Back to Top

Total Visitors

Web statistics

Locations of visitors to this page

Research Programme

Table of Contents
Objectives of the research programme

Crimea is an exceptional region for the investigation of late Neanderthal behaviour. Together with Southern Spain it is the region with the youngest Middle Palaeolithic sites: the Middle Palaeolithic of Crimea finds its end during Denekamp Interstadial (e.g. Buran-Kaya III, Level B/B1, Kabazi II, Level A3A) parallel to the occurrence of early Aurignacian occupations in the same region attested at the multi-layered site Siuren 1. The research programme’s aim was not only to describe and compare these late Middle Palaeolithic assemblages which span a time from the Eemian Interglacial till Denekamp Interstadial, but also to reconstruct and explain the evolution of specific types of assemblages (CHABAI ET AL. 2002, 442 ff.). Between 1999 and 2006 the German Research Foundation (Deutsche Forschungsgemeinschaft / DFG) brought forward the research programme “Functional Variability in the late Middle Palaeolithic at the Crimean Peninsula, Ukraine” which was carried out in cooperation between the Institute of Prehistoric Archaeology of the University of Cologne and the Crimean Branch of the Ukrainian Academy of Sciences in collaboration with the Southern Methodist University in Dallas (Prof. A. E. Marks).

The research programme, targets to focus on the functional variability of the ‘material culture’ of late Neanderthals. Research over the last 15 years has emphasized the high variability of Middle Palaeolithic assemblages which is due to the time of occupation, the site function and the annual season. The fundamental cultural interpretation of Middle Palaeolithic variability (“Inventartypen sind Einheiten in Raum und Zeit”; “techno-complexes are temporal and special” units / Bosinski 1967) during the second half of the 20th century produced in many cases contradictions. Before this background, research not only has to describe and to compare the evolution of different techno-complexes, but also to explain it as detailed as possible. By reclining concepts of Systems Theory, the research programme was targeting to put into context all measureable parameters determined by culture and environment. The research programme is not only focused on regional studies of a key area of Palaeolithic research. On a bigger scale it is aiming to point out new perspectives for the Palaeolithic Archaeology. (research aim of the research programme quoted after the application in 1999)
Fig. 1: Map of Crimean Middle Palaeolithic sites of Crimea mentioned in this text (BATAILLE 2010).

Fig. 1: Map of Crimean Middle Palaeolithic sites of Crimea mentioned in this text (BATAILLE 2010).

There are excellent archaeological, environmental and topographical conditions on the Crimean Peninsula to follow the research aim. Not at least, since more than 100 sites with assemblages of late Neanderthals are known until today. On one hand, those sites are distributed over approximately the same height, within the internal second ridge of the Crimean Mountain Massif which falls in southeast-northwest direction from more than 1545 m a.s.l. to only 200 m a.s.l. On the other hand, these sites cluster in two regional “groups” west and east of the river Salgir which crosses the Province capital Simferopol in north-south direction. The “eastern group” is characterized in most cases by a divers hunting fauna whereas the “western group” shows in most cases a limited fauna with a specialized hunt on wild ass (Equus hydruntinus) and Saiga antelope (Saiga tatarica). These palaeo-environmental conditions enable the researchers to obtain information by investigating significant sites of both “regional groups” in order to find answers to the above mentioned research target. Within the research programme those sites were investigated which represent the different cultural entities, the different environmental zones and the resulting ecological areas (eastern and western group) of Crimea (CHABAI ET AL. 2002, 442). At some of those sites excavations were conducted under the direction of V. P. CHABAI and A. I. YEVTUSHENKO. The methods of the research programme had a twofold aim: 1) excavations of new and old sites and 2) analyses of the archaeological material to describe the functional variability of these assemblages by means of attribute analysis carried out by the Ukrainian team and of the new methodological approaches Transformation Analysis and Reconstruction of Operational Chains (‘Arbeitsschritt-Analyse’) carried out by the German team (Methods).

Excavations in the course of the research programme (1999-2006)

The basis of information concerning activities conducted in the course of the DFG-research programme reported in the following, is the unpublished “Final Report 2006” of the research programme written by Th. Uthmeier (2007) and is in parts quoted as literal translation not further indicated by citations.

Excavations were conducted at the six Middle Palaeolithic multi-layered sites Kabazi V, Kabazi II, Chokourcha 1, Sary-Kaya, Karabai I and Karabi Tamchin. Altogether 109 archaeological assemblages embedded in these sequences were excavated during the summer campaigns between June and September of every year (Fig. 1).

Sites Excavation Area in m2 Number of archaeological horizons Number of excavated artefacts
Sites of the “Western Group”
Kabazi V 34 21 308.396
Kabazi II 22 36 1.355
Sites of the “Eastern Group”
Chokourcha I 12 20 9.089
Sary-Kaya 6 6 586
Karabai 1 70 16 3.177
Karabi-Tamchin 24 10 22.829
Total 109 345.432

Tab. 1: Excavation activities in the course of the research programme “Functional variability in the late Middle Paleolithic of Crimea Peninsula, Ukraine” financed by the German Science Foundation (DFG) (after UTHMEIER 2007).

Sondages and more extensive excavations were carried out according to lithological layers. Additional, geological and archaeological layers were subdivided. The excavated sediment was sieved with 5 mm and 1.5 mm screens. All artefacts, pebbles, faunal remains and rock slabs bigger 3 cm were measured three-dimensionally. Together with structures like pits, fireplaces or discolourations those pieces were drawn in a scale of 1:10 in surface and profile maps.

Sites of the ‘Western Group’

In the course of the research programme excavations were carried out at two important multi-layered sites of the ‘Western Group’: Kabazi II and Kabazi V. Both sites are situated at Kabazi Mountain, only 200 m away from each other. Kabazi II and V exhibit stratigraphies which inherit Micoquian and Mousterian occupations (Fig. 1). In Kabazi II Micoquian archaeological layers can be found in the lower units (Unit IIA/4 – VI) and Mousterian archaeological layers in the upper units of the sequence (Units II - IIA). In contrast to that, in Kabazi V interstratifications of Micoquian (Sub-Units III/5, III/2, III/1) and Mousterian (Sub-Unit III/3, Unit IV) archaeological layers could be observed.

Kabazi V

The rock-shelter site Kabazi V is situated above the Kalinovaya Balka, a tributary of the river Alma at the eastern part of the Kabazi Massiv. In 1983 the site, which was exposed in the course of construction works, was discovered by Yu. ZAITSEV. In 1986 a test pit and following excavations of 12 m2 dimensions were carried out under the direction of Yu. KOLOSOV and V. CHABAI without reaching whether the rock-shelter’s back wall neither the bed rock (YEVTUSHENKO 1998). In the course of the excavations a 1 m grid-system was established and orthogonally oriented towards the visible wall of the cliff. During the 1990 excavation campaign the grid system had to be changed since it became clear that the back wall of the rock-shelter, which was localized about two meters away from the original test pit, was oriented in an angle of 90° to the cliff (CHABAI 2007).

During the campaign of 1986 the sequence was separated within four geological units which were separated from each other by layers of debris from different rock fall episodes. Within these units few horizons with Middle Palaeolithic artefacts were recognized. According to their setting within the geological units these horizons were separated into four archaeological horizons. According to the Middle Paleolithic nomenclature of this time the three upper layers were associated with the Starosele industry (today Starosele facie of the Crimean Micoquian). The assemblage of the lowest recognized layer was too small for a secure techno-typological assignment.

In 1990 the excavations were continued under the direction of V. P. CHABAI and A. I. YEVTUSHENKO. The presence of four geological units could be verified. The both upper units (Units I and II) were further separated and supplemented by horizons I/A and II/A. Unit II was subdivided in accordance with different lenses of stone artefacts and faunal remains into different archaeological levels which were subdivided by sterile strata from each other (sub-units II/3, II/4, II/4a). Due to the presence of different structures like fire places and pits these sub-units were interpreted as in situ (YEVTUSHENKO 1998, 273). In the course of the same campaign the rock shelter’s back wall and the bed rock were reached.

In the years 1993 and 1995 the excavations were continued under the direction of A. I. YEVTUSHENKO (YEVTUSHENKO 1998a, 274). During these field seasons the trench of the 1990 campaign was deepened in order to properly document the known lower Units II and III and to collect samples in order to achieve material for age determination. Furthermore, a small area of the 1986 campaign was excavated. Altogether seven additional archaeological horizons were exposed (Sub-Units II/4a, II/7, III/1, III/1a, III/2, III/3 and III/4). In contrast to earlier assumptions, some archaeological layers were recognized as the result of post-sedimentary processes. The layers of Units I, I/A and II/A as well as the upper section of Unit II (Sub-units II/1 and II/2) within the main excavation area were disturbed by slope erosion.

In the course of the DFG-research program further excavations were carried out between 2002 and 2003 in cooperation between the Crimean Branch of the Academy of Sciences of Ukraine and the Institute of Prehistoric Archaeology of the University of Cologne. The excavations were directed by V. P. CHABAI and A. I. YEVTUSHENKO. The excavation target in the course of the DFG-research program was the exposure of bigger connected excavation surfaces and artefact clusters. Altogether, 34 m2 were opened and excavated down to the bed rock at a depth of 7 m.

The geological and lithological history can be summarized as follows. The site formation was strongly influenced by two phases of rock fall. Before the first rock fall episode the natural inclination was too steep for anthropogenic usage. After a first phase of accumulation of collapsed rocks and debris a part of the surface was levelled – afterwards it transformed into the intensely frequented area in front of the rock-shelter. The roof of that rock-shelter collapsed and a second phase of deposition of debris destroyed this area. Archaeological horizons with artefacts deposited in primary position were solely found between those later on partially altered boulders of those two phases of rock fall. The lower section below the debris of the first rock fall episode was sterile. In the upper section of the stratigraphy 10 lithological layers could be determined consisting of colluvial sediments and thus bearing only re-deposited artefacts. 21 of the archaeological horizons have a thickness between 10 and 12 cm and stretch across the complete area in front of the rock-shelter (AH III/1B, III/1, III/1A, III/2, III/5-3B1, III/5-B2).

Those sediments bear between 6.000 and 47.000 artefacts. Fireplaces were fed by wood or mammoth bones. Another important class of features are pits. The high number of artefacts and the presence of latent structures speak for intensive occupational occurrences. In comparing those observations with detailed archaeozoological investigations at least a part of the excavated horizons more rich in archaeological finds turned out to be palimpsests of several occupations. For instance, diverging states of preservation and times of death for the two most numerous taxa of Level III/2 Saiga tatarica and Equus hydruntinus could be attested (PATOU-MATHIS 2007, 105f.). This evidence proves two different hunting events during two different annual seasons – one in connection with hunt on Saiga antelope and another event connected with the hunt on Equides. Archaeozoological investigations were conducted on inventories of Units III and IV which inherit assemblages of both Crimean Micoquian and Western Crimean Mousterian industries (PATOU-MATHIS 2007, 97ff.). To shortly summarize the results, usually the body parts of high nutritional value stemming from animals killed and butchered pre-site were brought to Kabazi V.

It is likely that the more thick archaeological layers are the result of recurrent occupations, while thin artefact lenses represent single events which were separated from each other by thin sterile sediments (UTHMEIER 2007). These small lenses are characterized by clear borders of artefact distribution and a weak thickness (1-2 cm). Nevertheless, pits were features of those ephemeral artefact clusters. The pit of Level III/4-2 is very particular: the remains from the production of a bifacial tool were found here. A nearly complete refitting of the production process consisting of numerous chips and flakes was undertaken by A. VESELSKY. This refitting made possible a replica of the missing foliate (VESELSKI 2008a).

In Kabazi V archaeological horizons bearing assemblages attributable whether to Crimean Micoquian or Western Crimean Mousterian industry are interstratified: Crimean Micoquian assemblages occur within Sub-units III/1, III/2 and III/5, while Western Crimean Mousterian inventories were localized in Sub-unit III/3 and in Unit IV (Sub-units IV/1, IV/2, IV/3 and IV/4) (VESELSKY 2008b, CHABAI 2008c, DEMIDENKO 2008, YEVTUSHENKO 2008, CHABAI 2008b). At the same time most archaeological horizons of Sub-Units III and IV of Kabazi V show affinities with both industries. In many archaeological horizons both techno-complexes occur at the same occupational surface (e.g. Sub-units III/7, Unit IV) (VESELSKY 2008c, 391 ff., CHABAI 2008b, 421f.). This led to the conclusion that some of these layers are the result of palimpsests by Mousterian and Micoquian occupations (CHABAI 2008b, 421f.). Other archaeological horizons, like the two layers of Sub-unit III/6 were exposed to natural taphonomical processes which caused vertical mixing. The co-occurrence of Mousterian and Micoquian attributes was interpreted as “the remnants of a number of eroded Micoquian and WCM occupations” (CHABAI 2008a, 374 f.).

First absolute dates have been obtained in 1996/8: RINK ET AL. measured samples by ESR and MSUS and McKINNEY measured samples by U-series. RINK ET AL. gave age determinations for horizon III/1 of 26-32 ka BP (94301 IIA, 94303 IIA, 94303 IIIA) and for the stratigraphical lower horizon III/1A of <41 ka BP (94301 IIIA) (RINK ET AL. 1998, 339). In contrast to these dates, McKINNEY’s age determinations of four tooth enamels coming from layer III/1 obtained by Uran-Thorium gave a much older age: 73.3±6.0 ka BP (59-300-E-1; 210-300-E-3; 64-304-E-3; 214-312-E-3; 202-311-E3) (McKINNEY 1998, 351). During field campaign 2004 a series of samples was taken for OSL, TL and AMS determination in the course of the Natural Environment Research Council’s EFCHED initiative which gave different results (HOUSLEY ET AL. 2007, 51 ff.). In general, Luminiscence data for Unit III gave ages between 60 ka and 100 ka BP, the ages for Unit IV range around 200 ka BP.

The dating of the site can be summarized as follows (UTHMEIER 2007):

  • The sedimentation of most archaeological horizons occurred under Stadial conditions which encompass a phase with Interstadial conditions (AH Level III/5-3B1).
  • Absolute dates correlate the warm phase with Hengelo-Interstadial.
Archaeological Horizon (AH) Number of Artefacts Geological Horizon (GH) Classification
Level III/1B 10.555 Geological Horizon (GH) 12 Crimean Micoquian
Level III/1 47.884
Level III/1A 22.668
Level III/1C 4.173
Level III/2 10.046
Level III/2A 2.786
Level III/3-1 4.032 Western Crimean Mousterian
Level III/3-1A 2.108
Level III/3-1B 260
Level III/3-1C 232
Level III/3-1D 226
Level III/3-2 1.990
Level III/3-2A 612
Level III/3-3 2.653
Level III/3-3o 188
Level III/3A 3.666
Level III/4-1 13.094 Crimean Micoquian /
Western Crimean Mousterian
Level III/4-2 10.779
Level III/4-3 5.911
Level III/4-4 5.851
Level III/4-5 7.646
Level III/4-6 5.119
Level III/5 619
Level III/5-1A 9.745
Level III/5-1 10.236
Level III/5-1B 5.702
Level III/5-2-1 1.808 Geological Horizon (GH) 12A Crimean Micoquian
Level III/5-2 15.154
Level III/5-3 25.801
Level III/5-3B 18.247
Level III/5-3B1 3.465
Level III/5-3B2 14.135
Level III/6-1-2 7.740 Geological Horizon (GH) 14
Level III/6-3 1.474
Level III/7-1 3.809
Level III/7-2 4.264
Level III/7-3 250
Level 3RF 1.006
Level IV/1 12.994 Geological Horizon (GH) 14A Western Crimean Mousterian
Level IV/2 5.488
Level IV/3 3.808
Level IV/4 172
Total 308.396

Tab. 2: Kabazi V. Summary of artefacts in archaeological horizons (AH) and their classification (after UTHMEIER 2007).

Kabazi II

The open-air site Kabazi II is situated at a height of 90 m a.s.l. at the right bank of the Alma River at the southern slope of the Kabazi Massive. This cuesta belongs to the second ridge of the Crimean Mountains. “The site is on the upper part of the slope, 90 m above the Alma River channel, and 70 m from a limestone cliff which towers 33 m over the site” (Chabai 2005, 1f.).

Its stratigraphy, which is the longest Upper Pleistocene one in Eastern Europe, was accumulated by series of rock-falls, alluvial and pedogenic processes. During the Last Interglacial a nearly 12 m high limestone block collapsed from the towering rock wall. This boulder was deposited outside the southernmost edge of the later accumulated artefact concentrations where it functioned as sedimental trap. Due to that, between rock wall and bolder a 14 m deep stratigraphy of sediments was preserved (CHABAI 2005, 4ff.). At a height of 280 m and 270 m a.s.l the remnants of a third terrace of the Pleistocene Alma River could be determined, while the lowest occupation of Kabazi II, which was not affected by natural dislocation processes, was discovered at an elevation of 290.9 m (Chabai 2005, 1f.). The lowermost archaeological horizons (Unit VI) could be found in sediments belonging to Eemian Intarglacial (MIS 5d). Due to its high temporal depth (Eemian Interglacial till Denekamp Interstadial), the deposition of various in-situ assemblages and the good preservation of faunal and palynological remains Kabazi II is among the most instructive Middle Palaeolithic sites in Europe. The stratigraphy has its base within the fluvial sediments of the Alma in a depth of minus 14 m. Within 26 lithopedological strata 76 archaeological levels could be determined, 55 of them in primary position (CHABAI 2005, CHABAI 2006).

First sondages were carried out in 1986. The following excavation campaigns were continued until 1994 by the Academy of Sciences of Ukraine in cooperation with different partners. In the course of the excavation campaigns 1993 till1995 the surface was expanded to 22 m2. In the course of the research program new excavations started in 1999. At this point of time the upper section with geological horizons 2 till 10 had already been excavated over 60 m2. 22 archaeological horizons were embedded within those geological strata. Except few layers, which were classified as belonging to the Crimean Micoquian situated in the lower section of Unit IIA, all discovered assemblages of Units II and IIA were attributed to the Western Crimean Mousterian. Additionally, some test-pits were dug up-slope in order to clarify the situation of the youngest layers.

The excavations done in the course of the research program led to the following results (after UTHMEIER 2007):

  1. The archaeological sequence ends with Denekamp Interstadial. Younger Pleistocene colluviums could not be trapped anymore by the now completely covered limestone block.
  2. The archaeological sequence starts at the end of the last Interglacial.
  3. Altogether 31 archaeological horizons in primary position could be documented within Units III, IV, V and IV (Tab. 3).
  4. A change is attestable within the archaeological sequence of Kabazi II: unifacial assemblages of the Western Crimean Mousterian are represented in the upper section of the sequence – this facie first occurs in Kabazi II during Hosselo Stadial before 45.000 uncal. BP (Level IIA/2). The older assemblages of the lower section can be attributed to the Crimean Micoquian characterized by bifacial technology – in Kabazi II this facie starts in Eemian Interglacial and ends in Moershoofd Interstadial.
  5. Most of the archaeological layers yield the remains of killing-butchering sites for the dismemberment of hunted Equus hydruntinus individuals. During the glacial periods complete family herds were killed and meat bearing parts were exported to near-by residential camps. Beside traces of dismemberment of hunting fauna only the oldest layers in the Eemian period militate in favour for the on-site consumption of parts of wild ass individuals. For the Western Crimean Mousterian Level II/8 (Hengelo Interstadial) the presence of a fireplace beside the excavated area was assumed – this could indicate on-site consumption of game (PATOU-MATHIS 2006, 37ff.).
Archaeological Horizon (AH) Number of Artefacts Pollen Classification
Level III/4 47 Rederstall (?), Pryluki pl1b2-b1 Crimean Micoquian
Level III/5 70
Level III/6 90
Level III/7 79
Level III/8 55
Level III/8A 57 Brörup, Saint-Germain I, Pryluki pl1b1
Level III/8B 51
Level III/8C 41
Level III/8D 34
Level III/8E 33
Level IV/1 redeposited not classifiable
Level IV/2
Level IV/3
Level IV/4
Level IV/5
Level V/1 268 Tyasmin, Herning Stadial Crimean Micoquian
Level V/2 87
Level V/3 17 Eemian (E6b)
Level V/4 10
Level V/5 17
Level V/6 28
Level VI/1 17 Eemian (E6a), Kaydaky kd3b2+c
Level VI/2 15
Level VI/3 21
Level VI/4 13
Level VI/5 10
Level VI/6 39
Level VI/7 17
Level VI/8 25
Level VI/9 35
Level VI/9A 30
Level VI/10 13
Level VI/11-14 124
Level VI/15 7
Level VI/16 3
Level VI/17 2
Total 1.355

Tab. 3: Kabazi II. Number of artefacts within the archaeological horizons (AH), and their temporal and vegetational classification (After Uthmeier 2007).

Sites of the ‘Eastern Group’

Four Middle Palaeolithic sites investigated in the course of the research program belong to the ‘eastern group’: Chokourcha I, Sary-Kaya, Karabai and Karabi Tamchin. Due to economical and ecological questions the multi-layered site Karabi Tamchin is of special interest, since it belongs to the few examples of sites situated within the main ridge of the Crimean Mountains at an altitude of around 750 m a.s.l. Moreover, this site is quite remote from water and lithic raw material sources what has implications on the composition of lithic and faunal assemblages.

Chokourcha I

The rockshelter site Chokourcha I is situated in the north-eastern suburbs of the Crimean provincial capital Simferopol. It is located at the foot of a 20 m high limestone cliff in the Second Ridge of the Crimean Mountains. The rockshelter faces to the north about 8 m above the present level of the Malyi Salgir, a small tributary of the River Salgir which crosses Simferopol. A relatively large area of 30 x 40 m is located in front of the rockshelter. The rock shelter’s back wall is situated approximately 75 m away from the present river (CHABAI 2004, 343ff.).

Intensive excavations have already been carried out in the first half of the 20th century by N. L. ERNST and by B. I. TATARINOV. The site was first investigated by the local amateur archaeologist S. I. ZABNIN in 1927. He excavated a test pit in which he found archaeological remains (ZABNIN 1928, after CHABAI 2004, 344f.). The excavation of 1928-31 carried out under the direction of N. L. ERNST showed intensive occupations in the area in front of the cave entrance. He distinguished five archaeological horizons with altogether three Middle Palaeolithic horizons (Layers 2, 3, and 4). The upper layer contained Holocene Sediments with four archaeological horizons attributed to Bronze Age and the medieval time. The layer below the Bronze Age layer and the uppermost Middle Palaeolithic horizons of Layer 2 was sterile. ERNST subdivided four Pleistocene layers (Layers 2-5). The diverging degree of ashy remains within layers 2, 3 and 4 was the base for the subdivision of those layers which were all separated by sterile deposits of “cemented silt” (CHABAI 2004, 344f.). The lowermost Layer 5 was described as free of artefacts. Outside the rock shelter area more pronounced numbers of big limestone blocks could be documented. “One of them, found ‘on the border between Layers 3 and 4’, which covered about 50 square meters of the site area (Ernst 1934:189, 190). (...) Based on the 50 m2 block and the cliff wall configuration, Ernst decided that the Chokourcha I shelter had a quite different shape before the series of rockfalls which destroyed it” (CHABAI 2004, 344f.). Due to that, ERNST concluded that the site originally must have been a cave.

Unfortunately, apart from the preliminary report mentioned above, nothing is known about these extensive excavations since the documentation as well as all faunal remains and most of the lithic artefacts got lost during the Second World War (CHABAI 2004, 343f.). However, the schematic profile drawings of the three Pleistocene layers subdivided by ERNST (1934) indicate features like fireplaces and lenses that have not been recognized during the excavation. Due to that, in 1996 a test pit with an extension of 12 m2 was established within the preserved sediments. During this campaign, only the lowermost layer of Ernst’s excavations was still preserved. During the following excavation campaign, this profile section could be subdivided into 24 geological layers, bearing 21 archaeological horizons (Tab 4). The high stratigraphical resolution is due to recurrent flooding events by the nearby situated river, which caused a regular covering of the area in front of the cave with fine-grained sandy sediments. At the same time, the presence of numerous fireplaces with preserved charcoal remains speaks for a low current velocity. Due to that, most of the horizons could be found in primary position. All assemblages can be attributed to the Crimean Micoquian. The lithic raw material was brought to the site from sources of 30 km distance. Level IV-O delivered a radiocarbon age of (OxA-10877) > 45.500 BP. The analysis of faunal remains indicates at least one occupation during springtime. The presence of individual game of Saiga tatarica and Equus hydruntinus implies a specialized hunting strategy well known for other sites of Crimea. Nevertheless, important differences can be seen. Saiga antelopes and wild ass show unusual low individual numbers per archaeological layer. At the same time, old animals are overrepresented. Moreover, the preserved skeletal elements of mammoth, red deer and bovid emphasize scavenging of already dead animals as additional nutritional resource.

Archaeological Horizon (AH) Number of Artefacts Classification
Level IV-A 79 Crimean Micoquian
Level IV-A2 82
Level IV-B 442
Level IV-D 52
Level IV-F 908
Level IV-G 264
Level IV-I 2.144
Level IV-I2 264
Level IV-K 89
Level IV-L 88
Level IV-L2 166
Level IV-M 1959
Level IV-N 95
Level IV-O 2.038
Level IV-P 15
Level IV-Q 123
Level IV-S 104
Level IV-T 8
Level IV-U 128
Level IV-V 41
Total 9.089

Tab. 4: Chokurcha I. Overview of artefacts in archaeological horizons (AH) and their vegetatio-historical classification (After Uthmeier 2007).


The open-air station Sary-Kaya is situated at the feet of a steep rock wall. This cliff is part of the Second Ridge of the Crimean Mountains and closes the escarpment (cuesta) to the south. In the course of first excavations under the direction of Y. G. KOLOSOV between 1976-1977 and 1985-1986 a surface of 300 m2 was excavated. In order to test the few unpublished informations concerning stratigraphy and archaeological horizons in summer 2004 a sondage of 6 m2 extension was deepened till a depth of 5 m. Thus, the stratigraphical resolution could be improved. Instead of two, five archaeological horizons could be determined (Tab. 5). The oldest layers (AH Level III, Level IV and Level V) are embedded in sediments, which were exposed to soil formation processes. Probably we are dealing here with the end of the Last Interglacial (MIS 5a). The two younger archaeological horizons (AH Level I and II) were embedded in Loesses, most probably from the beginning of the Last Glacial.

Archaeological Horizon (AH) Number of Artefacts Datation
Level 0 5 Holocene
Level I 247 1. Glacial Maximum
Level II 164
Level III 103 Eemian Interglacial
Level IV 14
Level V 53
Total 586

Tab. 5: Sary Kaya I. Overview of artefacts in archaeological horizons (AH), their datation and classification (After Uthmeier 2007).

Karabai I (Kara-Bey)

The site Karabai I is situated approximately 2 km south-east of the small village Mironovka in the northern part of the second chain of the Crimean Mountains. At the upper border of a dry valley, which today seasonally drains the precipitations from the plateau into the river Bijuk-Karasu, a sequence with Palaeolithic finds was cut by heavy rain falls in 2001. First sondages gave sight onto a spacious area (Karabai I) in front of a big partially collapsed rock shelter (Karabai II). The area in front of the rock shelter was excavated during three summer seasons under the direction of A. I. YEVTUSHENKO and V. P. CHABAI. Due to its wide extension Karabai I can be understood as an open-air site. Two trenches were excavated. The first one, situated in the southern area of a fireplace, was excavated till a depth of -4 m on a surface of altogether 56 m2. A second trench was dug with a maximum depth of 4.5 m and an extension of 14 m2 in the west. The oldest archaeological assemblages were found in Last Interglacial sediments. They are covered by numerous soil horizons alternating with Aeolian deposits. Those altogether 11 archaeological horizons (AH’s Level 2 till 7-29) embedded in Last Interglacial sediments, which are separated by sterile deposits from each other, are of special interest (Tab. 6). The more comprehensive assemblages of Levels 3-2, 4-1, 4-2, 4-3 and 5-1 are characterized by bifacial tools and belong to the Crimean Micoquian. We can assume the same for the smaller assemblages with less numerous surface shaped tools. Where those pieces are missing, the waste of surface shaping gives distinct hints on the presence of bifacial technology. The bifaces were either transported to a not excavated part of the site or exported off-site. Only the assemblages of AH Level 2 can be attributed to the Western Crimean Mousterian.

Archaeological Horizon (AH) Number of Artefacts Excavated area
(each level)
Datation Classification
Level 0 7 56 m2 Holocene n.a.
Level 1 8 Glacial
Level 1A 10
Level 2A 21
Level 2 131 Archaeological horizons
in primary position:
Eemian Interglacial
Western Crimean Mousterian
Level 3-1 184 Crimean Micoquian
Level 3-2 287
Level 4-1 571
Level 4-2 767
Level 4-3 331 40 m2
Level 5-1 474
Level 6-1 210 33 m2
Level 6-2 142
Level 7-1 15  6 m2
Level 7-2 14
Level 7-3 5
Total 3.177

Tab. 6: Karabai I. Diagram showing the sizes of investigated excavated surfaces of archaeological horizons (AH), the numbers of artefacts, the classification and dating of these horizons (After Uthmeier 2007).

Karabi Tamchin

Karabi Tamchin is with 740 m a.s.l. the most elevated Middle Palaeolithic site of Crimea known today. It was discovered by A. I. YEVTUSHENKO and V. P. CHABAI in 1996. There is only one other highland site, admittedly with Palaeolithic artefacts in mixed stratigraphic context: the cave site Adji-Koba. “All other known highland locations are open air find-spots near the meteorological station on Karabi Yaila, and on Yaltinskaya Yaila” (Yevtushenko et al. 2004, 278 f.). Karabi Tamchin is situated at the north-eastern edge of the Karabi Yaila, the largest Plateau of the Main Ridge of Crimean Mountains which is located 30 km east of the Crimean provincial capital Simferopol. Karabi Tamchin is a collapsed rock shelter with original dimensions of 18 m to 6 m. The rock shelter opens to the northwest in the direction of the Adijiskli Valley which is crossed by the Tamchin, a tributary of the Burulcha River. This canyon-like valley is cut into the plateau. The nearest known raw material sources are situated more than 30 km away from the site.

In the same year of discovering, a first sondage of 2 x 1 m extension was dug which brought to light first Middle Palaeolithic artefacts. The sediments could be separated into five lithological strata in which four archaeological horizons were embedded with Middle Palaeolithic and faunal remains (YEVTUSHENKO ET AL. 2004, 278). Between 1999 and 2001 a trench with the extension of altogether 27 m2 was opened in the central part of the site. Thus, most of the formerly rock shelter area could be investigated and a stratigraphy of 1.5 m thickness was recovered.

The stratigraphy bears 10 archaeological horizons, which are usually separated from each other by thin sterile layers of limestone debris. Irrespective the Holocene finds from the recent humus of AH Level 0 there are some small assemblages which, due to processes of mechanical mixing and displacement, cannot be interpreted (Levels I/A, I, II/1 and II/2) (YEVTUSHENKO ET AL. 280ff.). On the contrary, five archaeological horizons gave enough data. Archaeological horizons Level V and Level IV/2 are characterized by bifacial tools and typologically are associated with the Crimean Micoquian industry. Level IV/A, which is embedded between the lowest horizons with artefacts of the Crimean Micoquian, exhibits an assemblage that is attributed to the Western Crimean Mousterian. At the top of the sequence there are inventories belonging to the Western Crimean Mousterian, as well (Levels II/2 and III) (Tab. 7). After Kabazi V, also Karabi Tamchin inherits interstratifications of both important Middle Palaeolithic entities. At Karabi Tamchin we are dealing with functional identical or similar occupations in greater distance to raw material sources. The weak artefact densities in all layers and the composition of the hunting fauna in different horizons, which is characterized by a high diversity in connection with a small number of individuals, militate in favour for small functional differences between these layers. This is interpreted as a result of ephemeral occupations, in connection with the opportunistic hunting events on single animals in the vicinity of the site. The following dismemberment and consumption took place on-site.

The stone artefacts of the undisturbed Levels II/2, III, IV, V/1 and V/2 show similarities concerning the small sizes of blank products and cores. Another common feature is the nearly complete absence of primary cortical flakes and only a few examples of blank product partially covered by cortex (CHABAI ET AL. 2002, 446 ff.). At the same time many tools and resharpening flakes could be observed within all mentioned assemblages. Due to that, it is obvious that at least a part of the tools was produced pre-site, imported and rejuvenated on-site. The similarities between these assemblages can be explained by raw material scarcities – the imported raw material of Karabi Tamchin derives from a source which is located 20-25 km away from the site (CHABAI ET AL. 2002, 446).

Environmental reconstructions by the analysis of molluscs and small mammals indicate a mixed vegetation of open steppe and forest-steppe around the plateau. Different from deeper elevated areas of the Crimean Mountains, the grasses, which were the preferred diets of the herbivores, were, due to the high elevation, present later in the year. Following this argument, occupations more likely occurred at the end of the warm season. Although environmental data show chronological differences, climatic oscillations did never cross the border of the climatic tolerance of the big mammals. For instance, Levels III and IV were deposited under warm Interstadial conditions while Level II/2 was accumulated in a much colder environment. The absolute dating of the archaeological horizons is problematic. Due to the unpronounced thickness of the sediments, data based on ESR and Uran-Thorium could not be obtained. Thus, only 14C dates are available (YEVTUSHENKO ET AL. 2004, 281ff):

  • AH Level III
  • (OxA-10883)
  • >42,400 uncal BP
  • 13C = -19.8%)
  • AH Level IV/2
  • (OxA-10884)
  • >41,200 uncal BP
  • 13C = -20.1%)
  • AH Level V
  • (OxA-11387)
  • >29,800 uncal BP
  • 13C = -21.2%)
  • AH Level II/2
  • (TO-10990)
  • 26,640±210 uncal BP
  • 13C = -25%)
  • AH Level II/2
  • (TO-10991)
  • 28,720±240 uncal BP
  • 13C = -25%)

Due to the low degree of collagen of the bone samples, the dates obtained for Level II/2 are considered as minimum ages. Sediments, molluscs and small mammals speak for an accumulation of Levels II/2, III and IV/2 under Interstadial conditions. For Level II/2 a chronological setting in the Arcy (Denekamp) Interstadial seems to be most appropriate. Respecting sedimentology, Levels III and IV/2 do not belong to the same Interstadial. Nevertheless, the represented small mammal taxa speak for a comparably contemporary deposition of those Levels, possibly during Hengelo Interstadial. In contrast to that, Level V is possibly much older, though it gave a quite young radiocarbon-age: the presence of the primitive small mammal Ellobius talpinus gives hint to a hiatus within the sequence, which possibly reaches the Last Interglacial. Such a long hiatus is indicated by the presence of the weathered surface between the breccias-like horizon 7c, which yields Level V, and the upper layers. Level V can be best settled in a broader time range between the Last Interglacial and one of the Early Glacial Interstadials (Amersfoort, Brörup, Odderade) (YEVTUSHENKO ET AL. 2004, 282f.).

With the Western Crimean Mousterian layers III and IVA, which are dated older 41 ka BP and that fore are chronologically associated with Moershoofd Interglacial or even older, these assemblages possibly represent the oldest known representatives of the Mousterian industry in Crimea - in Kabazi II the WCM starts during the Stadial between Moershoofd and Hengelo Interstadial. The association of the lowest archaeological horizons, belonging to the Crimean Micoquian, is in accordance with the earliest occurrence of this industry in the Kabazi II sequence. Nevertheless, this chronological position has to be verified. Due to its setting within the Crimean highland region Karabi Tamchin is of special importance for Middle Palaeolithic settlement studies.

Archaeological Horizon (AH) Number of
Datation Classification
0 82 Holocene mixed / small amount of artefacts
IA 4
I 78
II/1 56
II/2 1.028 Denekamp Western Crimean Mousterian
III 1.825 Stadial
IV/1 583 Moershoofd small amount of artefacts
IV/2 4.918 Crimean Micoquian
IV/A 1.035 Western Crimean Mousterian
V 13.220 Eemian Interglacial Crimean Micoquian
Total 22.829

Tab. 7: Karabi Tamchin. Diagram showing the sizes of investigated excavated surfaces of archaeological horizons (AH), the numbers of artefacts, the classification and dating of these horizons (After Uthmeier 2007).

Analyses of artefacts: Attribute Analysis, Transformation Analysis (‘Transformations-Analyse’ & Reconstruction of Operational Sequences (‘Arbeitsschritt-Analyse’)

Altogether 345.432 stone artefacts stemming from 109 archaeological horizons were excavated in course of the research program. These artefacts were investigated according to attribute analysis: besides the classification of blank products and tool classes qualitative and quantitative features were taken into consideration (V. P. Chabai, A. I. Yevtuchenko, Y. E. Demidenko 2004, 2005, 2006, 2007, 2008). Additionally, for answering technological questions refittings were made (V. I. Usik 2006 and A, P. Veselsky 2008a). In another analytical step the impact of dynamic processes on the production and discard of stone artefacts was reconstructed by new methodological approaches. This mentioned dynamic has impact on the analysis of stone artefacts on three levels:

  1. Assemblages: there composition is a result of the transport of artefacts and raw material between the raw material sources, the camp sites of the “group” and the places of hunt and collection.
  2. Operational Sequences: the operational sequences of the transformation of raw material, of which all artefacts are part of.
  3. Formal Tools: the tools are not only the result of the intended end-product. Furthermore, they are intensely used artefacts, which are the result of a complex process of different steps of rejuvenation. Those rejuvenations change shape, thickness and the state of the (working) edges, especially of surface shaped (bifacial) tools.

Thus, central aim of interest of the analyses is not the classification of states of deposition, but the causal concepts of land use, raw material transformation and tool use. Methods that help to understand these concepts are Transformation Analyses, Reconstruction of Operational Sequences for raw material transformation and the Reconstruction of Operational Sequences of surface shaped tools (Methods).

Altogether 61 assemblages were analyzed according to these methods. These assemblages stem from the sites Kabazi V, Kabazi II, Starosele, Chokourcha 1, Buran-Kaya III and Kiik-Koba. A general assumption of the results will be given below.

Sites investigated by technological analysis

Selected assemblages of three further sites were investigated in the course of the research program: Buran-Kaya III rock-shelter, Starosele rock-shelter and Kiik-Koba cave. The results obtained by Transformation Analysis and the reconstruction of operational sequences were introduced in the Habilitation of Th. Uthmeier (2006) and the Master work of M. Kurbjuhn (2003).


The rock-shelter site Starosele is situated in the south-west of Crimea near the city Bakchisaray. It is located at the entrance of the steep and narrow, about 1 km long Kanly-Dere Canyon, which cuts into the soft limestone in southward direction. It opens to the bigger west-east directed Bakchisaraiskaya Valley.

This site obtained international attention in the context of the excavations that were conducted by A. A. Formozov between 1952 and 1956 (DEMIDENKO 1998). Altogether 12000 artefacts, 60000 faunal remains and 15 fire places could be observed (CHABAI ET AL. 2002, 464f.). The remains of a Homo sapiens sapiens individual (‘child of Starosele’) found in Middle Palaeolithic context led to a controversial dispute. New excavations between 1993 and 1995 by a joint Ukrainian-American team under the direction of A. E. Marks brought new light on the stratigraphy and the archaeological context. Instead of only two archaeological horizons altogether four could be observed (level 1 at the top till level 4 at the bottom of the sequence). Two further burials of Homo sapiens within level 1 gave way to the intuition, that all finds of human remains belonged to a medieval cemetery which has been deepened into the Pleistocene sediments.

In the course of the new excavations, instead of the two archaeological horizons postulated by FORMOZOV altogether 12 geological layers (Geological Horizons A till F) could be observed which inherited four archaeological horizons (Levels 1 till 4).

The lithic artefacts of Starosele, Levels 1, 2 and 3 were subject to new investigations by the means of Transformation Analysis and the Reconstruction of Operational Chains (‘Arbeitsschritt-Analyse’). Level 1 assemblage was dealt with in the Master thesis of M. KURBJUHN (2003) while the results of Levels 2 and3 were presented in the State Doctorate of Th. UTHMEIER (2006). The results of Level 3 shall be summarized here in short, since only this assemblage could not be associated with one of the known Crimean Middle Palaeolithic industries. The 2337 of this archaeological horizon artefacts stem from an excavation surface of 48 m2 (CHABAI ET AL. 2002, 464 ff.). The good preservation of the associated faunal remains and the presence of different concentrations of lithic artefacts led to the conclusion that the material of Level 3 was deposited in primary position (MARKS &MONIGAL 1998, quoted after CHABAI ET AL. 2002, 464 f.). Chronologically this layer is today attributed into the context of an Early Glacial Stadial (“Frühwürm”) (CHABAI, MARKS &MONIGAL 1999). However the faunal remains are quite diverse, the taxa Equus hydruntinus and Equus caballus are dominating the assemblage. Also Saiga tatarica, Vulpes Crocuta, and Ursus are present. Cervus elaphus, Bos sp., Sus sp. and Rangifer tarandus are represented in only small numbers. The carnivores’ impact on the composition of the faunal assemblage is estimated as low. The high share of horse long bones was that fore interpreted as a result of anthropogenic activities (CHABAI ET AL. 2002, 464).

In order to understand and describe the blank production of Siuren 1, Level 3, which had already been estimated as not classifiable by MARKS and MONIGAL (1998, cited after CHABAI ET AL. 2002, 466), the methodological aims ‘reconstruction of operational chaines’ (‘Arbeitsschrittanalyse’) and Transformation Analysis helped to understand the technological concept of this assemblage (see Methods). The results of the reconstruction of operational chains gave way to the assumption that a special variant of the discoidal method is present. This variant is probably due to the lack of round raw nodules which would deliver enough raw volume to strictly follow this concept. As a result, from some of these raw nodules the discoidal sequences could only be produced at one platform. In other cases flakes were used as cores for a secondary blank production. Investigations of A. BONCH-OSMOLOWSKI lead to the assumption that a similar system of blank production is also present within the lower layer of Kiik-Koba (CHABAI ET AL. 2002, 466f.).

Results of the Transformation Analysis (CHABAI ET AL. 2002, 466-470)

In Level 3 among 481 stone artefacts bigger 3 cm, 48 raw material units (RMUs) could be observed. With help of literature the origin of a part of the raw material could be estimated at the base of varieties.

  1. Local origin: the rock walls below the rockshelter inherit small round nodules with black coarse grained cleavage plains with numerous white inclusions. A brownish-white chalky cortex of 1 cm thickness is characteristic.
  2. Local origin: about 500 m away from the entrance of the Kanly-Dere Canyon nodules are inherited within the rock walls which feature light brown cleavage plains.
  3. Regional origin: about 7 km west of the Starosele site a secondary raw material source is located. The fine grained honey coloured flint can be collected from the river gravels.
  4. Regional origin: Approximately 7 km away of Starosele is the village Partisanskoje, where flint of grey-blue colour can be found. Probably raw material units (RMUs) 12, 26 and 29 derive from this source.
  5. Supra-regional origin: dark grey flat flint nodules with white cortex resemble the Kabazi II raw material. According to the thickness of these plaquettes the might derive either from Alma or Bodrak Valley. In both cases the raw material sources are located about 20 km away from Starosele.

A methodological problem of the Transformation Analysis is the account for complete assemblages. The lithics of Starosele which were analyzed according to this method derive from the 1993-1995 excavations in which course only an incomplete section with the extension of 47 m2 has been excavated. The biggest part (250 m2) has been excavated by A. A. FORMOZOV during the 1950s (DEMIDENKO 1998, after CHABAI ET AL. 2002, 466). Admittedly, the lithic artefacts and faunal remains of the excavations 1993-1995 were gathered around a fire place. According to the excavators’ opinion, the assemblage is the result of one single occupation (MARKS &MONIGAL 1998, after CHABAI ET AL. 2002, 467). Obviously, a concentration of finds has been uncovered. Taking into consideration what is known about the distribution of finds on the surface of the old excavations, similar concentrations might have been present there, as well.

In order to receive an impression on the completeness of workpieces, artefacts belonging to one RMU were mapped horizontally. In cases that artefacts belonging to one RMU are clustering in one concentration of the incomplete excavation surface in combination with the representation of a reasonable operational chain they are assumed as complete units. According to this assumption, 18 of the 48 workpieces and single pieces can be interpreted as complete. For 15 workspieces consisting only of a few pieces, which exhibit no technologically reasonable context, it is possible that further belonging artefacts can be found within the not excavated area (RMU 1, 3, 7, 11, 17-21, 25, 27, 30, 44, 58).The same can be assumed for some of the ‘single pieces’; probably some of the belonging pieces might be found within the material excavated by FORMOZOV. This is especially likely for the unmodified artefacts introduced as ‘single pieces’ (RMUs 55, 57-64).

Differences between those workpieces interpreted as units are refer to the state of the imported nodules. A combination of the distances to raw material sources and the quality of the raw material seems to be an important factor. Beside the workpiece of RMU 5 of less good quality which was collected below the site, also fine grained workpieces deriving from the Kacha Valley 7 km away have been brought to the side as raw nodules (RMUs 35-42). In the course of the reduction according to the above mentioned variant of the discoidal method, in an advanced state the lateral and distal convexities were produced by steep blows in order to obtain last thick flakes with crested remnants.

Such a way of exploitation is lacking within workpieces imported as already decorticated (RMUs 6, 8, 15, 23, 28, 31) or partially reduced (RMUs 10, 14, 16, 22). In spite of the fact that that raw material is of higher quality than for instance RMU 5 these workpieces have not been reduced completely. Most of those RMUs might probably explained by deriving from a remote source, what is indicated by the higher degree of reduction when imported to the site.

In general, nearly all bigger flakes resulting from blank production were modified. The high share of formal tools (“W” = “Werkzeug”) speaks for an intensive usage of the assemblages. This can be investigated by workpieces of ‘good’ and ‘bad’ raw material quality (e. g. RMU 5). The distinct differences in raw material treatment between archaeological assemblages like Kabazi II, Unit III/1 and Starosele are best explained from the point of view of the Transformation Analysis as the result of a longer sojourn at Starosele, Level 3. In this context, the diverging states of imported lithic artefacts may indicate peripheral activities of different durations in the nearer and broader surrounding area:

  1. Not decorticated nodules are brought without an intermediate sojourn from the source to Starosele.
  2. Partially decorticated nodules indicate a stay at the raw material source.
  3. Reduced cores and single pieces derive from stations where blanks were produced and used.

The results of the Transformation Analysis give confirmation to the assumption that Starosele, Level 3 was occupied for a longer time span than the levels of Kabazi II (CHABAI ET AL 2002, 470). In the context of blank production the quality of the used raw material was only one among other criteria. It is possible that simply the need for flakes at a specific point of time was more important than the quality convenient for reduction.


The site Kiik-Koba was discovered within the internal range of the Crimean Mountains by G. A. BOCH-OSMOLOWSKI (1934) in 1924. The rockshelter is located about 25 km east of Simferopol near the small settlement Taou-Kiptshak. At this point a small valley is connected to the entrance of the wider Zouia Valley. By the Zouia Valley the Crimean Mountains are connected with the northern steppe zone stretching only 3 km away. The rockshelter which was formed by erosional processes caused by changing atmospherique conditions faces to the south.

Excavations took place between 1924 and 1926. For the first time, human fossils could be discovered at the territory of the former Soviet Union. Moreover, altogether 15.000 unretouched blanks, 2000 tools, 50 retouchers, dozens of faunal remains with use traces and 680 faunal remains belonging to 27 different taxa were found in the course of these campaigns (UTHMEIER 2006, 203f.).

According to BONCH-OSMOLOWSKI (1940, 177; after UTHMEIER 2006, 205) the only 1 m deep stratigraphy of the rockshelter contained 6 archaeological layers, which could easily be separated from each other. In reality BONCH-OSMOLOWSKI recognized two geological horizons which were further subdivided by archaeological horizons of darker colour: one clayish Holocene stratum with archaeological layers Couche I, Couche IIa and Couche II and one Pleistocene stratum consisting of yellowish debris with archaeological layers Couche III, Couche IV, Couche V and Couche VI (UTHMEIER 2006, 205). Layers IV and VI are of importance for Middle Paleolithic investigations.

Layer VI (‘couche à foyers inférieur’) is the lowest archaeological layer directly covering the bedrock. It extended over 70 m2 within the rockshelter and was also found on the area in front of the rockshelter. This layer contained many lithic artefacts, but badly preserved faunal remains. The upper archaeological layer IV (‘couche à foyers supérieur’) is separated from layer VI by layer V (‘entre foyers’) which contained some displaced artefacts of layer VI.

Layer IV of Kiik-Koba was investigated in the course of the research programm. Due to stratigraphical uncertainties caused by the mixing of artefacts belonging to the upper and lower layer no ‘Transformation Analysis’ was carried out. But workpieces realized as conform with that layer were interpreted as belonging to one assemblage and operational sequences of bifacial pieces were reconstructed (‘Arbeitsschrittanalyse’) (UTHMEIER 2006, 210 ff.).

Buran-Kaya III

The collapsed rockshelter Buran-Kaya III belongs to a group of four rockshelters and caves located at the right bench of the Burulcha River. All sites are situated at a distance of only a few meters from each other. Buran-Kaya III is situated abaout 20 km northeast of Simferopol and 4 km southwest of the small town Aromatnoye. The Burulcha River, which drains in northern direction into the steppe zone, has cut here deep into the Crimean Mountains. The site was occupied approximately 4 meters behind the drip line of the rock shelter’s roof. During time of occupation the roof must have been 3 till 4 meters high. Originally, the rock shelter opened to the 10 m distant located steep river bank (UTHMEIER 2006, 212 f.).

The stations Buran-Kaya I, Buran-Kaya II and Kilse-Kobe have already been discovered in 1936 by O. N. BADER. At these stations first excavations were conducted between 1957 and 1958 and brought to light Mesolithic and Neolithic finds (MONIGAL 2004a, 3, cited after UTHMEIER 2006, 212). In 1990 A. A. YANEVICH discovered Buran-Kaya III. In the same year he carried out first sondage excavations which were continued from 1994 in cooperation with M. YAMADA (YANEVICH, STEPANCHUK &COHEN 1996; YANEVICH 1998; YAMADA 1996; after UTHMEIER 2006, 212). This test pit was enlarged to an extension of 4 m2 and dug till the bed rock in a depth of 3.5 m. In 1994 a 1 m grid-system was established and the excavations were carried out according to artificial spits of 5 till 10 cm. More extensive excavations were carried out under cooperation with A. E. MARKS and M. OTTE in the years 1996, 1997 and 2001 (MARKS &MONIGAL 2000; MONIGAL 2004a, 2004b, 2004c; after UTHMEIER 2006, 212). In 1996 the new excavators changed then excavation method: archaeological horizons were not excavated anymore according to artificial spits. Instead, the sediment was excavated in 1 m grids according to natural strata. The soil was screened with 5 mm screens. All finds bigger 1 cm were measured three-dimensionally and drawn into maps with a scale of 1:10. Find bearing layers were denoted by the letters A till E. On the contrary, sterile strata were not denoted. Since there is no reliable correlation of the layers of the 1990-1996 excavations with the layers of the 1996-2001 campaigns, these assemblages have to be investigated separately (UTHMEIER 2006, 213). Between 1996 and 2001 five layers were excavated: Layers A, B/B1, C, D, and E.

The stratigraphy can be described as follows:

  • Layer 1
  • Middle Ages
  • Layer 2
  • Middle Ages
  • Layer 3
  • Bronze Age
  • Layer3a
  • Neolithic
  • (5.070±40, 5.180±50)
  • Layer 4
  • Late Palaeolithic: Swidrien
  • Layer 4a
  • Late Palaeolithic: Swiderian
  • (5.070±40, 5.180±50)
  • Layer 5 – 6-2
  • "Epi-Gravettian"
  • Layer 6-3 – 6-5
  • Aurignacian / Gravettian?"
  • (30.740±460; 11.900±150; 11.950±130; 28.700±620; 34.400±1.200)
  • Layer A
  • unknown Middle Palaeolithic with bifacial pieces
  • Layer A1
  • sterile
  • Layer B
  • Middle Palaeolithic: Kiik-Koba facie of Crimean Micoquian
  • Layer B1
  • Middle Palaeolithic: Kiik-Koba facie of Crimean Micoquian
  • (28.840±460; 28.520±420)
  • Layer B2
  • sterile
  • Layer C
  • “Initial Upper Palaeolithic”: "Eastern Szelettian"
  • (32.350±700; 32.200±650; 36.700±1.500)
  • Debris of collapsed rockshelter roof
  • Layer D
  • unknown non-Levallois flake-based Middle Palaeolithic assemblage
  • (post-depositionally disturbed)
  • Layer E
  • unknown blade-based Middle Palaeolithic assemblage

While Layer D shows signs of post-depositional disturbance, all other archaeological layers were found in primary position. The 4.5 m deep stratigraphy of Buran-Kaya III is composed of limestone debris which is whether embedded within sandy and silty sediments in the lower section (layers E till 5) or in argil and loam in the upper section (layers 4a till 1). Only the lowest sediments deposited at the bed rock are composed of thin layers of fluviatile sands of orange and yellow-orange colour.

Due to the presence of the Early Upper Palaeolithitc Level C situated below the Middle Palaeolithic Level B/B1 the stratigraphy of Buran-Kaya III is of high importance for the interpretation of the Middle to Upper Palaeolithic Transition. For the first time in Europe, a secure interstratification of Middle and Early Upper Palaeolithic / Transitional industries is attested. While the lower Middle Palaeolithic occupations D and E are too small and at the same time lacking diagnostic for a secure attribution to one of the known Eastern European Middle Paleolithic industries, the assemblage of Level C was paralleled with the Streletskaya industry known from Kostenki 1, Layer 5 (YANEVICH, STEPANCHUK & COHEN 1996; MARLS &MONIGAL 2000; MONIGAL 2004; CHABAI, MARKS &MONIGAL 2004). Furthermore, Level D is not in primary deposition. Contrary to that, the artefacts of Level C come from a thin in-situ surface. The assemblage is characterized by flat bifacial tools, most of the time shaped on plaquettes. The blanks which were used for edge modified pieces are most of the time blanks from surface shaping. A peculiarity are bifacially shaped edge modified trapezoidal microliths (MONIGAL 2004, ). Absolute C14-dates for that layer range between 36 ka BP and 32 ka BP. According to pollenanalytical studies Level C can be settled within the Stadial between Hengelo and Denekamp Interstadial.

In the course of the research program Level B1 was investigated by Transformation Analysis (UTHMEIER 2004). The assemblage of this layer is attributed to the Kiik-Koba facie of the Crimean Micoquian (DEMIDENKO 2004). The archaeological horizons Level B and B1 are separated from upper Level A and lower Level C by the sterile strata A1 and B2. Levels B and B1 can be distinguished from each other by the dark, nearly black colour of Level B1 which is interpreted as the remains of a former fireplace. Due to the lack of visible artefact concentrations, DEMIDENKO (2004, 115) assumed that we are dealing in both layers with former occupational surfaces, in combination with latent structures destroyed in the course of intensive occupational activities. In contrast to that, KURBJUHN (2004) and UTHMEIER (2004; 2006, 216 f.) draw the attention on the presence of different concentrations which were indicated by the mapping of artefacts bigger 3 cm belonging to different workpieces. For a primary position of the artefacts speaks the good preservation of lithic artefacts and faunal remains, as well. The results of the pollen-analysis which settles Levels B and B1 within Denekamp Interstadial are confirmed by two C14 dates of (OxA-6673) 28.840±460 BP and (OxA-6674) 28.520±420 BP (GERASIMENKO 2004, Fig. 2-2; MONIGAL 2004a, Tab. 1-1). Accordingly, Levels B and B1 are among the latest known Middle Palaeolithic assemblages in Europe.

Results: the explanation of the functional variability in the late Middle Palaeolithic of Crimean
according to Uthmeier (2007) (translated by G. Bataille)

The project’s aim of interest was traced by Th. UTHMEIER in the course of his state doctorate (UTHMEIER 2006). The results introduced in the following were presented in this not yet published State Doctorate. The terminus ‘concept reservoir’ (“Konzeptreservoir”) introduced by the German archaeologist W. WEISSMÜLLER is the key term in this context (WEISSMÜLLER 1995). It addresses the knowledge of a group, which was delivered from generation to generation. In context of Palaeolithic research ‘concept reservoirs’ is looked after in lithic production processes. Since in Middle Palaeolithic context no division of labour transcending small family groups and larger aggregations of such family groups is likely, it is assumed that nearly every Middle Palaeolithic adult was firm with the group’s concept reservoir. According to Uthmeier (2006) three concept reservoirs can be distinguished for the Crimean Middle Palaeolithic:

Concept reservoir A: plano-convex surface shaped blanks with symmetric shapes. The assemblages of Kabazi II, Units V and VI belong to concept reservoir A. Plano-convex surface shaped foliates are characteristic. During processes of rework and rejuvenation the shape of the primary tool is preserved. Despite of surface shaping no other concept of blank production could be determined. For the production of simple tools waste from surface shaping/ surface shaping flakes are modified.

Concept reservoir B: Levallois and blade concept, plano-convex surface shaped blanks with asymmetric shape. The assemblages of “Kiik-Koba-industry”, “Ak-Kaya-industry” and “Starosele-industry” that together form the Crimean Micoquian and the assemblages of the Western Crimean Mousterian belong to concept reservoir B. Blank products which derive from cores of Levallois concept or of uni- and bipolar blade concept occur beside plano-convex surface shaped blank products. Among the simple tools according to blank type simple and transversal side scrapers are most common. In the course of advancing time of usage it is possible that simple side scrapers are transformed into double and déjeté side scrapers. The group of surface shaped tools can be summarized in two classes:

  1. Surface shaped tools of different sizes, depending on the sizes of the used raw nodules; often produced on thin plaquettes
  2. Standardized surface shaped items, which are produced on plaquettes, round flat nodules and flakes.

Artefacts of the first group involve ‘backed knives’ and different foliate pieces, including Faustkeilblätter’ and ‘Halbkeile’. These pieces are often discarded in an early stage of reduction. Another possibility is that during reduction processes the smaller secondary forms repeat the primary form – in that way the shape of the initial form is conserved. In context of the second group of hafted items, the conservation of the function over a long time-span is intended. In this context, the pieces are treated similar to flakes with edge modification: after the production of a blank, despite of the ventral thinning all edge modifications are solely made on the artefact’s dorsal surfaces. Since most of the time the right edge is in use and thus asymmetrically reworked, secondary and tertiary forms clearly differ from each other. It can be concluded that at the beginning of a reduction sequence ‘Halbkeile’ are produced which are transformed to ‘blattförmige Schaber’ and ‘Fäustel’.

Concept reservoir C: Here a variant of the discoidal concept for blank production but no concept of surface shaping is present. Up to now, concept reservoir C is only known from Starosele, Level 3. Features are the discoidal concept and an edge modification which is more or less dependent from the blank’s shape (see Industries’).

Concept reservoirs A and C only represent small sections of the former land use spectrum. Up to now, it is not possible to establish a reliable picture about the nature of this segment of the Crimean Middle Palaeolithic land use system. The majority of assemblages investigated in the course of the research program belong to Concept Reservoir B. In that latter case the reconstruction of land use patterns was possible:

  1. Summer months. During the warm season mainly wild ass was hunted in the region of the second range of the Crimean Mountains. Since these animals are increasingly dependent from water sources with increasing draught, there was a good opportunity to predict their presence at specific times. Since places of nutritional resources were easy to estimate, a stabile system of longer frequented camp sites was established, in connection with stations which had the function of resource acquisition by task-groups. It is not out of the question and from a logistical point of view it is sensible that due to the good supply with meat during the warm season two or more core families temporarily spent the time at such a place. The stations were situated either in direct vicinity of the camp site or in a distance of few kilometres. Depending on how many stations in the direct vicinity to the camp could (recurrently) deliver nutrition, the time of occupation of such camp sites probably was up to several weeks. A second limiting factor was the natural carrying capacity of the environment. Probably, there was the need to leave a logistical range and enter into another, after having hunted or banished whole wild ass families (and other taxa). Depending to the distance to the main camp, field camps (Außenlager) can be understood as ephemeral camps with occupations of some hours to maximum few days duration.
  2. Winter months. During wintertime several taxa were hunted in the vicinity of the first mountain range and at the border of the plain extending north of the Crimean Mountains. Reason might be the seasonal migration behaviour of the animals and the worse food supply. It is assumed, that migratory game / steppe species aggregated in lower levels of the Crimean Mountains and at the border to the steppe plain, to survive the winter. But, also the presence of Equus hydruntinus throughout the whole year within the Second Ridge is assumed (BATAILLE 2010). Wild ass, Saiga antelope and in case of nutritional shortages steppe bison were hunted while Mammoth obviously was added to the nutrition by scavenging activities of already dead animals. Generally, it is assumed that during the cold period of the year a similar land use system existed as in summer, with residential camps (Hauptlager) for the whole group and field camps (Nebenlager) respectively locations (Stellen) which were used by task groups for resource acquisition.
  3. Spring / autumn. During spring time people tried to capture herds of migratory game while migrating to their summer ranges. In this course everything in need was transported over greater distances, possibly as long as groups encountered hunting game. Target of such behaviour was not only to overpass the critical time of herd migration but also to reach the summer ranges of the main hunting game Equus hydruntinus. In case it was possible to delay the decampment into early summertime, predominantly Saiga tatarica was hunted shortly after the time of calving. For sure, spring time was from an economical point of view the most critical time; also resources of less quality like old animals and carrion were supplied to diet.

Stone artefacts were preserved in good condition during all annual cycles: at residential camps and most other site types stone artefacts were produced, changed or prepared for usage. In cases that no local raw material was at hand, lithic material was brought to the sites. This is an important observation: stone material was not a limiting factor for the choice of a specific location to establish a camp site. Short termed activities near raw material sources show only scarce or no record of bifacial tools and belong to the Western Crimean Mousterian. Nevertheless, unifacial surface shaping is present and attestable in undisputed Western Crimean assemblages (e. g. Kabazi II, Level II/8) (BATAILLE 2006b, 2010). Occupations without local raw material procurement belong to the Crimean Micoquian, even in cases of only short-term activities.

The land use pattern equates during summer and winter months to the behaviour of modern hunter-gatherers with a logistical pattern (“collectors”). In the past such a pattern was accepted for Upper Palaeolithic modern humans but refused for Neanderthals. First hints for such ‘modern’ behaviour of Neanderthals were known from Southern Germany and gave confirmation that their underestimation in that regard was unjustified (RICHTER 1997). With help of the extraordinary good data base of the Crimean Middle Palaeolithic it is possible to illustrate that much of the innovations attributed to the Upper Palaeolithic already existed and were well-established in the Middle Palaeolithic. Differences existed in part in the kind of equipment that was needed for the land use patterns. Concepts and methods of stone production are of Middle Palaeolithic type in Crimea; no tendency to an Upper Palaeolithic development is observable.

Besides these main results of the research program, which makes Crimea to a case study for the meaning of cultural entities of Neanderthals and its relation to land use patterns, numerous further results and features could be added, which have in parts already been published. Particularly, features like the numerous pits in the main archaeological horizon of Kabazi V are to mention. The site Buran-Kaya III which was accessible for the research program by cooperation with A. A. Yanevich and A. E. Marks an interstratification of an Early Upper Palaeolithic “Streletskaya” horizon with a Kiik-Koba horizon at the top of the sequence was attestable (MONIGAL 2004a, 8f.; MONIGAL 2004b, 57ff.; DEMIDENKO 2004, 113ff.). The Micoquian assemblage (layer B) is comparable with the one of the upper horizon of the eponymous site Kiik-Koba, only 8 km away, which yielded a Neanderthal burial. At several sites the series of TL-dates could be expanded by new datings conducted by the British colleagues D. Sanderson, Ch. Burbidge and R. Housley, financed by the NERC (Natural Environment Research Council). Furthermore, in the lower units of Kabazi II and in the lower layers of Sary-Kaya, Karabi-Tamchin and Karabai I for the first time Last Interglacial occupations could prove evidence in Crimea.

Guido Bataille


BINFORD, L. R. & S. BINFORD (1966) A preliminary analysis of functional variability in the Mousterian of Levallois-facies. American Anthropologist 68, 1966, 238-295.

BONCH-OSMOLOWSKI, G. A. (1934) The results of the investigations in the Crimean Paleolithic. In: The Proceedings of the Second International Congress of the Association for the Quaternary Investigations in Europe 5. Leningrad, Moskau, Nowosibirsk 1934, 114-183. (in Russian).

BORDES, F. (1961) Typologie du paléolithique ancient et moyen. Publications de l’Institut de Préhistorique Francaise 50, 1953, 457.466.

BOSINSKI, G. (1967) Die mittelpaläolithischen Funde im westlichen Mitteleuropa. Köln Graz 1967.

CHABAI, V. P. (1998) The History of Crimean Middle Paleolithic Studies. In: MARKS, A. E. & V. P. CHABAI (eds.) The Middle Paleolithic of Western Crimea, Vol. 1. ERAUL 84. Liège 1998, 1-16.

CHABAI, V. P. (2004) The Middle Paleolithic of Crimea: stratigraphy, chronology, typological variability & Eastern European context. Simferopol 2004. (in Russian).

CHABAI, V. P. & DEMIDENKO, Y. E. (1998) The Classification of Flint Artifacts. In: MARKS, A. E. & V. P. CHABAI (eds.) The Middle Paleolithic of Western Crimea, Vol. 1. ERAUL 84. Liège 1998, 35-38.

CHABAI, V. P. & UTHMEIER, T. (2006). Settlement systems in the Crimean Middle Palaeolithic. In: CHABAI, V. P., RICHTER, J. & Th. UTHMEIER (eds.) Kabazi II: The 70000 Years since the Last Interglacial. Palaeolithic Sites of Crimea. Vol. 2. Simferopol-Cologne 2008, 297-359.

DEMIDENKO, Y. E. (2008) The Early and Mid-Upper Paleolithic of the North Black Sea region: an overview. Quartär 55 (2008), 99-114.

FERRING, C. R. (1998) The Geologic Setting of Mousterian Sites in Western Crimea. In: MARKS, A. E. & V. P. CHABAI (eds.) The Middle Paleolithic of Western Crimea, Vol. 1. ERAUL 84. Liège 1998, 17-30.

FORMOZOV, A. A. (1954) The Starosele site near Bakchisarai, as the place of discovery of a fossil human. Soviet Ethnography 1, 1954, 11-22. (in Russian).

GLADILIN, V. N. (1966) The distinct types of stone industries in the Mousterian oft he Russian Plain and Crimea and their place in the Early Paleolithic oft he USSR. In: RYBAKOV, B. A. (Red.) Proceedings of the VIIth international congress of proto- and prehistorians in Prag. Moskau 1966, 14-18. (in Russian).

GLADILIN, V. N. (1980) The principles and criteria oft the Early Paleolithic subdivision. In: ARTEMENENKO, I. I. (ed.) Prehistoric Archeology – research and finds. Kiew 1980, 21-38. (in Russian).

KLEIN, R. G. (1966) The Mousterian of European Russia. Ph. D. Dissertation, University of Chicago 1966.

KOLOSOV, V. G., STEPANCHUK, V. N. & V. P. CHABAI (1993) The Early Paleolithic of the Crimea. Kiev 1993. (in Russian).

MARKS, A. E. & V. P. CHABAI (2006) Stasis and Change during the Crimean Middle Paleolithic. In: HOVERS, E. & S. L. KUHN (eds.) Transitions Before the Transition. Evolution and Stability in the Middle Paleolithic and Middle Stone Age. Tucson, Arizona 2006, 121-135.

MONIGAL, K. (2004) Introduction to the Site of Buran-Kaya III. In: CHABAI, V. P., K. MONIGAL & A. E. MARKS (eds.) The Middle Paleolithic and Early Upper Paleolithic of Eastern Crimea, Vol. 1. ERAUL 104. Liège 2004, 34-18.

RICHTER, J. (1997) Sesselfelsgrotte III: Der G-Schichten-Komplex der Sesselfelsgrotte – Zum Verständnis des Micoquien. Quartär-Bibliothek 7. Saarbrücken 1997.

UTHMEIER, Th. (2006) Landnutzungsmuster im Mittelpaläolithikum der Halbinsel Krim, Ukraine. Ein Beitrag zu Ökonomie und Soziologie der Neandertaler. Habilitation. Philosophische Fakultät der Universität zu Köln 2006. (not published).

YEVTUSHENKO, A. I. (2004) The Site of Karabi Tamchin: Introduction. In: CHABAI, V. P., K. MONIGAL & A. E. MARKS (eds.) The Middle Paleolithic and Early Upper Paleolithic of Eastern Crimea, Vol. 1. ERAUL 104. Liège 2004, 277-282.