Middle Paleolithic Neanderthal Open-Air Camp and Hyena Den Westeregeln (D)—Competition for Prey in a Mammoth Steppe Environment of Northern Germany (Central Europe)

Abstract

A gypsum karst sinkhole at Westeregeln (north-central Germany) was filled during the Late Pleistocene, first by fluvial flooding, then by solifluctation, and finally with wind-transported loess. Pleistocene mollusks and bones of snakes, birds, micro- and macromammals, and hyena coprolites were accumulated, often mixed in gravel or sand layers with Middle Paleolithic artifacts, whereas ice wedges reach deep into the sinkhole. The high amount of small flint debris prove on-site tool production by using 99% local Saalian transported brownish-to-dark Upper Cretaceous flint, which could have been collected from the Bode River gravels near-site. Only a single quartzite and one jasper flake prove other local gravel sources or importation. A large bifacial flaked knife of layer 4 dates to the early/middle Weichselian/Wuermian (MIS 5-4), similar to two triangular handaxes in the MTA tradition and an absolutely dated woolly rhinoceros bone (50,310 + 1580/−1320 BP). A cold period of Late Pleistocene glacial mammoth steppe megafauna is represented, but the material is mostly strongly fragmented and smashed by humans. Neanderthal camp use on the gypsum hill is indicated also by small charcoal pieces, burned bone fragments, and fire-dehydrated flint fragments. Crocuta crocuta spelaea (Goldfuss) hyenas are well known from Westeregeln, with an open-air commuting den site, which was marked with feces.

1. Introduction

Bone remains of Ice Age animals became known from the Kalkberg near Westeregeln first already in the middle of the 19th century, whereas the oldest labeled one was from a lion found by Beyrich in 1844 in one of the three main Bergling’sche Gipsbruch (Figure 1; cf. [1,2,3,4]). The first larger excavations and bone acquisitions from quarry workers who collected bone material were made by Giebel from 1843 to 1849 [5]; this bone material is still preserved but not redescribed yet. Giebel was the first to interpret the site as a hyena site. He published [6,7] some woolly rhinoceros remains and figured some of the teeth, which have recently been revised and partly refigured in the European rhino revision [8].

Later, more macro- but also many micromammals were excavated and collected or bought by Nehring, who mainly focused on horse and micromammal remains but also mentioned the first stone artifacts, which seem to have been found in the uppermost soil layers together with Neolithic and Bronze Age pottery, a dog canine tooth necklace (still preserved in the MB), and dog remains, but also stone artifacts [9], and finally charcoal pieces [10,11,12,13,14,15]. The only regular blade being lost seems to be from the Holocene Neolithic/Bronze Age cultures (most probably not of “Upper Paleolithic” age); therefore, perhaps no Paleolithic stone tools were known at that time.

Figure 1. (A). Middle Paleolithic and Late Pleistocene hyena den sites in Saxony-Anhalt and the position of the gypsum karst site Westeregeln. (B). The historic finds of Beyrich, Giebel, and Nehring are all, or mainly, from the large Berling’sche gypsum quarry, (C). Northeastern wall excavated in 2009–2010 in a deep sinkhole fill. (D). Brick factory, (E). Berling’sche gypsum quarry; around the corner is the excavated sinkhole.

Figure 1. (A). Middle Paleolithic and Late Pleistocene hyena den sites in Saxony-Anhalt and the position of the gypsum karst site Westeregeln. (B). The historic finds of Beyrich, Giebel, and Nehring are all, or mainly, from the large Berling’sche gypsum quarry, (C). Northeastern wall excavated in 2009–2010 in a deep sinkhole fill. (D). Brick factory, (E). Berling’sche gypsum quarry; around the corner is the excavated sinkhole.

Toepfer [16] composed the old descriptions without studying the original material or carrying out new fieldwork. Some more megafauna material determined here as Crocuta crocuta spleaea skull remains, Mammuthus primigenius, Bison priscus, and Rangifer tarandus are without exact documentation context. Newer excavations outside the Berling’sche quarry (Figure 1) in a large southeastern sinkhole depression delivered some new microfauna of Weichselian Late Pleistocene age [17]. Already in the 1950s, the first Middle Paleolithic discovery—a “Micoquid” handaxe found by the Westeregeln teacher Mrs. Breithaupt and used as demonstration material for the school children but never reported to the State Archaeology—were the first proof of Neanderthal presence.

Following the discovery of one more hyena skull braincase, a Coelodonta antiquitatis antiquitatis lower jaw, and a Canis lupus sp. lower jaw in the British Museum Natural History London (BMNH, originally from the Museum Magdeburg; obviously deported in WWII as “robbery art”) in 2005 [18] within the “Ice Age spotted hyena project” and the relocating of the “lost Giebel 1848 collection”, including the most complete hyena skull from Westeregeln, new prospections relocated the original locality.

2. Materials and Methods

Three months in the summer of 2009 and three weeks in the summer of 2010, a small team made a full cross-section eight meters long and seven meters deep of the sinkhole fill (Figure 2, Figure 3, Figure 4, Figure 5 and Figure 6). First, the dump was removed, in which more than 50 stone artifacts and 50 megafauna bone fragments had already been discovered without stratigraphic context. Two larger flakes and one nucleus were reworked by modern foxes and rabbits that left their burrows (Figure 7), in which also modern Canniculus bone remains were in place, and remains of some micromammals.

In 2000, two handaxes were found in a small new brickyard pit nearby [19] (Figure 1B). These pieces, together with some white patinated flint flakes, were collected outside the gypsum area surrounding Buntsandstein in the paleosoil and in the ice wedges, which were filled with gravel, artifacts, and bone remains northwest of the brick factory and gypsum area. In the Berling’sche quarry sinkhole, remains of a woolly rhinoceros skull were found in 2005 in the weathered section, and finally the first two larger flakes within first cleanings of the Quaternary sinkhole section. Those finds in an eight-meter-wide and fifteen-meter-deep sinkhole filling resulted in a two-year excavation campaign from 2009 to 2010.

The artifacts from the dump in front of the cleaned sinkhole section were reworked by rabbits, badgers, or foxes, which made their den tunnel systems intensively in the Paleolithic layers (layer 5, Figure 7, Figure 8 and Figure 9). One of the handaxes (handaxe B, Figure 1) must have been removed from those layers by the carnivores in modern Holocene times. The excavations concentrated later on the central part of the sinkhole fill, where a final seven-meter-deep and four-meter-wide section reached the last artifact layers at a depth of 530 to 570 cm (Figure 7, Figure 8 and Figure 9).

The sediments (Figure 10, Figure 11 and Figure 12) themselves were checked macroscopically by the gravel content (erratic/local rock types and possible flint raw material) and were studied in their sediment structures (ice wedges, cross-bedding, and microtectonic) such as their fossil contents (reworked Oligocene fossils: Figure 13), to understand the origin of the sinkhole sediments.

Figure 2. Trench I–II, squares A–G, and plana with main measuring points (Gauß-Krüger Krasowski Ellipsoid 3° stripes and GPS height).

Figure 2. Trench I–II, squares A–G, and plana with main measuring points (Gauß-Krüger Krasowski Ellipsoid 3° stripes and GPS height).

Excavation Surfaces (=Plana)

The excavation surfaces (=plana) are not always parallel to the layers as a result of the complex sinkhole filling (cf. Figure 1, Figure 2, Figure 3, Figure 4, Figure 5 and Figure 6) and are excavated mostly in about 15 cm thickness. Stone artifacts, bones, and coprolites were mapped. Also, all gypsum block and other rock types were included to reconstruct the sedimentological history. In the first plana, eight square meters were mapped (squares A–H; Figure 3). Planum VII was documented only photographically. From Planum XIV to XXI, only one surface was documented because of a single interesting finding situation. In total, 11 plana were drawn (Figure 3, Figure 4 and Figure 5). Starting in Planum VI, only four square meters (A–D) were continued in lowering. The squares and position within Trench I–II and measuring points are given in Figure 2.

In total, 25 plana were lowered (Figure 3, Figure 4, Figure 5 and Figure 6). First, Plana II-V were excavated in 10 cm thickness over a length of eight square meters starting below the loess in layer 3 (=paleosoil). In those, all gypsum blocks and other large erratic or regional sediment blocks were mapped, including all finds (artifacts, bones, and hyena coprolites). With Plana VI to XIII, the trench was reduced to four square meters. With Planum XIV, those were extended with a half square to the margin because the original wall was conical and wider at the bottom. From Planum XVIII, the methodology changed to mainly sieving because of the clear identification of redeposited sediments and artifacts, which were found mainly accumulated in fine-grained gravels (sediments partly deposited by running water and ice wedge fills). The artifact sizes even correlated to the gravel size. Only one planum was finally documented in the second half of the complete section, where many artifacts and bones, including a woolly rhinoceros tooth, were found. This section was prepared at a depth of about 750 cm, whereas the last artifacts reached no deeper than 570 cm. Layers below, including a final gravel layer at a depth of 730 cm, did not contain any artifact remains but did contain the same reworked Oligocene and enriched Pleistocene mollusks.

Figure 3. Plana II to V (layers 3–4) of Westeregeln. The backflaked knife was found in Planum III (square C) at the top of layer 4. Legend: see Figure 5.

Figure 3. Plana II to V (layers 3–4) of Westeregeln. The backflaked knife was found in Planum III (square C) at the top of layer 4. Legend: see Figure 5.

Nearly all the sediments (several tons) of Trench I to II were sieved. The loess (layer 2) was checked for snails, which indeed are absent. In the upper layers 3 to 4, the fractions above 1 mm were sieved. Later, at a depth of about four meters (layer 5), the fraction was chosen to be larger (two mm). In those fractions, the main micromammals, snakes, frog remains, and also freshwater and land snails or bivalves (Figure 14 and Figure 15), and Oligocene reworked marine fossils (Figure 13), such as all small artifact flakes, were selected on-site. The sieving itself was carried out with two sieves on top of each other of one, two, and four millimeters to save as many fragile mollusks as possible. In several cases, snail specimens floated or were damaged, which has to be taken into account for statistics.

The locality area was prospected by geophysics to understand the morphology of the sinkhole and extensions of sediment types. Here, a larger area of about 45 × 60 m of surface to the NE of the quarry margin and section was prospected.

Bones of megafauna from the plana (Figure 16 and Figure 17) were only partly prepared, because most of them were encrusted in caliche, which results from the Post-Quaternary decarbonatization of the loess layer. In several cases, it was impossible to remove those. Coprolites (Figure 18) were similarly encrusted and therefore survived well, whereas non-calcified coprolites were nearly impossible to save and were damaged by weathering close to the section wall. The bones were only prepared mechanically without treating them for further possible geochemical analyses. Two coprolites were finally dated by isotope ¹⁴C.

The former isolated handaxe finds and newly excavated backed knife, such as excavated and sieved flakes (Figure 19 and Figure 20), are compiled in the artifact terminology. Finally, the complete sinkhole refill history is reconstructed (Figure 21). Nearby Westeregeln and other Neanderthal camp sites in and around the Harz Mountain Range, which are from different time periods and interglacials and glacials (MIS 9 to 3), are composed for a hunter-gatherer seasonal model with migration paths along rivers (Figure 22). To understand the competition, especially with Ice Age spotted hyenas, their den sites, which overlap in several places, are also mapped.

Figure 4. Plana VIII to X (layers 4–5) of Westeregeln. Legend: see Figure 5. A. Plana VI, VIII, IX, and X.

Figure 4. Plana VIII to X (layers 4–5) of Westeregeln. Legend: see Figure 5. A. Plana VI, VIII, IX, and X.

Figure 5. Plana XI to XIX (layers 4–5) of Westeregeln. (A). Planum XI. (B). Planum XII. (C). Planum XIII (with hyena fecal place 2). (D). Planum XIX with ice wedge fill of redeposited Saalian gravels including wind-polished blocks (Block, LDA no. WE-2933-1212). The gravels of layer 5, strongly burrowed in modern times, seem to have delivered the handaxe.

Figure 5. Plana XI to XIX (layers 4–5) of Westeregeln. (A). Planum XI. (B). Planum XII. (C). Planum XIII (with hyena fecal place 2). (D). Planum XIX with ice wedge fill of redeposited Saalian gravels including wind-polished blocks (Block, LDA no. WE-2933-1212). The gravels of layer 5, strongly burrowed in modern times, seem to have delivered the handaxe.

Figure 6. (A). Overview of the gypsum block surface of Planum 7 (−290 cm) in layer 7. (B). Gypsum block surface of Planum XXV at a depth of 6.5 m of layer 9.

Figure 6. (A). Overview of the gypsum block surface of Planum 7 (−290 cm) in layer 7. (B). Gypsum block surface of Planum XXV at a depth of 6.5 m of layer 9.

3. Geology

3.1. Gypsum Karst Sinkholes

Northeast of the section, a georadar prospecting of about 45 to 60 m of surface allowed a morphological separation between Pleistocene sediments, gypsum, and Buntsandstein red clay sediments, which was not that informative. The most important result is the presence of three sinkhole depression structures, of which the section is in the middle of one half-destroyed and Pleistocene-filled sinkhole (Figure 1). Northeast of this, another larger sinkhole is present, in which much more archeological material can be expected in the future. Also of importance is the result that those three depressions are on the most northern margin of the gypsum hill. North of those other sediments, such as red clays from the Lower Triassic, a smooth surface is built on which loess was deposited in direct contact.

Figure 7. Section of the middle sinkhole filling of cut I–II (squares -AA—Treppe) at Westeregeln. (A). Detail of the “artifact gravels” and artifact in the section (layer 5). (B). Overview of the central and deepest part of the sinkhole. (C). Boundary between artifact-rich layers (layer 5) in the upper part and artifact-free layers (layers 7–8) in the lower part. Modern red fox and rabbit burrows destroyed many parts, but explain the artifacts in the “dump cone” in front of the section, including the handaxe.

Figure 7. Section of the middle sinkhole filling of cut I–II (squares -AA—Treppe) at Westeregeln. (A). Detail of the “artifact gravels” and artifact in the section (layer 5). (B). Overview of the central and deepest part of the sinkhole. (C). Boundary between artifact-rich layers (layer 5) in the upper part and artifact-free layers (layers 7–8) in the lower part. Modern red fox and rabbit burrows destroyed many parts, but explain the artifacts in the “dump cone” in front of the section, including the handaxe.

The bone remains, which are mostly fragments of the megafauna but also nearly complete bird bones, indicate well a secondary position in most cases, but the transport cannot have been far. The paraautochthon megafauna bone fragments are not rounded and have sharp breaking corners. The bones seem to be a mixture, because an articulated cranium and lower jaw of the ground squirrel Spermophilus rufescens and areas of two hyena fecal places prove autochthonous depositions. Spermophilus is known to have made burrows; hyenas seem to have used the depression for marking their den. Possibly, even the hyenas had a den burrow system for the cubs, but a clear burrow system could not be found; instead, modern large tunnels of foxes and badgers and medium-sized ones of rabbits were found (Figure 6).

Figure 8. Section of the middle sinkhole filling of cut I–II along the squares A–D (see Figure 2) at Westeregeln. Middle Paleolithic artifacts and bones or hyena coprolites were found in layers 3 to 5. Redrawing in Figure 9.

Figure 8. Section of the middle sinkhole filling of cut I–II along the squares A–D (see Figure 2) at Westeregeln. Middle Paleolithic artifacts and bones or hyena coprolites were found in layers 3 to 5. Redrawing in Figure 9.

The recently studied sinkhole has preserved a more complete section and is on the northern margin of the gypsum hill southwest from the Pre-Bode River valley and its former terraces, but not far from the recently preserved (in traces) glacigenic and glacifluvial Elsterian and Early Saalian sediments. All these sediments seem to be eroded in the surroundings of the gypsum hill and are represented only in the depression with the reworked layers.

Figure 9. Redrawn section of Figure 8. The red line indicates the deepest artifact finds.

Figure 9. Redrawn section of Figure 8. The red line indicates the deepest artifact finds.

Figure 10. Prepared surface of the periglacial soil (top of layer 3, surface) with solifluctation structures, ice wedge polygons, and a frost-cracked wind-polished erratic block. (A). Photo. (B). Redrawing.

Figure 10. Prepared surface of the periglacial soil (top of layer 3, surface) with solifluctation structures, ice wedge polygons, and a frost-cracked wind-polished erratic block. (A). Photo. (B). Redrawing.

Figure 11. Sediment types of layers 3 to 5 of the sinkhole filling in Westeregeln. (A). Layers 2–3A. Upper part: loess; middle part: sandy loess with ice wedges and polygons; bottom: periglacial soil. (B). Layer 3B. Gypsum blocks in non-stratified sandy loess. (C). Layers 3B and 4. Upper part: gypsum blocks in non-stratified sandy loess. Lower part: sand with clay layers that are deformed. (D). Layer 5: alternating silt and sand/gravel layers with some microtectonic structures. (E). Gravels from layer 4, which contain many artifacts. (F). Cross-bedding structures in a sand layer of layer 4.

Figure 11. Sediment types of layers 3 to 5 of the sinkhole filling in Westeregeln. (A). Layers 2–3A. Upper part: loess; middle part: sandy loess with ice wedges and polygons; bottom: periglacial soil. (B). Layer 3B. Gypsum blocks in non-stratified sandy loess. (C). Layers 3B and 4. Upper part: gypsum blocks in non-stratified sandy loess. Lower part: sand with clay layers that are deformed. (D). Layer 5: alternating silt and sand/gravel layers with some microtectonic structures. (E). Gravels from layer 4, which contain many artifacts. (F). Cross-bedding structures in a sand layer of layer 4.

Figure 12. (A). Artifact-rich gravels in the lowermost artifact “layer”, which might represent ice wedge refills (bottom layer 5). (B). Gypsum residual clay without stratifications and archeological or faunal remains (layer 7a). (C).Rhizome-negative clay bed (layer 6). (D). Artifact- and fauna-free medium coarse brownish sands with fine gravels (layer 7). (E). White sand (layer 8) and fine clay layers alternating with microtectonic horst/graben structures in contact with layer 9 (bottom). (F). Non-stratified mixed fine gravel medium loess-rich sand of layer 9.

Figure 12. (A). Artifact-rich gravels in the lowermost artifact “layer”, which might represent ice wedge refills (bottom layer 5). (B). Gypsum residual clay without stratifications and archeological or faunal remains (layer 7a). (C).Rhizome-negative clay bed (layer 6). (D). Artifact- and fauna-free medium coarse brownish sands with fine gravels (layer 7). (E). White sand (layer 8) and fine clay layers alternating with microtectonic horst/graben structures in contact with layer 9 (bottom). (F). Non-stratified mixed fine gravel medium loess-rich sand of layer 9.

3.2. Stratigraphy, Sedimentology, Dating, and Bone/Artifact Taphonomy

The 7.5 m deep excavated gypsum karst section can be subdivided in the center into about nine main layers (Figure 8 and Figure 9). The upper 60 cm (A/B horizons) are dark brown soils that originated from the underlying 60 cm thick loess. The basal periglacial MIS2 (LGM)-dated soil contains some frost-fractured erratic blocks (Figure 10), and local blocks from the Harz Mountains are the stone components. Those blocks are from the Elsterian or Saalian glacial, which must have been reworked fluvial from the Elsterian or Saalian river terraces, tills, and glaciofluvial sediments by a pre-Bode River in the surroundings. Typical are also large reworked Tertiary freshwater quartzite blocks. The gravels (Figure 11E) contain regional Harz mountain rock types and reworked Middle Pleistocene Scandinavian material, and also northern German Upper Cretaceous flint and some belemnite rostra fragments of Belemnitella sp. from originally Campanian layers, which are rock types found along the southern East Sea coast. Within those gravels, not rarely must wind-edged facetted pebbles (Figure 5) of some centimeters to 40-centimeter-large blocks (Figure 5) have been first deposited by the Saalian or Elsterian glaciations in the region, then polished by wind abrasion, and finally transported by solifluction and/or periodical flooding, but is seems that layer 5 is also a periglacial soil and contains again ice wedges and fills (Figure 5). In the larger gravels, bigger bones and artifacts accumulated, which is well documented in Planum XIX (Figure 5). Layer 5 seems to represent less a solifluidal layer, such as a predominantly fluvial-like flood layer which was further cryoturbated, especially by ice wedges, which were again filled fluvially with large pebbles and other material during stronger flooding events. In the gravels (fluvioform and ice wedge fills) and in the sand layers, many Oligocene mollusk fragments and diverse marine fauna are mainly also present, which originate from Paleogene marine deposits in the near surroundings of the gypsum karst area. The fauna date the reworked sediments to the Upper Oligocene (see species in Figure 13). In those gravels (Figure 11E), the stone artifacts are enriched. The gravels found in layer 5 do not include a large percentage of local pieces and represent reworked material of Elsterian or Saalian origin (S. Wansa pers. com., 2010). The TGZ (=indicator erratic count) length of 15.24^o and a TGZ breadth of 58.08^o indicate reworked Drenthe material, which, however, is quite weathered and perhaps already influenced by Elsterian-II sediments (K.-D. Meyer pers. com., 2010). There are comparable Drenthe fluvioglacial sands in the Hadmersleben gravel pit, 5–6 km northwest from the Westeregeln site (cf. [20]), a section which is very different to the WesteregelnWeichselian/Wuermian fill.

Figure 13. Reworked Upper Oligocene (Paleogene) marine fossils from the Late Pleistocene gravel layers (layers 3 to 5) of Westeregeln. (A). Caryophyllia sp. (Anthozoa) (WE2933-792). (B). Rhynchonella sp. (Brachiopoda) (WE2933-677.3). (C). Eucalathis sp. (Brachiopoda) (WE2933-677.2). (D). Odontaspidae indet. (Selachii) (WE2933-677.4). (E). Balanus zonalis (Cirripedia) (WE2933-677.5). (F). Turitella sp. (Gastropoda), (WE2933-677.6). (G). Haustator sp. (Gastropoda) (WE2933-677.7). (H). Astraea sp. (Gastropoda) (WE2933-677.8). (I). Ficus sp. (Gastropoda) (WE2933-791). (J). Dentalium kicksi (Gastropoda) (WE2933-677.9). (K). Crassostrea sp. (Bivalvia) (WE2933-797.1). (L). Spondylus tenuispina (Bivalvia) (WE2933-677.11). (M). Pecten (Hilberia) cf. hoeninghausi (WE2933-797.2). (N). Glycimeris sp. (Bivalvia) (WE2933-677.12). (O). Cyclocardia sp. (Bivalvia) (WE2933-677.10). (P). Laevicardium sp. (Bivalvia) (WE2933-677.13). (Q). Astarte sp. (Bivalvia) (WE2933-677.14).

Figure 13. Reworked Upper Oligocene (Paleogene) marine fossils from the Late Pleistocene gravel layers (layers 3 to 5) of Westeregeln. (A). Caryophyllia sp. (Anthozoa) (WE2933-792). (B). Rhynchonella sp. (Brachiopoda) (WE2933-677.3). (C). Eucalathis sp. (Brachiopoda) (WE2933-677.2). (D). Odontaspidae indet. (Selachii) (WE2933-677.4). (E). Balanus zonalis (Cirripedia) (WE2933-677.5). (F). Turitella sp. (Gastropoda), (WE2933-677.6). (G). Haustator sp. (Gastropoda) (WE2933-677.7). (H). Astraea sp. (Gastropoda) (WE2933-677.8). (I). Ficus sp. (Gastropoda) (WE2933-791). (J). Dentalium kicksi (Gastropoda) (WE2933-677.9). (K). Crassostrea sp. (Bivalvia) (WE2933-797.1). (L). Spondylus tenuispina (Bivalvia) (WE2933-677.11). (M). Pecten (Hilberia) cf. hoeninghausi (WE2933-797.2). (N). Glycimeris sp. (Bivalvia) (WE2933-677.12). (O). Cyclocardia sp. (Bivalvia) (WE2933-677.10). (P). Laevicardium sp. (Bivalvia) (WE2933-677.13). (Q). Astarte sp. (Bivalvia) (WE2933-677.14).

4. Fauna

4.1. Mollusks

Several hundreds of Late Pleistocene mollusks were sieved, dominated by snails and a few 2–3 mm small thin-walled bivalves, but their small numbers are not sufficient for a good quantitative analysis per layer. The loess snail Pupilla muscorum (Figure 14A) was found in layers 4–5, but those were only abundant around the coprolite fecal areas (Planum IX) and were completely absent in the loess of layer 2. Fewer were the terrestrial snails such as Arianta sp. (Figure 14B). Especially in layers 4–5, freshwater mollusks are typically represented by Lymnaea sp.(Figure 14C), Succinea cf. oblonga (Figure 14D), Planorbarius sp. (Figure 14E), and Planorbis planorbis (Figure 14F), besides some tiny bivalve Pisidium sp. (Figure 14G), with the latter being the main indicator for pre-Bode River floods.

Figure 14. Late Pleistocene mollusks from layers 3 to 5 of Westeregeln. (A). Pupilla muscorum (layer 4, WE2933-741.1). (B). Arianta sp. (layers 3–4, WE2933-672.1). (C). Lymnaea sp. (layers 3–4, WE2933-672.2). (D). Succinea oblonga (layers 3–4, WE2933-672.3). (E). Planorbarius sp. (layer 4, WE2933-671.1). (F). Planorbis planorbis (layers 4–5, WE2933-741.2). (G). Pisidium sp. (layer 4, WE2933-671.2). 1, 3–7: lateral views; 2: dorsal view.

Figure 14. Late Pleistocene mollusks from layers 3 to 5 of Westeregeln. (A). Pupilla muscorum (layer 4, WE2933-741.1). (B). Arianta sp. (layers 3–4, WE2933-672.1). (C). Lymnaea sp. (layers 3–4, WE2933-672.2). (D). Succinea oblonga (layers 3–4, WE2933-672.3). (E). Planorbarius sp. (layer 4, WE2933-671.1). (F). Planorbis planorbis (layers 4–5, WE2933-741.2). (G). Pisidium sp. (layer 4, WE2933-671.2). 1, 3–7: lateral views; 2: dorsal view.

4.2. Amphibians

Amphibian remains are represented only by about 120 frog remains of Bufo sp., Rana sp., and possibly other genera, which are currently difficult to distinguish on the few postcranial materials (Figure 15A,B). The frog remains were mainly found in layer 4 but also in the deposits of layer 5, whereas a single bone was also found at a depth of 730 cm in a gravel layer of layer 9.

4.3. Reptiles

About 80 vertebrae are the only remains of snakes in some cases; those were found in layer 4, and some were even articulated, which proves a more or less autochthonous origin in some cases at least. Possibly only two genera are represented, at least with Columber sp. and Natrix sp. (Figure 15C–E), whereas it remains unclear if it is a water-preference subspecies or not.

4.4. Birds

There are only nine, mostly incomplete, bones of birds, mainly found only in layer 5. One half clavicle was mapped in Planum XIX (square AA-1, find AA-1-2; Figure 15D), but some small pieces of eggshells were found due to sieving in layer 4. Smaller birds are represented by ulna bones (Figure 15H), but most material is from large-sized species (Figure 15F,G,I). The material has not yet been determined taxonomically.

Figure 15. Frog, snake, and micromammal remains from layers 3 to 5 of the sinkhole in Westeregeln, Central Germany. (A). Frog femur from layer 4 (LDA no. WE2933-648c), lateral. (B). Frog tibiofibula from layer 4 (LDA WE2933-649), cranial. (C). Two articulated snake thoracic vertebrae from layer 4 (LDA WE2933-648a), dorsal. (D). Columber sp. snake thoracic vertebrae from layer 4 (LDA WE2933-648b), dorsal. (E). Natrix sp. thoracic vertebra from layer 4 (LDA WE2933-648c), dorsal. (F). Clavicle from a large bird from layer 5, Planum XVIII, square AA-1 (LDA WE2933-1159), dorsal. (G). Half-ulna of a large bird from layer 5, Treppe (LDA WE2933-1141), dorsal. (H). Ulna of a small bird from layer 5, Planum XIX, square A-1 (LDA WE2933-1143), cranial. (I). Incomplete humerus of a large bird from layer 5, Planum XIX, square AA-1 (find AA-1-13, LDA WE2933-1159), cranial. (J). Spermophilus rufescens deformed skull from layer 5, Planum XIII, (Find A0-16, LDA WE2933-483), (Ja). ventral. (Jb). lateral. (K). Alactaga saliens humerus from layer 4 (LDA WE2933-626). (L). Arvicula amphibius skull from layer 4 (LDA no. WE2933-626.1). (M). Arvicula amphibius mandible from layer 4 (LDA WE2933-630.1). (N). Lemmus lemmus mandible from layer 4 (LDA WE2933-630.2). (O). Martes sp. right mandible of a juvenile individual from layer 4 (LDA WE2933-490), (Oa). lateral, (Ob). dorsal.

Figure 15. Frog, snake, and micromammal remains from layers 3 to 5 of the sinkhole in Westeregeln, Central Germany. (A). Frog femur from layer 4 (LDA no. WE2933-648c), lateral. (B). Frog tibiofibula from layer 4 (LDA WE2933-649), cranial. (C). Two articulated snake thoracic vertebrae from layer 4 (LDA WE2933-648a), dorsal. (D). Columber sp. snake thoracic vertebrae from layer 4 (LDA WE2933-648b), dorsal. (E). Natrix sp. thoracic vertebra from layer 4 (LDA WE2933-648c), dorsal. (F). Clavicle from a large bird from layer 5, Planum XVIII, square AA-1 (LDA WE2933-1159), dorsal. (G). Half-ulna of a large bird from layer 5, Treppe (LDA WE2933-1141), dorsal. (H). Ulna of a small bird from layer 5, Planum XIX, square A-1 (LDA WE2933-1143), cranial. (I). Incomplete humerus of a large bird from layer 5, Planum XIX, square AA-1 (find AA-1-13, LDA WE2933-1159), cranial. (J). Spermophilus rufescens deformed skull from layer 5, Planum XIII, (Find A0-16, LDA WE2933-483), (Ja). ventral. (Jb). lateral. (K). Alactaga saliens humerus from layer 4 (LDA WE2933-626). (L). Arvicula amphibius skull from layer 4 (LDA no. WE2933-626.1). (M). Arvicula amphibius mandible from layer 4 (LDA WE2933-630.1). (N). Lemmus lemmus mandible from layer 4 (LDA WE2933-630.2). (O). Martes sp. right mandible of a juvenile individual from layer 4 (LDA WE2933-490), (Oa). lateral, (Ob). dorsal.

4.5. Micromammals

Those are not represented in large amounts, but about 300 bones and teeth (few jaws or cranial fragments) were sieved only from layers 4 and 5. Quantitative analyses for paleoenvironment reconstructions are difficult, but a qualitative one can be presented for the climate and fluvial influence interpretation. The largest find is a deformed and incomplete skull of the ground squirrel Spermophilus rufescens (Figure 15J) from Planum XIII in layer 5 close to the hyena fecal place 1 (Figure 13C). A humerus (Figure 15K) is from the typical cold-climate-adapted Alactaga saliens. Another skull, consisting mainly of the maxillaries and isolated lower jaws and some single teeth from layers 4/5, can be attributed to the water rat Arvicula amphibius (Figure 15L,M). Most isolated lower jaws and teeth are from other cold-fauna mice such as Lemmus lemmus (Figure 15N) or water-related Microtus gregalis.

4.6. Megamammals

The problem of the megafauna remains is their taphonomic situation. The bones often have plant rhizome surface dissolutions, which destroyed cut and bite mark information in many cases. The main problem is the attached caliche, which often cannot be removed without bone surface damage. Not all bone fragments can be determined to their exact bone origin or the genus. The bone fragmentation degree, calculated on 143 megamammal bones and teeth, is very high at 98%. Distal leg horse (Equus ferus cf. przewalskii) bones are the only few complete bones, such as one astragalus from layer 4, square D (Figure 12B and Figure 17H), or vertebral column remains, such as a cervical vertebra of a steppe bison Bison priscus (Figure 17A). In general, longbone material and even larger ribs seem not to have been fragmented by water transport, which can be proven by the sharp breakage surfaces, corners, and non-rounded bone material at all. Those bones must have been fragmented by humans (Neanderthals) or hyenas (Ice Age spotted hyenas). A single black-colored bone fragment shows burning contact (from a fireplace), but not dehydration.

The total count of megafauna from layers 3–5 is 143 (104 non-determined fragments and 39 determined bones). The bones can be attributed as belonging to the herbivorous Mammuthus primigenius (NISP = 11; Figure 16E–Q), Coelodonta antiquitatis antiquitatis (NISP = 6; Figure 16A–D), Bison priscus (NISP = 5; Figure 17A,B), Equus ferus cf. przewalskii (NISP = 12; Figure 17C–I), and Rangifer tarandus (NISP = 3; Figure 17J,K). The carnivores are present with bone material only from Martes sp. (NISP = 2). The hyenas Crocuta crocuta spelaea are not proven from this excavation by their bones but with two fecal places of about 19 pellets (Figure 18A,B). The micromammal pellet is of unclear predator bird origin. This small number of bones and especially the few determinable bones are not a good base for a quantitative analysis but they allow first trend estimations and first comparisons between hyena- and human-accumulated bone assemblages, and with the former historically selected bones, especially many hyena remains; cf. [4,18].

4.7. Hyena Coprolites

The excavations delivered, for the first time in Germany, hyena coprolites in situ, and those were at two documented fecal places in two different elevations of Plana IX and XIII and layers 3 and 4. The first fecal place 1 (Planum IX; Figure 5C) had many single pellets, which originate from pellet aggregates. Therefore, those are disarticulated and no longer100% autochthonous. The weathered coprolite samples did not deliver small bone fragments rounded or perforated by stomach acid, and the presence of spongiosa in the sieve residuals seem to belong to the rib fragment spongiosa. In one case, a rib was on a coprolite but not embedded. The high number of pellets and gypsum blocks possibly weathered by the urine acid nearby would indicate a more or less autochthonous fecal place, such as the second one. At the second fecal place 2 (Planum XIII; Figure 14C), which was found directly below the first place, but 56 cm deeper in squares A–B (cf. Figure 13C), two smaller coprolites were saved complete (Figure 18B); the others were again only sampled but again did not deliver small swallowed bone fragments. Other pellet finds consist of two attached pellets, which were found between the river gravels 640 cm deep in the “Treppe” area (Figure 18A), indicating a short-event transportation by water.

Figure 16. Megafauna remains of (A–G) woolly rhinoceros Coeleodonta antiquitatis antiquitatis and (G–Q) woolly mammoth Mammuthus primigenius from layers 3 to 5 of Westeregeln. (A–D). Selected fragments of a weathered skull from the section (layer 3, LDA no. WE2933-475). (A). Occipital and condyle, lateral. (B). Temporal arch, lateral. (C). Molar fragment, lateral. (D). Premolar fragment, lateral. (E). M² of a high adult animal from the “artifact gravels” (Planum XVIII, square AA-1, find AA-1-2, layer 5, LDA WE2933-1155). (F). Femur proximal joint (layer 4, LDA WE2933-347). (G–Q). Smashed (or cracked) rib fragments from the hyena fecal place 1 (Planum IX, squares A0-B0, layer 5, cf. Figure 5C). (H). Anterior rib fragment from the section wall (layer 5, LDA WE2933-476). (I). Middle rib fragment from the section wall (layer 5, LDA WE2933-344a). (J). Middle rib fragment (square A0, find A1-9, layer 5, LDA WE2933-372). (K). Middle rib fragment from the section wall (layer 5, LDA WE2933-344b). (L). Middle rib fragment (square A0, find A1-7, layer 5, LDA WE2933-374). (M). Middle rib fragment (square A0, find A1-3, layer 5, LDFA WE2933-379). (N). Middle rib fragment from the section wall (layer 5, LDA WE2933-480). (O). Middle rib fragment (square A0, find A1-8, layer 5, LDA WE2933-373). (P). Middle rib fragment (square A0, find A1-1, layer 5, LDA WE2933-367). (Q). Middle rib fragment (square A0, find A1-6, layer 5, LDA WE2933-375).

Figure 16. Megafauna remains of (A–G) woolly rhinoceros Coeleodonta antiquitatis antiquitatis and (G–Q) woolly mammoth Mammuthus primigenius from layers 3 to 5 of Westeregeln. (A–D). Selected fragments of a weathered skull from the section (layer 3, LDA no. WE2933-475). (A). Occipital and condyle, lateral. (B). Temporal arch, lateral. (C). Molar fragment, lateral. (D). Premolar fragment, lateral. (E). M² of a high adult animal from the “artifact gravels” (Planum XVIII, square AA-1, find AA-1-2, layer 5, LDA WE2933-1155). (F). Femur proximal joint (layer 4, LDA WE2933-347). (G–Q). Smashed (or cracked) rib fragments from the hyena fecal place 1 (Planum IX, squares A0-B0, layer 5, cf. Figure 5C). (H). Anterior rib fragment from the section wall (layer 5, LDA WE2933-476). (I). Middle rib fragment from the section wall (layer 5, LDA WE2933-344a). (J). Middle rib fragment (square A0, find A1-9, layer 5, LDA WE2933-372). (K). Middle rib fragment from the section wall (layer 5, LDA WE2933-344b). (L). Middle rib fragment (square A0, find A1-7, layer 5, LDA WE2933-374). (M). Middle rib fragment (square A0, find A1-3, layer 5, LDFA WE2933-379). (N). Middle rib fragment from the section wall (layer 5, LDA WE2933-480). (O). Middle rib fragment (square A0, find A1-8, layer 5, LDA WE2933-373). (P). Middle rib fragment (square A0, find A1-1, layer 5, LDA WE2933-367). (Q). Middle rib fragment (square A0, find A1-6, layer 5, LDA WE2933-375).

Figure 17. Megafauna steppe bison, horse, and reindeer remains from layers 3 to 5 of Westeregeln. (A). Bison priscus middle cervical vertebra from the section wall (layer 4, LDA no. WE2933-1154), cranial, lateral left. (B). Bison priscus lower jaw molar of a very old animal from the section wall (layer 4, LDA WE2933-366). (C). Equus ferus subsp. (cf. przewalskii) upper jaw I¹ of a very old animal (Planum XVI, square AA-1, layer 5, LDA WE2933-1193), lateral. (D). Equus ferus subsp. (cf. przewalskii) upper jaw M³ of an adult animal from the section wall (layer 3, LDA WE2933-689), labial. (E). Equus ferus subsp. (cf. przewalskii) mandible M₂ (Planum V, layer 4, find E1-12, LDA WE2933-356), labial. (F). Equus ferus subsp. (cf. przewalskii) distal half humerus (Treppe, layer 5, LDA WE2933-549), ventral. (G). Equus ferus subsp. (cf. przewalskii) middle costa collected from the section wall (layer 4, LDA WE2933-336), lateral. (H). Equus ferus subsp. (cf. przewalskii) astragal (Planum VIII, layer 4, find D1-1, LDA WE2933-380), dorsal. (I). Equus ferus subsp. (cf. przewalskii) half-phalanx I (Treppe, layer 4, LDA WE2933-335), dorsal. (J). Rangifer tarandus tibia fragment (Planum III, layer 3, find C1-3, LDA WE2933-349), lateral. (K). Rangifer tarandus mandible M₃ of a calf (Planum VIII, layer 4, find D1-2, LDA WE2933-331), labial.

Figure 17. Megafauna steppe bison, horse, and reindeer remains from layers 3 to 5 of Westeregeln. (A). Bison priscus middle cervical vertebra from the section wall (layer 4, LDA no. WE2933-1154), cranial, lateral left. (B). Bison priscus lower jaw molar of a very old animal from the section wall (layer 4, LDA WE2933-366). (C). Equus ferus subsp. (cf. przewalskii) upper jaw I¹ of a very old animal (Planum XVI, square AA-1, layer 5, LDA WE2933-1193), lateral. (D). Equus ferus subsp. (cf. przewalskii) upper jaw M³ of an adult animal from the section wall (layer 3, LDA WE2933-689), labial. (E). Equus ferus subsp. (cf. przewalskii) mandible M₂ (Planum V, layer 4, find E1-12, LDA WE2933-356), labial. (F). Equus ferus subsp. (cf. przewalskii) distal half humerus (Treppe, layer 5, LDA WE2933-549), ventral. (G). Equus ferus subsp. (cf. przewalskii) middle costa collected from the section wall (layer 4, LDA WE2933-336), lateral. (H). Equus ferus subsp. (cf. przewalskii) astragal (Planum VIII, layer 4, find D1-1, LDA WE2933-380), dorsal. (I). Equus ferus subsp. (cf. przewalskii) half-phalanx I (Treppe, layer 4, LDA WE2933-335), dorsal. (J). Rangifer tarandus tibia fragment (Planum III, layer 3, find C1-3, LDA WE2933-349), lateral. (K). Rangifer tarandus mandible M₃ of a calf (Planum VIII, layer 4, find D1-2, LDA WE2933-331), labial.

Figure 18. Phosphatic hyena excrements (=coprolites) from layers 4 to5 of Westeregeln. (A). Pellet aggregate (layer 5, LDA WE2933-477). (B). Single pellet from the hyena fecal place 2 (Planum XIII, layer 5, find A1-2, LDA WE2933-597). (C). Fox or owl pellet with many micromammal bones (Planum XVI, square AA1, LDA 1194).

Figure 18. Phosphatic hyena excrements (=coprolites) from layers 4 to5 of Westeregeln. (A). Pellet aggregate (layer 5, LDA WE2933-477). (B). Single pellet from the hyena fecal place 2 (Planum XIII, layer 5, find A1-2, LDA WE2933-597). (C). Fox or owl pellet with many micromammal bones (Planum XVI, square AA1, LDA 1194).

5. Middle Paleolithic Stone Artifacts

5.1. Stone Tool Raw Material

It is evident that several flakes (Figure 19) are mostly incomplete with ancient breaking damage (Figure 19A). For the most part, the damages seem to be explainable by frost corrosion, because some of the flint nodules have frost cracks, which make them today (and even in the Paleolithic) unusable for any tool production. In one artifact case, such damage happened most probably by cryoturbation. In some cases, the damage had taken place after artifact production, and worked pieces themselves are fragmented in more than 50% of all cases (Figure 19A). The strong artifact damage is of natural and not human origin, at least in most cases.

5.2. Artifact Distribution

In more than 99%, the raw material is a light darkbrown–black transparent flintstone (Figure 19A), which is typical of Upper Cretaceous Coniacian–Santonian transparent flints. The raw material can be found partly in the neighboring gravels, which originate from Saalian glacigenic material deposited in the region. It could have been collected easily by Neanderthals in the river terraces surrounding the gypsum hill. Two of the handaxes (both MTA triangular flat types; Figure 20B,C) are made of grayish Upper Cretaceous Campanian middle-gray inhomogeneous non-transparent flint. Those two handaxes were most probably imported as tools and might reflect migration directions or the use of raw material and import from a more distant source in the north/north-eastern German direction. A single yellowish flake seems to be a yellow jasper type, whereas only three quartzite flakes (two very small ones and a large one; see Figure 19H) represent a fourth raw material.

Figure 19. Artifact types and raw material from layers 3–5 at the Middle Paleolithic site Westeregeln, Central Germany. (A). Brownish and partly patinated Flinstone flakes, typically fragmented from the gravels of layer 5 (sieved from Planum XVII-XIX, Treppe; LDA WE 2933-847). (B). Very lightly patinated flake from square AA-1, Planum XIX, layer 5 (Find AA-1-1, LDA WE 2933-931). (C). Lightly patinated flake from layer 3 (LDA WE 2933-686). (D). Strongly patinated flake from square A-1, Planum XVIII, layer 5 (Find A-1-4, LDA WE 2933-763). (E). Strongly patinated flake from square AA-1, Planum XVI, layer 5 (Find AA-1-1, LDA WE 2933-762). (F). Strongly patinated flake from the bottom of the fox den burrow in square A1, Planum XX, layer 5 (LDA WE 2933-915). (G). Thick-blade-like flake from square AA-1, Planum XIX, layer 5 (LDA WE 2933-926). (H). Large quartzite flake from layer 3 (LDA WE 2933-368; see also Figure 13D). (I). Jasper flake from Planum XIX, square A-1, layer 5 (Find A-1-11, LDA WE 2933-853).

Figure 19. Artifact types and raw material from layers 3–5 at the Middle Paleolithic site Westeregeln, Central Germany. (A). Brownish and partly patinated Flinstone flakes, typically fragmented from the gravels of layer 5 (sieved from Planum XVII-XIX, Treppe; LDA WE 2933-847). (B). Very lightly patinated flake from square AA-1, Planum XIX, layer 5 (Find AA-1-1, LDA WE 2933-931). (C). Lightly patinated flake from layer 3 (LDA WE 2933-686). (D). Strongly patinated flake from square A-1, Planum XVIII, layer 5 (Find A-1-4, LDA WE 2933-763). (E). Strongly patinated flake from square AA-1, Planum XVI, layer 5 (Find AA-1-1, LDA WE 2933-762). (F). Strongly patinated flake from the bottom of the fox den burrow in square A1, Planum XX, layer 5 (LDA WE 2933-915). (G). Thick-blade-like flake from square AA-1, Planum XIX, layer 5 (LDA WE 2933-926). (H). Large quartzite flake from layer 3 (LDA WE 2933-368; see also Figure 13D). (I). Jasper flake from Planum XIX, square A-1, layer 5 (Find A-1-11, LDA WE 2933-853).

A large number of the artifacts were collected from the dump in front of the sinkhole fill, whereas the stratigraphic result proves their origin in artifact layers 3–5 only. After some finds of some modern reworked larger flakes at the bottom of a rabbit’s/fox’s den (Figure 3), all made in layer 5, the main origin of the non-stratified finds might be attributed primarily to layer 5. This is the layer that is richest in artifacts, perhaps including the one MTB triangular flat handaxe find (Figure 20C). Many artifacts were excavated and are in stratigraphic and positional context, but most of the small flakes and fragments (2–8 mm) were found while the sediments of the excavated squares were sieved (Figure 19A). Comparing the Westeregeln artifacts with pieces from other Middle Paleolithic sites of Central Germany, the sieving technique used in Westeregeln with its special results—a large amount of the smallest flakes– should be taken into consideration for further analyses of tool production techniques.

5.3. Stone Tool Debris

In total, 808 Middle Paleolithic pieces were excavated and collected/excavated/sieved in 2009/2010. There are only a few cores. Those are also mostly fragmented, also due to frost damage. The only well-preserved one belongs to the Levalloisian nucleus type. A large amount of flakes (chips—often very small flake pieces) are in more than 50% of cases fragmented (frost damage). For a similarly large number of “wrecked pieces”/thermal debris, it is difficult to decide if they have been brought to the site as “manuports” by Neanderthals. The rather large flake pieces may be classified as “microlithic industry”. Blades are not represented. It is not possible to reconstruct an intentional systematic blade production, although several pieces have a length/width ratio of more than 2. There are no pieces with parallel ridges between the dorsal negatives struck from the same direction. The artifacts can be classified as being of only Middle Paleolithic character in all three artifact layers 3–5.

5.4. Bifacial Stone Tools

In Westeregeln, there are at least four large tools, three handaxes and one bifacial flaked knife, but all are from different outcrop parts and are therefore not necessarily from the same Middle Paleolithic settlement period.

Acheuléenne-handaxe type: A bifacial non-flat worked Acheuléenne handaxe type tool (Figure 20A) was discovered (pers. com. Geisler) northwest of the “Alte Ziegelei” (Figure 1), where a clay quarry was opened by the brickyard management in the 1990s. In one of the well-developed ice polygons, the handaxe was found together with a number of small flakes. This handaxe is 123 mm long, 67 mm broad, and relatively thick at 39 mm (Figure 20A)—as it is elongated and even relatively thick, it possibly belongs to a “Micoquid” settlement phase. The raw material is a transparent brownish material, with a flint type resembling most of the flakes from the sinkhole fill artifact material.

Micoque-handaxe type: In the 1950s, in the Westeregeln (pers. com. Breithaupt), a “Mousterian” MTA triangular and flat bifacial worked handaxe type (Figure 20B) was found in the gypsum karst, which was discovered in a quarry during that time.

Mousterian MTA-triangular handaxe type: In 2001, a third, smaller, MTB triangular flat retouched biface type was found (Figure 20C) in an ice wedge (pers. com. Wächter), where additionally three fragments of face-retouched flake tools were saved from the same clay pit. From a typological point of view, another handaxe, which is 105 mm long and 97 mm wide (Figure 20B), seems to be a more or less triangular or “cordiform” biface with flat face retouch. Similar is the last handaxe left during the same settlement event, and they are more or less of similar age (Figure 20C); this last handaxe is also similarly worked and in non-transparent gray flint.

Bifacially flaked knife: A single coarse bifacially flaked knife (Figure 20D) seems to be of “microlithic character”.

Figure 20. Handaxes and backed knives from Westeregeln (cf. Figure 1). (A). Acheuléennehandaxe type from the ice polygons northwest of the Alte Ziegelei in the clay pit area (coll. Breithaupt, LDA without no.). (B). Micoque-handaxe type (MTA) of non-patinated, non-transparent gray flint from an ice wedge found in 2001 in the small Bunter clay pit near the Alte Ziegelei (coll. Geisler, LDA without no.). (C). Small Micoque-handaxe type (MTA) of patinated flint from the dump in front of the sinkhole (collection Wächter, LDA without no.). (D). Backflaked knifeof patinated flint from layer 3 of the sinkhole (LDA no. WE 2933-270).

Figure 20. Handaxes and backed knives from Westeregeln (cf. Figure 1). (A). Acheuléennehandaxe type from the ice polygons northwest of the Alte Ziegelei in the clay pit area (coll. Breithaupt, LDA without no.). (B). Micoque-handaxe type (MTA) of non-patinated, non-transparent gray flint from an ice wedge found in 2001 in the small Bunter clay pit near the Alte Ziegelei (coll. Geisler, LDA without no.). (C). Small Micoque-handaxe type (MTA) of patinated flint from the dump in front of the sinkhole (collection Wächter, LDA without no.). (D). Backflaked knifeof patinated flint from layer 3 of the sinkhole (LDA no. WE 2933-270).

6. Discussion

6.1. Sinkhole Sediment Filling History and Stratigraphy

The general fill of the sinkhole can be explained by the redeposition of sediments of the surrounding marine Oligocene (e.g., near Wolmirsleben/Egeln-Nord and Schneidlingen) and Middle Pleistocene Elsterian and Saalian fluvial, glacigenic, and glacifluvial sediments [21], from which most of the sand and even typical fossils were transported into the sinkhole. The typical Upper Oligocene mollusk remains found in the deposits of layers 3–9 are dated by comparison to fossil material from the Leipziger (cf. [22]) and Osnabrücker Meeressande (cf. [23]). The few rounded shell fragment corners and complete marine snails indicate short-distance transport.

Other regional materials of the sinkhole sediments are the Tertiary freshwater quartzite blocks (cf. [1]) and metamorphic and magmatic rocks from the Harz Mountain transported by the Pre-Bode River into its terrace sediments surrounding the Gypsum Karst, already starting in the Elsterian. Scandinavian magmatic/metamorphic rocks, fossils, and flintstones from the Upper Cretaceous have already been deposited by the Elsterian and Saalian glaciers, and were finally reworked in the Weichselian. The mixed gravels were washed into gravel layers and ice wedges together with local artifacts and bones inside the sinkhole and outside in the gypsum karst area surroundings.

The Middle Paleolithic stone artifacts are important, with different finds in layers 4 and 5. The artifact technology, including the bifacial flaked knife find of layer 4, would indicate a rather early Weichselian Age comparable with other inventories from the “Upper Pleistocene technocomplex”, especially the finds from Petersberg near Halle/Saale, Central Germany [24,25]. Layer 3, a periglacial soil, is a marker bed in the region, which is dated to the beginning of the Maximum Glaciation of the Late Weichselian/Wuermian, in which period large ice polygon soils and ice wedges were formed, described from the Pleistocene soils overlaying the Lower Buntsandstein sediments of the Marl pit of Westeregeln north of the gypsum karst; cf. [26].

6.2. C¹⁴ Dating

One of the Coelodonta antiquitatis antiquitatis rib fragments from layer 4, Planum IX (WE 2933-367), has been dated as 16,847 ± 316 BP (KIA-40740: 1σ 18,350–17,817 and 17,747–17,627; 2σ 18,750–17,421 BP). A hyena coprolite, WE 2933-596, from the same stratigraphic situation, was dated as a result of a second attempt of carbon extraction as 14,100 ± 70 BP (KIA-40326: 1σ 15,076–14,662; 2σ 15,246–14,426 BP);cf. [24]. Both dates are much too late for their stratigraphic position (together with the “bifacial flaked knife” in layer 4) clearly below the (probably LGM) loess. Those were impregnated by calcium carbonate from the above loess and have partly been encrusted by caliche. The data seem to reflect the age of the movement of carbonates from the loess into the underlying layers 3–5 and encrustation/impregnation, whose age between 14,100 and 16,847 corresponds well to an expected periglacial loess (layer 2) accumulation time.

In 1993, during a small rescue excavation some 300 m southeast of the sinkhole described here, the uppermost layers of a karst cavity (70 cm deep) contained a fragment of a Coelodonta antiquitatis antiquitatis bone compacta, which has been dated as 50,310+1580/−1320 BP (KIA 10337; [25,26]). This date correlated well with the stratigraphic dating of the mammoth steppe megafauna into the early-to-middle Late Pleistocene periglacial landscape (cf. [27]) and Neanderthal artifact record.

6.3. Mollusks as Paleoenvironmental Indicators

Whereas Nehring [15] mentioned that mollusks were rare in Westeregeln, this seems to result from the non-sieving of the sediments. Mollusks are generally common but are often lost during sieving by floating or due to the breakage of their fragile shells. The water bivalves Pisidium sp. (obtusale/stewarti), which are found only in streams (cf. [15,28]), argue for cool water conditions of the floating water during layer 4–5 periods.

The terrestrial Succinea sp., which lives close to water environments [29], seems to have dropped and floated to the sinkhole. Another terrestrial snail is the loess-snail Pupilla muscorum (a typical grass steppe and cold climate indicator), which was mostly found around both hyena fecal places, but also rarely in layers 3–5. The relationship between those snails and the excrement places is obvious but unclear in the interpretation. Finally, another terrestrial snail, Arianta sp. [29], seems to have lived in the gypsum karst. Vallonia tenuilabris from layers 3–5 represents a clear land snail from a cold climate [29].

6.4. Micromammals as Paleoenvironmental Indicators

The newly excavated and sieved microfauna, which is quite rare, consist of different faunal compositions. In layers 3–4, there is a dominance of snake (even articulated vertebrae) and frog remains (Figure 15A–E). Such enrichments of snake and frog remains in karstic regions in the Tertiary or Pleistocene are also known from other sites and are interpreted as hibernation places [30,31,32]. Snakes and frogs must have used the Westergeln gypsum karst cavities or depressions for hibernating, which is also supported by the bone record of articulated snake vertebral column pieces and well-preserved frog bones.

The taphonomy and interpretation of micromammals is even more complex. After his excavations, Nehring found articulated skeletons of Spermophilus rufescens (Nehring 1876) (cf. “S. altaicus” [11]) which had their burrows in the gypsum karst sediments; cf. [15]. A skull and lower jaw from the sinkhole excavation would be congruent with this. The other micromammal remains are not as abundant in the sediments as Nehring described for his sites in the karstic areas;cf. [15]. This medium-sized burrowing animal is a cold period indicator and was found at different loess sections of the Late Pleistocene of Germany or the Czech Republic, e.g. [15,33,34]. Other micromammals of the sinkhole sediments have not yet been studied in detail, but Nehring [15] listed specimens of a cold period fauna from Westeregeln, namely, Lagomys pusillus, Arctomys bobac, Alactaga saliens, Arvicula amphibius, Arvicola gregalis, Arvicula ratticeps, Plecotus auritus, Lemus lemmus, and Microtus torquatus, and others can be partly recognized on the cranial material from the sinkhole sediments: Spermophilus refuscens, Alactaga saliens, Arvicula amphibius, Lemmus lemmus, Microtus torquatus, and Arvicola gregaris. Finally, a single micromammal pellet was sieved from layer 5 (Figure 18C), which must have been left by a fox, most probably. The origin of micromammals is highly complex and seems to be a mixture of fox (?owl) pellet remains and autochthonous remains. Such a mixture does not allow for simple paleoclimatologic and environmental interpretation as that given by Nehring; cf. [11,15]. Prey selection by foxes (or owls) could give a false picture of the surrounding landscape. The general interpretation of the surroundings as a “mammoth steppe” landscape by Nehring [15] is agreed upon, including the megafauna information presented here and in another paper dealing with the historical bone finds [18].

6.5. The Megafauna—Hyena Den Versus Human Import

The megafauna of Westeregeln from historical (cf. [18]) and new finds presented here has the following biodiversity: the megaherbivorous Mammuthus primigenius, Coelodonta antiquitatis antiquitatis, Bison priscus, Equus ferus przewalskii, and Rangifer tarandus, and the carnivores Ursus spelaeus subsp., Panthera leo spelaea, Crocuta crocuta spelaea, Canis lupus cf. spelaeus, Meles meles, Vulpes vulpes, Alopex lagopus, and Martes sp.;cf. [4].

Nehring [11] found or bought many horse remains from Westeregeln, whereas experiments with this old material demonstrate that, in some cases, there must have been complete legs of E. f. cf. przewalskii, which are figured and reinterpreted as horse body part prey remains of specialized horse killers imported by hyenas;cf. [4,18]. By contrast, this and also possible originally articulated woolly rhinoceros legs were not found in the excavation of the sinkhole. All long bones are strongly fragmented (Figure 16 and Figure 17), and teeth of C. a. antiquitatis, B. priscus, E. f. cf. przewalskii, and R. tarandus appear isolated; only vertebrae are more or less complete (B. priscus vertebrae). The longbones and even the mammoth rib fragments have more smashing (Figure 16D–Q) than “carnivore cracking” characters and seem to be products mostly (or only) of human activities. One horse phalanx I is cracked in the median line (Figure 17I), with similar remains typically left by Late Paleolithic Late Magdalenians in several places; e.g., they are well figured for the Teufelsbrücke and Kniegrotte sites in Thuringia (cf. [35,36]), where humans smashed those massive phalangeal bones which split mostly median into two halves. In hyena dens, such damaged horse phalanges are never preserved as such and are in most cases untouched, for example, all the lower legs (metapodial to phalanx 3), because they are not of feeding interest for hyenas, like all distal leg elements [4,18]. No bone fragments have any characteristic structures of hyena activities (no chewing or bite marks, crushing structures, or zigzag margins, which are typical hyena gnaw marks and bone damages [18]), nor does the half horse humerus (Figure 17F), which also seems to have been smashed by humans compared to other human/hyena bone assemblage studies;cf. [37,38,39]. The herbivorous megafauna remains (C. a. antiquitatis, B. priscus, E. f. cf. przewalskii, and R. tarandus) of the sinkhole fill of layers 3–5 must be all or mainly attributed to human import activities, even if puncture marks of tools are not well documented as a result of the surface dissolution or caliche encrustations. The comparison of the historical and more recent findings in hyena dens attributed faunal composition [18] to the new material from the sinkhole, underlining this result. Typical at hyena dens (in most cases) are the high amounts of remains of the carnivore itself [40], but this was revised to be only typical at commuting/communal den sites; cf. [18].

Figure 21. Reconstruction of the Westeregeln sinkhole fill history, hyena/Neanderthal presence, and bone statistics.

Figure 21. Reconstruction of the Westeregeln sinkhole fill history, hyena/Neanderthal presence, and bone statistics.

The Westeregeln (Central Germany) locality is not the first to be discussed as both a hyena den and a human campsite. In southern England, a Middle Paleolithic (Moustérian) open-air site was similarly described [41]. At the Balve Cave (northwestern Germany, Sauerland Karst), another famous European large cave portal locality, the share of a cave bear, hyena den, and Middle-to-Late Paleolithic use over the last Late Saalian (Achéeuleen) to Late Pleistocene (Magdalenian) time frame was found, with repeated periodical hyena den usage in different periods of the Late Pleistocene [42]. At such sites, the bones can be partly separated according to their accumulation origin, human or not, which is in further discussion at Pleistocene localities [37,38]. Those often mixed assemblages can be separated better with the “European Ice Age spotted hyena project”, which focuses on non-human impact hyena den sites in Central Europe [42,43,44,45,46,47,48,49,50,51,52,53,54,55].

6.6. Hyena Fecal Places Around Smashed Ribs

Hyena den marking by partly enormous amounts of phosphate pellets is well known for the modern African spotted hyenas, e.g. [56], and for the Late Pleistocene Ice Age spotted hyenas, e.g., [40,43]. The two hyena fecal places are at different elevations but both are in layer 5. The pellets have an autochthonous character and are associated with many rib fragments of M. primigenius (Figure 13 and Figure 16D–Q). Similar fecal places must have been present at the open-air loess den site Bad Wildungen, a hyena prey depot and cub-raising den site where many woolly rhinoceros postcranial bones were found with chewing marks but not cracked ribs; cf. [44,46]. When the coprolites are not reworked, as documented by their fragility, then hyenas must have left those after the flooding and deposition and under dry sinkhole conditions. For hyenas, the fragmented woolly mammoth ribs associated with the coprolites, but also the artifacts, are unusual. Those rib fragments do not have dissolution structures and are therefore not swallowed and undigested bone fragments, which was the first interpretation. It seems, rather, that the ribs must have been smashed by humans, and hyenas subsequently marked their den around them.

6.7. Artifacts

The artifacts consist mainly of Coniacian–Santonian brownish transparent flintstone, which was used especially in the Mesolithic but partly also in the Neolithic in northern Germany. Only a few quartzite flakes (Figure 19H) and a single jasper (Figure 19I) flake seem to result from importation, such as two handaxes made of Campanian non-transparent gray flintstone. Those two latter Moustérian-type handaxes indicate a north/north-eastern relationship. The stone artifacts can be integrated in the Late Pleistocene glacial (Weichselian) Middle Paleolithic sites known in Central Germany [57]. Those show an affinity to early Late Pleistocene inventories; cf. [57]. The inventories from the Central German Lichtenberg (Wendland, Lower Saxony) and Oppurg-Gamsenberg (Orla Valley, Thuringia) are dated to the “Odderade Interstadial” [58]. The dating of the “Königsaue Interstadial” with several larger backflaked knivesis, after Mania and Toepfer [59], the oldest Weichselian/Wuermian interstadial record in Saxony-Anhalt, dated to the Brörup Interstadial, whereas Jöris [60,61] favors a parallelization of this “KönigsaueIb interstadial” with the Greenland interstadial 21 (perhaps Odderade) possibly 60 ka ago. The artifacts from the earliest early Late Pleistocene from the lake shores of Neumark-Nord Lake 1, with several backflaked knives (all smaller), are dated to the earliest Late Pleistocene. The backflaked knife of Westeregeln can be classified as “Bocksteinmesser” of the KMG C group dated into the MIS 3, around 60 ka (cf. [60,61]), and comparable and similar sized tools, even such small bifacially flaked knives as were excavated in the Early Weichselian/Wuermian (early Late Pleistocene) Neumark-Nord 2 site [62].

Figure 22. (A). Megafauna assemblages at cave and open air hyena den and Neanderthal camp sites along the Bode River. (B). Weichselian Ice Cap maximum extension and Middle Paleolithic Neanderthal camp sites (and Ice Age spotted hyena dens) in the Harz Mountain region along the rivers. Seasonal model for summer, autumn, and winter hunting camps and game specialization hunting.

Figure 22. (A). Megafauna assemblages at cave and open air hyena den and Neanderthal camp sites along the Bode River. (B). Weichselian Ice Cap maximum extension and Middle Paleolithic Neanderthal camp sites (and Ice Age spotted hyena dens) in the Harz Mountain region along the rivers. Seasonal model for summer, autumn, and winter hunting camps and game specialization hunting.

The handaxes are in three different shapes, namely, an Acheuléenne type, a Mousterien MTA type, and a Micoque type, all well known for Central Europe, e.g. [63], whereas the shape depends more on their resharpening; cf. [64,65,66].

6.8. Seasonal Summer and Winter Camp Migration Model

Westeregeln is an important locality to understand many other similar and older-aged Neanderthal camp sites, which are in different elevated topographic positions but mostly along the main rivers running down from the Harz Mountain ridge into the lowlands (Figure 22). General summer and winter camp models have already been presented for French Middle Paleolithic Neanderthal sites.

There, winter camps were proposed at lowland river terraces and summer camps in the mountain regions, especially in cave entrances/rock shelters [67,68]. These models did not include “cave bear hunt” possibilities but rather big game mountain hunts (Ibex, Cervus, etc.). Following these models, the Baumann’s Cave rock shelter site (Eemian Interglacial, MIS 5e) upstream of the Bode River would reflect a short-term hunting summer camp with a target of valley migratory or mountain game (cf. [69]), but only if cave bear hunts did not take place (which is at any rate expected in winter during their hibernation). In spring, cave bears used the cave entrances to breed and protect their cubs, so conflicts must be expected with Neanderthals, even though the Neanderthals probably did not choose those cave bear den entrances as shelters. If there was a summer campsite at Baumann’s Cave, sensu Binford’s definitions [67], some of the bovid/cervid herbivore prey remains (Bison, Megaloceros, and Cervus) might be attributed to their hunts and a short-term camp, where prey specialization cannot be seen on the few bone materials [69] (Figure 22). At the Late Pleistocene (Weichselian/Wuermian Glacial, MIS 5d-3) Westeregeln site, the overlap of a well-frequented hyena den [4,18] with a Neanderthal camp (discussed herein) makes it hard to present human prey bone (or “kitchen rubbish”) NISP amounts. Neanderthals hunted differently in Westeregeln (at least in the studies of the sinkhole). Mostly male reindeer remains were found at the lowland human campsite Salzgitter-Lebenstedt during autumn [70]. At Westeregeln, the guilt spectrum and percentages of the historical and newly excavated bone material are similar to the hyena den Seveckenberge, which is along the pathway of the Neanderthals along the Bode River (cf. [69]) (Figure 22). If the historical material is excluded and only the few excavated materials (NISP = 39) are used, there is a dominance of woolly mammoth, woolly rhinoceros, and horse (cf.

Table 1. Fauna bone NISP from the sinkhole excavation of all layers.

Species	NISP
Mammuthus primigenius	11
Coelodonta antiquitatis antiquitatis	6
Bison priscus	5
Equus ferus cf. przewalskii	12
Rangifer tarandus	3
Martes sp.	2
Micromammals	200
Birds	10
Snakes	65
Anures	45

), but not reindeer. Finally, the older Late Middle Pleistocene (Dömnitz Interglacial, MIS 7) Neanderthal lake site Schöningen, with its famous nine wooden spear finds, and “kitchen rubbish” found along the lake shore [71,72], consists nearly only (90%) of at least 20 individuals of the large horse Equus f. mosbachensis [73], which was followed by red deer and bison [74,75]. Both cave sites, Baumann’s Cave and Unicorn Cave, complete for these possibly similar or slightly earlier time frames the “mountain fauna”, which is present mainly with cave bear-dominated material from those cave bear dens [52,53]. Compiling all those Middle-to-Late Pleistocene Neanderthal sites of northern Germany (north-central Europe), such as Salzgitter-Lebenstedt [70,75], a full-year migration model is presented parallel with the site Westeregeln [69]; there, NISP data preliminarily range from lowlands (river terrace and lake sites) to mountains (cave entrance/rock shelter sites), targeting different game, possibly even cave bears (Figure 22).

7. Conclusions

The Westeregeln gypsum karst of Central Germany was mostly a dry place (after layer 4), situated slightly above the Pre-Bode River terraces. Periodical flooding (layers 5–9) was responsible for larger sinkhole fill starting from the bottom up to layer 5, which finally ended in solifluctation in layer 3, and at the end, in wind loss sediment trap in layer 2. A larger sinkhole, cut historically in the quarry of Berling that also delivered historical hyena and other megafauna remains between 1848 and 1878, was excavated from 2009 to 2010 in a 7.5 m deep and 15 m wide cross-section of the sinkhole fill. This and two other depressions explored by geophysics are the most northern ones of the partly destroyed hill oriented to the northern Pre-Bode River terrace sediments of Saalian Age. The filling of the sinkhole first started probably within the Late Saalian. Artifacts were found only above a probably Eemian aged soil layer marker bed, whereas the oldest artifact layers reach 570 cm deep (layer 5) just above this soil layer. The Nordic Flintstone artifacts were found mainly within deposits influenced by flooding (layer 5 gravel and sand sequences). They are often damaged by frost cracks. Sediment deformation in layer 5 resulted further from ice wedges, which are filled up with coarser gravels and several artifacts, as an alluvial sorting effect and washing into the ice wedges. On the top of layer 5, to the solifluctation-dominated layer 4 series, two hyena fecal places were found associated with woolly mammoth rib fragments, which seem to have been smashed by humans. This situation of autochthonous places within the allochthonous sediments must have resulted in the high complexity of the sinkhole fill and use by humans and hyenas. Many coprolites are the only clear record of those carnivores in this material up to now, whereas historical material represents the largest ever known European open-air site population of Crocuta crocuta spelaea (Goldfuss). Nearly all of the megafauna remains of layer 5 consist of bone fragments and isolated teeth and seem to have been flooded by periodic floods into the depression. None of the bones can be clearly referred to hyenas but to human activities, whereas a few bones are even burned by fire, such as some dehydrated flint artifact fragments. The use of the depression during the cold period is also indicated by Spermophilus rufescens articulated skull/lower jaw remains. Microtectonic structures reaching from the post-?Late Saalian layers to layer 5 prove very small tectonic or gravitational activities until layer 4 (=?middle Late Pleistocene). With layer 4, the still more or less horizontal deposited solifluctation-dominated sediments change to reddish sandy loess deposits very rich in Marienglas crystals and medium-sized gypsum blocks. This layer has again fresh water mollusks (but mainly of stagnant water), but now relatively high amounts of serpentes and anura remains. This terrestrial and fluvial mixed invertebrate and vertebrate fauna represents a period of less or even no flooding, and proved the use of the depression as a hibernation area for snakes and frogs. In this layer, charcoal pieces, burned bone fragments, and larger bones of steppe bison or horses were found together with flakes and a bifacial flaked knife, which latter seems to date to one of the first Weichselian interstadials, possibly prior to 50 ka BP. Above this, a paleosoil layer (layer 3) changes in the sediments. Here, even large Tertiary Quartzite blocks and other erratic blocks and gravels are mixed within the ice wedge and polygonal cracked sediments. In those, a fragmented skull of woolly rhinoceros C. a. antiquitatis, some bone fragments, and artifact flakes represent the last Paleolithic finds or hyena prey remains within a paleosoil (periglacial soil in the maximum glaciation). The gypsum hill was used within the Early-to-Middle Late Pleistocene repeatedly by hyenas as a commuting and prey storage den and less as a cub raising site, and by Neanderthals at different times when the typical mammoth steppe megafauna was present in the near surroundings. Perhaps within the maximum cold around 22,000 BP (or even later in the uppermost Weichselian, as data from the Weichselian loess underlying a Magdalenian site near Förderstedt, some 20 km southeast from Westeregeln, probably show), this paleosoil was covered by 1.2 m of thick sterile loess sediments. Those finally became black soils in the Holocene, when this hill was reused by Neolithic/Bronze Age populations, also as a burial place and even as a settlement. What the historical and most recent finds demonstrate is that the hill was used during different periods by humans (no clear guilt dominance on the few NISP), and even at modern times with the brick and gypsum factories. We have only started to understand Westeregeln but it is a very important site along the Bode River, where other camp and hyena den sites prove well the competition for territory and guilt. Westeregeln is placed into a Neanderthal winter campsite, comparable to other sites of a similar age in and around the Harz Mountain Range, that overlaps strongly with an Ice Age spotted hyena den.