Archaeological excavation
Fig 1 Bad Kösen-Lengefeld. Engraving of an animal, probably a woolly rhino. Detail from limestone slab ID 18398 (see Fig. 2 and 3; photo: courtesy Archäologisches Landesmuseum Sachsen-Anhalt, J. Lipták)
Fig2 Limestone slab ID 18398, found 2015 in square 56/33b of the Bad-Kösen-Lengefeld Magdalenian settlement site (photo: Florian Sauer, University of Cologne)
Fig 3 Bad Kösen-Lengefeld. Sketch of engraved lines representing a woolly rhino (red) and further, yet unidentified motives (blue). Black lines represent natural scratches and fissures. Object ID 18398, size 25 x 15 cm, indicated scale 3 cm (sketch: Anja Rüschmann, University of Cologne)
Fig 4 Bad Kösen-Lengefeld. The same detail as Fig. 1, here taken by RTI imaging system, specular enhancement function in 70 single views, combined (RTI image: Sebastian Hageneuer; processing Sebastian Hageneuer, Florian Sauer, University of Cologne)
The large limestone slab bearing the rhino image (Fig. 1-4) was found by the archaeological excavation team on August 19th, 2015, in square Q56/33 and on Planum 2.2.1. At the moment of excavation, the finely engraved image remained invisible, hidden by the dust still covering the surface of the slab (Fig. 5). As it turned out after cleaning, the upside surface of the slab displayed an engraved image (Fig. 1). Because, during excavation, all limestone slabs must be expected possibly yield engravings, the slabs were all processed in the same careful way during excavation. All slabs were superficially dry brushed in situ, photographed as part of a planum map, and located in a 3D positioning system by several data points. After its contour line taken, we removed the slab from the sediment, the lower side carefully dry brushed, and the piece bagged and labeled. After excavation, we carefully washed all slabs in pure water to guarantee the best preservation of the original surfaces. Thousands of slabs are to be processed, and slabs have happened to be inspected in detail some years after excavation, in this case, inspection and drawing of the slab were completed by July 6, 2018 (Fig. 3). Excavations since ongoing until now we were able to collect more information about the neighborhood of the rhino slab in 2019 and 2021 (Fig. 5) and about the age of the Magdalenian settlement (Fig. 6).
Fig 5Bad Kösen-Lengefeld. General site plan of limestone structures with a large cooking place (feature 15) in the central-western Part of the site. The insert shows the limestone slab (cf. Fig. 2), in its original in-situ position and at the moment of excavation, as discovered from the southern fringe of feature 15 (see yellow frame). Yellow stars indicate 14C-samples close-by (see Fig. 6). Site plan by Jürgen Richter/Joel Orrin, arranged by Anja Rüschmann, University of Cologne
Local find context
A Cologne-Erlangen team has carried out archaeological research at the Magdalenian open-air site of Bad Kösen-Lengefeld (Saale River Valley) since 2008 (Richter, Uthmeier & Maier, eds., 2021; Uthmeier & Richter, 2012). Surveys and corings indicated the settlement area covering 110 m2 in total. Excavations uncovered so far 91 m2 (Fig. 5) of what is now the upper occupation horizon, since a second, very limited lower occupation surface occurred during the 2017 excavations.
The geological sequence comprises 9 m of pure Loess (bottom not yet known), with two archaeological horizons in the upper 0.4 m of the sequence (see Uthmeier & Richter, 2012). Whereas the lower archaeological layer seems to appear within one primary deposition phase of Loess, the upper (main) archaeological layer is interlacing with a series of thin solifluction horizons, separating the lower and the upper layer by 30-40 cm of mostly reworked Loess and sandy Loess. Still, the archaeological remnants of the upper (main) layer appear as preserved in the primary position, with hundreds of limestone slabs brought in by humans and representing former settlement features. The primary position of the preserved structures, however, does not imply the completeness of all remnants. A certain lack of small pieces indicates the smaller fraction of lithic finds to have been over-proportionally discharged by low-energy sheet flow events (Richter, Uthmeier & Maier, eds., 2021).
The following observations concern the upper (main) layer (Fig. 5): the northern area of the site (several horse hunting episodes) comprises one well-preserved feature, partially connected with traces of charcoal. Three further stone scatters are visible at the outermost northern periphery, possibly from an earlier occupation phase which would have been exploited by later occupants in order to set up the aforementioned, well-preserved feature of almost quadrangular shape. The northern area of the excavated site displays mostly horse remnants, with parts of the skeletons in anatomical connex. The horse bones indicate on-site killing and dismembering of the animals. Based on the different states of preservation of the stone scatters or structures, we are dealing with one „early horse“ and one „late horse“ occupation phase at the present state of research (Richter, Uthmeier & Maier, eds., 2021).
The central area of the site shows three more well-preserved stone features (Fig. 5). At the south-eastern corner of the central feature, we excavated a complex fireplace (feature 15), consisting of a shallow pit with some quartz pebbles, covered by a large limestone slab bearing about a dozen of quartz pebbles and then surrounded or covered by further limestone slabs. The central feature is surrounded by something like 30 postholes attesting that a tent had been set up here and was possibly renewed 3-5 times, given that 6 to 10 postholes were freshly dug for each new tent. In the same area, the predominant animal prey included horses and reindeer along with many ice foxes, indicating the preferred acquisition of these animals and possible use of their furs.
The southern area (one reindeer hunting episode) is particularly interesting because of one large concentration of limestone slabs, stone tools (many backed bladelets and burins) and reindeer bones (Richter, Uthmeier & Maier, eds., 2021). The representation of body parts would again indicate on-site killing and dismembering of the animals, as already mentioned for the northern area. By contrast, bone preservation is much better in the southern than in the northern area. The space intermediate between the central and southern areas yielded a surprisingly large number of engraved limestone slabs, at least five of them with multiple lines and motives.
Radiocarbon measurements
According to more than twenty 14C-dates produced from the site (Fig. 6), the principal occupation (upper layer) took place around 15,350 cal BP would place Bad Kösen into the consolidation phase of the Central European Magdalenian (Magdalenian V), about 500 years after the expansion phase (Maier, 2015; Küßner, 2010; Küßner & Jäger, 2015; Jöris, 2021). The above mentioned lower layer produced three dates of the same age. There are no differences between the southern (5 values) and northern area (2 values), but dates from the central area (12 values) show a somewhat broader range, beginning with 15,500 cal BP and ending with 15,000 cal BP. This might reflect the redeposition of sediments caused by humans digging pits and by post-occupational refill of the pits. This would mean that the occupation surface (or parts of it) remained exposed for as long as 350 years before solifluction sediments buried the surface. Consequently, the above-mentioned occupation episodes (1) early horse, (2) late horse, and (3) reindeer (three episodes at minimum) would outperform the resolution currently achieved by 14C measurements. Currently, the rhino depiction cannot be tied to one of these occupation phases, and we would estimate 15,350 cal BP as the most probable absolute date connected to the depiction.
Fig 6 Bad Kösen-Lengefeld. MAMS radiocarbon measurements, all from faunal remains found in a stratified context. Dates cluster around 13,400 cal BC (15,350 a cal. BP), including an earlier and a later group of dates, each around 150 a BP distant from the intermediate cluster. All three groups appear as almost synchronous and would match with a short time span of human occupation around 15,350 a cal. BP. Some younger dates (see MAMS-31530 and below) would be explained as either being influenced by bad resolution (cau sed by the 13–12.5 ka BC plateau of atmospheric radiocarbon) or by the C/N ratio being less optimal. MAMS radiocarbon measurements: Ronny Friedrich/Curt-Engelhorn-Centrum für Archäometrie, Mannheim
RTI imaging of the rhino depiction
In order to capture the intricate detail of the engraving, we decided to capture the rhino image with Reflectance Transformation Imaging (RTI). RTI designates a multi-imaging technique using photos of an object taken from a static position with variable lighting. Each photo shows the object at constant position but illuminated from a different direction. These photos get processed into one single digital image file, in which the user can interactively re-light the object from any direction with the help of viewer software (Duffy et al., 2013; Earl et al., 2010). The advantage of this technology is not only the possibility to interactively re-light the digital image of the object but also to apply certain filters that rely on the calculated normals (perpendicular vectors to the objects’ topography) to make the slightest details visible.
To achieve the best results, we used an RTI-dome, a closed hemisphere with pre-installed LED lights, with a diameter of 50 cm to completely exclude any ambient light. We combined the RTI-dome with a mirror-less Nikon Z7 camera, equipped with a 50 mm Nikkor lens. A mirrorless camera can drastically reduce the vibration while taking photos. The 50 mm lens can minimize the distortion of the resulting images. For each dataset, 64 photos were taken, color-corrected, and then processed in the RTIBuilder (v2.0.2) software offered by Cultural Heritage Imaging. We produced four datasets: A complete recording of the slab, and three detailed datasets of the rhino itself.
The detailed datasets of the rhino showed the best results, given the delicacy of most of the engraving, and certain lines becoming only visible when changing the direction of the light within the viewer software. Additionally, we overlayed a specular enhancement filter and combined different views into one single image, that shows the complete engraving very clearly (Fig. 4). The detailed documentation by Reflectance Transformation Imaging has since substantiated that recording of the smallest details on paleolithic limestone slabs can be achieved, resulting in detailed visualizations of delicate engravings otherwise hard to detect.