and all architectural remains so far discovered were built. The Late Bronze Age occupation can be further divided into two phases: the LB 1 phase (17th–15th century BCE) and the LB 2 phase (14th–13th century BCE). The site was abandoned in the final Late Bronze Age or perhaps as late as the beginning of the Early Iron Age (1200–1000 BCE). The fortified area of the citadel is topo‑ graphically and architecturally divided into several sectors, in which excavations have explored northern sections of the fortification system itself, acentral inner citadel and its surrounding slopes, and abroad southern terrace (Fig. 2), with varied activities taking place in each. The inner citadel on top of anatural hill represents the heart of the site and is secured within three arcs of fortification walls. It has been excavated in areas 93.545, 97.541, and 98.531. This sector was dedicated to storage, at least during the end of LB 1 and throughout the LB 2 period, as well as other regular residential and small ‑scale productive activities, such as food preparation, textile production, and other crafting activities. An extensive southern terrace is divided by awide street into two segments to the northeast and southwest. It has been explored through excavations in area 99.526 to the west and 108.522/109.523 near its center. Residential, household, and workshop activities took place here, including food processing, small ‑scale crafting, and storage. Fortifications are an especially dominant feature of the site’stopography, and their contours can be readily traced with the aid of satellite imagery. The area west of the inner citadel was most heavily and extensively fortified, and key sections of the fortifications were excavated in areas 81.551 and 95.555. EARTHEN MATERIALS AT KAYMAKÇI The primary objective of the pilot field study was to identify and document the wide range of structural and utilitarian earthen features and objects at Kaymakçı, to identify their structure and composition as well as function and distribution across time and space. Therefore, despite the near total absence of built features at the site above foundation levels, the excavation team collected fragmentary earthen materials from the onset of the fieldwork, as it was hoped that new methods and wider scholarship on earthen materials would allow for identification of new datasets that inform on architecture, the use of space, household and production activities, as well as technological choices. As we will demonstrate later, this effort was extremely pro‑ ductive and resulted in identification of abroad range of materials and activities. This section describes the building blocks of our field activities– the sample collection procedures– and introduces the categories of extant structural and utilitarian earthen features and objects, their basic morphologies, fabrics, and functions. Mud‑ and clay ‑based features and objects are typically made in relatively flexible man‑ ners; mixing recipes vary and are usually location specific (evidenced by both ethnographic and archaeological studies: Morgenstein– Redmount 1998; van Beek– van Beek 2008; Nodarou– Frederick– Hein 2008; Claasz Coockson 2010; Homsher 2012; O’Gradyetal. 2018; Devolder– Lorenzon 2019). They contain sediment, aggregates, and binding agents, and the remaining portion is usually complemented by vegetable temper and stone aggre‑ gate. Similarly, structural and utilitarian earthen features and objects are simply amodelled mixture of sand, clay, silt, and organic (straw, chaff) and inorganic (stone, calcium carbonate, shells, pottery) temper or aggregate. The differential representation of these components influences physical properties, such as tensile strength, shrinkage, and plasticity (Rosen 1986; Houben– Guillaud 1994, 23–27; Kemp 2000, 79–83; Guillaud 2008, 21–26). Specialized knowledge is needed in order to produce amaterial that performs well given aspecific envi‑
35JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY ronment or context of use. Most items, such as mudbricks and hearth fragments, are made of locally available materials to fit specific conditions (Homsher 2012, 2–5; Cammas 2018, 161). This high degree of locality in terms of resource procurement and manufacture also means that earthen features and objects can vary significantly between sites. Studying these features yields not only technological, but also valuable cultural information (Love 2013). In general, archaeological remains made of earthen materials are challenging to detect and excavate if they are not baked by intentional or accidental firing, as they tend to disin‑ tegrate quickly without constant maintenance. According to excavations to date, Kaymakçı appears to have never suffered wide ‑spread burning and the end of its occupation is marked by athorough abandonment. Additionally, Kaymakçı has arelatively shallow stratigraphy, and buried ancient remains have experienced post ‑depositional degradation resulting from environmental conditions as well as modern agropastoral activities, such as sheep and goat herding and plowing. Acombination of these factors results in generally poor preservation of earthen architecture and other features at the site. In this and similar cases, it is easier to identify the remains of features commonly identified in the wider region, such as mudbrick superstructures, than features that are fairly understudied, such as ad hoc working platforms made of mud, poorly preserved hearths, and the like. Despite these challenges, collection protocols established at the onset of the excavations at Kaymakçı prescribed collecting fragments of incompletely preserved as well as well ‑preserved earthen materials, thus creating an extensive and consistent dataset. Earthen features and objects were used widely throughout the site, and the collection strategy aimed to capture their variability. Precise stratigraphic and contextual information was documented for each collected sample in the field, and full sample descriptions including basic characteristics were produced in the excavation laboratory. Increasingly, archaeologists have successfully applied quantitative methods to study mineral and chemical compositions using portable X ‑Ray fluorescence (pXRF) (Emery– Mor‑ genstein 2007; Love 2012; Costi de Castrillo– Philokyprou– Ioannou 2017; Lorenzon– Iacovou 2019), X ‑Ray diffraction (XRD) (Tung 2005; Devolder– Lorenzon 2019), neutron activation analysis (NAA) (Nodarou– Frederick– Hein 2008), granulometry (Goldberg 1979; Jerome 1991; Homsher 2012; Love 2017), and micromorphology and petrography (Cam‑ mas 2018; Lorenzon– Iacovou 2019; Devolder– Lorenzon 2019). 2 While scientific practices are becoming increasingly streamlined, implemented analyses vary based on research goals. These types of analysis have demonstrated their usefulness, especially when there is good preservation and comparative datasets; most of them are, however, costly and often require transfers of samples to laboratories that must follow country ‑specific guidelines, producing increasing logistical and administrative challenges. Yet, simpler field ‑based methods– such as macroscopic qualitative observations of asample’smorphology, fabric, and aggregate com‑ bined with particle ‑size analysis– are able to provide useful insights into the differential use of resources and manufacturing techniques, because the vast majority of earthen material manufacture is local, resources are pooled from small catchment areas, or the materials are reused and recycled. This suggests that limited variation in quantitative signatures can be expected. These field ‑based methods, moreover, can be much more viable for researchers in cases where greater administrative challenges and/or smaller research budgets prevail. This rationale is also the primary reason why these techniques were piloted at Kaymakçı. 2 Preliminary unpublished studies were conducted as part of the Central Lydia Archaeological Survey including micromorphology of asmall number of hardened mudbrick fragments found during pedestrian survey in the Marmara Lake basin. See Wolff 2008 and Curtis 2009.
36 STUDIA HERCYNIA XXIV/1 SAMPLE COLLECTION AND ANALYSIS Samples for this study derive from over four seasons of excavations (2014–2016, 2018). Sam‑ ples were collected from all excavated areas and comprised both substantially preserved as well as fragmentary items, including those showing evidence of manufacturing or functional characteristics such as fingerprints and reed impressions, while at the same time ensuring collection of representative ranges of temper, shapes, and color. This strategy was chosen in order to facilitate future analyses and maximize the potential of the dataset for comparative purposes. Most earthen features and artifacts came from secondary or tertiary deposits and were thus very fragmented; only afew features were retrieved from primary contexts. Such intact features were carefully excavated and documented in place, with smaller samples extracted from clean fresh surfaces to avoid contamination. At the end of the fieldwork day, they were taken to the excavation laboratory for further analysis. The samples were brushed off and left to dry in the shade in open plastic sample bags to allow gradual evaporation of moisture. Macroscopic analysis recorded the color, dimension, fabric description, shape, and state of preservation of individual samples. The color was determined by astandard Munsell color chart, using Pantone Capsure units to ensure standardization across the dataset. Any evidence of manufacturing technique, if present, was carefully noted and documented (e.g., mold and finger impressions, intentional baking). All samples were photographed and weighed. Visual inspection of earthen materials is highly informative (especially concerning aggre‑ gate), but particle ‑size analysis provides aquantitative differentiation of composition. Granu‑ lometric analysis was conducted on selected non ‑hardened baked samples from interesting ar‑ chaeological contexts or with visually striking characteristics. While particle ‑size analysis using the hydrometer method has been the established practice, it requires export of archaeological samples to an institutional laboratory. Afield ‑based procedure, however, was implemented in Kaymakçı’sexcavation laboratory following the field protocol of S. Love (2017), which allows for aquantitative identification of properties of earthen materials with good precision directly in the field laboratory. This protocol, summarized below, was chosen as the most efficient avenue for analysis of the composition of structural and utilitarian earthen features and objects. Granulometric analysis through wet sieving with the aid of nested sieves (using aScience‑ ware® mini ‑sieve micro sieve set) was conducted in the excavation laboratory. An approximately 25 g sample was lightly crushed with aceramic mortar. Although P.Goldberg and R. Macphail suggest treatment of samples with asolution of hydrogen peroxide before wet sieving (Gold‑ berg– Macphail 2006, 336–339), the authors followed Love (2017, 356) in considering that such pre ‑treatment was unnecessary because most organic matter in the earthen material had already decayed. Moreover, dry samples were not pre ‑sieved through alarge mesh so that the original matrix of the samples was preserved, even at the cost of more heterogeneous sand fractions; the heterogeneity of coarseness was in itself reflective of sourcing and manufacturing choices. The crushed samples of 25.0 g were dissolved in 100 ml of deflocculant made of 20.0 g of laundry soap dissolved in 500 ml of deionized water. The samples were left in the solution for at least seven hours in order to loosen individual particles. The solution was then poured over aseries of nested sieves with mesh sizes of 500 μm, 250 μm, 125 μm, and 63 μm to distinguish between coarse, medium, fine, and very fine sand. Particles smaller than 63 μm comprised amixture of silt and clay, which were not processed any further by the hydrometer method, because this would have required exporting the samples to an institutional laboratory. Sieved samples were left to dry slowly in ashaded area over the course of afew days. Each fraction was weighed to a0.01 g precision, and visual observations and photographs then documented dry samples using aLeica stereoscopic binocular microscope under 12.5 magnification.
37JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY RESULTS In total, 364 fragments were recovered and saved for study, of which 40 samples were processed with granulometric analysis. Analyzed samples and fragments varied remarkably in terms of shape, context, and morphology, and although they were initially all classified as ‘mudbrick’, aclose examination showed they in fact represented an array of different morphological and functional categories. These fragments comprised well ‑known as well as previously uniden‑ tified classes of unfired to low ‑fired structural and utilitarian earthen features and objects. As they were made of earthen mixtures, they shared some common visual characteristics. Samples were usually reddish in color, ranging from red (2.5YR 5/6) and reddish brown (2.5YR 5/4), to weak red (10R 4/2) and brown (7.5YR 6/4). While some samples were preserved in their original forms and were either unfired or low ‑fired, some bore traces of both oxidation and reduction, aremnant of uneven firing, perhaps through accidental as opposed to intention‑ al contact with higher temperatures. In such cases, gray reduced fabrics ranged from dark reddish gray (5YR 4/2) to gray (2.5YR 5/1). Macroscopic analysis combined with particle ‑size analysis revealed that sample fabrics were usually coarse, with varying additions of organic and inorganic tempering agents that improved the physical properties of the raw earthen materials. Chaff and straw were common additives in architectural materials. Other organic materials, such as bone and shell, were used only very occasionally. Stone inclusions were very common, too. Micas (biotite, muscovite, and mica ‑schist), limestone, and calcium carbonate were recognized in virtually all samples, while quartz was less common. CLASSIFICATION OF FUNCTIONAL CATEGORIES The collected samples were identified as belonging to one of the three categories: architecture, installation, and portable item. These were further subdivided into more specific categories common across these broader functional categories. Commonly, the original function of recovered samples could not be identified with certainty due to poor state of preservation and the disturbed nature of some of the excavated deposits. These samples were assigned an ‘unclear’ identifier, which could be allocated to any of the major functional categories. Asignificant portion of the original dataset– 116 out of 364 samples, thus 31.9% of the entire assemblage– was identified only as ‘earthen’ in material, but ‘unclear’ in functional category. The following section introduces all the major categories of earthen features and objects found at Kaymakçı and provides an overview of their characteristics and function. ARCHITECTURE The buildings at Kaymakçı consisted of stone socles with mudbrick superstructures, which were preserved only rarely and in rather fragmentary states. However, detailed documen‑ tation of standing features and collapse deposits around walls allowed an understanding of much of the architectural composition at the site (Tab.1). 3 Architectural fragments from more deeply buried levels of the site tend to be sun ‑dried and fairly well preserved– aclear distinction from those buried more shallowly. Mudbricks and other earthen objects found in upper levels closer to topsoil and in pits tend to be accreted and exposed to fire. This distinc‑ 3 Munsell colors provided in the tables reflect interior measurements, whenever possible, as those colors were more indictive than exterior readings, which often combined the color of the object/ feature and accretion.
38 STUDIA HERCYNIA XXIV/1 Excavation Area Context Sample Colour Architecture – mudbrick 81.551 9 4 7.5YR 5/2 81.551 9 5 10YR 4/2 81.551 9 7 5YR 5/2 81.551 9 18 7.5YR 5/2 81.551 16 8 5YR 5/4 81.551 38 19 10YR 5/1 93.545 7 4 5YR 5/4 93.545 7 134 5YR 5/4 93.545 19 1 7.5YR 6/4 93.545 27 1 7.5YR 6/4 93.545 54 5 10YR 7/4 93.545 114 5 5YR 5/6 93.545 135 6 7.5YR 6/4 93.545 149 6 5YR 6/6 93.545 158 1 7.5YR 6/4 93.545 191 6 2.5YR 5/6 93.545 196 392 10YR 6/4 93.545 197 1 7.5YR 5/4 93.545 199 9 5YR 6/6 93.545 204 194 7.5YR 5/4 93.545 211 5 7.5YR 7/4 93.545 216 6 10YR 6/4 93.545 220 6 7.5YR 6/4 93.545 223 5 10YR 6/2 93.545 234 1 10YR 6/4 93.545 280 7 10YR 7/4 93.545 289 1 10YR 5/1 93.545 291 7 10YR 4/2 93.545 294 1 2.5YR 6/2 93.545 330 7 5YR 6/4 93.545 331 1 10YR 6/2 93.545 340 6 5YR 6/4 95.555 46 7 10YR 6/4 95.555 60 10 7.5YR 6/4 95.555 60 338 5YR 5/4 95.555 63 11 7.5YR 4/4 95.555 80 1 5YR 6/4 95.555 103 1 7.5YR 5/2 95.555 130 11 10YR 4/1 97.541 15 1 7.5YR 6/2 97.541 148 1 7.5YR 6/4 97.541 289 56 7.5YR 5/2 Excavation Area Context Sample Colour Architecture – mudbrick 97.541 305 1 5YR 6/4 97.541 374 1 10YR 4/2 97.541 384 1 10YR 4/1 97.541 16 6 7.5YR 6/2 97.541 97 1 10YR 4/2 97.541 104 1 5YR 5/4 97.541 128 1 5YR 6/4 97.541 146 1 n/a 97.541 237 66 5YR 5/4 97.541 265 1 7.5YR 5/4 97.541 330 9 5YR 5/4 98.531 8 8 5YR 6/4 98.531 8 95 5YR 6/4 98.531 8 106 5YR 6/4 98.531 34 3 2.5YR 5/2 99.526 4 5 7.5YR 7/4 99.526 4 26 10YR 6/2 99.526 7 8 7.5YR 6/4 99.526 11 10 10YR 5/1 99.526 71 6 5YR 6/4 99.526 113 1 10YR 6/2 99.526 122 89 n/a 99.526 149 6 n/a 99.526 194 2 10YR 6/4 99.526 238 1 10YR 5/1 99.526 300 1 10YR 6/2 99.526 338 1 5YR 6/4 99.526 350 4 5YR 5/4 99.526 445 7 7.5YR 6/2 99.526 515 5 7.5YR 5/2 99.526 573 10 5YR 5/4 99.526 630 1 7.5YR 5/4 99.526 660 5 10YR 6/2 99.526 671 2 7.5YR 6/4 108.522 7 2 7.5YR 6/4 108.522 14 4 5YR 5/4 108.522 30 6 5YR 6/4 108.522 52 5 7.5YR 5/2 109.523 115 1 7.5YR 5/2 109.523 162 1 10YR 5/4 109.523 195 1 7.5YR 5/2 109.523 281 1 10YR 5/4
45JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY stratigraphic layers corresponding to the LB 1 phase. This phase of activity was not associated with more formal architectural features, which appeared at the transition of Kaymakçı’sLB1 to LB 2 phase (the late 15th century BCE). In LB 2, the investigated strata suggest that the area was dedicated to household production and food preparation activities. Despite this function, collected fragments of earthen features were often sun dried or soft ‑baked at low tempera‑ tures, and were thus suitable for granulometric analysis, the results of which are discussed further below. In addition to area 99.526 on the southern terrace, earthen installations have been found also in the inner citadel in area 93.545. Installation– surface and substructure This category contains arange of features that vary in visual characteristics but are united by function. They often take acircular form and are characterized by aflat hardened surface supported by arelatively fragile mudbrick substructure. They are rich in coarse stone ag‑ gregate and were exposed to low heat. This exposure is most likely aresult of their function as hearths, ovens, and working platforms. Although the surfaces of such features and their underlying substructures can be easily distinguished visually and were classified accordingly into distinct subcategories, they were often discovered in very fragile and crumbly condition, meaning that they could not easily be excavated separately. They are presented together here for this reason, reflecting their conditions upon excavation. In total, 26 surface and four sub‑ structure fragments of such features have been studied. Fig. 8: Well -preserved oven (93.545.171) built near awall of the inner citadel during excavation. Sectioning reveals its pebble and cobble base layer (©Gygaia Projects).
46 STUDIA HERCYNIA XXIV/1 Fragments of surfaces come from the topmost layer of abuilt earthen feature. Surface fragments comprise ahardened cap ranging in color from grayish brown (10YR 5/2) and brown (7.5YR 6/4) to light reddish brown (5YR 6/4), and occasionally even weak red (10R 5/4). Fragments of substructure, the layer below the surface of abuilt earthen feature, are strikingly different in terms of color and aggregate. They can be reddish brown (2.5YR 4/4 or 5YR 4/4) or dark reddish brown (5YR 4/2), and all recovered samples are rich in stone aggregate, especially micas and occasionally also limestone. Some samples, however, visually resemble the hard cap surface with their yellowish brown (10YR 5/4) color. Both surfaces and substructures lack chaff and straw inclusions, in contrast to mudbrick, and are compact, dense, and heavy. One better preserved oven was discovered in alate LB 2 level of the inner citadel in area 93.545 (93.545.171; Rooseveltetal. 2018, 183–184). The surface of the circular installation was preserved in situ (Fig. 8). Upon excavation, it was possible to determine that the structure was likely originally covered, as two vertically set upright slabs in the enclosing stone circle seemed to define the east ‑facing mouth of the oven. The surface of the structure was well preserved. This uppermost layer consisted of two parts: athin and relatively hard earthen cap (0.5–0.9 cm) and, immediately beneath it, asoft‑ er layer 2.0–5.0 cm thick. Although the cap is baked hard, the soft subsurface immediately beneath it is prone to disintegration, hence the two could not easily be separated during ex‑ cavation, suggesting that both the surface and the layer immediately beneath were originally constructed together. Both are rich in sand and micaceous aggregate. The surface hardened as aresult of activities that took place on top of it, with fairly low temperatures perhaps reach‑ ing around 200o C and penetrating only shallowly. Beneath these upper layers– the surface and immediate subsurface together– the continuing substructure of the oven consisted of progressively coarser layers of earthen material approximately 6.0–8.5 cm thick. These layers were pressed into adense layer of schist and limestone pebbles and cobbles at the base of the feature. Fragmented materials of similar composition were found in significant numbers in the fill of acircular feature (93.545.63) in the same area, suggesting that alarge portion of the surface and substructure of asimilar installation was discarded there. Feature 99.526.79 in area 99.526 on the southern terrace presents an example of aworking platform, or perhaps adifferent type of hearth or oven, supported by alayer of cobbles and enclosed by amudbrick rim (Fig. 9; Rooseveltetal. 2018, 191). It consists of aseries of al‑ ternating hard packed, clay ‑rich surfaces with earthen substructure: the earlier surface 276 (light reddish brown, 5YR 5/4) was supported by two fills, 286 and 295 (reddish brown, 5YR 5/4), superimposed by alater surface 264 (light brownish grey, 10 YR 6/2), and fill 267 (reddish brown, 2.5YR 4/4). Its surfaces are relatively heterogeneous as they were built at different times. The topmost hardened surface had avery hard, ca. 0.8 cm thick lens, with aslightly softer and wider ca. 1.4 cm thick lens immediately below it, similar to the uppermost layers of the installation in 93.545. The feature is surrounded by three vertically positioned mudbricks creating arim that delineates the feature on its south side. These concentrically arranged rim mudbricks differ in color, and their temper consists mostly of stone aggregate– micas and limestone– as well as calcium carbonate, with little visible chaff/straw, thus distinguishing them from structural mudbricks used for architectural purposes. These types of rim mudbricks are described fully in the following section.
47JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY Installation– mudbrick Architecture comprises only one possible use of mudbrick at Kaymakçı. The edges of earthen surfaces, platforms, and hearths/ovens were sometimes defined by mudbrick rims. Eighteen mudbricks from such features have been identified, all of which were placed on end to create an elevated rim defining the edge of an installation. Mudbricks used in such contexts had recognizably rectangular forms and fabrics lacking straw or chaff that differ from those of architectural mudbricks. They served as ameans of delineation rather than as components of wall superstructures. Differences from architectural mudbrick in form, composition, and function thus suggest special purpose manufacture. Moreover, it seems that bricks of differ‑ ent dimensions were used to frame installations, but their complete dimensions cannot be identified without difficulty, because most were found very abraded around their edges or only partially preserved. Mudbricks used in installations were found almost exclusively in area 99.526 on the south‑ ern terrace. Features 99.526.385 and 99.526.386, identified as hearths, ovens, or some other type of cooking platforms, showcase the use of these types of mudbricks (99.526.393.2, 426.1, 428.1) (Fig. 10). Both features were of oval shape and delineated on all sides by mudbricks that, together, created an elevated rim surrounding the functional surface. Each rim was constructed of three individual rectangular mudbricks, each stood on their short ends. Each mudbrick bore different visual characteristics and varied compositionally in terms of sand, clay, and silt ratios. Here, individual bricks ranged from brown (7.5YR 5/2) to light brownish gray (10YR 6/2). Fig. 9: Large platform/oven (99.526.79) during excavation. Sectioning reveals its cobble base layer as well as amudbrick rim that delineates its southwestern edge (bottom right corner of image) (©Gygaia Projects).
48 STUDIA HERCYNIA XXIV/1 Fig. 10: Sherd hearth (99.526.385) and hearth (99.526.386) during excavation. Amudbrick rim lines the northern edge of the feature (©Gygaia Projects). Likewise, the relatively large and incompletely preserved platform/oven discussed above concerning its surface and substructure characteristics (99.526.79), included three succes‑ sively built rims of mudbrick (99.526.271, 299, and 305) (Fig. 9). Each of the rims had adistinct color and coarseness, as revealed by particle size analysis (discussed in the next section). The outermost ring (271) had alight gray exterior (10YR 7/2) and brown interior (7.5YR 5/4); the middle ring (299) apinkish gray exterior (7.5YR 7/2) and light reddish brown interior (5YR 6/4); and the inner ring included two varieties of mudbrick excavated separately: 305.1 had alight reddish brown exterior (5YR 5/4) and light brown interior (7.5YR 6/4), and 305.2 had alight reddish brown exterior (5YR 6/4) and light brownish gray interior (10YR 6/2). Together, they provide ahighly visible and functional separation between the installation and the room in which it was used. Installation– tray This category comprises tray ‑shaped features that constituted the permanent equipment of certain buildings. Most of the identifiable examples exhibit aflat base with raised angular edges with rounded rims, presumably designed to serve as large but shallow receptacles or mud ‑made containers. Their size and weight distinguish them from portable trays (see be‑ low); they are heavy, quite thick and large, and have uneven bottom surfaces that bear the impressions of the surfaces on which they sat. Only five fragments have been identified, four of which were found broken and discarded in fills in area 93.545 in the inner citadel. The size and shape of non ‑portable trays cannot be determined with certainty owing to their fragmentary state of preservation, but it is possible that they had rectangular or semi‑
49JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY ‑rounded plans. Their raised exterior edge walls are generally low and are rounded on top. Their colors range from light brown (7.5 YR 6/4) to light reddish brown (5YR 6/4). Micas and limestone are common stone inclusions; chaff and straw are present in small quantity. Even color and hardness throughout suggest controlled firing at low temperatures. The bottom bears impressions of the surface on which the installation would have been set, including impressions of straw and chaff, suggesting they were fired in situ (cf. ‘utilitarian ceramic trays’ from Mitrou, dating to the Middle Helladic II–Protogeometric period; Jazwa forthcoming). An illustrative example of atray with araised rim is fragment 99.526.350.5 (Figs. 11). The tray is only 3.1 cm thick and its straight rim with rounded top rises to aheight of 9.65 cm. Its interior and exterior surfaces are smoothed. Its bottom surface is rough, slightly irregular, and seems to have taken the shape of the surface it was placed on when wet. Although the fragment is accreted on its interior and exterior surfaces (including the bottom), it is clear that all surfaces were smoothed and evenly treated with aslip, unlike other recovered examples. Fig. 11: Example of an installation– tray with straight raised edges and slip (99.526.350.5). Photograph and section (N. Gail and P.Demján; ©Gygaia Projects). Fig. 12: Example of aroughly smoothed installation– tray (97.541.68.17) (N. Gail; ©Gygaia Projects).
50 STUDIA HERCYNIA XXIV/1 Most other trays have similarly flat bottoms and straight raised edges but are only rough‑ ly smoothed and bear no other treatment. An example of this more typical type is fragment 97.541.68.17 (Fig. 12), the maximum preserved bottom thickness of which is 3.2 cm, while the raised edge rises to aheight of 6.4 cm. The fabric of this tray includes micaceous and quartz particles, calcium carbonate, and only asmall amount of chaff and straw. Installation– impressed This category is rather narrowly defined and includes mud‑ or clay ‑rich linings of negative fea‑ tures in area 99.526. So far, only four samples have been recovered. One sample was excavated in situ (Fig. 13), while the other three were collected from fills of pits or other negative features. Fig. 13: Clay -rich lining of apit (99.526.604), classified as belonging to the installation– impressed category during excavation. The lining is only partially preserved and visible near the right edge of the photo (©Gygaia Projects). Each is composed of athick layer of evenly fired and finely processed clay up to 6.8 cm thick and is usually dark red in color (5YR 4/4) with ahard, light brown lens along its uneven exterior surface (7.5YR 6/4). The irregularity of the surface likely results from the mud‑ or clay ‑rich mixture being pressed into place by hand, or perhaps by asmall flat implement, prior to firing. The composition of these linings is rich in very well ‑sorted clay and bears very little stone aggre‑ gate and almost no calcium carbonate particles. Small pieces of charcoal, however, can be seen throughout the matrix. It is not immediately clear whether they were fired intentionally, but they are very hard, suggesting they were exposed to prolonged heat in the course of their use.
51JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY Installation– unclear Alarge proportion of this category, including 20 samples in total, could not be identified beyond their association with installations. While highly recognizable as belonging to such features based on their shape, texture, and color, their size and state of preservation or their position within individual features leaves their function or original form unclear. PORTABLE ITEM Portable items were found in very fragmentary state, but they are identifiable by their shape, worked top surfaces, coarse fabrics, and well ‑sorted dense matrices, often resembling very coarse pithoi (Tab.3). Yet, they are clearly flat and are not processed or fired with the same care allotted to coarse ceramics, such as pithoi. Tab.3: List of all fragments of earthen portable items identified at Kaymakçı. In some cases, color could not be identified due to severe accretion on the surface. Portable item– tray Flat and relatively thin fragments of hardened portable earthen materials were found primar‑ ily in the inner citadel excavations. In total, 11 fragments are characterized primarily by their flatness and thinness: their top surface is smooth, their bottom either smooth or rough, and curved raised edges terminate in rounded rims. Their form and sturdy lightness suggest their function as utilitarian and portable trays. They were likely fashioned and fired at low tempera‑ ture in the same place, even though most have smooth bottoms, suggesting they were handled during shaping. Representative samples are noticeably lighter than non ‑portable ‘installation– tray’ fragments. Their coarse fabric has well sorted inclusions, the most common of which are micas and calcium carbonate. Their color ranges from reddish brown (2.5YR 5/4) to pinkish grey (7.5YR 6/2). Their thickness ranges from 1.9–4.9 cm, with an average around 3.2–3.6 cm. Their overall size and shape cannot be determined with certainty owing to their fragmentary state of preservation, but they were likely either rectangular or semi‑rounded in plan: most examples have curved rather than straight edges in plan with straight raised edges. (Fig. 14). Examples of asimilar class of objects have been documented recently at the Late Bronze Age site of Mitrou in eastern Lokris, Greece. The abundance of ‘utilitarian trays’ at Mitrou does not Excavation Area Context Sample Colour Portable item – tray 93.545 98 158 5YR 5/4 93.545 185 6 7.5YR 5/2 93.545 196 394 5YR 6/4 93.545 240 8 7.5YR 6/2 93.545 265 12 7.5YR 6/2 93.545 278 6 5YR 5/1 93.545 300 7 5YR 5/4 93.545 334 60 10R 5/4 97.541 68 15 5YR 5/4 97.541 199 5 7.5YR 5/2 99.526 169 9 2.5YR 5/4 Excavation Area Context Sample Colour Portable item – unclear 93.545 149 11 n/a 95.555 109 7 5YR 4/4 97.541 25 641 2.5YR 6/4 97.541 111 139 7.5YR 6/4 98.531 6 80 2.5YR 4/4 108.522 63 5 10YR 4/1
52 STUDIA HERCYNIA XXIV/1 appear to be an exception; rather, according to K. Jazwa, similar artifacts have been recovered from other Late Bronze Age sites in Greece but remain unpublished or have been misidentified as roofing tiles or crucibles (e.g., at Kynos, Nemea, Alimos, Kalapodi, and Tsoungiza; see Jazwa forthcoming, with bibliography). Unlike examples from Kaymakçı, however, Mitrou’strays tend to be of arectangular shape, and are rich in organic aggregate and prone to cracking. Much later examples of similar trays– so ‑called bread trays– are well ‑known from Lydian Sardis, yet no relation to Kaymakçı’sBronze Age varieties can be established (c.f. Ramage 1978, 8, fig. 18; these were of rectangular shape with flat central part and straight raised edges). Portable item– unclear Asmaller subset of portable items, six fragments altogether, could not be identified with precision, although they are flat and have fabrics similar to those of trays. As their raised edges are not preserved, however, and thus their shape cannot be ascertained, they remain categorized as portable, yet unclear in form. UNCLEAR As the name indicates, samples that could be classified as neither architecture, installation, nor portable item have been grouped in aseparate ‘unclear’ category (Tab.4). The vast majority of fragments were too poorly preserved for any functional identification– in total, 116 fragments or 31.9% of the entire dataset. The sheer number of fragments in this category demonstrates that much of our ability to identify form and function depends on preservation as well as context. However, an additional 31 samples could be assigned to afunctional subcategory– Fig. 14: Example of aportable tray (93.545.265.12) (N. Gail; ©Gygaia Projects).
53JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY mudbrick or impressed. Of these, 29 samples could be identified as some type of mudbrick. They were poorly preserved or lacked contextual clues, which made more specific associa‑ tion with architecture or installations impossible. Their morphology, color, and inclusions match both those of mudbricks associated with architecture and installations, but because of their extreme fragmentation, it was decided not to sort them further based solely on the presence– common in bricks associated with architecture– or absence– usually absent in bricks associated with installations– of chaff and straw. Similarly, two ‘impressed’ fragments bearing impressions of reeds, planks, and rocks were too fragmentary for their function to be determined; they might have been part of either architectural features or installations. GRANULOMETRIC ANALYSIS In the previous section major functional classes of earthen materials were outlined, explaining macroscopic differences in terms of color, fabric, and aggregate composition. In this section (cf. also Tab.5; Pl. 1/1), qualitative macroscopic characterization is supplemented with the results of quantitative granulometric (or particle ‑size) analysis. As described earlier, particle‑ ‑size analysis evaluates the ratio of silt, sand, and clay in sun ‑dried, non ‑hardened earthen materials. This, in turn, offers insights into resource needs and manufacturing processes and reflects manufacturer choices and user preferences. In total, 40 samples were selected based on two primary factors: context– samples were selected from arepresentative range of well ‑preserved features or stratigraphically significant fills; and preservation– samples were selected from fragments that were only sun ‑dried (and not exposed to elevated temperatures) so that they could be dissolved in adeflocculant solution. Most common inclusions seem to be of local origins: micas (biotite, mica ‑schist, and occasionally muscovite), limestone, and occasionally also grog and shells. In afew samples, quartz was detected as well. While chaff and straw inclusions are common in mudbricks, they preserve as voids or impressions in mudbrick fabrics only and thus are undetectable by this analysis. Architecture Impressed earthen architectural fragments were often too fragmentary and/or hardened to allow particle ‑size analysis, resulting in aselection of architectural fragment samples deriving only from mudbricks and unclear fragments. Mudbricks present the most informative dataset because of abundant comparative research. The ideal ratio of sand to clay and silt in mudbricks is posited to be less than 50% sand and 25–45% clay. This ratio yields the best combination of tensile strength (assured by coarse sand and aggregate), binding strength between particles (assured by silt and clay), and minimal shrinkage during drying (assured by the addition of temper) (Rosen 1986; Emery 2009, 2). Mudbrick compositions at Kaymakçı varied greatly, as five of the nine processed samples contained more than 50% silt/clay, while the remaining four samples range between 16.6% to 35.2% silt/clay content. Unclear architectural fragments were less coarse in general, with higher percentages of silt/clay (41.2–70.9%) than sand. Installations Installations comprise aspecial category of evidence at Kaymakçı, as many were recorded in better preserved lower stratigraphic levels, thus providing rare glimpses of the site’sprimary contexts. Moreover, much of these features– including hearth, ovens, and/or working plat‑ forms, with their mudbrick rims– did not harden, which, on the one hand, made excavating them more challenging, but, at the same time, allowed for abetter characterization of their morphology and composition.
54 STUDIA HERCYNIA XXIV/1 Excavation Area Context Sample Colour Unclear– unclear 81.551 3 46 7.5YR 6/4 81.551 3 96 7.5YR 5/2 81.551 4 5 n/a 81.551 20 92 5YR 6/4 81.551 31 10 n/a 93.545 2 3 n/a 93.545 2 6 n/a 93.545 7 8 n/a 93.545 7 30 n/a 93.545 18 5 n/a 93.545 36 9 7.5YR 6/4 93.545 43 64 5YR 6/4 93.545 53 7 n/a 93.545 68 8 5YR 4/4 93.545 124 6 n/a 93.545 124 21 5YR 5/6 93.545 138 6 n/a 93.545 189 8 n/a 93.545 245 5 n/a 93.545 249 9 5YR 5/4 93.545 263 6 n/a 93.545 270 6 n/a 93.545 275 6 n/a 93.545 277 3 n/a 93.545 280 9 10YR 6/4 93.545 286 6 n/a 93.545 292 1 n/a 93.545 296 6 n/a 93.545 297 6 n/a 93.545 298 8 n/a 93.545 299 6 n/a 93.545 300 25 7.5YR 6/4 93.545 302 6 n/a 93.545 303 4 n/a 93.545 304 6 n/a 93.545 307 6 n/a 93.545 308 7 n/a 93.545 315 6 n/a 93.545 333 8 n/a 93.545 335 7 n/a 95.555 2 8 n/a 95.555 2 164 n/a Excavation Area Context Sample Colour Unclear– unclear 95.555 41 46 n/a 95.555 91 6 10YR 2/1 95.555 115 6 n/a 95.555 120 5 n/a 95.555 130 7 n/a 95.555 141 11 n/a 97.541 2 8 n/a 97.541 6 86 n/a 97.541 6 87 n/a 97.541 10 3 n/a 97.541 16 614 n/a 97.541 17 4 n/a 97.541 22 2 n/a 97.541 22 9 n/a 97.541 33 3 n/a 97.541 38 6 n/a 97.541 43 3 n/a 97.541 46 6 n/a 97.541 49 5 n/a 97.541 51 6 n/a 97.541 51 7 n/a 97.541 98 8 n/a 97.541 100 5 n/a 97.541 159 3 n/a 97.541 218 7 n/a 97.541 261 1 10YR 3/2 97.541 268 6 n/a 97.541 278 72 n/a 97.541 384 2 n/a 97.541 486 5 n/a 98.531 1 63 2.5YR 5/6 98.531 3 67 7.5YR 6/4 98.531 6 10 n/a 98.531 15 3 n/a 98.531 26 3 n/a 98.531 34 7 2.5YR 4/4 98.531 101 9 n/a 99.526 1 35 10R 4/4 99.526 5 54 2.5YR 6/6 99.526 6 4 n/a 99.526 7 3 2.5YR 5/6 99.526 10 4 n/a
61JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY activities, and earthen materials and objects here were broadly represented by the fills of the pits. In fact, much of the material was highly fragmented. Remains of architectural fragments, parts of installations, and portable items were discarded as part of secondary and tertiary fills, perhaps as aresult of periodic clearing of the area. On the southern terrace, residential, household, and workshop activities, such as food processing, small ‑scale crafting, and storage, took place in areas 99.526 and 108.522/109.523. Area 99.526, in particular, was important for the present study, as it was the only area in which earthen installations, such as hearths and ovens, were preserved in situ (as part of late LB 1 and early LB 2 strata). Despite the relatively good preservation of features that were buried in the deeper stratigraphic levels, much of the material remains of unclear form and function. CONCLUSIONS This study presented an overview of different categories of earthen materials discovered at the site of Kaymakçı. It aimed to demonstrate the importance of studying datasets in their en‑ tirety and in areflective manner, grappling even with remains of unknown form and function. Even if one in three collected fragments from the excavated areas at Kaymakçı could not be identified beyond being purposely manufactured from amixture of sediment and aggregate, careful documentation of both the context of discovery and the characteristics of excavated features and objects allowed for new information and interpretation. While this number may seem large at first, the proportion of unidentifiable fragments was much higher in the first year of the study. By tackling these traditionally unclassified materials, incorporating increasing knowledge of Kaymakçı’sarchaeological remains, and re ‑studying the earthen assemblage, we gained abetter understanding of the entire dataset and were able to interpret its function in 81.551 93.545 95.555 97.541 98.531 99.526 108.522/109.523 Total Architecture – mudbrick 6 26 7 14 4 19 14 90 Architecture – unclear 000005 0 5 Architecture – impressed 0 22 1 1 0 0 4 28 Installation – surface 0 15 0 0 0 11 0 26 Installation – substructure 000004 0 4 Installation – mudbrick 0 1 0 0 0 16 1 18 Installation – tray 030101 0 5 Installation – impressed 000004 0 4 Installation – unclear 041239 1 20 Portable item – tray 080201 0 11 Portable item – unclear 011210 1 6 Unclear – unclear 5 35 8 24 7 19 18 116 Unclear – mudbrick 194201 12 29 Unclear – impressed 000200 0 2 Total 12 124 22 50 15 90 51 364 Tab.6: Overall counts of the different types of earthen features and objects per excavated area.
62 STUDIA HERCYNIA XXIV/1 more specific terms. This result was possible only because of collaboration across the various specialisms of the excavation team, bringing into conversation excavators, ceramics and small finds specialists, as well as analytical scientific and conservation crews. While the multiyear analysis of earthen materials and features produced ageneral classifi‑ cation for Kaymakçı, the morphology and function of earthen features and objects at sites both near and far away might differ significantly because of the ‘hyper ‑local’ nature of this class of archaeological material. We therefore stress the need of increased site ‑specific research and publication that would allow data comparisons. It is not only the diversity of local forms and their poor preservation, however, that makes inter ‑site comparisons difficult. Issues of recognition and misidentification, too, are pervasive, because this is achallenging class of material to excavate, identify, analyze, and curate. Yet, it is precisely this very challenging nature of the material that makes its increased study and publication imperative in order to reveal the diversity of ways it was used every day by Late Bronze Age communities in the eastern Mediterranean. ACKNOWLEDGEMENTS This work would not have been possible without substantial collaboration with our local col‑ laborators, funding from the US National Endowment for the Humanities (Award RZ5155613), US National Science Foundation (Award 1261363), and Merops Foundation, and permissions and support from the Manisa Museum of Archaeology and Ethnography, and General Di‑ rectorate of Cultural Heritage and Museums, Ministry of Culture and Tourism, Republic of Turkey. Moreover, we would like to acknowledge and thank all members of the Kaymakçı Archaeological Project. BIBLIOGRAPHY Aurenche, O. 1981: La maison orientale. L’architecture du Proche ‑Orient ancien des origines au milieu du quatrième millénaire. Paris. Avrami, E.– Guillaud, H.– Hardy, M. eds. 2008: Terra Literature Review. An Overview of Research in Earthen Architecture Conservation. Los Angeles. Cammas, C. 2018: Micromorphology of Earth Building Materials. Toward the Reconstruction of Former Technological Processes (Protohistoric and Historic Periods). Quaternary International 483, 160–179. Claasz Cookson, B. 2010: Living in Mud. Istanbul. Curtis, C. 2009: Geoarchaeological Analyses of Ancient and Modern Mud ‑brick in Central Lydia, Western Turkey. Unpublished BA Research Paper. Boston. Costi de Castrillo, M.– Philokyprou, M.– Ioannou, I. 2017: Comparison of Adobes from Pre ‑History to ‑Date. Journal of Archaeological Science, Reports 12, 437–448. Devolder, M.– Lorenzon, M. 2019: Minoan Master Builders? ADiachronic Study of Mudbrick Architecture in the Bronze Age palace at Malia (Crete). Bulletin de correspondance hellénique 143, 63–123. Emery, V.L. 2009: Mud ‑Brick. In: W. Wendrich (ed.): UCLA Encyclopedia of Egyptology. Los Angeles, 1–10. Emery, V.L.– Morgenstein, M.E. 2007: Portable EdXRF Analysis of aMud Brick Necropolis Enclosure: Evidence of Work Organization, El Hibeh, Middle Egypt. Journal of Archaeological Science 34, 111–122. Forgetetal. 2015 = Forget, M.C.– Regev, L.– Friesem, D.E.– Shahack ‑Gross, R.: Physical and Mineralogical Properties of Experimentally Heated Chaff ‑Tempered Mud Bricks. Implications for Reconstruction of
63JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY Environmental Factors Influencing the Appearance of Mud Bricks in Archaeological Conflagration Events. Journal of Archaeological Science, Reports 2, 80–93. French, C.A.I. 1984: ASediments Analysis of Mudbrick and Natural Features at El ‑Amarna. In: B.J. Kemp (ed.): Amarna Reports I/1. Egypt Exploration Society Occasional Publications. London, 189–201. Friesemetal. 2011 = Friesem, D.– Boaretto, E.– Eliyahu ‑Behar, A.– Shahack ‑Gross, R.: Degradation of Mud Brick Houses in an Arid Environment. AGeoarchaeological Model. Journal of Archaeological Science 38, 1135–1147. Friesem, D.E.– Karkanas, P.– Tsartisidou, G. 2014: Sedimentary Processes Involved in Mud Brick Degra‑ dation in Temperate Environments. AMicromorphological Approach in an Ethnoarchaeological Context in Northern Greece. Journal of Archaeological Science 41, 556–567. Goldberg, P.1979: Geology of the Late Bronze Age Mudbrick from Tel Lachish. Tel Aviv 6, 60–67. Goldberg, P.– Macphail, R.2006: Practical and Theoretical Geoarchaeology. Malden, MA. Guest ‑Papamanoli, A. 1978: La brique crue en Egée au Néolithique et à l’Âge du Bronze. Bulletin de cor‑ respondance Hellénique 102, 3–24. Guillaud, H. 2008: Characterization of Earthen Materials. In: E. Avrami– H. Guillaud– M. Hardy (eds.): Terra Literature Review. An Overview of Research in Earthen Architecture Conservation. Los Angeles, 21–31. Homsher, R.S. 2012: Mud Bricks and the Process of Construction in the Middle Bronze Age Southern Levant. Bulletin of the American Schools of Oriental Research 368, 1–27. Houben, H.– Guillaud, H. 1994: Earthen Constructions. AComprehensive Guide. London. Jazwa, K.A. forthcoming: Utilitarian Ceramic Trays from Mitrou. Introducing aNew Artifact Class. In: Pro‑ ceedings of the 5th Archaeological Meeting of Thessaly and Central Greece, 2012–2014. Volos. Jerome, P.S. 1991: Analysis and Conservation of Mudbrick Construction in Bronze Age Crete. Unpublished MA Thesis. New York. Kemp, B. 2000: Soil (Including Mud ‑Brick Architecture). In: P.Nicholson– I. Shaw (eds.): Ancient Egyptian Materials and Technology. Cambridge, 78–103. Lorenzon, M.– Iacovou, M. 2019: The Palaepaphos ‑Laona Rampart. APilot Study on Earthen Architecture and Construction Technology in Cyprus. Journal of Archaeological Science, Reports 23, 348–361. Love, S. 2012: The Geoarchaeology of Mudbricks in Architecture. AMethodological Study from Çatalhöyük. Geoarchaeology 27, 140–156. Love, S. 2013: An Archaeology of Mudbrick Houses from Çatalhöyük. In: I. Hodder (ed.): Substantive Technolo‑ gies at Çatalhöyük. Reports from the 2000–2008 Seasons. Çatalhöyük Research Project Series 9. London– Los Angeles, 81–96. Love, S. 2017: Field Methods for the Analysis of Mud Brick Architecture. Journal of Field Archaeology 42/4, 351–363. Luke, C.– Cobb, E. 2013: Dwelling in Hacıveliler. Social Engineering Policies in the Context of Space, Place and Landscape in Rural, Western Turkey. Anatolian Studies 63, 155–173. Luke, C.– Roosevelt, C.H. 2009: Central Lydia Archaeological Survey. Documenting the Prehistoric through Iron Age Periods. In: S.W. Manning– M.J. Bruce (eds.): Tree ‑Rings, Kings, and Old World Archaeology and Environment. Papers Presented in Honor of Peter Ian Kuniholm. Oxford– Oakville, 199–217. Luke, C.– Roosevelt, C.H.– Scott, C.B. 2017: Yörük Legacies. Space, Scent, and Sediment Geochemistry. International Journal of Historical Archaeology 21/1, 152–77. Mielke, D.P.2009: Alte Paradigmen und neue Erkenntnisse zur hetitischen Holtz ‑Lehmziegel ‑Architektur. In: M. Bachmann (ed.): Bautechnik im antiken und vorantiken Kleinasien. Internationale Konferenz, 13.–16. Juni 2007 in Istanbul. Istanbul, 81–106. Morgenstein, M.E.– Redmount, C.A. 1998: Mudbrick Typology, Sources, and Sedimentological Compo‑ sition. ACase Study from Tell el ‑Muqdam, Egyptian Delta. Journal of American Research Center in Egypt 35, 129–146. Naumann, R.1971: Architektur Kleinasiens von ihren Anfängen bis zum Ende der hethitischen Zeit. Tübingen.
64 STUDIA HERCYNIA XXIV/1 Nodarou, E.– Frederick, C.– Hein, A. 2008: Another (Mud)brick in the Wall. Scientific Analysis of Bronze Age Earthen Construction Materials from East Crete. Journal of Archaeological Science 35, 2997–3015. O’Gradyetal. 2018 = O’Grady, C.R.– Luke, C.– Mokrišová, J.– Roosevelt, C.H.: Interdisciplinary Approaches to Understanding and Preserving Mudbrick Architecture in Regional and Diachronic Contexts. Cogent Arts & Humanities 5. Peinettietal. 2017 = Peinetti, A.– Aprile, G.– Caruso, K.– Speciale, C.: Looking for aScientific Protocol in Prehistoric Daub Experimental Project. In: R. Alonso– J. Baena– D. Canales (eds.): Playing with the Time. Experimental Archaeology and the Study of the Past. Madrid, 307–311. Ramage, A. 1978: Lydian Houses and Architectural Terracottas. Archaeological Exploration of Sardis 5. Cam‑ bridge, MA. Roosevelt, C.H.– Luke, C. 2008: Central Lydia Archaeological Survey: 2006 Results. Araştırma Sonuçları Toplantısı 25/3, 305–326. Roosevelt, C.H.– Luke, C. 2009: Central Lydia Archaeological Survey: 2007 Results. Araştırma Sonuçları Toplantısı 26/2, 433–450. Roosevelt, C.H.– Luke, C. 2010: Central Lydia Archaeological Survey: 2008 Results. Araştırma Sonuçları Toplantısı 27/1‑2, 1–24. Roosevelt, C.H.– Luke, C. 2011: Central Lydia Archaeological Survey: 2009 Results. Araştırma Sonuçları Toplantısı 28/3, 55–74. Roosevelt, C.H.– Luke, C. 2012: Central Lydia Archaeological Survey: 2010 Results. Araştırma Sonuçları Toplantısı 29/1, 383–400. Roosevelt, C.H.– Luke, C. 2013: The Central Lydia Archaeological Survey: 2011 Work at Kaymakçı and in the Marmara Lake Basin. Araştırma Sonuçları Toplantısı 30/1, 237–54. Roosevelt, C.H.– Luke, C. 2017: The Story of aForgotten Kingdom? Survey Archaeology and the Historical Geography of Central Western Anatolia in the Second Millennium BC.European Journal of Archaeology 20/1, 120–147. Rooseveltetal. 2014 = Roosevelt, C.H.– Luke, C.– Cobb, P.– O’Grady, C.R.– Sekedat, B.: The Central Lydia Archaeological Survey. 2013 Work at Kaymakçı and in the Marmara Lake Basin. Araştırma Sonuçları To‑ plantısı 31/1, 333–355. Roosevelt, C.H.– Luke, C.– Sekedat, B. 2016: The Central Lydia Archaeological Survey. 2014 Work at Kay‑ makçı and in the Marmara Lake Basin. Araştırma Sonuçları Toplantısı 33/2, 251–262. Rooseveltetal. 2018 = Roosevelt, C.H.– Luke, C.– Ünlüsoy, S.– Çakırlar, C.– Marston, J.M.– O’Grady, C.R.– Pavúk, P.– Pieniążek, M.– Mokrišová, J.– Scott, C.– Shin, N.– Slim, F.: Exploring Space, Economy, and Interregional Interaction at aSecond ‑Millennium BCE Citadel in Central Western Anatolia: 2014–2017 Research at Kaymakçı. American Journal of Archaeology 122/4, 645–688. Rosen, A. 1986: Cities of Clay. The Geoarchaeology of Tells. Chicago. Rosenstock, E. 2009: Tells in Südwestasien und Südosteuropa. Urgeschichtliche Studien 2. Remshalden ‑Grunbach. Sauvage, M. 1998: La brique et sa mise en œuvre en Mésopotamie. Des origines à l’époque achéménide. Paris. Schwandner, E.L. 1999: Einzelprobleme. In: W. Hoepfneretal.: Die Epoche der Griechen. In: W. Hoepfner (ed.): Geschichte des Wohnens 1. Stuttgart, 526–536. Seeher, J. 2007: AMudbrick City Wall at Hattuša. Diary of aReconstruction. Istanbul. Shaffer, G. 1993: An Archaeomagnetic Study of Wattle and Daub Building Collapse. Journal of Field Archae‑ ology 20, 59–75. Shaw, J.W. 2009: Minoan Architecture. Materials and Techniques. Studi di Archaeologia CreteseVII. Padova. Shell, C.A. 1997: Appropriate Geophysics and Excavation Strategy. From Mud Brick to Masonry in the East Mediterranean Region. In: A.J. Sinclair– E.A. Slater– J. Gowlett (eds): Archaeological Sciences 1995. Proceedings of aConference on the Application of Scientific Methods to Archaeology. Monographs in Archaeology 64. Oxford, 333–342.
65JANA MOKRIŠOVÁ– CHRISTOPHER H. ROOSEVELT– CHRISTINA LUKE – CAITILIN R. O’GRADY Stevanovic, M.C. 1997: The Age of Clay. The Social Dynamics of House Destruction. Journal of Anthropological Archaeology 16, 334–395. Tung, B. 2005: APreliminary Investigation of Mudbrick in Çatalhöyuk. In: I. Hodder (ed.): Changing Mate‑ rialities at Çatalhöyuk. Reports from the 1995–99 Seasons. Cambridge, 215–219. van Beek, G.W.– van Beek, O. 2008: Glorious Mud! Ancient and Contemporary Earthen Design and Construction in North Africa, Western Europe, the Near East, and Southwest Asia. Washington, D.C. Wolff, N. 2008: Geoarchaeology of Iron Age Burial Mounds in Lydia, Western Turkey. Unpublished MA Thesis, Boston. Wright, G.R.H. 2005: Ancient Building Technology. Leiden. Jana Mokrišová Research Associate Faculty of Classics, University of Cambridge Sidgwick Avenue, Cambridge, CB3 9DA [email protected].uk Christina Luke Department of Archaeologyand History of Art Koç University Rumelifeneri, Sarıyer İstanbul, TR‑34450 [email protected] Christopher H. Roosevelt Department of Archaeologyand History of Art Research Center for Anatolian Civilizations Koç University Rumelifeneri, Sarıyer İstanbul, TR‑34450 [email protected] Caitlin R. O’Grady Institute of Archaeology University College London 31‑34 Gordon Square London, WC1H 0PY, UK caitlin.r.[email protected].uk
197PLATES Pl. 1/1: Percentage representation of sand and silt/clay particles in earthen features from Kaymakçı. Pl. 1/2: Percentage representation of sand and silt/clay particles in constituent components of aplatform/oven (99.526.79).
198 STUDIA HERCYNIA XXIV/1 Pl. 1/3: Distribution of the different types of earthen features and objects per excavated area.