Late Neandertals in Southeastern Iberia: Sima de las

Palomas del Cabezo Gordo, Murcia, Spain

Michael J. Walkera, Josep Gibertb, Mariano V. Lópeza, A. Vincent Lombardic, Alejandro Pérez-Pérezd, Josefina Zapataa,
Jon Ortegaa, Thomas Highame, Alistair Pikef, Jean-Luc Schwenningere, João Zilhãof, and Erik Trinkausg,1
aÁreade Antropología Física, Departamento de Zoología y Antropología Física, Facultad de Biología, Campus Universitario de Espinardo, Universidad de
Murcia, 30100 Murcia, Spain; bInstitut Paleontológic ‘‘Dr. M. Crusafont,’’ Carrer de l’Escola Industrial 23, 08201 Sabadell, Barcelona, Spain; c2584 Blossom
Lane, New Castle, PA 16105; dSección de Antropología, Departamento de Biología Animal, Facultad de Biología, Universitat de Barcelona, Avenida Diagonal
645, 08028 Barcelona, Spain; eResearch Laboratory for Archaeology and the History of Art, University of Oxford, Oxford OX1 3QY, United Kingdom;
fDepartment of Archaeology and Anthropology, University of Bristol, 43 Woodland Road, Bristol BS8 1UU, United Kingdom; and gDepartment of

Anthropology, Washington University, Saint Louis MO 63130

Contributed by Erik Trinkaus, November 7, 2008 (sent for review October 17, 2008)

Middle Paleolithic fossil human remains from the Sima de las Palomas pendix, Figs. 1 and 2). The main shaft is 18 m deep, opening
in southeastern Iberia (dated to <43,000 – 40,000 calendar years beneath overhanging rock at ⬇123 m a.s.l. (SI Appendix, Figs. 3
before present) present a suite of derived Neandertal and/or retained and 4). The shaft’s brecciated contents were largely emptied by
ancestral morphological features in the mandibular symphysis, man- 19th century miners, who left a sediment column down one side
dibular ramus, dental occlusal morphology, and distal hand phalanx. and scattered fossiliferous rubble on the hillside. Systematic
These traits are combined with variation in the mandibular corpus, collection of disturbed remains and excavation of the upper
discrete dental morphology, tooth root lengths, and anterior dental breccia deposits were begun by M.J.W. and the late J. Gibert
size that indicate a frequency difference with earlier Iberian and more after discovery of a crushed facial skeleton (Palomas 1) from the
northern European Neandertals. The Palomas Neandertals therefore uppermost breccia by J. C. Blanco in 1991 (14, 15). Fossil human
confirm the late presence of Neandertals associated with the Iberian remains have been collected from the mine rubble (1992–1999)
persistence of the Middle Paleolithic, but suggest microevolutionary and excavated in situ (1994–present) (SI Appendix, Table 1).
processes and/or population contact with contemporaneous modern A number of the in situ human remains come from above,
humans to the north. within, and slightly below a fusiform lens of dark-gray sediment
(burnt, according to X-ray diffraction and fluorescence analy-
dentition 兩 mandible 兩 Middle Paleolithic 兩 postcrania ses). It attained a maximum thickness of ⬇20 cm in the angle of
an L-shaped excavation area in the Upper Cutting, covering a
thin, oblong marble slab (⬇40 ⫻ 30 ⫻ 10 cm) between levels 2k
I t has become apparent that the transition from a Europe
populated by Neandertals to one occupied by early modern
humans during marine (oxygen) isotope stage (MIS) 3 consisted
and 2m. The lens petered out, both near the open shaft and
below the entrance where it covered part of the foot of a
of a westward spread of modern humans, variably absorbing at downward and inwardly steeply sloping (30°– 40°) éboulis or
least some local Neandertal populations (1, 2). However, the scree of marble blocks ⱕ50 kg. A more consistent and wide-
detailed aspects of this population process remain obscure as a spread lower layer of dark-gray sediment lies beneath level 2l,
result of the scarcity of well-dated late Neandertals and early fading away beside the open shaft (Fig. 1; SI Appendix, Figs. 5
modern humans, despite the recent direct dating of several and 6).
specimens from both samples (3–9). If we are to understand the Only human bones and teeth from sediments postdating the
paleobiogeography of this process, and hence the potential éboulis (from levels no deeper than 2l) are dealt with here.
patterns of interactions between these two morphologically Although these later sediments contain few large rocks and were
defined human groups, then it is essential to document the laid down horizontally, they often have a coarse texture implying
biology of both the earliest modern humans and the latest that they were washed from the hillside into the cave and down
Neandertals. the steep interior scree slope; the slope had developed neither an
It has been recognized for some time that the Middle Paleo- eroded surface nor a calcrete crust, in contrast to the thin calcite
lithic, generally presumed in Europe to have been the product of deposit on the hillside that eventually sealed off the éboulis.
Neandertals, persisted substantially longer in Iberia south of the These later sediments contained Middle Paleolithic artifacts,
Pyrenees (south of the ‘‘Ebro Frontier’’) than elsewhere in animal bones (some of which are burnt), and pollen indicative of
Europe, to ⬇34,000 (⬇34 ka) calendar years before present (cal mild climatic conditions (14, 16, 17). Because of carbonate
BP) (⬇30 ka radiocarbon years before present (14C BP)] (10; precipitation, the éboulis is in part cemented into irregular
supporting information (SI) Appendix, Fig. 1). Even though there brecciated conglomerate masses, in which there are Middle
are Iberian Neandertal remains that have been referred to this Paleolithic artifacts, faunal remains, and variably articulated and
age (11), their purported late age has been placed in doubt (12). crushed human remains. The human fossils include the Palomas
At present, the most recent, securely dated, diagnostic Nean- 92 and 96 partial skeletons, the Palomas 93 to 95 teeth, and the
dertal fossil from the region is the Oliveira 1 middle manual remains of at least 2 more individuals, including crania with
phalanx at ⬇43,500 cal BP (⬇39 ka 14C BP) (13), about the same mandibles.
age as the northern Spain El Sidrón remains (7). It is in this
context that we present a series of Neandertal remains from the Author contributions: M.J.W., J.G., and E.T. designed research; M.J.W., J.G., M.V.L., A.V.L.,
upper levels of the Sima de la Palomas in southeastern Spain. A.P.-P., J. Zapata, J.O., T.H., A.P., J.-L.S., J. Zilhão, and E.T. performed research; M.J.W.,
M.V.L., J. Zapata, J.O., T.H., A.P., J.-L.S., and E.T. analyzed data; and M.J.W., T.H., A.P., J.-L.S.,
Results
ANTHROPOLOGY

J. Zilhão, and E.T. wrote the paper.

Context of the Palomas Human Remains. The Sima de las Palomas The authors declare no conflict of interest.
(Rock-Dove Hole) is a karstic shaft (37° 47⬘ 59⬙ N, 0° 53⬘ 45⬙ W) 1To whom correspondence should be addressed. E-mail: [email protected].
in the Permo-Triassic marble of a hill (Cabezo Gordo) in Torre This article contains supporting information online at www.pnas.org/cgi/content/full/
Pacheco township, Murcia, reaching 312 m a.s.l., overlooking the 0811213106/DCSupplemental.
Mar Menor coastal lagoon of the Mediterranean Sea (SI Ap- © 2008 by The National Academy of Sciences of the USA

www.pnas.org兾cgi兾doi兾10.1073兾pnas.0811213106 PNAS 兩 December 30, 2008 兩 vol. 105 兩 no. 52 兩 20631–20636

Fig. 1. Schematic drawing of the current profiles of the Upper Cutting (see SI Appendix, Figs. 5 and 6). (A) The éboulis above northwestern corner of the
excavated cutting. (B) Levels 2m-2o breccia containing human bones. (C) Projection of the éboulis scree slope, which is less perceptible in this profile than it had
been in now removed sections parallel to it in the foreground. (D) Uppermost limit (levels ⬇2h-2i) of lens of burnt sediment mainly in the northeastern area of
the cutting. (E) Lower limit of burnt ashy sediment (levels 2m-2o) in the northern and eastern area of the cutting.

Chronological Age of the Palomas Human Remains. The age of the dark-gray sediment level in the northeastern corner of the Upper
human fossils of concern here is primary, because the issue is the Cutting was dated by using optically stimulated luminescence
morphology of the latest Neandertals in Europe. The Palomas (OSL) (SI Appendix, Section V). The sample provided an age
artifacts are all Middle Paleolithic, made on retouched flakes of estimate of 54,700 ⫾ 4,700 cal BP for these sediments. This
flint, quartz, rock crystal, and marble (14–16, 18). Given current determination increases confidence in the U-series dates for
reliable dates for Iberian Middle Paleolithic assemblages (10, stratigraphically similar specimens and hence in the age of the
19), this indicates an age of ⱖ34 ka cal BP (ⱖ30 ka 14C BP) for stratigraphically younger deposits dated by 14C.
the deposits. These dating assessments therefore combine to indicate an age
Collagen preservation is poor in the Palomas human remains, for the upper sediment fill ⬇40 ka cal BP but possibly slightly
and direct 14C dating of the human bones has not been possible. older (i.e., ⬇43 ka cal BP). The subsample of the Palomas human
However, a combination of dating techniques permits chrono- remains from this portion of the sediment includes 63 elements,
logical control of the Upper Cutting (see SI Appendix, Figs. 7 and 54 or 85.7% of which are isolated teeth or tooth fragments. They
8, for the stratigraphic positions of the samples). make these Palomas Neandertals the most recent, and largest,
A burnt faunal bone directly adhering to the Palomas 59 sample of southern Iberian late Neandertals currently known.
mandible in level 2f provided a date of 34,450 ⫾ 600 14C BP The other Neandertals close in age are the Initial Upper
(OxA-10666) [95% CI: 40,950–37,622 cal BP using the Cari- Paleolithic ones from Spy (9), and probably Saint-Césaire,
aco06 14C curve (20)]. A burnt lagomorph bone from the deeper Arcy-Renne, and Vindija G1 (22–24). These Palomas fossils are
level 2l provided a statistically identical age of 35,030 ⫾ 270 14C also approximately the same age as the earliest modern humans
BP (OxA-15423) (95% CI: 40,986–38,850 cal BP). Given the in Europe (4, 25), albeit at the other end of Europe. If the earliest
natures of the dating specimens and the moderately high C:N phases of the Aurignacian were indeed made by modern humans
atomic ratios for these samples (albeit an expected elevation (26, 27), then the Palomas remains should overlap in time with
given the burning), it is likely that these ages fall closer to the modern humans as close as the northern Pyrenees (10).
upper limits of the 95% confidence intervals for these dates (SI
Appendix, Section III). This interpretation of these dates from The Palomas Human Remains. The Palomas fossil human remains
the upper sediment fill is supported by the associated palynology therefore consist of 3 samples. There are the undated and
(17), which shows relatively temperate conditions, probably isolated remains discovered in the miners’ rubble. There are the
during GIS9 and therefore slightly before the onset of the partial face, partial skeletons, and isolated remains from the
severely cold Heinrich 4 oscillation ⬇40 ka cal BP (21). brecciated éboulis. And there are the 63 isolated remains from
To assess further the ages of these deposits and the strati- the excavated deposits at or above the levels dated to ⬇40–43 ka
graphically older éboulis deposits, U-series (LA-ICP-MS) dates cal BP.
were obtained from 3 bones (SI Appendix, Section IV). A date The diagnostic remains from the first 2 samples can all be
of 43,800 ⫾ 750 cal BP (APSLP4) is from a faunal specimen from attributed to Late Pleistocene Neandertals. For the remains out
level 2i; it is slightly earlier than the likely 14C calibrated ranges of context, the relevant aspects include supraorbital torus pres-
from levels 2f and 2l. Two much older and statistically similar ence (Palomas 11, 12, and 62), retreating mandibular symphyses
U-series dates come from the stratigraphically deeper, steeply and lateral corpus thickness (Palomas 6 and 23), incisor and
sloping éboulis: one on a Palomas 96 metacarpal from level 2e molar occlusal morphology and incisor root length (Palomas 24
of 54,000 ⫾ 3,850 cal BP (APSLP1) and the other on a faunal and 50), manual middle phalanx breadth (Palomas 65), and
bone from level 2l of 51,000 ⫾ 1,250 cal BP (APSLP6). These femoral diaphyseal shape (Palomas 52). For the brecciated
U-series estimates presume that the samples were closed sys- remains, the isolated teeth appear undiagnostic, and Palomas 96
tems, which cannot be verified. They should therefore be re- and the other associated skeletons are still largely in breccia. Yet,
garded only as corroborating the 14C and paleoclimatological the Palomas 1 mandible has a retreating symphysis, a retromolar
assessments of the age of the upper sediment fill and the earlier space, a prominent coronoid process, and an asymmetrical
age of the éboulis level, including some mixed material in the mandibular notch. The Palomas 92 partial skeleton aligns with
burnt dark-gray sediment horizon. the Neandertals in distal humeral, proximal ulnar and femoral
In addition, a sediment sample (X2509) from the top of level diaphyseal morphology, as well as inferred body proportions.
2k directly overlying the marble slab and hence below the Although many of the isolated remains from Palomas are

20632 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0811213106 Walker et al.

 Fig. 3. Bivariate plot of external bi-deciduous canine (dc1) dental arcade
breadth versus developmental age for immature Late Pleistocene human
mandibles. Symbols: black diamond, Palomas 49; gray circles, Neandertals;
black squares, MPMH; black triangle, EUP modern human (La Quina-Aval 4);
open triangles, MUP modern humans. Ages are based on dental calcification
Fig. 2. Occlusal view of the immature Palomas (SP) 49 mandibular corpus and relative to extant humans.
lateral views of the Palomas 59, 80 and 88 mandibles. Palomas 59 is in norma
lateralis, and Palomas 80 and 88 are in the planes of their lateral corpori. Scale
in centimeters. 2 high outlier trimmed, 1.97 standard deviations with Kebara 2
included).
undiagnostic as to human group in a Late Pleistocene context,
The Dentitions. The abundance of isolated teeth, plus those in
there are sufficient indicators to align these remains with the
Palomas 59 and 80, make several observations possible. All four
Neandertals. In no case do any of the remains exhibit uniquely
of the maxillary central incisors (I1s: Palomas 34, 73, 79, and 90)
derived characteristics of modern humans (cf., ref. 28).
exhibit moderate to marked labial convexity, and the 3 lingually
Given the concern with late Neandertal paleobiology, the
preserved ones have large marginal ridges and lingual tubercles
considerations here are limited to the remains found in situ at or
(Fig. 5). The 2 maxillary canines (C1s: Palomas 35 and 74) have
above the 40–43 ka cal BP levels. The comparative samples
very small lingual tubercles, and Palomas 35 has little if any
consist of MIS 5–3 western Eurasian Neandertals, MIS 5 Middle
Paleolithic early modern humans (MPMH), circum-Mediterra- shoveling (Fig. 5). Of the 4 P4s (Palomas 57, 59, 78, and 87), 3
nean Early Upper Paleolithic modern humans (EUP) (⬎33 ka each exhibit a transverse crest even though only the Palomas 59
cal BP), and western Eurasian Middle Upper Paleolithic modern crest is pronounced (Fig. 6). All of them have a mesially
humans (MUP) (⬇33–24 ka cal BP). displaced metaconid, and 3 have extra lingual cusps. Only one of
the P4s, Palomas 59, has lingual asymmetry, but it probably
The Mandibles. Four partial mandibles were found in situ in the
lacked extra lingual cusps. Yet, there are no consistent associ-
younger levels (Palomas 49, 59, 80, and 88) (Fig. 2). Palomas 59 ations between these traits across the 4 Palomas P4s. Of the 5 first
is a left corpus lacking the full symphysis, and the others are and second lower molars (M1s and M2s: Palomas 29, 80, and 84,
variably complete immature specimens. with one each from Palomas 59), all exhibit anterior fovea but 3
The preserved bone of Palomas 59 indicates that it had a lack midtrigonid crests (Fig. 7). The one M2 (Palomas 36) has a
relatively vertical symphysis but no prominent development of skewed profile and centrally placed cusps.
either a tuber symphyseos or lateral tubercles [probably mentum Most of these dental occlusal traits occur in both the Nean-
osseum rank 3 (29)]. The immature Palomas 49 has a similar or dertals and other Pleistocene (and recent) human samples, and
more retreating symphyseal profile. The dental arcade is only
intact for Palomas 49, and as with other very young Neandertals
(30), its bi-dc1 external arcade diameter (35.2 mm) is beyond
those of similarly aged early modern humans (Fig. 3), including
that of the earlier Aurignacian La Quina-Aval 4 mandible (⬇30.0
mm). Palomas 59 had a retromolar space, and Palomas 80 has a
prominent coronoid process and an asymmetrical mandibular
notch, but an open mandibular foramen. Most of these features
align them principally with the Neandertals among MIS 5–3
humans (8).
At the same time, the mental foramina of the adult Palomas
59 (at P4M1) and the infant Palomas 49 and 88 (at dm1) are
moderately mesial (8, 31), but the juvenile Palomas 80 mandible
has an unusually mesial mental foramen, because it was distinctly
mesial of the P3P4 break. There is little difference in lateral
mandibular corpus height between Neandertals and early mod-
ern humans (Kruskal–Wallis P ⫽ 0.251), but there is a significant
ANTHROPOLOGY

Fig. 4. Bivariate plot of mandibular corpus breadth versus height at the

(P ⫽ 0.0002) difference in breadth (SI Appendix, Table 4). The mental foramen, for Late Pleistocene mature mandibles. Symbols as in Fig. 2;
other Palomas mandibles (1, 6, and 23) are with other Nean- numbered symbols are for Palomas 1, 6, 23, and 59, the first 3 of which are
dertals in corpus breadth (Fig. 4). Palomas 59, however, is among geologically older than Palomas 59 or undated. The early modern humans
the EUP and MUP modern humans; its corpus breadth is 2.21 with high corpus breadths are Qafzeh 9 and Skhul 4 (MPMH), Nazlet Khater 2
standard deviations from the Neandertal mean (with the Kebara (EUP), and Cro-Magnon 1 (MUP); the high Neandertal outlier is Kebara 2.

Walker et al. PNAS 兩 December 30, 2008 兩 vol. 105 兩 no. 52 兩 20633
Fig. 5. Lingual views of Palomas (SP) maxillary central incisors (I1s) (SP 34, 79
and 90) and maxillary canines (C1s) (SP 35 and 74). Scale in millimeters.

Fig. 7. Occlusal views of Palomas (SP) molars. SP 36, maxillary M2; SP 59,
not all Neandertals exhibit all of them (32). However, these are mandibular M1 and M2; SP 80, mandibular M2; SP 29, mandibular M2; SP 84,
all traits that occur in high frequencies among the Neandertals, mandibular M1. Scale in millimeters.
and Neandertals in particular have high frequencies, compared
with other Pleistocene samples, of these traits occurring in
combination. Of these Palomas teeth, only the I1s and the M2 phalanx. It has a broad, rounded apical tuberosity lacking ungual
exhibit the full suite of features increasingly considered charac- spines (Fig. 8), an archaic Homo configuration (36, 37). Nean-
teristic of the Neandertals. dertal distal phalangeal breadths are significantly absolutely
Root lengths, especially of anterior teeth, have been shown to broader than those of almost all early modern humans (Fig. 8;
largely differentiate Neandertals and Upper Paleolithic humans Kruskal–Wallis P ⫽ 0.00003) (SI Appendix, Table 7), and most
(33); additional data (SI Appendix, Table 5) indicate a significant of them are broader relative to phalangeal length despite the
comparative sample difference in root lengths for all but the I2. relatively longer distal phalanges of the Neandertals (38). The
Of the 13 Palomas anterior teeth providing root lengths, 4 are distal breadth of the Palomas 28 phalanx (9.7 mm) falls abso-
below the Neandertal ranges, 4 are ⬎2␴ from the Neandertal lutely and relatively among those archaic humans (Fig. 8 and SI
means, and 77% are below the Neandertal means. Appendix, Table 7).
The Palomas 59 M1 exhibits supraradicular taurodontism, and
the M2 has pronounced radicular endotaurodontism, the latter Discussion
especially being characteristic of the Neandertals (34). Yet, the Neandertal Affinities. These considerations of the in situ human
Palomas 29 M2 lacks any pulp chamber expansion. remains from the upper levels of the Sima de las Palomas
Neandertals and early modern humans have similar postca- confirm that they are best seen as late southwestern European
nine dental dimensions (35), but the former have greater I1 and Neandertals. There is a suite of features, including mandibular
I2 labiolingual crown diameters (Kruskal–Wallis P ⬍ 10⫺6 for symphyseal configuration, ramal shape, dental occlusal mor-
each) but contrast less in C1 breadths (Kruskal–Wallis P ⫽ 0.025) phology, and manual distal phalanx shape, that places them with
(SI Appendix, Table 6). One of the in situ Palomas I1s (Palomas archaic Homo and separate from early modern humans. More-
21) and 3 of the 5 C1s (Palomas 26, 54, 59) are below the over, the I1s and the P4s, M2, M1s and M2s exhibit apparently
Neandertal range, and the remainder of the Palomas anterior derived Neandertal occlusal traits or combinations of traits.
mandibular teeth are at or below the Neandertal means [2-tailed Other retained plesiomorphous aspects lost among early modern
Wilcoxon P ⫽ 0.047 (I1), 0.312 (I2), 0.013 (C1)]. humans or autapomorphous traits of the Neandertals (28) are
not preserved or evident only on undated or older Palomas
The Postcrania. In the postcrania, despite multiple elements (SI
fossils. These considerations should nonetheless be sufficient to
Appendix, Table 1), only 1 diagnostic bone is stratigraphically
secure in the more recent deposits, the Palomas 28 distal hand

Fig. 8. Bivariate plot of distal phalangeal breadth versus articular length for
Palomas 28 and Late Pleistocene samples. Symbols as in Fig. 2. Given uncer-
tainties in digit assignment for isolated ray 2– 4 distal phalanges, values are
Fig. 6. Occlusal views of Palomas (SP) mandibular second premolars (P4s). averaged for those individual preserving multiple distal phalanges to provide
Scale in millimeters. an individual value.

20634 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0811213106 Walker et al.

confirm that at ⬇40–43 ka cal BP, in southeastern Iberia, the the presence of autapomorphous modern human traits (e.g., a
Middle Paleolithic population consisted of Neandertals. narrow mandibular corpus, reduced C1 lingual tubercles, short
anterior tooth roots, or small anterior dental crowns) in a
Morphological Variability. At the same time, it is evident that the Neandertal population. The second explanation would imply
late Palomas Neandertals exhibit a complex mix of Neandertal that some of these mandibular and dental features confer an
and more ‘‘modern’’ features. Three of the mandibles have advantage on these populations, although variation in some of
mental foramina that are moderately mesial for Neandertals, but the features may well be selectively neutral. The third consid-
one is unusually so. Distinctive dental features such as large C1 eration would require that the earliest Aurignacian of the
lingual tubercles, M1 and M2 midtrigonid crest, P4 transverse northern Pyrenees and elsewhere be the product of modern
crest, lingual tubercles, mesial metaconid and lingual asym- humans, to provide a geographically proximate source for gene
metry, long anterior tooth roots, anterior crown dimensions, flow. It would also invoke human biological contact across the
and lower molar taurodontism are reduced or variably present ‘‘Ebro Frontier’’ during a time when there is little evidence for
in the sample. None of the P4s has all 4 of the ‘‘Neandertal’’ cultural diffusion (10). In this scenario, the cultural contrasts
configurations. across the ‘‘Ebro Frontier’’ would be due to behavioral choices,
It is possible to find individual Neandertal teeth or mandibles possibly ecologically driven, rather than isolation of the southern
that exhibit one or more of most these ‘‘non-Neandertal’’ Iberian populations.
aspects, and biological variation across the Neandertals (includ-
Conclusion
ing trends through time and clinal variation in space) has been
noted (7, 39–42). However, the level of variation in these The human remains from the Sima de la Palomas in southeastern
features in the Palomas late Middle Paleolithic sample is unusual Iberia therefore document the presence of Neandertals, rela-
for a group of Neandertals. It is possible that these contrasts with tively late in the Middle Paleolithic. They help to substantiate
other Neandertals represent: (i) late Neandertal genetic drift in that the Middle Paleolithic of the region was the product of
the direction of modern human morphology through isolation- Neandertals, even though diagnostic human remains associated
by-distance in the cul-de-sac of southern Iberia, (ii) an adaptive with the very latest phases of this technocomplex in Europe
shift to local environmental constraints in some of these features, remain elusive. At the same time that the Palomas humans
and/or (iii) the product of gene flow from early modern human exhibit a suite of derived Neandertal features and archaic Homo
populations to the immediate north. configurations long since lost among early modern humans, their
morphological variation indicates that they deviate from the
The first explanation would emphasize regional (perhaps
expected Neandertal ranges of variation. This pattern may be
clinal) variation among the Neandertals (cf., refs.7 and 40).
result of genetic drift in relative isolation, directional change or,
Securely dated earlier MIS 4–3 Neandertals from south of the
perhaps more likely, population contact to the north.
Pyrenees [e.g., Banyoles (43, 44), Cova Negra (45), Gegant (46),
Valdegoba (47), Zafarraya (48), plus the earlier Palomas re- ACKNOWLEDGMENTS. We thank the landowners of Cabezo Gordo and the
mains] exhibit the Neandertal pattern in comparable elements management of the Cabezo Gordo S.A. quarry for allowing the excavations to
(wide lateral mandibular corpori, relatively posterior mental occur, T. Rodríguez and J. L. Polo for providing X-ray diffraction and fluorescence
foramina, large anterior teeth, anterior foveae and midtrigonid analyses, and the many colleagues and volunteers over the years for enthusias-
tically providing assistance and expertise. This work was supported by the Spanish
crests on M1–M3, and asymmetrical P4s with mesial metaconids government (CGL2005/02410/BTE; BOS/2002/02375; PB/98/0405; PB/92/0971), the
and lingual tubercles) with minor variation in mental foramen Murcian government (PSH93/52; 05584/ARQ/07; CTC/DGC/SPH/063/2001; CCE/
position and a couple of the dental traits. These fossils thus imply DGC/IPH/SAR0/1998; CCE/DGC/IPH/SAR/1997; CCE/DGC/IPH/SAR/011/1996; CCE/
DGC/IPH/SAR/1995; CCE/DGC/IPH/SAR/1994; PSH/93/52), the municipality of Torre
that the Palomas variation is not merely the result of long-term Pacheco, and the Earthwatch Institute (1994 –2001). Analysis of the human
isolation of Iberian Neandertals through the earlier Late Pleis- remains was supported by the Spanish government (CGL2007– 60802/BTE)
tocene. Moreover, such isolation would not necessarily explain (A.P.-P.) and Washington University (E.T.).

1. Smith FH, Janković I, Karavanić I (2005) The assimilation model, modern human origins 15. Walker MJ, et al. (1999) Excavations at new sites of early man in Murcia, Sima de las
in Europe, and the extinction of the Neandertals. Quat Int 137:7–19. Palomas del Cabezo Gordo and Cueva Negra del Estrecho del Río Quípar de la
2. Trinkaus E (2007) European early modern humans and the fate of the Neandertals. Proc Encarnación. Hum Evol 14:99 –123.
Natl Acad Sci USA 104:7367–7372. 16. Walker MJ (2001) A Very Remote Period Indeed, eds Milliken S, Cook J (Oxbow,
3. Schmitz RW, et al. (2002) The Neandertal type site revisited: Interdisciplinary investi- Oxford), pp 153–159.
gations of skeletal remains from the Neander Valley, Germany. Proc Natl Acad Sci USA 17. Carrión JS, et al. (2003) Glacial refugia of temperate, Mediterranean and Ibero-North
99:13342–13347. African flora in south-eastern Spain: New evidence from cave pollen at two Neander-
4. Trinkaus E, et al. (2003) An early modern human from the Peştera cu Oase, Romania. thal man sites. Glob Ecol Biogeog 12:119 –129.
Proc Natl Acad Sci USA 100:11231–11236. 18. Walker MJ, et al. (2004) Settlement Dynamics in the Middle Palaeolithic and Middle
5. Wild EM, et al. (2005) Direct dating of Early Upper Palaeolithic human remains from Stone Age 2, ed Conard NJ (Kern, Tübingen), pp 461–511.
Mladeč. Nature 435:332–335. 19. Zilhão J, Pettitt P (2006) On the new dates for Gorham’s Cave and the late survival of
6. Trinkaus E (2005) Early modern humans. Annu Rev Anthropol 34:207–230.
Iberian Neanderthals. Before Farming 2006(3):95–122.
7. Rosas A, et al. (2006) Paleobiology and comparative morphology of a late Neandertal
20. Hughen K, Southon J, Lehmanc S, Bertranda C, Turnbull J (2006) Marine-derived 14C
sample from El Sidrón, Asturias, Spain. Proc Natl Acad Sci USA 103:19266 –19271.
calibration and activity record for the past 50,000 years updated from the Cariaco
8. Soficaru A, Doboş A, Trinkaus E (2006) Early modern humans from the Peştera Muierii,
Basin. Quat Sci Rev 25:3216 –3227.
Baia de Fier, Romania. Proc Natl Acad Sci USA 103:17196 –17201.
21. Sepulchre P, et al. (2007) H4 abrupt event and late Neanderthal presence in Iberia.
9. Semal P, et al. (2008) New data on the late Neandertals: Direct dating of the Belgian
Earth Planet Sci Lett 258:283–292.
Spy fossils. Am J Phys Anthropol, in press.
22. Mercier N, et al. (1991) Thermoluminescence dating of the late Neanderthal remains
10. Zilhão J (2006) Chronostratigraphy of the Middle-to-Upper Paleolithic transition in the
from Saint-Césaire. Nature 351:737–739.
Iberian peninsula. Pyrenae 37:7– 84.
11. Hublin JJ, Barroso Ruiz B, Medina Lara P, Fontugne M, Reyss JL (1995) The Mousterian 23. Bailey SE, Hublin JJ (2006) Dental remains from the Grotte du Renne at Arcy-sur-Cure
site of Zafarraya (Andalucia, Spain): Dating and implications on the Paleolithic peo- (Yonne). J Hum Evol 50:485–508.
pling processes of western Europe. C R Acad Sci 321:931–937. 24. Higham T, Bronk Ramsey C, Karavanić I, Smith FH, Trinkaus E (2006) Revised direct
12. Michel V, et al. (2003) El Pleistocene Superior del la Cueva del Boquete de Zafarraya, radiocarbon dating of the Vindija G1 Upper Paleolithic Neandertals. Proc Natl Acad Sci
ANTHROPOLOGY

ed Barroso-Ruíz C (Consejería de Cultura, Junta de Andalucía, Sevilla), pp 115–155. USA 103:553–557.

13. Trinkaus E, Maki J, Zilhão J (2007) Middle Paleolithic human remains from the Gruta da 25. Zilhão J, et al. (2007) Rethinking the Human Revolution, eds Mellars P, Boyle K,
Oliveira (Torres Novas), Portugal. Am J Phys Anthropol 134:263–273. Bar-Yosef O, Stringer C (McDonald Institute of Archaeology, Cambridge, UK), pp
14. Walker MJ, et al. (1998) Two SE Spanish Middle Palaeolithic sites with Neanderthal 249 –262.
remains: Sima de la Palomas del Cabezo Gordo and Cueva Negra del Estrecho del Río 26. Henry-Gambier D, Maureille B, White R (2004) Vestiges humains des niveaux de
Quípar (Murcia Province). Internet Archaeol 5. Available at http://intarch.ac.uk/ l’Aurignacien ancien du site de Brassempouy (Landes). Bull Mem Soc Anthropol Paris
journal/issue5/). Accessed October 24, 2008. 16:49 – 87.

Walker et al. PNAS 兩 December 30, 2008 兩 vol. 105 兩 no. 52 兩 20635
27. Bailey SE, Hublin JJ (2005) Who made the Early Aurignacian? A reconsideration of the 39. Smith FH (1984) The Origins of Modern Humans, eds Smith FH, Spencer F (Liss, New
Brassempouy dental remains. Bull Mem Soc Anthropol Paris 17:115–121. York), pp 137–209.
28. Trinkaus E (2006) Modern human versus Neandertal evolutionary distinctiveness. Curr 40. Trinkaus E (1991) Les hommes fossiles de la Grotte de Shanidar, Irak: Évolution et
Anthropol 47:597– 620. continuité parmi les hommes archaïques tardifs du Proche-Orient. L’Anthropol 95:535–
29. Dobson SD, Trinkaus E (2002) Cross-sectional geometry and morphology of the man- 572.
dibular symphysis in Middle and Late Pleistocene Homo. J Hum Evol 43:67– 87. 41. Trinkaus E, Churchill SE, Ruff CB, Vandermeersch B (1999) Long bone shaft robusticity
30. Mallegni F, Trinkaus E (1997) A reconsideration of the Archi 1 Neandertal mandible. J and body proportions of the Saint-Césaire 1 Châtelperronian Neandertal. J Archaeol
Hum Evol 33:651– 668. Sci 26:753–773.
31. Coqueugniot H (1999) Le crâne d’Homo sapiens en Eurasie: Croissance et variation 42. Caramelli et al. (2006) A highly divergent mtDNA sequence in a Neandertal individual
depuis 100,000 ans. Brit Archaeol Rep S 822:1–197.
from Italy. Curr Biol 16:630 – 632.
32. Bailey SE (2006) Beyond shovel-shaped incisors: Neandertal dental morphology in a
43. Maroto J (ed) (1993) La mandíbula de Banyoles en el context dels fòssils humans del
comparative context. Period Biol 108:253–267.
Pleistocè. Ctr Invest Arqueol Girona 13:1–194.
33. Bailey SE (2005) Current Trends in Dental Morphology Research, ed Zadzinska E
44. Grün R, et al. (2006) ESR and U-series analyses of enamel and dentine fragments of the
(University of Lodz Press, Lodz), pp 201–210.
Banyoles mandible. J Hum Evol 50:347–358.
34. Kallay J (1970) Krapina 1899-1969, ed Malez M (Jugoslavenske akademije znanosti I
45. Arsuaga JL, et al. (2007) New Neandertal remains from Cova Negra (Valencia, Spain).
umjetnosti, Zagreb), pp 165–176.
35. Trinkaus E (2004) in Miscalánea en Homenaje a Emiliano Aguirre, ed Rubio S (Museo J Hum Evol 52:31–58.
Arqueológico Regional, Alcalá de Henares), pp 393–398. 46. Daura J, et al. (2005) A Neandertal mandible from the Cova del Gegant (Sitges,
36. Trinkaus E (1983) The Shanidar Neandertals (Academic, New York). Barcelona, Spain). J Hum Evol 49:56 –70.
37. Rosas A (1985) Falanges humanas de la Sima des los Huesos (Cueva Mayor), Sierra de 47. Quam RM, et al. (2001) Human remains from Valdegoba Cave (Huérmeces, Burgos,
Atapuerca (Burgos). Estudio anatomico y comparativo. Actas IV Cong Español Spain). J Hum Evol 41:385– 435.
Antropol Biol 4:557–566. 48. Barroso-Ruíz C, Lumley MA de, Caparros M, Verdu L (2003) El Pleistocene Superior del
38. Villemeur I (1994) La Main des Néandertaliens (Centre National de la Recherche la Cueva del Boquete de Zafarraya, ed Barroso-Ruíz C (Consejería de Cultura, Junta de
Scientifique, Paris). Andalucía, Sevilla), pp 327–387.

20636 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0811213106 Walker et al.