The use of lithic raw materials at the Early Mesolithic open‐air site Feuersteinacker (Vogelsbergkreis, Germany)

Abstract The open‐air site Feuersteinacker near Stumpertenrod has yielded one of the largest lithic assemblages in Central Germany. It repeatedly served as a workshop for the production of stone tools during an early phase of the Mesolithic. The range of lithic raw materials is extremely diverse, but until today, there is only a limited number of archaeological studies on the occurrence of lithic resources in the area. The following study presents the first in‐depth investigation of the use of different rock types by Mesolithic hunter–fisher–gatherers at the site. Provenance analyses using petrographic methods permit raw materials to be assigned to a specific source and provide new insights into their formation. Furthermore, this study explores the way in which the materials were processed throughout the reduction sequence. A comparison of topographic parameters suggests that the location was situated on an important transit route during prehistoric times. The presented results contribute to a better understanding of mobility patterns and subsistence strategies of Early Mesolithic groups in Central Germany.

; Figure 3). Several rivers-flowing radially in all directions-rise in this mountain range. Due to its topographic features, the site is located on a strategically important junction.
At the same time, it is protected from storms and floods. Finally, the slightly elevated position presumably allowed prehistoric hunter-gatherers to observe and approach their prey without being detected (Krüger & Taute, 1964, p. 21).

| Research history
The uncannily apt name «Feuersteinacker» («flint field» or «flint acre») derives from the fact that the inhabitants of the nearby village collected siliceous rocks in the field to use as strike-a-lights in historical times (Krüger & Taute, 1964, p. 19). After the first antiquarian, collectors had  (Krüger & Taute, 1964, pp. 24-27). The entire excavated sediment was screened.
As a consequence of ploughing, the archaeological features were not in situ anymore and the excavators did not detect any further structures.
However, the site that was typologically dated to the Beuronien A (∼9000 BC) forms an important reference point for Mesolithic research in Germany. According to palynological data and radiocarbon dates from the nearby Lahn Valley, there was an increased human impact during the Preboreal and the early Boreal (Bos & Urz, 2003 archaeological record in the region. In addition to a report summarising the results of the field campaign (including a description of the stratigraphic sequence), the site is mentioned in Taute's habilitation treatise which was aimed at creating a chronological scheme for the Mesolithic in Southwestern Germany and adjacent areas (Taute, 1971, pp. 282-284).
During the following years, the amateur archaeologist Horst Quehl conducted extensive field surveys with the permission of the responsible authorities. Until 2016, he collected more than 7000 lithic artefacts (Fiedler, 2017). Between 2019 and 2020, the assemblage discovered during the excavation, as well as the surface finds were systematically

| Provenance analysis
Provenance analyses of the lithic raw material are of particular interest for the reconstruction of prehistoric territories and mobility patterns (Affolter & Nielsen, 2006;Eriksen, 2002;Högberg & Olausson, 2007;Kind, 2006;Moreau et al., 2015). Especially in case of sites with poor preservation of organic materials, the study of stone tools is fundamental to address the mentioned issues. Lithic sourcing was conducted using a Zeiss Stemi 2000 optical microscope with a magnification of up to ×80.
Each artefact was described petrographically with reflected light and subsequently compared to a reference collection. An important advantage of this method is the fact that the analysis was nondestructive.
The reference collection consists of rock samples that were collected in the framework of field surveys in 2019 and were supplemented with specimens provided by the Hessian State Museum in Kassel. Currently, it comprises 23 different raw materials, including primary sources and secondary deposits, such as river gravels. Each outcrop was documented photographically, and the coordinates were measured with a handheld global positioning system device. Geological maps (Rösing, 1976) and previous work on the issue (Pflug, 1993) served as a starting point for investigations in the field. For a more detailed description of the rocks, selected samples were studied in thin section.
Several scholars working on the site have pointed out the difficulties in distinguishing the lithic raw materials based on their macroscopic appearance (cf., Fiedler, 2017, pp. 3-4;Taute, 1971, p. 19). Therefore, it was crucial to study the artefacts under a microscope (Figure 4). In the case of biogenic sedimentary rocks, the microfacies analysis was applied (Flügel, 1982(Flügel, , 2010. Originally developed for the study of carbonate rocks, it involves a detailed description of the texture and the components of a sample. The occurrence of microfossils and their degree of preservation allow a reconstruction of the depositional setting in which the rock was formed. Thus, it is possible to distinguish between a nonmarine environment, such as a freshwater lake, and different marine zones, for e.g. a coastal milieu, a reef, a continental slope, or a pelagic milieu. As flint and chert occur within carbonate rock formations and the original texture of the parent material is preserved when it is replaced by silica, the criteria for their petrographic description are mostly the same. The mentioned raw material samples were classified according to the work of Dunham (1962). Archaeological applications of microfacies analysis have already produced interesting results concerning the procurement and circulation of lithic raw materials in Switzerland and Southwestern Germany (Affolter, 2002;Hess, 2019;Kaiser, 2013).
The petrographic description of silicified sandstone (Tertiärquarzit) was based on classification schemes developed for the study of clastic sedimentary rocks (Füchtbauer, 1974;Tucker, 1992). In this case, it was possible to distinguish different raw materials based on parameters like grain size, sorting, and roundness. As these aspects are considered to be a function of the transport distance of components, they also contain spatial information.
In addition to the texture of lithic materials, the properties of the cortex were considered. It was possible to distinguish between a primary cortex (i.e., fresh or chalky), a battered surface-which is typical for transport by rivers or moraines-and a cortex that is the result of chemical weathering (Affolter, 2002, p. 19;Kaiser, 2013, p. 81). The colour of the raw materials was determined using the Munsell Color Chart.

| Sample
The archaeological sample considered in the framework of this study consists of surface finds in the collection of the Hessian State Museum in Kassel and objects that are displayed in the permanent exhibition of the museum. It includes 8089 artefacts (≥1 cm). The objects were sorted by raw material and assigned to different artefact classes. For a better F I G U R E 4 Comparison between the macro-and microscopic appearance of two end-scrapers made of different raw materials (magnification: ×40). Siliceous shale (left) and Cretaceous flint (right) [Color figure can be viewed at wileyonlinelibrary.com] understanding of the raw material economy, the number of cortical pieces was determined for each rock type. The weight of the pieces was measured with a precision scale (accuracy: 0.1 g). Although it has been subject to postdepositional processes, the assemblage dates almost exclusively to an early phase of the Mesolithic, corresponding to the Beuronian A in Southwestern Germany (Taute, 1971, p. 282 (Table 1). These consist of microliths, notched pieces, laterally retouched pieces, end-scrapers, burins, truncations, semifinished products of microliths, perforators, multifunctional tools, and denticulates ( Table 2).

| Provenance analysis
The lithic raw material from Feuersteinacker is extraordinarily diverse in terms of its composition. Besides different types of silicified sandstone, chert, and siliceous shales, there is so-called chalcedony. The spectrum of colours ranges from black through red to white and grey. There are even green materials within the assemblage. In the following, the present rock types will be discussed in more detail (Figures 5 and 6).

| Silicified sandstone
Silicified sandstone (Tertiärquarzit) was formed under semiarid conditions during the a, by the silicification of sand layers. The silica derives from Tertiary volcanic rocks and was dissolved by groundwater (Freyberg, 1926). In the study area, the material occurs at several places in the shape of large boulders of up to 1 m, as well as smaller nodules. The coloration of silicified sandstones is mainly dependent on the composition of the original sediment. Iron oxide leads to a red colour, while green pieces contain the mineral olivine or epidote. An important outcrop is situated near Lenderscheid, where a particularly fine-grained and well-sorted variant is found. Typical colours are white (10.0Y 9/2), grey (5.0Y 5/2), orange (5.0Y 7/20, 2.5YR 6/18), and pink (5.0R 6/12, 5.0R 7/8) (Figures 5c-g and 6a). Other sources are located in Hausen, Ziegenhain, and Rörshain.

| Siliceous shale
Local siliceous shale (Kieselschiefer) is the most common lithic raw material in the study area. The term refers to a range of different rocks that naturally occur in gravels of major rivers and streams. Primary outcrops can be found in the Rhenish Massif ( Figure 3) and the Thuringian Highlands. Siliceous shales are formed in a pelagic environment below the carbonate compensation depth (Füchtbauer, 1974). Petrographically, they can be described as Palaeozoic radiolarian cherts. As a consequence of tectonic stress, there are differences concerning the preservation of microfossils. The colour of the material is a product of carbonaceous pigments (Carozzi, 1960) and ranges from black to dark olive green (5.0Y 2/2, 5.0Y 6/4, 5.0GY 3/4; Figures 5a,b and 6b).

| Chalcedony
Another fascinating raw material that is abundant in the study area is chalcedony. It occurs at several locations in Hesse and is associated T A B L E 1 Overview over the assemblage (n = 8089)  (Behn, 1925;Deecke, 1933, p. 5), as well as Homberg near the river Ohm. Chalcedony is usually translucent and has a characteristic lustre that is comparable to flint.
At the same time, it was possible to demonstrate that a raw material, which is commonly referred to as "Basalthornstein" in the Germanspeaking literature (Fiedler, 2017;Krüger & Taute, 1964;Pflug, 1993), belongs in fact to the same group. As the volcanic rocks in Hesse are the product of multiple eruptive events, chalcedony sometimes occurs as alternating layers within basalt quarries. Elster glaciation north of 51°N latitude (Floss, 1994, p. 103). The nearest occurrence of Cretaceous flint is in Lower Saxony, where the raw material played an important role during the Mesolithic (Grote, 1993). Other potential sources that were evidentially used by Mesolithic hunter-gatherers are alluvial terraces of the rivers Rhine and Maas (Floss, 1994, p. 98;Gelhausen et al., 2003). Due to their characteristic shape, the pebbles are also called Maaseier. The latter often display concentric zoning (so-called Liesegang rings). They can be classified as wackestones, including triaxon sponge spicules and planktonic foraminifera. Cretaceous flint is usually strongly silicified and translucent. Colours include dark reddish brown (5.0YR 3/4), grey (2.5Y 8/2), and yellow (10.0YR 8/11). Some pieces show a white patination which is the result of fire.

| Jurassic chert
In the study area, Jurassic chert occurs in alluvial terraces of the river Main. The nearest primary sources of the material are limestone formations of the Swabian and the Franconian Jura. Due to the transport by alluvial processes, the nodules generally show a battered cortex. The rocks can be described as mud-and wackestones pointing to a neritic environment. Microfossils include remains of sponges and corals, bivalves, gastropods, foraminifera, ostracods, bryozoans, as well as crinoids.
Typical colours are grey (10.0YR 7/2) and white (5.0Y 9/2). As the geo- 3.1.7 | Other raw materials from river gravels (radiolarian chert, quartz, and chert from Flysch formations) Several lithic raw materials with primary sources in alpine regions are present in river gravels of the Rhine to the south of the site. They include radiolarian chert of Mesozoic age, quartz, as well as chert from Flysch formations.

| Jasper (Kellerwald region)
Jasper from the Kellerwald region (Kellerwald Jaspis/Eisenkiesel) occurs as secondary deposits in gravels of the river Lahn to the north of Stumpertenrod. It is of Palaeozoic (Devonian/Carboniferous) origin and owes its red and green colour to iron oxide and manganese, respectively. The material is often fissured and contains spherical inclusions. Petrographic analysis suggests that the genesis of the rock is linked to hydrothermal vents (Schneiderhöhn, 1941). It seems to have played only a minor role as a raw material for the production of tools during prehistoric times (see also Pflug, 1993, p. 80).   Floss, 2002). Several watercourses originate in the

| Raw material economy
Vogelsberg mountains and what is nowadays a rather remote, rural area seems to have been a major transportation hub during the Mesolithic, and a persistent place in the early Holocene landscape (cf., Barton et al., 1995, p. 81). In this context, it is also interesting to mention archaeobotanical evidence for intentional woodland clearance by fire setting (Bos & Urz, 2003, pp. 31-33).
Large numbers of cores, preparation flakes, microburins, and semifinished products suggest that the site was repeatedly used as a workshop for the serial production of stone tools. Furthermore, several elongated sandstone pebbles that served as retouching tools were discovered. The lithic assemblage analysed in the framework of this study has a weight of more than 16 kg (!). This is a substantial quantity of material, considering that Mesolithic technology was based on the production of microliths. It should also be added that the sample only reflects a small part of the entire lithics discarded at the site. Laminar blanks are more regular compared to contemporary sites in adjacent areas and appear to be standardised. The fact that various raw materials were processed in a different way implies a good understanding of their physical properties and a certain degree of specialisation. Furthermore, the presence of Jurassic chert and Cretaceous flint argue for a far-reaching contact network. In this context, it is possible to suggest seasonal gatherings of otherwise dispersed groups (cf., Baales, 2001). A technological system involving composite tools offers new possibilities concerning the production and exchange of material culture (Finlay, 2003). As in the case of other Mesolithic sites, it is possible that the colour of siliceous rocks was an additional criterion for their selection. While Jurassic chert in the Early Mesolithic record of Southwestern Germany was intentionally heat treated (Eriksen, 2006;Hess, 2019), lithic raw material that is available in the wider region naturally occurs in a variety of different colours. Together with pigments, rocks were among the most colourful materials in prehistoric times, and ethnographic analogies suggest that in addition to functional aspects, they might have had a symbolic meaning (Hess, 2019;Taçon, 2008).
Finally, there are a number of landscape archaeological implications derived from the observations detailed above. First, and in line with previous studies in neighbouring regions (e.g. Eriksen, 2002;Floss, 2002;Jochim, 1998), the results argue for new patterns of land use emerging with the onset of the Mesolithic in the study area.
Transportation routes, linking raw material outcrops and settlements, followed the river systems of Lahn, Main, Fulda, Schwalm, Ohm, and Eder. In addition, mountain ranges (e.g., the Vogelsberg and the Rhön mountains) could have served as important landmarks (cf., Hess, 2019). Mesolithic sites are often situated on elevated terraces that are protected from floods, near small streams leading to the tributary waters of larger rivers (cf., Fiedler, 1994;Pflug, 1993, pp. 35-37;Quehl, 1994). One of them is the site Niederweimar in the Lahn Valley (Schön, 2006;Urz et al., 2002, p. 275). Archaeological excavations yielded microliths as well as pebbles that served as retouching tools comparable to the finds from Feuersteinacker. At this site, the percentage of local raw materials (mainly siliceous shale) is much higher, which indicates shorter stays. The faunal assemblage consists of roe deer, red deer, wild boar, and aurochs (Bos & Urz, 2003, p. 32). Other sites with a similar age that probably served as small hunting camps were discovered at Lahrbach and Kleinsassen (Pflug, 1993, pp. 15-17, 35-37). In both cases, the proportion of local raw materials is higher than 60%. Whereas the assemblage found at Lahrbach consist mainly of Triassic chert from Muschelkalk formations, most lithic artefacts from Kleinsassen are made of chalcedony. Second, it can be shown that the technology of Early Mesolithic groups in Central Germany allowed the use of a broader spectrum of lithic resources compared to previous time periods. Large assemblages dating to the Late Palaeolithic are usually dominated by siliceous shale, which contributes between 50% and 90% to the raw material (Fruth, 1994;Hofbauer, 1992;Loew, 2005;Riede, 2012). The observed differences are connected with shifting settlement dynamics and a reduction of residential mobility (cf., Binford, 1980). In case of the Hessian Late Palaeolithic/Early Mesolithic, these are likely to reflect a shift from a rather ephemeral human presence in the region during the final Pleistocene towards a more sustained land use in the early Holocene.
As rock types associated with Tertiary volcanism are frequently weathered when exposed to the surface, it is even possible to suggest that a simple form of mining (such as digging shallow pits) was used to obtain the material. On a large scale, the processes described above can be interpreted as an adaptation to changing environmental factors. In Central Germany, the combined impact of the Laacher See volcanic eruption around 13,000 cal BP, the subsequent Younger Dryas cooling, and the equally dramatic warming in the early Holocene likely led to this reorganisation of mobility and raw material procurement (cf., Riede, 2016).
Future research could focus on integrating the results of provenance analysis into an existing geographic information system-based predictive model (Sauer et al., 2018) to detect new sites dating to the Late Palaeolithic and the Mesolithic in the study area. This could contribute to a better understanding of the Pleistocene-Holocene transition in the region and lead to an archaeological definition of the respective cultures.