How many elephant kills are 14?: Clovis mammoth and mastodon kills in context

Abstract

Recent evaluation of the use of Pleistocene megafauna by Clovis hunter-gatherers has suggested that a small number of reliable associations between Clovis artifacts and the remains of Proboscideans are documented, with perhaps as few as 14 occurrences currently known. Specifically, we ask whether 14 is a large or a small number of associations given the spatio-temporal dimensions of the Clovis period in North America. To place these 14 occurrences into context, we compare the time–space density and relative frequency of Clovis Proboscidean-bearing sites to those of Old World contexts. We develop models to identify the variables contributing to the archeological record of Proboscidean site creation, destruction, and sampling. While acknowledging potential biases in the record, comparative analysis suggests that the Clovis archeological record, with the possible exception of Lower Paleolithic of Iberia, indicates the highest frequency of subsistence exploitation of Proboscidea anywhere in the prehistoric world.

Full text

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright

Author's personal copy
Quaternary International 191 (2008) 82–97
How many elephant kills are 14?
Clovis mammoth and mastodon kills in context
Todd A. Surovell

, Nicole M. Waguespack
Anthropology Department 3431, University of Wyoming, 1000 East University Ave., Laramie, WY 82071, USA
Available online 15 December 2007
Abstract
Recent evaluation of the use of Pleistocene megafauna by Clovis hunter-gatherers has suggested that a small number of reliable
associations between Clovis artifacts and the remains of Proboscideans are documented, with perhaps as few as 14 occurrences currently
known. Specifically, we ask whether 14 is a large or a small number of associations given the spatio-temporal dimensions of the Clovis
period in North America. To place these 14 occurrences into context, we compare the time–space density and relative frequency of Clovis
Proboscidean-bearing sites to those of Old World contexts. We develop models to identify the variables contributing to the archeological
record of Proboscidean site creation, destruction, and sampling. While acknowledging potential biases in the record, comparative
analysis suggests that the Clovis archeological record, with the possible exception of Lower Paleolithic of Iberia, indicates the highest
frequency of subsistence exploitation of Proboscidea anywhere in the prehistoric world.
r2007 Elsevier Ltd and INQUA. All rights reserved.
1. Introduction
Clovis projectile points and other artifacts have been
recovered in association with the remains of mammoths
and mastodons from throughout the continental United
States (e.g., Sellards, 1952;Haury, 1953;Haury et al., 1959;
Leonhardy, 1966;Warnica, 1966;Hester, 1972;Rayl, 1974;
Graham et al., 1981;Frison and Todd, 1986;Laub et al.,
1988;Hannus, 1989;Overstreet, 1996;Overstreet and
Kolb, 2003). From this simple observation, it has been
induced that Clovis peoples hunted Proboscideans. This
much we can agree is true. In this paper, we accept this
basic premise, and address one particular facet of the
Paleoindian–Proboscidean relationship; how frequently
were mammoths and mastodons taken? One can envision
the answer to this question as a continuum, varying from
Proboscidean meat being a dietary staple to being one that
a person might enjoy once or twice in a lifetime. In order to
establish the limits of such a continuum, we quantitatively
compare the Clovis record of elephant exploitation with
similar records from other parts of the world.
Using the most lenient and problematic standard of
Proboscidean use, simple presence in zooarcheological
assemblages, we previously estimated that at least 91
individual mammoths and mastodons are known from a
total of 26 Clovis sites (Waguespack and Surovell, 2003,
Table 2). Based on available data, no other taxon is present
in as many sites or is represented by as many individuals.
These findings suggested that Proboscideans were utilized
more frequently than other types of prey. Using more
stringent and taphonomically rigorous standards, Grayson
and Meltzer (2002, 2003) found that there are only 14 sites
and 15 Clovis components showing secure associations
with Proboscideans (see also Cannon and Meltzer, 2004).
Two of these contain mastodon and 13, mammoth.
Although Grayson and Meltzer’s (2002, 2003) studies were
concerned primarily with evaluation of Martin’s overkill
model (Martin, 1973, 1984;Mosimann and Martin, 1975;
Martin and Steadman, 1999), they suggest that the rarity of
reliable associations between Clovis artifacts and Probos-
cidean remains indicates that there is ‘‘little support for the
assertion that big-game hunting was a significant element
in Clovis-age subsistence patterns’’ (Grayson and Meltzer,
2002, p. 348). From the Paleoindian faunal record then,
two interpretations widely divergent on the Proboscidean
use continuum, have been proposed.
ARTICLE IN PRESS
1040-6182/$ - see front matter r2007 Elsevier Ltd and INQUA. All rights reserved.
doi:10.1016/j.quaint.2007.12.001

Corresponding author. Tel.: +1 307 766 3239; fax: +1 307 766 2473.
E-mail addresses: surovell@uwyo.edu (T.A. Surovell),
nmwagues@uwyo.edu (N.M. Waguespack).

Author's personal copy
By contrast, Haynes (1999, p. 23; 2002, p. 183) has
argued that the number of Clovis elephant kill sites is
extremely high in comparison to the record of elephant
hunting from Africa, a considerably larger region and one
where humans have coexisted with Proboscideans for a
significantly longer period of time:
The United States contains more megamammal
killsites than there are elephant killsites in all of
Africa—a land mass that is much larger than the United
States. Not only is Africa much larger, but its hominin
presence extends back at least 100 times the chonos-
tratigraphic span of the human presence in North
America. Yet there are fewer than a dozen probable
killsites, spanning a time range from Plio-Pleistocene
to mid-Holocene y[emphasis in original] (Haynes,
2002)
To Haynes (2002, p. 183), the Clovis data ‘‘show how
breathtakingly abundant the associations are.’’ These
clear differences in opinion are about a relative simple
question: is the archeological record of Proboscidean
hunting in North America characterized by abundance
or scarcity? Examined from a North American-centric
perspective it is difficult to evaluate just how many
Proboscidean kills equate to ‘‘a lot’’ versus ‘‘a little’’
subsistence use.
In this paper, we use the Grayson and Meltzer (2002,
2003) estimate of 14 reliable subsistence associations with
Proboscideans as a benchmark for comparison. In
particular, we ask whether 14 is a large number when
viewed in a comparative context. Humans have also hunted
and likely scavenged Proboscideans in contexts beyond
Clovis in Africa, Europe, Asia, and Central and South
America. We focus specifically on the record of subsistence
use of Proboscideans from the Old World in comparison to
Clovis by quantifying the spatio-temporal density and
relative frequency of elephant-bearing sites. Following
Haynes (1999, 2002), we argue that by using a comparative
framework we are able to better evaluate counts of
Paleoindian–Proboscidean associations. At this time, we
cannot derive an absolute estimate of the frequency of
elephant hunting in Clovis times, but we can address
whether elephant hunting by Clovis foragers was relatively
common or rare as compared to other slices of hunter-
gatherer space and time manifested archeologically.
Making such quantitative comparisons is not simple
or straightforward because numerous sampling issues
interfere. The organization of the paper is as follows. We
first build a comparative Old World dataset, and to
highlight the sampling issues involved and to guide our
analyses, we construct formal models of the quantitative
abundance of Proboscidean-bearing sites. Finally, we
compare quantitative measures of spatio-temporal densi-
ties and relative frequencies of Proboscidean kill/scavenge
(PKS) sites for the Clovis period and the Old World
Paleolithic.
2. The archeological of record subsistence exploitation of
elephants in the Old World
By our estimates, hundreds of archeological sites from
Africa, Europe, and Asia are known to contain the remains
of elephants. For example, Gamble (1986, Tables 7.3 and
7.4) records approximately 100 occurrences of Probosci-
deans from Middle and Upper Paleolithic cave assemblages
from various regions of Europe alone (see also Haynes,
1991, pp. 195–263). Furthermore, archeological associa-
tions with elephants are found from the early phases of the
Lower Paleolithic (Leakey, 1971;Berthelet and Chavaillon,
2001;Delagnes et al., 2006) to the Holocene (MacCalman,
1967;Haynes, 1991, p. 195), and the archeological history
of such finds stretches back to at least the end of the 17th
century (Grayson, 1983, pp. 7–8). As in the Clovis case,
there has been considerable debate concerning the roles of
Proboscideans in various places and times in Old World
prehistory, best exemplified by questions of hunting,
scavenging, and taphonomy at the Spanish Lower Paleo-
lithic sites of Torralba and Ambrona (Howell, 1966;
Freeman, 1973;Howell and Freeman, 1983;Shipman and
Rose, 1983;Binford, 1987;Klein, 1987;Villa, 1990;
Haynes, 1991;Villa et al., 2005), and the role of mammoth
in the Eastern Gravettian and Epigravettian economies of
Central and Eastern Europe (Klein, 1973;Kozlowski,
1986;Soffer, 1985, 1993;Haynes, 1991;Svoboda et al.,
1996, pp. 154–157). These debates underscore the difficulty
of reliably establishing the identification of subsistence use
of Proboscideans from the archeological record (Haynes
and Stanford, 1984;Haynes, 1991;Grayson and Meltzer,
2002;Cannon and Meltzer, 2004).
We suggest that the best evidence for subsistence use of
Proboscideans comes from kill/scavenge sites (Haynes,
2002, p. 183), what Gamble (1999, p. 344) has referred to as
‘‘gatherings around single carcasses,’’ or what others have
called ‘‘single carcass sites’’ (Delagnes et al., 2006, p. 448).
These sites often contain the remains of one to a few
animals, often partially articulated, associated with small
artifact assemblages typically less than 10 to a few hundred
pieces. Limiting the sample this way, we are able to take a
vast number of Old World sites containing elephants and
pare them down to a manageable sample that we argue
provide the best evidence for the subsistence use of
elephants. By focusing on ‘‘gatherings around carcasses’’,
a phrase that is intentionally ambiguous with respect to the
question of hunting or scavenging, the frequency of such
occurrences throughout time and space enable large-scale
comparisons for identifying differences in Proboscidean
use by prehistoric foragers. We also include in our sample
two special cases, La Cotte de St. Brelade (England) and
Lugovskoye (Russia) that are discussed in greater detail in
the following text.
It is important to note, however, that by applying this
standard to the archeological record of the Old World, we
are not applying one that the Clovis record of mammoth
and mastodon exploitation cannot meet. All but one of the
ARTICLE IN PRESS
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–97 83

Author's personal copy
14 sites that Grayson and Meltzer (2002, Table VII)
identify as reliable evidence of human predation of North
American Proboscideans are sites with small numbers of
animals well-associated with artifact assemblages, the
exception being Pleasant Lake (Fisher, 1987), which lacks
artifacts. Semantically, it might be difficult to fit Naco
(Haury, 1953) and possibly Escapule (Hemmings and
Haynes, 1969) into the category of ‘‘gatherings around
carcasses’’ since they are sometimes interpreted to repre-
sent animals that were targets of human predation but
escaped (Haynes, 1966;Grayson and Meltzer, 2002,
p. 348). However, the association of these animals with
Clovis projectile points is undeniable. Also, Lehner (Haury
et al., 1959), Colby (Frison and Todd, 1986), and Dent
(Figgins, 1933;Brunswig and Fisher, 1993) could be
interpreted as not meeting our criteria since they contain
at least 7, 13, and 12 animals, respectively. However, at
Lehner and Colby, we find the association with Clovis
weaponry convincing evidence that at least some of the
animals at these sites were hunted, but we agree with
Grayson and Meltzer (2002, p. 337) that of these, Dent is
certainly the most problematic. Therefore, put another
way, in attempting to make a quantitative comparison
between Clovis and Old World elephant exploitation, we
are only asking of the Old World record that it be
qualitatively comparable to that of Clovis.
In sum, we are able to identify a total of 25 sites,
localities, or components showing evidence of hunting or
scavenging of elephants from Africa, Europe, and Asia
(Table 2,Fig. 1). We include sites that could be considered
to have somewhat weak associations in our sample (e.g.,
Torralba, Skaratki, and Zoo Park), and by doing so, we
may be artificially inflating our sample of Old World sites
relative to Clovis. There are a number of similar sites from
North America that could be included in a comparable
sample, such as Boaz (Palmer and Stoltman, 1975),
Martins Creek (Brush and Smith, 1994;Brush et al.,
1994), McLean (Ray and Bryan, 1938;Ray, 1942), and
Guest (Rayl, 1974;Hoffman, 1983), but these are excluded.
Although we have been criticized for ignoring taphonomic
considerations in the past (Cannon and Meltzer, 2004,
p. 1956), by including sites with questionable associations
only from Old World contexts, we are biasing our analysis
against a position we have previously taken, that early
Paleoindians in North America preferentially targeted
Pleistocene megafauna (Waguespack and Surovell, 2003;
Surovell and Waguespack, in press).
Proboscidean kill/scavenge sites are most common in the
Lower Paleolithic, representing 56% of the sample. These
include five sites dating to the Early Pleistocene, including
the remains of two Proboscideans tightly associated with
artifact assemblages from two levels of the FLK North
Locality at Olduvai Gorge (Leakey, 1971, pp. 64–66,
85–86) a single Elephas recki at Barogali in the Republic of
Djibouti (Berthelet and Chavaillon, 2001), an E. recki
associated with more than 2000 Acheulean artifacts at
Site 15, Olorgesailie Member I in Kenya (Potts et al., 1999,
pp. 768–769), and a similar association at Mwanganda’s
Village in Malawi (Clark and Haynes, 1970;Kaufulu,
1990). Eight Lower Paleolithic sites dating to the Middle
Pleistocene are known. The youngest known kill/scavenge
site of an African E. recki, dating to ca. 700 ka, is Nadung’a
4 in West Turkana, Kenya, where a single animal was
recovered with almost 7000 artifacts (Delagnes et al., 2006).
The remaining Lower Paleolithic associations occur out-
side of Africa in southwest Asia or western Europe and are
all associated with Paleoloxodon antiquus. These include
Gesher Benot Ya’aqov, Israel (Goren-Inbar et al., 1994),
Notarchirico, Italy (Piperno and Tagliacozzo, 2001),
Aridos I and II (Villa, 1990;Santonja et al., 2001) and
Torralba and Ambrona (Howell, 1966;Freeman, 1973,
1975;Villa, 1990;Villa et al., 2005), Spain, and the
Southfleet Road and Ealing sites in Great Britain (Brown,
1889;Wenban-Smith et al., 2006).
Four Middle Paleolithic occurrences are known, repre-
senting 16% of the sample. All occur in northern Europe
and likely date to interglacial periods. Two sites, Lehringen
and Gro
¨bern, are located in northern Germany. At
Lehringen, a single P. antiquus was recovered in association
with a 2.5 m long wooden spear (apparently found between
ARTICLE IN PRESS
Age
(yr)
106105104
Fig. 1. World map showing the spatio-temporal distribution of Proboscidean kill/scavenge sites for Africa, Europe, Asia, and North America. Site
location is indicated by the base of the pin and site age is indicated by pin length. Ages (and pin lengths) are logarithmically scaled.
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–9784

Author's personal copy
the ribs of the elephant) and 24 chipped stone artifacts
(Movius, 1950;Adam, 1951;Thieme and Veil, 1985;
Weber, 2000). This one of the few sites in the Old World
for a which a strong argument for a kill can be made due to
the association with weaponry (but see Gamble, 1987).
Gro
¨bern also produced a single P. antiquus in association
with 28 lithic artifacts and one possible bone artifact,
interpreted to be to a projectile point (Mania et al., 1990;
Weber, 2000). At La Cotte de St. Brelade (Channel Islands,
England) two levels contain what appear to be cultural
piles of wooly mammoth and rhinoceros remains (Scott,
1980, 1986;Callow, 1986). This site does not qualify as a
‘‘gathering around a carcass’’, but we include it in our
sample because the site is generally believed to present
evidence of two episodes of the killing of mammoths by
driving them over a cliff, or by funneling them into a trap
(e.g., Scott, 1980, 1986;Lister and Bahn, 1994, p. 129;
Mellars, 1996, p. 229).
Seven Upper Paleolithic and Later Stone Age sites occur,
accounting for 28% of the sample. Three sites associated
with woolly mammoth (Mammuthus primigenius) cluster in
the early Upper Paleolithic of eastern Europe. This include
Nowa Huta (Kozlowski et al., 1970) and Skaratki
(Chmielewski and Kubiak, 1962) in Poland, and Halich,
Ukraine (Wojtal and Cyrek, 2001). Two M. primigenius
sites occur in the late Upper Paleolithic of Sibera,
Lugovskoye, Russia (Maschenko et al., 2003;Zenin et
al., 2003) and Shikaevka II (Vasil’ev, 2003). Lugovskoye
(ca. 13.5 ka) is the only site of this group that is not a
‘‘gathering around a carcass’’, but instead, mammoth
hunting is inferred from the presence of a steeply retouched
blade fragment embedded within a thoracic vertebra of an
adult mammoth (Maschenko et al., 2003;Zenin et al.,
2003). The Upper Paleolithic site of Lake Nojjiri, Japan
has yielded three distinct aggregations of Palaeoloxodon
naumanni remains and artifacts (Kondo et al., 2001). The
youngest site in the Old World sample, Zoo Park,
Namibia, believed to date ca. 10 ka contains two elephants
(Loxodonta sp.) roughly 10 m apart, both associated with
artifacts (MacCalman, 1967).
3. Modeling spatio-temporal density
Having compiled a dataset, it is tempting to simply begin
comparing relative site frequencies with the aim of
constructing a continuum of Proboscidean use. Unfortu-
nately, there are numerous sampling issues that make it
extremely difficult to directly and meaningfully compare
the number of elephant kills between any two regions.
Greater numbers of sites evidencing the subsistence use of
Proboscideans are expected in regions characterized by
greater land areas and/or greater archeological temporal
depth. These issues come into play when making compar-
ison between a portion of North America and the
significantly larger combined area of Africa, Europe, and
Asia. Similarly, prehistoric humans and hominids lived
side-by-side with Proboscideans in the Old World for more
than 1 million years, while in the New World, people only
coexisted with mammoths and mastodons for a few
hundred to a few thousand years. Therefore, we might
expect a priori that there would be considerably more
elephant kills in the Old World than North America. To
correct these problems, it is possible to use spatio-temporal
densities of elephant kill sites as a proxy for per capita rates
of elephant hunting, but a number of additional complica-
ting factors must be considered, such as spatio-temporal
variation in past human population densities, taphonomy,
and archeological research intensity. In the proceeding
section we examine this multitude of variables involved in
attempting to answer the question; how many elephant
kills are 14?
In the simplest sense, the number of PKS, those sites
showing evidence of the subsistence utilization of Probos-
cideans, known from a region (n
a
) can be modeled as a
function of two variables: the number of sites in existence
today (n
e
), and the intensity of archeological research (i):
na¼nei(1)
where irepresents the proportion of sites in existence today
that have been discovered and investigated. The number of
PKS in existence today (n
e
) can be expressed as
ne¼ncx(2)
where n
c
is the total number of PKS created by human
hunters in a region throughout time, and xis the
proportion of PKS surviving in the archeological record
to the present. Combining these two equations yields a
simple model of the total number of kill sites known from a
region (n
r
):
nr¼ncxi (3)
Therefore, the current archeological sample of PKS
can be expressed simply as a function of the total number
created, the proportion surviving destructive taphonomic
processes, and the archeological research intensity in a
given region. The spatio-temporal density of PKS (d)is
modeled as
d¼
nr
tha¼
ncxi
tha(4)
where t
h
is the time depth of elephant hunting and ais the
land area of that region.
Eq. (4) is an extremely simplistic portrayal of the system,
and most of the terms in the equation can be further
decomposed. For example, we are not interested as much
in the number of subsistence sites created, but instead how
they relate to the relative frequency of elephant use for
prehistoric humans, the per capita rate of Proboscidean
exploitation. Therefore, we can decompose the term n
c
, the
total number of PKS created, into more appropriate
variables. The variable of interest is the number of
elephants used per person per year (r
h
). The variable n
c
can be expressed as a function of r
h
:
nc¼pharhth(5)
ARTICLE IN PRESS
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–97 85

Author's personal copy
where p
h
is the average human population density over
time t
h
in a region of area a. Substituting this Eq. (5) into
Eq. (4) gives
d¼pharhthxi
tha
d¼phrhxi (6)
Eq. (6) shows that when counts are divided by space
and time, both variables cancel out of the equation,
showing that spatio-temporal densities, unlike simple
counts, are insensitive to the spatio-temporal dimensions
of the region sampled. The critical variables become the
average human population density (p
h
), the per capita rate
of elephant use (r), site preservation (x), and research
intensity (i).
There are many ways that we could decompose the term
i, archeological research intensity. The term is meant to
express the concept of sample size. Hypothetically, if we
had a 10% sample of all of the Clovis sites created, but we
only have a 1% sample of all of the Paleolithic sites ever
created, then even if the rate of elephant use was the same
for Clovis and Paleolithic times, the spatio-temporal
density of Clovis PKS would appear 10 times greater.
The variable research intensity (i) can be modeled as the
total number of archeological sites excavated (s
e
) relative to
the total number of sites created (s
c
):
i¼
se
sc
(7)
Following Eq. (5), the total number of archeological sites
created (n
c
) is modeled as
nc¼ptarttt(8)
where p
t
is the average human population density over the
entirety of archeological time, ais the land area, r
t
is the
mean per capita rate of archeological site creation, and t
t
is
the entire archeological time depth. The variables p
t
,r
t
, and
t
t
differ from p
h
,r
h
, and t
h
in that the former refer to all of
archeological time, while the latter refer to only that time
when humans coexisted with Proboscideans in a particular
region. Therefore, archeological research intensity can be
modeled as
i¼
se
ptarttt
(9)
To highlight what we see as one of the most serious
sampling problems affecting this analysis, we can also
decompose the term s
e
, the number of archeological sites
excavated to date:
se¼paarata(10)
where p
a
is the mean population density of archeologists
over the history of archeological research, r
a
is the mean
per capita rate of site excavation (the number of sites
excavated per archeologist per year), and t
a
is the time
depth of archeological research (the amount of time
archeologists have been at work). Combining Eqs. (9)
and (10) yields
i¼paarata
ptarttt
i¼
parata
ptrttt
(11)
Except for the variable land area (a), which cancels out of
the equation, six variables remain. Research intensity is a
function of the population density of archeologists, the per
capita rate of excavation, the time depth of archeological
research, the average prehistoric population density, the
rate of archeological site creation, and the time depth of
prehistoric occupation. The most troubling aspect of this
equation is that if all variables are assumed to be constant,
except the time depth of human occupation, there is a
problem of ‘‘temporal research dilution’’ in comparing
Clovis to the Paleolithic.
Humans have occupied the New World for approxi-
mately 10
4
yr, while elephant-using hominids have occu-
pied Africa, Europe, and Asia for approximately 2 10
6
yr,
a difference of two orders of magnitude. If we hold all
variables constant, except prehistoric time depth, the
research intensity in New World archeology would be
expected to be roughly 100 times greater than that of the
Old World. To put this problem another way, we can think
of 10 archeologists working in two regions: A and B.
In Region A, there are 1000 years of prehistory, and in
Region B, there are 1,000,000 years of prehistory. In
Region A, there is one archeologist for every 100 years of
prehistory. In Region B, there is one archeologist for every
100,000 years of prehistory. Therefore, even if the
archeologists in both regions work equally hard, the
research intensity for Region B will be 1000 times less
than for Region A, simply because research in Region B is
temporally diluted. Therefore, if we were to simply
compare the spatio-temporal densities of elephant kills in
Clovis and Old World Paleolithic contexts, then we would
necessarily bias our results in favor of Clovis.
Combining Eqs. (6) and (11) provides a model that is
considerably more complex than the one we started with:
d¼phrhxparata
ptrttt
(12)
Furthermore, we have not even approached to question of
site preservation. Some of the factors that likely would go
into modeling the variable x(the proportion of sites not
surviving to the present) might be hunting practices, animal
behavior, age, ecology, geomorphology, and climate, but
rather than reducing this equation further, we leave it in its
current state because it underscores the difficulties involved
in using spatio-temporal densities of archeological sites to
monitor rates of Proboscidean use.
In the simplest sense, from Eq. (12), we can state that all
things being equal, the spatio-temporal density of PKS (d)
will be proportional to the per capita rate of Proboscidean
use (r
h
) The key to that statement, of course, is the phrase
ARTICLE IN PRESS
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–9786

Author's personal copy
‘‘all things being equal’’. Those ‘‘things’’ include: (1) the
average prehistoric human population density during
the time in which humans and elephants coexisted, (2)
the proportion of kill sites surviving to the present, (3) the
average population density of archeologists over the
history of archeological research, (4) the per capita rate
of site excavation, (5) the time depth of archeological
research, (6) the average prehistoric human population
density over all archeological time, (7) the average per
capita rate of archeological site creation, and (8) the time
depth of human occupation. In order to directly compare
spatio-temporal densities between any two regions, then,
would require one to control all of these factors, a
seemingly daunting task.
Our point is not to suggest that such an analysis is
impossibly difficult or hopeless. We only wish to point out
that such analyses are far more complex than they are
commonly portrayed to be. The most serious problem in
this case, we argue, is temporal research dilution. There is
simply too much time in the Old World and too few
archeologists to treat research intensities as equivalent for
Clovis and the Paleolithic. While vastly more Paleolithic
sites have been excavated in comparison to Clovis, it is
likely that the known Clovis archeological record is
actually a larger sample in comparison to the total number
of sites in existence. Another problem that we think
presents a similar obstacle to comparing spatio-temporal
densities is the population density of archeologists.
Considering the vastness of Africa and East Asia (including
Siberia) and the relatively small numbers of archeologists
working in these regions, again in comparison to Clovis,
these areas have yet to be sampled archeologically to the
same extent as North America. Site preservation and
destruction is another term that is highly problematic. We
have treated it as a constant, and yet it undoubtedly could
be modeled as a function of time, in that the probability of
a site surviving to the present should be inversely
proportional to its age (Surovell and Brantingham, 2007),
and in this sense, we suspect that huge numbers of
Paleolithic sites have been lost to destructive geomorphic
and taphonomic factors in comparison to Clovis. Likewise,
the probability of finding evidence of the use of elephants
from the Paleolithic of central Africa or southeast Asia (or
eastern North America) is very low due to geochemical
conditions not conducive bone preservation. One factor
that might act to bias the analysis in the opposite direction
is prehistoric human population density. We suspect that
average population densities for Clovis were significantly
lower than for much of the Upper Paleolithic, and in this
sense large numbers of elephant-bearing sites in Clovis
contexts suggests greater frequencies of exploitation in
comparison to the Upper Paleolithic.
If these problems are not difficult enough, there are
additional assumptions built into the model. For example,
it is assumed that archeological research is unbiased and
produces a representative sample of time–space patterning
in the archeological record, an assumption unlikely to be
valid. Our efforts to mathematically model the compound-
ing factors influencing the archeological record of Probos-
cidean subsistence use may appear to produce an
overwhelming array of ‘‘what if ’s’’, it is important to
consider that all variables must be addressed before
declarations regarding the relative abundance or paucity
of elephant hunting in particular times and places are
made. After all, any archeological interpretation regarding
the frequency of particular behaviors through time or
across space must account for the variables outlined in our
model.
3.1. Quantifying spatio-temporal density
In order to estimate spatio-temporal densities, we must
also control for space and time. Beginning with the
question of space, we can simply sum the areas of the
continents of Africa, Asia, and Europe, but to do so would
be to ignore two important factors. First, the repeated
growth and recession of glaciers over during the Pleisto-
cene resulted in regular fluctuations in the area of habitable
land.
1
Also, humans did not occupy all of Europe, Africa,
and Asia during all of this time. While the question of the
timing of initial migration out of Africa continues to be
debated (Roebroeks and Kolfschoten, 1994;Swisher et al.,
1994;Gibert et al., 1998;Klein, 1999;Roebroeks, 2001;
Klein and Edgar, 2002;Dennell, 2003), it is clear that
hominids entered Eurasia prior to the Bruhnes/Matuyama
boundary at 780 ka (Carbonell et al., 1995;Gabunia et al.,
2001;Belmaker et al., 2002). Also, humans apparently did
not colonize arctic environments until 30 ka (Pitulko et al.,
2004). Various species of Proboscideans did, however,
occupy virtually all of this area during the Pleistocene
(Todd and Roth, 1996). For Clovis, the same problems
could apply, especially if Clovis is assumed to be a
colonizing population. With respect to time, there is the
question of the timing of the extinction of elephants in the
Old and New Worlds, and that they did not go extinct in
portions of Africa and Asia. The timing of Proboscidean
extinctions appears to have varied considerably across
Eurasia (Tchernov, 1984;Soffer, 1993, pp. 41–44; Stuart
and Lister, 2001;Stuart et al., 2002, 2004;Baryshnikov,
2003;Kuzmin et al., 2003;Ugan and Byers, in press), while
in North America, extinction was temporally more
constrained with many of the latest occurrences of
mammoth and mastodon appearing in association with
Clovis artifacts or are contemporaneous with Clovis
(Meltzer and Mead, 1985;Haynes, 1993;Grayson and
Meltzer, 2002, 2003;Haynes, 2002).
We opted to simplify our calculations and proceed
bearing in mind these potential complicating factors
because we are trying to identify coarse differences at large
scales. For our Old World sample, the total area of Africa,
ARTICLE IN PRESS
1
Changes in sea level have also increased habitable area at time in the
past, but because these areas have not been sampled, they have no impact
on spatio-temporal densities.
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–97 87

Author's personal copy
Europe, and Asia is 84,852,000 km
2
, and temporally, the
sites stretch from approximately 1.75 ma to 13.5 ka, a span
of 1.74 million years. For the New World sample, the
area of the 48 continental United States is roughly
8,081,000 km
2
. The age range for Clovis was calculated
using age difference for the calibrated ages of the oldest
and youngest sites. The Hebior mammoth produced three
dates ranging between 12,480760 (CAMS 28303) and
12,590750
14
CyrBP(Overstreet and Kolb, 2003), and the
Clovis occupation at Murray Springs is dated to approxi-
mately 10,900750
14
C yr BP based on the average of eight
charcoal dates (Haynes, 1992). After calibration,
2
the sites
span a total of approximately 2000 years. The total sample
of elephant associations from Clovis contexts includes 14
sites (Table 1), but because Hebior lacks artifacts, we
exclude it from our sample. The North American sample
remains 14 because Blackwater Draw contains two
separate localities (Table 1;Fig. 1). Our Old World
sample includes a total of 25 localities, components, or
sites (Table 2;Fig. 1).
Using these data, it is possible to calculate spatial,
temporal, and spatio-temporal densities for the Clovis and
Old World samples (Fig. 2). Beginning with spatio-
temporal densities for Clovis, there are approximately
0.87 sites per billion km
2
yr, and 0.0002 sites per billion k-
m
2
yr for the Old World sample (Fig. 2a). In other words,
elephant subsistence sites are roughly 5000 times more
abundant in Clovis when standardized to space and time.
Looking only at time, for the Clovis sample there are 7.0
sites per thousand years, and for the Old World sample
there are 0.014 sites per thousand years (Fig. 2b). There-
fore, Clovis elephant subsistence sites are approximately
600 times more abundant with respect to time. Of course,
both of these analyses are problematic because of the
temporal research dilution problem discussed previously.
One solution to this dilemma is to ignore time, using only
spatial densities. Although this analysis should in theory
bias the sample toward the Old World, where considerably
more time is represented, Clovis still comes out on top.
There are approximately 1.73 Clovis sites per million km
2
as compared to only 0.29 per million km
2
in the Old World
(Fig. 2c).
From Fig. 1, it is clear that sites with evidence for
subsistence use of elephants are considerably more
common in Europe than in Africa or Asia, despite that
Europe represents the smallest land area. This pattern is
likely due in part to differences in research intensity.
Consequently, it seems fruitful to repeat the analysis
comparing only the European record to that of Clovis.
The continent of Europe is roughly 10,000,000 km
2
in area,
and our European sample includes 14 sites, spanning a
period of roughly 650,000 years. In comparison to Clovis,
the European record of elephant use is quite impoverished
when time is figured into the equation (Fig. 2b and c). For
example, there are only 0.02 PKS per thousand years in the
European record as compared to 7 for Clovis, a difference
of two orders of magnitude. The spatial densities of the two
regions however, are quite comparable with 1.4 sites per
million km
2
for Europe versus 1.73 sites per million km
2
for
Clovis. Furthermore, during multiple glaciations, a larger
proportion of Europe was covered by ice than for North
America in Clovis times, suggesting that the true spatial
densities for these two regions are likely quite similar. It
remains difficult to make conclusive interpretations of
these findings due to complicating factors such as a much
greater time span of human occupation in Europe.
Although admittedly course-grained, according to this
preliminary analysis the archeological record of Clovis
appears to be especially rich in sites showing the
subsistence use of elephants with 14 occurrences known
from a relatively small slice of space and time. The record
of elephant use in Europe, which spans a time period more
than 300,000 times that of Clovis, approaches Clovis in
spatial density. There are numerous assumptions that must
ARTICLE IN PRESS
Table 1
Clovis sites showing secure subsistence associations with Proboscideans
Site and locality State Latitude Longitude Age
a
(ka) Taxon
Hebior Wisconsin 43N 88W 12.5 Mammuthus primigenius
Colby Wyoming 44.0N 107.9W 11.2 Mammuthus columbi
Domebo Oklahoma 35.0N 98.4W 11.1 Mammuthus imperator
Lubbock Lake Texas 33.6N 101.9W ca. 11 Mammuthus columbi
Blackwater Draw Mammoth Pit New Mexico 34.3N 103.3W ca. 11 Mammuthus columbi
Escapule Arizona 31.6N 110.2W ca. 11 Mammuthus columbi
Blackwater Draw El Llano Dig No. 1 New Mexico 34.3N 103.3W ca. 11 Mammuthus columbi
Kimmswick Missouri 38.4N 90.4W ca. 11 Mammut americanum
Miami Texas 35.6N 100.6W ca. 11 Mammuthus columbi
Naco Arizona 31.4N 109.9W ca. 11 Mammuthus columbi
Lehner Arizona 31.3N 110.1W 10.9 Mammuthus columbi
Murray Springs Arizona 31.6N 110.2W 10.9 Mammuthus columbi
Lange-Ferguson South Dakota 43N 103W 10.8 Mammuthus sp.
Dent Colorado 40.3N 104.8W 10.8 Mammuthus columbi
a
Uncalibrated radiocarbon ages.
2
Radiocarbon dates were calibrated using OxCal v. 3.9 (Ramsey, 2003).
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–9788

Author's personal copy
be made if one wishes to take these results at face value,
and many of those assumptions are tenuous. We simplify
the problem considerably by supplementing this initial
analysis with one based on the relative proportions of
elephant-bearing sites for various divisions of space and
time. As we show in the following text, such an analysis
allows us to eliminate many of the problematic, yet
necessary, assumptions when comparing spatio-temporal
densities.
4. Modeling relative site frequencies
Instead of standardizing counts of elephant-bearing sites
to space and time, if we standardize them relative to the
total sample of sites known from a given spatio-temporal
region, we can make fewer and perhaps more justifiable
assumptions. To model the relative frequency of
PKS, we need only two terms, one describing the
number of PKS excavated to date and a second term
describing the total number of sites excavated to date.
From above, the numerator of Eq. (6) describes the
total number of elephants subsistence sites excavated to
date (n
e
) for a spatio-temporal region of area (a) and time
span (t
h
):
ne¼pharhthxi (13)
where p
h
is the average human population density, r
h
is the
per capita rate of elephant site creation, dis the proportion
of sites not surviving to the present, and iis the research
intensity. The total number of sites (n
s
) excavated within
ARTICLE IN PRESS
Table 2
Old World sites showing evidence of subsistence use of Proboscideans
Site and locality Country Latitude Longitude Phase Age Taxon
FLK North Upper Bed I Tanzania 3S 35E LP 1.7–1.8 ma Elephas recki
FLK North Lower Bed II Tanzania 3S 35E LP 1.2–1.7ma Deinotherium sp.
Barogali Djibouti 11N 42E LP 1.3–1.6 ma Elephas recki
Olorgesaille Member I Site 15 Kenya 2S 36E LP 974–992 ka Elephas recki
Mwanganda’s Village Malawi 10S 34E LP 0.6–1.2 ma (?) Elephas sp. (?)
Gesher Benot Ya’aqov Israel 33.0N 35.6E LP ca. 750 ka Palaeoloxodon antiquus
Nadung’a 4 Kenya 4N 36E LP ca. 700 ka Elephas recki
Notarchirico Italy 41N 16E LP 600–740 ka Palaeoloxodon antiquus
Aridos I Spain 40.3N 3.5W LP 300–500 ka (?) Palaeoloxodon antiquus
Aridos II Spain 40.3N 3.5W LP 300–500 ka (?) Palaeoloxodon antiquus
Southfleet Road UK 51.4N 0.3E LP ca. 400 ka Palaeoloxodon antiquus
Ealing UK 51.5N 0.3W LP? 300–500 ka (?) Palaeoloxodon antiquus (?)
Ambrona Spain 41N 2.5W LP 200–400 ka (?) Palaeoloxodon antiquus
Torralba Spain 41N 2.5W LP 200–400 ka (?) Palaeoloxodon antiquus
La Cotte de St. Brelade Level 3 UK 49.2N 2.3W MP 120–200ka Mammuthus primigenius
La Cotte de St. Brelade Level 6 UK 49.2N 2.3W MP 120–200ka Mammuthus primigenius
Lehringen Germany 52.8N 9.5E MP ca. 120 ka Palaeoloxodon antiquus
Gro
¨bern Germany 51.8N 12.5E MP ca. 120 ka Palaeoloxodon antiquus
Lake Nojiri Japan 36.8N 138.2E UP? 33–39 ka
a
Elephas naumanni
Nowa Huta Poland 50N 20E UP 35 ka (?)
a
Mammuthus primigenius
Skaratki Poland 52.0N 19.8E UP? 29–32 ka (?)
a
Mammuthus primigenius
Halich Ukraine 49.1N 24.7E UP 20–24 ka
a
Mammuthus primigenius
Shikaevka II Russia 56.3N 66.4E UP 18 ka
a
Mammuthus primigenius
Lugovskoye Russia 61.0N 68.5E UP ca. 13.5 ka
a
Mammuthus primigenius
Zoo Park Namibia 22.5S 17.1E LSA 10 ka (?)
a
Loxodonta sp. (?)
a
Uncalibrated radiocarbon age.
Old World
Region
Spatio-Temporal Density
(Sites •10-9km-2yr-1)
10-3
10-4
ClovisEurope
100
10-1
101
10-2
10-1
10-2
100
Temporal Density
(Sites•10-3yr-1)
Old World
Region
ClovisEurope
100
10-1
101
Spatial Density
(Sites •10-6km-2)
Old World
Region
ClovisEurope
Fig. 2. Spatio-temporal (a), temporal (b), and spatial (c) densities of Proboscidean kill/scavenge sites for the Old World (Africa and Eurasia), Europe, and
Clovis.
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–97 89

Author's personal copy
that spatio-temporal region can be modeled the same way:
ne¼pharsthxi (14)
where r
s
is the per capita rate of archeological site
formation. Taking the ratio of these two equations then
allows us to model the relative frequency of elephant
subsistence sites (f):
f¼pharhthxi
pharsthxi
f¼rh
rs
(15)
Therefore, the relative frequency of elephant subsistence
sites (f) should be a function of only two variables, the per
capita rate of elephant subsistence site creation (r
h
) and the
per capita rate of site creation. All of the remaining
variables cancel out of the equation because they are
assumed to be constant.
Are these assumptions reasonable? With respect to
population density, area, and time these assumptions are
unquestionable since time and space are held constant. The
assumptions that research intensity (i) and site preservation
(x) are constant for PKS and all other sites are clearly
tenuous. For example, geomorphic and taphonomic factors
might bias preservation for or against sites showing
evidence for the use elephants relative to other site types.
Also, research bias might increase or decrease the like-
lihood that a PKS will be excavated relative to other site
types. However, in contrast to using spatio-temporal
densities as a measure of the relative frequency of
Proboscidean exploitation, when using site proportions,
these are the only two assumptions that are necessary.
4.1. Quantifying relative site frequencies
Data reporting the percentage of sites containing
Proboscidean remains from various spatio-temporal re-
gions of the Old World were compiled from four sources
(Soffer, 1985;Gamble, 1986, Tables 7.3–7.5; Borziyak,
1993;Vasil’ev, 2003).
3
These data (Table 3) are based on
the presence or absence of Proboscideans with no concern
for taphonomy. Therefore, we begin by comparing these
data to our estimate for Clovis, which is also based solely
on presence or absence in faunal assemblages (Waguespack
and Surovell, 2003). In our study, we found that
Proboscidean remains are present in 79% of Clovis sites
that have faunal assemblages. Some regions of the Old
World show comparable and even higher frequencies of
elephant remains (Fig. 3). For example, elephants are
found in 98% of Upper Paleolithic sites from the Central
Russian Plain (Soffer, 1985), and in 90% of cave
assemblages from the Upper Paleolithic (35–20 ka) of
Southern Germany (Gamble, 1986). On the other end of
the scale, fewer than 10% of Middle and Upper Paleolithic
sites from Cantabrian Spain and southern France contain
elephant remains, and various spatio-temporal regions
show intermediate values (Gamble, 1986).
On the surface it appears that the frequency of elephant
exploitation seen in Clovis may not be particularly unique.
However, these data, including those for Clovis, are
problematic since they include all occurrences of elephants,
whether they represent reliable subsistence associations or
not. Consider for example, the two cases where Probosci-
deans are more common than Clovis; many doubts linger
concerning whether elephants played any role in the
subsistence economies of these areas. For example, for
the Central Russian Plain, Soffer reports:
At present, therefore, I conclude that mammoth did not
play the major role in subsistence that has been widely
attributed to them in both Soviet and Western literature.
While some mammoth were undoubtedly hunted, how
many remains a question for future research, as does
determination of the frequency of mammoth kills y
(Soffer, 1985, p. 281)
Unfortunately, no unequivocal kill sites have been
found in the study area. (Soffer, 1985, p. 309)
In fact, none of Soffer’s sites would qualify as a subsistence
association using our criteria, essentially turning the 98%
frequency figure for the Central Russian Plain to 0%
(see also Soffer, 1985, Table 5.16). Likewise, for the Upper
Paleolithic of southern Germany, Niven reports:
Indisputable evidence of mammoth hunting has not yet
been documented in southern Germany although
mammoth remains are found in 90% of cave localities
in this region during the Upper Paleolithic. (Niven,
2001, p. 323)
Therefore, the 90% frequency data point for Southern
Germany, like the Central Russian Plain, can potentially be
reduced to 0%. What about Clovis?
Clovis contrasts starkly with these areas, and we argue
that it is like few other spatio-temporal regions arche-
ologically documented. In contrast to the regions from
Table 3 where Proboscidean remains are quite common in
the archeological record (e.g., the Central Russian Plain,
Siberia, England, the Crimea, and southern Germany) in
Clovis not only are elephant remains common, but Clovis
sites that include elephant remains in their faunal
assemblage frequently exhibit evidence of subsistence
exploitation of these animals (Haynes, 1999).
If we were to revise our original estimate of 79% by
eliminating questionable associations, 46.7% of Clovis sites
ARTICLE IN PRESS
3
The data in Tables 2 and 3 were compiled independently, and for those
in Table 3, in particular, we relied on the work of others (as cited). As
much as we would have liked to include spatio-temporal regions in Table 3
that included all of the PKS in Table 2, it would have been extremely
difficult to do so. This would require compiling nearly complete samples of
excavated sites with faunal remains from many parts of Africa, Europe,
and Asia. There are numerous difficulties involved in doing so, most
notably, it would require an impressive level of linguistic competence that
we admittedly lack. As such, we decided in would be prudent to rely on
compilations already produced by regional experts.
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–9790

Author's personal copy
or localities still show strong evidence for the subsistence
use of elephants. To derive this second figure, we exclude
Boaz, Schaefer, Leikem, and McLean because if we remove
the association with mammoths or mastodons in these
sites, they effectively lack culturally associated faunal
remains of any kind. Also, we count Blackwater Draw
twice since two localities provide evidence of mammoth
hunting (or scavenging?). We estimate that approximately
one-half of Clovis sites with preservation of bone show
strong evidence for the subsistence use of Proboscideans.
A second independent estimate is derived from Grayson
and Meltzer (2003). By querying the FAUNMAP database
ARTICLE IN PRESS
Table 3
The relative frequency of the occurrence of Proboscideans in faunal assemblages from various Old World spatio-temporal regions in comparison to Clovis
Region Time period % of sites or components
with Proboscideans
Sites and
components
Reference
Central Russian Plain Upper Paleolithic 93 16 Soffer (1985), Table 5.8
Southern Germany 35–20 kyr 90
a
20 Gamble (1986), Table 7.4
North America Clovis 79 33 Waguespack and Surovell (2003), Table 2
Crimea 118–35 kyr 69
a
13 Gamble (1986), Table 7.3
England 35–20 kyr 66
a
13 Gamble (1986), Table 7.4
Southern Germany 118–35 kyr 63
a
20 Gamble (1986), Table 7.3
Dnestr River, Moldova Late Upper Paleolithic 44 25 Borziyak (1993), Table 1
Siberia Middle Upper Paleolithic 35 40 Vasil’ev (2003), Table 4
Siberia Early Upper Paleolithic 32 31 Vasil’ev (2003), Table 3
Siberia Middle Paleolithic 30 33 Vasil’ev (2003), Table 2
Bu
¨kk Mountains, Hungary 35–20 kyr 30
a
20 Gamble (1986), Table 7.4
Southern Germany 20–10 kyr 19
a
47 Gamble (1986), Table 7.5
SW France 35–20 kyr 17
a
46 Gamble (1986), Table 7.4
Siberia Late Upper Paleolithic 15 235 Vasil’ev (2003), Table 5
England 20–10 kyr 15
a
33 Gamble (1986), Table 7.5
Switzerland 118–35 kyr 14
a
14 Gamble (1986), Table 7.3
SW France 20–10 kyr 9
a
70 Gamble (1986), Table 7.5
SW France 118–35 kyr 8
a
59 Gamble (1986), Table 7.3
S. France 118–35 kyr 7
a
38 Gamble (1986), Table 7.3
Cantabrian Spain 20–10 kyr 6
a
54 Gamble (1986), Table 7.5
Cantabrian Spain 118–35 kyr 5
a
19 Gamble (1986), Table 7.3
Cantabrian Spain 35–20 kyr 3
a
19 Gamble (1986), Table 7.4
a
Sample includes only cave assemblages.
Relative Frequency of Proboscideans
0%
25%
50%
75%
100%
Central Russian Plain (UP)
Southern Germany (EUP)
North America (Clovis)
Crimea (MP)
England (EUP)
Southern Germany (MP)
Dnestr River (LUP)
Siberia (MUP)
Siberia (EUP)
Siberia (MP)
Bükk Mtns, Hungary (EUP)
Southern Germany (LUP)
SW France (EUP)
Siberia (LUP)
England (LUP)
Switzerland (MP)
SW France (LUP)
SW France (MP)
Southern France (MP)
Cantabrian Spain (LUP)
Cantabrian Spain (MP)
Cantabrian Spain (EUP)
Spatio-Temporal Region
% of Assemblages
Fig. 3. The relative frequency of the occurrence of Proboscideans in faunal assemblages from various Old World spatio-temporal regions in comparisonto
Clovis. UP ¼Upper Paleolithic, EUP ¼Early Upper Paleolithic, MUP ¼Middle Upper Paleolithic, LUP ¼Late Upper Paleolithic, MP ¼Middle
Paleolithic.
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–97 91

Author's personal copy
under the terms ‘‘Clovis’’ and ‘‘Paleoindian’’, Grayson and
Meltzer (2002, p. 317) identified 76 sites in which ‘‘extinct
mammal remains of any kind were found in what appeared
to be a Clovis or Paleoindian archeological context.’’ Of
these, 47 ‘‘failed to produce minimally acceptable evidence
of an association between artifacts and extinct mammals’’
(Grayson and Meltzer, 2002, p. 321), leaving a total sample
of 29 sites. The 47 sites eliminated were cut due to
insufficient data, the presence of only bone tools, and/or
doubtful archeological status. Using solely these data, it is
difficult to determine precisely how many of these sites are
indeed archeological and Clovis, but their hesitance to even
include them in their careful analysis does warrant some
caution in using them to estimate the frequency of elephant
subsistence sites. Using their total sample (n¼76), 18.4%
of sites have reliable associations with Proboscideans.
However, according to Grayson and Meltzer (2003) in this
sample, 21 sites have archeological status that is ‘‘doubt-
ful’’, and an additional five sites are characterized only by
the presence of bone tools. Eliminating these sites provides
a minimum estimate of 28% of Clovis faunal assemblages
showing reliable evidence for the use of Proboscideans.
Alternatively, using their reduced sample of 29 sites, 46.7%
show reliable use of mammoth or mastodon (again
counting Blackwater Draw twice), precisely the figure we
derived from our sample.
4
Therefore, combining the
estimates derived from Grayson and Meltzer (2002) and
Waguespack and Surovell (2003), we can say with some
degree of confidence that between 28% and 46.7% of
Clovis sites with faunal remains show strong evidence
of the use of Proboscideans. Again, we are left with
the difficult task of determining whether these values
represent ‘‘a lot’’ or ‘‘a little’’ evidence of Proboscidean
subsistence use.
Comparing this to our Old World sample, Clovis
appears to be somewhat unique with respect to the
abundance of reliable subsistence associations with ele-
phants. Fig. 1 shows the spatio-temporal distributions of
sites showing strong or possible evidence for the use of
Proboscideans for the Old World and Clovis samples. The
Clovis record shows a clear cluster in space and time, while
the record of elephant exploitation in the Old World is
considerably more dispersed. Perhaps the spatio-temporal
region most similar to Clovis is the Lower Paleolithic of the
Iberian Peninsula where four localities, Aridos I and II,
Torralba, and Ambrona cluster fairly tightly in space and
time. Other possible regions include the Lower Paleolithic
of East Africa, with four sites represented (over a period
of more than 1 million years), the Middle Paleolithic of
northern Germany with two sites, the Upper Paleolithic of
Eastern Europe with two sites, and the late Upper
Paleolithic of Siberia with two sites.
Do any of these regions even meet the minimum estimate
of Clovis with 28% of sites with faunal remains being
elephant subsistence sites? In part, the answer to this
question depends on how the spatio-temporal regions are
defined. If a region is defined very narrowly encompassing
one or two elephant-bearing sites and little else, then
elephant subsistence sites will appear very common. If a
region is defined broadly, then such sites become diluted in
larger samples. Using our sample of Old World sites, there
is at least one reasonably sized region nears Clovis in terms
of the relative frequency of Proboscidean exploitation
sites—the Lower Paleolithic of the Iberian Peninsula where
four sites are known. A compilation Lower Paleolithic sites
and components from this region from Straus (1992),
Carbonell et al. (1995),Gibert et al. (1998), and Santonja
and Villa (1990) yielded a total of 15 additional arche-
ological components with lithics and fauna.
5
Including,
Torralba and Ambrona, 21% of the Lower Paleolithic sites
from this area are elephant subsistence sites, and in
comparison to the minimum and maximum estimate for
Clovis, these differences are not statistically significant
(Fisher’s exact test, Iberia versus Clovis Min., p¼0.398;
Iberia versus Clovis Max., p¼0.064).
The Middle Paleolithic of Germany, like the Lower
Paleolithic of Spain, has two sites, Lehringen and Gro
¨bern,
showing strong evidence for the use of Proboscideans, but
unlike Iberia, it has a much greater number of sites that do
not (e.g., Gamble, 1986, 1999;Conard and Prindiville,
2000). The same is true of the Late Upper Paleolithic of
Siberia where two sites, Lugovskoye and Shikaevka II,
suggest the subsistence use of mammoth, but Vasil’ev
(2003) reports 234 additional archeological components
with faunal remains. Similarly, in the Upper Paleolithic of
Eastern Europe, despite mammoth bone being very
common in sites of this region, sites with evidence for the
subsistence use of mammoth (Nowa Huta, Halich, and
Skaratki according to our definitions) are swamped out by
those that do not (e.g., Klein, 1973;Soffer, 1985;
Kozlowski, 1986). Lower Paleolithic East Africa also
contains far more localities with artifacts and faunal
assemblages that do not show subsistence use of elephants
than do (e.g., Leakey, 1971;Isaac, 1977;Chavaillon et al.,
1979;Bunn, 1982;Clark et al., 1984).
Therefore, with the single possible exception of the
Lower Paleolithic of Iberia, Clovis appears to be uniquely
rich in evidence of the subsistence use of elephants.
Although we have identified other possible regions
where there appear to be minor clusters in elephant use,
such as Lower Paleolithic East Africa, Middle Paleolithic
Germany, and Upper Paleolithic Eastern Europe,
ARTICLE IN PRESS
4
The reduced Grayson and Meltzer (2002, 2003) sample, what they call
‘‘sites that made the first cut’’, includes four sites that contain only
proboscidean remains (Boaz, Schaefer, Burning Tree, and Pleasant Lake).
If the associations at these sites are eliminated, there is no evidence for the
use of any other taxa at these sites, and removing these sites from the
sample increases the Clovis estimate to 53%.
5
These are El Castillo (Levels 24, 25a, 25b, 26), Pinedo, Arganda II,
Oxı
´geno, San Isidro, Cu´ llar-Baza I, Solan del Zambotino (Multiple
Levels), Cueva Hora
´, Gran Dolina (Levels TD4 and TD6), Barranco
Leo
´n-5, and Fuentenueva-3a.
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–9792

Author's personal copy
elephants and hominids have occupied most of Africa,
Europe, and Asia for hundreds of thousands of years;
yet, for most of archeological time and space, there is
little if any evidence of the subsistence utilization of
Proboscideans.
5. Discussion
In recent years, it has become commonplace to question
many of the things that we once ‘‘knew’’ about Clovis. The
‘‘old Clovis model’’ can be succinctly stated as follows.
Clovis peoples were highly mobile, prodigiously reprodu-
cing, specialized hunters of large game, who migrated to
the unpopulated continent of North America from Asia via
the ice-free corridor approximately 12,000 years ago, and
upon their arrival promptly hunted the Pleistocene mega-
fauna to extinction. Virtually every descriptive phrase in
this statement has been questioned despite having once
been widely accepted. While we strongly believe that
questioning and of reevaluating entrenched ideas is a good
thing, we suggest that it is time to step back and truly
evaluate what the archeological record of Clovis represents
and detach issues of Clovis adaptations from Clovis origins
(Meltzer, 1989, p. 477). With this so-called ‘‘paradigm-
shift’’ (Bonnichsen and Turnmire, 1999) in Paleoindian
archeology, it has become routine, for example, to discount
the record of Clovis large game hunting using various
arguments ranging from the theoretical to the empirical
(Meltzer and Smith, 1986;Meltzer, 1993;Dixon, 1999;
Dillehay, 2000;Adovasio and Page, 2002;Grayson and
Meltzer, 2002, 2003;Cannon and Meltzer, 2004;Chilton,
2004;Byers and Ugan, 2005). One way of doing this is to
point out a perceived rarity of evidence of the exploitation
of extinct elephants:
That there are so few actual sites of mammoth kills
(and virtually none of horse, camel, or sloth kills)
is a major embarrassment to the overkill theory y
[emphasis added] (Adovasio and Page, 2002, p. 127)
Clovis is the only firmly documented New World
archaeological complex positively associated with mam-
moth procurement. However, only about twelve archae-
ological sites have been reported where Clovis projectile
points are associated with the bones of these large
extinct elephant-like creatures yAlthough there is no
doubt that some Clovis people hunted mammoth and
possibly mastodon, current research suggests that these
rather spectacular kill sites are probably not typical of
Clovis culture. [emphasis added] (Dixon, 1999, p. 216)
Obviously, given our arguments above, we should not be
arguing that there are ‘‘so few’’ or ‘‘only about twelve’’
mammoth kills, but instead asking why there are so many?
In over 1 million years of archeology spread over four
continents, we have attempted to demonstrate that there is
likely nothing that has yet to be documented archeologi-
cally that compares to Clovis in terms of the frequency of
Proboscidean exploitation, with the single possible excep-
tion of the Lower Paleolithic of Iberia.
Certainly, 14 sites do not seem like a very large number,
but when viewed in a comparative context, it is in fact a
very large number. Furthermore, the number of elephant
kill sites in Clovis is truly remarkable when we consider the
total number of elephant kills documented from four
continents. Of the 21 sites we have identified in the Old
World, only two or three have weaponry associated with
carcasses. The two best cases are Lehringen and Lugovs-
koye, and the third case, Gro
¨bern, is questionable. There-
fore, in the entire archeological records of Africa, Europe,
Asia, and North America, there are a total of 16 strong
cases for hunting of elephants, and 14 of these are found in
Clovis. Furthermore, between of 28% and 46.7% of
excavated Clovis sites that have preserved fauna are
mammoth or mastodon kill sites. Why?
According to our model of the archeological frequency
of elephant subsistence sites, there are three possible
answers. The first possibility is that Clovis hunter-gatherers
regularly hunted mammoth and mastodon, or at least did
so more frequently than people did in any other time and
place in the world. The second possibility is that the
archeological record of North America is strongly biased
toward elephant kills, and/or the record of the Old World
is biased against elephant kills. In other words, our sample
of archeological sites from the Clovis record is not random,
and in particular, it is biased toward the discovery
elephant-bearing sites over sites that do not contain
elephants. Such biases can conceivably come from many
different factors, such as site preservation, discovery, or
excavation. The third possibility is that both factors are
operating. Clovis peoples frequently hunted Probosci-
deans, and the record is biased toward the discovery of
these sites.
At this point, we are somewhat at a loss. Despite
attempts to show that the Clovis faunal record is biased
(Cannon and Meltzer, 2004), it is impossible to directly test
a hypothesis about sample bias without a theoretical or
empirical null model of the population that is sampled (see
Surovell and Waguespack, in press). Therefore, the most
that we are willing to conclude with any confidence is that
the archeological record of Clovis Proboscidean exploita-
tion is unique. Whether that is a function of the behavior of
Paleoindians, Paleoindian archeologists, or geological
systems is very difficult to address, and ultimately may
have to be resolved with large sample sizes and/or different
methods of sampling the archeological record. For the time
being, we are, by necessity, forced to make arguments
concerning the plausibility of sample bias and the frequent
killing of Proboscideans.
Elsewhere, we have presented theoretical arguments
concerning the plausibility and likelihood of specialized
large mammal hunting in late Pleistocene North America
(Waguespack, 2003;Waguespack and Surovell, 2003). We
argue that a focal large game subsistence economy during
Clovis times is not only plausible but also likely and
ARTICLE IN PRESS
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–97 93

Author's personal copy
theoretically expected (but see Byers and Ugan, 2005).
Concerning the issue of sample bias, we begin with
discovery bias, perhaps the most frequently invoked bias
in discussions of Clovis subsistence. Grayson (1988) and
Meltzer (1989) have argued that Clovis elephant kill sites
have a greater probability of discovery than other Clovis
sites because elephant bones are extremely visible in
cutbanks, backhoe trenches, or surface exposures (see also
Cannon and Meltzer, 2004). In contrast, an exposed
Clovis-aged flake or rabbit bone does not have the same
visibility and therefore probability of discovery. This would
inevitably bias discovery toward elephant-bearing sites,
and we agree. However, this bias should not be unique to
Clovis, and there is no reason to believe that it would not
apply to our Old World sample as well. In fact, many of the
sites we have identified were discovered this way- first the
elephant bone, then the artifacts. In other words, there very
well may be a bias operating but in this comparative
framework, it should be more or less constant.
Is there a preservation bias? Grayson and Meltzer (2002,
2003) have pointed out that four of the sites in the Clovis
sample are from one very small corner of southeastern
Arizona, and Holliday (2003, p. 1373) attributes this cluster
to ‘‘fortuitous circumstances of preservation and expo-
sure’’. Similarly, Grayson (1988) has written:
In short, the history of the discovery of Clovis sites in
Southwest and Plains is such as to suggest that we have
an extraordinarily biased sample of those sites, one that
may have given us a greatly exaggerated impression of
the importance of large mammals in the daily lives of
those people. (Grayson, 1988, p. 115)
There is no question that the geochemical and geomorphic
conditions in Eastern North America and the Far West
leave a lot to be desired with respect to the preservation of
Paleoindian faunal assemblages. Certainly, the vast major-
ity of Clovis sites with faunal remains come from a limited
portion of North America, and extrapolating what we see
in the Plains and Southwest to the remainder of the North
America is a precarious practice (Meltzer and Smith, 1986;
Meltzer, 1988, 1989, 1993). However, sites in both the East
and the Far West evidencing the possible subsistence use of
elephants do exist (e.g., Rayl, 1974;Gustafson et al., 1979;
Graham et al., 1981;Laub et al., 1988;Overstreet and
Kolb, 2003). Furthermore, the sample we have is the
sample with which we must work, and in that sample,
elephant kill sites are remarkably common.
Could it be that Clovis elephant kills have a greater
chance of preservation than Old World elephant kills?
Certainly, because Clovis sites are considerably younger
than virtually our entire Old World sample, they have
withstood considerably less time over which destructive
forces have operated. However, when we quantify the
relative frequency of elephant subsistence sites for various
spatio-temporal regions, site preservation is a constant and
factors out of the equation (Eq. (15)). The critical question
is whether Clovis elephant kills relative to other Clovis sites
have a greater probability of preservation than the same
ratio for Old World contexts. This question is extraordi-
narily difficult to answer. Certainly, African elephants
often die near water (Conybeare and Haynes, 1984;
Haynes, 1987, 1991), and archeologically, elephant sub-
sistence sites are found near ancient water sources, but this
is true for the Old and New World samples. Obviously, wet
lowland settings act as sediment traps increasing the
likelihood of burial and preservation over upland settings.
However, campsites, with or without elephant bone, are
often found in similar contexts. At this point, we are
content to leave this question unresolved.
Let us surmise for a moment that the abundance of
elephant kill sites in the Clovis record is due to human
behavior. Is it reasonable to expect that one out of two, or
even one out of four archeological sites created during
Clovis times were elephant kills? We suspect not. To us,
this number seems very high and strongly suggests that
some biases are likely operating. However, those same
biases should operate in the Old World as well, and
therefore, we suspect that a strong behavioral signal is
coming through our sample, biased or not. In other words,
the abundance of mammoth and mastodon kills known
from Clovis contexts relative to Old World contexts is
largely a product of past human behavior. Does this mean
that mammoth meat was served on the dinner table every
night in the latest Pleistocene of North America? Abso-
lutely not. From our analysis, we cannot estimate an
absolute frequency of elephant hunting. All we can say is
that in comparison to the Old World record, Clovis peoples
seem to have exploited elephants with much greater
frequency than in any other time and place.
Acknowledgments
We would like to thank Matt Hill, Joaquin Arroyo, and
Oscar J. Polaco for providing us the opportunity to
contribute to this volume, and Bob Kelly and David
Meltzer, who provided valuable comments on an earlier
version of this paper.
References
Adam, K.D., 1951. Der waldelefant von Lehringen, eine jagdbeute des
diluvialen Menschen. Quarta
¨r 5, 79–92.
Adovasio, J.M., Page, J., 2002. The First Americans: In Pursuit of
Archaeology’s Greatest Mystery. Random House, New York.
Baryshnikov, G., 2003. Mammuthus primigenius from the Crimea and the
Caucasus. Deinsea 9, 41–56.
Belmaker, M., Tchernov, E., Condemi, S., Bar-Yosef, O., 2002. New
evidence for hominid presence in the Lower Pleistocene of the
Southern Levant. Journal of Human Evolution 43, 43–53.
Berthelet, A., Chavaillon, J., 2001. The Early Palaeolithic butchery site of
Barogali (Republic of Djibouti). In: Cavarretta, G., Gioia, P., Mussi,
M., Palombo, M.R. (Eds.), La Terra Degli Elefanti. Consiglio
Nazionale delle Ricerche, Rome, pp. 176–179.
Binford, L.R., 1987. Were there elephant hunters at Torralba? In: Nitecki,
M.H., Nitecki, D.V. (Eds.), The Evolution of Human Hunting.
Plenum Press, New York, pp. 47–105.
ARTICLE IN PRESS
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–9794

Author's personal copy
Bonnichsen, R., Turnmire, K.L., 1999. An introduction to the peopling of
the Americas. In: Bonnichsen, R., Turnmire, K.L. (Eds.), Ice Age
Peoples of North America: Environments, Origins, and Adaptations of
the First Americans. Center for the Study of the First Americans,
Corvallis, Oregon, pp. 1–26.
Borziyak, I.A., 1993. Subsistence practices of Late Paleolithic groups
along the Dnestr River and its tributaries. In: Soffer, O., Praslov, N.D.
(Eds.), From Kostenki to Clovis: Upper Paleolithic: Paleo-Indian
Adaptations. Plenum Press, New York, pp. 67–84.
Brown, J.A., 1889. On the discovery of Elephas primegenius associated
with flint implements at Southall. Proceedings of the Geologist’s
Association 10, 361–372.
Brunswig Jr., R.H., Fisher, D.C., 1993. Research on the Dent mammoth
site. Current Research in the Pleistocene 10, 63–65.
Brush, N., Smith, F., 1994. The Martins Creek mastodon: a paleoindian
butchery site in Holmes County, Ohio. Current Research in the
Pleistocene 11, 14–15.
Brush, N., Newman, M., Smith, F., 1994. Immunological analysis of flint
flakes from the Martins Creek Mastodon site. Current Research in the
Pleistocene 11, 16–18.
Bunn, H., 1982. Meat-eating and human evolution: studies on the diet and
subsistence patterns of Plio-Pleistocene hominids in East Africa.
Unpublished Ph.D. Dissertation. Department of Anthropology,
University of California Berkeley.
Byers, D.A., Ugan, A., 2005. Should we expect large game specialization
in the late Pleistocene? An optimal foraging perspective on early
Paleoindian prey choice. Journal of Archaeological Science 32,
1624–1640.
Callow, P., 1986. An overview of the industrial succession. In: Callow, P.,
Cornford, J.M. (Eds.), La Cotte de St. Brelade 1961–1978: Excava-
tions by C.B.M. Burney. Geo Books, Norwood, pp. 219–230.
Cannon, M.D., Meltzer, D.J., 2004. Early Paleoindian foraging: examin-
ing the faunal evidence for megafaunal specialization and regional
variability in prey choice. Quaternary Science Reviews 23, 1955–1987.
Carbonell, E., Castro, J.M.B.d., 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, 826–829.
Chavaillon, J., Chavaillon, N., Hours, F., Piperno, M., 1979. From the
Oldowan to the Middle Stone Age at Melka-Kunture
´(Ethiopia).
Understanding cultural changes. Quaternaria 21, 87–114.
Chilton, E.S., 2004. Beyond ‘‘Big’’. In: Barton, C.M., Clark, G.A., Yesner,
D.R., Pearson, G.A. (Eds.), The Settlement of the American
Continents. The University of Arizona Press, Tucson, AR,
pp. 162–172.
Chmielewski, W., Kubiak, H., 1962. The find of mammoth bones at
Skaratki in the Lowicz District. Folia Quaternaria 9, 1–29.
Clark, J.D., Haynes Jr., C.V., 1970. An elephant butchery site at
Mwanganda’s Village, Karonga, Malawi, and its relevance for
Palaeolithic archaeology. World Archaeology 1, 390–411.
Clark, J.D., Asfaw, B., Assefa, G., Harris, J.W.K., Kurashina, H., Walter,
R.C., White, T.D., Williams, M.A.J., 1984. Palaeoanthropological
discoveries in the Middle Awash Valley, Ethiopia. Nature 307,
423–428.
Conard, N.J., Prindiville, T.J., 2000. Middle Paleolithic hunting econo-
mies in the Rhineland. International Journal of Osteoarchaeology 10,
286–309.
Conybeare, A., Haynes, G., 1984. Observations on elephant mortality and
bones in water holes. Quaternary Research 22, 189–200.
Delagnes, A., Lenoble, A., Harmand, S., Brugal, J.-P., Prat, S., Tiercelin,
J.-J., Roche, H., 2006. Interpreting pachyderm single carcass sites in
the African Lower and Early Middle Pleistocene record: a multi-
disciplinary approach to the site of Nadung’a 4 (Kenya). Journal of
Anthropological Archaeology 25, 448–465.
Dennell, R., 2003. Dispersal and colonisation, long and short
chronologies: how continuous is the Early Pleistocene record for
hominids outside East Africa? Journal of Human Evolution 45,
421–440.
Dillehay, T.D., 2000. The Settlement of the Americas: A New Prehistory.
Basic Books, New York.
Dixon, E.J., 1999. Bones, Boats, and Bison: Archeology and the First
Colonization of Western North America. University of New Mexico
Press, Albuquerque, USA.
Figgins, J.D., 1933. A further contribution to the antiquity of man in
America. Proceedings of Colorado Museum of Natural History XII(2).
Fisher, D.C., 1987. Mastodont procurement by Paleoindians of the Great
Lakes region: hunting or scavenging. In: Nitecki, M.H., Nitecki, D.V.
(Eds.), The Evolution of Human Hunting. Plenum Press, New York,
pp. 309–421.
Freeman, L.G., 1973. The analysis of some occupation floor distributions
from Earlier and Middle Paleolithic sites in Spain. In: Freeman, L.G.
(Ed.), Views of the Past: Essays in Old World Prehistory and
Paleoanthropology. Mouton, The Hague, The Netherlands,
pp. 57–115.
Frison, G.C., Todd, L.C., 1986. The Colby Mammoth Site. University of
New Mexico Press, Albuquerque, USA.
Gabunia, L., Anto
´n, S.C., Lordkipanidze, D., Vekua, A., Justus, A.,
Swisher III, C.C., 2001. Dmanisi and dispersal. Evolutionary Anthro-
pology 10, 158–170.
Gamble, C., 1986. The Paleolithic Settlement of Europe. Cambridge
University Press, Cambridge.
Gamble, C., 1987. Man the shoveler: alternative models for middle
Pleistocene colonization and occupation in northern latitudes. In:
Soffer, O. (Ed.), The Pleistocene Old World, Regional Perspectives.
Plenum Press, New York, pp. 81–98.
Gamble, C., 1999. The Paleolithic Societies of Europe. Cambridge
University Press, Cambridge.
Gibert, J., Gibert, L., Iglesias, A., Maestro, E., 1998. Two ‘Oldowan’
assemblages in the Plio-Pleistocene deposits of the Orce region,
southeast Spain. Antiquity 72, 17–25.
Goren-Inbar, N., Lister, A., Werker, E., Chech, M., 1994. A butchered
elephant skull and associated artifacts from the Acheulian site of
Gesher Benot Ya’aqov, Israel. Pale
´orient 20, 99–112.
Graham, R.W., Haynes, C.V., Johnson, D.L., Kay, M., 1981. Kimms-
wick: a Clovis–Mastodon Association in eastern Missouri. Science 213,
1115–1117.
Grayson, D.K., 1983. The Establishment of Human Antiquity. Academic
Press, New York.
Grayson, D.K., 1988. Perspectives on the archaeology of the first
Americans. In: Carlisle, R.C. (Ed.), Americans Before Columbus:
Ice-Age Origins. Ethnography Monographs 12, Department of
Anthropology, University of Pittsburgh, Pittsburgh, Pennsylvania,
pp. 107–122.
Grayson, D.K., Meltzer, D.J., 2002. Clovis hunting and large mammal
extinction: a critical review of the evidence. Journal of World
Prehistory 16, 313–359.
Grayson, D.K., Meltzer, D.J., 2003. A requiem for North American
overkill. Journal of Archaeological Science 30, 585–593.
Gustafson, C.E., Gilbow, D., Daugherty, R.D., 1979. The Manis
mastodon: early man on the Olympic Peninsula. Canadian Journal
of Archaeology 3, 157–164.
Hannus, L.A., 1989. Flaked mammoth bone from the Lange/Ferguson
site White River Badlands area, South Dakota. In: Bonnichsen, R.,
Sorg, M.H. (Eds.), Bone Modification. Center for the Study of the
First Americans, University of Maine, Orono, ME, pp. 395–412.
Haury, E.W., 1953. Artifacts with mammoth remains, Naco, Arizona.
American Antiquity 19, 1–14.
Haury, E.W., Sayles, E.B., Wasley, W.W., 1959. The Lehner mammoth
site, southeastern Arizona. American Antiquity 25, 2–42.
Haynes Jr., C.V., 1966. Elephant-hunting in North America. Scientific
American 214, 104–112.
Haynes, G., 1987. Where elephants die. Natural History 96, 28–33.
Haynes, G., 1991. Mammoths, Mastodonts, and Elephants. Cambridge
University Press, Cambridge.
Haynes Jr., C.V., 1992. Contributions of radiocarbon dating to the
geochronology of the peopling of the New World. In: Taylor, R.E.,
ARTICLE IN PRESS
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–97 95

Author's personal copy
Long, A., Kra, S. (Eds.), Radiocarbon After Four Decades. Springer,
New York, pp. 355–374.
Haynes Jr., C.V., 1993. Clovis–Folsom geochronology and climatic
change. In: Soffer, O., Praslov, N.D. (Eds.), From Kostenki to
Clovis: Upper Paleolithic-Paleo-Indian Adaptations. Plenum Press,
New York, pp. 219–236.
Haynes, G., 1999. The role of mammoths in rapid Clovis dispersal.
Deinsea 6, 9–38.
Haynes, G., 2002. The Early Settlement of North America: The Clovis
Era. Cambridge University Press, Cambridge.
Haynes, G., Stanford, D., 1984. On the possible utilization of Camelops by
early man in North America. Quaternary Research 22, 216–230.
Hemmings, E.T., Haynes Jr., C.V., 1969. The Escapule Mammoth and
associated projectile points, San Pedro valley, Arizona. Journal of the
Arizona Academy of Science 5, 184–188.
Hester, J.J., 1972. Blackwater Locality No. 1: A Stratified, Early Man Site
in Eastern New Mexico. Fort Burgwin Research Center, Southern
Methodist University, Rancho de Taos, NM.
Hoffman, C.A., 1983. A mammoth kill site in the Silver Springs Run. The
Florida Anthropologist 36, 83–87.
Holliday, V.T., 2003. Where have all the mammoth gone? Science 300,
1373–1374 (review of: The Early Settlement of North America by G.
Haynes).
Howell, F.C., 1966. Observations on the earlier phases of the European
Lower Paleolithic. American Anthropologist 68, 88–201.
Howell, F.C., Freeman, L.G., 1983. Ivory Points from the earlier
Acheulean of the Spanish meseta. In: Homenaje al Professor Martin
Almagro Basch. Ministerio de Cultura, Madrid, pp. 41–61.
Isaac, G.L., 1977. Olorgesaile: Archeological Studies of a Middle
Pleistocene Lake Basin in Kenya. University of Chicago Press,
Chicago.
Kaufulu, Z.M., 1990. Sedimentary environments at the Mwanganda site,
Malawi. Geoarchaeology 5, 15–27.
Klein, R.G., 1973. Ice-Age Hunters of the Ukraine. The University of
Chicago Press, Chicago.
Klein, R.G., 1987. Reconstructing how early people exploited animals:
problems and prospects. In: Nitecki, M.H., Nitecki, D.V. (Eds.), The
Evolution of Human Hunting. Plenum Press, New York, pp. 11–45.
Klein, R.G., 1999. The Human Career. University of Chicago Press,
Chicago.
Klein, R.G., Edgar, B., 2002. The Dawn of Human Culture. Wiley,
New York.
Kondo, Y., Mazima, N., Group, N.-K.R., 2001. Palaeoloxodon naumanni
and its environment at the Paleolithic site of Lake Nojiri, Nagano
Prefecture, central Japan. In: Cavarretta, G., Gioia, P., Mussi, M.,
Palombo, M.R. (Eds.), La Terra Degli Elefanti. Consiglio Nazionale
delle Ricerche, Rome, pp. 284–288.
Kozlowski, J.K., 1986. The Gravettian in Central and Eastern Europe. In:
Wendorf, F., Close, A. (Eds.), Advances in World Archaeology, vol. 5.
Academic Press, New York, pp. 131–200.
Kozlowski, J.K., Kubiak, H., Welc, A., 1970. Palaeolithic site with
mammoth remains at Nowa Huta (Cracow, Poland). Folia Quaternar-
ia 36, 1–19.
Kuzmin, Y.V., Orlova, L.A., Zolnikov, I.D., 2003. Dynamics of the
mammoth (Mammuthus primigenius) population in Northern Asia:
radiocarbon evidence. Deinsea 9, 221–237.
Laub, R.S., DeRemer, M.F., Duport, C.A., 1988. The Hiscock site: a rich
late Quaternary locality in western New York State. In: Laub, R.S.,
Miller, N.G., Steadman, D.W. (Eds.), Late Pleistocene and Early
Holocene Paleoecology and Archaeology of the Eastern Great Lakes
Region, vol. 33. Bulletin of the Buffalo Society of Natural Sciences,
Buffalo, pp. 67–81.
Leakey, M.D. (Ed.), 1971. Olduvai Gorge, Excavations in Beds I and II,
1960–1963, vol. 3. Cambridge University Press, Cambridge.
Leonhardy, F.C., 1966. Domebo: A Paleo-Indian Mammoth Kill in the
Prairie-Plains. Contributions of the Museum of the Great Plains,
Lawton, OK.
Lister, A., Bahn, P., 1994. Mammoths. Box Tree, London.
MacCalman, H.R., 1967. The Zoo Park elephant site, Windhoek.
Palaeoecology of Africa 2, 102–103.
Mania, D., Thomae, M., Litt, T., Weber, T. (Eds.), 1990. Neumark-
Gro
¨ben. Beitra
¨ge zur Jagd des Mittelpala
¨olithischen Menschen.
Deutscher Verlag de Wissenschaften, Berlin.
Martin, P.S., 1973. The discovery of America. Science 179, 969–974.
Martin, P.S., 1984. Prehistoric overkill: the global model. In: Martin, P.S.,
Klein, R.G. (Eds.), Quaternary Extinctions. The University of Arizona
Press, Tucson, AR, pp. 354–403.
Martin, P.S., Steadman, D.W., 1999. Prehistoric extinctions on islands
and continents. In: MacPhee, R. (Ed.), Extinctions in Near Time.
Kluwer Academic/Plenum Publishers, New York, pp. 17–55.
Maschenko, E.N., Pavlov, A.F., Zenin, V.N., Leshchinskiy, S.V., Orlova,
L.A., 2003. The Lugovskoe site: relations between the mammoth
assemblage and late Paleolithic Man (L). In: Proceedings of 3rd
International Mammoth Conference—Abstracts Volume, Dawson
City, Yukon Territory, Canada, pp. 77–80.
Mellars, P., 1996. The Neanderthal Legacy: An Archaeological Perspec-
tive from Western Europe. Princeton University Press, Princeton.
Meltzer, D.J., 1988. Late Pleistocene human adaptations in eastern North
America. Journal of World Prehistory 2, 1–52.
Meltzer, D.J., 1989. Why don’t we know when the first people came to
North America? American Antiquity 54, 471–490.
Meltzer, D.J., 1993. Is there a Clovis adaptation? In: Soffer, O., Praslov,
N.D. (Eds.), From Kostenki to Clovis: Upper Paleolithic: Paleo-
Indian Adaptations. Plenum Press, New York, pp. 293–310.
Meltzer, D.J., Mead, J.I., 1985. Dating late Pleistocene extinctions:
theoretical issues, analytical bias, and substantive results. In: Mead,
J.I., Meltzer, D.J. (Eds.), Environments and Extinctions: Man in Late
Glacial North America. Center for the Study of Early Man, Orono,
ME, pp. 145–174.
Meltzer, D.J., Smith, B.D., 1986. Paleoindian and Early Archaic
subsistence strategies in eastern North America. In: Neusius, S.
(Ed.), Foraging, Collecting, and Harvesting: Archaic Period Sub-
sistence and Settlement in the Eastern Woodlands. Center for
Archaeological Investigations, Southern Illinois University, Carbon-
dale, pp. 3–31.
Mosimann, J.F., Martin, P.S., 1975. Simulating overkill by Paleoindians.
American Scientist 63, 304–313.
Movius Jr., H.L., 1950. A wooden spear of the third interglacial age from
lower Saxony. Southwestern Journal of Anthropology 6, 139–142.
Niven, L., 2001. The role of mammoths in the Upper Palaeolithic
economies of southern Germany. In: Cavarretta, G., Gioia, P., Mussi,
M., Palombo, M.R. (Eds.), La Terra Degli Elefantis. Consiglio
Nazionale delle Ricerche, Rome, pp. 323–327.
Overstreet, D.F., 1996. Still more on cultural contexts of mammoth and
mastodon in the southwestern Lake Michigan basin. Current Research
in the Pleistocene 13, 36–38.
Overstreet, D.F., Kolb, M.F., 2003. Geoarchaeological contexts for late
Pleistocene archaeological sites with human modified woolly mam-
moth remains in southeastern Wisconsin, USA. Geoarchaeology 18,
91–114.
Palmer, H.A., Stoltman, J.B., 1975. The Boaz Mastodon: a possible
association of man and mastodon in Wisconsin. Midcontinental
Journal of Archaeology 1, 163–177.
Piperno, M., Tagliacozzo, A., 2001. The elephant butchery area at the
Middle Pleistocene site of Notarchirico (Venoa, Basilicata, Italy).
In: Cavarretta, G., Gioia, P., Mussi, M., Palombo, M.R. (Eds.),
La Terra Degli Elefanti. Consiglio Nazionale delle Ricerche, Rome,
pp. 230–236.
Pitulko, V.V., Nikolsky, P.A., Girya, E.Y., Basilyan, A.E., Tumoskoy,
V.E., Koulakov, S.A., Astakhov, S.N., Pavlova, E.Y., Anismov, M.A.,
2004. The Yana RHS site: humans in the arctic before the last glacial
maximum. Science 303, 52–56.
Potts, R., Behrensmeyer, A.K., Ditchfield, P., 1999. Paleolandscape
variation and Early Pleistocene hominid activities: members 1 and 7,
Olorgesailie Formation, Kenya. Journal of Human Evolution 37,
747–788.
ARTICLE IN PRESS
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–9796

Author's personal copy
Ramsey, B., 2003. OxCal Program, v. 3.9. University of Oxford
Radiocarbon Accelerator Unit, Oxford.
Ray, C.N., 1942. Ancient artifacts and mammoth’s teeth of the McLean
site. Bulletin of the Texas Archaeological and Paleontological Society
14, 137–146.
Ray, C.N., Bryan, K., 1938. Folsomoid point found in alluvium beside a
mammoth’s bones. Science 88, 257–258.
Rayl, S.L., 1974. A Paleo-Indian mammoth kill site near Silver Springs,
Florida. Unpublished M.A. Thesis, Department of Anthropology,
Northern Arizona University.
Roebroeks, W., 2001. Hominid behaviour and the earliest occupation of
Europe: an exploration. Journal of Human Evolution 41, 437–461.
Roebroeks, W., Kolfschoten, T.v., 1994. The earliest occupation of
Europe: a short chronology. Antiquity 68, 489–503.
Santonja, M., Villa, P., 1990. The lower Paleolithic of Spain and Portugal.
Journal of World Prehistory 4, 45–94.
Santonja, M., Pe
´rez-Gonza
´lez, A., Villa, P., Soto, E., Sese
´, C., 2001.
Elephants in the archaeological sites of Aridos (Jarama Valley,
Madrid, Spain). In: Cavarretta, G., Gioia, P., Mussi, M., Palombo,
M.R. (Eds.), La Terra Degli Elefanti. Consiglio Nazionale delle
Ricerche, Rome, pp. 602–606.
Scott, K., 1980. Two hunting episodes of Middle Paleolithic age at La
Cotte de Saint Brelade, Jersey (Channel Islands). World Archaeology
12, 137–152.
Scott, K., 1986. The bone assemblages of layers 3 and 6. In: Callow, P.,
Cornford, J.M. (Eds.), La Cotte de St. Brelade 1961–1978: Excava-
tions by C.B.M. Burney. Geo Books, Norwood, pp. 159–183.
Sellards, E.H., 1952. Early Man in America: A Study in Prehistory.
University of Texas Press, Austin.
Shipman, P., Rose, J., 1983. Evidence of butchery and hominid activities
at Torralba and Ambrona; an evaluation using microscopic techni-
ques. Journal of Archaeological Science 10, 465–474.
Soffer, O., 1985. The Upper Paleolithic of the Central Russian Plain.
Academic Press, San Diego.
Soffer, O., 1993. Upper Paleolithic adaptations in Central and Eastern
Europe and man-mammoth interactions. In: Soffer, O., Praslov, N.D.
(Eds.), From Kostenki to Clovis: Upper Paleolithic: Paleo-Indian
Adaptations. Plenum Press, New York, pp. 31–49.
Straus, L.G., 1992. Iberia Before the Iberians. University of New Mexico
Press, Albuquerque, USA.
Stuart, A.J., Lister, A.M., 2001. The Late Quaternary extinction of woolly
mammoth (Mammuthus primigenious) and straight-tusked elephant
(Palaeoloxodon antiquus) and other megafauna in Europe. In:
Cavarretta, G., Gioia, P., Mussi, M., Palombo, M.R. (Eds.), La Terra
Degli Elefanti. Consiglio Nazionale delle Ricerche, Rome,
pp. 722–723.
Stuart, A.J., Sulerzhitsky, L.D., Orlova, L.A., Kuzmin, Y.V., Lister, A.,
2002. The latest woolly mammoths (Mammuthus primigenius Blumen-
bach) in Europe and Asia: a review of the current evidence. Quaternary
Science Reviews 21, 1559–1569.
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, 684–689.
Surovell, T.A., Brantingham, P.J., 2007. A note on the use of temporal
frequency distributions in studies of prehistoric demography. Journal
of Archaeological Science 34, 1868–1877.
Surovell, T.A., Waguespack, N.M., in press. Human prey choice in the
Late Pleistocene and its relation to megafaunal extinctions. In: Haynes,
G. (Ed.), American Megafaunal Extinctions at the End of the
Pleistocene. Springer.
Svoboda, J., Lozˇ ek, V., Vlcek, E., 1996. Hunters Between East and West:
The Paleolithic of Moravia. Plenum Press, New York.
Swisher Jr., C.C., Curtis, G.H., Jacob, T., Getty, A.G., Suprijo, A.,
Widiasmoro, 1994. Age of the earliest known hominids in Java,
Indonesia. Science 263, 1118–1121.
Tchernov, E., 1984. Faunal turnover and extinction rate in the Levant. In:
Martin, P.S., Klein, R.G. (Eds.), Quaternary Extinctions. The
University of Arizona Press, Tucson, AR, pp. 528–552.
Thieme, H., Veil, S., 1985. Neue Untersuchungen zum eemzeitlichen
elefanten-jagdplatz Lehringen, Ldkr. Verden. Die Kunde 36,
11–58.
Todd, N.E., Roth, V.L., 1996. Origin and radiation of the Elephanitidae.
In: Shoshani, J., Tassy, P. (Eds.), The Proboscidea: Evolution and
Paleoecology of Elephants and Their Relatives. Oxford University
Press, Oxford, pp. 193–202.
Ugan, A., Byers, D.A., in press. Geographic and temporal trends in
proboscidean and human population histories during the late
Pleistocene. Quaternary International.
Vasil’ev, S.A., 2003. Man and mammoth in Pleistocene Siberia. In:
Cavarretta, G., Gioia, P., Mussi, M., Palombo, M.R. (Eds.),
La Terra Degli Elefanti. Consiglio Nazionale delle Ricerche, Rome,
pp. 363–366.
Villa, P., 1990. Torralba and Aridos: elephant exploitation in middle
Pleistocene Spain. Journal of Human Evolution 19, 299–309.
Villa, P., Soto, E., Santonja, M., Pe
´rez-Gonza
´lez, A., Mora, R.,
Parcerisas, J., Sese
´, C., 2005. New data from Ambrona: closing the
hunting versus scavenging debate. Quaternary International 126–128,
223–250.
Waguespack, N.M., 2003. Clovis hunting and the organization of
subsistence labor. Unpublished Ph.D. Dissertation, Department of
Anthropology, University of Arizona.
Waguespack, N.M., Surovell, T.A., 2003. Clovis hunting strategies, or
how to make out on plentiful resources. American Antiquity 68,
333–352.
Warnica, J., 1966. New discoveries at the Clovis site. American Antiquity
31, 345–357.
Weber, T., 2000. The Eemian Elephas antiquus finds with artefacts from
Lehringen and Gro
¨bern: are they really killing sites? Anthropologie et
Pre
´histoire 111, 177–185.
Wenban-Smith, F.F., Allen, P., Bates, M.R., Parfitt, S.A., Preece, R.C.,
Stewart, J.R., Turner, C., Whittaker, J.E., 2006. The Clactonian
elephant butchery site at Southfleet Road, Ebbsfleet, UK. Journal of
Quaternary Science 21, 471–483.
Wojtal, P., Cyrek, K., 2001. The Upper Paleolithic mammoth site at
Halich (Ukraine). In: Cavarretta, G., Gioia, P., Mussi, M., Palombo,
M.R. (Eds.), La Terra Degli Elefanti. Consiglio Nazionale delle
Ricerche, Rome, pp. 373–375.
Zenin, V.N., Maschenko, E.N., Leshchinskiy, S.V., Pavlov, A.F.,
Grootes, P.M., Nadeau, M.-J., 2003. The first direct evidence of
mammoth hunting in Asia (Lugovskoye site, western Siberia). In:
Proceedings of 3rd International Mammoth Conference—Abstracts
Volume. Dawson City, Yukon Territory, Canada, pp. 152–155.
ARTICLE IN PRESS
T.A. Surovell, N.M. Waguespack / Quaternary International 191 (2008) 82–97 97