Introduction

Construction materials, referred to as ‘earthen construction materials’, have been investigated worldwide using a variety of methods including archaeological soil micromorphology (e.g. Matthews 1995, 2017; Matthews et al. 1996; Middleton 2004; Barba 2007; Karkanas and de Moortel 2014; Kanthilatha et al. 2014, 2017; Friesem et al. 2017; Amadio 2018; Macphail and Goldberg 2018; Lisá et al. 2020; Lorenzon 2023). Among the many different possible approaches to archaeologically studying earthen construction materials, this paper combines archaeological soil micromorphological and phytolith analyses of floors, daub and hearth construction material with a chaîne opératoire analysis. In doing so, this paper attempts to combine the application of natural scientific techniques to archaeological materials, with a social interpretation which sees the built environment (e.g. house, oven, hearth etc.) as the result of socially and culturally conditioned choices. Microscopic techniques have already been successfully applied to investigate social aspects of built environments (Boivin 2000; Karkanas and Efstratiou 2009; Anderson et al. 2014; García-Suárez et al. 2021; Tomé et al. 2024), and, in particular, building techniques (e.g. study of daubs, re-plasterings, renovation) (Cook et al. 2006; Kreiter et al. 2013, 2014; Pető and Vrydaghs 2016; Pető et al. 2015; Pecci et al. 2017). In this study, we emphasise a precise description of the ‘earthen’ building materials used at the analysed sites. The term earthen building materials mostly refers to soils and sediments (Friesem et al. 2017). Wherever the origin of ‘earthen building material’ has been investigated, mostly sediments, but also soils have been used for construction (Goldberg 1979; Rosen 1986; Macphail and Goldberg 2018; Wattez et al. 2020). However, in many cases there is no focus on pedological and geological differentiations (Lorenzon 2023). Depending on the availability of sources in the surrounding environment, the material can be soil, muddy sediments or parent sediment (parent material); each has completely different properties as a consequence of different origins. Whether weathered or unweathered, the initial choice of material from a particular environment can provide indications of specific techniques and expertise. Similar to many other prehistoric archaeological sites worldwide, earthern materials also predominate Middle Bronze Age architecture in Hungary (Courty et al. 1989; Nicosia and Stoops 2017; Cammas 2018; Knoll and Klamm 2020), and require a precise description and categorization to better understand their formation, utilisation and social aspects. To date the literature on Hungarian Bronze Age architecture lacks such a precise description. For example, when floor building matter is discussed in the Hungarian archaeological record two main types are differentiated based on the macroscopically observable composition and colour (for further reference see Kovács and Vicze 2022): 'Yellow clay’ and ‘earthen’ floors are distinguished, which in our recent study was further specified in the case of Százhalombatta and Borsodivánka sites (see Röpke et al. 2024b), and the traditionally termed ‘clay’ floors were differentiated into silt (loess) floors, being very fine (but not clay size) in composition and yellowish in colour, and earthen floors, being beige or brownish-greyish in colour with a coarser matrix (loam/sandy loam texture).

A chaîne opératoire approach – an archaeological analytical technique that attempts to reconstruct the technical and social processes that occurred in connection with the production, use and discarding of archaeological objects – is one way to achieve the goal of viewing building materials as social or craft products, so that rather than applying little to no analysis to them, building materials are treated as archives of social and environmental information (Love 2013; Lorenzon 2021). The production of a floor, wall or hearth includes similar processes when compared to other craft products (Sofaer 2006, 2011), such as vessels, tools or metal objects. The building materials are composed of soil or sediment, with water and various other materials included (as temper...). The building materials undergo a complex chaîne opératoire which complete transforms them (Lorenzon 2023). This implies that the knowledge of the raw materials’ characteristics, their availability, and the desired properties of the finished material are similarly important when preparing a floor, just as in the case of a pot for example. Reconstructing the building materials’ chaîne opératoire − the sequence of decisions made by builders around the sourcing, preparation and application of them − makes social choices visible (Kalogiropoulou et al. 2023; Uzdurum et al. 2023). These choices are often specific to a cultural unit, site or even smaller social units that display certain idiosyncrasies while building, i.e. faced with a particular decision, one group made one choice and another group made another for reasons that can be difficult to reconstruct, but are best researched within the given cultural context, not compared to our modern notions of ‘best practice’ in building (Anvari 2021: 5–6, 167−170). Knowledge regarding construction also needs to be transferred and from the available archaeological data it is evident that different societies have differences in construction during the Bronze Age (Kovács 1977; Bóna 1992). Most chaîne opératoire approaches to earthen building materials are based on mudbrick architecture from arid environments (e.g. Love 2012; Lorenzon et al. 2023; Uzdurum et al. 2023), but here we intend to apply it to wattle and daub constructions and sediment-based construction materials in a more humid continental climate. In Hungary, the majority of data regarding Middle Bronze Age construction is mainly from the archaeological perspective (e.g. Poroszlai 1996; Vicze 1992; Jaeger et al. 2018) with a few more recent exceptions (Vicze 2013a, 2013b; Parkinson et al. 2018; Kovács 2018; Kovács et al. 2020, 2023), so the size of the dwellings, their orientation, inner division, some indication on the position of the entrance, the material of floors (being mainly ‘clay’ or ‘earth’), wattle and daub walls and reed/straw roof on wooden frame is mainly recorded macroscopically. With limited in-depth data, it is hard to assess how building materials can be socially significant or indicative in this region, so in this study we will investigate this aspect. Floors, daub and hearth construction materials will be studied from three Middle Bronze Age tell sites in Hungary – Százhalombatta-Földvár, Borsodivánka-Marhajárás-Nagyhalom and Kakucs-Turján (Fig. 1) – to gain and compare data on construction materials and techniques that will enable us to investigate socio-technological differences. Although all the sites are located on the Great Hungarian Plain (Fig. 1), the immediate environment of each site has distinctive qualities with differences in the inhabiting communities those being, Vatya at Százhalombatta-Földvár and Kakucs-Turján, and Otomani-Füzesabony Cultural Circle (OFCC) at Borsodivánka-Marhajárás-Nagyhalom. Archaeological soil micromorphological research on building materials at these sites has primarily focused on floors (Kovács 2018; Kovács et al. 2020; Kovács and Vicze 2022; Kovács et al. 2023; Röpke et al. 2024b), and constitutes the core of this paper, but daub and hearth materials are also discussed.

Fig. 1
figure 1

Location of the study sites: a location of Hungary within Europe; b location of the three study sites within the territory of Hungary; c drone image of Százhalombatta-Földvár tell settlement (white bracket); d geological environment of Százhalombatta-Földvár displayed on 1:100.000 geological map, oriented to the North (red bracket: location of the site) (legend: eQp3l – loess, fQh – Holocene fluvial sediment (general); e drone image showing the surroundings of Borsodivánka-Marhajárás-Nagyhalom tell settlement and close-up image of the excavation trench (white bracket enlarged); f geological environment of Borsodivánka-Marhajárás-Nagyhalom displayed on 1:100.000 geological map, oriented to the North (red bracket: location of the site) (legend: fQh2al- Holocene fluvial sediment (aleurite), fQp3al - Pleistocene fluvial sediment (aleurite); g edited Google Maps satellite image of Kakucs-Turján settlement and its structure displayed on magnetometry image (enlarged white bracket); geological environment of Kakucs-Turján (red bracket: location of the site) displayed on 1:100.000 geological map, oriented to the North (legend: eQp3l – loess, eQp3lh – loessy sand, eQp3h – sand; lQh2mi – lime mud)

Full size image

Research sites and archaeological context

Százhalombatta-Földvár with its ca. 5 m thick cultural layers is a Bronze Age fortified tell settlement studied by the SAX Project (Poroszlai and Vicze 2000; Sørensen et al. 2020) since 1998. The settlement is one of a series of tell sites of the Vatya cultural complex situated along the western high bank of the River Danube South from Budapest. It is located on a plateau overlooking the Danube, bordered by natural water drains from the South and West and enclosed by fortification ditches in the North. The lifespan of the tell is approximately 700-800 years. A Roman watchtower and a Word War II rifle-pit are in close proximity attesting to the advantageous position of the location (Vicze and Sørensen 2023). The excavation on the central part of the site is conducted in a 20 m by 20 m trench and, so far, has revealed 4 m of complex occupation levels with several partially-preserved houses. The site was first settled during the Early Bronze Age. It is still under excavation, therefore the 400-500 years of the Middle Bronze Age contexts (ca. 2000/1900–1500/1450 cal BC), belonging to the Vatya cultural group will be examined here. This group is one of the so-called tell-forming communities occupying the eastern half of the Carpathian Basin from the second half of the Hungarian Early Bronze Age (ca. 2300/2200 cal BC). The Vatya social group, together with its Early Bronze Age Nagyrév cultural predecessor, is the only such community that crosses the Danube towards the west, thus encompassing both sides of the Danube in Central Hungary. This culture has been identified via its characteristic pottery style, distinct hierarchical settlement system (Bóna 1975; Kovács 1982; Vicze 2000), and a very consistent cremation burial custom. Százhalombatta-Földvár and Kakucs-Turján represent two out of the several distinct settlement types attributed to this social group.

With cultural layers of ca. 3 m thickness surviving into modern times Borsodivánka-Marhajárás-Nagyhalom belongs to the group of proper tell sites in the Bronze Age Borsod plain studied by the BORBAS project (e.g. Kienlin et al. 2018: 163–169; P. Fischl et al. 2022a, b). The settlement is situated on the northern edge of a flood-free peninsula surrounded by marshland and old river arms of today’s Rima, Kánya and Eger streams. The modern vegetation is composed of a mosaic of oak forest with pastures, agricultural and ruderal areas as well as more swampy regions covered with reed. It comprises the central multi-layer tell section mentioned, enclosed by a ditch that connected to the adjacent riverbed or swamp and a horizontal outer settlement on the slightly elevated southern island area. The excavation on the central tell part carried out between 2015 and 2024 exposed a complex succession of occupation phases on the margin of the tell, separated by use of the plot as a midden comprising multiple phytolith layers and intermediate waste layers. While a comprehensive dating programme of the excavated house phases is still pending, from the profile initially cleaned at the start of the excavation there is evidence of about 3 m of Otomai-Füzesabony Cultural Circle stratigraphy building up over possibly ca. 300 years during the 19th to 17th centuries cal BC (Kienlin 2020: 88–90, 104 fig. III-73). The Otomani-Füzesabony Cultural Circle (OFCC) is a Bronze Age cultural unit that encompasses parts of present-day Hungary, Slovakia, Ukraine, Romania, and Poland. The OFCC was primarily distinguished from other communities in the Carpathian Basin by its burial customs and its material culture, distinct ceramic forms and decorative motifs, which prominently featured concentric circles, spirals, semicircles, and garlands throughout its distribution area (P. Fischl and Kienlin 2019).

The site of Kakucs-Turján is also a multi-layered settlement of the Vatya cultural group investigated by the KEX project (Jaeger et al. 2018). It is located in Central Hungary, southeast of Budapest, in the valley of the ancient Danube riverbed. Although there is some evidence of Early Bronze Age habitation, the lifespan of the settlement is mostly related to the Middle Bronze Age Vatya period (2000/1900 − 1500/1450 (Jaeger et al. 2022)). The area can be characterised by gentle slopes and slightly undulating plains covered by fluvio-aeolian sands that are interrupted by the mosaics of loess derivates and former meanders, filled up with alluvial sediments (Pető et al. 2019; Niebieszczański et al. 2019).

Materials and methods

In total 21 samples (representing 23 micro-contexts) were targetedly selected for investigation to represent three major construction elements, namely floors, wattle and daub walls and fire installation materials (i.e. hearth, oven, fireplace) (Table 1). Floor samples were selected based on our previous studies in which we investigated and compared the so-far known variety of floor matters, and found more precise terminology for the the traditionally termed clay floors based on their petrographic characteristics (Röpke et al. 2024b) (Table 1). Daub and dauby materials including the material of the hearths and oven platforms (i.e. baking surface) were also chosen to represent a selection of materials (i.e. burnt under oxidative and reductive conditions), although no systematic thin section analysis has so far been carried out in these cases. Although only a limited number of samples were available from the Kakucs site, it was important to investigate them in order to identify the cultural and/or geographical differences, as Bronze Age tell contexts for comparison are rare in Hungary. For this study, only the relevant contexts of the thin sections will be analysed and discussed as the aim of this work is solely related to construction materials. In the case of floors only the passive zone (Gé et al. 1993) also termed preparation layer (e.g. Devos et al. 2013) (i.e. the construction material) is the focus here. Wattle and daub walls are only investigated via the daub matter, while in relation to the hearths, platforms will be analysed.

Table 1 Inventory of the investigated samples: Százhalombatta-Földvár (SZHB-FV); Borsodivánka-Marhajárás-Nagyhalom (BorN); Kakucs-Turján (KT)
Full size table

Undisturbed anthropogenic sediment samples were gathered for archaeological soil micromorphological analysis from all three sites in so-called Kubiena boxes during excavation. After air drying, samples were impregnated with polyester resin under vacuum, and polished down to approximately 30 microns (Murphy 1986; Beckmann 2007) in order to analyse them under a polarising microscope (Zeiss Jenalab Pol and Leika DM2700 P). Plane- (PPL) and cross- (XPL) polarised light at various magnifications, ranging between 25–400×, were employed for the identification of the various components (e.g. minerals, organics, anthropogenic inclusions) and traceable processes. The description of the thin sections is in accordance with Bullock et al. (1985), Stoops (2021) and Verrecchia and Trombino (2021). For further identification, Nicosia and Stoops (2017) were also consulted. The phytolith analysis was performed on the prepared thin sections as well, no further bulk samples were available for such purpose. Phytolith analysis within the thin sections followed the methodological procedure applied in previous studies in Hungary (Kovács et al. 2020, 2023). Plant opal particles, silicified tissue elements, as well as amorphous charred plant material were analysed in the thin sections at a magnification ranging between 100× to 400×. Description of phytoliths was based on the International Code for Phytolith Nomenclature 2.0. (ICPT 2019). A modern reference collection of inflorescence bract elements of cereals and relevant scientific literature (Metcalfe 1960; Haraszty 1979; Miller Rosen 1992; Ball et al. 1996, 1999, 2017; Dickinson 2000; Lisztes-Szabó 2019; Lisztes-Szabó et al. 2023) was used to identify the anatomical origin of the observed morphologies. As numerous criteria influence the visibility of phytoliths within a thin section (see Kreiter et al. 2013, 2014; Pető and Vrydaghs 2016; Vrydaghs et al. 2017) the phytolith analysis of the thin sections employed five direct indices (for details consult Vrydaghs and Devos 2018; Kovács et al. 2020):

  1. 1.

    Observed/Not Observed Index (O/NO),

  2. 2.

    Distribution Index (D),

  3. 3.

    Morphotype Identification Index (MI),

  4. 4.

    Aspect Index (A),

  5. 5.

    Associated Material Index (AM).

The description protocol of the micromorphological analysis approach provides semi-quantitative data based on the identification of the relative amount of plant material (organic and inorganic) observed within a sample. This is complemented by the identification of the relative abundance of plant temper in the fabric material based on the presence of pseudomorphic plant voids (Matthews 1995; Shillito et al. 2011), phytolith aggregates, as well as on the presence of not fully decayed plant material in the thin section. Empty voids (i.e. pseudomorphic plant voids) of the thin sections represent the complete decay and loss of plant organic material decayed out (in contrast to the ceramic fabric, where voids are representative of plant material burnt out during the process, see Pető and Vrydaghs 2016 for methodological considerations and Kreiter et al. 2013; 2014 for application). Still, phytoliths and various forms of charred – and in many cases amorphous – plant material might also occur in the matrix and in voids as well. Charred organic plant material is referred to as POM (Plant Organic Matter) within this study. As phytoliths are produced in the tissues and cell lumens of the plant, the primary in-put of phytoliths into the clay material is intentional plant tempering. As a secondary process, plant parts may be incorporated into the clay when the raw material is stored or deposited during and between the production and/or utilisation process.

Results

In order to be able to make observations on the similarities and differences of the various construction materials of the three sites, the micromorphological characteristics were recorded and summarised in ESM#1. Results of the phytolith observations and analyses are summarised in ESM#2. In the following the results are broken down to the separate archaeological sites. Both micromorphological and phytolith observations are described and a microphoto table (Figs. 2, 3 and 4) is also presented for each site. Codes in ESM#2 refer to codes in ESM#1 and to the relevant microphotographs in Figs. 2, 3 and 4.

Fig. 2
figure 2

Microphotographs of the investigated materials from Százhalombatta-Földvár (for letter coding see ESM#1) (magnification 25x or indicated separately; PPL/XPL). A-D matrix of the various silt (‘clay’) floors; E-F matrix of the earthen floors; N-P matrix of the daub matter; Pa-Pb Microphotographs of daub sample id 3280. Pa) general view of the daub with intensive plant tempering; Pb) Elongate entire morphotypes in anatomical sound position in a void of the daub fragment; U-V matrix of the hearth platforms; Ua: Void left over after the charring and shrinking of a grass leaf as well as the charred tissue of the leaf in sample id 3322; Va: Void left over after the charring of grass stem and charred organic material hiding phytoliths within the void in sample id 3321

Full size image
Fig. 3
figure 3

Microphotographs of the investigated building materials from Borsodivánka-Marhajárás-Nagyhalom) (for letter coding see ESM#1) (magnification 25x, except for Sa, X and Y; PPL/XPL). G matrix of the clayey silt, plant tempered floor; H-J matrix of the earthen floors; Q-R matrix of the daub matter; S matrix of the wall decoration; Sa dissolved phytoliths and their remnants, coated by secondary calcite needles in the wall decoration W matrix of the hearth platform; X crinoids; Y articulated Elongate dendritic phytoliths as remnants of the temper (cerealia)

Full size image
Fig. 4
figure 4

Microphotographs of the investigated building materials from Kakucs-Turján (for letter coding see ESM#1) (magnification 25x; PPL/XPL). K-L matrix of the calcareous silt floors; M matrix of the loam floor; T matrix of the daub matter

Full size image

Százhalombatta-Földvár

Floors

Based on the on-site observations and the previous thin section soil micromorphological observations, two very distinctive floor types can be distinguished at the site of Százhalombatta so far (Poroszlai 2000; Vicze 2013a; Kovács et al. 2020, 2023). The archaeologically described yellow(ish) coloured clay floors are the most common, while the presence of brownish-greyish earthen floors were also recorded. Four samples, representing the major types, were chosen (MS39 2004, MS40 2004, MS41/1 2004 and MS13 2004; six micro-horizons in total (representing House ids 3181, 3497, 3076, and 3147)) for this study to characterise floor materials at Százhalombatta-Földvár (Table 1 and ESM#1, Fig. 2).

After the micromorphological analysis a more precise characterization of these matters could be presented (see Röpke et al. 2024b). The ‘clay’ floors mainly consist of silt with some sand grains and a lower amount of clay content in them (ESM#1, Fig. 2A-D). They were prepared from the locally available loess. This type is not tempered with plant matter, although occasionally some plant remains (in the form of phytoliths and pseudomorphic voids) are visible. (The only exception was seen in sample MS41/1 2004 (d), where a higher amount of plant material could be observed, but it still does not reach the amount and intensity of the earthen floors.) Only a few general grass vegetative indicators, such as Elongate entire and Elongate echinate, could be observed within these microlayers (ESM#2).

Such matters can be described as clean building material, as no anthropogenic inclusion or any other type of addition is detectable in them. All of them are very compact, mainly mineral and very fine in composition. The investigated matters are rather calcareous (by nature), but their sorting and the amount of the coarser elements indicate different origins and sources.

The material of the earthen floors is best described as loam, but in these cases some addition of plant matter is detected. In MS39 2004 (g) microlayer, as well as in MS13 2004 (k) microlayer large amounts of in situ plant material could be observed. The identified phytolith morphotypes, which were found within their anatomical position, represent grass leaf (e.g. Elongate sinuate), stem (e.g Elongate entire) and inflorescence (e.g. Elongate dendritic) material as well. Both grass silica short cells (so-called GSSC) and grass silica long cells (so called GSLC) were found; moreover, trichomes were also identified in the silica skeletons (ESM#2). In the samples analysed to date, the addition of ash (including some small, embedded charcoal fragments as well) was also identified (Fig. 2E-F). This building material is considered to be clean, as it excludes macro size anthropogenic inclusions in most of the cases. Bone, burnt bone, ceramic fragments etc. are only occasionally present and tiny in size. Regarding the building technique, renewal of the floor was observed in both cases regardless of the used material. The only ‘universal’ building technique seems to be the order of the floors, ‘earthen’ being always the initial one. Furthermore, until now it seems that no earthen floor remains unplastered/uncovered. Although the discussion of activity remains (i.e. floor build-up) is not included in this paper, it needs to be noted that on the surface of the earthen floors in some of the cases activity remains are present, which shows that such surfaces are not preparational surfaces/foundations, but ‘living’ floors. There is, however, always a silt floor on top of the earthen one if the latter is present. Only silt floors can remain un-renewed (i.e. being single floor horizons). Other than this, no general pattern can be observed. There are houses with a series of silt floor renewals (e.g. House id 3076) and in one case this was also observed in relation to the earthen ones (House id 3147). Silt floors are more frequently renewed in those cases where no earthen one is present. No activity remains were observed in those cases, where the floor was renewed frequently (regardless of being silt or earthen), which indicates that such residues were removed prior to the new flooring (Kovács and Vicze 2022). So far there is no evidence of the surface treatment of the silt floors (i.e. fine re-plasterings, see wall re-platerings below), but previously some microscopic traces of coverings, such as matting were detected (see Kovács 2023).

Daub wall

Three samples were selected (MS8 2009, MS13/1 2002, and id 3280 (representing Debris ids 4693, 1388 and House id 3147)) to investigate and present some of the daub material at Százhalombatta (Table 1 and ESM#1), where the most typical wall building technique is wattle and daub, which is common to all of the investigated sites. Only this building material and technique will be discussed here, although variation in wall building technique is also evidenced at this site (Vicze 2013a). Burnt and unburnt daub fragments have been exposed at the site in large quantities. The material of the daub at Százhalombatta is a more clayey one (best described as clayey loam) as opposed to the silty character of the floor materials, indicating the utilisation of a different type of raw material for wall building. A systematic coring project has already been set up to identify possible sources. When it comes to daub there is always a large amount of plant matter added that we can trace in the thin sections as voids (pseudomorphs of plant temper) and sometimes phytoliths are also present (ESM#1–2 and Fig. 2N-P). It is interesting to note that sample MS8 2009 shows no oxidative part within the daub; the entire thin section exposes a reductive micro-environment, and as a consequence of this plant opal particles cannot be observed. However, an extremely high amount of charred non-arboreal plant material can be observed. On one hand charred grass leaf tissues, on the other hand voids most probably left by grass stems are present all over the thin section. Due to the reductive state of the charred plant tissue material in sample MS8 2009, the more precise identification of the plant material is not probable. This observation – namely that high amounts of plant material derived from grasses are used for intensive tempering – was underlined by sample MS13/1 2002 and also by the observations made during the analysis of sample id 3280 (cf. ESM#2, Fig. 2P, Pa, Pb). In the latter, large amounts of fragmented, mostly non-identifiable phytoliths were observed in the voids, while very few organic materials were present. In those cases where phytoliths were identifiable Elongate entire, Elongate echinate and Trichomes were observed. It is interesting to note that in contrast to the generally observed characteristics we could not prove (or observe) the presence of cereal-related chaff phytoliths in the thin sections. At Százhalombatta, daubs are made of materials with a considerable amount of clay in them, being clayey loams in texture. Thin section analysis showed that walls are frequently maintained, resulting in a series of fine plasterings on the initial surface. No vegetal tempering is visible in the case of the fine layers, they are being silt as opposed to clay (Fig. 2O).

Hearths, ovens

Two samples from two different types of fireplaces (sample id 3322 from the rimmed hearth 3322, and sample id 3321 from oven 3439, both belonging to House 3497) will be used to illustrate the building material of such features (Table 1 and ESM#1, Fig. 2U, V). At Százhalombatta, in the analysed samples the material of the hearth platform is similar to daub as it contains large amounts of plant tempering. (From the on-site observations it is evident that this is only one type of mixture as some of the platforms have no plant temper in them. Unfortunately, no thin section is currently available for examination.) The present samples show that the mixture used for the hearth platform contains more sand in comparison to floor or daub material, but has less clay content in comparison to daub. Both daub and hearth platform material contain more clay as opposed to floor material, but while in the case of the daub longer and bigger plant matter fragments are added, in the case of the hearth platforms it seems that only finer plant parts were mixed in. Both samples show intensive plant tempering, which is seen in the pseudomorphic voids within the fabric, but also in the form of phytoliths. Both leaf (Fig. 2Ua) and stem (Fig. 2Va) cross sections could be identified in the fabric. The voids contain shrunk and charred organic matter of the plant. Phytoliths are usually not visible within these structures. Based on the size of the stem and leaf cross sections it can be put forward that these voids derive from grass species with thin stems and most probably do not represent old-world cereals. This is underlined by the observed phytolith morphotypes as no Elongate dendritic was found in the samples (ESM#2). It seems that the hearth material was tempered with grass of non-cereal origin. In other words, while the general tempering action at Százhalombatta is linked to cereal by-products, it seems that grass species from the surrounding natural environment could have been used for these platforms, so instead of chaff and straw, hay was utilised in this case. As noted above, on-site observation registered a different type of platform matter, where no indication of plant addition is detectable, showing various mixes for different firing instruments. In respect of the hearth building techniques there is evidence that both the wall and the platform of the hearth were renewed from time to time in some of the cases at Százhalombatta (Kovács et al. 2020).

Borsodivánka-Marhajárás-Nagyhalom

Floors

The most common floor on the tell is a compacted beige to greyish coloured floor composed of a poorly sorted sandy loam. Three floors from House A (samples S1080, S1101 and S1094; for the stratigraphy see P. Fischl et al. 2022a, Table 3) have been analysed micromorphologically (Table 1 and ESM#1, Fig. 3H-J). Although a variation in colour was observed during the excavation, on the microscopic level no distinct difference in the composition of sediment could be found. For the preparation of these floors the use of local fluvial sediments is very likely. The soil beneath the tell has a similar composition as the construction material, but it has been decalcified and leached. This indicates that for the construction material unweathered parent material must have been chosen. The occurrence of rock fragments of different origin such as siltstone and pumice together with marine indicators (lithoclast of marine limestone with foraminifera together with silicified crinoids (Fig. 3X) and silicified disc corals) point to a mixing of sediments of different age and origin, typical for the shore sediments of the Pannonian Lake (Karátson 2000; Seghedi et al. 2004; Lexa et al. 2010). Only a few accidental inclusions, phytoliths (burnt and unburnt), charcoal, micro charred plant material, ash, burnt daub, and bones were observed (see also ESM#1). The sporadic phytolith occurrence is limited to articulated single specimens, no evidence of plant deposition and in situ tissue elements were observed. The only exception is sample S1094 where a few silica skeletons with only a few cells could be observed; these might represent cereals, but the visibility of the remains was not good enough for the secure morphological identification (cf. Elongate dendritic). In contrast to well-prepared floors with plastering or temper and evidence of renovation, we classify these in the ‘earthen floors’ category.

A very special floor type (sample S36, sequence of different floor layers of House J, the oldest Füzesabony House; Table 1, ESM#1, Fig. 3G) occurs at the beginning of the first floor sequence. The sequence is composed of several formal plant tempered floors separated by domestic refuse layers. For its construction a platform was built sealing a dark brownish humic layer modified by anthropogenic activity (P. Fischl and Kienlin 2024). The prepared, beige-coloured floors are composed of clayey silt tempered with chaff. It is distinctly finer than the coarser sandy loam floors, this is why a different sediment source is assumed. The sediments contain some calcium carbonate, rounded and partly rounded silt stone/quartz as well as microfossils (crinoids), which points to a fluvial origin. The deposit of the sediments is unclear; however, a local origin is very likely. The voids are mostly derived from pseudomorphs of the plant temper. There are many tubular-shaped voids (Karkanas 2007) but also chambers are present; many articulated phytoliths (many Elongate dendritics) are left, indicating chaff as temper. Only a few accidental components are included (some charcoal/micro plant coal).

Daub wall

The burnt daub (samples S35-5 and Bor8_S35, house G in the mid-section; Table 1, ESM#1, Fig. 3Q, R) from Borsodivánka is part of a typical wattle and daub construction (Mancini 2024). In Borsodivánka the same local source has been used as for the floors, coarse sandy loam of fluvial origin (see above). For the walls they added chaff and straw (cereal by-products of a non-wetland habitat) as well as calcium carbonate to make it more stable. The source of the calcium carbonate is not yet known. Articulated phytoliths (many Elongate dendritics) occur frequently in the pseudomorphic voids (Fig. 3Y). Most of the daub fragments are reddened, vitrification has only been observed in areas of higher concentration of organic components probably causing reductive conditions. But mostly phytoliths do not show heat alteration, some are coloured. Wall plasters have been observed on daub fragments at Borsodivánka, but not yet investigated in detail.

Wall decoration

The decorated wall (sample S35-3, burnt House G; Table 1, ESM#1, Fig. 3S) is a very special architectural detail in Borsodivánka. As known from the archaeological excavation, the decorations formed spirals and it is assumed that the decorated wall belonged to the inner part of a house (P. Fischl et al. 2022b). It consists of clayey-silty material lacking the typical ingredients of the fluvial sediments, only one fragment of volcanic material and some sand-sized quartz and siltstone has been evidenced. The fine material has been mixed with calcium carbonate as well as micritic calcite that occurs in the groundmass. Furthermore, chaff has been added as a temper (Fig. 3Y). Here, neither the source of calcium carbonate nor that of the sediment is known. The material is horizontally aligned, showing a smoothing effect comparable to pottery sherds. In the horizontal orientated pseudomorphic voids articulated phytoliths are present from which Elongate dendritics represent cereal chaff. Due to the high calcium carbonate content some phytoliths have been partly dissolved, in some cases secondary needle shaped calcium carbonate has been formed around them (Fig. 3Sa). On the surface of the decoration a calcium carbonate lining is present, visible in the thin section.

Fire installation

The reddened platform of the fire installation (sample S1104 of House A (for stratigraphy see P. Fischl et al. 2022b Tab. 3), Table 1, ESM#1, Fig. 3W) was built on a thick layer of the earthen floor material, showing a slight brown-reddish colouration due to heat impact. The platform was constructed from clayey silt; the material is distinctly finer than that of the daub and common floors. It is finely sorted in contrast to the other building materials and most comparable with plant tempered floors. The microstructure is very compact, only few pseudomorphic voids occur. The texture is dominated by silt with few sand-sized quartz and the fine groundmass is composed of fine silt and clay responsible for the granostriated b-fabric. The chosen sediment is of fluvial origin containing a lot of heat modified silicified crinoids and disc corals. They appear blackened and vitrified with many internal vesicles. Only some phytoliths and POM were found, this is why an intentional tempering cannot be assumed.

Kakucs–Turján

Floors

Based on the analysis of the Kakucs samples (samples T2 (b), T10 (c), T4 (b), Table 1, ESM#1, Fig. 4K-M) at least two main types of floors can be differentiated at the site.

One of them is a sandy loam type (Fig. 4M) and the other type is a highly calcareous one, in which silt size fragments dominate with some sand (Fig. 4K, L). The compact, very fine matrixed, highly calcareous loam type of floor (sample T4 (b)) showed no addition of plant tempering and hardly any anthropogenic inclusions were detected in it. The other type showed no signs of plant tempering either. Excluding a tiny bone fragment, no other anthropogenic inclusions were detected in the matrix. Out of the two calcareous samples (T2 (b), T10 (c)), one exhibited a very distinctive white colour on site (sample T10 (c), Fig. 4L) as it was made of very fine calcareous matter. The crack system (platy structure) typical of trampled surfaces was clearly visible in this case as well. A number of bioclasts (molluscs) were also observed. No plant tempering was detected in this case. It seems that clean building matter was used for construction. Some evidence of floor renewal was also observed in one of the analysed samples, where as overlying of the loam type a calcareous one was prepared. No indication of activity residues was detected between the two floor horizons, just as in the case of some of the Százhalombatta contexts.

Daub walls

Similarly to the other two sites, the daub matter (sample T3 (b), Table 1, ESM#1, Fig. 4T) is part of a wattle and daub wall construction. The matrix of the daub is very fine and calcareous. The loose nature of the material is due to the added plant temper that is visible in the form of pseudomorphic plant voids. Nevertheless, no phytoliths could be observed in the voids (ESM#2). Occasional tiny fragments of charcoal appeared in the matrix. No other anthropogenic inclusions were detected during the analysis. Wall-plasterings were also observed, just as in the case of the Százhalombatta samples.

Unfortunately, no hearth context (in thin section) is available for investigation from this site.

Discussion: Construction materials and building techniques made visible through a chaîne opératoire analysis

Floors

Sourcing raw materials

One of our overall results is that unweathered local sediment (loess or fluvial sediments) and not soils or so called ‘mud’ have been chosen as raw material for the floors (Fig. 5). In most of the documented cases – with the exception of floors made from recycled materials at Százhalombatta (see below) – unweathered sediments were freshly sourced for the construction of floors. The freshly sourced floor building materials of the investigated sites varied greatly in grain size composition and mineral content, but all of them display the characteristics of their immediate environment. Using sediments from the vicinity of the settlement site for construction is a common practice shown by many other studies (e.g. Rosen 1986; Goldberg 1979; Macphail 2007; Friesem et al. 2017). At Százhalombatta, loess was used for the silt floors; and according to the geological map, loess (eQp3l) is available in close vicinity to the site (Fig. 1). At Kakucs-Turján, the most common floor type is made of loam, very likely derived from sandy loess (Fig. 1). Also here, the site lies within an area characterised by fluvio-aeolian sandy loess and an extended marshy area (called ‘turján’ in Hungarian – hence the site name) is also nearby. In both cases the raw material was specifically chosen, especially in light of the fact that the Danube, the Danube-Tisza-interfluve and its fluvial sediments were just as accessible. The occurrence of loess and its derivatives is widely spread in the Carpathian Basin (Lehmkuhl et al. 2020) and it was also used for flooring at other Bronze Age sites (Parkinson et al. 2018; Lie et al. 2024).

Fig. 5
figure 5

Comparison and summary of the main characteristics of the investigated building materials

Full size image

Meanwhile at the wetland site in Borsodivánka different fluvial sediments were used, which always contain different rock fragments as well as glassy minerals related to former volcanic activity (Fig. 5). Miocene rhyolite volcanic activity is present in the near vicinity of Borsodivánka at the Bükkalja region (Karátson 2000; Seghedi et al. 2004; Lexa et al. 2010). Other ingredients are calcitic lithoclasts with foraminifera and silicified microfossils (probably crinoids and disc corals) deriving from marine origin of the former Pannonian Lake. This composition is typical for redeposited Pannonian sediments (Karátson 2000; Head and Gibbard 2015). Due to the overall occurrence of these natural components in the floor material the source of the raw material can be easily recognised and the components can be used as markers in thin sections. The use of natural wetland materials is very well known from other sites in different climate zones (e.g. Rosen 1986; Macphail 2007; Viklund et al. 2013; García-Suárez et al. 2021).

As a common practice at all three sites, calcareous raw material was purposefully chosen for flooring. Calcite acts like a binder, hardening after the floor had been constructed with wet material, and was therefore a good material for floor construction (Thiemeyer and Fritzsch 2011; Uzdurum et al. 2023). Calcite seems to be naturally part of the sediments (loess, fluvial sediments) used at Százhalombatta, Kakucs-Turján and Borsodivánka. In Kakucs-Turján, there is even a floor type that is made of highly calcareous material. Obtaining calcitic sediments involved removing about 1 m of weathered soil and creating pits or using a natural exposure. At Borsodivánka, they did not use the decalcified and leached soil available adjacent to the site, but its unweathered fluvial parent material. In Százhalombatta, sourcing materials for floors also meant a choice by the builders to use loess instead of the nearby fluvial sediments. This selection of sediments illustrates the builders’ decision to use available sediments from the surrounding area with different backgrounds but similar construction characteristics. So far unique in our sample of sites, Százhalombatta features a floor type made from recycled accumulated anthropogenic sediment from the settlement itself. It appears that these sediments were removed somewhere on the site and then during processing the visible fragments of the material culture in them were removed. This can be deduced because anthropogenic inclusions, such as bone, burnt bone or pottery fragments, only occur sporadically and they are very small in size. The presence of such components in the floor, together with the addition of ash and charcoal fragments, indicates that not natural soils/sediments were transported to the site in this case, but the deliberately modified and processed anthropogenic sediment was used for flooring. It needs to be noted here that the evenly distributed ash and micro-charcoal that is observable within the earthen floors conspicuously distinguishes them from the general anthropogenic sediments’ characteristics. Although micro-charcoal is rather abundant all over the site, in the general matrix it appears without detectable pattern mixed with other types of materials (e.g. bone, ceramic sherd, daub etc.) or in very distinct horizons, depending on the ongoing activities.

Floor tempering and preparation techniques

The three sites vary as to the existence of temper in floor building materials (Fig. 5). Kakucs floors were not tempered; at Százhalombatta, only floors made from recycled materials (‘earthen’ floors) were tempered, but not those derived from fresh sediment sources (silt floors). At Borsodivánka the earliest floors in the excavated sequence were tempered while later floors were not. In both cases of tempered floors – Borsodivánka and Százhalombatta – temper appears to have been added to reach the floor consistency desired by the builders, but in both cases, there are interesting social choices attached to the tempering. To create the ‘earthen’ floors at Százhalombatta, ash and charcoal were deliberately added to the modified (cleaned) recycled site matrix. Creating these floors included digging up and/or collecting accumulated sediments on site, removing larger cultural inclusions (‘cleaning’ the sediment) and then adding ash as temper. There were numerous cases observed during the excavation when large amounts of ash, visibly from different sources, had been collected in one area or deep depression. It is interesting to think about possible reasons for the creation of the earthen floors at Százhalombatta. Clearly, other types of flooring that used freshly extracted sediments with either no temper or chaff temper (see below) worked well in this region − including the Százhalombatta silt floors, which due to their calcite content are hard and durable. It is possible that the addition of ash to the material of the earthen floors had the purpose of raising the calcite content in these floors’ mixture to approximately re-create the composition of the calcitic silt floors. Wood ash is highly calcareous, i.e. alkaline (Karkanas et al. 2000). This property makes ash a stabiliser in building materials as demonstrated by modern experiments (Cultrone 2022; Leitão et al. 2017). If the hypothesis that the addition of ash to earthen floors was meant to mimic some properties of silt floors is correct, then this makes the decision by Százhalombatta people to build earthen instead of silt floors even more baffling. However, there must have been perceived advantages to using ash-tempered recycled anthropogenic sediments for flooring. It is possible that using materials already on site was sometimes more achievable than obtaining fresh sediments that needed to be extracted outside the settlement. It is also possible that people perceived an advantage in adding ash to a floor mixture. Alkaline ash can be used as a disinfectant (Hakbijl 2002) so ash temper in floors might have served hygienic purposes as well. Furthermore, ash has a high capacity of water absorption, which makes it useful to reduce dampness, which is considered favourable within the internal living spaces (Milek 2006). Ash can also be used for odour elimination therefore it was commonly spread on earth floors (Milek 2012). In this context, it is also relevant to mention that ash was sometimes spread on Százhalombatta floors (Kovacs et al. 2020), possibly also for hygienic reasons. Taken together the ash on floors and in floors could show a local preference for using ash for floor building and house hygiene.

Another interesting feature of floor preparation at Százhalombatta is the alternate use of recycled (‘earthen’) and silt floors in houses, but it is not currently possible to interpret this particular choice made by the builders: The earthen floors made from ‘recycled’ (i.e. modified) anthropogenic sediment in some of the cases were prepared first and at a later phase of the house they were renewed with silty material. Both types were recorded either as single horizons (i.e. unrenewed) or as part of sequences of floors. It is interesting to note that so far within the investigated samples, all earthen floors were at some point covered by a silty floor, while not all silty floors had earthen floor layers beneath them. Indicated by activity remains (refer to Kovács et al. 2020 for further details on this topic), earthen floors are not foundations of silt floors, but ‘living floors’, similar to the silt ones. Series of earthen floors were only found in one house context so far (House id 3147), but the ongoing sampling and analysis will show that this was a unique technique. It is without question that the preparation of earthen floors was part of the practice at Százhalombatta, but more data is needed for a better understanding of the social or individual preference behind this practice.

The second variety of tempered floors in this study is the chaff tempered clayey silt floor sequence in Borsodivánka. This first phase of floor sequences is of a completely different composition compared to later floors, having been made with finer sediment and tempered with chaff. It seems that vegetal temper needed to be added as a stabiliser to the fine fluvial sediment to be able to construct the floor. This is a common practice for fine clayey-silty material to keep it stabilised and has been also evidenced in other regions where sediments from wetlands were used (Macphail et al. 2017, García-Suárez et al. 2018). However, the chaff temper at Borsodivánka is not from the immediate surrounding of the tell itself, as it was a wetland, but it proves threshing on site and cereal cultivation in the wider area (Fig. 1). On a macroscopic level, the plant tempering of the prepared fine floor sequence of Borsodivánka cannot be recognized. But microscopically pseudomorphic voids with phytoliths testify to its former chaff temper (by-products of crop processing such as cereal husks, which are glumes, palea and lemma from the botanical point of view, so parts of the cereal inflorescence) (Macphail and Goldberg 2018; Shillito et al. 2011). This is different to the silty floors of loess origin, where tempering seems not to be necessary as evidenced in Százhalombatta, Kakucs-Turján and Vésztő-Mágor (Parkinson et al. 2018). In general, descriptions of plant tempered floors are quite rare in Hungary. This might be due to the utilisation of loess as raw material or because of the missing microscopic record. Further analyses have to be done on this subject to get a clearer picture of the occurrence of plant tempered floors on tells in the Carpathian Basin.

Later floors within the Borsodivánka sequence were rather thick earthen floors made from the local loamy fluvial sediment, very different from early floors. This leads us to question the origin of the technique of the first floors and why it was abandoned. According to the ongoing archaeological analyses societal change is not very likely (P. Fischl et al. 2022b). So, the more elaborate and time-consuming technique used for the first floor could have been replaced by a simpler one that fulfilled the same desired requirements. It is also possible that the earlier, finer floors had less durability. The abandonment of an inherited technique to something more functional with a good availability of the source material could be a form of adaptation to local conditions. Moreover, looking at this first phase of the settlement might also reveal details about the architectural knowledge the settlers brought to the site. The preparation technique can be regarded as a technique of the community, which had been developed before bringing their building traditions to the wetland, maybe from other parts of the Borsod plain.

Maintaining floors

Different types of floor maintenance were documented at our sites. A common type of maintenance is the periodical renewal of floors (construction of a new floor layer). Here, the topic of maintenance partially overlaps with the construction. In addition to the renewal of entire floors, we could also observe local renovation of only parts of floors at Százhalombatta (for further discussion see Kovács and Vicze 2022). In many cases silt floors at Százhalombatta are renewed indicating long-term usage of houses and probably activities that result in the contamination and/or wearing of the floors. Silt floors were frequently scraped prior to the new flooring (for more details see Kovács and Vicze 2022), a practice that was not observed at the other two sites, although floor renewal was practised at all three sites. At Kakucs-Turján, some evidence of floor renewal was also observed in one of the analysed samples, where overlying a loam floor, a calcareous floor was prepared. No indication of activity residues were detected between the two floor horizons, similar to some of the Százhalombatta contexts. Apart from floor renewal, there are other aspects to be explored in this respect. The first one is the use of mats for floor coverage that could be identified at Százhalombatta in two cases so far and also at Borsodivánka above one loamy floor. It could be argued that floor coverage is not strictly part of the maintenance procedure, but the lifespan of a floor can be extended by mat coverage similar to renovation, so in this respect such coverage is a maintenance practice. Two types of evidence can be put forward in relation to floor coverage. The first and more straightforward one is mat imprints observed both at Százhalombatta and the upper part of the tell at Borsodivánka. In the Borsodivánka example, layers of phytoliths deriving from reed were present with fine dust-like accumulations underneath (Fig. 6). We interpret this as fine dirt that infiltrated through the mat coverage, leaving a very characteristic powdery accumulation (Röpke et al. 2018). Based on this, areas of different utilisation can be identified.

Fig. 6
figure 6

Imprint of the former matting at Százhalombatta-Földvár site (House id 127) (12.5x, PPL, left) and remnants of a former reed mat evidenced by a phytolith layer with fine dirt underneath at Borsodivánka (S107) (PPL, right)

Full size image

Daub

Sourcing raw materials

The sediments used for daub building at the three sites differ in their grain sizes and also differ in whether or not they are congruent with the sediments used to build floors at the same site (Fig. 5). The Százhalombatta daub was made from clayey loam sediments different to the floor sources; although the earthen floors are also loamy, but with a lower amount of clay compared to the raw daub material. At Kakucs-Turján, the matrix of the daub is loamy, very fine and calcareous – different from the floor material. In Borsodivánka, for daub the same local source was used as for the floors, coarse sandy loam of fluvial origin. This coarse material necessitated lots of temper to become suitable for wall building, but it is likely that finer sediments were rarer in the surrounds of the site. This might have prompted Borsodivánka people to build walls with coarse material and to reserve the finer sediments for other building components. Interestingly, the material used to form wall decorations is very different from the daub, being clayey and lacking the typical ingredients of the fluvial sediments. This material was therefore derived from a different source than the main daub material, but the source is unknown. The builders apparently went through the effort of obtaining this different material to create a much finer matrix that could be easily shaped into thin spiral decorations.

Daub tempering and preparation techniques

The Százhalombatta and the Borsodivánka daubs were tempered with chaff evidenced micromorphologically by pseudomorphorphic voids with phytoliths, which is known as stabiliser (García-Suárez et al. 2021). This is a common occurrence in prehistoric earth building during which builders presumably used a refuse product (chaff from cereal processing) present in large quantities in the settlement, at least around harvest time. At Kakucs-Turján, the one analysed piece of daub also exhibited plant tempering.

At Borsodivánka, beside chaff, calcium carbonate was added as temper to the daub. Probably due to its coarser texture, builders chose to add calcium carbonate as mineral temper to act as binder as evidenced at some other sites (e.g. Thiemeyer and Fritzsch 2011; Uzdurum et al 2023). The same applies to the clayey material used for wall decoration, which was also tempered with chaff and calcium carbonate. We had expected to see smaller plant pieces used as tempering for the wall decoration material, which needed to be a lot more plastic than the daub, but this does not seem to be the case. Similar to the daub, the material for the wall decoration was tempered with calcium carbonate, but with distinctly higher concentrations, probably for stabilising purposes of the more fragile elements.

Wattle-and-daub building can be observed in Hungary since the Neolithic. Neolithic settlements are located along rivers, and this might be why they often used reed both as structural element (wattle) in walls and as temper for the building material (Carneiro and Mateiciucová 2007). There is micromorphological evidence at Borsodivánka that under the Bronze Age tell burnt daub with reed tempering occurs which was typically used during the Neolithic period (e.g. Carneiro and Mateiciucová 2007). In an interesting contrast that highlights culturally different approaches in the same environment, reed has not been observed in walls in the Bronze Age settlements under investigation now, although Százhalombatta is close to the river and Borsodivánka and Kakucs are even within wetlands.

As a last stage in wall construction at Százhalombatta and Kakucs, daub walls were covered with a thin yellow plaster. The Százhalombatta wall plaster is − different from the daub sediments − silty in nature, similar to floors, and was not plant-tempered. The plaster must therefore have been created in a separate process of raw material procurement and material mixing. The choice of fine silty material without temper, applied in thin layers, might have been prompted by the ease of applying such material to the wall and possibly the desire to create smooth looking walls, and at the same time it made the plaster more resistant to weather conditions e.g. rain. Although we have some evidence from Kakucs that wall re-plastering was also practised, no systematic analysis has been done yet. Wall plasters have been observed on daub fragments at Borsodivánka as well, but not yet investigated in detail. A final stage in some Borsodivánka buildings was the creation of wall decorations that are so far unique within our sample of sites. It can be seen as an interior design element that extends beyond the pure function of architectural construction. In the Carpathian Basin only a few other examples are known like an Early Bronze Age house at Tiszaug-Kéménytető site (Hungary), which belongs to the Nagyrév culture (Csányi and Tárnoki 1992, Abb. 76), ‘Dealul Vida’ in Sălacea (Romania) (Ordentlich et al. 2014) which is also an OFCC settlement, and Feudvar near Mošorin (Serbia) (Hänsel and Medović 1991) belonging to the Early and Middle Bronze Age Vatin Culture.

Maintaining walls

Evidence for wall plastering was found at all three investigated sites, but more research has been done in this respect at Százhalombatta-Földvár. Here the repeated plastering of the house walls − both the interior and the outer wall surface − seems to be a common practice. These re-plastering layers are between 0.5–1.5 mm thick and must therefore have been created by applying a very watery silty mixture (Kovacs et al. 2020). This replastering was probably done to fill cracks to prevent moisture from entering and damaging the walls. It could also be assumed, but cannot be proved by our data, that a sense of beauty and/or orderliness also played a role besides the practical needs. In one case (House id 3497) seven replasterings were observed (Kovács and Vicze 2022), indicating a regular and careful maintenance activity.

Fire installations: Sourcing raw materials, tempering, construction and maintenance

Fire installation contexts were only available from two of the sites, Százhalombatta and Borsodivánka. They clearly differ in the used raw material and in the preparation technique (Fig. 5). At Százhalombatta, macroscopic observations on site showed the existence of different types of platform building materials related to various fireplace types (Vicze and Sørensen 2023), but only two samples of House 3497, id 3321 (belonging to the 3439 oven) and id 3322 (of the 3322 rimmed hearth) are presented here. In the analysed platforms, the sediment is sandier than that used for floors or daub, and less clayey than the daub. Similar to daub, the fire installation platform material was plant-tempered, but plant fragments seem to be smaller compared to daub. Reasons for the use of the sandier sediment and finer plant material could be that a horizontal fire installation platform does not need to exhibit the same stability as a vertical wall, hence less clayey material and less long vegetal temper are sufficient. Material with a higher sand content has different thermal properties compared to more clayey material, it is less affected by heat which makes it more durable (Röpke and Dietl 2017). In contrast to other ‘dauby’ materials, it seems that some fire installation platforms were tempered with grass of non-cereal origin, presumably dried hay. In other words, while the general tempering action at Százhalombatta is linked to cereal by-products, this study shows that in the case of fire installation platforms grass species from the surrounding natural environment could have also been used. This decision is interesting given that Százhalombatta must also have had chaff available seasonally as a by-product of cereal processing. To use grass as temper, they had to go through the additional work of cutting grass − which then also would not have been available as animal fodder (fresh or as hay) if it was used in building. Possible reasons for this could include the timing of building processes − chaff would have been readily available during harvest time, but may not be during the rest of the year, while grass was available in the surrounds of the settlement during most of the year. Also, there could have been a limited availability of chaff (straw), which was also used for roof thatching (Vicze 2013a) and possibly also for other purposes in the settlement. On-site observation registered a different type of platform matter, where no indication of plant addition is detectable. In some Százhalombatta hearths, both wall and the platform had been renewed from time to time.

At Borsodivánka, the material of the fire installation platforms is finer than the wall daub and the typical floor, consisting of clayey silt of fluvial origin, and most comparable with the fine raw material of the plant tempered floor. Another difference compared to wall and floor materials is that it was not tempered. In the reddened material, silicified crinoids and disc corals are distinctly heat modified. This is in contrast to the well preserved phytoliths. It seems that these silicified micro-fossils are less heat tolerant than the phytoliths. Usually, phytoliths vitrify above 600°C–800°C (Macphail and Goldberg 2018; Fritzsch et al. 2019) depending on the plant part, but embedded in sediment they are more heat resistant due to isolation of the minerogenic material (Canti and Linford 2014; Röpke et al. 2024a). To conclude, the preparation techniques of fire installations vary, and different sediments could be used to fulfil the purposes of a fire installation. All of them show the typically intense reddening of the sediment due to the recurring firing process.

Conclusions

In this paper ‘earthen’ construction materials from floors, walls and fire installations from three Middle Bronze Age tell settlements in Hungary have been investigated. The combined methods of archaeology and microscopic analyses (archaeological soil micromorphology and phytolith analysis within thin sections) offered the opportunity to examine these materials as social products and reconstruct their chaîne opératoire; and compare across sites as well as between phases of different sites. There are three main outcomes from our investigation in this paper. The first is a need for clearer terminology and actual investigation of ‘earthen’ building materials in our research region, during the Hungarian Bronze Age. As ‘earthen construction material’ is an umbrella term including different types of soils and sediments of varying origin and genesis (Friesem et al. 2017), our analyses demonstrate how important it is to determine the origin of the raw material. Depending on the source of the material and at which depth it is extracted, completely different properties can be expected. From a pedological and geological perspective this often-used term is even misleading because ‘earth’ or ‘earthen’ is only associated with the topsoil or humic horizon (e.g. Amelung et al. 2018; IUSS Working Group WRB 2022). However, in many cases muddy sediments or other fluvial sediments were used for construction such as in the Middle East (e.g. Goldberg 1979; Rosen 1986; Shillito and Matthews 2013; García-Suárez et al. 2021). In our case there seems to be a preference across all three sites to avoid weathered material (soil), and choose calcareous parent sediment of loess or fluvial origin from the vicinity of the settlement, although the builders might have had to remove the soil before reaching the parent material (sediment).

Our second conclusion is that even small variations in the environmental and social setting lead to differences in building materials and techniques between settlements. It is therefore important to investigate building materials at every site instead of assuming similarities between roughly contemporary sites in approximately the same environment. Our data sample set currently contains more information on floors than walls or fire installations, and floor data demonstrates large variety in the chaîne opératoire of floors in different parts of one site as well as between sites. They occur without temper, with chaff as temper or tempered with ash and charcoal. Not only the construction of the floors, but their renewal and maintenance are also diverse, indicating different social practices. At Százhalombatta it became evident that different raw material sources (the yellow silt (loess) and the greyish anthropogenic sediment of the settlement) and different preparation techniques were used, sometimes even within the lifespan of a single house. At Borsodivánka the different floor types (the coarse beige to greyish, sandy-loam and the fine beige, clayey silt) with different preparation techniques (use of temper) were used for different phases. At Kakucs some of the floor materials are quite unique with their high calcareous content and in one case with its distinct white colour. At all sites locally available sources were used and their processing was adjusted accordingly. Very different materials were utilised from fluvial, aeolian or anthropogenic sediments, and from rather heterogeneous coarse loam to fine sediments consisting of silt or clayey silt. But the presence of calcareous materials in floors is persistent at all sites and could be interpreted as a shared practice and sphere of knowledge. Furthermore, preparation of hearths did not follow a consistent line as seen in the presented data. Even within one site and one household differences have been observed macroscopically, which points to individual production by and for households and for different functions. However, more samples need to be analysed microscopically. Available samples from Borsodivánka and Százhalombatta show that different sediments were used compared to daub, and that some hearth materials were tempered and some were not. This leads us to the idea that the creation of hearths might allow a greater design creativity causing individual constructions and should be compared in greater detail.

Our third conclusion is that even within the same micro-landscape and with similar or identical environmental conditions, different groups of people can make different decisions concerning construction materials. This highlights that construction materials are social products that embody cultural choices and knowledge. One example is the change at Borsodivánka from floors made from fine sediment with plant temper to floors made from loamy sediment without temper. At Százhalombatta and Kakucs, there was a lack of build-up between the subsequent floor horizons in many of the renewal phases (even the scraping of floor surfaces could be detected at Százhalombatta (Kovács and Vicze 2022)). This might hint at some differences on a social group level compared to Borsodivánka, but more data is necessary to substantiate this tentative interpretation. It is also interesting to note that at the Vatya culture sites (Százhalombatta and Kakucs) rather frequent maintenance of house walls with plastering could be observed, while at Borsodivánka (an OFCC site) an elaborate spiral decoration with excessive plant tempering was added to the wall surface. It seems that while the construction material of walls was very similar all over the Carpathian Basin – chaff-tempered daub –, the outer appearance of walls seems to display cultural group specific characteristics. For example, the subtle and restrained decorative program of the pottery tradition of the Vatya cultural group seems to be reflected on the smooth, clean and even wall surfaces; whereas elaborate spirals for the Füzesabony group (Borsodivánka), or the angular geometric designs of the Nagyrév cultural group (see above mentioned Tiszaug) that are integral part of their ceramic decorative motifs became also visible in their wall decorations. The concurrent use at Százhalombatta of floors made from fresh and floors made from somewhat laboriously recycled materials showcases another aspect of construction materials as social products: the reasons why builders made certain choices are not always obvious to a modern observer. In this particular case, the floors made from recycled sediments and ash apparently had some desirable characteristics that we can only speculate about at this time. Another example is the choice to use grass as temper instead of chaff in some Százhalombatta fire installations.

In summary, this paper highlights the value and potential of combined micromorphological and phytolith investigations of Bronze Age construction materials in a humid continental climate. This research on construction materials shows that – as in many parts of the world – raw materials were available in the vicinity of each site, but also which of the many possible materials were selected, combined and transformed by people. It therefore delivers information both on the natural environment and about knowledge, and choices available to the builders within their social and cultural setting.