New fossil remains of Elephas from the southern Levant: Implications for the evolutionary history of the Asian elephant

Full text

New fossil remains of Elephas from the southern Levant: Implications for the
evolutionary history of the Asian elephant
Adrian M. Lister
a,
⁎, Wendy Dirks
b
, Amnon Assaf
c
, Michael Chazan
d
, Paul Goldberg
e,h
,
Yaakov H. Applbaum
f
, Nathalie Greenbaum
f
, Liora Kolska Horwitz
g
a
Earth Sciences Department, Natural History Museum, London SW7 5BD, UK
b
Centre for Oral Health Research, School of Dental Sciences, Newcastle University, Newcastle upon Tyne NE2 4BW, UK
c
Prehistoric Man Museum, Kibbutz Ma'ayan Baruch, Israel
d
Department of Anthropology, University of Toronto, 19 Russell St., Toronto, ONT M5S 2S2, Canada
e
Department of Archaeology, Boston University, 675 Commonwealth Ave., Boston, MA 02215, USA
f
Department of Radiology, Hadassah-Hebrew University Medical Center, Jerusalem 91230, Israel
g
National Natural History Collections, Faculty of Life Sciences, The Hebrew University, Jerusalem 91904, Israel
h
Eberhard Karls University Tübingen, The Role of Culture in Early Expansions of Humans (ROCEEH), Rümelinstr. 23, D-72070 Tübingen, Germany
abstractarticle info
Article history:
Received 11 February 2013
Received in revised form 3 May 2013
Accepted 5 May 2013
Available online 20 May 2013
Keywords:
Elephas hysudricus
Elephas maximus
Asian elephant
'Ain Soda (Jordan)
Ma'ayan Baruch (Israel)
Middle Pleistocene
We describe new fossil remains of elephant (Elephas cf. hysudricus) from archaeological sites in the Levant:
Ma'ayan Baruch (Israel) and 'Ain Soda (Jordan). Both sites date to the Middle Pleistocene based on stone ar-
tefacts typical of Levantine Late Acheulian assemblages. The elephant remains show ‘primitive’dental fea-
tures reminiscent of E. hysudricus from the Plio-Pleistocene of the Siwaliks (northern India), the species
thought to be ancestral to Asian elephant E. maximus. Regionally, the new fossils are chronologically interme-
diate between an earlier (ca. 1 Ma) record of Elephas sp. from Evron Quarry (Israel), and Holocene remains of
E. maximus from archaeological sites in NW Syria, Turkey, Iraq and Iran. It is unclear at present whether this
represents continuity of occupation or, more plausibly, independent westward expansions.
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
The ancestry of the living Asian elephant Elephas maximus L. is
poorly understood. While the generic name Elephas was formerly
applied to many different kinds of fossil elephant, only a few fossil
species are now included within Elephas sensu stricto (Maglio,
1973). Of these, the Pleistocene species E. hysudricus of the Indian
subcontinent and E. hysudrindicus of SE Asia are clearly, from their
morphology, closest to the ancestry of the living species. However,
the history of these species, their temporal and geographical extent,
and the mode of transformation of one or both of them into the mod-
ern species, are poorly known.
Elephas maximus is today restricted to the Indian subcontinent and
SE Asia. In historical times, however, its range extended eastward to
the Pacific coast of China, and westward to the Levant (Shoshani
and Eisenberg, 1982; Sukumar, 2012). Until recently, earlier fossil
evidence of Elephas s.s. in the western extremity of the distribution
was restricted to an Early Pleistocene molar from Evron Quarry (Israel),
referred to Elephas sp. by Tchernov et al. (1994).
This article describes new fossil remains from the Levant that are
referable to Elephas and are of Middle Pleistocene age: two elephant
teeth found at Ma'ayan Baruch (Israel), and three partial molars
from 'Ain Soda (Jordan). Other Elephas specimens from the region
are revised, and the place of all of this material in the evolutionary
history of the genus is assessed.
2. Materials
2.1. Ma'ayan Baruch
The Late Acheulian locality of Ma'ayan Baruch is a large, open-air
site at the northern end of the Hula Valley (Israel) (Fig. 1). The locality
comprises numerous small find spots as well as three dense concen-
trations of lithic artefacts that were exposed by ploughing in the
‘Hamara’fields of Kibbutz Ma'ayan Baruch. The artefacts lie within
and on top of a terra rossa soil. Since the 1960s, some 8000 artefacts,
predominantly handaxes, have been collected from an area of ca.
0.3 km
2
(Stekelis and Gilead, 1966; Gilead, 1977; Ronen et al., 1980;
Grosman et al., 2008). The ‘Hamara’find locality has yielded a few
bone (probably proboscidean) and tusk fragments (Stekelis and
Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130
⁎Corresponding author. Tel.: +44 207 942 5398.
E-mail address: A.Lister@nhm.ac.uk (A.M. Lister).
0031-0182/$ –see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.palaeo.2013.05.013
Contents lists available at SciVerse ScienceDirect
Palaeogeography, Palaeoclimatology, Palaeoecology
journal homepage: www.elsevier.com/locate/palaeo

Gilead, 1966:12), but it has not been possible to identify them to
genus due to their extreme fragmentation.
No excavations have been conducted at the Ma'ayan Baruch local-
ity, but in 1974 and 1977 one of the authors (AA) recovered 28 arte-
facts from the walls of a naturally formed trench ca. 0.1 m deep, that
had been created as a result of local high-velocity winter runoff. At
the base of this trench, partly embedded in the north wall, a large
elephant tooth was found lying on top of yellowish sediment. A
small fragment of a second tooth was found in the eastern part of
the same trench —the direction of the water flow (Ronen et al., 1980).
The large tooth was removed by consolidating it in a block of sediment
cradled in a fibreglass jacket (see Suppl. S1). No other osteological
remains were found in their vicinity.
The soil surrounding the large elephant molar shows micromor-
phological features typical of an oxisol (Marcelino et al., 2010), and
micromorphological analysis confirmed that the sediment in which
the tooth was embedded had been waterlogged (see Suppl. S2).
Researchers (e.g. Gilead, 1977; Bar-Yosef, 1994; Bar-Yosef and
Belmaker, 2011) have grouped Ma'ayan Baruch with other Late
Acheulian assemblages, such as Oumm Qatafa D1, that are dominated
by cordiform bifaces with few ovates, pointed bifaces and cleavers.
Analyses agree that all the Ma'ayan Baruch lithic assemblages display
Fig. 1. Map showing location of Pleistocene sites mentioned in the text (Letters) and Holocene Middle Eastern sites (Numbers) with the type of elephant remains found.
Map Site Country Material
1 Ulu Burun (Kaş) shipwreck Turkey Tusk
2 Acemhoyuk Turkey Tusk
3 Sirkeli Tepe Turkey Bone
4 Gavur Lake Swamp Turkey Bone & tooth
5 Zincirli Turkey Tooth
6 Chatal Hoyuk
(Amuq Valley)
Turkey Bone
7 Tel Tayinat Turkey Tusk
8 Tel Atchana-Alalakh Turkey Bone & tusk
9 Minet el Beida Syria Tooth
9 Ras Shamra-Ugarit Syria Bone, tooth &
tusk
10 Kamid el-Loz Lebanon Bone
11 Mishrife/Qatna Syria Bone
12 Arslantepe Turkey Bone
13 Tel Sabi Abyad Syria Bone
14 El Qitar Syria Bone & tooth
15 Munbaqa Syria Bone
16 Emar Syria Bone & tooth
17 Tel Sheikh Hamad Syria Bone
18 Chagar Bazar Syria Tusk
19 Nimrud Iraq Bone & tusk
20 Nuzi Iraq Bone
21 Babylon Iraq Bone
22 Haft Tepe Iran Bone & tusk
Note: Tusk = whole tusks or sawn but otherwise unworked sections of the tusk.
120 A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

a high degree of standardization and show careful control of symme-
try. As early as 1966, Stekelis and Gilead argued that they should be
considered the product of a single industry. Ronen et al. (1980)
reported some variation between artefacts recovered from the natu-
ral trench where the elephant teeth were found (and an assemblage
recovered from a second trench), and artefacts recovered from the
‘Hamara’field, with the trench lithics showing varying degrees of
abrasion, alteration and rolling. They suggested that the artefacts re-
covered from the lowermost part of the trenches had been swept
down the walls by the runoff water. A new study of the bifaces from
the natural ‘elephant’trench is presented in Suppl. S3, and concludes
that there are no significant signs of long-distance transport on any
artefacts in the sample currently available for analysis. Most artefacts
are in very fresh condition and two bifaces show a clear difference in
patina between the two faces suggesting that that were recovered
from an undisturbed context. Although a number of handaxes in the
trench were recovered lying in horizontal position, others were
found in vertical orientation.
Dating of the Ma'ayan Baruch find locality has relied on geological
correlations with overlying or underlying basalts and travertines. The
terra rossa soil in which the artefacts are found interdigitates with the
Kefar Yuval Travertine, as shown by the travertine coating on many of
the artefacts (Schwarcz et al., 1980). Heimann and Sass (1989) postu-
lated that this travertine began to accumulate ca. 1 Ma and continued
until ca. 25 ka. Based on pollen content, Horowitz (1979) estimated
the age of the Kefar Yuval travertine as ~0.15 Ma, an age corroborated
by radiometric measurements (Th-230/U-234) that gave an age of
189 ± 49 ka (Gur et al., 2002). Direct U-series dates on the Kefar
Yuval travertine gave ages in excess of 350 ka, but this age was
rejected as too old, the result of contamination (Schwarcz et al., 1980;
Horowitz, 2001:561). As summarised by Horowitz (2001:559–560),
the Kefar Yuval Travertine is overlain, unconformably, by the Ma'ayan
Baruch basalt dated at 73 ± 14 ka (Seidner and Horowitz, 1974) and is
underlain by the Hasbani and Dalwe basalts dated to ca. 1 Ma. Hence,
the Ma'ayan Baruch Acheulian occurrence is older than 189 ± 49 ka
BP but younger than 1 Ma. New chronological work is required to accu-
rately date this travertine.
Palaeomagnetic analysis was carried out on four soil samples taken
from the block of matrix after removal of the elephant tooth at the Paleo-
magnetic Laboratory of the Institute of Earth Sciences, The Hebrew Uni-
versity of Jerusalem, by the late Prof. Hagai Ron. The original orientation
of the block was reconstructed based on the field diaries of AA. Although
this is certainly not an ideal way to sample for palaeomagnetism, all four
samples were correctly oriented and gave similar Normal signals (Suppl.
S4). Soil samples were also taken from the inner part of the block for OSL
dating, but all proved to be unusable since the block had been exposed to
light for several years since it was removed from the trench. Based on the
geological context, the palaeomagnetic signal —with some reservation,
and most clearly the character of the associated lithic artefacts, it may
be concluded that the Ma'ayan Baruch elephant teeth are Late Acheulian
in age.
The time span covered by the Late Acheulian in the Levant is
currently under debate. We support a time range of ca. 500–220 ka
based on a compilation of currently available chronometric ages,
as published in Porat et al. (2002). Other researchers however
(e.g. Gopher et al., 2010; Bar-Yosef and Belmaker, 2011), constrain
the Late Acheulian to ca. 600–400/350 ka, with the Acheulo–
Yabrudian industries as a later and separate phase spanning the
period 400/350–250/220 ka.
2.2. 'Ain Soda
'Ain Soda is an open-air site located in the wetlands of the Azraq
Basin in eastern Jordan. Excavations, co-directed by Gary Rollefson,
Philip Wilke and Leslie Quintero, were initiated in 1977 as an archae-
ological field school for students from San Juan College (Farmington,
New Mexico, USA). 'Ain Soda (Rollefson et al., 1997a, 1997b, 2006;
Quintero et al., 2004), like the neighbouring sites of 'Ain el-Assad
(Rollefson, 1983), C-Spring (Copeland and Hours, 1989) and those
in the Al-Jafr basin (Rech et al., 2007), all reflect the distribution
and movement of Middle Pleistocene hominins in the desert interior
of Eastern Jordan.
The 'Ain Soda site lies on the edge of a large pool, originally creat-
ed by a spring fed by an underground aquifer, bringing water from as
far away as Jebel Druze in southern Syria and Zarqa in western Jordan.
In the Pleistocene, the site lay along the shore of what was once a
large lake. Four trenches (1–4) were excavated along the northern,
southern and western edges of the 'Ain Soda pool. Altogether some
60 m
2
were sampled (Rollefson et al., 1997a, 1997b). The site yielded
evidence of in situ Epipalaeolithic/Late Upper Palaeolithic, Early
Mousterian and Late Acheulian occupations. In places the sediments
were waterlogged due to the high water table (Quintero et al., 2004).
Analysis of the artefacts from the Late Acheulian layers revealed
an extremely high proportion (>90%) of bifacial tranchet cleavers
(Quintero et al., 2004:3, 2005; Wilke et al., 2005; Rollefson et al.,
2006), similar to that found in Late Acheulian localities in the Al-Jafr
basin (Rollefson et al., 2006). However, compared to sites in the
Mediterranean region, such as the Late Acheulian of Tabun Cave
LayerE,'AinSodahasarestrictedrangeofartefacts,ahighratio
(3:2) of flake tools to bifaces, and a slightly higher frequency of bi-
faces which are also larger and narrower (relative to length), with
extensive use of Levallois technology.
The 'Ain Soda locality has been identified by the excavators as a
butchering site (Rollefson et al., 2006). The faunal preservation was
good in both the Late/Final Acheulian and Mousterian deposits
(Quintero et al., 2004), despite the fact that the site was an open-air
locality and the silt dunes at the edges of the pool contain salts which
are detrimental to bone preservation. Two trenches yielded Mousterian
artefacts associated with aurochs (Bos primigenius) and equid remains.
Two other trenches produced Late Acheulian artefacts together with
faunal remains including those of rhinoceros (Stephanorhinus cf.
hemitoechus), Equus hydruntinus, and an extinct elephant identified as
Elephas cf. hysudricus (Dirks et al., 1998; Rollefson et al., 2006;theele-
phant remains were incorrectly listed in Rollefson et al., 1997a as
Elephas planifrons). The three elephant teeth were recovered from the
south trench (termed the “Elephant Trench”), and were found nearly
at water level.
Given the uniqueness of the elephant remains, the teeth were
taken to the Mammoth Site of Hot Springs, South Dakota, USA and
three sets of polyurethane casts made. Subsequently, the original
teeth were lost in transit. Fortunately the casts survive and these
were used in the current study in conjunction with notes and photo-
graphs made during direct observations of the teeth by WD.
No radiometric dates are available for the site of 'Ain Soda, so
dating of the elephant teeth is based on the characteristics of the
associated lithic artefacts, which indicate a Late Acheulian age. As
such, it falls within the same general age range as the Ma'ayan Baruch
elephant remains.
3. Analytical methods
3.1. Measurements
Measurement of elephant molars follows Maglio (1973),modified by
Lister (2012: 207).Talons(‘x’)andplatelets(‘p’) are not included in
lamellar counts. Lamellae (‘plates’) are numbered ‘l1’,‘l2’, etc., counting
from the natural anterior end of the tooth, or ‘L1’,‘L2’etc., counting
from the posterior end. Comparative data of modern Elephas maximus
teeth is from original measurements of material at The Natural History
Museum and other UK collections (see Acknowledgements). Roth and
Shoshani (1988) also provide useful comparative data, but their lamellar
counts have not been used as they included the vestigial structures here
121A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

termed talons and platelets, so the relationship of their scores to ours,
which exclude these structures, is unclear.
3.2. CT scanning
Even after cleaning and consolidation, the occlusal surface of
the large molar from Ma'ayan Baruch was poorly preserved. To
facilitate examination and measurement, a CT scan was made of the
tooth following Roth (1989). We used a Philips Brilliance 64ME
(dual energy) CT scanner (Radiology Department, Hadassah Hospital,
Jerusalem) to obtain contiguous CT slices in a coronal (transverse)
planes passing from root to occlusal surface, and parasagittal (vertical)
planes parallel to the vertical plane of the molar (for complete set of
scans, see Suppl. S5). Scans were made at two settings: (i) 140 kV,
0.67 mm × 0.33 mm at 550 mas and (ii) 140 kV, 0.9 mm × 0.45 mm
at 250 mas.
Abbreviations: FAD: First Appearance Datum; LAD, Last Appearance
Datum; NHM, Natural History Museum, London; UMZC, University
Museum of Zoology, Cambridge; P, plate (lamellar) number; LF, lamel-
lar frequency; H, crown height, HI, hypsodonty index; W, crown width;
L, crown length; e, enamel thickness; M
3
,upperthirdmolar;M
3
,lower
third molar; CT, computed tomography.
4. Morphological and metric study of Maya'an Baruch specimens
4.1. The larger specimen (catalogue MB1)
The specimen MB1 (Fig. 2) is a very large elephantid molar from
the left side, almost certainly an upper and very probably, from its
size (Fig. 3), an M
3
. The preserved specimen represents the posterior
9 or 10 lamellae, the anterior part of the molar being partly lost. The
interpretation of the specimen is challenging because of damage
and likely distortion, but has been greatly aided by the CT scans
(Fig. 4; Suppl. S5).
The identification of the tooth as an upper is based, firstly, on the
angle of the occlusal surface to the vertical plane of the lamellae. It is
not perpendicular to the lamellae as in a lower, but forms an obtuse
angle to them. Posterior to the occlusal region, the surface of the spec-
imen tilts rootward for the last 3–4 lamellae (L3 to p in Fig. 2a),
forming a ‘tented’shape with the occlusal surface —again like an
upper. This interpretation depends, however, on the exposed surface
being natural and undistorted. Most parts of the exposed enamel
appear naturally rounded through wear, suggesting that this is
approximately the natural occlusal surface. The lamella labelled L1
in Fig. 2a shows an unworn and only slightly broken peak of a
‘digit’, positioned rootwards of the occlusal area, again consistent
with the slope of the unworn posterior part of an upper. However,
the two lamellae immediately behind the proposed occlusal surface
(L2–L3) appear to have naturally worn enamel too; this is difficult
to interpret since their apices are rootward of the main worn (occlu-
sal) area. Conceivably they were part of the occlusal surface but have
been moved rootward through crushing or slippage.
In medial view, the lamellae converge from bottom to top of the
crown, generally considered characteristic of lower rather than
upper molars. The convergence is seen in the vertical CT sections
(Fig. 4a,b) and is considered genuine, a separate phenomenon from
the distorted orientation of some lamellae discussed below. However,
the lamellae are quite straight and not S-shaped as in M
3
. Moreover a
similar degree of overall convergence can be seen in some M
3
—e.g. in
E. maximus UMZC H.4692 (Fig. 5).
Fig. 2. The large molar (left M
3
) MB1, Elephas cf. hysudricus, from Ma'ayan Baruch, in
medial (above) and occlusal (below) views. Anterior is to the left.
120
140
160
180
200
220
240
Crown height (H), mm
A
100
65 70 75 80 85 90 95 100
Crown width (W), mm
3
4
5
6
7
8
9
10
60 70 80 90 100 110 120
Lamellar Frequency (LF)
Crown width (W), mm
B
Fig. 3. Dimensions of the upper molar teeth from Ma'ayan Baruch, 'Ain Soda, Siwalik
Elephas hysudricus, and modern E. maximus. In both graphs, molar width (horizontal
axis) is an index of the tooth size. A: crown height (H) against molar width. B: Lamellar
Frequency (LF) against molar width. Blue diamonds: E. maximus M
3
; green triangles: E.
maximus M
2
; red squares: 'Ain Soda M
2
or M
3
; black circle: Ma'ayan Baruch M
3
; blue
dashed box: M
3
range of E. hysudricus from Maglio (1973); green dashed box: M
2
range of E. hysudricus from Maglio (1973). In B, logarithmic regression lines have
been fitted to the E. maximus M
3
(blue) and M
2
(green).
122 A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

The posterior medio-lateral narrowing of the crown suggests we
are dealing with the last molar (M
3
). The medial sides of the posterior
lamellae are complete, and are positioned increasingly toward the
tooth mid-line as we move posteriorly across the last three plates.
The lateral sides are too broken to tell. In transverse scans (Fig. 4c,
d), L3 is much narrower than L4–6; its medial and lateral ends both
seem to be complete, and both are closer to the midline than in L4–
6. Moreover, the transverse CT scans suggest that the successive dis-
placement of the medial edges is not due to slippage of the lamellae
in the medio-lateral plane, because homologous points (idiosyncrasies
of enamel rings and wiggles) seem to line up antero-posteriorly
between successive lamellae.
The narrowing of the crown across its posterior three lamellae
thus suggests an M
3
. However, anterior of L4, there seems to be no
further increase in lamellar width (Table 1); this is unexpected as
the widest part of an M
3
is normally more anteriorly placed, with a
longer zone of narrowing toward the posterior end. However, the
morphology of the Ma'ayan Baruch tooth does not fit a typical M
2
ei-
ther. In an M
2
, the last true lamella and talon can be narrowed, but
not as much as in the Maya'an Baruch tooth. Another diagnostic dif-
ference between M
3
and M
2
is the root, which tapers posteriorly in
M
3
but is widest at the very back of the tooth in M
2
. Unfortunately,
the CT scans show that there is hardly any root left in the Maya'an
Baruch tooth, so this feature cannot be determined. Overall, the sig-
nificant narrowing across the preserved posterior three lamellae bet-
ter fits an M
3
.
The narrowness of the most posterior preserved lamella, seen on
the CT scans (Fig. 4c,d), suggests that it is probably very close to the
natural back end of the tooth. Additionally, the crown base of the pos-
terior three lamellae rises in the direction of the crown apex. This is
visible on the medial side of the tooth though not on the lateral side
where the lamellae are very crushed. Such a trend in the crown
base is common in elephant molars, and confirms that the last lamel-
lae in the Maya'an Baruch specimen are close to, or at, the natural
posterior end of the tooth. The vertical CT scan, Fig. 4b, suggests
that the posterior ‘platelet’(a reduced lamella analogous to the
Fig. 4. Selected CT scans of the Ma'ayan Baruch upper molar MB1. A–B, vertical scans; C–D, transverse scans. A: scan 95, close to midline of the tooth; B: scan 120, parasagittal scan
between midline and medial edges of the tooth; C: scan 163, close to top of the crown; D: scan 133, about half-way down crown.
Fig. 5. Left M
3
of E. maximus, UMZC H.4692 in medial view, showing features similar to
the large Ma'ayan Baruch molar; specifically: lamellae strongly convergent from base
to apex, and posterior end lacking a long taper. Although unusual, these features in a
demonstrable M
3
corroborate the attribution of the Ma'ayan Baruch tooth to an
upper third molar. Because of the distortion this specimen has not been plotted in
Fig. 3b.
Table 1
Measurements of Ma'ayan Baruch molar MB1. p = platelet, L1, L2, L3 etc. = position
of lamella starting from posterior end. P, lamellar (plate) formula; L, crown length;
LF, lamellar frequency across L3-L5; W, crown width (no cement); H, crown height.
p L1L2L3L4 L5 L6
P−10p
L>273
LF 4.76
W?42 ?65 ?75 ~ 90 ~ 103 ~103 ~105
H75 108 140 150 146 Worn Worn
123A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

talon of more anterior teeth; Lister and van Essen, 2003) is preserved
behind L1, so the tooth is essentially complete posteriorly.
Lamellae immediately in front of L1 (seen especially in the vertical
CT sections, Fig. 4a) are quite high, suggesting that the posterior
tapering of crown height was not pronounced in this specimen.
Although posterior tapering is typical of third molars in elephantids,
lesser tapering similar to the Maya'an Baruch example can be found
in some M
3
and even M
3
of E. maximus (e.g. UMZC H.4692: Fig. 5),
with very rapid lowering of crown height only in the posterior-most
one or two lamellae. The preserved unworn ‘digit’at the apex of the
penultimate lamella (L1), mentioned above, if more or less in its
correct place (as suggested by vertical CT scan, Fig. 4b), confirms a
lowering of crown height, since its apex is rootward of those of
more anterior lamellae. However, the restriction of tapering to the
last few lamellae means this feature cannot be used to distinguish
M
2
from M
3
in this specimen.
The strongest evidence for post-mortem distortion of this tooth is
the orientation of the lamellae seen in medial and lateral views. The
wide spacing and backward sloping of the lamellae in the middle
region of the crown gives an unnatural appearance. First, the sloping
of the lamellae becomes less marked toward the back of the crown,
rather than more marked as would be the case with the natural slop-
ing of an M
3
. Second, the odd backward slope of the plates from L5
forward is not seen in the vertical CT scan (compare medial view
Fig. 2a with midline scan Fig. 4a), suggesting that it is in large part
due to crushing or movement of the lateral sides of the tooth.
Overall, the form of the tooth suggests that it may have been partly
liquified post-depositionally, and re-set in a slightly distorted shape.
Spaces between the enamel ridges, that should have been filled with
dentine or cement, are either hollow or appear to be filled with consol-
idated sediment. The tooth may have been partly dissolved –starting
with the softer cement and dentine –so the remaining enamel plates
were somewhat mobile –accounting for their odd orientation plus
some possible up–down displacement. This interpretation is consistent
with evidence for long-term waterlogging of the sediment (see above
and Suppl. S2).
Measurements of the molar are given in Table 1, but because of
the damage and distortion of the crown, some discussion of the valid-
ity of the measurements is necessary.
4.1.1. Lamellar number and crown length
Seven lamellae are clearly demarcated on the medial side of the
molar, but the vertical CT scan shows additional lamellae in the dam-
aged anterior portion, with a total of 8 lamellae clearly visible and the
remains of probably two more, progressively dropped rootwards, in
the crushed anterior end (Fig. 4b). These two lamellar remnants can
just be seen on the medial side of the tooth itself. The total of 10 or
so preserved lamellae is, however, unlikely to represent the original
count, lamellae having been lost anteriorly through natural wear in
life or breakage post-mortem. There are no preserved roots allowing
us to assess position in relation to the original anterior end (cf. Sher
and Garutt, 1987). However, in the vertical CT scans (Fig. 4b), even
the anterior-most preserved plate has a significant height of crown
remaining, suggesting that the loss of more anterior lamellae is due
to breakage, not wear in life. The preserved length of the crown,
273 mm, is therefore also less than the original. Among Siwalik
E. hysudricus and modern E. maximus, maximal recorded lengths for
M
3
are 300–340 mm, for a width of 90–100 mm (Maglio, 1973;
Roth and Shoshani, 1988; Suppl. S6). With an estimated original
width of 115–120 mm, the Maya'an Baruch molar, if of similar
length/width proportion, could have been as much as 380 mm long
when complete. Assuming a lamellar frequency of 4.76 as measured
on the least-distorted region of the preserved molar (Table 1), origi-
nal plate count can be estimated at ca. 18. The significance of this
value is discussed below.
4.1.2. Crown width
The maximum preserved width of the crown, 105 mm, was mea-
sured on L5 and L6. This is considered to be a true measurement be-
cause the medial ends of the lamellae appear undamaged, and while
the lateral ends are damaged, the transverse CT scan shows three
subequal enamel loops, with the centre of the middle loop at the pre-
served midline of the tooth, suggesting that little or nothing has been
lost at the lateral side. No external cement is preserved and a molar of
this size typically would have had 10–15 mm of cement (medial and
lateral sides combined), indicating an original width of 115–120 mm.
A value of 115 mm has been plotted on the graph (Fig. 3b).
4.1.3. Crown height
The maximum preserved crown height, 150 mm measured just
behind the occlusal surface on L3, is certainly not the original maxi-
mum of the tooth. Sher and Garutt (1987) showed that in elephantid
M
3
s, there is a ‘zone of maximum (unworn) crown height’in the
central region of the tooth, with lower heights anterior and, espe-
cially, posterior to this, due to the posterior taper of the tooth. The
maximum height of the Maya'an Baruch tooth would have been
in the region that has been naturally worn, and is therefore not
measurable.
4.1.4. Lamellar frequency
Because of the severe damage to the anterior part of the preserved
tooth, and the distortion to the middle portion discussed above, the
region of lamellae L3–L5 is considered to give the most accurate esti-
mate of original lamellar frequency. Lamellar frequency measured on
all seven plates clearly visible on the medial side (i.e. including those
considered to be distorted) gives a value of 4.54. The three lamellae
L3–L5 give 4.76, and this is taken as the best estimate.
4.1.5. Enamel thickness
This can be measured at a few points on the specimen, and also on
the vertical and transverse CT scans, and gives values in the region of
2.8–3.0 mm.
4.1.6. Individual age
Roth and Shoshani (1988: Fig. 7) presented a scheme of dental
eruption versus age in known-age Asian elephants. In the Maya'an
Baruch M
3
, all but the posterior four lamellae are in wear, and the
M
2
would have been naturally lost. Assuming an original lamellar
count of 18 (see above), 14/18 or 78% of the lamellae were in wear,
translating to an age of around 50 years in Roth and Shoshani's
(1988) scheme.
4.2. The smaller specimen (MB2)
This specimen is a crushed part of an elephantid molar crown
(Fig. 6). It comprises six lamellae which are uncemented and no lon-
ger in their orderly alignment. The lamellae are apparently from the
middle part of a molar, but their identity as upper or lower, left or
right, is uncertain. Their apices are unaffected by wear in life, so the
specimen represents a molar all or part of which was unerupted
at death. The large size of the tooth –the widest lamella measures
96 mm without cement –indicates an upper or lower M
2
or M
3
.
Enamel thickness is ca. 3.5 mm. If the larger, worn molar from
Maya'an Baruch is an M
3
as suggested above, the smaller, unworn
specimen cannot be from the same individual. If the larger specimen
were an M
2
, the smaller tooth could be part of the unfinished M
3
or
M
3
of the same individual. The lamellae appear to be either hollow
or sediment-filled like the larger molar, suggesting a similar history
of erosion or dissolution.
124 A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

4.3. Generic identification
Possible candidate taxa for the Maya'an Baruch teeth are Elephas,
Palaeoloxodon,Mammuthus, and conceivably Loxodonta (although
the latter genus has never been definitively identified outside Africa).
Characters for the separation of these genera are given in Maglio
(1973),Albayrak and Lister (2012) and elsewhere. The lack of a
loxodont lamellar sinus rules out Loxodonta, while the hypsodont
crowns rule out Mammuthus rumanus or M. meridionalis.All
Mammuthus, moreover, have relatively unfolded enamel.
Several features, however, point strongly to identification as
Elephas. The pattern of wear of the lamellae, seen in the transverse
CT scans of the large tooth (Fig. 4c,d), progresses from a row of
small rings at the apex, to three subequal rings, which then fuse to
form the lamella as the tooth wears, a typical configuration for
Elephas. The transverse CT scans (Fig. 4c,d) also show the enamel
band strongly folded into a series of tight loops along the lamella,
again typical for Elephas. The smaller specimen also shows a row of
small, equal digits at the apex, that would have worn into small
rings, while the unfinished base of each lamella, seen in ‘root’view
(although there are no roots), shows strong, tight folding of the
enamel band. Neither tooth shows any sign of the typical features of
Palaeoloxodon, where the enamel folds are concentrated into a
major fold in the midline of the tooth, flanked by smaller folds on ei-
ther side of it, and additional minor (and not so strongly plicated)
folds. The early wear pattern in Palaeoloxodon also typically shows a
long central enamel loop flanked by two subcircular loops at the lat-
eral and medial sides, not seen in the Maya'an Baruch molars.
Palaeoloxodon is also characterised by relatively narrow crowns, un-
like the markedly wide crowns of the Maya'an Baruch specimens.
5. Morphological and metric study of 'Ain Soda specimens
The sample comprises three partial molars (Fig. 7), incomplete but
well-preserved and undistorted. Measurements are given in Table 2.
5.1. Upper molar M92449
This is the anterior part of a right upper molar, from its size M
2
or
M
3
. Its width, greater than the range of modern E. maximus M
2
(Fig. 3), and its distinct curvature from front to back (concave medi-
ally, convex laterally), makes M
3
more likely, but given the very
large size of the Maya'an Baruch M
3
(Fig. 3b), M
2
cannot be excluded.
Only the first three plates are worn, so the unworn plates behind give
the true maximum crown height of the molar. Similarly the measured
crown width is in the maximal region. If M
3
, the individual was in its
early 30s at death; if M
2
, in its mid-teens.
5.2. Upper molar M92450
This specimen represents the anterior part of a left upper molar,
from its size M
2
or M
3
. Its width, greater than the range of modern
E. maximus M
2
(Fig. 3), and the beginning of distinct curvature from
front to back, makes M
3
more likely, but as above, M
2
cannot be exclud-
ed. Only the first three plates are worn, so the unworn plates behind
give the true maximum crown height of the molar. Similarly the mea-
sured crown width is in the maximal region. If M
3
, the individual was
in its early 30s at death; if M
2
, in its mid-teens.
5.3. Lower molar M92451
This is a small segment of a left lower molar, of uncertain position
in the crown. From its width and evident curvature (concave lateral-
ly) it is probably M
3
, but M
2
cannot be excluded. The occlusal surface
of the piece is in very early wear.
The two upper molars M92449 and M92450, from their respective
crown widths and wear stages, cannot be from the same individual. It
is not excluded, however, that the lower molar M92451 could be from
the same individual as one of the uppers.
5.4. Generic identification
Specimens M92449 and M92450 show subequal enamel rings in
early wear, with no sign of the Palaeoloxodon features described
above. In M92449 the second plate shows fusion of the rings into a lon-
ger lateral and shorter medial loop, commonly seen in early to midwear
of E. maximus (Albayrak and Lister, 2012). This specimen further shows
the beginnings of strong enamel folding, and in all three specimens the
ends of the enamel loops, seen in medial and lateral views of the teeth,
are longitudinally grooved, an expression of strongenamel folding com-
monly seen in Elephas and Palaeoloxodon, but not Mammuthus or
Loxodonta. The 'Ain Soda sample is therefore referred to Elephas.
6. Morphometric comparison and specific identification
of all material
The most important variables for specific identification within the
Elephas hysudricus–E. maximus lineage are lamellar number and
crown height, both of which increase between the two species. The
roughly estimated original lamellar number of 18 for the Maya'an
Baruch M
3
compares to observed ranges of 12–17 in E. hysudricus
(Maglio, 1973, n = 10) and 21–26 in E. maximus (n = 8; Suppl. 6),
uppers and lowers pooled. If anything, the figure of 18 may be an
overestimate because the estimated crown length on which it was
based is greater than for any Elephas molar ever recorded. Overall,
Fig. 6. The small elephant molar from Ma'ayan Baruch, MB2. A: lateral view, B: apical view.
125A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

therefore, lamellar number appears more consonant with E. hysudricus
than E. maximus.
All of the 'Ain Soda teeth are too incomplete to allow estimation of
original lamellar number. In this situation, lamellar frequency (LF),
the number of lamellae in a 10 cm length of tooth, can be used as a
proxy, provided the size-related nature of lamellar frequency is
taken into account (Lister and Joysey, 1992). LF is inversely related
to molar size, here represented by crown width (W). In Fig. 3b, LF is
therefore regressed against width (W), with a logarithmic fit applied
to the modern comparative sample, extrapolated to encompass the
large widths of the fossil specimens. Modern M
3
and M
2
are both
plotted, because of the uncertain positional identity of the 'Ain Soda
specimens. The graph shows both the two 'Ain Soda M
2/3
plotting
clearly below the trends of the modern sample, especially that of
M
3
, indicating lower LF even taking their large size into account,
and implying an originally lower lamellar number (P) than in modern
E. maximus. The Maya'an Baruch M
3
similarly falls below the extrapo-
lated modern LF/W trend, corresponding to its lower estimated la-
mellar number than the living species.
Hypsodonty cannot be estimated for the Maya'an Baruch tooth,
but in the two 'Ain Soda M
3
s it clearly falls below the modern sample
in a plot of crown height (H) versus width (W) (Fig. 3a). This corre-
sponds to hypsodonty indices (H/W) of 1.35 and 1.44 (Table 2), com-
pared to a range of 1.98–2.71 in E. maximus (n = 12; Suppl. S6).
The LF and H values of the 'Ain Soda specimens, lower than mod-
ern E. maximus, fall, however, within the range of E. hysudricus M
3
from the Siwaliks (Fig. 3b; data from Maglio, 1973). The LF of the
Maya'an Baruch M
3
also falls within this range, with a value some-
what below those of 'Ain Soda which can be accounted for by its
very large size and the inverse relation of LF to tooth size (Lister
and Joysey, 1992). All of the studied molars, therefore, are at an evo-
lutionary level comparable to Siwalik Plio-Pleistocene E. hysudricus.
Because of the limited nature of the material, and the lack of charac-
ters outside the dentition, they are referred to Elephas cf. hysudricus.
7. Discussion and conclusions
7.1. Significance of the Levantine Pleistocene fossils in the evolution of the
Asian elephant
The most numerous fossil remains from the lineage of the Asian
elephant have come from the Indian subcontinent, especially the
Siwalik Group of India and Pakistan. Of the two Elephas species
known from the Siwaliks, E. planifrons Falconer & Cautley, 1845 and
E. hysudricus Falconer & Cautley, 1845, the latter is clearly, from
its cranial and dental morphology, closest to the ancestry of the
living species E. maximus L. (Maglio, 1973). A related fossil species,
E. hysudrindicus Dubois, 1908, is known from insular SE Asia. Finally,
the more distantly related Palaeoloxodon namadicus (Falconer &
Cautley,1846), originally and still sometimes named Elephasnamadicus,
also occurs in the Pleistocene of the Indian subcontinent.
In order to assess the position and significance of the new Levan-
tine finds in the context of Asian elephant evolution, existing evi-
dence on the LAD of E. hysudricus and FAD of E. maximus will be
reviewed (see also Vidya et al., 2009: supplementary information
10). This evidence is complicated by both stratigraphic and dating
uncertainties, and issues of the correct identification of remains.
In the Siwalik sequence, E. hysudricus is characteristic of the Pinjor
Formation; the appearance of the species, and the base of the Forma-
tion, are currently placed at 2.6–2.7 Ma (Nanda, 2002). The upper
limit of the Pinjor is time-transgressive but its youngest date, and
Table 2
Measurements of 'Ain Soda molars.
Abbreviations as in Table 1, plus: HI = hypsodonty index. Width measurements (W)
exclude cement, but are plotted (Fig. 3) with an allowance of 5 mm for missing ce-
ment. LF of the lower molar measured at base (medial and lateral averaged) only
(see Lister, 2012).
Specimen Position P L LF W H HI
M92449 RM
2/3
x8–>164 5.82 89 135 1.44
M92450 LM
2/3
x6–>120 5.35 92 131 1.35
M92451 Lm
2/3
−4–>95 4.73 ≥87 ≥110 –
Fig. 7. Elephant molars, Elephas cf. hysudricus, from 'Ain Soda. A: right upper molar M92449 in occlusal and lateral views; B: left upper molar M92450 in occlusal and lateral views;
C: left lower molar M92451 in occlusal and medial views.
126 A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

that of E. hysudricus in the Siwaliks, is placed at 0.6 Ma (Nanda, 2002,
2008).
Based on available published information, Chauhan (2008) and
Nanda (2008) note the persistence of E. hysudricus into the post-
Siwalik (Middle to Late Pleistocene) Narmada (= Narbada) and Go-
davari beds of Peninsular India, although Nanda (2008, p. 9) cautions
that “the post-Siwalik faunal lists provided by various workers are not
always supported by descriptions, line diagrams or photographs”.
According to Chauhan (2008), the Lower Narmada group (Middle
Pleistocene, i.e. ca. 780–125 ka) yielded both E. hysudricus and
E. namadicus.Cameron et al. (2004) indicate a date of >236 ka for
Narmada fauna associated with Late Acheulian industry. Chauhan
(2008, p. 27) further indicates that the Upper Narmada Group
(early Late Pleistocene) ‘is thought to yield’both E. hysudricus and
P. namadicus, while Nanda (2008) lists E. hysudricus in the Late
Pleistocene (ca. 125–12 ka) of Peninsular India and the Indo-Gangetic
region. Other authors, however, such as Deraniyagala (1955),Khatry
(1966),Maglio (1973) and Badam (1979) list only E. namadicus for
Narmada and Godavari, implying that E. hysudricus was extinct,
at least regionally, by the end of Pinjor times (i.e. 0.6 Ma).
The earliest date for E. maximus is also problematic. Badam (1988,
cited by Chauhan, 2008) is of the opinion that the species is not found
in the older part of the Narmada sediments (Middle Pleistocene), but
is present only in the Late Pleistocene of the Narmada and other Dec-
can fluvial systems. Nanda (2008), similarly, lists the species as pres-
ent in the Late Pleistocene of Peninsular India and the Indo-Gangetic
region.
On the available evidence it must be admitted that the date and
mode of the transition from E. hysudricus to E. maximus within the
Middle to Late Pleistocene is uncertain (Dennell, 2004; Vidya et al.,
2009: supplementary information 10; Sukumar, 2012). If the identifi-
cation of both E. hysudricus and E. maximusin t he Late Pleistocene is cor-
rect, this could reflect either (i) strict chronological co-occurrence,
implying earlier speciation of E. maximus elsewhere, followed by dis-
persal into the Indian range of E. hysudricus (cf. Eurasian Mammuthus:
Lister et al., 2005); or (ii) that the remains are of different ages within
the Late Pleistocene and that E. maximus chronologically replaced
E. hysudricus. However, the persistence of E. hysudricus beyond the
Middle Pleistocene still remains to be rigorously demonstrated.
The likely age of Maya'an Baruch and 'Ain Soda (in the range ca.
500–220 ka), and the correspondence of their dental morphology to
E. hysudricus, make them potentially the youngest dated remains at-
tributable either to that species, or at least to a transitional form not
yet at the level of E. maximus.IfE. maximus evolved anagenetically
from E. hysudricus, this suggests 500 ka as a maximal age for the tran-
sition. However, it cannot be excluded that E. maximus arose earlier in
another part of the E. hysudricus range (presumably further east), so
that the Levantine population evidenced at Maya'an Baruch and 'Ain
Soda represents a relict, more primitive population. At present,
since we lack dated early remains of E. maximus from the Indian sub-
continent or SE Asia, it is impossible to choose between these options.
Vidya et al. (2009) showed that the two major mitochondrial DNA
clades within modern E. maximus originated 1.6–2.1 Myr ago, the mi-
tochondrial split therefore probably occurring within E. hysudricus
and plausibly in isolated populations during a period of refugial con-
traction in the regions of Myanmar and India/Sri Lanka. They further
suggest that the current complex pattern of distribution of the
mtDNA clades in E. maximus resulted from a series of contractions
and expansions during the climatic oscillations of the Quaternary, im-
plying a complex distributional history for the living species. An alter-
native, intriguing possibility is that the two palaeo-species recognised
on morphology, E. hysudricus from mainland south and southeast
Asia, and E. hysudrindicus from insular southeast Asia, could be the or-
igin of the two clades in modern E. maximus, so that the latter repre-
sents a ‘hybrid’of the two forms. Vidya et al. (2009) consider this
less likely because of the Late Pleistocene age assigned to fossils of
E. hysudrindicus. However, as summarised here, the age of this mate-
rial, and the date of origin of the species, are poorly-constrained, so its
contribution to the modern species remains a possibility.
The earliest known Elephas in the Levant comprises an upper
molar from Evron Quarry (Fig. 8) reported by Lister in Tchernov et
al. (1994) and dated to between ca. 1.0 and 0.78 Ma (Ron et al.,
2003). Its size (crown width 73 mm including cement) is too large
for dP
4
and makes identification as M
1
highly probable. In this case
its morphology is closer to E. hysudricus than E. maximus (Table 3),
with some caution because of damage to the specimen (Tchernov et
al., 1994). The plate count of 10 assumes the preserved anterior-
most root is the true anterior root. Crown height cannot be measured
at its highest point (at the posterior end in an M1), but the value
of 87 mm on plate 6 suggests that a true maximum in the range
of 108–111 (E. hysudricus) is much more likely than 127–142
(E. maximus) (see Table 3). The LF value for the Evron tooth is valid
and unaffected by the breakage, and corroborates a morphology
more primitive than the living species.
Other records of ‘Elephas’in the Levant are not supported by
detailed morphological study. Pliocene remains from Bethlehem
formerly referred to E. planifrons (Hooijer, 1958) are now considered
probably to pertain to Mammuthus (Markov, 2012). Bate (1937:222)
listed a tusk fragment from Tabun Layer E (Israel) as ‘Elephas sp.’,
but the identity of this and other tusk fragments from Tabun in the
NHM collection cannot be determined beyond Elephantidae indet. A
rolled, partly mineralized posterior crown fragment of an elephantid
upper molar (cf. right M
1
or M
2
) from the Late Acheulian locality of
Oumm Zinat (Israel), listed as Elephas by Horwitz and Tchernov
Fig. 8. Elephant molar, cf. Elephas hysudricus, from Evron Quarry. A: occlusal view, B:
lateral view.
127A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

(1989), is of uncertain identity but shows features suggestive of
Palaeoloxodon sp.
7.2. Ecology of Levantine Pleistocene proboscideans
The ecology of Elephas hysudricus is essentially unknown, al-
though its dental morphology –with moderate development of
hypsodonty and lamellar number –suggests a mixed-feeder taking
both browse and graze.
In the southern Levant, there is overlap in the chronological range
of several proboscidean taxa in the late Early to Middle Pleistocene.
E. cf. hysudricus at Evron Quarry (1.0–0.78 Ma) co-occurs with
Stegodon (Tchernov et al., 1994), while Gesher Benot Ya'akov (Israel),
dated to slightly younger than 0.78 ka (Goren-Inbar et al., 2000),
has yielded both Stegodon and the earliest occurrence in the region
of the straight-tusked elephant Palaeoloxodon (Tchernov and Shoshani,
1996). Eurasian Palaeoloxodon is a migrant from Africa, a derivative of
P. recki (formerly Elephas recki). The Gesher Benot Ya'akov elephantid
skull was identified as the European species P. antiquus (Goren-Inbar
et al., 1994; Shoshani et al., 2001), but has been considered a
possible P. recki by Saegusa and Gilbert (2008).Possiblecontemporane-
ity of Mammuthus during this interval also cannot be excluded
(M. meridionalis at 'Ubeidiyeh, Israel; M. trogontherii at Latamne, Syria:
Lister, 2004). Lister (2004) discussed the ecology of Palaeoloxodon in
Eurasia and its niche separation from the Mammuthus lineage, the latter
moving from browser/mixed feeder M. meridionalis toward a more
grazing adaptation (M. trogontherii) after the entry of the browsing/
mixed-feeding Palaeoloxodon. However, on current evidence it is diffi-
cult to ascertain if any of the three genera were precisely contempora-
neous (and hence competed) at any time in the Levant.
The next known occurrence of Palaeoloxodon in the region
(as P. antiquus) is in the Late Acheulian beginning ca. 500 ka. Records
include Holon (Davies and Lister, 2007), Revadim (Rabinovich et al.,
2012), Oumm Zinat (Horwitz and Tchernov, 1989; see above), and
most recently the Zuq Fawqani find locality near to Ma'ayan Baruch
(see Suppl. S7). Although the Late Acheulian Palaeoloxodon remains
span the same general time period as the Elephas finds from Ma'ayan
Baruch and 'Ain Soda, their remains are not found in the same sites,
so it is unclear if they coexisted and presumably exploited different
habitats or resources (niche separation). Alternatively, they may
have been separated chronologically, a possibility given the lengthy
time span of the Late Acheulian and the lack of refinement in dating
the localities.
7.3. Elephas in the Holocene of the Levant
It is clear from archaeological and documentary evidence that in
relatively recent times the distribution of Elephas maximus extended
much further west than it does today. However, the western limit of
the distribution has been unclear, with authors varyingly placing it
in Iraq, NE Syria or SE Turkey (e.g. Hofman, 1974; Shoshani and
Eisenberg, 1982; Sukumar, 1989, 2012; Santiapilli and Jackson,
1990). Faunal remains of elephants from Holocene archaeological
sites in Southwest Asia have been summarised by several researchers
(Miller, 1986; Caubet and Poplin, 1987, 2010; von den Driesch, 1996;
Albayrak and Lister, 2012), and by Becker (2005, 2008) who has also
reviewed the ancient written and iconographic records dealing with
elephants. Fig. 1 presents an updated map of these finds.
Some researchers have supported the idea that a living E. maximus
population inhabited the Euphrates–Tigris River region in the late Ho-
locene (e.g. Miller, 1986; Becker, 2005, 2008). The disappearance of
this population has been linked to human predation and hunting,
changing climate (aridification) and/or diminishing natural habitats,
especially deforestation (Miller, 1986; Sukumar, 2012). The last ap-
pearance has been dated to the early 1st millennium BC, this being
the latest written reference to the hunting of live elephants in north
Syria by the Assyrian king Shalmaneser III (858–824 BC) (Winter,
1973: 265; Miller, 1986; Moorey, 1994; Becker, 2005, 2008).
However, other researchers view the same finds as remains of live
animals and/or raw material that originated in the Indian subconti-
nent and were traded, sent as tribute, or dispatched for help in mili-
tary campaigns to areas in the west (Deraniyagala, 1955; Winter,
1973; Colon, 1977; Vila, 2010). Reconstructing the Holocene history
of E. maximus in Southwest Asia is further complicated by imprecise
species identifications, as in the case of material from the site of
Kamid el Loz, Lebanon (Bökönyi, 1985), or problematic contexts, for
example the contested Early Bronze Age date for elephant remains
from Ras Shamra (Ugarit) in Syria (Hooijer, 1978; Moorey, 1994:118;
Caubet and Poplin, 1987, 2010).
The Elephas cf. hysudricus remains from the Middle Pleistocene sites
of Ma'ayan Baruch and 'Ain Soda, spanning the period 500–220 ka, are
chronologically intermediatebetween cf. Elephas sp.from Evron Quarry,
dated to between ca. 1.0 Ma and 0.78 Ma, and the mid-Holocene
E. maximus of the Near East. However, available data are too scanty to
assess whether this represents continuity of occupation, independent
westward expansions from further east, or importation of some or all
of the Holocene material (see above). The Pleistocene records do, how-
ever, provide a precedent for the natural expansion of Elephas as far as
the Near East. While not proving the existence of an indigenous Holo-
cene population, it makes it at least ecologically plausible.
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.palaeo.2013.05.013.
Acknowledgements
Special thanks to Ms. Gali Beiner for her invaluable and painstak-
ing restoration work on the elephant teeth from the Ma'ayan Baruch
trench, and to Mr. Vladimir Nakhlin for the photography of the
Ma'ayan Baruch and Evron Quarry teeth and lithics. We are grateful
to Prof. G.O. Rollefson and Dr. Leslie Quintero for the permission to
study and publish the 'Ain Soda remains and to Gary Sawyer (Amer-
ican Museum of Natural History) and Malon Anderson (Mammoth
Site of Hot Springs, South Dakota) for the assistance with preparation
of the casts. We would like to acknowledge the contribution of the
late Prof. Hagai Ron, who undertook the palaeomagnetic analysis of
the Ma'ayan Baruch sediment block, and Dr. Naomi Porat (Geological
Survey of Israel) for trying to date the Ma'ayan Baruch sediment
block using OSL. Thanks also to Yael Ebert for assistance with
the palaeomagnetic results. For access to modern Asian elephant
molars we thank Roberto Portela-Miguez (Natural History Museum,
London), Mark Carnall (Grant Museum, UCL, London), Matthew
Lowe and Ann Charlton (University Museum of Zoology, Cambridge),
Milly Farrell (Hunterian Museum, Royal College of Surgeons, London)
and Natasja den Ouden (Naturalis, Leiden). Work on the Ma'ayan
Baruch specimens was funded by grants from the Canadian Social Sci-
ences and Humanities Research Council to MC.
References
Albayrak, E., Lister, A.M., 2012. Dental remains of fossil elephants from Turkey. Quater-
nary International 276–277, 198–211.
Badam, G.L., 1979. Pleistocene Fauna of India, with Special Reference to the Siwaliks.
Deccan College Postgraduate and Research Institute, Pune, India.
Table 3
Comparison of M
1
from Evron Quarry with E. hysudricus and E. maximus. Evron data
from Tchernov et al. (1994); Siwalik E. hysudricus from Maglio (1973: Table 23); mod-
ern and subfossil E. maximus (Supplementary S6). Numbers in brackets are sample
sizes.
Taxon P (uppers & lowers pooled) LF (uppers) H (uppers)
Evron 10 7.5 >87
E. hysudricus M1 9–10 (5) 5.5–8.2 (2) 108–111 (2)
E. maximus M1 12–15 (7) 7.8–10.5 (6) 127–143 (4)
128 A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

Bar-Yosef, O., 1994. The Lower Paleolithic of the Near East. Journal of World Prehistory
8, 211–265.
Bar-Yosef, O., Belmaker, M., 2011. Early and Middle Pleistocene faunal and hominins
dispersals through Southwestern Asia. Quaternary Science Reviews 30, 1318–1337.
Bate, D.M.A., 1937. Palaeontology: the fossil fauna of the Wady el-Mughara caves. In:
Garrod, D.A.E., Bate, D.M.A. (Eds.), Stone Age of Mount Carmel. Cambridge University
Press, pp. 135–233.
Becker, C., 2005. Small numbers, large potential—new prehistoric finds of elephant and
beaver from the Khabur River/Syria. Munibe 57, 445–456.
Becker, C., 2008. Die tierknochenfunde aus Tall Seh Hamad/ Dur-Katlimmu: eine
zoogeographischhaustierkundliche studie. In: Kuhne, H. (Ed.), Umwelt und
Subsitenz der Assyrischen Stadt Dur-Katlimmu am Unteren Habur, BATSH 8.
Harrassowitz Verlag, Wiesbaden, pp. 61–132.
Bökönyi, S., 1985. Subfossil elephant remains from southwestern Asia. Paléorient 11,
161–163.
Cameron, D., Patnaik, R., Sahni, A., 2004. The phylogenetic significance of the Middle
Pleistocene Narmada hominin cranium from central India. International Journal
of Osteoarchaeology 14, 419–447.
Caubet, A., Poplin, F., 1987. Matières dures animales: étude du matériaux. In: Yon, M.
(Ed.), Ras Shamra Ougarit III, Le centre de la ville, pp. 273–306.
Caubet, A., Poplin, F., 2010. Réflexions sur la question de l'éléphant syrien. In: Kühne, H.
(Ed.), Dur-Katlimmu 2008 and Beyond. Otto Harrassowitz GmbH & Co., KG, Wies-
baden, pp. 1–10.
Chauhan, P.R., 2008. Large mammal fossil occurrences and associated archaeological
evidence in Pleistocene contexts of peninsular India and Sri Lanka. Quaternary In-
ternational 192, 20–42.
Colon, D., 1977. Ivory. Iraq 39, 219–222.
Copeland, L., Hours, F., 1989. The Hammer on the Rock: studies in the Early Palaeolithic
of Azraq, Jordan. British Archaeological Reports, International Series, 540. Oxford
University Press, Oxford.
Davies, P., Lister, A.M., 2007. Palaeoloxodon. In: Chazan, M., Horwitz, L.K. (Eds.), Holon
A Lower Paleolithic Site in Israel. : American School of Prehistoric Research Bulle-
tin, 50. Peabody Museum of Archaeology and Ethnology, Harvard University, Cam-
bridge Mass, pp. 123–131.
Early hominin landscapes in Northern Pakistan: investigations in the Pabbi Hills. In:
Dennell, R.W. (Ed.), British Archaeological Reports, International Series, 1265.
Oxford University Press, Oxford.
Deraniyagala, P.E.P., 1955. Some Extinct Elephants, Their Relatives and the Two Living
Species. Government Press, Ceylon.
Dirks, W., Watson, R.P., Schnurrenberger, D.W., 1998. Preliminary account of Pleisto-
cene mammals from 'Ain Soda, Azraq Basin, Jordan. Journal of Vertebrate Paleon-
tology 18 (Suppl. 3), 38A–39A.
Gilead, D., 1977. Some metrical studies of Acheulian assemblages in Israel. Eretz Israel
13, 38–49.
Gopher, A., Ayalon, A., Bar-Matthews, M., Barkai, R., Frumkin, A., Karkanas, P., Shahack-
Gross, R., 2010. The chronology of the late Lower Paleolithic in the Levant based on
U–Th ages of speleothems from Qesem Cave, Israel. Quaternary Geochronology 5,
644–656.
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.
Paléorient 20, 99–112.
Goren-Inbar, N., Feibel, C.S., Verosub, K.L., Melamed, Y., Kislev, M.E., Tchernov, E.,
Saragusti, I., 2000. Pleistocene milestones on the out-of-Africa corridor at Gesher
Benot Ya'aqov, Israel. Science 289, 944–947.
Grosman, L., Smikt, O., Smilansky, U., 2008. On the application of 3-D scanning technol-
ogy for the documentation and typology of lithic artifacts. Journal of Archaeological
Science 35, 3101–3110.
Gur, D., Sass, E., Bar-Matthews, M., Heimann, A., 2002. Tufa deposits and spring and
river water in the northern Hula Valley. Israel Geological Society Annual Meeting
Field Trip Guidebook, pp. 83–103 (in Hebrew).
Heimann, A., Sass, E., 1989. Travertine in the northern Hulla Valley, Israel. Sedimentol-
ogy 36, 95–108.
Hofman, I., 1974. Die Artzugehörigkeit des syrischen Elefanten. Säugetierkundliche
Mitteilungen 22, 225–232.
Hooijer, D.A., 1958. An Early Pleistocene mammalian fauna from Bethlehem. Bulletin of
the British Museum (natural history). Geology 3 (8), 265–292.
Hooijer, D.A., 1978. The Indian elephant at Bronze Age Ras Shamra, Ugarit. In:
Schaeffer, C.F.-A. (Ed.), Ugaritica VII. Mission de Ras Shamra 18. : Bibliotheque
Archeologique et Historique, 99. Paul Geuthner, Paris, pp. 187–188.
Horowitz, A., 1979. The Quaternary of Israel. Academic Press, New York.
Horowitz, A., 2001. The Jordan Rift Valley. Taylor and Francis.
Horwitz, L.K., Tchernov, E., 1989. The Late Acheulian fauna from Oumm Zinat. Journal
of the Israel Prehistoric Society 22, 7–14.
Khatry, A.P., 1966. The Pleistocene mammalian fossils of the Narmada River valley and
their horizons. Asian Perspectives 9, 113–133.
Lister, A.M., 2004. Ecological interactions of elephantids in Pleistocene Eurasia:
Palaeoloxodon and Mammuthus. In: Goren-Inbar, N., Speth, J.D. (Eds.), Human Pa-
leoecology in the Levantine Corridor. Oxbow Books, Oxford, pp. 53–60.
Lister, A.M., 2012. Quantitative analysis of mammoth remains from Lynford, Norfolk,
England. In: Boismier, W.A., Gamble, C., Coward, F. (Eds.), Neanderthals Among
Mammoths: Excavations at Lynford Quarry, Norfolk English Heritage Archaeological Re-
ports, pp. 205–214 (http://www.english-heritage.org.uk/publications/neanderthals-
among-mammoths/).
Lister, A.M., Joysey, K.A., 1992. Scaling effects in elephant dental evolution—the exam-
ple of Eurasian Mammuthus. In: Smith, P., Tchernov, E. (Eds.), Structure, Function
and Evolution of Teeth. Freund, Jerusalem, pp. 185–213.
Lister,A.M., van Essen, H., 2003. Mammuthus rumanus(Stefanescu),the earliest mammoth
in Europe .In: Petcu lescu, A., Ştiucă, E. (Eds.), Advances in Palaeontology'Hen to Panta.
Romanian Academy. ‘Emil Racoviţă’Institute of Speleology, Bucharest, pp. 47–52.
Lister, A.M., Sher, A.V., van Essen, H., Wei, Guangbiao, 2005. The pattern and process of
mammoth evolution in Eurasia. Quaternary International 126–128, 49–64.
Maglio, V.J., 1973. Origin and evolution of the Elephantidae. Transactions of the Amer-
ican Philosophical Society of Philadelphia, New Series 63 (3), 1–149.
Marcelino, V., Stoops, G., Schaefer, C.E.G.R., 2010. Oxic and Related Materials. In:
Stoops, G., Marcelino, V., Mees, F. (Eds.), Interpretation of Micromorphological Fea-
tures of Soils and Regoliths. Elsevier, Amsterdam, pp. 305–327.
Markov, G.N., 2012. Mammuthus rumanus, early mammoths, and migration out of
Africa: some interrelated problems. Quaternary International 276–277, 23–26.
Miller, R., 1986. Elephants, ivory, and charcoal: an ecological perspective. Bulletin of
the American Schools of Oriental Research 264, 29–43.
Moorey, P.S., 1994. Ancient Mesopotamian Materials and Industries: The Archeological
Evidence. Oxford University Press, Oxford.
Nanda, A.C., 2002. Upper Siwalik mammalian faunas of India and associated events.
Journal of Asian Earth Sciences 21, 47–58.
Nanda, A.C., 2008. Comments on the Pinjor mammalian fauna of the Siwalik Group in
relation to the post-Siwalik faunas of peninsular India and Indo-Gangetic Plain.
Quaternary International 192, 6–13.
Porat, N., Chazan, M., Schwarcz, H., Horwitz, L.K., 2002. Timing of the Lower to Middle
Paleolithic boundary: new dates from the Levant. Journal of Human Evolution 43,
107–122.
Quintero, L.A., Wilke, P.J., Rollefson, G.O., 2004. The Eastern Levant, the Pleistocene and
paleoanthropology. ACOR Newsletter 16 (1), 1–3.
Quintero, L.A., Wilke, P.J., Rollefson, G.O., 2005. The Acheulian of 'Ain Soda and al-Jafr
Basin, Jordan. Talk Presented at the 69th Annual Meeting. Society for American
Archaeology, Salt Lake City, UT.
Rabinovich, R., Ackermann, O., Aladjem, E., Barkai, R., Biton, R., Milevski, I., Solodenko,
N., Marder, O., 2012. Elephants at the Middle Pleistocene Acheulian open-air site of
Revadim Quarry, Israel. Quaternary International 276–277, 183–197.
Rech, J.A., Quintero, L.A., Wilke, P.J., Winer, E.R., 2007. The Lower Paleolithic landscape
of 'Ayoun Qedim, al-Jafr Basin, Jordan. Geoarchaeology 22 (2), 261–275.
Rollefson, G.O., 1983. Two seasons of excavation at 'Ain el-Assad, eastern Jordan. Bulle-
tin of the American Schools of Oriental Research 252, 25–34.
Rollefson, G.O., Quintero, L., Wilke, P., Schnurrenburger, D., Low, R., Watson, R., 1997.
Excavations at 'Ain Soda in Azraq. Occident and Orient (Newsletter of the German
Protestant Institute of Archaeology in Amman Eastern Jordan), 2 (2) 16–17.
Rollefson, G.O., Schnurrenberger, D., Quintero, L.A., Watson, R.P., Low, R., 1997. 'Ain
Soda and 'Ain Qasiya: New late Pleistocene and early Holocene sites in the Azraq
Shishan area, eastern Jordan. In: Gebel, H.G.K., Kafafi, Z., Rollefson, G.O. (Eds.), Pre-
history of Jordan II: Perspectives from 1997 Studies in Early Near Eastern Produc-
tion, Subsistence and Environment, 4. Ex Oriente, Berlin, pp. 45–58.
Rollefson, G.O., Quintero, L.A., Wilke, P.J., 2006. Late Acheulian variability in the South-
ern Levant: a contrast of the western and eastern margins of the Levantine corri-
dor. Near Eastern Archaeology 69 (2), 61–72.
Ron, H., Porat, N., Ronen, A., Tchernov, E., Horwitz, L.K., 2003. Magnetostratigraphy of
the Evron Member—implications for the age of the Middle Acheulian site of
Evron Quarry. Journal of Human Evolution 44, 633–639.
Ronen, A., Ohel, M., Lamdan, M., Assaf, A., 1980. Acheulean artifacts from two trenches
at Ma'ayan Barukh. Israel Exploration Journal 30, 17–33.
Roth, V.L., 1989. Fabricational noise in elephantdentitions. Paleobiology 15 (2), 165–179.
Roth, V.L., Shoshani, J., 1988. Dental identification and age determination in Elephas
maximus. Journal of Zoology (London) 214, 567–588.
Saegusa, H., Gilbert, W.H., 2008. Elephantidae. In: Gilbert, W.H., Asfaw, B. (Eds.), Homo
erectus: Pleistocene Evidence from the Middle Awash, Ethiopia. : The Middle
Awash Series, volume 1. University of California Press, Berkeley and Los Angeles,
pp. 193–226.
Santiapilli, C., Jackson, C., 1990. The Asian Elephant: An Action Plan for its Conservation.
International Union for Conservation of Nature, Gland, Switzerland.
Schwarcz, H.P., Blackwell, B., Goldberg, P., 1980. Uranium series dating of archaeologi-
cal sites in Israel. Israel Journal of Earth Sciences 29, 157–165.
Seidner, G., Horowitz, A., 1974. Radiometric ages of late Cainozoic basaltsfrom northern
Israel: chronostatigraphic implications. Nature 250, 23–26.
Sher, A.V., Garutt, V.E., 1987. New data on the morphology of elephant molars.
Doklady: Transactions of the USSR Academy of Sciences Earth Science Sections
285, 195–199.
Shoshani, J., Eisenberg, J.F., 1982. Elephas maximus. Mammalian Species 182, 1–8.
Shoshani, J., Goren-Inbar, N., Rabinovich, R., 2001. A stylohyoideum of Palaeoloxodon
antiquus from Gesher Benot Ya'aqov, Israel: morphology and functional inferences.
In: Cavarretta, G., Gioia, P., Mussi, M., Palombo, M.R. (Eds.), The World of Elephants
—Proceedings of the 1st International Congress. Consiglio Nazionale delle Ricerche,
Rome, pp. 665–667.
Stekelis, M., Gilead, D., 1966. Ma'ayan Barukh: a Lower Paleolithic site in Upper Galilee.
Journal of the Israel Prehistoric Society 8, 1–23.
Sukumar, R., 1989. The Asian elephant: ecology and management. Cambridge Studies
in Applied Ecology and Resource Management.Cambridge University Press,
Cambridge.
Sukumar, R., 2012. The Story of Asia's Elephants. Marg Foundation, Mumbai.
Tchernov, E., Shoshani, J., 1996. Proboscidean remains in southern Levant. In: Shoshani,
J., Tassy, P. (Eds.), The Proboscidea: Evolution and Paleoecology of Elephants and
their Relatives. Oxford University Press, Oxford, pp. 225–233.
Tchernov, E., Horwitz, L.K., Ronen, A., Lister, A., 1994. The faunal remains from Evron
Quarry in relation to other Lower Paleolithic hominid sites in the southern Levant.
Quaternary Research 42, 328–339.
129A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130

Vidya, T.N.C., Sukumar, R., Melnick, D.J., 2009. Range-wide mtDNA phylogeography
yields insights into the origins of Asian elephants. Proceedings of the Royal Society
B 276, 893–902.
Vila,E.,2010.TheelephantinSyria:animportationfromtheMiddleAsia?Some
considerations about a finding of elephant bones at Mishrife/Qatna in Late
Bronze Age Syria. Abstract of Paper presented at the 11th ICAZ Meeting, 23rd–
28th August, Paris (http://alexandriaarchive.org/bonecommons/items/show/
1632).
von den Driesch, A., 1996. Faunenhistorische Untersuchungen am prähistorischen
Tierknochenmaterial vom Sirkeli Höyüuk, Adana/Türkei. Istanbuler Mitteilungen
46, 27–39.
Wilke, P.J., Quintero, L.A., Rollefson, G.O., 2005. Technology, use-life history, and typol-
ogy of Acheulian bifaces from 'Ain Soda, Jordan. 69th Annual Meeting, Society for
American Archaeology, Salt Lake City, UT.
Winter, I., 1973. North Syria in the Early First Millennium B.C., with Special Reference to Ivory
Carving. Doctoral dissertation, Columbia University University Microfilms, Ann Arbor.
130 A.M. Lister et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 386 (2013) 119–130