CENTRAL PLATEAU 517
1. Discovery and early research
The Torralba and Ambrona sites (Santonja et
al., in: Santonja and Pérez-González, 2005 (eds.):
18-39) are 150 kilometres north east of Madrid on
the watershed between the Ebro, Duero and Tagus
Rivers, in the southern part of Soria Province. This
is a strategic transit zone between the highlands
of the Iberian Meseta (plateau) and the Jalón River
Valley. An oblique perspective generated by a
digital terrain model shows the two sites between
reliefs drained by tributaries of the Atlantic slope
Duero and Tajo Rivers. To the east, the clear outline
of the Jalón River canyon, which flows into
the Ebro and eventually into the Mediterranean
(Fig. 1).
1.1. First work by Marquis of Cerralbo (1909-1916)
The discovery of both sites began with the detection
of large elephant bones in 1888 at the Torralba
railway station. Between 1909 and 1913, Enrique de
Aguilera y Gamboa (1845-1922), the 17th
Marquis
of Cerralbo, excavated over 2000 m2
in Torralba.
From 1914 to 1916, he continued his work at the
Loma de los Huesos in Ambrona site, 2.5 km to the
north. His results had a great impact at the time and
drew visits by leading researchers. Cerralbo’s conclusions,
paradoxically from a creationist’s ideological
perspective (Santonja and Vega, 2002), conjugated
the key aspects around which the site was later interpreted
–organised hunting of elephant herds by a
group of human settlers on the banks of a lake– and
provided a glimpse of the potential importance of
these sites for the study of human behaviour (Isaac,
1977: 3-4).
Ambrona and Torralba archaeological
and palaeontological sites, Soria Province
Manuel Santonja*,Alfredo PérezGonzález*,
Joaquín Panera**, Susana
Rubio-Jara**, Carmen Sesé***, Enrique
Soto***, Laura Sánchez-Romero*, ****
1.2. Research resumed by F.C. Howell (1960-1963)
In Spain after the Civil War, in 1936, Lower
Palaeolithic research was abandoned almost entirely.
Only occasional visits to international congresses
awakened memories of Torralba and Ambrona
and spurred individual initiatives such as the
palynological studies by J. Menéndez Amor and F.
Florschütz in 1959 and 1963, which focused on
the Middle Pleistocene chronology of the sites
(Santonja and Vega, 2002).
Contacts between Luis Pericot and Clark Howell
at the Panafrican Prehistory Congresses led to
the recommencement of the research work, nearly
50 years later of the Cerralbo’s work. Howell proposed
systematic full-cover, multidisciplinary excavations
for Torralba and Ambrona. By the end of
the decade, research into the African Pleistocene
had implanted this model, but in 1960 it was a
novelty in Europe.
When this work began, the Torralba site was
estimated to cover approximately 3800 m² (Howell
et al., 1962), of which over 2000 m2
were to
be preserved, of which 1026 m2
dug between
1961 and 1963 (González Echegaray and Freeman,
1998). Loma de los Huesos in Ambrona was
estimated to cover roughly 6000 m2
, of which
1243 m2
were excavated in 28 weeks during the
1962 and 1963 digs (Howell, 1965). Both sites
were studied by the same team –K.W. Butzer, E.
Aguirre, P. Biberson and L.G. Freeman– who used
a similar methodology and reached common conclusions.
Butzer’s geological survey attributed an
identical age and formation processes to both
sites and defined a morphosedimentary unity, the
“Torralba Formation”, which integrated the strati*
Centro Nacional de Investigación sobre la Evolución Humana. (CENIEH); 09002 Burgos manuel.santonja@cenieh.es
** IDEA (Instituto de Evolución en África), Museo de los Orígenes, Plaza de San Andrés 2, 28005 Madrid.
*** Departamento de Paleobiología. Museo Nacional de Ciencias Naturales. Madrid.
**** Escuela Interuniversitaria de Posgrado en Evolución Humana, Universidad de Burgos. C/ Juan de Austria 1, 09001 Burgos.
PLEISTOCENE AND HOLOCENE HUNTER-GATHERERS IN IBERIA AND THE GIBRALTAR STRAIT:
THE CURRENT ARCHAEOLOGICAL RECORD
518
graphic sequences of the two sites, described at
the time as twin (Butzer, 1965).
The published studies of the lithic tools, all
preliminary, interpreted the industry at either
sites as early (Freeman, 1975) or middle (Biberson,
1964) Acheulean. Biberson and Aguirre also
noted the existence of worked bone, an issue debated
subsequently and pending a systematic review
(Domínguez-Rodrigo, in Santonja and PérezGonzález
(eds.), 2005: 282-287).
This first research stage led to an interpretation
of the sites which in some general aspects matched
Cerralbo’s imaginative foresight. The Mansegal or
La Mentirosa stream valley connected the highlands
plateau of the North with the Ebro Basin, and
was probably a corridor frequented by herbivores
during their seasonal migration (Butzer, 1971).The
presence of these herds led groups of hominins –in a
display of premonitory behaviour– to burn the vegetation
in order to drive them into swampy zones
where weeds and mud hampered the animals’
movement. In these conditions would have been
easy to kill them, dismembered in nearby spaces
and prepared for consumption (Howell, 1966).
1.3. The Howell-Freeman period (1980-1983)
The complete aperture of East Africa to Pleistocene
research led Howell to interrupt his Spanish
work in 1963. The large interdisciplinary teams
which started to work at African sites proved decisive
and triggered profound changes to the methods
used in Palaeolithic archaeology from the 1970’s onwards.
In this context, the interpretations of Torralba
and Ambrona were reviewed by Binford, who found
no arguments in support of organized hunting, expressed
doubts about the presence of fireplaces and
questioned whether the areas containing bones had
Figure 1. Geographic location of the Torralba and Ambrona sites in the south of Soria province, 150 km NE of Madrid
(Spain). Both sites are in the valley of La Mentirosa arroyo Rivulet, also known as Mansegal, a tributary of the Jalón River
which flows into the Ebro. The Bordecorex stream flows into the Duero River, while south of Olmedillas and Ventosa del
Ducado, the river network flows into the Henares River (Tagus basin).
CENTRAL PLATEAU 519
remained in a primary position. Even in the absence
of different data from those published by Howell
and his team, Binford suggested that natural agents
and trampling by elephants had caused intense modifications
(Binford, 1987).
The debate was in full swing when a new phase
of research at Ambrona began in 1980. With the
addition of 207 m2
dug by E. Aguirre in 1973, by
the start of the 1980 season, almost 1450 m2
had
been excavated at Ambrona. During this stage, under
the joint leadership of Howell and Freeman
and the management of M. Almagro –responsible
for channelling relations with the Spanish government
and facilitating funding for the project under
the Spain-US Cultural Cooperation Programme–,
an additional 1267 m2
were dug in 203 days. By
1983, 2717 m2
of the estimated 6000 m2
of the
Ambrona site had been excavated. In addition,
work continued on the hillside opposite Loma de
los Huesos (Camp North). Here, although published
references are quite vague (Howell and
Freeman, 1982), we know that a little over 200 m2
were dug, 55m2
in 1963 and 162 m2
in 1981 and
1983, according to unpublished documents held
at the Numantine Museum in Soria.
The Ambrona assemblage remained attributed to
Butzer’s “Torralba Formation”, with new aspects in
the stratigraphic interpretation (Howell et al., 1995),
primarily the differentiation of two members, the
“Lower complex” and the “Upper complex”. The
former included the characteristic concentrations of
megafauna –particularly elephant– and Acheulean
industry.An intermediate occupation was defined in
the central part of the site, with fauna and sporadic
industry considered close or equivalent to Camp
North, where deer, aurochs and elephant were recorded
along with some Acheulean handaxes.At the
“Upper Complex”, on the levels of alluvial and colluvial
origin, a more frequent and more evolved lithic
industry than the Lower complex was detected,
also identified as Acheulean. Scarcely any elephant
remains were detected, replaced in importance by
Equus (Howell et al., 1995).
At the end of this stage, the older interpretations
of Ambrona’s Lower complex were accepted
with certain nuances. The faunal remains were basically
regarded as residue from deliberate hunting
activity and the processing and consumption
of the food. The hominins had also shifted substantial
portions of this prey to their base camps,
whose location was imagined –in the absence of
evidence– to be on high ground overlooking the
area, “overlooking the valley” (Freeman, 1994).
2. Current state of research at Torralba and Am-
brona
2.1. Digs from 1990 to 2000 and subsequent work
The controversy over the nature of these sites
continued into the late 1980’s, and the published
information was still insufficient to be able to test
the hypotheses proposed in the light of the previous
excavations. Substantial unknowns still remained
about the general sedimentary processes
and also the microstratigraphy and the spatial distribution
of the remains.
It was known for certain that there were still
large unexcavated areas in Ambrona and probably
in Torralba as well. Consequently, in 1990 –following
an excavation permit granted to E. Carbonell
in 1988 which was not continued and there are no
known results– another project headed by M. Santonja
and A. Pérez-González was begun using geoarchaeological,
taphonomic, and techno-economic
methods. In summary, its aims were to understand
the morphodynamic and sedimentary processes in
order to contextualize the megafauna assemblages
and interpret the human activity in this area (Santonja,
1989).
An initial stage between 1990 and 1991 ascertained
the basic features of the local geomorphological
evolution and situated the sites in a geological
framework at the local and regional scale
(Pérez-González et al., 1991).
Annual digs were conducted at Ambrona from
1993 to 2000, focused on the “Lower complex”,
with 688 m2
excavated. In Torralba, work was
hampered by the large volume of waste material
from previous excavations dumped on the site itself,
and thus only limited test pits were dug (Santonja
et al., in Santonja and Pérez-González (eds.),
2005: 104-123).
In 2001, a vertical electrical sondage was lowered
into a sinkhole in Jurassic dolomites 200 m
NE of Ambrona, where Pleistocene fauna was recorded.
In 2001 and 2002, samples were removed
for magnetostratigraphy as well as dating by luminescence
and aminoacid racemisation (Parés et
al., in Santonja and Pérez-González (eds.), 2005:
190-199; Falguères et al., 2006).
Finally, in 2013 the National Human Evolution
Research Centre (CENIEH) began fresh digs focused
initially on the middle stratigraphic member –partly
equivalent to the Upper Complex of Howell (PérezGonzález
et al., in: Santonja and Pérez-González,
2005 (eds.): 176-199)–, in Ambrona and Torralba.
PLEISTOCENE AND HOLOCENE HUNTER-GATHERERS IN IBERIA AND THE GIBRALTAR STRAIT:
THE CURRENT ARCHAEOLOGICAL RECORD
520
González et al., in Santonja and Pérez-González
(eds.), 2005: 176-199). A cut and fill phase of the
Mansegal stream, and another of the +35 m terrace
development and subsequent incision on this terrace
separate the two sites (Fig. 2b), categorically
invalidating hypotheses which merged their stratigraphies
in a single unified sequence (Butzer, 1965).
The chronological distance between Ambrona
and Torralba has also been checked by numerical
Figure 2. A: Stratigraphy of Ambrona, lower and middle members, in the central sector of Loma de los Huesos. B: Relative
geomorphological position of the Torralba and Ambrona archaeological sites.
2.2. Results from Ambrona and Torralba sites between
1990 and 2000
2.2.1. Geomorphology and chronology of the sites
Ambrona is at the bottom of the karst valley or
polje between Torralba, Ambrona and Conquezuela,
while Torralba is in the valley of the Mansegal,
built up from the bottom of the polje and set
on the + 35 m terrace of the current valley (Pérez-
CENTRAL PLATEAU 521
dating. The middle stratigraphic unit of Ambrona
has been dated at c. 350 ky by ESR/U series (Falguères
et al., 2006), suggesting an age of around
400 ky for the lower unit of this site. In the case
of Torralba, OSL datings currently in press (N.
Mercier) indicate around 200 ky, similar to the
220 ky and 240 for U series (Howell et al., 1995),
obtained on a +20-25 m terrace (comparable to
Torralba) in the nearby Alto Henares area. This
reinforces the geomorphological interpretation
which refutes the contemporaneity of the sites
and instead suggests a sequence of occupations in
the area. The Torralba site at the bottom of the
Mansegal River valley is close to MIS 7 (243-192
ky), while the lower and middle stratigraphic units
of Ambrona correspond to MIS 11 (424-375 ky)
and MIS 10 (374-338 ky) or 9 (337-301 ky) respectively,
in positions related to the small ponds
and the drainage network which developed at the
bottom of the Conquezuela polje.
2.2.2. Ambrona stratigraphy
The stratigraphic units defined in Ambrona
correspond to fluvial and fluvio-lacustrine environments
(Pérez-González et al., in: Santonja and
Pérez-González, 2005 (eds.): 176-199), each one
with a different preservation potential for remains.
A systematic, integrated stratigraphic interpretation
has been proposed for the site, which we
have called “Ambrona formation”, consists of the
lower, middle and upper members. Excavation
campaigns from 1993 to 2000 were carried out
in the lower member, that includes the following
levels (Fig. 2a):
– A fluvial (AS1) and another fluvio-lacustrine
level (AS4) which contains the highest
density of lithic and faunal remains, albeit
in a derivative position, partly brought from
outside the preserved site.
– Level AS3, composed of mud built up at the
bottom of a shallow pond with occasionally
input via small channels. This level has the
best conservation conditions for industry
and fauna remains, mainly found in a primary
position.
– Other minor levels in the lower stratigraphic
unit are of a fluvial nature (AS1/2 and
AS2). At the top of this unit is the fluvio-lacustrine
level AS5.The presence of archaeopalaeontological
remains is sporadic in all
these cases.
All these levels were mapped, defining the vertical
and horizontal relationships of the stratigraphy
across more than 400 linear metres. While
AS1, AS5 and AS6 are spread across almost the
entire site, other levels cover smaller areas, resulting
in different stratigraphic sequences in each
zone. Thus, in the Central and Western sectors
there are several areas where AS1/2, AS2 and AS3
are absent and the AS4 level lies directly in contact
with AS1.
The middle member of the Ambrona formation
(AS6) includes fluvio-lacustrine deposits in
the Central and Western Sectors and fluvial deposits
in the Eastern Sector. The latter are rich
in lithic industry and they also contain fauna. In
some cases they occur in overbank facies, accumulated
in low energy conditions which enabled the
remains be found in an almost primary position.
The Ambrona formation is completed by the upper
member (AS7), composed of channelled and
edaphized facies. It is archeologically and palaeontologically
sterile.
2.2.3. Ambrona palaeontology
The results from the 1990-2000 stage almost
coincide with those obtained previously, with
some further details, particularly about birds and
small vertebrates. The mammal association identified
between 1993 and 2000 (Sesé and Soto, in:
Santonja and Pérez-González 2005 (eds.): 258-
281) is the following: Crocidura sp.; Microtus
(Iberomys) brecciensis; Arvicola aff. sapidus; Apodemus
aff. sylvaticus; Oryctolagus sp.; Canis lupus
cf. mosbachensis; Panthera (Leo) cf. fossilis; Palaeoloxodon
antiquus; Stephanorhinus hemitoechus;
Equus caballus torralbae; Cervus elaphus; Dama
cf. dama; Capreolus sp. and Bos primigenius. Other
taxa identified previously in the lower Ambrona
member must be added to this list (Howell et al.,
1995): Vulpes sp., Crocuta crocuta aff. praespelaea
and Megaloceros aff. savini.
In the middle member, almost the only species
recorded to date from the 1993-2000 excavations
is Equus caballus and very occasional remains of
Palaeoloxodon antiquus.
The presence of avian fauna is recorded
(Sánchez Marco, in: Santonja and Pérez-González
2005 (eds.): 248-257): Anser anser; Tadorna ferruginea;
Tadorna sp.; Anas acuta; Anas strepera; Anas
sp.; Mergus merganser; Anseriformes indet.; Fulica
cf. atra; Otis tarda and Vanellus vanellus; A herpetological
sample (Martínez-Solano and Sanchiz,
PLEISTOCENE AND HOLOCENE HUNTER-GATHERERS IN IBERIA AND THE GIBRALTAR STRAIT:
THE CURRENT ARCHAEOLOGICAL RECORD
522
in: Santonja and Pérez-González 2005 (eds.) : 232-
239) includes: Bufo bufo; Bufo calamita; Discoglossus
cf. jeanneae; Hyla arborea; Pelobates cultripes;
Pelodytes punctatus; Rana perezi; Rinechis scalaris;
Natrix sp.; Lacertidae and Colubridae indet.There
was also some ichthyologic remains (Perea and
Doadrio, in: Santonja and Pérez-González 2005
(eds.): 240-247) of Chondrostoma arcasii.
The macromammal series is considered characteristic
of the advanced but not final Middle Pleistocene,
with more modern elements than the peninsular
faunas from early Middle Pleistocene sites
such as Cúllar de Baza I (Granada) or Buenavista,
Campo de Tiro and Polígono Industrial (Toledo),
characterized by Mammuthus trogontherii. The micromammals
show a similar chronology. The morphology
and size of Microtus (I.) brecciensis teeth
imply a previous age to those found in late Middle
Pleistocene populations of the same species. The
presence of a relatively large form of the Arvicola
aff. sapidus species indicates a more modern phase
than the fauna from Cúllar de Baza I, with Arvicola
mosbachensis. The evolutionary stage of the Ambrona
microfauna corresponds to the third ensemble
of Middle Pleistocene associations defined by Sesé
and Sevilla (1996), which include fauna such as
that found in Áridos.
2.2.4. Palaeoenvironments
In addition to the study of the fauna and nanofauna
–ostracods–, palynology and biomineralizations
–phytoliths– have provided information about
the environmental conditions when the deposits of
Ambrona formation were accumulated (Baltanás
et al., 2005; Ruiz Zapata et al., 2005; Pinilla et al.,
2005, in Santonja and Pérez-González (eds.), 2005:
200-231).All the conclusions indicate the existence
of climatic constants which were comparable to the
current conditions, albeit with certain nuances.
The macromammals from the lower member
of Ambrona indicate a good representation of forest
environments and open lands, with meadows
and areas with abundant water in relatively warm
and moist climatic conditions. Birds confirm the
proximity of wetlands, flooded zones and shallow
water bodies. Some species like the common goose
and lapwing denote flat or gently undulating grassland.
Taxa such as swamphen, coot and northern
pintail require thick patches of vegetation around
water bodies. The swamphen, a sedentary animal,
is incompatible with very low temperatures. The
herpetofauna corroborates these interpretations
and indicates that the conditions were similar to
today, with a more or less contrasted seasonality,
less dry summers and less wet springs and winters,
with slightly higher winter temperatures.
Locally, the presence of ostracods as Leucocythere
cf. mirabilis in several levels indicates a
lake system in oligotrophic conditions. The taphocenosis
found in the ostracods is similar to current
conditions in shallow ponds and lakes in southern
Europe. Sometimes, Heterocypris salina became
predominant at Level AS6 (middle member), suggesting
a drier and colder period, a trend also suggested
by the dominion of Equus caballus in replacement
of Palaeoloxodon antiquus.
The silicophytoliths in the lower member suggest
a temperate climate. The diatoms are often
epiphytes, indicating a frequency of aquatic plants.
Biominerals are less abundant at the top of the
lower member due to the changed environmental
conditions. Silicophytoliths are more abundant,
but with many spicules and diatoms reduced to
the Amphora genus, reflecting quite stressful environmental
conditions for microorganisms. The
reduction in the number and variety of biominerals
is greater in the middle member: virtually
all are silicophytoliths and almost all C3 grasses.
Hantzschia amphioxys, a species that can survive
in a wide range of saline environments, predominates
amongst the diatoms. There are almost no
biominerals at the top of AS6, suggesting that the
lake may have dried up.
Pollen analysis describes vegetation mainly
consisting of pine (Pinus), juniper/sabine (Juniperus)
and grasses (Poaceas). The riparian taxa such
as alder (Alnus), willow (Salix), elm (Ulmus) and
characteristic swamp species (Cyperaceas, Ranunculaceas,
Typha...) were present. Temperate trees
such as deciduous oak (Quercus), birch (Betula),
chestnut (Castanea), hazel (Corylus) and walnut
(Juglans) were also detected. These results corroborate
the predominance of a milder climate than
today’s conditions at the time of the lower member
accumulation.
2.2.5. Human presence
The technical characteristics of the lithic industry
and the sources of the raw material have
fostered debate about the mode and intensity of
the site’s use. With regard to bone industry, recent
studies (Domínguez-Rodrigo, in: Santonja
and Pérez-González, 2005 (eds.): 282-287) do
not support the hypothesis suggested primarily by
CENTRAL PLATEAU 523
Aguirre and Biberson about an intense transformation
of elephant bones by shaping and retouch.
The possibility that the fragments corresponding
to the tip of infantile elephant tusks found at both
sites deriving from any kind of manufacturing has
also been rejected (Villa and d’Errico, in: Santonja
and Pérez-González, 2005 (eds.): 288-305), arguing
that they broke off naturally in the course of
the elephants’ lives.
To understand the significance of the Ambrona
lithic industry, we must take into account its taphonomic
history (Santonja et al., in: Santonja and Pérez-González,
2005 (eds.): 306-333).AS1 and AS4,
the richest levels in the lower member, are fluvial
deposits with a degree of energy. The industry they
contain is not in a primary position, having been
dragged from its original positions in the immediate
vicinity of the site. This material and the fauna
found on the same levels was carried and classified
by size by the watercourse. The technological imbalance
in the series from AS1 and AS4 are not due
to any palaeoeconomic or functional factors but
rather to the natural process which formed record.
OnAS3, however, the industry is mainly in a primary
position, albeit with some items deposited along
the small streams leading to the pond. It is very
low-density and essentially contains final elements
of chaînes opératoires with little or no shaping, such
as non-retouched flakes selected by size and form,
and also bifacial macrotools brought from outside
and left on the site.
EXCAVATED AREAS IN THE LOWER STRATIGRAPHIC
MEMBER OF AMBRONA IN 1993-
2000
LITHIC INDUSTRY
Level AS1: 535m2
. Only 35m2
complete. Only the
surface of the level in the rest
235 items, including 9 handaxes and 5 tools.
Density in 35 m2
: 5 items/m2
, 1 handaxe/5 m2
and
1.25 tools/m2
(= 5 items/4 m2
)
Level AS3: 250 m2
72 items, including 2 handaxes and 17 tools
Density in 250 m2
: 1 item /3.5 m2
; 1 handaxe/
125 m2
and 1 tool/15 m2
Level AS4: 379 m2
339 items, including 1 handaxe and 56 tools.
Density in 379 m2
: nearly 1 item/m2
;1 handaxe in
379 m2
and 1 tool/7 m2
Table 1. Excavated areas and lithic industry in the Lower Stratigraphic Member of Ambrona.
The industry at AS1 and AS3 can definitely be
ascribed to the Acheulean technocomplex, as with
the other levels in the lower stratigraphic member,
given that in AS4, hydraulic factor is responsible
for the deficit in medium and large format of lithic
items, which explains the absence of bifacial tools.
The industry in AS6, previously defined as a
more advanced Acheulean type than the lower
levels (Howell et al., 1995), is characterized by
the lack of true handaxes and cleavers, the development
of retouched tools on flake and the presence
of Levallois debitage (Fig. 3).This level corresponds
to the early Middle Palaeolithic (Santonja
and Pérez-González, 2006).
2.2.6. Ambrona palaeoeconomy
The study of the sedimentation processes has
enabled us to establish significant differences in
relation to the meaning of the presence of remains
in each stratigraphic context. A unified interpretation
of the site, accepted until 1993, now
seems inappropriate. The process by which the sequence
was built up is a millenarian time period,
but each level also comprises a major diachrony.
Consequently, interpretations contextualised in
short time intervals can only be applied to specific
stratigraphic and spatial units. Moreover, they are
obviously only meaningful for each case in point,
and may differ even in areas that are part of the
same level.
The low lithic density suggests that human
activity did not reach great intensity in the lower
stratigraphic member. Although the small amount
of Acheulean evidence in the surrounding area supports
this interpretation (Rodríguez de Tembleque,
in: Santonja and Pérez-González, 2005 (eds.): 334-
PLEISTOCENE AND HOLOCENE HUNTER-GATHERERS IN IBERIA AND THE GIBRALTAR STRAIT:
THE CURRENT ARCHAEOLOGICAL RECORD
524
Figure 3. Lithic industry of the middle member of Ambrona site. Various sidescrapers: convergent pedunculate (1),
angled-convergent (10), alternate angled-convergent (17), alternate with partial low retouch (6), doubles (8 and 12),
concave (3), straight opposite a cortical back (11), sub-transversal straight (5), straights with invasive retouch (7 and 14
on débordant flake), straight with stepped retouch and Kombewa removal on ventral face (15) and sub-transversal convex
(16). Denticulates (2 and 9). Levallois flakes (4, 12 –sidescraper– and 13). Flint, except for 8 (lidite) and 16 (quartzite).
CENTRAL PLATEAU 525
351). Under these circumstances it must be stressed
that the Ambrona area was a recurrent point of attraction
for several millennia, given that lithic industry
has been recorded on all levels in the lower
member. The general pattern for the procurement
of raw material coincides in all of them, and it must
also be noted that both the flint and the quartzite
used here was brought from elsewhere, in the case
of the flint, from sources up to several dozen kilometres
away (Freeman, 1991; Parcerisas, 2006).
The relationships between fauna –particularly
elephants– and humans has not been established
in all cases. The low incidence of freshly broken
bones and cut marks indicates that human groups
did not play a major role in the accumulation of
these fauna remains. Current studies of the behaviour
patterns of herds of elephants and other
herbivores eloquently define the environments
where remains of these animals build up in Africa:
around ponds and springs, and during prolonged
droughts (Haynes, 1991).
A natural scenario such as the one indicated in
the previous paragraph is what we propose for Ambrona
(Villa et al., in: Santonja and Pérez-González,
2005 (eds.): 352-381). The concentration alpha at
LevelAS3,mainly consisting of a well-circumscribed
adult elephant carcass, has been studied in depth
(Fig. 4).This case provides an example of an individual
which died from natural causes and was buried
in mud, without evidence of human intervention.
The presence of lithic industry and the few recorded
cut marks indicates that the Palaeolithic groups only
acted in the site in a marginal way on the fauna.
2.2.7. Torralba
Work at Torralba between 1990 and 2000 has
been much more limited than in Ambrona (Santonja
et al., in: Santonja and Pérez-González,
2005 (eds.): 104-123). Nevertheless, stratigraphic
checks, numerical dating and reviews of the industry
confirm that this is an Acheulean site situated
chronologically toward MIS 7 (vid. § 2.2.1).
Figure 4. Remains of adult male Palaeoloxodon antiquus scattered across less than 60 m2
on Level AS3, Central Sector.
Excavation campaign of 1995. Bones not detected in the excavated assemblage marked in black (upper left).
PLEISTOCENE AND HOLOCENE HUNTER-GATHERERS IN IBERIA AND THE GIBRALTAR STRAIT:
THE CURRENT ARCHAEOLOGICAL RECORD
526
Figure 5. A: Torralba site. Zone partly dug by the Marquis of Cerralbo (1909-1913) and areas excavated by F. C. Howell
(1961-1963). Equidistance of level curves: 25 cm. B: Torralba: composite stratigraphic profile (adapted from Butzer
1965).
CENTRAL PLATEAU 527
The oldest formations on which the Torralba
Pleistocene deposits lie are, as in Ambrona, red
clays and gypsum from the Keuper and TriassicJurassic
carbonate deposits. Despite the large area
of the site, the only detailed stratigraphy is the
N-S section in the western sector of the site (Fig.
5a), published by K.W. Butzer (1965). This is a
composite profile (Fig. 5b) which starts with up to
3-4 m deep red colluvial facies lying on the Keuper.
These deposits disappear to the north and the
sequence continues with gray sand facies, interspersed
with angular and sub-angular gravel with
a carbonate composition, sized 1-3 cm along the
major axis, reaching a maximum depth of 1 m in
the northern half of profile. At the top there is a
fairly continuous unit of grey-green marl, somewhat
more sandy at the base, with a maximum
depth of roughly 2 m.Above these facies, of a shallow
lacustrine nature, there is a red alluvial-colluvial
deposit, between 0 and 1.5 m deep. In this
sector, the Pleistocene and Keuper levels are affected
by reverse faults with movements of more
than one metre.
The above-mentioned river sand and gravel facies
contain the main concentrations of fauna and
industry found during the excavations by Cerralbo
and Howell. The evidence of fluvial rolling found
on the fauna and industry is typical of this type of
context, and implies movement or disturbance of
some intensity (Sánchez-Cervera et al., e. i. p.).The
former interpretation of some of the Torralba stratigraphic
units (Freeman and Butzer, 1966; Freeman,
1994) as occupation sites thus seems irrelevant, as
they did not take into account the fact that these
deposits were built up in fluvial contexts with
enough energy to move the material. The same is
true of the findings of the taphonomic studies of the
Torralba fauna (Díez et al., 1985), which assumed
the unitary nature of the assemblages on each level,
which is are inconsistent with the secondary position
(non-autochthonous) of the remains.
The composition of the Torralba macrofauna
closely resembles Ambrona, although there is a
clear imbalance in the frequency of certain representative
taxa such as Equus and Elephas. A
parallel of any greater scope would be premature
at this stage, given that only a few taxa –including
Equus (Prat, 1977), poorly represented
in the lower member of Ambrona– have been
studied in depth, and the microfauna and small
vertebrates –except for birds (Tadorna ferruginea;
Mergus serrator; Anatidae indet. and Porphyrio
porphyrio)– are still largely unknown also at
Torralba.
In Torralba, the lithic industry has a density of
less than 1 item per m3
throughout the levels excavated
by Cerralbo and Howell. As in Ambrona,
such low frequencies suggest low-intensity human
presence and interventions.
The raw material collection patterns are also
similar to those observed in Ambrona. The presence
of cores and flakes shows that in Torralba,
quartzite and flint blocks brought from elsewhere
were exploited. But the lack of cortical flakes suggests
also that the material brought to the site may
have been previously scabbled. Similarly, the lack
of cores used to produce the flake supports for
handaxes and cleavers indicates that this toolkit
was already configured when it was brought to the
site (Sánchez-Cervera et al., i.p.).
The handaxes and cleavers on flake set lets us
include Torralba in the Acheulean technocomplex.
However, progressive technological items have
also been observed, such as handaxes and cleavers
with retouch (bifaces support of tools), making
further work necessary to check for technological
traditions which may be ascribed to the Middle
Palaeolithic, as in the case of Ambrona. In the
general context of the European Acheulean, it is
important to stress the presence of true flint hachereaux,
since there is a clear tendency to link these
items closely to the availability of quartzite. In the
light of the numerical chronologies now available,
Torralba is one of the most recent Acheulean sites
in southern Europe.
ROBERT SALA RAMOS (EDITOR)
EUDALD CARBONELL | JOSÉ MARÍA BERMÚDEZ CASTRO | JUAN LUIS ARSUAGA
(COORDINATORS)
PLEISTOCENE AND HOLOCENE
HUNTER-GATHERERS IN IBERIA
AND THE GIBRALTAR STRAIT:
THE CURRENT ARCHAEOLOGICAL RECORD