Late Mousterian Persistence near the Arctic Circle

See allHide authors and affiliations

Science  13 May 2011:
Vol. 332, Issue 6031, pp. 841-845
DOI: 10.1126/science.1203866


Palaeolithic sites in Russian high latitudes have been considered as Upper Palaeolithic and thus representing an Arctic expansion of modern humans. Here we show that at Byzovaya, in the western foothills of the Polar Urals, the technological structure of the lithic assemblage makes it directly comparable with Mousterian Middle Palaeolithic industries that so far have been exclusively attributed to the Neandertal populations in Europe. Radiocarbon and optical-stimulated luminescence dates on bones and sand grains indicate that the site was occupied during a short period around 28,500 carbon-14 years before the present (about 31,000 to 34,000 calendar years ago), at the time when only Upper Palaeolithic cultures occupied lower latitudes of Eurasia. Byzovaya may thus represent a late northern refuge for Neandertals, about 1000 km north of earlier known Mousterian sites.

Most of the Russian Arctic was free of glacier ice throughout the past 50,000 years, including during the Last Glacial Maximum (LGM) (1). A varied herbivorous fauna existed in high Arctic areas that are presently wet tundra or almost barren Arctic deserts (2). Recent archaeological evidence demonstrates that Ice Age humans also at least temporarily lived and hunted in these northern landscapes beginning around 35,000 to 36,000 14C years before the present (yr B.P.) [≥40,000 yr B.P. in calibrated/calendar (cal) years] (37) (fig. S1). It has not been clear whether the early visitors were members of a fossil population [such as Homo sapiens neanderthalensis and affiliated groups (8, 9)] or whether modern humans (H. sapiens sapiens) expanded northward into a previously uninhabited area.

This question is related to the Middle Palaeolithic (MP) to Upper Palaeolithic (UP) cultural transition in Eurasia. This transition, which has been considered to have taken place about 40,000 to 37,000 yr B.P. in most of Eurasia, saw the global extinction of the Neandertals and thus the end of their specific MP (Mousterian) culture. The Neandertals were replaced by modern humans, who were the bearers of all known UP cultures.

Here we describe lithic technology and age constraints from the Byzovaya site near the Polar Urals and show that humans bearing MP stone technology persisted to 32,000 to 34,000 cal yr B.P. in the Eurasian Arctic (Fig. 1). Byzovaya, which is among the northernmost known Palaeolithic sites, was previously considered to be an Early Upper Palaeolithic (EUP) site mainly on the basis of a few radiocarbon dates that suggested an age of about 27,000 14C years or younger.

Fig. 1

Map showing the location of the Byzovaya site, close to the northern Urals. Other sites mentioned in the text are also shown. The red stippled line circumscribes the area with known Neandertal sites; here only some selected young sites are marked. The maximum extent of the Eurasian Ice Sheet during the LGM (about 26,000 to 20,000 cal yr B.P.) is also shown (1).

The Byzovaya site (65°01′25′′N, 57°25′12′′E) is located on the right bank of the Pechora River, which flows northward across the lowland areas west of the Ural Mountains (Fig. 1). First described in 1965 by Guslitser et al. (10), the locality was investigated several times by Russian archaeologists (11); later by a Norwegian-Russian team, since 1996 (6, 12); and by a French-Russian team since 2007. More than 300 stone artefacts and 4000 animal remains have been uncovered during the various excavations, which together cover an area of approximately 550 m2.

The site is located in a paleo-gully that is incised into Triassic sandstone. The gully, which is about 250 m long and 300 m wide, starts from a shoulder in the river valley at an altitude of about 100 m above sea level (a.s.l.) and ends near the shore of the Pechora River about 44 m a.s.l. Artefacts are encased in a layer of sandy gravel up to 2.5 m thick that rests directly on the bedrock at the bottom of the paleo-gully. This layer crops out along the river cliff (Fig. 2 and fig. S7). The sediments that contain the archaeological finds accumulated during a series of successive debris flows. We were not able to determine whether there were breaks between each flow or if they all happened during one single event, although some individual flows are separated by thin layers of silt. Several meters of well-sorted windblown (aeolian) sand overlay and interfinger with the debris flow deposits (Fig. 2). Most of the bones and artefacts are well preserved. Both the excellent preservation of the bones and the inferred paleotopography indicate that the debris flows moved only a short distance, most likely less than a few tens of meters or so. The bones and artefacts occur throughout the gravelly layers. Several well-preserved mammoth bones and tusks were found on the surface of this formation, draped by the overlying windblown sand (fig. S8). Some bones and artefacts are concentrated within one of the mentioned silt horizons in the middle of the layer and may represent in situ deposits.

Fig. 2

A slightly simplified cross section of the Byzovaya site, from the Pechora River and up the gully with Paleolithic finds. All 14C dates from the find layer are plotted as a ±1σ-long line in calibrated years in order to compare with OSL ages from the overlying strata. The calculated mean age is 28,450 ± 820 14C yr B.P., corresponding to the age interval 34,580 to 31,370 cal yr B.P. for the 95% confidence interval (±2σ). This interval is marked by a vertical gray zone on the figure. All calculations are explained in the SOM text.

More than 97% of the identified faunal remains are from mammoth (Mammuthus primigenius) and represent at least 21 individuals. In addition, several other animals have been identified [supporting online material (SOM)]: woolly rhinoceros (Coelodonta antiquitatis), reindeer (Rangifer tarandus), musk ox (Ovibos moschatus), horse (Equus callabus sp.), brown bear (Ursus arctos), wolf (Canis lupus), and polar fox (Alopex lagopus). Many of the animal remains from the top deposit have cut marks that indicate processing by humans (figs. S10 and S11).

We obtained 33 14C dates from the bones (table S1). The mean age for all samples is 28,570 ± 1370 14C yr B.P., but 28,450 ± 820 when 12 dates with large uncertainties are omitted (SOM text). The distribution of the 14C ages is almost Gaussian (fig. S2), and the simplest interpretation is that the spread of ages is mainly an artefact of random counting uncertainties and that all samples should therefore have a similar age. We conclude that the age of 28,450 ± 820 14C yr B.P. is the best age of the human occupation (fig. S2). Conversion of 14C ages to sideral (cal) years is problematic for such old samples, but this age corresponds to the calibrated age intervals of 33,650 to 31,670 and 34,580 to 31,370 cal yr B.P. for the 68 and 95% confidence intervals, respectively, when using the IntCal09 (13) data set. As a further test, we obtained four optically stimulated luminescence (OSL) dates from the windblown sand draping the find-bearing strata. The ages were all between 30,000 and 33,000 yr B.P., with an uncertainty (1σ) of 2000 to 3000 years (Fig. 2 and table S2) and a mean of 32,000 ± 2000 yr B.P. The interfingering contact between the debris flow layer and the sand suggests that the ages of the layers are overlapping in time. Indeed, the OSL ages overlap with the calibrated 14C ages of the bones.

These ages demonstrate that the human occupation of Byzovaya happened during the latter part of marine isotope stage (MIS) 3, well before the buildup of the Barents-Kara Ice Sheet during MIS 2. The age fits roughly into the period when the Bryansk soil developed on the Russian Plain farther to the south (14). This interstadial (31,000 to 24,000 14C yr B.P.) was slightly warmer than the preceding and following periods, but the climate was still much colder than at present. The radiocarbon chronology indicates that the human visits at Byzovaya took place a few hundred years after the mild Ålesund interstadial in Scandinavia, an interval that has been correlated with Greenland interstadial (GI) 8 to 7 (15). Taken at face value, the calibrated age of Byzovaya corresponds to the ice core interval GI 6 to 5.

A total of 313 lithic artefacts have been collected during the various excavations. The artefacts are mainly composed of local raw materials coming from the gravels of the Pechora River (table S3). These rocks are abundant on the river shore and also in thick glacial and glaci-lacustrine sediments in the river bank (12). Our experimental in situ knapping shows that the local raw materials are applicable for any kind of flaking technology.

This industry consists of flakes, 11 cores, and as many as 80 defined typological tools. End products are well represented for a collection of this size (80 out of 313). The cores were used exclusively for flake production. Four pieces are typical Levallois cores, a flaking method that is considered to be a distinctive feature in MP assemblages (Fig. 3, no. 2). This technology is also represented by some well-defined Levallois flakes. The remaining seven cores are based on discoid (fig. S5, no. 3) and polyhedric flakings, which are a common combination in MP contexts. The applied stone technology is based on Levallois and discoid flake production. There are neither technological blades nor bladelets nor other elements that could be ascribed to the technical representation of UP technologies. We conducted experiments on the same raw materials and were able to make similar flakes only by direct hard hammer percussion. The preserved flakes could have been used to make side-scrapers of various types and sizes. Some flakes were directly used without any retouching, because some of their edges are worn (fig. S12).

Fig. 3

Photographs of two stone artefacts from Byzovaya. 1, Keilmesser; 2, Levallois core, preferential method. [Photos by Hugues Plisson]

None of the 313 artefacts reflects a tool production technology typical of UP cultures. Furthermore, diagnostic tools that are common in any UP industry of Eurasia such as burins, backed tools, pointed blades, or bladelets are not represented. There are 11 end-scrapers, but none of these were prepared from UP blades. Varieties of end-scrapers, prepared from flakes, are common elements in any European MP industry, known since the first Mousterian typological analysis (16). Typological tools are mainly members of the Mousterian group (16), dominated by distinctive side-scrapers made out of flakes (fig. S5, nos. 1 and 2) that are typical for MP industries (17) (fig. S6 and table S4). Six of these tools have been retouched to form a bifacial tool. Most of the bifacial tools are thick, with a plano-convex section: one face shaped by large flakes and the opposite face formed by a semi-abrupt retouch. This way of shaping has been used for producing so-called Keilmesser tools (plano-convex and backed bifacial tools, Fig. 3, no. 1), which are considered to be specific artefacts of some archetypical MP industries of Central and Eastern Europe (1820). Two of the bifacial tools from Byzovaya present a thin regular transformation of their faces that illustrates the technological similarities between this industry and the Eastern European MP (18, 19), where the so-called Blattspitzen (short foliate) tools occur frequently.

The lithic industry of Byzovaya is technically and typologically homogeneous and shows a combination of diagnostic features that belong exclusively to MP traditions from Central and Eastern Europe, previously known from sites dating from MIS 3 to 4 (19). All lithic tools and bones are stratigraphically associated, and 22 faunal remains show anthropogenic modifications. This is in accordance with the use wear of lithic tools related to butchering activities. Two of our dates are on a cut marked mammoth bone and a modified reindeer antler (table S1).

The EUP complexes that succeeded the MP complexes present specific archaeological features that distinguish them from the MP industries. Blades, bladelets, and projectile points constitute the lithic background for EUP assemblages in Eurasia. For example, at Kostenki 14, along the Don River in southern Russia (Fig. 1), these EUP elements appear in layer IV, dated to 45,000 to 41,000 cal yr B.P., which is about 10,000 years earlier than the human occupation at Byzovaya (2123). At Zaozer’e, farther to the north and closer to Byzovaya, blades and bladelets were produced by humans sometime between 39,000 and 37,0000 yr B.P. (Fig. 1) (4, 6). Russian EUP sites that are more or less contemporaneous with Byzovaya, such as Sungir (24) and Garchi (6), are characterized by lithic triangular projectile points. A rich bone and ivory technology with spear points, figurines, ornaments, and beads is common in most EUP Eurasian sites where organic materials are well preserved (as at Byzovaya). It has been recently proposed that some of the Kostenki sites show some archaic traits resulting from a specialized tool kit made by UP hunters at a kill-butchery site (25). However, tools from the Gorotsovian and Streletskian layers at Kostenki do not resemble the Byzovaya assemblage but are dominated by typical UP traits reflecting their specific cultural associations (standardized triangular foliated spear points, in the Streletskian layers, and rich bone and ivory technologies in the Gorotsovian).

There is no evidence that the assemblage at Byzovaya can be explained by site-specific activities such as mammoth exploitation and butchering. At all well-dated sites with abundant mammoth remains, the artefacts that are found with the animal remains also reveal the cultural affiliation of the humans who made them. This is true for MP sites such as Mont-Dol (26) or Salzgitter-Lebenstedt (27) and for EUP Gravettian sites of Central Europe, such as Milovice in Moravia and Krakow-Spadzista in Poland (28, 29). Theoretically, one may also wonder if the lack of some UP elements, such as blades and bladelets, can be explained by the quality of the raw material at hand. However, various siliceous rocks of good quality were easily available and were also used for tool production at an adjacent Mesolithic site containing almost exclusively well-prepared and regular blades and bladelets.

Most researchers agree that classical Mousterian industries in Europe were exclusively produced by Neandertals (30, 31). However, whether Byzovaya represents a Neandertal site or not cannot be demonstrated beyond doubt until human bones or DNA are found. If the Byzovaya artefacts were struck by modern humans, this would have major implications for understanding the MP-UP transition, as it would imply that these Arctic H. sapiens groups preserved older, traditional MP cultures far after the full expansion of UP modern societies in the rest of Eurasia.

Neandertal fossils and unambiguous MP industries are known from a large part of the Eurasian continent (fig. S1), stretching from the Atlantic Ocean to the Altai Mountains in the east (32). However, other unambiguous MP artefacts have not been reported from areas north of about 55°N latitude, which is about 1000 km south of Byzovaya (33). (Fig. 1). One of the most recent Neandertal remains is a skeleton found in Croatia (Vindija G1) and dated to ~32,000/33,000 14C yr B.P. (≥37 cal yr B.P.) (34). Younger Mousterian artefacts have been reported from the southern part of the Iberian Peninsula, including layer IV of Gorham’s cave, Rock of Gibraltar, where 103 artefacts have been attributed to a late Mousterian culture between 33,000 and 24,000 14C yr B.P. (35). However, these ages have been debated (3638).

If, like elsewhere in Europe, these technologies were realized by Neandertals, then such late persistence of Mousterian traditions at this high latitude as late as about 10,000 years after the UP first appearance at lower latitudes in Russia (Kostenki) would imply that the two human species coexisted over a longer time period than has been thought. In any case, the site shows that Neandertals or another relic population of humans with Mousterian technology were fully capable of coping with the cold climate and dark winters that prevailed at the highest latitudes. The site also challenges the hypothesis that there was a complete replacement of the MP societies in all of Europe as early as around 37,000 cal yr B.P. (37).

Supporting Online Material

SOM Text

Figs. S1 to S12

Tables S1 to S4

References 39 to 57

References and Notes

  1. Acknowledgments: L.S., H.P., and A.B. analyzed the archaeological material; J.I.S, J.M., and H.P.H. performed the stratigraphic, paleoenvironmental, and chronological descriptions and interpretations; and P.Y.P. organized the archaeological field missions. All contributed to the writing of the paper. Supported by the Fondation des Maisons des Sciences de l’Homme, the French-Russian Center for Social and Human Sciences in Moscow, the French Ministry of Foreign and European Affairs, the Research Council of Norway, and Russian Foundation for Humanitarian Research grant 08-01-94952a. Kh. A. Arslanov and S. Gulliksen performed the radiocarbon dating and A. Murray the OSL dating. We thank the Komi Scientific Centre in Syktyvkar; the museums of Pechora and of Syktyvkar; and the Institute of History of Material Culture of St. Petersburg, especially A. A. Sinitsyn and A. E. Matjuhin for access to their collections and N. N. Skakun for her help. We also thank E. Bjørseth for preparing Figs. 1 and 2 and figs. S1 and S2 and J. E. Lewis for improving the English.
View Abstract

Navigate This Article