Early human presence in the Arctic: Evidence from 45,000-year-old mammoth remains

See allHide authors and affiliations

Science  15 Jan 2016:
Vol. 351, Issue 6270, pp. 260-263
DOI: 10.1126/science.aad0554

Earliest human Arctic occupation

Paleolithic records of humans in the Eurasian Arctic (above 66°N) are scarce, stretching back to 30,000 to 35,000 years ago at most. Pitulko et al. have found evidence of human occupation 45,000 years ago at 72°N, well within the Siberian Arctic. The evidence is in the form of a frozen mammoth carcass bearing many signs of weapon-inflicted injuries, both pre- and postmortem. The remains of a hunted wolf from a widely separate location of similar age indicate that humans may have spread widely across northern Siberia at least 10 millennia earlier than previously thought.

Science, this issue p. 260


Archaeological evidence for human dispersal through northern Eurasia before 40,000 years ago is rare. In west Siberia, the northernmost find of that age is located at 57°N. Elsewhere, the earliest presence of humans in the Arctic is commonly thought to be circa 35,000 to 30,000 years before the present. A mammoth kill site in the central Siberian Arctic, dated to 45,000 years before the present, expands the populated area to almost 72°N. The advancement of mammoth hunting probably allowed people to survive and spread widely across northernmost Arctic Siberia.

Evidence for human habitation in the Arctic before the Last Glacial Maximum (LGM), which spans 26.5 to 19 thousand years ago (ka) (1), is very rare (Fig. 1A and fig. S1). It has only become available in the past 20 years, with the discoveries of the Mamontovaya Kurya site (2) in the European Arctic and the Yana Rhinoceros Horn Site (RHS) in Arctic Siberia (3). Before these discoveries, researchers held that humans could not have started populating the Arctic regions until the Pleistocene-Holocene boundary (4). The recent discoveries indicate the presence of people in the Arctic at least at the end of marine isotope stage 3 (MIS 3), around 28,000 14C years before the present (yr B.P.) or slightly earlier, whereas older sites are found south of 55°N.

Fig. 1 Geographic, stratigraphic, and dating overview for the Siberian Arctic.

(A) Dated archaeological contexts (2, 3) and geological sites mentioned in the text: SKMS, SK mammoth site; KML, Kastyktakh mammoth site (29). (B) Location of the SK mammoth site. (C) General view of the exposure before the excavations. (D) Lithology and cryostratigraphy of the SK mammoth site.

Archaeological evidence for human dispersal throughout northern Eurasia dated to the first half of MIS 3 is known from different areas (510). In Siberia, this evidence is concentrated far south of the Arctic circle (711). In addition to some dozen sites with age estimates within ~45 to 30 ka, which are well supported by reliable data, there are several controversial localities in northeast Europe (12, 13), the Urals (14), and Siberia (9, 10). Thus, the relevant archaeological record for all of northern Eurasia amounts to 15 to 20 contexts with a northern latitudinal limit of 55°N (fig. S1). In west Siberia, a human fossil with an early modern human genome was recently found at 57°N and produced a direct radiocarbon date of 45,000 yr B.P.; however, these remains were not found in an archaeological context (15). Thus, the archaeological record for northern Eurasia has not suggested human presence in the Arctic that early, although there is evidence firmly dated to ~30 ka (3).

Ice sheet expansion during MIS 4 did not affect most of the Siberian Arctic (16). Interglacial (MIS 3) environmental conditions varied across the Eurasian Arctic (1721) but were overall beneficial for the growth of late Pleistocene large herbivore populations, including mammoths, in various parts of the region (2225). At the end of MIS 3, mammoths moved into deglaciated areas (26) and offered an unlimited food source, supporting the pre-LGM human settlement now corroborated for both the European and eastern Siberian Arctic (2, 3, 25, 26).

The central Siberian Arctic was also populated by mammoths for a long time (22, 27). Reconstructed fluctuations of their population numbers (22) largely follow the pattern seen among the mammoth populations in the New Siberian Islands and Northern Yana-Indighirka lowland. One of the highest peaks of the mammoth population numbers occurred around 44 to 42 ka (25).

In 2012, a team led by one of us (Tikhonov) excavated a partial carcass of a woolly mammoth (Mammuthus primigenius, Blumenbach, 1799) from frozen sediments exposed in a coastal bluff on the eastern shore of Yenisei Bay, 1.8 km north of the Sopochnaya Karga (SK) meteorological station, at 71°54′19.2″N and 82°4′23.5″E (Fig. 1B). The exposure stratigraphy there is composed of several clear structural elements (28).

Several radiocarbon dates help anchor this sequence in time (Fig. 1, C and D, and table S1). The peat layer (bed 3) is dated to 36,000 ± 2500 14C yr B.P. (LE-9822). A small willow branch fragment from bed 3 yielded an age of 47,600 +10,200/–4400 14C yr B.P. (LE-9821). Bed 4, which marks the beginning of the taberite sediments, produced a date of 18,200 +2600/–2200 14C yr B.P. (LE-9823). This sequence is above bed 1, which holds the mammoth carcass. Considering these dates, the age of the SK mammoth can be estimated at about 40 ka. A direct date on tibia bone, 44,570 +950/–700 14C yr B.P. (table S2), is consistent with the stratigraphy and the surrounding deposit dates (Fig. 1D). Thus, the SK mammoth carcass is a rare in situ specimen from the early phase of MIS 3 when mammoth populations in Taimyr were slowly growing (22, 25). Additional dates (28) independently received on mammoth remains concur (table S3).

The SK mammoth is an exceptionally complete mammoth skeleton (28) with a small amount of preserved soft tissue, including the remains of the fat hump and the penis. It is more complete than other recent finds from Taimyr, known as the Kastyktakh (29) and Jarkov mammoths (30). The large amount of fat at the hump indicates that the mammoth was in a good physical condition. This was a young male around 15 years old, according to the tooth change model (31).

Its bones exhibit a number of unusual injuries (Figs. 2 and 3 and figs. S2 to S4). Damage is seen on the left scapula (Fig. 2B and fig. S2), the left jugal bone (Fig. 3), the fifth left rib (fig. S3, A to C), and the second right rib (fig. S3, D and E). Unequivocal postmortem damage is also recorded on the tip of the right tusk and the mandible, the left coronoid process of which is broken off (28).

Fig. 2 Taphonomy of the SK mammoth and the inventory map for perimortem (red arrows) and postmortem damages (blue arrows) found on bones.

(A) Mammoth skeleton, right view. (B) Mammoth skeleton and internal organs, left view. (C) Excavations of the carcass from channel deposits unit. (D) Right mammoth tusk with worked tip. (E) Undamaged alveolar part of the tusk, with extensive vivianite crust. (F) Unpatterned traces on the right tusk surface left by hard particles of the matrix sediments.

Fig. 3 Human-inflicted lesion on the left jugal bone of the SK mammoth head.

(A) Mammoth skull with the marked direction of the penetrating wound (solid red arrow) documented by a lesion on the jugal bone [(B) to (F)] and desired direction of the impact (dashed red arrow) in the area of the nasal opening. (B) Jugal bone, external side. (C) Jugal bone, internal side with a hole. (D) XCT image of the impact area. (E) Longitudinal section of the impact area. (F) Cross section of the impact area. (G) Three-dimensional model of the tool tip.

A small rounded hole, filled with sediment particles, was discovered on the internal side of the left jugal bone after it was separated from the skull during the excavation (Fig. 3). The bone is uniformly patinated and has no evidence of recent damage. To clarify the character of this unusual injury, the bone was studied with x-ray computed tomography (XCT) (Fig. 3, D to F).

The shape of the injury (Fig. 3G) suggests that the tip of the weapon that damaged the jugal bone had a thinned symmetric outline (the cross-tip section at the point of entry measures 3.4 mm at the short axis and 5.1 mm at the long axis) and was relatively sharp (the height of an equilateral triangle with a base of 5.1 mm is 3.42 mm). In most cases, bone or ivory weapons have a conical tip that is symmetric and quite acute (~30° to 40°) at the end (3, 32, 33), but they are relatively fragile and often break as they penetrate bones (33). In this case, the tool resisted breaking and inflicted injury on the cranial bones. The blade retained its weapon characteristics and kept enough energy to go through the cheekbone surface, penetrating deeply into the bone. The blow was evidently very strong and was suffered by the animal from the left back and from top down (Fig. 3A), which is only possible if the animal was lying down on the ground.

This injury itself is probably the result of a missed blow, targeting the base of the trunk. This specific hunting method is still practiced in Africa by elephant hunters, who target the base of the trunk to cut major arteries and cause mortal bleeding (34). This blow becomes necessary after the animal has been sufficiently injured, and the SK mammoth displays numerous injuries in the thoracic area (to ribs and the left scapula).

The most remarkable injury is to the fifth left rib, caused by a slicing blow, inflicted from the front and somewhat from above downward (Fig. 2B and fig. S3C). The incision is filled in with well-rounded sand particles from the matrix deposits. Although it was a glancing blow, it was strong enough to go through skin and muscles and damage the bone. A similar but less powerful blow also damaged the second right front rib (Fig. 2A and fig. S3, D and E). Such blows were aimed at internal organs and/or blood vessels.

The SK mammoth was also hit in the left scapula at least three times. Two of these injuries, judging from the location of the dents at the distal (top, in an anatomic skeletal position) edge of the scapula, left of the spine, were imparted by a weapon, which went downward through the skin and muscles, moving from the top and side (Fig. 2B and fig. S2, A and D). These markings indicate injuries evidently left by relatively light throwing spears.

A much more powerful blow damaged the spine of the left scapula (Fig. 2B and fig. S2). It may have been imparted by a thrusting spear, practically straight from the front at the level of the coracoid process. The weapon went through the shoulder skin and muscle, almost completely perforating the spine of the scapula. The blow produced a lattice of cracks, and the bone curved in the direction of the impact. Taking into account the scapula’s location in the skeleton and the estimated height of this mammoth, the point of impact would be approximately 1500 mm high; i.e., at the height of an adult human’s shoulder. This area is the weakest on the scapula, but it usually remains intact even on bones found separately, including bones of younger mammoth individuals.

The SK mammoth remains also display postmortem damage, which indicates human activity. First, the ramus of the mandible is broken. These bones are quite strong: even when found isolated in much harsher taphonomic conditions, they often remain complete. However, in anthropogenic contexts, mammoth mandibles are mostly lacking one or both coronoid processes. Such mandible treatment may reflect tongue extraction. The Yana collection showed that mammoth tongue was often consumed by the hunters, perhaps as a ritual food or a delicacy (33).

The only preserved tusk of the SK mammoth (the right tusk) shows traces of human modification. These could be expected to be focused around the alveolar region, reflecting an attempt to separate the outside of the tusk by chopping. Instead, that portion remained intact (Fig. 2E), whereas the tip of the tusk shows evidence of impact. Several relatively thin subparallel spalls were removed (fig. S4); the tip of the tusk, normally quite blunt (35, 36), served as a platform for these removals.

The natural breakage pattern found on mammoth tusks (36) and on those of recent elephants (37) is completely different from the pattern on the SK mammoth tusk. At the same time, such a technique is quite different from the ivory technology known in arctic Siberia in late MIS 3 (38). However, the same approach might have been used to produce ivory tools with asymmetric working parts that are known in the younger archaeological sites nearby (39).

These removals aimed to obtain long thin slivers of ivory with sharp edges, which were usable as butchering tools in places where quality lithic raw material is hardly available. In the modern estuarine area of the Yenisei, it is completely absent. Ivory flakes have a sharp cutting edge and can be used as cutting or scraping implements. Rarely, such objects are found in Siberian Arctic Stone Age sites (39), specifically in regions where lithic raw material is available in small nodules only.

Thus, the SK mammoth bone remains provide well-supported evidence for human involvement in its death. These are: (i) a lesion on the jugal bone, (ii) a butchery mark on the fifth left rib (the most solid evidence), and (iii) the worked distal end of the right tusk. The same butchery pattern is well known for mammoth bones from the Yana RHS complex (33). In particular, it is repeatedly documented for mammoth ribs (Fig. 4). The bone lesion found on the rib of the SK mammoth fully replicates that pattern (Fig. 4G). All together, these findings leave no doubt that people were present in the central Siberian Arctic by 44,570 +950/–700 14C yr B.P. (GrA-57723), or 49,150 to 47,100 calendar yr B.P.

Fig. 4 Mammoth ribs with hunting lesions collected at Yana RHS [(A) to (F)], showing the mechanism for the formation of bone injury on the fifth rib of the SK mammoth (G).

(A) Mammoth rib with embedded lithic tool fragment. (B) Mammoth rib with two injuries that retain lithics. (C) View of the upper cut at (B). (D) View of the lower cut at (B). (E) Bone injury with no lithic in it but clearly left by the same action as for (A) and (B). (F) Bone injury that resulted from sliding of the lithic implement that removed part of the bone. (G) Hunting lesion on the fifth left rib of the SK mammoth; compare to (A) to (F) and note the same scale for all images.

Independently, another locality in the eastern Siberian Arctic (Fig. 1A) produced evidence for human habitation that also dates to early MIS 3 (28). The Bunge-Toll site yielded remains of bison, rhinoceros, and mammoth. Among those, a left humerus of a Pleistocene wolf exhibited an unusual pathology, resulting from an injury, on its external lateral surface. Standard x-rays and XCT show that the pathology is a result of a penetrating injury inflicted by a sharp weapon (fig. S5). Such trauma could only have been caused by a human. Direct 14С dating of the bone indicates that this episode occurred 44,650 +950/–700 (GrA-57022); i.e., contemporaneously with the SK mammoth kill. These two incidents suggest that even during the early phases of MIS 3, humans inhabited the Arctic quite widely, although the population was probably small and remained sparse for a long time.

The SK mammoth kill, along with the Bunge-Toll site, confirms the presence of humans in the Arctic much earlier than previously suggested. This is a rare case of unequivocal evidence for clear human involvement, even if there is no artifact association (40). Apparently, humans’ ability to survive in the Arctic environment, and their spread within the region as early as 45 ka, represents an important cultural and adaptational shift. We speculate that adaptation changes that ensured human survival there may be related to innovations in mammoth hunting. Sustained development of the populations, secured by an abundant food source, could have led to their rapid spread across the Siberian Arctic. The early arrival of humans in the area close to the Bering land bridge may have provided an opportunity for humans to enter the New World before the LGM.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S5

Tables S1 to S3

References (4157)


  1. See the supplementary text on Science Online.
  2. ACKNOWLEDGMENTS: We thank E. Solinder and A. Bystrov (SK weather station), the Taimyr Region administration, and Norilsk Nickel mining company. We are also grateful to P. Ivanov (photography); A. Mashezerskaya (drawings); V. Stegantseva (illustrations); N. Slobodina (translation); and N. Burova, A. Sementsov, and J. van der Plicht (radiocarbon dating). The SK mammoth remains are currently stored at the Zoological Institute, St. Petersburg, Russia. When the study ends, they will be moved to the Taimyr Regional Museum for permanent storage and display (Dudinka, Taimyr Autonomous District, Russia), and they can then be resampled if needed while a bulk sample of the organic deposit from the inside of the SK mammoth stays at the Zoological Institute in St. Petersburg. The Yana site materials and finds from the Bunge-Toll site are stored at the Institute for the History of Material Culture, St. Petersburg, Russia), as a part of the materials collected under the Zhokhov-Yana research project. All collected skeletal elements of the Kastyktakh mammoth are with the Ice Age Museum, Moscow, Russia. This project was supported by the Russian Foundation of Basic Research (grant Nr 13-06-12044 to V.V.P.); publication is supported by the Rock Foundation (New York, USA).
View Abstract

Navigate This Article