Content uploaded by Pasquale Raia
Author content
All content in this area was uploaded by Pasquale Raia on Dec 21, 2018
Content may be subject to copyright.
Human influence on distribution and extinctions of
the late Pleistocene Eurasian megafauna
Diana Pushkina
a,
*, Pasquale Raia
b
a
Department of Geology, University of Helsinki, P.O. Box 64, 00014, Finland
b
Dipartimento STAT, Universita` degli Studi del Molise, Via Mazzini 10, 86170, Isernia, Italy
Received 20 February 2006; accepted 19 September 2007
Abstract
Late Pleistocene extinctions are of interest to paleontological and anthropological research. In North America and Australia, human occu-
pation occurred during a short period of time and overexploitation may have led to the extinction of mammalian megafauna. In northern Eurasia
megafaunal extinctions are believed to have occurred over a relatively longer period of time, perhaps as a result of changing environmental
conditions, but the picture is much less clear. To consider megafaunal extinction in Eurasia, we compare differences in the geographical distri-
bution and commonness of extinct and extant species between paleontological and archaeological localities from the late middle Pleistocene to
Holocene. Purely paleontological localities, as well as most extinct species, were distributed north of archaeological sites and of the extant spe-
cies, suggesting that apart from possible differences in adaptations between humans and other species, humans could also have a detrimental
effect on large mammal distribution. However, evidence for human overexploitation applies only to the extinct steppe bison Bison priscus. Other
human-preferred species survive into the Holocene, including Rangifer tarandus,Equus ferus,Capreolus capreolus, Cervus elaphus, Equus hem-
ionus, Saiga tatarica, and Sus scrofa. Mammuthus primigenius and Megaloceros giganteus were rare in archaeological sites. Carnivores appear
little influenced by human presence, although they become rarer in Holocene archaeological sites. Overall, the data are consistent with the
conclusion that humans acted as efficient hunters selecting for the most abundant species. Our study supports the idea that the late Pleistocene
extinctions were environmentally driven by climatic changes that triggered habitat fragmentation, species range reduction, and population
decrease, after which human interference either by direct hunting or via indirect activities probably became critical.
Ó2007 Elsevier Ltd. All rights reserved.
Keywords: Commonness; Rarity; Extinction; Human overexploitation; Steppe bison; Paleolithic
Introduction
Humans are often blamed for their imprudent exploitation
of the Earth’s living resources (Diamond, 2005). A myriad
of known and unknown species currently risk extinction
because of human activities. The current ineptitude of human
societies to adopt an environmentally sustainable lifestyle may
trace back to the end of the Pleistocene, when most megafauna
vanished in coincidence with the spread of Neolithic people
and their advanced tools (Martin, 1973; Barnosky et al., 2004).
In North America, the late Pleistocene megafaunal extinc-
tions were tightly grouped around 10,500e11,000 years ago
(kiloannum, ka), a period coinciding with the arrival of humans
and with widespread climate change. Consequently, these ex-
tinctions are often attributed to humans (Martin, 1973, 1984;
Alroy, 2001; Barnosky et al., 2004; Bulte et al., 2006; Pankovic
et al., 2006) although evidence is far from conclusive (Graham
and Lundelius, 1984; Beck, 1996; Grayson and Meltzer, 2002;
Guthrie, 2003, 2006; Grayson, 2006; Solow et al., 2006). The
same applies to Australia, where the extinction wave started
some 30 ka, following human arrival at 50 ka (Roberts et al.,
2001; Johnson, 2002; but see Trueman et al., 2005). In Eurasia,
late Pleistocene extinctions were spread over a longer period of
time, and human influence is less evident (Stuart et al., 2002,
* Corresponding author.
E-mail address: diana.pushkina@helsinki.fi (D. Pushkina).
0047-2484/$ - see front matter Ó2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jhevol.2007.09.024
A
vailable online at www.sciencedirect.com
Journal of Human Evolution 54 (2008) 769e782
2004; Stuart, 2005). Consequently, Eurasian megafaunal ex-
tinctions have been proposed to result from rapid environmen-
tal changes (Kvasov, 1977; Tormidiaro, 1977; Guthrie, 1984;
Stuart, 1991; Lister and Sher, 1995; Sher, 1997). At present
the largest species in the world have remained only in Africa
and, perhaps, in southeastern Asia (Nowak, 1999). Intrigu-
ingly, most African megafauna never went extinct although
this is the continent on which humans arose (Brook and Bow-
man, 2002).
Evidence of human species intervention is the keystone to
understanding how and to what extent people influenced
megafaunal demise. Consequently, most analyses of the late
Pleistocene extinction patterns have focused either on surveys
of the occurrence and extent of human modification (e.g., bone
manipulation, artifacts; Fisher, 1984; Grayson, 2001; Grayson
and Meltzer, 2002; Surovell et al., 2005) or on species biolog-
ical attributes such as reproductive rate (Johnson, 1998; Alroy,
2001; Cardillo and Lister, 2002) or sexual versus nutritional
selection pressures (Moen et al., 1999; Pastor and Moen,
2004). Species ecological and/or behavioral responses, such
as direct avoidance of predators (i.e., humans) and prey na-
ivety to a novel predator (humans, in this case), often leave
few, if any, traces in the fossil record and have been little in-
vestigated so far. Here we concentrate on a related attribute,
species commonness, to untangle the human influence on the
distribution and survival of the late Pleistocene Eurasian large
mammals.
We suppose that if humans drove megafauna to rarity and
extinction, these species should have been significantly more
abundant in human-occupied localities (where human hunters
were gathering their remains) than at paleontological sites;
such commonness would be an obvious sign of human interest.
These species of ‘‘human interest’’ should also have become
rarer in time as human presence became more evident. We
tested these hypotheses by comparing species commonness
and geographic distribution between paleontological and ar-
chaeological localities in areas with and without human pres-
ence during the late middle and late Pleistocene and Holocene,
taking extant species as the control group. We analyzed the
temporal trend in commonness patterns looking at the grand
scale pattern, not concentrating on any species in particular.
We also tried to evaluate possible biases on commonness fac-
tor such as species body size or locality environment type. The
expectation for the extinct megafauna is to become overrepre-
sented in archaeological sites as human technology improved
from paleolithic to neolithic. Otherwise, factors other than
direct human exploitation should have driven the demise of
Eurasian megafauna.
Materials and methods
Locality distribution data
Numerous literature sources and the existing databases: the
Neogene of the Old World (NOW, Fortelius, 2007), European
Quaternary Mammalia (Pangaea, available online at: http://
www.pangaea.de/), the marine oxygen isotope stage 3 project
(OIS-3; van Andel, 2002) and the Paleolithic sites of northern
Asia (available online at: http://sati.archaeology.nsc.ru/pna/)
were incorporated to obtain data on Eurasian localities and
species. The dataset comprises the late middle and late Pleis-
tocene as well as the Holocene sites from western and central
Europe and the former USSR along with several type localities
from eastern Europe and the Mediterranean. However, the
main emphasis was made on the former USSR territory be-
cause eastern Europe and Siberia were the last stand and key
survival area of many species that went extinct at the end of
the Weichselian and Holocene. Southeastern Asia (China, In-
dia) and Japan were excluded because of controversial taxon-
omy for many of their species (Chinese), high endemicity
(Japanese), or lack of data.
In order to better correlate localities across Eurasia where
more detailed temporal divisions are not available, sites
were arranged into six age groups: (1) the late middle Pleisto-
cene (or late Early Paleolithic sites, 400e130 ka; here called
late MP); (2) Eemian interglacial (or the early Middle Paleo-
lithic or early Mousterian sites, 130e115 ka); (3) early and
middle Weichselian Glaciation (or the late Middle Paleolithic
or late Mousterian sites along with the early Late Paleolithic or
Aurignacian sites, 115e24 ka); (4) late Weichselian or Last
Glacial Maximum (LGM) (or the middle Late Paleolithic sites,
24e15 ka); (5) latest late Weichselian or Late Glacial (LG) (or
the latest Late Paleolithic or Magdalenian sites, 15e10 ka);
and (6) Holocene (or the Mesolithic and Neolithic sites less
than 10 ka). Chronostratigraphic correlations were based on
marine oxygen isotope record (MIS/OIS) or chronometric dat-
ing (radiocarbon, TL, ESR, U-series methods), paleomagnetic
polarity and chron correlations, continental mammalian bio-
stratigraphy, along with archaeology (human cultural stages;
Shantser, 1982; Stuart, 1991; Kahlke, 1999; van Kolfschoten,
2000; Currant and Jacobi, 2001; Khromov et al., 2001;
Vangengeim et al., 2001; Vasil’ev et al., 2002; Sher et al.,
2005). The oldest age group consists of the Khazarian and Sin-
gilian, which are correlated to the Holsteinian and Saalian
sites (Azzaroli et al., 1988; Khromov et al., 2001; Vangengeim
et al., 2001). The Eemian interglacial (MIS5e) was excluded
from the main analyses because of a possible stratigraphic
mixing since only recently have biostratigraphers in western
and central Europe begun to discern the Eemian from the ear-
lier interglacial (MIS7) localities, many of which happened to
be often mistaken for the Eemian (Schreve and Thomas,
2001).
Originally we used a so-called loose temporal locality ar-
rangement, in which a locality with a temporal range was
attributed to an age group based on the majority of its temporal
range and the description of a locality. In the original ‘‘loose’’
version as ‘‘archaeological’’ we took the sites that were attrib-
uted to a human cultural stage or where human artifacts were
found. The rest of the localities were treated as ‘‘paleontolog-
ical’’. Although even when attributed to a cultural stage
archaeological sites may contain remains collected by nonhu-
man related processes or predators, the presence of lithics in-
dicates that humans perhaps exploited some of the remains
and it is this latter condition we were seeking.
770 D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
However, we also used a more conservative approach with
a stricter temporal resolution and different interpretation of
archaeological versus paleontological sites. In this conserva-
tive version we only used directly dated late Paleolithic local-
ities. The earlier archaeological sites were either dated or
correlated to the MIS/OIS and Pleistocene/Weichselian
stages. An overlap of 1,000 to 3,000 years between age
groups may be present in a few of archaeological localities
that are considered the latest Late Paleolithic (according to
Vasil’ev, 2003), whereas their lower time boundaries are
dated to slightly earlier than 15 ka. The rest of the localities
have clear-cut boundaries. We also left out sites with a larger
age overlap and those in which the dating did not correspond
to the archaeological description. A slight bias could exist
here toward recognizing an increased human effect during
the later periods, because humans could have reworked
some bones from earlier periods that were supposed to be
either paleontological or archaeological.
In the conservative version, we counted as archaeological
(rather than paleontological) any site with any indication of
human presence, including those with human skeletal remains
but without artifacts or cultural attributes even though it could
be argued that such sites might represent carnivore generated
accumulations. For that reason it was impossible to determine
the exact agent of accumulation in many of our localities from
all parts of Eurasia although our dataset consists of localities
with a firm indication that faunal remains were accumulated
by humans (‘‘kitchen’’ remains, camp sites). These latter sites
still could have contained scavenged bones of species, which
were not actively or passively (using natural traps) hunted in
a live form. In general, it is difficult to determine when
Homo species evolved as the ‘‘hunter’’ from the ‘‘hunted
upon’’ and the predation efficiency of human colonizers is
likely to be exaggerated. Thus, we also exaggerated a possible
human effect by ascribing to them a selection actually made
by other agents, biasing our analysis toward the recognition
of human influence.
In sum, the strict conservative database includes 564
Eurasian localities, of which 267 are archaeological and
297 paleontological. We anticipated and obtained strikingly
similar results in both approaches, indicating that the
patterns we found are robust irrespective of the particular
selection criteria we applied. Because of the similarity in
the results using both methods, we only show the conserva-
tive version below.
Commonness
Species commonness is the proportion of sites in which
a species was present during a given time period. Common-
ness is a proxy of abundance of the species remains in a fossil
locality or assemblage, because it is likely that within a given
region a species abundant in some sites will be present in most
of them (Jernvall and Fortelius, 2002, 2004; Raia et al., 2006).
It has been shown that the species present in 25% of local-
ities (that is, common species) show general evolutionary
trends more strongly as they are the ones that make the
most use of the available abundant habitats and resources
(Jernvall and Fortelius, 2002, 2004), thus, being better adapted
to the environment. The rare species, on the other hand, pro-
duce the opposite results as the common species and tend
only to add noise to the results (Jernvall and Fortelius, 2004;
Pushkina, 2006).
Estimates of commonness proved rather robust (Jernvall
and Fortelius, 2004), especially when using many localities
in a single time period because as proportion (or percentage)
commonness is not affected by the actual number of localities.
Even a misidentification of several individuals will not signif-
icantly change the results. Such a method is strengthened here
by using not only a single varying species, but several species
belonging to a certain assemblage or group. The differential
length of the temporal units we use should not significantly
affect the commonness results, especially when the statistical
differences between the archaeological and paleontological
sites are large.
Species
We analyzed the large ungulate and carnivore species in
Eurasia during times when the human effect could be observed
along with their movement from Europe to Siberia. We chose
ungulate species that belonged to the two best known faunal
assemblages in Europe and Siberia (Palaeoloxodon antiquus
interglacial and Mammuthus primigenius glacial) and the car-
nivores that accompanied them in Eurasia during the second
half of the Pleistocene and Holocene.
The size of megafauna is not universally defined and some
papers do not indicate any explicit size threshold (e.g., Brook
and Bowman, 2002; Johnson, 2005; Wroe et al., 2006). Brook
and Bowman (2004) interestingly noted that ‘‘.most authors
have restricted their discussion of extinct ‘megafauna’ by ref-
erence to some arbitrary body mass threshold, often set at
around 45 kg.despite there being no obvious functional ba-
sis for this threshold (Owen-Smith, 1988). A point often over-
looked is that many medium- and some small-sized mammals
also went extinct (Johnson, 2002).’’ We used a body size
limit of approximately 7 kg in this study.
We excluded ungulate species known to have had a re-
stricted distribution (e.g., sheep, goats, the antelope, Spiro-
cerus kiakhtensis,ortheextinctyak,Bos baikalensis). And
we excluded all domestic species, such as Bos taurus or
Canis familiaris. In the case of horses and pigs, the do-
mesticated forms were not osteologically discriminated
from wild forms in all localities. They may be included
in some localities. However, no species was used more
than once. For instance, if there were Bison priscus and
B. cf. priscus in a site list it was counted only once as
B. priscus.
We divided the species into extinct and extant groups
depending on whether the species was currently extant
andlivinginthewildinEurasia.Contemporaryhuman
reintroductions [e.g., Ovibos in Eurasia (Kahlke, 1999)]
were excluded. The extinct group includes 13 species: the
woolly mammoth, Mammuthus primigenius;thewoolly
771D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
rhinoceros, Coelodonta antiquitatis; the steppe bison, Bison
priscus;themuskox,Ovibos moschatus; the straight-tusked
elephant, Palaeoloxodon antiquus; Merck’s rhinoceros, Ste-
phanorhinus kirchbergensis; the narrow-nosed rhinoceros,
Stephanorhinus hemitoechus;theaurochs,Bos primigenius;
the giant deer or Irish elk, Megaloceros giganteus;the
Pleistocene ass, Equus hydruntinus; cave hyena, Crocuta
crocuta; cave lion, Panthera leo; and cave bear, Ursus
spelaeus.
We note that although the aurochs was still living in histor-
ical times and the giant deer and cave lion survived into the
Holocene, we had robust logical grounds to maintain these
species in the extinct category despite their presence in our
Holocene data. Species become rare before extinction
(Jernvall and Fortelius, 2004; Vrba and DeGusta, 2004; Raia
et al., 2006) and rare species often do not leave fossilized re-
mains. Hence, as we were dealing with species commonness
variation in the presence of humans, the zero commonness
data could in fact represent no species present or true extinc-
tion, a species doomed to extinction, or a local extinction that
could be or not be followed by new access to former habitats.
Sorting out among these alternatives is impossible in our data.
Yet, considering the aurochs, mammoth, and giant deer as ex-
tant or the fallow deer as extinct just because we had the first
two (and no fallow deer) in our Holocene data would deny the
mere fact that fallow deer are still living while the aurochs,
mammoth, and the giant deer are extinct. This would also
deny a period of rarity that species pass through prior to ex-
tinction. If we did that, we would have supported the illogical
conclusion that humans could force a species to extinction by
demographic collapse and then neglect extinction as a real
consequence of that collapse by assigning them to the category
of living species.
For the extant group we used the 17 species that survived in
the wild in Eurasia until present: reindeer, Rangifer tarandus;
saiga antelope, Saiga tatarica; fallow deer, Dama dama; roe
deer, Capreolus capreolus; European wild boar, Sus scrofa;
red deer, Cervus elaphus; moose, Alces alces; horse, Equus
ferus (including E. caballus, E. ferus, E. gmelini, E. latipes,
and E. lenensis taken as synonyms); kulan, Equus hemionus;
wolf, Canis lupus; red fox, Vulpes vulpes; brown bear, Ursus
arctos; dhole, Cuon alpinus; leopard, Panthera pardus; Arctic
fox, Alopex lagopus; and wolverine, Gulo gulo.
We tried to eliminate nomenclatural discrepancies as much
as possible. Species of horses, bison, and aurochs remain
rather enigmatic in their relation to modern species or time
of divergence into other species. Certainly, difficulties in spe-
cies description, especially in discrimination between the nu-
merous species of bison and/or horse and identification of
aurochs, may introduce errors into the database; however,
this analysis is offered as a broad guide of apparent differ-
ences in commonness of extinct and extant species in the sites
associated with humans. The case of the horse would be more
trivial if the many taxa used to describe its remains are, as
many authorities contend, only regional variants of the same
species. More detailed systematic analyses may produce
more accurate results, although we argue the current results
are robust enough to withstand any underlying taxonomic
mistakes.
Tests of possible sampling biases (body size, cave
environment) on commonness values
It is the case that ‘natural’ commonness could be altered
by taphonomic factors that were at play in bone preservation
in many of the sites we included in our database. A com-
plete survey of taphonomic biases for each of the 564 local-
ities in our data set is clearly impossible due to lack of
information. We would add here that risks imposed by
taphonomic factors are usually small in large data sets, given
the so-called Portfolio effect (Raia et al., 2005), which we
paraphrase as the relative risk associated with any given
taphonomic factor is decreased by the number of taphonomic
biases present in the dataset, hence by the number of local-
ities included.
Nonetheless, we tested for the effect on commonness
values of the two most relevant taphonomic factors: body
size and preservation in cave environments. Smaller bones
are less resistant to destruction by taphonomic agents because
of their higher surface to volume ratio. Consequently, larger
species are often overrepresented in fossil sites (Damuth,
1982). To consider whether body size influence commonness,
we calculated body size of the extinct species by applying re-
gression equations published in Damuth and MacFadden
(1990). The data for a few ‘‘living’’ species were not available
and we used a compilation of recent mammal body sizes
across continents (Smith et al., 2003). The complete lists of
data and body sizes can be found in Melore et al. (2007).
The regression analysis of body size against commonness
was performed on all species together and only on extinct
and living species, separately. A significant relationship would
suggest that large body sizes bias towards higher commonness,
whereas large species cannot be very common (Brown, 1984;
Gaston and Blackburn, 2000).
We also analyzed the most pervasive and important effect
of cave environments. Caves naturally provided shelters for
humans (Rolland, 2004) and other species (most notably the
carnivores: cave bear, cave lion, and cave hyena) and were
rather unsuitable for the megaherbivores or their bones accu-
mulations. Furthermore, accumulation of bones by carnivore
predators in caves could mistakenly increase commonness at
human-occupied sites for some species at the expense of
others. Similarly, the tendency of humans and some other spe-
cies to ‘‘frequent’’ caves might lead to overestimates of human
exploitation on these species. We discerned the layers with and
without human artifacts in a locality, where it was possible, ac-
cording to the latest updated versions of databases and cata-
logs (e.g., latest publication in Pangaea database; Paunovic
et al., 2001). We partitioned our localities into ‘‘cave’’ and
‘‘open-air’’ categories. Grottos and rock shelters were consid-
ered caves. We then performed a c-test to assess whether ar-
chaeological sites included more caves than expected, and
then computed commonness to be compared between caves
and open air sites with a Wilcoxon test.
772 D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
Table 1
Species commonness in archaeological and paleontological sites by time period. Body size is given in log
10
kg
Late middle Pleistocene Eemian Early and middle Weichselian Late Glacial Maximum Late Glacial HoloceneSpecies
paleo archaeo paleo archaeo paleo archaeo paleo archaeo paleo archaeo paleo archaeo
status mass
Alces alces 0.04 0.07 0.17 0.17 0.1 0.19 0.05 0.06 0.04 0.25 0.27 0.29 living 5.585
Alopex lagopus 0.01 0 0.04 0 0.11 0.21 0.14 0.45 0.11 0.23 0 0.03 living 3.798
Bison priscus 0.45 0.21 0.57 0.33 0.37 0.53 0.18 0.52 0.11 0.67 0.05 0.04 extinct 5.837
Bos primigenius 0.09 0.43 0.28 0.67 0.08 0.18 0 0.09 0.04 0.08 0.02 0.13 extinct 6.021
Canis lupus 0.1 0.57 0.3 0.67 0.23 0.66 0.27 0.45 0.11 0.43 0.32 0.32 living 4.568
Capreolus capreolus 0.02 0.64 0.26 1 0.11 0.27 0.05 0.09 0.04 0.28 0.37 0.56 living 4.364
Cervus elaphus 0.4 1 0.55 1 0.28 0.75 0.09 0.39 0.15 0.54 0.32 0.74 living 5.272
Coelodonta antiquitatis 0.19 0.07 0.21 0 0.37 0.52 0.27 0.55 0.11 0.13 0 0 extinct 6.429
Crocuta crocuta 0.09 0.29 0.26 0.33 0.21 0.47 0.14 0.09 0 0.03 0 0 extinct 5.009
Cuon alpinus 0.04 0.07 0 0 0 0.13 0 0 0 0 0.02 0.04 living 4.106
Dama dama 0.11 0.14 0.26 1 0.04 0 0 0 0 0 0 0 living 4.813
Equus ferus 0.49 0.36 0.17 0.33 0.34 0.62 0.55 0.61 0.15 0.64 0.2 0.5 living 5.699
Equus hydruntinus 0.02 0.21 0.04 0.33 0.04 0.15 0 0.09 0 0.02 0 0 extinct 5.362
Equus hemionus 0.02 0 0.02 0 0.03 0.15 0 0.24 0 0.21 0.05 0.12 living 5.322
Gulo gulo 0.02 0 0 0 0.07 0.2 0.14 0.24 0.07 0.18 0.15 0.06 living 4.29
Lynx lynx 0 0.14 0 0.17 0.07 0.19 0 0.03 0.07 0.03 0.12 0.03 living 4.254
Mammuthus primigenius 0.15 0.07 0.34 0 0.69 0.49 0.73 0.67 0.67 0.33 0.1 0 extinct 6.632
Megaloceros giganteus 0.13 0.29 0.43 0.67 0.15 0.22 0 0 0.11 0.02 0.07 0.01 extinct 5.589
Ovibos moschatus 0.02 0 0.02 0 0.25 0.05 0.32 0.06 0.11 0.05 0.05 0 extinct 5.496
Paleoloxodon antiquus 0.13 0.5 0.45 0.83 0.04 0.04 0 0 0 0 0 0 extinct 6.678
Panthera leo 0.29 0.5 0.32 0.67 0.27 0.35 0.14 0.09 0.07 0.1 0 0.01 extinct 5.262
Panthera pardus 0.01 0.5 0.02 0.67 0.03 0.13 0 0 0 0 0.05 0.04 living 4.778
Rangifer tarandus 0.11 0.14 0.15 0 0.34 0.47 0.23 0.64 0.22 0.59 0.05 0.09 living 4.935
Saiga tatarica 0.07 0 0.02 0 0.01 0.22 0 0.18 0 0.13 0 0.07 living 4.462
Stephanorhinus kirchbergensis 0.15 0.29 0.36 1 0.04 0.07 0 0 0 0 0 0 extinct 6.341
Stephanorhinus hemitoechus 0.04 0.21 0.21 0.33 0.03 0.01 0 0 0 0 0 0 extinct 6.429
Sus scrofa 0.07 0.29 0.21 0.83 0.1 0.16 0.05 0.09 0.07 0.08 0.24 0.69 living 5.069
Ursus arctos 0.04 0.43 0.23 0.33 0.18 0.38 0.05 0.27 0.07 0.26 0.44 0.34 living 5.223
Ursus spelaeus 0.04 0.36 0.17 0.67 0.23 0.44 0.14 0.09 0.07 0.07 0.02 0.06 extinct 5.439
Vulpes vulpes 0.13 0.5 0.17 0.33 0.2 0.52 0.09 0.24 0.11 0.21 0.34 0.44 living 3.778
773D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
Testing for human influence on commonness values
The northernmost human-occupied sites in our data reached
51N during the late MP, 52.87N during the Eemian (early
middle Palaeolithic or early Mousterian), 58.3N during the
early and middle Weichselian (late Middle Paleolithic or late
Mousterian and early Late Paleolithic), 60.35N during the
LGM (middle Late Paleolithic), 70.43N during the Late Gla-
cial (latest Late Paleolithic), and 65.07N during the Holocene
(Mesolithic and Neolithic). Many other species resided far
north of humans (see below). Theoretically, it is possible
that the northern ‘‘human free’’ areas acted as a refuge to spe-
cies ‘‘trying to avoid’’ human overexploitation in the south, or
for those already extirpated there (as suggested by Surovell
et al., 2005, for proboscideans). To consider the importance
of these northern areas as potential refugia for large mammal
species, we calculated species commonness both within and
beyond the northernmost limit of human occupation, regard-
less of whether they occurred in archaeological or paleontolo-
gial sites, and then compared the extinct and extant species
groups, using a nonparametric (Mann-Whitney) test. For these
following analyses we used only five time periods, excluding
the 130e115 ka as mentioned earlier because of the extremely
low frequency of archaeological sites.
Many late Pleistocene species of the cold-adapted Mammu-
thus primigenius assemblage or the mammoth steppe roamed
mostly northward of human settlements. However, the use of
northern refugia was by no means the only way to avoid people
for species were able to tolerate very cold conditions typical of
northern latitudes. Human-free areas were also available within
human ranges (Stewart and Lister, 2001; Surovell et al., 2005).
Thus, we calculated species commonness separately in archae-
ological and paleontological sites to test if and which species
were abundant in human-occupied sites (Table 1). We also
computed a ‘‘human preference’’ (HP) or human association
factor for each time period by calculating the difference in
commonness between archaeological and paleontological sites
for each species. Any strong association between humans and
currently extinct species (indicated by higher than expected
commonness in archaeological sites or higher HPs) may de-
pend on human selection and exploitation of these animals;
yet, it could also depend on geographical distribution and a pos-
sible habitat selection of the latter. An obvious bias would oc-
cur when comparing species with different geographic ranges
than humans, because latitudinal differences in commonness
within a species range can be substantial since species abun-
dance, and, hence, commonness is expected to decrease toward
the periphery of its range (Lawton, 1993; Brown et al., 1995;
Blackburn et al., 1999). Accordingly, we corrected for the spe-
cies distribution effects by taking the difference in mid-latitude
range point between the focal species and humans in each age
group. These differences in latitude were regressed against the
differences in HP (commonness) using a least squares regres-
sion analysis. HP residuals were then compared between living
and extinct species for each time period by a Mann-Whitney
test. If humans effectively altered commonness patterns of ex-
tinct species via hunting, HP residuals of extinct species should
be significantly higher (i.e., they should be more common than
expected in archaeological sites). The opposite result (i.e., liv-
ing species being more common than expected in archaeolog-
ical sites) would not support the influence of humans on
megafaunal extinction. Apart from geographical distribution,
habitat selection could similarly bias the HP residuals. Thus,
it is important to take into account if human-associated species
(i.e., species with high HP residuals) were present in the same
habitats as humans. If they were not, then the interpretation
that their abundance in archaeological sites was caused by
human selection on them would be strengthened.
Results
Archaeological versus paleontological sites
The number of human-occupied sites increases consider-
ably from the late MP into the Weichselian and Holocene.
The temporal shifts in mean latitudinal distribution demon-
strate that localities bearing signs of human intervention are
continuously situated southward of the localities with purely
paleontological associations [ANOVA effect of age group
F
5,552
¼21.98, p <0.001; effect of human presence F
1,552
¼
109.58, p <0.001; the effect of interaction between age group
and human presence F
5,552
¼3.51, p ¼0.004 (Fig. 1)]. This
suggests that earlier in time and prior to human northward
spread, purely paleontological localities would be situated
4,1682,7612,2337,1854768,914
A
g
e
g
roups
Holocene
LG
LGM
ea-md Weichs
Eem
late MP
Mean latitude
80
70
60
50
40
30
Localities
paleontological
with humans
n =
Fig. 1. Temporal changes in mean latitudinal distribution in archaeological and
paleontological localities. Error bars represent 95% confidence limits. Age
group are defined in Methods (some abbreviations: late MP, late middle Pleis-
tocene; Eem, Eemian; ea-md Weichs, early and middle Weichselian
glaciation).
774 D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
northwards of sites associated with humans. At all times, some
wild species were better adapted to cold conditions than were
humans and humans were occupying the southern end of the
spectrum of other species’ occupation.
In general, the human-occupied localities show a tendency
to shift to the north during the Weichselian or the late Paleo-
lithic (ANOVA F
5,261
¼18.71, p <0.001). Post-hoc Tukey dif-
ferences between age groups indicate that the most significant
Table 2
Results of regression of body size versus total commonness by time period
Age group Status Intercept Slope r
2
Fp
Late MP extinct 0.032077 0.000785 0.000233 0.002568 0.960496
living 0.09715 0.027049 0.189786 3.513629 0.080475
Eemian extinct 0.04209 0.014371 0.125931 1.584816 0.234129
living 0.04793 0.015703 0.144577 2.535185 0.132184
Early and middle Weichselian extinct 0.049699 0.00258 0.0034 0.03753 0.849922
living 0.01361 0.009773 0.063682 1.020206 0.328487
Last Glacial Maximum extinct 0.07935 0.019122 0.0673 0.793717 0.392049
living 0.02285 0.011927 0.0406 0.634773 0.438039
Late Glacial extinct 0.002889 0.003824 0.004616 0.05101 0.825457
living 0.05572 0.020115 0.114641 1.942272 0.183728
Holocene extinct 0.029179 0.00381 0.07437 0.883797 0.36735
living 0.06766 0.025638 0.089894 1.481594 0.242329
Middle Pleistocene
0
0.1
0.2
0.3
0.4
0.5
34567
3456734567
3456734567
34567
Log body size (g)
Commonness
Eemian
0
0.2
0.4
0.6
0.8
Log body size (g)
Early and Middle Weichselian
0
0.1
0.2
0.3
0.4
0.5
0.6
0
0.1
0.2
0.3
0.4
0.5
0.6
0
0.1
0.2
0.3
0.4
0.5
0.6
0
0.1
0.2
0.3
0.4
0.5
0.6
Log body size (g)
Commonness
LGM
0.7
0.7
Log body size (g)
Late Glacial
Lo
g
bod
y
size
(g)
Commonness
Commonness CommonnessCommonness
Holocene
Log body size (g)
Fig. 2. Species body size plotted versus commonness. Extinct species are represented by solid circles. Living species are represented by open circles. The
difference in body size between extinct and extant species is highly significant (Mann and Whitney U ¼18.000; p <0.0001).
775D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
northward shifts occurred between the late middle Pleistocene
and Weichselian ( p¼0.026) and between the early and
middle Weichselian and LGM ( p¼0.004). The shift between
the LGM and LG is not significant. The southward Holocene
trend possibly occurs due to a paucity of northern sites in our
dataset in this epoch.
Analysis of potential biases
Body size did not exert a significant effect on species
commonness (Table 2). The relationship was nonsignificant
in spite of the extreme commonness of the woolly mammoth
during the early and middle Weichselian or LGM. We antic-
ipate here that the woolly mammoth shows the least associa-
tion with humans (see below), and consequently, its high
commonness is further evidence that its scarce association
with humans is probably genuine. Most extinct megafauna
were significantly larger than their living counterparts
(Fig. 2). Thus, we point out that if the effect of body size
on commonness was indeed larger than with our interpreta-
tion, it should have favored the presence of large, extinct spe-
cies in association with people.
The human association with caves is significant (ctest,
c¼48.87, df ¼5, p <0.001; Table 3 ). Interestingly, this as-
sociation between humans and caves is very strong during
the early and middle Weichselian. Differences in common-
ness between caves and open sites are not pronounced
(Tab l e 4 ). Wilcoxon tests indicate living species to be more
common than expected in caves in all time periods, but
significantly so only during early and middle Weichselian.
The same does not apply to extinct species (Table 4 ). This
discrepancy should have increased the association of living
species with people, at least for the early and middle
Weichselian period.
Human interference
All extinct species were much more common north of the
human geographic range limit than they were within it. This
difference is especially apparent since the early and middle
Weichselian (Table 5). On the contrary, living species were
always more common within than beyond the human range.
Differences between extinct and living species are generally
nonsignificant both within and beyond human range, except
for the obvious rarity of extinct species during the Holocene
period. This commonness pattern is probably explained by
the better adaptation of majority of extinct species to cold con-
ditions as mentioned earlier, although it is also consistent with
human overexploitation (Table 5). Except for the late MP and
the early and middle Weichselian, latitudinal differences in
distribution between humans and other large mammals were
poor predictors of the difference in commonness between
archaeological and paleontological sites, especially in the
more recent periods (Table 6).
The analysis of HP residuals indicates that the differences
between the extinct and living species are significant during
the Weichselian glaciation. Yet, contrary to the hypothesis of
human-induced extinction, Mann-Whitney U tests performed
on the HP residuals indicate that the species with the highest
relative commonness in archaeological sites (high HP resid-
uals) are extant taxa (Tables 7 and 8), except for the extinct
steppe bison Bison priscus.
Contrary to the surviving late Pleistocene cold-adapted
reindeer and saiga that were strongly associated with humans,
the extinct cold-adapted ungulate species (woolly mammoth,
woolly rhino, musk ox) were extremely rare in human-occu-
pied sites, as expected by their distribution. The commonness
of extinct warm-adapted species (straight-tusked elephant,
hippopotamus, Merck’s and narrow-nosed rhinoceri, giant
deer, aurochs) is very similar with and without humans, again
indicating little human influence on them.
Table 3
Number and type of fossil assemblage and site type by time period
Age group Locality
type
Assemblage type Total
paleontological archaeological
Late middle
Pleistocene
open-air site 61 6 67
cave 18 8 26
unknown 10 0 10
Eemian open-air site 25 4 29
cave 11 2 13
unknown 11 0 11
Early and middle
Weichselian
open-air site 49 30 79
cave 19 52 71
unknown 3 3 6
Last Glacial
Maximum
open-air site 17 27 44
cave 5 4 9
unknown 0 2 2
Late Glacial open-air site 18 50 68
cave 7 9 16
unknown 2 2 4
Holocene open-air site 19 27 46
cave 19 33 52
unknown 3 8 11
Table 4
Results of Wilcoxon tests performed to compare species commonness between caves and open air sites. Positive signs indicate that commonness was higher in
caves
Age group Extinct Living
Z p sign of difference Z p sign of difference
Late MP 1.57339 0.115629 1.49136 0.135867 þ
Early and middle Weichselian 0.62757 0.530285 þ2.17177 0.029873 þ
LGM 0.1777 0.858955 0.38437 0.700703 þ
Late Glacial 0.05096 0.959354 þ0.84748 0.396726 þ
Holocene 0 1 0.84748 0.396726 þ
776 D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
The commonness of the extinct carnivores declined at
a similar rate in both paleontological and archaeological sites.
Among the surviving carnivores in the human-occupied sites,
the commonness of the brown bear and red fox declined dur-
ing the Weichselian (especially LGM and LG), but the com-
monness of the arctic fox increased during the late
Weichselian (LGM). The wolves were very abundant through-
out all stages. The carnivores’ HP residuals were very close to
zero for most of the late middle to late Pleistocene. Yet, the
carnivores become rarer in the Holocene archaeological sites,
perhaps, indicating that humans become increasingly able to
keep them out of their sites (Table 9).
The commonness in the extinct species is quite similar in
paleontological and archaeological sites within human ranges,
suggesting that humans did not intentionally avoid bringing
bones of the large species (that represent the extinct category)
back to their site (Table 10).
Discussion
Human distribution
Long before 100,000 years ago, humans coexisted with the
megafauna without causing any extinction. In western Europe,
there is archaeological evidence of human occupation as early
as about 800,000 years ago (Carbonell et al., 1995; Bermu
´dez
de Castro et al., 1997; Finlayson, 2004, 2005). The Acheulean
cultures are known from the Caucasus from about 583,000 years
ago (Baryshnikov, 1987; Molodkov, 2001). Humans were present
in northern Eurasia (Eurasian Plains) at about 45,000e40,000
years ago (Finlayson, 2005; Anikovich et al., 2007). Anatomi-
cally modern humans inhabited western Europe at about
34,000e36, 000 years ago (Smith et al., 1999; Paunovic et al.,
2001; Trinkaus et al., 2003)or31,000yearsago(Finlayson
and Carrio
´n, 2007). The earliest modern human occupation or
early Late Paleolithic occupation for southern Siberia is recorded
at 43e39,000 yrs ago and they appear to have occupied all of
northern Asia by 13,000 yr ago (Vasil’ev et al., 2002).
The difference in distribution between humans and other spe-
cies suggests that humans were rather warm-loving species, con-
sidering their place of origin in tropical Africa and the late
survival of archaic hominins in Southeast Asia (Finlayson,
2004, 2005). The majority of the large mammals that survived
the late Pleistocene were also adapted to temperate environments.
Nevertheless, humans dispersed northwards from the late MP to
the Weichselian (Praslov, 1984; this study). During the LGM hu-
mans moved northwards along the main Siberian rivers, perhaps,
fleeing from the deserted steppes in the south (Madeyska, 1992).
Much evidence exists on human/animal interactions in
a broad sense. Various mammoth products dated to 20e
14,000 yr BP are known from many countries; including Uk-
raine, Poland, and the Czech Republic (Pidoplichko, 1998;
Pe
´an, 2001). However, the evidence for hunting and/or scav-
enging may be much older as humans could also collect the
bones of already long dead animals (Vereshchagin and Barysh-
nikov, 1984; Vasil’ev, 2003). In this case, not all species from
the archaeological sites are necessarily dated to the corre-
sponding cultural stage and, thus, may increase the bias to-
wards human influence. Therefore, it is all the more striking
to find no or little effect of human exploitation in our study.
Biases due to cave sites and body size
Large body size did not translate into artificially high
commonness values. This structure of the commonness/
body size plot is very similar to the abundance/body size
Table 5
Commonness of extant and extinct species within and beyond human geographic range
Age group 400e130 ky ago 115e24 ky BP 24e15 ky BP 15e10 ky BP Holocene
Commonness (site occupancy) outside human range
Extinct 0.132 0.212 0.142 0.156 0.112
Living 0.057 0.055 0.052 0.026 0.039
p¼0.869083 0.650058 0.650058 0.156925 0.001473
Commonness (site occupancy) within human range
Extinct 0.152 0.270 0.197 0.128 0.032
Living 0.168 0.281 0.237 0.221 0.232
p¼0.015382 0.680075 0.772946 0.408285 0.591685
Species mid-range point (in degrees of latitude, range in parentheses)
Extinct 49.3 51.9 56.0 55.2 52.1
(70e39.0) (76e38.1) (75.5e40.1) (75.3e39.4) (74.5e40.3)
Living 46.7 49.3 53.3 53.0 48.5
(68.0e39.0) (74.5e8.1) (75.3e40.1) (70.5e39.4) (74.5e40.2)
p¼0.1 0.2 0.4 0.2 0.1
Table 6
Results of the least-squares regression of HP factor (commonness differences
between archaeological and paleontological sites) against differences in mid-
latitude geographical range between large mammal species and humans by
time period
Time period Intercept Slope r
2
Significance
Late MP 0.012 0.04 0.283 0.003
Early and middle Weichselian 0.04 0.04 0.448 0.001
LGM 0.001 0.003 0.122 0.061
LG 0.001 0.002 0.012 0.268
Holocene 0.005 0.001 0.051 0.153
777D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
plot in the macroecological literature (cf. Brown and Maurer,
1987; Gaston and Blackburn, 2000). Similarly, the significant
association of humans and cave environment only during the
early and middle Weichselian or the late Middle Paleolithic
(late Mousterian) and the early Late Paleolithic does not en-
tail relevant alteration of commonness of human-associated
fauna.
Intriguingly, this is consistent with the evidence that regular
cave occupation begins during the later Acheulean and be-
comes frequent during the Middle and Late Paleolithic (Roll-
and, 2004). Only during the late Paleolithic did humans
probably start to dominate open landscapes considering vast
territories of northern Eurasia given that the majority of the
Late Paleolithic sites in Siberia, for example, are open-air sites
on river terraces or not very high mountains (Derevianko et al.,
1997; Reinhart-Waller, 2000; Goebel, 1999). Beauval et al.
(2005) suggest that mobility becomes an important component
of the middle Late Paleolithic hunter-gatherer adaptation
across Europe.
However, according to our analysis, species commonness
was not significantly different between caves and open air
sites. Abundance of extant large mammals in caves seems lit-
tle influenced by human activity (see below). Instead, it seems
to reflect the preference of these species for warmer habitats
and their smaller body size. Indeed, the penchant of extant
species for caves becomes nonsignificant in the LGM and, es-
pecially, Late Glacial and Holocene, when most megafauna
had already vanished and climate became warmer. The
significant ‘‘preferences’’ of humans and extant species for
caves during the early and middle Weichselian certainly can
indicate that humans rarely brought megamammal bones
home to caves.
Intuitively, the relative rarity of remains of extinct species
in many archaeological sites could simply reflect a situation
in which remains of these heavy animals were seldom brought
back to any human site. Yet, the extinct large mammals were
equally rare in paleontological sites at similar latitudes. This
latter observation indicates that the rarity of extinct species
within human ranges (and not just in archaeological sites)
was a real ecological phenomenon, and strongly argues against
the notion that humans were responsible for their demise. Our
data suggest that if a direct avoidance of humans by these spe-
cies via gathering in northern areas was possible, it was not
due to human overexploitation in the south.
The association between extant species and humans was
particularly strong in middle-sized species like deer (reindeer,
roe deer, red deer), horse (horse, kulan), saiga, and, during the
Holocene, the wild boar. Although this association is partially
driven by some proclivity to inhabit caves or be collected there
by carnivore predators it holds for open-air sites as well, and
for some extant species (e.g., saiga, reindeer, and moose),
which were extremely rare in caves. Wolves are extremely
common at archaeological sites which may indicate the long
association between ‘‘man and his dog’’ and a confirmed do-
mestication of a Paleolithic dog known from Eurasia (Rein-
hart-Waller, 2000).
The exploitation of similar habitats was not responsible for
the strong human association with some species in our study.
Indeed, those most heavily-exploited by humans include spe-
cies of different habitat; forest species, such as red deer and
wild boar (although both can be met outside forests), and typ-
ical steppe inhabitants, such as reindeer, horse, and steppe bi-
son. In addition, typical steppe species appear both among the
human-preferred (bison, horse, reindeer) and the human-
avoided (woolly mammoth, musk ox) categories.
The tendency of commonness either to increase or to de-
crease is consistent between archaeological and paleontologi-
cal sites for most extinct megafauna (this study; Grayson and
Meltzer, 2002). A ‘‘peaked’’ trajectory in which species are
initially relatively uncommon, then increase in commonness,
and then become less common prior to extinction is consistent
with the ‘‘natural’’ course for the species through time (Raia
et al., 2006) and casts further doubts on the hypothesis that
humans doomed these megafauna to extinction.
Table 7
Comparison of HP residuals between extinct and living species
Age group Status nMean rank Mann-
Whitney U
Zp
Late MP extinct 13 15.154 106 0.188 0.851
living 17 15.765
total 30
Early and middle
Weichselian
extinct 13 11.462 58 2.197 0.028
living 17 18.588
total 30
LGM extinct 9 9.000 36 1.503 0.133
living 13 13.231
total 22
Late Glacial extinct 10 10.300 48 1.288 0.198
living 14 14.071
total 24
Holocene extinct 8 13.125 51 0.581 0.561
living 15 11.400
total 23
Table 8
Herbivore species with the three highest and lowest HP residuals by time period
Late middle Pleistocene Early and middle Weichselian LGM Late Glacial Holocene
high Capreolus capreolus Cervus elaphus Rangifer tarandus Bison priscus Sus scrofa
Ovibos moschatus Saiga tatarica Equus hemionus Equus ferus Cervus elaphus
Bos primigenius Equus ferus Bison priscus Equus hemionus Equus ferus
HPres
low Bison priscus Dama dama Ovibos moschatus Coelodonta antiqutatis Capreolus capreolus
Equus ferus Stephanorhinus hemitoecus Equus ferus Megaloceros giganteus Megaloceros giganteus
Dama dama Mammuthus primigenius Mammuthus primigenius Mammuthus primigenius Alces alces
778 D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
Only the data for the extinct steppe bison may indicate
a disproportionate selection by humans although more suffi-
cient and recently updated data are needed. The bison be-
comes more common in archaeological sites with time but
declines in paleontological sites. Shapiro et al. (2004) showed
that the bison’s genetic decline in Beringia started at 37,000
years ago. However, a genetic bottleneck does not always
lead to extinction. A good example in medium-sized mammals
are muskoxen that survived in North America despite suffering
a great loss of genetic diversity at the Pleistocene-Holocene
boundary that leaves them with very low genetic variability
(MacPhee et al., 2005). Other extant cold-adapted species
may reflect this drastic change in ecological conditions in their
genetic history. A study of the Late Paleolithic sites in the Bai-
kal and Altai regions indicated that the bison’s commonness
increased significantly from the late middle Weichselian
(30e24 ka) to the LGM (Pushkina, 2006), while humans
were present in southern Siberia slightly prior to 37,000 years
ago (Vasil’ev et al., 2002; Shapiro et al., 2004).
In our study too the extinction of the steppe bison appears
to be related to humans. Human subsistence activities in Sibe-
ria heavily relied on bison, reindeer, and horse during the
Late Paleolithic (Vasil’ev, 2003). The commonness of the ex-
tinct bison and surviving ungulate species (reindeer, horse,
saiga, and red deer) in archaeological sites strongly indicates
their intensive utilization by humans, especially during the
Weichselian glaciation or the late Middle and Late Paleo-
lithic. This can be related to increased population density (ar-
chaeological sites outnumber paleontological sites for the first
time during the early and middle Weichselian) and to improv-
ing hunting technologies, such as the appearance of the
microblade-producing populations of highly mobile hunters
across Siberia after the LGM who concentrated on a single
prey species of large or medium size (Goebel, 2002). The mi-
croblade technology appeared with the exploitation of smaller
less gregarious species probably because the big game species
were disappearing (Tankersley and Kuzmin, 1998). Overall, it
appears that humans were generalized, albeit effective, pred-
ators and efficiently concentrated on the most abundant prey
species, most of which survive today.
Mesolithic hunting was similar to Paleolithic hunting (Rein-
hart-Waller, 2000), but the range of hunted species was reduced
for the Mesolithic hunters (Tankersley and Kuzmin, 1998).
However, agriculture and domestication that appeared during
the Mesolithic could increase the pressure on exploitable ani-
mals, on one hand, and relieve the extinction pressure of some
species, on the other. In our case, almost all heavily-exploited
species that survived were either forest inhabitants or domesti-
cated. The reindeer was domesticated around 5,000 years
ago (Beringia series, 1992), and the horse rather late around
4e3,000 years ago (Reinhart-Waller, 2000; Bunzel-Dru
¨ke,
2001). The steppe bison gave rise to the two extant species:
Bison bonasus and Bison bison (Ricciuti, 1973). In striking con-
trast to domestication and long coexistence of ungulates and
people in Eurasia, prey naivety to humans has been claimed to
be relevant to the late Pleistocene faunal extinction on other
continents and later on islands (Burney and Flannery, 2005).
In summary, the evidence we found argues against the pos-
sibility that humans caused the late Pleistocene megafaunal
extinctions of Eurasia. In northern Eurasia extinctions at the
Pleistocene-Holocene boundary coincided with a rapid vegeta-
tion shift towards more humid and closed conditions (Guthrie,
1984, 1995; Sher, 1997). Highly productive grasslands or
steppe-tundra, that occupied vast territories in northern Eura-
sia during the late Pleistocene, disappeared in the Holocene
giving space to zonal vegetation, tundra in the north, and bo-
real forest or taiga in the south (Guthrie, 1984). Consistently,
typically forest taxa such as the moose, red deer, roe deer, and
wild boar, spread after the Weichselian glaciation. Likely, this
extensive environmental change (and the vanishing of ‘‘mam-
moth-steppe’’ in particular) was the most important determi-
nant of megafaunal extinction.
Conclusions
The relative commonnes of the large mammals of Eurasia
were influenced by human activity to some extent. People be-
came increasingly able to hunt abundant prey species, many of
which, however, are still living. Humans became able to ex-
clude large carnivores from their sites or defend their homes.
By the latest Late Paleolithic populations of large mammals of
the ‘‘mammoth-steppe’’ were already suffering from the deteri-
oration and contraction to the north of their preferred habitat,
while humans appeared to show little interest in the now-extinct
Table 9
Differences in HP residuals between herbivores and carnivores by time period
Late MP Early and middle Weichselian LGM Late Glacial Holocene
Mann-Whitney U 85 75 52 49 17
Mean Carnivore 17.273 18.182 12.000 10.444 7.200
Rank Herbivore 14.474 13.947 11.214 13.733 15.692
Asymp. Sig. (2-tailed) 0.401 0.204 0.785 0.270 0.003
Exact Sig. [2*(1-tailed Sig.)] 0.420 0.216 0.815 0.290 0.002
Table 10
Paired comparisons of extinct species commonness between archaeological
and paleontological sites, limited to sites located within human geographic
range
Middle
Pleistocene
Early and middle
Weichselian
LGM Late
Glacial
Holocene
Z0.319 0.339 0.276 0.143 0.052
p 0.750 0.734 0.783 0.886 0.958
779D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
species, even when a conservative archaeological approach
was used that should have favored finding human influence
on extinct fauna. Only the extinct steppe bison appears to
have been negatively influenced by humans. Our findings are
mainly consistent with the climatic explanation of the late
Pleistocene extinctions in Eurasia.
Acknowledgements
For the valuable comments on the manuscript and impor-
tant suggestions that greatly improved the quality of the man-
uscript, we would like to express our gratitude to Susan Anto
´n,
Paul Palmqvist, Robert Feranec, Mikael Fortelius, and two
anonymous reviewers.
References
Alroy, J., 2001. A multi-species overkill simulation of the end-Pleistocene
megafaunal mass extinction. Science 292, 1893e1896.
Anikovich, M.V., Sinitsyn, A.A., Hoffecker, J.F., Holliday, V.T.,
Popov, V.V., Lisitsyn, S.N., Forman, S.L., Levkovskaya, G.M.,
Pospelova, G.A., Kuz’mina, I.E., Burova, N.D., Goldberg, P.,
Macphail, R.I., Giaccio, B., Praslov, N.D., 2007. Early Upper Paleolithic
in eastern Europe and implications for the dispersal of modern humans.
Science 315, 223e226.
Azzaroli, A., de Giuli, C., Ficcarelli, G., Torre, D., 1988. Late Pliocene to early
Mid-Pleistocene mammals in Eurasia: faunal succession and dispersal
events. Palaeogeogr. Palaeoclimatol. Palaeoecol. 66, 77e100.
Barnosky, A.D., Koch, P.L., Feranec, R.S., Wing, S.L., Shabel, A.B., 2004. As-
sessing the causes of late Pleistocene extinctions on the continents. Science
306, 70e75.
Baryshnikov, G.F., 1987. Mammals of the Caucasus of the early Paleolithic.
Trudy Zool. Inst. AN. SSSR 68, 3e20 (in Russian).
Beauval, C., Maureille, B., Lacrampe-Cuyaube
`re, F., Serre, D., Peressinotto, D.,
Bordes, J.-G., Cochard, D., Couchoud, I., Dubrasquet, D., Laroulandie, V.,
Lenoble, A., Mallye, J.-B., Pasty, S., Primault, J., Rohland, N., Pa
¨a
¨bo, S.,
Trinkaus, E., 2005. A late Neandertal femur from Les Rochers-
de-Villeneuve, France. Proc. Natl. Acad. Sci. 102 (20), 7085e7090.
Beck, M.W., 1996. On discerning the cause of late Pleistocene megafaunal ex-
tinctions. Paleobiology 22 (1), 91e103.
Beringia series, 1992. Beringia Natural History Notebook Series, September,
1992. Available online at: http://www.nps.gov/bela/html/rangifer.htm.
Bermu
´dez de Castro, J.M., Arsuaga, J.L., Carbonell, E., Rosas, A.,
Martı
´nez, I., Mosquera, M., 1997. Hominid from the Lower Pleistocene
of Atapuerca, Spain: possible ancestor to Neandertals and modern humans.
Science 276, 1392e1395.
Blackburn, T., Gaston, K.J., Quinn, R.M., Gregory, R.D., 1999. Do local abun-
dance of British birds change with proximity to range edge? J. Biogeogr.
26, 493e505.
Brook, B.W., Bowman, D.M.J.S., 2002. Explaining the Pleistocene megafau-
nal extinctions: models, chronologies, and assumptions. Proc. Natl.
Acad. Sci. 99, 14624e14627.
Brook, B.W., Bowman, D.M.J.S., 2004. The uncertain blitzkrieg of Pleistocene
megafauna. J. Biogeogr. 31, 517e523.
Brown, J.H., 1984. On the relationship between abundance and distribution of
species. Am. Nat. 124, 255e279.
Brown, J.H., Maurer, B.A., 1987. Evolution of species assemblages: effects of
energetic constraints and species dynamics on the diversification of the
North American avifauna. Am. Nat. 130 (1), 1e17.
Brown, J.H., Mehlman, D.W., Stevens, G.C., 1995. Spatial variation in abun-
dance. Ecology 76, 2028e2043.
Bulte, E.H., Horan, R.D., Shogren, J.F., 2006. The economics of Pleistocene
megafauna extinction: early humans and the overkill hypothesis. J. Econ.
Behav. Organiz. 59, 297e323.
Bunzel-Dru
¨ke, M., 2001. Ecological substitutes for wild horse (Equus ferus
Boddaert, 1785 ¼E. przewalskii Poljakov, 1881) and aurochs (Bos primi-
genius Bojanus, 1827). Natur- und Kulturlandschaft, Ho
¨xter/Jena. Band 4
(WWF Large Herbivore Initiative).
Burney, D.A., Flannery, T.F., 2005. Fifty millennia of catastrophic extinctions
after human contact. Trends Ecol. Evol. 20, 395e401.
Carbonell, E., Bermudez de Castro, J.M., Arsuaga, J.L., Diez, J.C., Rosas, A.,
Cuenca-Bescos, G., Sala, R., Mosquera, M., Rodriguez, X.P., 1995. Lower
Pleistocene hominids and artifacts from Atapuerca-TD6 (Spain). Science
269, 826e832.
Cardillo, M., Lister, A.M., 2002. Death in the slow lane. Nature 419, 440e441.
Currant, A., Jacobi, R., 2001. A formal mammalian biostratigraphy for the
Late Pleistocene of Britain. Quatern. Sci. Rev. 20, 1707e1716.
Damuth, J., 1982. Analysis of the preservation of community structure in as-
semblages of fossil mammals. Paleobiology 8 (4), 434e446.
Damuth, J., MacFadden, B.J. (Eds.), 1990. Body Size in Mammalian Paleobi-
ology. Cambridge University Press.
Derevianko, A.P., Markin, S.V., Vasiliev, S.A., 1997. Introduction and the Fun-
damentals of Palaeolithic science, vol. 1. Institute of Archaeology and Eth-
nography, Novosibirsk. Siberian Archaeological Herald.
Diamond, J., 2005. Collapse: How Societies Choose to Fail or Succeed. Pen-
guin Group, New York.
Finlayson, C., 2004. Neanderthals and Modern Humans: An Ecological and
Evolutionary Perspective. Cambridge Univeristy Press, Cambridge.
Finlayson, C., 2005. Biogeography and evolution of the genus Homo. Trends
Ecol. Evol. 20, 457e463.
Finlayson, C., Carrio
´n, J.S., 2007. Rapid ecological turnover and its impact on
Neanderthal and other human populations. Trends Ecol. Evol. 22, 213e
222.
Fisher, D.C., 1984. Taphonomic analysis of late Pleistocene mastodon occur-
rences: evidence of butchery by North American Paleo-Indians. Paleobiol-
ogy 10, 338e357.
Fortelius, M., (coordinator) 2007. Neogene of the Old World Database of
Fossil Mammals (NOW). University of Helsinki. Available online at:
http://www.helsinki.fi/science/now/.
Gaston, K.J., Blackburn, T.M., 2000. Patterns and Processes in Macroecology.
Blackwell Science Ltd., Cambridge.
Goebel, T., 1999. Pleistocene human colonization of Siberia and peopling of
the Americas: an ecological approach. Evol. Anthropol. 8 (6), 208e227.
Goebel, T., 2002. The ‘‘Microblade adaptation’’ and recolonization of Siberia
during the Late Upper Pleistocene. In: Elston, R.G., Kuhn, S.L. (Eds.),
Thinking small: Global perspectives on microlithization. Archaeological
Papers of the Am. Anthropol. Assoc. 12, 117e131.
Graham, R.W., Lundelius Jr., E.L., 1984. Coevolutionary disequilibrium and
Pleistocene extinctions. In: Martin, P.S., Klein, R.G. (Eds.), Quaternary
Extinctions: A Prehistoric Revolution. University of Arizona Press, Tuc-
son, pp. 211e222.
Grayson, D.K., 2001. The archaeological record of human impacts on animal
populations. J. World Prehist. 15, 1e68.
Grayson, D.K., 2006. Ice age extinctions. Q. Rev. Biol. 81, 259e264.
Grayson, D.K., Meltzer, D.J., 2002. Clovis hunting and large mammal extinc-
tion: a critical review of the evidence. J. World Prehist 16, 313e359.
Guthrie, R.D., 1984. Mosaics, allelochemicals and nutrients: an ecological the-
ory of Late Pleistocene megafaunal extinctions. In: Martin, P.S.,
Klein, R.G. (Eds.), Quaternary Extinctions: A Prehistoric Revolution. Uni-
versity of Arizona Press, Tucson, pp. 259e298.
Guthrie, R.D., 1995. Mammalian evolution response to the Pleistocene-Holo-
cene transition and the break-up of the mammoth steppe: two case studies.
Acta Zool. Crac. 38 (1), 139e154.
Guthrie, R.D., 2003. Rapid body size decline in Alaskan Pleistocene horses
before extinction. Nature 426, 169e171.
Guthrie, R.D., 2006. New carbon dates link climatic change with human col-
onization and Pleistocene extinctions. Nature 441, 207e209.
Jernvall, J., Fortelius, M., 2002. Common mammals drive the evolutionary in-
crease of hypsodonty in the Neogene. Nature 417, 538e540.
Jernvall, J., Fortelius, M., 2004. Maintenance of trophic structure in fossil
mammal communities: site occupancy and taxon resilience. Am. Nat.
164 (5), 614e624.
780 D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
Johnson, C.N., 1998. Species extinction and the relationship between distribu-
tion and abundance. Nature 394, 272e274.
Johnson, C.N., 2002. Determinants of loss of mammal species during the late
Quaternary ‘megafauna’ extinctions: life history and ecology, but not body
size. Proc. R. Soc. Lond. B. 269, 2221e2227.
Johnson, C.N., 2005. What can data on the late survival of Australian mega-
fauna tell us about the cause of their extinction? Quatern. Sci. Rev. 24,
2167e2172.
Kahlke, R.-D., 1999. The History of the Origin, Evolution and Dispersal of the
Late Pleistocene Mammuthus-Coelodonta Faunal Complex in Eurasia
(Large Mammals). Mammoth site of Hot Springs, Fenske Companies,
Rapid City, USA.
Khromov, A.A., Arkhangelskij, M.S., Ivanov, A.V., 2001. Large Quaternary
mammals of the Central and Lower Volga region. International University
of Nature, Society and Human, Dubna.
Kvasov, D.D., 1977. An increase in climate moisture at the Pleistocene-
Holocene boundary as a cause of mammoth extinction. Trudy Zool. Inst.
AN SSSR 73, 71e76 (in Russian).
Lawton, J.H., 1993. Range, population abundance and conservation. Trends
Ecol. Evol. 8, 409e413.
Lister, A., Sher, A., 1995. Ice cores and mammoth extinction. Nature 378,
23e24.
MacPhee, R.D.E., Tikhonov, A.N., Mol, M., Greenwood, A.D., 2005. Late
Quaternary loss of genetic diversity in muskox (Ovibos). BMC Evol.
Biol. 5, 49.
Madeyska, T., 1992. Human occupation of the Old World during the last
Glaciation. In: Frenzel, B., Pe
´csi, M., Velichko, A.A. (Eds.), Atlas of
Paleoclimates and Paleoenvironments of the Northern Hemisphere, Late
Pleistocene-Holocene. Geographical Research Institute, Hungarian Acad-
emy of Sciences, Budapest, Gustav Fischer Verlag, Budapest-Stuttgart,
pp. 130e131.
Martin, P.S., 1973. The discovery of America. Science 179, 969e974.
Martin, P.S., 1984. Prehistoric overkill: the global model. In: Martin, P.S.,
Klein, R.G. (Eds.), Quaternary Extinctions: A Prehistoric Revolution. Uni-
versity of Arizona Press, Tucson, pp. 354e403.
Melore, C., Raia, P., Barbera, C., 2007. Effect of predation on prey abundance
and survival in Plio-Pleistocene mammalian communities. Evol. Ecol. Res.
9, 121.
Moen, R.A., Pastor, J., Cohen, Y., 1999. Antler growth and extinction of Irish
elk. Evol. Ecol. Res. 1, 235e249.
Molodkov, A., 2001. ESR dating evidence for early man at a Lower Palaeo-
lithic cave-site in the Northern Caucasus as derived from terrestrial
mollusc shells. Quatern. Sci. Rev. 20, 1051e1055.
Nowak, R.M., 1999. Walker’s Mammals of the World, sixth ed. The Johns
Hopkins University Press.
Owen-Smith, R.N., 1988. Megaherbivores: The Influence of Very Large Body
Size on Ecology. Cambridge University Press.
Pankovic, V., Glavatovic, R., Vunduk, N., Banjac, D., Marjanovic, N.,
Predojevic, M., 2006. A ‘‘quasi-rapid’’ extinction population dynamics
and mammoth overkill. Quant. Biol., 1e11. 0607033.
Pastor, J., Moen, R.A., 2004. Ecology of ice-age extinctions. Nature 431, 639e
640.
Paunovic, M., Jambresic, G., Brajkovic, D., Malez, V., Lenardic, J.M., 2001.
Last Glacial settlement of Croatia: catalog of fossil sites dated to the
OIS 2 and 3. Acta Geol. 26 (2), 27e70.
Pe
´an, S., 2001. Mammoth and subsistence practices during the Mid Upper
Palaeolithic of Central Europe (Moravia, Czech Republic). In:
Cavarretta, G., Gioia, P., Mussi, M., Palombo, M.R. (Eds.), The World
of Elephants Proceedings of the 1
st
International Congress. Consiglio
Nazionale delle Ricerche, Rome, pp. 701e703.
Pidoplichko, I.H., 1998. Upper Palaeolithic dwellings of mammoth bones in
the Ukraine: Kiev-Kirillovskii, Gontsy, Dobranichevka, Mezin and
Mezhirich. J. and E. Hedges, Oxford.
Pushkina, D., 2006. Dynamics of the mammalian fauna in southern Siberia
during the late Paleolithic. Vert. Pal. Asiat. 44 (3), 262e273.
Praslov, N.D., 1984. Palaeolithic cultures in Late Pleistocene. In:
Velichko, A.A. (Ed.), Late Quaternary Environments of the Soviet Union.
Minnesota University Press, Minneapolis, pp. 313e318.
Raia, P., Piras, P., Kotsakis, T., 2005. Turnover pulse or red queen? Evidence
from the large mammal communities during the Plio-Pleistocene of Italy.
Palaeogeogr. Palaeoclimatol. Palaeoecol 221, 293e312.
Raia, P., Meloro, C., Loy, A., Barbera, C., 2006. Species occupancy and its
course in the past: macroecological patterns in extinct communities.
Evol. Ecol. Res. 8, 181e194.
Reinhart-Waller, G., 2000. The Alexeev manuscript. Available online at: http://
www.drummingnet.com/alekseev/.
Ricciuti, E.R., 1973. To the Brink of Extinction. Holt, Rinehart & Winston,
New York and Chicago and San Francisco.
Roberts, R.G., Flannery, T.F., Ayliffe, A.K., Yoshida, H., Olley, J.M.,
Prideaux, G.V., Laslett, G.M., Baynes, A., Smith, M.A., Jones, R.,
Smith, B.L., 2001. New ages for the last Australian megafauna: conti-
nent-wide extinction about 46,000 years ago. Science 292, 1888e1892.
Rolland, N., 2004. Was the emergence of home bases and domestic fire a punc-
tuated event? A review of the Middle Pleistocene record in Eurasia. Asian
Perspectives 43 (2), 248e280.
Schreve, D.C., Thomas, G.N., 2001. Critical issues in European Quaternary
biostratigraphy. Quatern. Sci. Rev. 20, 1577e1582.
Shantser, E.V. (Ed.), 1982. Stratigraphy of the USSR, Quaternary system, half
vol. 1. Nedra, Moskva.
Shapiro, B., Drummond, A.J., Rambaut, A., Wilson, M.C., Matheus, P.E.,
Sher, A.V., Pybus, O.G., Gilbert, M.T.P., Barnes, I., Binladen, J.,
Willerslev, E., Hansen, A.J., Baryshnikov, G.F., Burns, J.A.,
Davydov, S., Driver, J.C., Froese, D.G., Harington, C.R., Keddie, G.,
Kosintsev, P., Kunz, M.L., Martin, L.D., Stephenson, R.O., Storer, J.,
Tedford, R., Zimov, S., Cooper, A., 2004. Rise and fall of the Beringian
steppe bison. Science 306, 1561e1565.
Sher, A.V., 1997. Late Quaternary extinction of large mammals in northern
Eurasia: a new look at the Siberian contribution (Past and Future Rapid
Environmental Changes: the Spatial and Evolutionary Responses of
Terrestrial Biota.). NATO ASI Series 147, 319e339.
Sher, A.V., Kuzmina, S.A., Kuznetsova, T.V., Sulerzhitsky, L.D., 2005. New
insights into the Weichselian environment and climate of the east Siberian
Arctic derived from fossil insects, plants and mammals. Quatern. Sci. Rev.
24, 533e569.
Smith, F.A., Lyons, S.K., Ernest, M., Jones, K.E., Kaufman, D.M., Dayan, T.,
Marquet, P.A., Brown, J.H., Haskell, J.P., 2003. Body mass of late Quater-
nary mammals. Ecology 84, 3403.
Smith, F.H., Trinkaus, E., Pettitt, P.B., Karavanic, I., Paunovic, M., 1999. Di-
rect radiocarbon dates for Vindija G1 and Velika Pecina late Pleistocene
hominid remains. Proc. Natl. Acad. Sci. 96 (22), 12281e12286.
Stewart, J.R., Lister, A.M., 2001. Cryptic northern refugia and the origins of
modern biota. Trends Ecol. Evol. 16, 608e613.
Solow, A.R., Roberts, D.L., Robbirt, K.M., 2006. On the Pleistocene extinc-
tions of Alaskan mammoths and horses. Proc. Natl. Acad. Sci. 103,
7351e7353.
Stuart, A.J., 1991. Mammalian extinctions in the late Pleistocene of northern
Eurasia and North America. Biol. Rev. 66, 453e562.
Stuart, A.J., 2005. The extinction of woolly mammoth (Mammuthus primige-
nius) and straight-tusked elephant (Palaeoloxodon antiquus) in Europe.
Quatern. Int. 126/128, 171e177.
Stuart, A.J., Kosintsev, P.A., Higham, T.F.G., Lister, A.M., 2004. Pleistocene
to Holocene extinction dynamics in giant deer and woolly mammoth.
Nature 431, 684e689.
Stuart, A.J., Sulerzhitsky, L.D., Orlova, L.A., Kuzmin, Y.V., Lister, A., 2002.
The latest woolly mammoths (Mammuthus primigenius Blumenbach) in
Europe and Asia: a review of the current evidence. Quatern. Sci. Rev.
21, 1559e1569.
Surovell, T., Waguespack, N., Brantingham, P.J., 2005. Global archaeological
evidence for proboscidean overkill. Proc. Natl. Acad. Sci. 102 (17), 6233e
6236.
Tankersley,K.B., Kuzmin, Y.V., 1998. Patternsof culture changein eastern Siberia
during the Pleistocene-Holocene transition. Quatern. Int 49/50, 129e139.
Tormidiaro, S.V., 1977. Change in physical-geographic environment on the
plains of northeast Asia at the Pleistocene-Holocene boundary as the
main reason for the mammoth theriofauna extinction. Trudy Zool. Inst.
AN SSSR 73, 64e71 (in Russian).
781D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782
Trinkaus, E., Moldovan, O., Milota, S., Bılgar, A., Sarcina, L., Athreya, S.,
Bailey, S.E., Rodrigo, R., Mircea, G., Higham, T., Ramsey, C.B., van
der Plicht, J., 2003. An early modern human from the Pestera cu Oase,
Romania. Proc. Natl. Acad. Sci. 100 (20), 11231e11236.
Trueman, C.N.G., Field, J.H., Dortch, J., Bethan, C., Wroe, S., 2005. Pro-
longed coexistence of humans and megafauna in Pleistocene Australia.
Proc. Natl. Acad. Sci. 102, 8381e8385.
van Andel,T.H., 2002.The climate and landscape of the middle part of the Weich-
selian glaciation in Europe: the Stage 3 Project. Quatern. Res. 57, 1e8.
van Kolfschoten, T., 2000. The Eemian mammal fauna of Central Europe. Geolo-
gie en Mijnbouw (Netherlands Journal of Geosciences) 79 (2/3), 269e281.
Vangengeim, E.A., Pevzner, M.A., Tesakov, A.S., 2001. Zonal division of the
Quaternary of eastern Europe based on small mammals. Stratigraphy,
Geological Correlation 9 (3), 76e88 (in Russian).
Vasil’ev, S., 2003. Faunal exploitation subsistence practices and Pleistocene
extinctions in Palaeolithic Siberia. Deinsea 9, 513e556.
Vasil’ev, S.A., Kuzmin, Y.V., Orlova, L.A., Dementiev, V.N., 2002. Radiocar-
bon-based chronology of the Palaeolithic in Siberia and its relevance to the
peopling of the New World. Radiocarbon 44, 503e530.
Vereshchagin, N.K., Baryshnikov, G.F., 1984. Quaternary mammalian extinc-
tions in Northern Eurasia. In: Martin, P.S., Klein, R.G. (Eds.), Quaternary
Extinctions: A Prehistoric Revolution. University of Arizona Press,
Tucson, pp. 483e516.
Vrba, E.S., DeGusta, D., 2004. Do species populations really start small? New
perspectives from the Late Neogene fossil record of African mammals.
Phil. Trans. R. Soc. Lond. B 359, 285e293.
Wroe, S., Field, J., Grayson, D.K., 2006. Megafaunal extinction: climate,
humans and assumptions. Trends Ecol. Evol. 21, 61e62.
782 D. Pushkina, P. Raia / Journal of Human Evolution 54 (2008) 769e782