Articles

https://doi.org/10.1038/s41559-018-0753-6

An early Aurignacian arrival in southwestern

Europe
Miguel Cortés-Sánchez1,2, Francisco J. Jiménez-Espejo   3,4*, María D. Simón-Vallejo1,2,
Chris Stringer   5, María Carmen Lozano Francisco   2, Antonio García-Alix4,6,
José L. Vera Peláez2, Carlos P. Odriozola1,2, José A. Riquelme-Cantal7, Rubén Parrilla Giráldez2,
Adolfo Maestro González8, Naohiko Ohkouchi3 and Arturo Morales-Muñiz9

Westernmost Europe constitutes a key location in determining the timing of the replacement of Neanderthals by anatomi-
cally modern humans (AMHs). In this study, the replacement of late Mousterian industries by Aurignacian ones at the site
of Bajondillo Cave (Málaga, southern Spain) is reported. On the basis of Bayesian analyses, a total of 26 radiocarbon dates,
including 17 new ones, show that replacement at Bajondillo took place in the millennia centring on ~45–43 calibrated thousand
years before the present (cal ka bp)—well before the onset of Heinrich event 4 (~40.2–38.3 cal ka bp). These dates indicate that
the arrival of AMHs at the southernmost tip of Iberia was essentially synchronous with that recorded in other regions of Europe,
and significantly increases the areal expansion reached by early AMHs at that time. In agreement with human dispersal sce-
narios on other continents, such rapid expansion points to coastal corridors as favoured routes for early AMH. The new radio-
carbon dates align Iberian chronologies with AMH dispersal patterns in Eurasia.

T
he replacement of Middle Palaeolithic Neanderthal popu-
lations by anatomically modern humans (AMHs), which
in Europe are associated with Early Upper Palaeolithic
(EUP) industries, constitutes a crucial and hotly debated issue in
Palaeolithic studies1,2. This biocultural turnover has been addressed
from various standpoints, including interspecies competition either
in isolation3 or combined with climate change4, environmental cri-
ses and episodic events such as volcanic eruptions5. In addition,
given that biological evidence, whether bones or biomolecules, is
transmitted through genetic processes whereas cultural materi-
als are transmitted via learning processes, another issue is to what
extent the biological and cultural transitions are coupled with, or
decoupled from, each other.
Great effort has been devoted to framing the spatiotemporal fea-
tures of Neanderthal replacement, as this may help to resolve the
processes that triggered population and technological turnovers6.
This includes determining any directionality of technological and
population changes that, according to genetic and archaeological
data, become recurrent events in western Europe from the Late
Pleistocene onwards, generally exhibiting an east to west trend.
Mousterian technocomplex replacement by Aurignacian popula-
tions is postulated to be one such east-to-west population turn-
over1. This same directionality is documented for the Gravettian
technocomplex that replaced the Aurignacian7, the genetically and
culturally documented Magdalenian–Azilian transition in western
Europe at the beginning of the Bölling–Alleröd interstadial8, and
replacements within the Epigravettian cultures of southern Europe.
Among European east–west population turnovers, the
Mousterian–Aurignacian transition has perhaps received the most
attention. This is because it is associated with the putative extinction
of Neanderthals, given that Aurignacian technocomplex elements
have now been securely associated with AMHs1.
In comparing the early stages of the Aurignacian dispersal (Fig. 1)
with subsequent transitions, two spatiotemporal anomalies emerge.
The first is recorded on the Italian Peninsula, where populations
manufacturing the Uluzzian industry (for some authors, a develop-
ment rooted in the Mousterian lithic tradition9) seemingly prevented
the expansion of the early Aurignacian10. The second anomaly is
documented in mid-southern Iberia, where the Aurignacian expan-
sion is postulated to have been delayed to the point of failure11. The
proposal that the Middle Palaeolithic technocomplex extended
to the end of Marine Isotope Stage 3 (Gorham’s Cave, Gibraltar:
~32.5 calibrated thousand years before the present (cal ka bp11) has
lent weight to the hypothesis that the EUP reached southern Iberia
at a comparatively late date. This reinforced the validity of a particu-
lar version of the east-to-west wave-of-advance mode that set apart
the ‘Iberian South’ from the rest of the Peninsula12.
Results and discussion
To reliably pin down the Neanderthal–AMH transition in southern
Iberia, 17 new dates restricted to levels covering the transition of the
Mousterian to the Aurignacian at the site of Bajondillo in the Bay
of Málaga (southern Spain) (Fig. 1, Supplementary Figs. 1–3 and
Supplementary Table 1) have been integrated with previous radio-
metric dates13, using a Bayesian approach. The Middle Palaeolithic
at Bajondillo lasts for ~120 kyr14, as represented by 6 archaeological
levels (that is, Bj/19–Bj/14) (Supplementary Fig. 1), the last of which
features a Denticulate Mousterian (~50–46 cal ka bp)13 (Fig. 2).

1
Departamento de Prehistoria y Arqueología, Facultad de Geografía e Historia, Universidad de Sevilla, Seville, Spain. 2HUM-949 Research Group,
Departamento de Prehistoria y Arqueología, Facultad de Geografía e Historia, Universidad de Sevilla, Seville, Spain. 3Japan Agency for Marine-Earth
Science and Technology, Yokosuka, Japan. 4Instituto Andaluz de Ciencias de la Tierra, CSIC-UGR, Armilla, Spain. 5Department of Earth Sciences,
Natural History Museum, London, UK. 6Departamento de Estratigrafía y Paleontología, Facultad de Ciencias, Universidad de Granada, Granada, Spain.
7
Departamento de Geografía y Ciencias del Territorio, Universidad de Córdoba, Córdoba, Spain. 8Instituto Geológico y Minero de España, Madrid, Spain.
9
Laboratorio de Arqueozooarqueología, Departamento de Biología, Universidad Autónoma de Madrid, Madrid, Spain. *e-mail: fjjspejo@ugr.es

Nature Ecology & Evolution | VOL 3 | FEBRUARY 2019 | 207–212 | www.nature.com/natecolevol 207
Articles Nature Ecology & Evolution

14
Europe
49° N
19

>42 cal ka BP
17

<42 cal ka BP 20
15
Atlantic Ocean 16
13 18
11
12 10
42° N 22
Latitude

8 21

1 7
5 6
4
3
2 Bajondillo
ODP-977A Mediterranean Sea
25
35° N 24

23
Africa

26 0 250 500 750 1,000

km

7° W 0° 7° E 14° E 21° E
Longitude

Fig. 1 | Selected archaeological sites in western Europe and North Africa. Map reconstruction with the coastline at 85 m below sea level, including
elevations (contours every 500 m), ice and permafrost correspond to LGM, ~20 cal ka bp (white and blue outlined areas). Referenced archaeological sites
in Europe and North Africa where EUP sites between 44 and 42 cal ka bp have been identified are labelled as follows: Bajondillo (this paper; 3), Romaní27
(9), Arbreda27 (10), Isturitz27 (11), Labeko27 (12), La Viña27 (13), Kent’s Cavern28 (14), Castanet27 (15), Pataud27 (16), Les Cottés27 (17), Riparo Mochi27 (18),
Geißenklösterle30 (19), Fumane27 (20), Serino27 (21) and Peshtera/Kozarnika8 (22). Potential associated areas are shown in orange. EUP sites with ages
<​42 cal ka bp in the Iberian Peninsula and North Africa are labelled as follows: Pego do Diabo29 (1), Gorham’s Cave11 (2), Ventanas and Carigüela
(this paper; 4), Antón12 (5), Mallaetes17 (6), Foradada16 (7) and La Boja12 (8). Potential associated areas are shown in yellow. The African sites Haua Fteah31
(23), Grotte des Pigeons32 (Taforalt) (24), Benzú33 (25) and Jebel Irhoud34 (26) are also included. See detailed references for each archaeological site in the
Supplementary Information. Base map reproduced from ref. 46, Elsevier.

The typological features of the 13,399 lithics from this Middle
Palaeolithic package evidence the stasis of a Mousterian technologi-
cal tradition dominated by scrapers, notches and denticulates lack-
ing the operative schemes and maintenance items of the nuclei that
typify Upper Palaeolithic technocomplexes13. Our new dates reveal
that the major change in technology occurs at level Bj/13 (~43.0−​
40.8 cal ka bp; Fig. 2 and Supplementary Table 1). The technological
novelties of the 353-item assemblage from this level include blades
and bladelet cores (1.1%), end-scrapers and one borer or blade with
continued retouch (see Supplementary chapter ‘Lithic industries’,
Supplementary Table 2 and Supplementary Fig. 4 for a detailed
discussion). Level Bj/13 also shows a 233% increase in the elonga-
tion index of flaked products: the number of elements of identified
blades and bladelets increases from 6% (Bj/14) to 14% (Bj/13)13.
Such a shift is no accident. At Bj/13, coincident with a striking
increase in the number of blades and bladelets, blade and bladelet
cores and their specific rejuvenation flakes are documented for the
first time (Fig. 2, Supplementary Table 2 and Supplementary Fig. 4).
These novelties signal the end of a Middle Palaeolithic technological
tradition that lasted from ~160–46 cal ka bp.
Given the absence of transitional Middle Palaeolithic–Upper
Palaeolithic technocomplexes in southern Iberia, the techno-
logical attribution of Bj/13 to the Aurignacian seems secure,
but we were not able to distinguish whether Bj/13 corresponds
to Proto-Aurignacian or Early Aurignacian technocomplexes.
The earliest Aurignacian technocomplexes in western Europe,
starting ≤​43 cal ka bp, have been traditionally classified as Proto-
Aurignacian (Mediterranean) or Early Aurignacian, which origi-
nally appears in Central Europe but later reaches its westernmost
regions1. To define Bj/13 as Proto- or Early Aurignacian is far
from straightforward. One reason for this is that the number of
tools from Bj/13 falls below the ≥​100 tools threshold required
for a statistically reliable assignment (Supplementary Table 2).
In Iberia, this is a recurrent problem in sites covering the Middle
Palaeolithic–Upper Palaeolithic transition. For Neanderthals from
Gorham’s Cave, Zafarraya Cave and Bajondillo Cave, such artefact
scarcity has been taken to reflect a sharp decline in their demog-
raphy towards the end of the Mousterian11,13,15. A sparse demogra-
phy, coupled with a territorial model where populations did not
settle for long at any particular place, is postulated to account for
the dearth of artefactual evidence from the earliest Iberian AMHs
at sites such as Foradada, Cendres, Mallaetes, La Boja and Pego do
Diablo12,16,17. A second problem at Bj/13 is a lack of the bone tools
normally associated with the Aurignacian—a problem shared with
other Mediterranean sites. One last crucial issue in the context of
this paper is that, according to the most recent studies, the Early
Aurignacian is a stage of prehistory that is not clearly defined from
technological and typological standpoints18,19.

208 Nature Ecology & Evolution | VOL 3 | FEBRUARY 2019 | 207–212 | www.nature.com/natecolevol
Nature Ecology & Evolution Articles
Bj/14: Middle Palaeolithic/Mousterian of notches and denticulates Bj/13: Upper Palaeolithic/Aurignacian Bj/11: Upper Palaeolithic/Evolved Aurignacian
Tools/typology

2
1 2

3
1

1 2 3
5
3 4
4 5 6

4 8
6
7 8
4 7 8
5
6 8 11
7

11
9 10
Production

9 10

11 12
8 9 10 11 13 14 12 13 14 15 16
15
Flakes Bladelets and blades Bladelets and blades
Cores

13 16 17
12 14 15 18
17 18
Discoid and levallois Core for bladelets/blades rejuvenation
Upper Palaeolithic-core flakes

Fig. 2 | Representative lithic industries from Bajondillo site archaeological levels Bj/14–Bj/11. From Bj/14 (Denticulate Mousterian, Middle Palaeolithic),
we present images of a Mousterian point (1), scrapers (2–4 and 7), a denticulate (5), a Tayac point (6), Levallois flakes (8 and 9), a Levallois point (10),
a knife with a natural back (11) and cores (12–15). From Bj/13 (Aurignacian, Proto/Early), we present a nosed end-scraper (1), a borer (2), retouched flakes
(3 and 8), a retouched blade (4), notches (5 and 6), core flanks (7 and 17), blades and bladelets (9–15), cores for bladelets (16) and a fragment of core
(18). From Bj/11 (Evolved Aurignacian), we present end-scrapers (1–8), a retouched blade (9), burins (10 and 11), retouched bladelets (12–16) and cores
(17 and 18). Scale bar: 2 cm.

The evidence that allows us to identify Bj/13 as Proto-Aurignacian
or Early Aurignacian results from a combination of chronological
and stratigraphic data, complemented by historiographic and tech-
nological data (Fig. 2 and Supplementary Fig. 4). Stratigraphically,
Bj/13 is firmly set above a well-defined Middle Palaeolithic pack-
age and overlaid by another Aurignacian level (Bj/11) lying below
a Gravettian deposit (Bj/10). The stratigraphic package is well
structured in sedimentological and micromorphological terms, and
shows no inconsistency in the seriation of absolute dates indepen-
dent of the long- or short-lived nature of the samples (see below).
The range of dates obtained from the Bayesian model appears
to be fully consistent with our hypothesis of the Bj/13 Aurignacian
(43.4–40.0 cal ka bp) lying chronologically between Heinrich events
5 (50–47 cal ka bp) and 4 (40.2–38.3 cal ka bp). However, techno-
logically, level Bj/11 (~37.6–32.4 cal ka bp; Fig. 2 and Supplementary
Table 1) corresponds to an Evolved Aurignacian with blade and
bladelet technologies and the characteristic tools of this techno-
complex. Hearths also testify to differences between the Middle
Palaeolithic and EUP. Hearths from Bj/16 and Bj/15 evidence an
exclusive use of grasses (Gramineae) as fuel, with temperatures
never reaching 500 °C, whereas those from EUP levels (that is, Bj/12
onwards) exhibit a more efficient pyrotechnology, with frequent use
of wood, and temperatures well in excess of 550 °C20.
At Bajondillo, all six of the new accelerator mass spectrom-
etry (AMS) dates for the Mousterian from Bj/14 derive from
short-lived (<​5 years) shells of marine and terrestrial molluscs,
and range between >​50 and 46 cal ka bp, the youngest ages being
based on a Bayesian model (Supplementary Table 1; the pulmo-
nate samples derive from a garden snail taxon (Otala species) that
feeds on fresh vegetal matter and thus does not exhibit the ‘old-
shell effect’). These dates calibrate this level to during or immedi-
ately after Heinrich event 5 (Fig. 3). The date obtained on a mussel
from the latest Middle Palaeolithic level at the nearby Abrigo 3
site, 20 km away from Bajondillo Cave on the Bay of Málaga, falls
squarely within this range of dates (that is, 50–43.1 cal ka bp21;
Supplementary Fig. 2b).
Until recently, very late Middle Palaeolithic chronologies in
southern Iberia were available for the sites of Carigüela (Granada),
Zafarraya (Granada) and Gorham’s Cave (Gibraltar) (Fig. 1).
Ongoing research by members of this team will put in question

Nature Ecology & Evolution | VOL 3 | FEBRUARY 2019 | 207–212 | www.nature.com/natecolevol 209
Articles Nature Ecology & Evolution
a in Europe, the last-known evidence of Neanderthals is documented
16
at Spy (Belgium) (41.0–39.3 cal ka bp)6, while the Italian Uluzzian
SST (°C)
14
12 disappears ~39.5 cal ka bp9,10.
10 The oldest of the six AMS dates from Bj/13 (EUP) come from
b long-lived taxa (sample charcoal/Ua-18270) (Supplementary Table
δ18O NGRIP (‰)

–38 1), which, although coherent among these taxa, may yield ages older
–40
–42
than the deposit from which they were retrieved. Samples for Bj/13
–44 that derive from short-lived (<​5 years) mussels and a terrestrial snail
–46 (Otala species) provide a Bayesian median age set between 42.5 and
c –70 41.2 cal ka bp (Fig. 3 and Supplementary Table 1). The taphonomic
features (that is, fragmentation and thermoalteration) of all marine
Red Sea
level (m)

–80
–90 shells from Bj/13 document shellfish processing contemporane-
–100 ous with the deposit, reinforcing the autochthonous nature of the

GI5.2
GI12

GI11

GI10

–110 GI8 radiocarbon-dated samples, thus their reliability as chrono-indica-

GI7

GI6
H5

H4

d tors. The dates available from Bj/14 and Bj/13 (which were derived
CNA−3819.1.1 Bj/10
Gravettian from the middles of the layers, away from their common bound-
Bj/11–Bj/10 boundary ary) show a potential discontinuity of ~2 kyr between these levels,
CNA−3821.1.1 Bj/11 one of which requires further research to be confirmed (Fig. 3 and
CNA−3820.1.1 Evolved Supplementary Table 1). The dates from Bj/13 coincide with others
Aurignacian
CNA−3878.1.1 from early Aurignacian sites in Europe (Fig. 1), such as the Italian
CNA−3817.1.1 rock shelter of Riparo Mochi, as well as sites from Central Europe
CNA−3883.1.1 (~43-42.5 cal ka bp)23 and (~43.5 cal ka bp)24. Additionally, the most
CNA−3876.1.1 recent Middle Palaeolithic/Upper Palaeolithic chronological review
Ua−18050 from the Cantabrian area (northern Spain)25 and Central Spain26
Ua−17150 yields essentially identical dates to those obtained at Bajondillo for
Bj/13–Bj/11 boundary
the Middle Palaeolithic–Upper Palaeolithic transition. In North
Africa, Upper Palaeolithic/Later Stone appear in Cyrenaica after
Bajondillo 14C samples

CNA−3218.1.2
CNA−3216.3.1
CNA−3873.1.1
CNA−3213.3.2
CNA−3882.1.2
Bj/14–Bj/13 boundaryy
CNA−3211.1.1
CNA−3881.1.1
CNA−3875.1.1
CNA−3880.1.1
50 45
cal ka BP
Fig. 3 | Comparison between the chronologies of the different
archaeological levels at Bajondillo Cave and a selection of
palaeoenvironmental proxies. a, Mediterranean sea surface temperatures
(SSTs) from Alboran Sea site ODP 977 (ref. 47) during Marine Isotope Stage
3. Blue bars show Heinrich events (H)48 and orange bars show Greenland
Interstadials (GI)17. b, δ​18O from the Greenland ice at the North Greenland
Ice Core Project (NGRIP) site at 20-year intervals obtained from the NGRIP
Greenland Ice Core Chronology 2005 annual-layer-counted chronology49.
c, Red Sea relative sea-level fluctuations, with data smoothing of a 1 kyr
moving Gaussian filter (black line) and maximum probability (orange line)50.
d, Calibrated radiocarbon dates from four archaeological levels of Bajondillo
Cave (see Supplementary Table 1). Age probability distributions are shown
in grey. Probability distributions of the Bayesian-modelled ages are shown in
black, with the 2σ​s.d. of the age uncertainty marked below and the median
value as a vertical line on the s.d. bar. The durations of the different phases
(darker intervals in each level) were calculated using the median values of
the Bayesian-modelled probability distribution for the boundaries.
the Carigüela dates, while Zafarraya has recently been reanal-
ysed, with its Middle Palaeolithic securely dated at >​43 cal ka bp22.
Accordingly, the youngest Mousterian dates from Abrigo 3 and
Zafarraya are ~43 cal ka bp, whereas for the Iberian Mediterranean
younger dates have been published for Cueva Antón (Murcia)
(>​37.1 cal ka bp12) and Gorham’s Cave (~32.5 cal ka bp11). Elsewhere
Bj/13
Proto-Early
~43 ka (Haua Fteah), but only after 30 ka in the Maghreb (Grotte
Aurignacian des Pigeons) (Fig. 1)8,12,17,27–34.
The new Bajondillo dates are crucial for several reasons. First,
they confirm the presence of a chronologically early Aurignacian
in southern Iberia at ~43 cal ka bp that now shows the first appear-
ance of the EUP to be an essentially synchronous event throughout
Bj/14 Europe (Fig. 3 and Supplementary Fig. 3). This suggests that the dis-
Mousterian
(Middle
persal of AMHs was much faster than hitherto postulated, and the
Palaeolithic) expansion of the earliest Aurignacians in Europe is now increased
westwards by >​1,000 km. These dates thus call into question both
the gradual ‘wave-of-advance’ and the ‘Ebro frontier’ models. They
40 35 30
also provide reference points for the attribution of early art work in
Iberian caves—something that remains highly controversial35. The
dates suggest either extremely high mobility of early Aurignacians
or well-developed networks of interchange. For early AMHs, rapid
dispersal was seemingly only possible over essentially ‘empty’ terri-
tories (that is, either completely depopulated areas, as some studies
suggest for southern Iberia at the time7, or areas featuring severely
depleted human populations).
The Aurignacian early expansion took place between Heinrich
events 5 and 4 (Fig. 3 and Supplementary Fig. 3), when cold and
steppic conditions prevailed throughout most of western Europe,
including the Iberian hinterland36. In light of this, it might be no
coincidence that a prevalence of European EUP sites has been
reported along shores or neighbouring lowlands where milder con-
ditions prevailed and more productive environments, in terms of
living resources, existed (Fig. 1 and Supplementary Fig. 2c). The
Aurignacian from Bajondillo Cave conforms with this pattern and
is not an isolated case in Iberia, where newly recorded Aurignacian
sites south of the Ebro river, such as Foradada16, Cendres17 and Pego
do Diabo12, are all located on the present-day coast or its adjoin-
ing lowlands (Fig. 1), nor in Italy (Riparo Mochi and Serino; see
Supplementary Information). Coasts and coastal lowlands as
instrumental for human dispersal and colonizing events are not a
phenomenon restricted to the European Aurignacian, since data
emerging from southern Arabia37, Australia38,39 and South America40
all point in the same direction. In eastern and southern Iberia, trav-
elling through coasts and coastal lowlands would have been con-
siderably easier than travelling through one of the most rugged

210 Nature Ecology & Evolution | VOL 3 | FEBRUARY 2019 | 207–212 | www.nature.com/natecolevol
Nature Ecology & Evolution
and mountainous hinterlands in Europe. The enhanced mobility
of Aurignacian populations can be inferred from their swift spread
over western Europe from Glacial Interstadial 12 onwards, thus
hinting at dispersals taking place across territories that were easy to
traverse (Supplementary Fig. 2a,b)1.
The onset of the Aurignacian cannot be detached from the
demise of the Neanderthals. Inferences about the mobility and
settlement patterns of southern Iberian Neanderthals are compli-
cated due to the restricted number of Mousterian sites, as well as
the prevalence of low sea-level stands during this period, which
means that an undetermined, yet probably substantial, fraction
of the evidence presently lies underwater. From this perspective
Bajondillo Cave is of great relevance since, due to its topogra-
phy, it was at all times located on or very close to the shore (that
is, not more than 5 km distant; Supplementary Fig. 2a), but never
flooded. The data from Bajondillo Cave and other coastal Iberian
locations41 reveal that stasis, as exemplified by around 120 kyr14 of
Neanderthal occupation with no clear traces of technological devel-
opments, prevailed during the Iberian Mousterian. Similar trends
are also documented—albeit in in a more restricted manner—at
the Abrigo 3 (ref. 21) and Gorham’s Cave sites11. All constitute evi-
dence that Neanderthals were settled along the coast well before
the onset of the Aurignacian. Although the Bajondillo Cave dates
do not indicate any coexistence of Neanderthals and Aurignacians,
the fact that the Middle Palaeolithic ceases at around 45 cal ka bp
(Supplementary Table 1) is also worth noting. Indeed, this date is
essentially synchronous with the cessation of Middle Palaeolithic
levels from sites in the province of Málaga, both coastal (Abrigo 3)21
and inland (Zafarraya)22. The idea of a regional, as opposed to local,
phenomenon seems compelling, as is the fact that (except for the
two Bay of Málaga sites) Late Mousterian settlements are located
at a substantial altitude (that is, >​400 m). A putative preferential
location of late Neanderthal sites on harsher and less productive
habitats than the coastal zones where older Middle Palaeolithic sites
occur and the earliest AMH sites are recorded hints at a scenario
of competitive displacement of Neanderthals by AMHs. This com-
plex issue is difficult to address when a substantial proportion of the
available evidence is circumstantial. In contrast, genetic data may
help confirm whether Neanderthal replacement was determined
by competition, migration and random species drift or if, as also
seems possible, AMHs and Neanderthal gene pools coalesced after
episodes of introgression42. However, it is notable that none of the
European late Neanderthal genomes sequenced so far shows evi-
dence of interbreeding with AMHs43.
Recent reviews highlighting the genetic complexities that under-
lie AMH expansion in Eurasia tend to focus on data from Asia, and
fail to consider the possibility that a crucial part of that evidence
may lie on the westernmost tip of the Mediterranean2. Given the
relevance of the palaeoanthropological and archaeological data that
are slowly emerging from this region, further efforts to determine
whether coastal dispersals played a major role in the arrival of mod-
ern humans in the region are crucial. Given the growing evidence
that humans were also capable of crossing bodies of water before
50 cal ka bp2, investigation into whether the Strait of Gibraltar
played any role as a connector of European Neanderthals and North
African AMHs is also a promising area of research. It is within this
interpretative framework that the presence of an early Aurignacian
at Bajondillo Cave may bear wider implications for the origin of
Upper Palaeolithic industries and the appearance of AMHs in the
European subcontinent than can be foreseen now.
Methods
The cultural definition of the archaeological sequence at Bajondillo has been
addressed from the technological standpoint of its lithic industries (materials
deposited in the Museum of Málaga)13. The ages of sedimentary fractions
from levels Bj/14–Bj/11 (Supplementary Table 1) were obtained from three
Nature Ecology & Evolution | VOL 3 | FEBRUARY 2019 | 207–212 | www.nature.com/natecolevol
Articles
laboratories (Ånström Laboratory at the University of Uppsala in Sweden, the
Centro Nacional de Aceleradores in Spain and the Dating and Radiochemistry
Laboratory of the Universidad Autónoma de Madrid in Spain) via a selection of
samples from thermo-altered and worked, chipped lithics and lithic industries
(thermoluminescence dating) and organic materials (14C/AMS). Whenever
possible, short-lived items from sealed contexts linked to human activities were
selected, such as terrestrial snail pulmonate samples. These specimens derive from
a garden snail taxon (Otala species) that feeds on fresh vegetal matter and thus
does not exhibit any ‘old-shell effect’. To remove surficial contamination of recent
carbonates, all dated shells were subjected to mechanical cleaning of their surfaces,
as well as acid and base leaching. OxCal 4.3 software (https://c14.arch.ox.ac.uk/
oxcal.html) IntCal13 and Marine13 curves were used to calibrate charcoals and
marine shells44,45. We selected the 2σ​(95.4%) probability distribution for age
uncertainties, and local variation of the reservoir age was not applied to marine
samples (see Supplementary Information for a detailed discussion). Subsequently,
a Bayesian model was applied, using the same software, to constrain the different
events in our archaeological sequence (that is, the start and end of different phases).
This model performs a multi-parameter Bayesian analysis using a Markov chain
Monte Carlo approach. We also selected the 2σ​(95.4%) probability distribution
here. The obtained model agreement index (Amodel) and individual agreement index
(Aoverall) were 91.4 and 93.4%, respectively, which improved the threshold values
(60%). To establish the boundary between the different archaeological phases,
we selected the median probability of the 2σ​phase-boundary modelled Bayesian
distribution.
Reporting Summary. Further information on research design is available in the
Nature Research Reporting Summary linked to this article.
Data availability
The described archaeological collections are housed at Museo Arqueológico de
Málaga, Spain.
Received: 25 January 2018; Accepted: 15 November 2018;
Published online: 21 January 2019
References
1. Hublin, J. J. The modern human colonization of western Eurasia: when and
where? Quat. Sci. Rev. 118, 194–210 (2015).
2. Bae, C. J. et al. On the origin of modern humans: Asian perspectives. Science
358, eaai9067 (2017).
3. Banks, W. E. et al. Neanderthal extinction by competitive exclusion. PLoS
ONE 3, e3972 (2008).
4. Stringer, C. et al. in Neanderthals and Moderns Humans in the European
Landscape During the Last Glaciation: Archaeological Results of the Stage 3
Project (eds van Andel, T. H. & Davies, W.) 233–240 (Univ. Cambridge,
Cambridge, 2003).
5. Giaccio, B. et al. High-precision 14C and 40Ar/39Ar dating of the Campanian
Ignimbrite (Y-5) reconciles the time-scales of climatic-cultural processes at
40 ka. Sci. Rep. 7, 45940 (2017).
6. Higham, T. et al. The timing and spatiotemporal patterning of Neanderthal
disappearance. Nature 512, 306–309 (2014).
7. Bicho, N. et al. Early Upper Paleolithic colonization across Europe: time and
mode of the Gravettian diffusion. PLoS ONE 12, e0178506 (2017).
8. Fu, Q. et al. The genetic history of Ice Age Europe. Nature 534,
200–205 (2016).
9. Peresani, M. et al. The Uluzzian technology of Grotta di Fumane and its
implication for reconstructing cultural dynamics in the Middle–Upper
Palaeolithic transition of Western Eurasia. J. Hum. Evol. 91, 36–56 (2016).
10. Douka, K. et al. On the chronology of the Uluzzian. J. Hum. Evol. 68,
1–13 (2014).
11. Finlayson, C. et al. Late survival of Neanderthals at the southernmost
extreme of Europe. Nature 443, 850–853 (2006).
12. Zilhão, J. et al. Precise dating of the Middle-to-Upper Paleolithic transition in
Murcia (Spain) supports late Neandertal persistence in Iberia. Heliyon 3,
e00435 (2017).
13. Cortés-Sánchez, M. in Cueva Bajondillo (Torremolinos). Secuencia
cronocultural y paleoambiental del Cuaternario reciente en la Bahía de Málaga
(ed. Cortés-Sánchez, M.) 93–138 (CEDMA, Málaga, 2007).
14. Cortés-Sánchez, M. et al. Earliest known use of marine resources by
Neanderthals. PLoS ONE 6, e24026 (2011).
15. Barroso Ruiz, C. & de Lumley, H. La grotte du Boquete de Zafarraya: Málaga,
Andalousie (Junta de Andalucía, Consejería de Cultura, Sevilla, 2006).
16. Morales, J. I. et al. Expanding the geography of the Middle to Upper
Palaeolithic transition: Foradada Cave (Calafell, Spain), a new site on the
Iberian Mediterranean coastline. Antiquity 351, 1–4 (2016).
17. Villaverde, V. et al. The early Upper Palaeolithic of Cova de les Cendres
(Alicante, Spain). Quat. Int. (in the press).
211
Articles
18. Tafelmaier, Y. Technological Variability at the Beginning of the Aurignacian:
Implications for the Proto- and Early Aurignacian Distinction (Neanderthal
Museum, Mettmann, 2017).
19. Bataille, G. et al. Living on the edge—a comparative approach for studying
the beginning of the Aurignacian. Quat. Int. 474, 1–98 (2018).
20. Albert, R. in Cueva Bajondillo (Torremolinos). Secuencia cronocultural y
paleoambiental del Cuaternario reciente en la Bahía de Málaga (ed.
Cortés-Sánchez, M.) 491–500 (CEDMA, Málaga, 2007).
21. Ramos Fernández, J. et al. Dating of the Middle to Upper Paleolithic
transition at the Abrigo 3 del Humo (Málaga, Spain). Mainake XXXIII,
275–284 (2012).
22. Wood, R. E. et al. Radiocarbon dating casts doubt on the late chronology of
the Middle to Upper Palaeolithic transition in southern Iberia. Proc. Natl
Acad. Sci. USA 110, 2781–2786 (2013).
23. Douka, K. et al. A new chronostratigraphic framework for the Upper
Palaeolithic of Riparo Mochi (Italy). J. Hum. Evol. 62, 286–299 (2012).
24. Nigst, P. R. et al. Early modern human settlement of Europe north of the
Alps occurred 43,500 years ago in a cold steppe-type environment. Proc. Natl
Acad. Sci. USA 111, 14394–14399 (2014).
25. Marin-Arroyo, A. et al. Chronological reassessment of the Middle to Upper
Paleolithic transition and Early Upper Paleolithic cultures in Cantabrian
Spain. PLoS ONE 13, e0194708 (2018).
26. Wolf, D. et al. Climate deteriorations and Neanderthal demise in interior
Iberia. Sci. Rep. 8, 7048 (2018).
27. Wood, R. E. et al. The chronology of the earliest Upper Palaeolithic in
northern Iberia: new insights from L’Arbreda, Labeko Koba and La Viña.
J. Hum. Evol. 69, 91–109 (2014).
28. Higham, T. F. G. et al. The earliest evidence for anatomically modern humans
in northwestern Europe. Nature 479, 521–524 (2011).
29. Zilhão, J. Pego do Diabo (Loures, Portugal): dating the emergence of
anatomical modernity in westernmost Eurasia. PLoS ONE 5,
e8880 (2010).
30. Higham, T. F. G. et al. Testing models for the beginnings of the Aurignacian
and the advent of figuratuve art and music: the radiocarbon chronology of
Geißenklösterle. J. Hum. Evol. 62, 664–676 (2012).
31. Douka, K. et al. The chronostratigraphy of the Haua Fteah cave (Cyrenaica,
northeast Libya). J. Hum. Evol. 66, 39–63 (2014).
32. Barton, R. N. E. et al. Rethinking the Human Revolution: New Behavioural and
Biological Perspectives on the Origins and Dispersal of Modern Humans (eds
Mellars, P., Boyle, K., Bar-Yosef, O. & Stringer, C.) 177–186 (McDonald
Institute, Cambridge, 2007).
33. Ramos, J. et al. The Benzú rockshelter: a Middle Palaeolithic site on the
North African coast. Quat. Sci. Rev. 27, 2210–2218 (2008).
34. Hublin, J.-J. et al. New fossils from Jebel Irhoud, Morocco and the pan-
African origin of Homo sapiens. Nature 546, 289–292 (2017).
35. Pearce, D. G. & Bonneau, A. Trouble on the dating scene. Nat. Ecol. Evol. 2,
925–926 (2018).
36. Burke, A. et al. Risky business: the impact of climate and climate variability
on human population dynamics in Western Europe during the Last Glacial
Maximum. Quat. Sci. Rev. 164, 217–229 (2017).
37. Groucutt, H. S. et al. Homo sapiens in Arabia by 85,000 years ago. Nat. Ecol.
Evol. 2, 800–809 (2018).
38. Clarkson, C. et al. Human occupation of northern Australia by 65,000 years
ago. Nature 547, 306–310 (2017).
39. Veth, P. et al. Early human occupation of a maritime desert, Barrow Island,
North-West Australia. Quat. Sci. Rev. 168, 19–29 (2017).
40. Llamas, B. et al. Ancient mitochondrial DNA provides high-resolution time
scale of the peopling of the Americas. Sci. Adv. 2, e1501385 (2016).
212 Nature Ecology & Evolution | VOL 3 | FEBRUARY 2019 | 207–212 | www.nature.com/natecolevol
Nature Ecology & Evolution
41. Haws, J. A. et al. Coastal wetlands and the Neanderthal settlement of
Portuguese Estremadura. Geoarchaeology 25, 709–744 (2010).
42. Kolodny, O. & Feldman, M. W. A parsimonious neutral model suggests
Neanderthal replacement was determined by migration and random species
drift. Nat. Comm. 8, 1040 (2017).
43. Mateja Hajdinjak et al. Reconstructing the genetic history of late
Neanderthals. Nature 555, 652–656 (2018).
44. Reimer, P. J. et al. IntCal13 and Marine13 radiocarbon age calibration curves
0–50,000 years cal bp. Radiocarbon 55, 1869–1887 (2013).
45. Ramsey, C. B. Bayesian analysis of radiocarbon dates. Radiocarbon 51,
337–360 (2009).
46. Ehlers, J., Gibbard, P. L. & Hughes, P. D. Quaternary Glaciations—Extent and
Chronology: A Closer Look Vol. 15 (Elsevier, Amsterdam, 2011).
47. Martrat, B. et al. Four climate cycles of recurring deep and surface water
destabilizations on the Iberian Margin. Science 317, 502–507 (2007).
48. Sanchez-Goñi, M. F. & Harrison, S. P. Millennial-scale climate variability and
vegetation changes during the Last Glacial: concepts and terminology. Quat.
Sci. Rev. 29, 2823–2827 (2010).
49. Rasmussen, S. O. et al. A stratigraphic framework for abrupt climatic changes
during the Last Glacial period based on three synchronized Greenland
ice-core records: refining and extending the INTIMATE event stratigraphy.
Quat. Sci. Rev. 106, 14–28 (2014).
50. Grant, K. M. et al. Rapid coupling between ice volume and polar temperature
over the past 150,000 years. Nature 491, 744–747 (2012).
Acknowledgements
M.C.-S. was provided access by the Consejería de Cultura of the Junta de Andalucía
(Spain) to analyse the Bajondillo Cave fauna (UPPH/49/06). The research was sponsored
by grants HAR2013-44269-P and HAR 2016-77789-P from the Spanish Ministerio de
Economía y Competitividad. A.G.-A. acknowledges a Ramón y Cajal Fellowship (RYC-
2015-18966) of the Spanish Government (Ministerio de Economía y Competitividad).
The research of C.S. is supported by the Calleva Foundation and Human Origins
Research Fund. The paper constitutes contributions from the HUM-949 Research Group
(Universidad de Sevilla, Spain) and ICArEHB (University of Algarve, Portugal).
Author contributions
M.C.-S., F.J.J.-E., A.M.-M. and C.S. conceived and designed the experiments
and wrote the manuscript. All authors analysed the data. M.C.-S., M.D.S.-V.,
C.P.O. and J.A.R.-C. performed the archaeology and lithic technology. A.M.-M.
performed the archaeozoology. R.P.G., A.M.G. and F.J.J.-E. produced the figures and
palaeoreconstruction. A.G.-A. and N.O. performed the geochronology. M.C.L.F. and
J.L.V.P. performed the archaeomalacology.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/
s41559-018-0753-6.
Reprints and permissions information is available at www.nature.com/reprints.
Correspondence and requests for materials should be addressed to F.J.J.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature Limited 2018
 nature research | reporting summary
Corresponding author(s): Francisco J. Jiménez-Espejo

Reporting Summary
Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency
in reporting. For further information on Nature Research policies, see Authors & Referees and the Editorial Policy Checklist.

Statistical parameters
When statistical analyses are reported, confirm that the following items are present in the relevant location (e.g. figure legend, table legend, main
text, or Methods section).
n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement
An indication of whether measurements were taken from distinct samples or whether the same sample was measured repeatedly
The statistical test(s) used AND whether they are one- or two-sided
Only common tests should be described solely by name; describe more complex techniques in the Methods section.

A description of all covariates tested

A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons
A full description of the statistics including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND
variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)

For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted
Give P values as exact values whenever suitable.

For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes
Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated

Clearly defined error bars

State explicitly what error bars represent (e.g. SD, SE, CI)

Our web collection on statistics for biologists may be useful.

Software and code

Policy information about availability of computer code
Data collection Authorized database: ASTER GDEM v2 Worldwide Elevation Data (1 arc-second Resolution) (https://gdex.cr.usgs.gov/gdex/); EMODnet
Bathymetry data (7.5 arc-second Resolution) (http://portal.emodnet-bathymetry.eu/?menu=19); Alboran Sea bathymetry (100 m
resolution) Instituto Español de Oceanografía (IEO)-Secretaría General de Pesca Marítima and Gibraltar Strait (25 m resolution) Sociedad
Española de Estudios para la Comunicación Fija a través del estrecho de Gibraltar, SA (SECEGSA).

Data analysis Software Global Mapper v.16, Oxcal 4.3 software,

For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors/reviewers
upon request. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information.
April 2018

1
Data

nature research | reporting summary

Policy information about availability of data
All manuscripts must include a data availability statement. This statement should provide the following information, where applicable:
- Accession codes, unique identifiers, or web links for publicly available datasets
- A list of figures that have associated raw data
- A description of any restrictions on data availability
Described archaeological collections are housed at Museo Arqueológico de Málaga, Spain. Publisher’s note: Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Field-specific reporting
Please select the best fit for your research. If you are not sure, read the appropriate sections before making your selection.
Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/authors/policies/ReportingSummary-flat.pdf

Ecological, evolutionary & environmental sciences study design

All studies must disclose on these points even when the disclosure is negative.
Study description Based on Bayesian analyses, a total of 26 radiocarbon dates, including 17 new ones, show that replacement at Bajondillo took place
in the millennia centering on ~45-43 cal ka BP, well before the onset of Heinrich event 4 (~40.2 to 38.3 cal ka BP).

Research sample Different marine shells, Burnt flint and Carbonaceous/ash sediments.

Sampling strategy In archeological studies one or two radiocarbon dates are enough in order to obtain the age of one archeological level. In this studie
we present a total of 26 radiocarbon ages, and key archeological levels has been dated by more than 4 radiocarbon dates and these
performed at different laboratories and also used different methods.

Data collection Samples were collected by Dr. Miguel Cortes-Sanchez during archeological excavations.

Timing and spatial scale 49048 to 31540 yrs Cal BP from Bajondillo Cave, compared with South Iberian archeological sites.

Data exclusions Yes, we obtained one abnormal radiocarbon date not included in the manuscript, but can be included if necessary. Radiocarbon date
correspond to archeological level Bj/13 and age obtained is 8.2 ka cal BP (samples details: Mytilidae; Laboratory code: CNA-3215.2.2;
radiocarbon age 7631+42 Calibrated age 8253-7980). We interpretated that this matherial was contaminated by contact with
present day radiocarbon matherial during sampling or laboratory treatments.

Reproducibility We include a detail figure that indicate where samples were taken with high resolution. Archeological samples and levels still present
and a re-sampling is posible.

Randomization We perfomed analyses in different organism and matherials using different tecniques in order to get a robust chonological control.

Blinding We used three different blinded laboratories in order to obtain a robust age control and avoid single laboratory methodological
byass. Sample selection tried to be equidistance along each archological level.

Did the study involve field work? Yes No

Field work, collection and transport

Field conditions Samples were taken during one month during summer.

Location 36.622694 N 4.496608 W Bajondillo Cave is a ca. 30 m long rock shelter that opens within a 30 m high travertine formation in
the city of Torremolinos

Access and import/export Bajondillo Cave is located in a cliff and required scaffold and safety infraestructure.
April 2018

Disturbance Five cubic meter of sediments were excaved

Reporting for specific materials, systems and methods

2
 nature research | reporting summary
Materials & experimental systems Methods
n/a Involved in the study n/a Involved in the study
Unique biological materials ChIP-seq
Antibodies Flow cytometry
Eukaryotic cell lines MRI-based neuroimaging
Palaeontology
Animals and other organisms
Human research participants

Palaeontology
Specimen provenance The Bajondillo cave faunal analyses were allowed to one of the authors (MC-S) by the Consejería de Cultura of the Junta de
Andalucía (Spain) (UPPH/49/06).

Specimen deposition Described archaeological collections are housed at Museo Arqueológico de Málaga, Spain

Dating methods Different faunal remains (marine shells) and charcoal fragments were selected from different archeological levels and preventing
contamination from modern carbon (Radiocarbon dates) or light (thermoluminescence dates). A chemical pretreatment for
Radiocarbon Dating samples acid-base-acid (ABA) was performed. Laboratory codes used were Ua: Ånström laboratory,
University of Uppsala (Sweden); CNA: National Center for Accelerators (Spain); MAD: Dating and Radiochemistry Laboratory,
Universidad Autónoma de Madrid. Radiocarbon has been calibrated using Oxcal 4.3 software (https://c14.arch.ox.ac.uk/
oxcal.html) along with Intcal13 and marine13 curves (Reimer et al., 2013; Ramsey, 2009). The local variation of the reservoir age,
estimated from recent samples, in the westernmost Mediterranean is 280+36 yr (Siani et al., 2000). However, it is not applied to
the calibration of marine samples from Bajondillo site because this value is unknown for the Mediterranean Sea during from 50
to 20 ky (glacial period), and apply the present reservoir effect have a nil effect on the obtained calibrated values.

Tick this box to confirm that the raw and calibrated dates are available in the paper or in Supplementary Information.

April 2018