Pleistocene freshwater ostracods from the Homo erectus site at Bilzingsleben, Germany—Review of historic collection and unpublished manuscript material for palaeoenvironmental reconstruction

We provide a review of micropalaeontological research on Ostracoda from the Middle Pleistocene (MIS 11, Holstein interglacial) hominin site Bilzingsleben in Thuringia in Central Germany from 1963 to the 1990s. Samples from four sections inside and six search pits outside the excavation area were investigated and, in total, 49 ostracod species were identified. The ostracod assemblages of the sections mirror the complex and small‐scale palaeoenvironmental evolution of the site from a seeping‐spring to fluviatile, lacustrine and finally seeping‐spring habitat in which a massive tufa layer formed and prevented erosion of the sediments beneath. Pleistocene index fossils are represented by Ilyocypris quinculminata from search pit 3/sample 9933 and Scottia browniana from section 70. Both species indicate the age dating of MIS 11 for the tufa deposit. The results of this study facilitate new insights into site formation processes, enable refinement of the interpretation of the archaeological record and shed light on the question: Does the find‐bearing layer at the Bilzingsleben site contain in situ remains of a camp site of Homo erectus or not? Our results suggest that the site is not unaffected at least.

relief inversion caused by the massive tufa bank that formed c. 370,000 years ago in the Holstein interglacial period (Mania & Altermannn, 2004). It prevented erosion of the unconsolidated tufa sand beds and lacustrine lime beneath and preserved the archaeological record enclosed in these sediments (Mania, 1990).

| Scientific controversy
For more than four decades, from 1963 to 2007, thousands of finds were excavated at the Bilzingsleben site: bones of bears, horses, woodland elephants, rhinos, bovids, suids and many more  (Mania & Mania, 2001). The most important and significant finds were the remains of H. erectus and his legacies in the form of flint and other stone and bone tools. Based on the H. erecuts bone finds, at least four individuals could be distinguished (Vlček, 2002).
The excavations carried out over decades continually brought new findings and refined the scientists' ideas about the site.
Sedimentological and palaeontological investigations of the layer sequences enabled the identification of an alluvial fan and a shore area (Figure 2b). In the shore area numerous mammal bones, elephant tusks, large stones and charcoal remains were found and interpreted as an almost in situ record of a Lower Palaeolithic hunting camp with outlines of huts with fireplaces, distinct F I G U R E 2 (a) Location of the excavation site and the search pits in the outcrop area of the Bilzingsleben tufa deposits. The search pits (SPs) are indexed with their numbers. SP 2-6 and 11 yielded ostracods. (b) Spatial interpretation of the excavation site with a shore area (grey) and an alluvial fan (white). The main excavations were carried out between 1969 and 2002. Areas A-C were excavated in [2003][2004][2005][2006][2007], of area C the southern and south-western quadrants only (arrows). Compiled from Mania (1980),  and Beck et al. (2007) and a current topographic map. DANIEL and FRENZEL | 447 activity zones and an area which is supposed to have been deliberately paved with bones and stones (Mania, 1990). A large proportion of the archaeological finds, on the other hand, were found in the alluvial fan close to the shore area. The latter, moreover, was incised by small streams whose channels are filled with tufa sand. The rising water level of a forming lake later flooded the shore area and the archaeological record was covered by tufa sand and lacustrine lime (Mania, 1990). A massive tufa bank protected the sediments from erosion and thus the archaeological findings have been preserved until our time.  Beck et al. (2007), Müller andPasda (2011), Pasda (2012) and Liebermann and Pasda (2014). They doubt that the distribution of finds at the shore area represents an in situ record and instead make natural environmental and sedimentation processes accountable for the vertical and lateral scatter of stones, bones and lithics. A gravitational mass flow theory was developed, according to which these should have led to the chaotic distribution of finds (Beck et al., 2007). Brasser (2017) stated, after investigating bones excavated in 1971-2002, that the hominids were certainly part of the natural environment but no signs of zonal activities in this area could be deduced from the distribution of mammal bones. The inventory of the site was subjected to fluviatile relocation and to a series of other taphonomic processes during and after being frequented by animals and hominids (Brasser, 2017). The question now is: Does the Bilzingsleben site contain in situ remains of a prehistoric hunting camp or do the archaeological finds have to be considered as relocated according to the gravitational mass flow theory?

| Micropalaeontological approach
Since ostracods are reliable indicators for various environmental parameters in aquatic sediments, for example, salinity, temperature, habitat structure, water chemistry and transport conditions, they have a great potential for ecological monitoring and palaeoenvironmental analyses (Frenzel & Boomer, 2005;Griffiths & Holmes, 2000).
In Quaternary sediments, the value of ostracods as bioindicators is augmented because most species are extant and can be studied alive in their natural environment. For this reason, they are excellent implements for an actualistic approach for the reconstruction of palaeoenvironments and are increasingly used for the clarification of such questions (Quante et al., 2022).
Although numerous papers and books about the Bilzingsleben excavation site have been published- Mania and Mania (2001) summarize 30 years of Bilzingsleben research and Brühl (2015) provides a summary of the most important articles and books-only a few dealt with freshwater ostracods: Unger (1963), Diebel (1979) and Diebel and Pietrzeniuk (1980). Since then, Erika Pietrzeniuk (1935-2015 had conducted more research on ostracod assemblages and the tufa sequence at the Bilzingsleben site. For various reasons, however, these results were never published.
This study aims to contribute to the refinement of the interpretation of the archaeological record by examining the diverse freshwater ostracod fauna studied and documented by Erika Pietrzeniuk to reconstruct the environmental conditions of the site affecting aquatic species and their distribution in space and time. With this palaeoenvironmental data set, the interpretation of site formation processes can be substantiated and improved.

| MATERIALS AND METHODS
The micropalaeontological material investigated in this study is housed in the micropalaeontology collections of the Museum of Natural History in Berlin (MfN). More than 650 microslides with picked and identified ostracod valves from sections 70,171,, five search pits 2-6 (dredge excavations, hereafter referred to as pit or abbreviated SP) from Holsteinian deposits and one from Eemian sediments (pit 11, sample 2478) of the Bilzingsleben site ( Figure 2) were reviewed, checked at the museum and also loaned for further investigation and photographic documentation. The valves were counted semiquantitatively and listed in six abundance classes: 1-2, 3-5, 6-10, 11-25, 26-100 and >100 specimens, as already done by Pietrzeniuk. The documentation of ostracod analyses (unpublished data), for example, drawings of sections, photographs of ostracods and compiled tables, was reviewed, digitized and edited in this study on behalf of E. Pietrzeniuk. The taxonomical identification of ostracods was checked based on Diebel and Pietrzeniuk (1975, 1977, 1978, 1980, 1984, Meisch (2000) andFuhrmann (2012). Stratigraphic, ecological and habitat interpretation relies on Griffiths (1995), Meisch (2000) and Fuhrmann (2012). The taxonomy is based mainly on Meisch (2000). An exception is made for Cyclocypris taubachensis, which refers to the taxonomic opinion of Fuhrmann (2012 (2007), Horne and Mezquita (2008) and Horne et al. (2012), as well as calibrated data by Dave Horne (personal communication, January 5, 2012), were applied to the ostracod assemblages. Salinity estimation was conducted with data from Frenzel et al. (2010) and Pint et al. (2015Pint et al. ( , 2017; it is given in PSUs (practical salinity units), which corresponds to per mille (‰).
Since the information value of the ostracod assemblages of the samples from the five Holsteinian pits is limited, this study focuses on the sections from the excavation area ( Figure 2b) regarding palaeoenvironmental reconstruction. The ostracod assemblage of sample 2478 from pit 11, which is situated c. 270 m southeast of the excavation area, was included for palaeo air temperature estimation of the Eem interglacial.

| Preservation of microfossils
The preservation of ostracod valves is exceptionally good due to conservation in a tufa setting and the covering of soft and loose sediment layers by a compact tufa bank. In parts, some signs of superficial dissolution occur. Internal moulds were predominantly found in the sandy layers rich in siliciclastics and in covering layers consisting of tufa sands. Most of the moulds within the siliciclastics-rich layers belong to Mesozoic ostracods eroded from the surrounding areas. They consist of siliceous material as well. The moulds from tufa sands are composed of calcite and represent filled and then broken and/or dissolved Quaternary ostracod valves. In the lime abundant remnants of charophytes were found together with the documented ostracods. Charophytes have been preserved as hollow calcified oospores and stem remnants of Chara hispida and minor Chara vulgaris.

| List of ostracod taxa
In total, 49 freshwater ostracod species were identified from the Bilzings-le-ben site and are listed alphabetically in

| Sections
The location of the four sections investigated refers to the excavation map by  and is indicated with the numbers of the excavation grid squares 70, 171, 218 and 362 (Figure 2b). F I G U R E 3 Distribution of ostracod taxa in section 70, modified from by Diebel and Pietrzeniuk (1980).  (Table 2) is important for the dating of the site since it is an index fossil for the Lower and Middle Pleistocene (Fuhrmann, 2012). I. bradyi and Heterocypris reptans become more abundant in the sandy lime, and species diversity increases, although most species occur with low valve counts only ( Table 2). The total number of valves peaks at 2000 in sample 9197, which also showed the highest species diversity.

| Palaeo air temperature
By applying the MOTR method by Horne (2007), Horne and Mezquita (2008) and Horne et al. (2012) to the ostracod assemblages of the excavation area, the mean July air temperature lay somewhere between +16°C and +20°C, while the mean January air temperature was somewhere between −4°C and +4°C ( Figure 9). The most indicative species are F. hyalina, P. zenkeri and S. pseudobrowniana, which are described by Meisch (2000) as oligothermophilic. With regard to the MOTR method, it should be noted that the MOTR Temperature estimations (Mania, 1990; Figure 9) based on fossil floral elements from the Bilzingsleben tufa point to a Central European to sub-Mediterranean climate with a temperature range of −0.5°C to +3°C in January and +20°C to +25°C in July with a mean of +20.5°C. Both the winter and summer temperature ranges overlap the equivalent MOTR ranges, even though only the minimum of the July floral range matches the maximum of the July MOTR range.
For palaeo air temperature reconstruction of the Eemian deposit (search pit 11; Figure 10), only two species were indicative: C. subterranea and P. zenkeri. The species reflect a temperature range T A B L E 1 Alphabetical list of ostracod taxa from the Bilzingsleben site, Thuringia, Germany compiled from Diebel (1979), Diebel and Pietrzeniuk (1980), unpublished data from Erika Pietrzeniuk and new investigations (Daniel et al., in preparation).

Cyclocypris serena (Koch, 1838)
Cyclocypris taubachensis (Diebel & Pietrzeniuk, 1984) Cypria ophtalmica (Jurine, 1820) Scottia pseudobrowniana (Kempf, 1971) Note: Not all species occur in the species distribution diagrams of this study, but were reported in the literature, found in search pits outside the excavation area or documented by new investigations on ostracods (Daniel et al., in preparation of −4°C to +4°C in January and +14°C to +20°C in July. Compared to the temperature range of the Holsteinian, the average temperature in July deduced from ostracods from the Eemian interglacial deposit shows a minimum two degrees lower and may indicate slightly colder summers compared to the Holsteinian palaeoclimate. This temperature estimation, however, is subject to greater uncertainties since the indicative ostracod species do not occur with abundant valve counts.

| Salinity
The salinity range estimation of the former water body (Figure 11) was conducted with ecological data given by Frenzel et al. (2010) for the southern Baltic Sea area and by Pint et al. (2015Pint et al. ( , 2017 for Central Germany. Species limiting the salinity range from 0.4 (freshwater) to 3.6 PSU (oligohaline) are C. torosa for the lower limit and P. lobipes for the upper one. C. torosa is well known from brackish waters and lives in shallow coastal waters as well as in saline inland waters (Fuhrmann, 2012). It can tolerate slightly brackish to hypersaline conditions (Meisch, 2000). Neale (1988)  C. torosa is used as an indicator for brackish waters and as an index species in palaeosalinity reconstructions relying on water chemistrybound morphological changes such as nodes and sieve-pore shapes (Pint et al., 2012). Since the species predominantly form nodes within a salinity range of 0.4-2 PSU (Frenzel et al., 2010), the occurrence of the species with smooth, unnoded valves in sample 9191 of section 70, although with very few valve counts, points to an increased salinity and may refer to a population of the species in more saline waters at a site close by. Due to the dissolution of salt by the subrosion processes, such sites are quite common in Central Germany (Pint et al., 2012). P. lobipes occurs in freshwater and β-oligohaline waters. It tolerates salinities up to 3.6 PSU (Frenzel et al., 2010;Meisch, 2000) and indicates the maximum salinity of the site. Some ostracod species with specific requirements to their environment were very indicative (Table 3). In contrast, less importance in the reconstruction was given to other species with less restrictive environmental requirements. Unfortunately, sedimentological data such as grain-size analyses were not available for these sections, so we had to rely on the schematic section drawings and the ostacod data alone.

| Section 70
The ostracod assemblage of section 70 (Figure 3 and Table 2 with additional taxa) points to an initially shallow lacustrine (III-A), respectively, a fluviatile (II) and then fully lacustrine (III) habitat. C.
vavrai and P. fallax indicate springs and spring-related waters (Meisch, 2000). H. salina can be an indicator for slightly saline conditions, but Meisch (2000), Fuhrmann (2012) and Pint et al. (2012) document the species in pure freshwater also. C. torosa from sample 9191 is exotic within the faunal association (Table 2). It can tolerate freshwater conditions but occurs in brackish water bodies in coastal environments in general. The tufa sand is interpreted as a springrelated sediment deposition into a shallow, lacustrine water body that was influenced by flowing water (II).
The ostracod fauna of the sandy lime (samples 9193-9195, habitat III-A) is more diverse than within the tufa sand. D. stevensoni, C. ophtalmica, Cypridopsis vidua and E. pigra point to a beginning stagnation phase of the water body which induced the sedimentation of pure lacustrine lime. D. stevensoni and C. ophtalmica, for instance, prefer stagnant water bodies but can also occur in flowing water (Fuhrmann, 2012;Meisch, 2000). The occurrences of C. vavrai and P.
fallax indicate the influx of water coming from a spring, as both species live in springs (Fuhrmann, 2012). Internal moulds of Mesozoic (Anisian, Ladinian) ostracods occur in sample 9195 for the last time and indicate the decrease and then end of sediment input from external sediment sources.
The fauna of the pure lime is diverse and abundant and mirrors a standing water body (III). The dominant taxa, such as Cyclocypris sp., P. lobipes, P. marchica and F. hyalina, are found in swampy and muddy habitats with decaying organic matter (Meisch, 2000). Stagnant conditions without water movement are indicated by C. ophtalmica, P. euplectella and N. monacha. C. ophtalmica colonizes, besides a wide range of other aquatic habitats, ponds with fallen leaves and organic F I G U R E 7 Ostracod taxa from the Bilzingsleben site. Photographs 1-22 by Diebel and Pietrzeniuk (1980) detritus, and P. euplectella lives in small muddy and swampy water bodies (Meisch, 2000). N. monacha feeds on the residue film of the water surface and only occurs in still water areas (Meisch, 2000). The ostracod fauna of the tufa sand lens intercalated in the uppermost massive tufa bank shows abundant M. zimmeri, which points to seeping-spring habitats (IV) in which the dense and more cemented tufa bank formed. This tufa bank protected the underlying sediments from erosion and probably existed in all profiles. It was partially removed by centuries of quarrying.

| Section 171
The faunal association of section 171 (Figure 4) mirrors flowing water conditions for the tufa sand of samples 10-023 to 10-025. The rheophilic species P. zenkeri (Meisch, 2000) and the low abundance of ostracods are significant indicators of the fluviatile stage II. The sandy lime is diverse and rich in individuals with diversity peaking in sample 10-027. The lacustrine habitat III with sandy and a fine sandy to silty lake bottom can be deduced due to the presence of P. marchica, P. lobipes, Cyclocypris sp., C. ophtalmica and P. euplectella. P. marchica and P. lobipes both live in small, permanent water bodies (Meisch, 2000). The increased abundance of H. salina in sample 10-028 is possibly a result of a change in the chemistry of the water body caused by higher evaporation rates, decreasing water inflow and stagnant water conditions, the latter indicated by N.

| Section 218
The ostracod assemblage obtained from the tufa sand of sample 10-016 ( Figure 5) indicates a sandy bottom in a shallow lacustrine habitat (III-A). The occurrence of H. salina may indicate increased salinity, possibly caused by evaporation and reduced water exchange. The faunal association changes insignificantly in sample 10-017 but the abundance of ostracods grows rapidly in a now fully established lacustrine habitat with a permanent water coverage (habitat III) indicated by P. marchica, P. lobipes and C. ophtalmica. M. zimmeri and F. breuili in sample 10-018 may originate from springs fed by groundwater and flowing into the lake as M. zimmeri occurs in seeping springs and F. breuili in the interstitial groundwater (Meisch, 2000). The inflow of the spring water is not verifiable for samples 10-019 and 10-021 but is detectable again above the lacustrine lime in the loose tufa. The habitat can be described as a shift from lacustrine (III) to a setting with seeping springs (IV) with humid to wet moss where tufa precipitated at the plant remains. M. zimmeri, S. pseudobrowniana and F. breuili are indicators of this habitat which was fed by freshwater-seeping springs.

| Section 362
A unique characteristic of section 362 (Figure 6) is the faunal assemblage with S. pseudobrowniana and M. zimmeri dominating in loose tufa at the base of the section and in the sandy limonitic layers intercalated in a massive tufa bank. In conjunction with F. breuili, C.

| Ostracod taxonomy and geological age
As the taxonomic scheme for this study is based on Meisch (2000), C.
ovum and C. taubachensis have to be discussed in brief. Diebel and Pietrzeniuk (1984) described C. taubachensis from sediments of the Eemian Parkhöhle (park cave) in Weimar (Thuringia, Central Germany) and therefore distinguished C. ovum and C. taubachensis among ostracods from the Bilzingsleben site. In contrast, Meisch (2000) stated that since C. ovum has a considerable variability of length and shape of the carapace in lateral and dorsal views, and shows variations in the colour of the valves, several separate species were described by different authors. Because of this variability, Meisch (2000) and Matzke-Karasz (1995) assign the carapace of C. taubachensis to C. ovum. On the other hand, Fuhrmann (2012) confirms C. taubachensis and distinguishes the species from C. ovum by a prominent dorsal angle. Meisch (2000) argues that the relationship of C. taubachensis with C. ovum must be clarified by the description of the appendages. However, in this article, since the species diagrams were compiled by Pietrzeniuk and species identity is not finalized, C. taubachensis remains in the diagrams (Figures 4-6) and in the taxa list as a separate taxon.
Sample 9933 from pit 3 yielded a new species, Potamocypris sp. A ( Figure 8, 17-18). This species is going to be described in a future publication. browniana lived in standing water bodies (Kempf, 1971). Fuhrmann (2012) states for S. browniana that no recent occurrence is known, and the species is probably extinct. For this reason, he characterizes it as an index fossil of the Lower and Middle Pleistocene. Besides Bilzingsleben, S. browniana was found only in two other localities in Central Germany: Lützensömmern and Kalbsrieth (Diebel, not published, cited by Kempf, 1971). I. quinculminata was found in T A B L E 2 Rare species from section 70 were documented by Diebel and Pietrzeniuk (1980) and according to collection material from the Natural History Museum Berlin (MfN).  Mania (1990) indicate average air temperatures of −0.5°C to +3°C in January and +20°C to +25°C in July (dark grey rectangle). The WorldClim database values for Bilzingsleben are −0.1°C in January and +17.6°C in July. Recent average air temperatures were estimated at +0.8°C in January and +19.9°C in July at the nearby Kindelbrück meteorological station. Wildschütz, Saxony, Germany (Fuhrmann, 1991) and by Kempf (1997) in Wohnbach near Berstadt, Hessen, Germany. Diebel and Pietrzeniuk (1980) already mentioned finds of I. quinculminata from the Bilzingsleben site but did not picture them. As no recent occurrence is known so far, Fuhrmann (2012) assumes that I.

Sample
quinculminata is also extinct and therefore represents an index fossil probably of the Lower and for sure the Middle Pleistocene. According to Fuhrmann (2012), it lived in standing water bodies. Morgan (1973) assigned it to slowly flowing water with calcareous silt as sediment. Griffiths (1995)

| Section correlation
The section correlation (Figure 12) could be conducted based on the sediment layers, the ostracod associations and the interpreted palaeoenvironmental conditions. As with the interpretation of the zenkeri (Fuhrmann, 2012). Since the species assemblage of the lower part of section 70 is also indistinct to some extent, it is correlated to the fluviatile setting likewise, even though P. zenkeri was not found here. F I G U R E 11 Salinity reconstruction of the Bilzingsleben site based on ostracod environmental data given by Frenzel et al. (2010) for the Baltic Sea area, Neale (1988) for Cyprideis torosa, and Meisch (2000), light grey bars. Further observations by the authors and/or unpublished data from third parties indicate higher salinity tolerances for some species (dark grey bars). The black bar indicates dominant populations of C. torosa at a salinity range from 0.5 to 2 PSU (Frenzel et al., 2010). The mutual salinity range of the Bilzingsleben site is estimated at 0.4-3.6 PSU. PSU, practical salinity units.
euplectella (Meisch, 2000). The deposition of tufa sand at the top of sections 218 and 362 can be assigned to a developing seeping-spring habitat (IV). Initially, tufa sand from the spring(s) was deposited in the forming lake since the ostracod record is still diverse and species abundance is high. For this reason, a further mixing of habitats is indicated (III/IV). Later, the deposition of tufa sand increased and the seeping-spring habitat (IV) developed fully and a massive tufa bank formed due to the ongoing precipitation of tufa.

| Site development
The ostracod assemblages of the sections investigated by Diebel and Pietrzeniuk (1980, section 70) and Pietrzeniuk in the 1980s (sections 171, 218 and 362) indicate a natural development from a fluviatile to lacustrine and seeping-spring setting or, in the case of section 362, from seeping springs to lacustrine and finally seeping-spring habitats.
Section 218 reflects a shallow-water environment with components from seeping springs merging into a lacustrine and then fully seepingspring habitat. Section 171 shows a development from fluviatile to lacustrine conditions ( Figure 13).
The ostracod species corroborate their facies and habitat dependence. They are reliable indicators for stratification and correlation of the sections as well as the reconstruction of the palaeoenvironments ( Table 3) (Fuhrmann, 2012;Meisch, 2000). Springs and spring-related flowing waters (habitat I-A; Figures 12 and 13) are indicated by F. breuili, C.
vavrai and N. faba (Meisch, 2000). As flowing water plays a significant role in these habitats, archaeological finds may also have been  +20°C. It would be of great benefit for the reconstruction of the site development to search for these sediments and to sample and analyse them fine stratigraphically as well.

| CONCLUSIONS
By investigating the diverse freshwater ostracod associations of four sections and six search pits, it was possible to substantially refine the analysis of the natural development of the Bilzingsleben site.
Palaeoenvironmental parameters like mean air temperature with a range of −4°C to +4°C in January and +16°C to +20°C in July point to mild winters and warm summers with temperatures similar to the present-day climate. For the Eemian deposits, a temperature range of −4°C to +4°C in January and +14°Cto +20°C in July could be F I G U R E 13 Site excavation map with the location of the examined sections 70, 171, 218 and 362 and their habitat interpretation. Compiled from Mania (1980),  and Beck et al. (2007). DANIEL and FRENZEL | 463 For further investigations, it is essential to carry out finestratigraphic analyses of the preserved sections for the specification of the statements made in this study. This is especially true for the deposits on the shore area, which show a high density of finds.
Further analyses could clarify to what extent the find layer, which was doubted to be in situ, was influenced by water-induced transport, resedimentation and other taphonomical processes.