Stratigraphy and Formation Processes of the Upper Pleistocene Deposit at Gruta da Oliveira, Almonda Karstic System, Torres Novas, Portugal

Abstract

Gruta da Oliveira is a cave located in the Almonda karstic system in central Portugal. Its Upper Pleistocene fill is made up of gravitational sediments with varied biogenic and anthropogenic inputs. The archaeological sequence (35–70 ka) is fairly homogeneous, reflecting a predominantly moist climatic context modulated by millennial-scale variations. Human occupation was almost continuous throughout. In the richer archaeological layers (8–14), the anthropogenic components are in primary position, and post-depositional processes are restricted, although penecontemporaneous biological activity—such as hyena scavenging—overprinted the record. In layers 15–19, the excavated area corresponds to the footslope of a talus, where accumulation proceeded mainly through runoff and where the lithics and bones derive from occupations situated ca. 5m outward. The base of the deposit is >2m from the current base of the excavation, suggesting that Oxygen Isotope Stage (OIS)-4 and OIS-5 sediments underlay the deposit already exposed. This study confirms the archaeological integrity of the Mousterian assemblages found in the different layers, and, given the dates for layer 8, supports the hypothesis that Neanderthals survived in western Iberia until at least 35,000–38,000 calendar years ago. © 2009 Wiley Periodicals, Inc.

Full text

Stratigraphy and Formation Processes
of the Upper Pleistocene Deposit at
Gruta da Oliveira, Almonda Karstic
System, Torres Novas, Portugal
Diego E. Angelucci1,* and João Zilhão2
1Laboratorio di Preistoria B. Bagolini, Dipartimento di Filosofia, Storia,
e Beni Culturali, Università degli studi di Trento, via S. Croce 65, 38100
Trento, Italy
2Department of Archaeology and Anthropology, University of Bristol, 43
Woodland Road, Bristol BS8 1UU, UK
Gruta da Oliveira is a cave located in the Almonda karstic system in central Portugal. Its Upper
Pleistocene fill is made up of gravitational sediments with varied biogenic and anthropogenic
inputs. The archaeological sequence (35–70 ka) is fairly homogeneous, reflecting a predomi-
nantly moist climatic context modulated by millennial-scale variations. Human occupation
was almost continuous throughout. In the richer archaeological layers (8–14), the anthro-
pogenic components are in primary position, and post-depositional processes are restricted,
although penecontemporaneous biological activity—such as hyena scavenging—overprinted
the record. In layers 15–19, the excavated area corresponds to the footslope of a talus, where
accumulation proceeded mainly through runoff and where the lithics and bones derive from
occupations situated ca. 5 m outward. The base of the deposit is ⬎2 m from the current base
of the excavation, suggesting that Oxygen Isotope Stage (OIS)-4 and OIS-5 sediments under-
lay the deposit already exposed. This study confirms the archaeological integrity of the
Mousterian assemblages found in the different layers, and, given the dates for layer 8, sup-
ports the hypothesis that Neanderthals survived in western Iberia until at least 35,000–38,000
calendar years ago. © 2009 Wiley Periodicals, Inc.
INTRODUCTION
Recently, the notion of a delayed survival of Middle Paleolithic Neanderthals in
southwestern Iberia has been the object of much controversy. As originally modeled,
that survival manifested itself in the fact that, south of the Cantabro-Pyrenean
Mountains, late Mousterian assemblages in long cave sequences occupied the
chronostratigraphic slot of the Protoaurignacian and the Aurignacian I of Iberian and
western European regions to the north of that range, while the occurrences of the
Aurignacian II in Portugal and southern Spain signaled replacement by (or assimi-
lation into) in-dispersing modern humans (Zilhão, 1993, 2000, 2006; Walker et al.,
2008). Given current dating evidence, the duration of this “Ebro Frontier” would
Geoarchaeology: An International Journal, Vol. 24, No. 3, 277–310 (2009)
© 2009 Wiley Periodicals, Inc.
Published online in Wiley Interscience (www.interscience.wiley.com). DOI:10.1002/gea.20267
*Corresponding author; E-mail: diego.angelucci@lett.unitn.it.
GEA243_04_20267.qxd 4/3/09 4:36 PM Page 277

have been of some five millennia, from ca. 42 to ca. 35–38 ka (all ages here and below
are in calendar years, with age correspondence of radiocarbon dates effected through
CalPal 2007; Weninger & Jöris, 2004). An even later persistence, until as late as ca.
29 ka, has also been proposed for at least Gibraltar and the surrounding areas
(Finlayson et al., 2006; contra, see Zilhão & Pettitt, 2006).
Others, however, have questioned the very reality of any visibly extended sur-
vival of Mousterian Neanderthals in Iberia (e.g., Jöris, Álvarez, & Weninger, 2003;
Vaquero, 2006). These objections can be grouped in two main families. On the one
hand, the radiometric determinations supporting the Ebro Frontier are deemed to
be inaccurate, reflecting incomplete sample decontamination and representing min-
imum ages only. On the other hand, it is postulated that, even if the dates are accu-
rate, the association of the samples with the archaeological assemblages may only
be apparent. For instance, complex depositional histories may generate palimpsests
that combine material accumulated over extended periods of time (e.g., Mousterian
lithics and collagen-depleted bones, with later carnivore-accumulated faunal mate-
rial producing the more recent dates); and post-depositional disturbance may cause
intrusions of post-Mousterian bones accumulated by carnivores or later Upper
Paleolithic humans.
The issues of dating accuracy eventually will be solved by the replication of radio-
carbon results and their corroboration by alternative techniques. Ultimately, there-
fore, the crucial issues are those of stratigraphic integrity and sample association.
One of the few long stratigraphic sequences where young dates have been obtained
for the latest Mousterian of southwestern Iberia is the Gruta da Oliveira (Cave of the
Olive) in central Portugal. Under excavation since 1989, this site contains an Upper
Pleistocene deposit already exposed over ca. 9 m. Spanning the ca. 35–70ka interval
and yielding rich lithic and bone assemblages, this sequence is thus an ideal testing
ground for the issues of site formation and sample taphonomy underlying current
Ebro Frontier debates. Moreover, it provides detailed evidence on the long-term
trends of change in environments, technology, and human adaptation underpinning
the hypothesized persistence of Neanderthals in the region after ca. 42 ka.
The overall aims of our ongoing geoarchaeological work at Gruta da Oliveira are
to assess the mode of accumulation of the deposit, the extent to which its archaeo-
logical content is in primary or derived position, and the degree of integrity of the
different stratigraphic units recognized. In this paper, we present the first results
concerning stratigraphy, chronology, and soil micromorphology. Besides clarifying
the context of the samples that yielded the recent Mousterian dates, these results also
support preliminary insights concerning the expression of global climatic patterns
in regional paleoenvironmental archives.
GRUTA DA OLIVEIRA
Site Description
The Gruta da Oliveira (coordinates 39°30'23"N and 08°36'49"W, European Datum;
altitude ca. 115 m; Figures 1, 2) belongs to the Almonda karstic system (Torres
Novas, central Portugal), which consists of an extensive labyrinthine complex of
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
278
GEA243_04_20267.qxd 4/3/09 4:36 PM Page 278

underground cavities excavated at different elevations along a 70 m-high escarp-
ment (Figure 3). This escarpment rises above the karstic spring of the Almonda, a
right-bank tributary of the Tagus, which runs a few kilometers to the south. The
water flows out from a gallery (Galeria da Nascente) where two underground
drainages converge: One comes from the north and drains the Serra d’Aire massif,
particularly the synclinal valley of Vale da Serra; the other comes from the west
and is one of the outlets through which the Mira-Minde polje is drained (Figure 2).
Figure 4 provides a schematic N–S profile of the Almonda escarpment with the
location of the main caves of archaeological interest hidden behind it.
Galeria da Cisterna is a fossil spring, where, between 1937 and 1942, limited
archaeological work resulted in a collection of Bronze and Iron Age materials (Paço,
Vaultier, & Zbyszewski, 1947). The more extensive excavations carried out in
1988–1989 uncovered and explored Upper Paleolithic remnant deposits and an Early
Neolithic funerary context (Zilhão, 1997, 2001). Subsequent speleo-archaeological
research led to the identification of several collapsed entrances located at different
elevations in the escarpment, two of which have been reopened for excavation of
the sealed Middle and Upper Pleistocene deposits they contain: at the top of the
escarpment, the Gruta da Aroeira/Galerias Pesadas/Brecha das Lascas complex of
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
279
Figure 1. Morphostructural sketch of littoral central Portugal (modified after Ribeiro, 1970).
Key: 1: Mountain massifs and plateaux. 2: Faults. 3: Granite. 4: Slate. 5: Quartzite ridges. 6: Palaeozoic lime-
stone. 7: Foreland depressions. 8: Jurassic limestone. 9: Other Mesozoic sedimentary rocks. 10: Tertiary
sedimentary basins and rocks. The circle marks the location of the Almonda karstic system; the rectan-
gle corresponds to the area in Figure 2.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 279

Acheulian sites; and, at mid-elevation, the Middle Paleolithic site of Gruta da Oliveira
(Zilhão, Maurício, & Souto, 1991, 1993; Zilhão & McKinney, 1995; Zilhão, 2000; Marks,
Monigal, & Zilhão, 2001; Marks et al., 2002; Chabai, Sitlivy, & Marks, 2000–2001;
Trinkaus et al., 2003; Trinkaus, Maki, & Zilhão, 2007).
Work at Oliveira began in 1989 with the discovery of the Mousterian Cone deposit.
Tested in 1990, the 0.25m3excavated yielded some 250 artifacts and more than 200
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
280
Figure 3. The Almonda escarpment from the southeast. The circles indicate the Gruta da Aroeira/Galerias
Pesadas (top) and the Gruta da Oliveira (bottom). In the foreground, River Almonda.
Figure 2. Digital terrain model of the southern edge of the Estremadura Limestone Massif (see Figure 1
for location; © Crivarque, Lda.).
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 280

ibex, red deer, rhino, and horse teeth and bone fragments (some 30% of which were
burned), as well as numerous fragments of tortoise shell. The deposit was clearly an
“hourglass accumulation” formed via fissures leading, higher up, to a filled chamber
or gallery. Further exploration eventually led to the location and reopening, in 1991,
of the corresponding entrance: the Gruta da Oliveira. Some 15 m further down the
escarpment and containing a late Middle Pleistocene deposit with stone tools and
fauna in derived position, the locus designated Entrada Superior (Figure 4) proba-
bly corresponds to the topographic and stratigraphic base of the speleogenetic stage
of the Almonda system, when Oliveira itself functioned as a spring.
Geological and Geomorphological Context
The Almonda system is located at the edge of the Meso-Cenozoic western border
of the Iberian Peninsula, which constitutes the bulk of the littoral and perilittoral
belt of central Portugal (Figure 1; Ribeiro et al., 1979; Manuppella et al., 1985). The
southern limit of this morphostructural domain is represented by the Serra d’Aire,
a slightly folded and uplifted limestone relief belonging to the Estremadura Limestone
Massif (Martins, 1949). The massif’s southern margin corresponds to a major regional
reverse fault, which causes the Serra d’Aire limestone to overthrust the siliciclastic
sediments of the Lower Tagus Tertiary Basin, dating from the Eocene to the Miocene
(Zbyszewski, Manuppella, & Veiga Ferreira, 1971). The tectonic contact is morpho-
logically emphasized by an escarpment, named Arrife, continuous over tens of km
(Cabral, 1995). Several springs exist along the Arrife’s base, among which the Almonda
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
281
Figure 4. Schematic profile of the Almonda karstic system behind the escarpment.
Key: 1: Mousterian Cone. 2: 27-S Chamber. 3: Reopened entrance of the Gruta da Oliveira. Dates pre-
sented: For Entrada Superior, U-Th on horse tooth enamel (Zilhão & McKinney, 1995); for Gruta da Aroeira,
ESR/230Th/234U on horse tooth (Trinkaus et al., 2003). Elevations in meters above modern sea level.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 281

is one of the most important in terms of water discharge and extent of the associ-
ated karstic network.
The bulk of the Serra d’Aire, as well as Gruta da Oliveira’s bedrock, is formed of
Middle Jurassic well-stratified limestone, including white to light gray oolitic, micro-
crystalline or massive limestone, with subordinate marly limestone, calcirudite, and
calcilutite. Bedrock at Oliveira is formed of calcilutite and massive limestone, some-
times recrystallized, affected by a dense network of discontinuities generated by
tectonic factors. Upslope from the Almonda escarpment, late Tertiary sedimentary
covers composed of sand, conglomerate, and clay are found in the synclinal valley
of Vale da Serra (Figure 2) and along the gentler slopes of the Serra d’Aire
(Zbyszewski, Manuppella, & Veiga Ferreira, 1971). Terra rossa–like soils have devel-
oped from these sediments and the Mesozoic limestone, and are mainly preserved
inside karstic depressions.
Cave Morphology
The original entrance to Oliveira was completely filled with sediment, and
only through continued archaeological excavation has it been possible to gradually
reveal the topography of the encasing rock walls. Basically, the excavated part of the
site comprises passages that developed from two roughly perpendicular sets of frac-
tures with approximate NW–SE and NE–SW orientations (Figure 5). These joints
acted as preferential axes for dissolution and erosion.
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
282
Figure 5. Gruta da Oliveira: Map of the cave, with indication of the topographic grid and the main sectors.
Key: a: Stratigraphic cross sections described in detail [those drawn in Figure 9 are numbered:
(1) N20-W, (2) O19-N, (3) PR18-W]. b: Outline of the cave wall at the base of unit 10. c: Approximate outline
of the cave wall in unexcavated areas. d: Excavated area (2008). e: A-A'topographic profile (Figure 8).
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 282

The current entrance leads to a fossil meander of the underground stream that orig-
inally shaped the cave: the Access Corridor. This meander was fed by two passages
that worked, depending on the fluctuations of the water table, as free-surface mean-
ders or as conduits: the Side Passage and the 27-S Chamber, a slight enlargement at
the west corner of the excavated area (27-S stands for 27 September, the day of the
year 1990 when this passage was discovered). Mechanical erosion related to run-
ning water is evident in the excavation-exposed bedrock of the 27-S Chamber, which
also retains karren-like features due to dissolution processes acting in a vadose envi-
ronment.
Archaeological excavation has so far explored most of the Side Passage and 27-
S Chamber, and is ongoing in the Access Corridor (Figure 6). At the time of reopen-
ing in 1991, a test trench was also dug in the now roofless, external part of the cave;
because of the huge collapsed boulders soon encountered, it was not continued, but
work at this part of the site—named, for descriptive purposes, Exterior—is planned
for the future.
Archaeological Background
Marks, Monigal, and Zilhão (2001) published a preliminary description of the stone
tool assemblages from layers 8–9. Layer 8 yielded 95 tools and debitage exceeding
2.5 cm (mostly quartzite, followed by flint and quartz); the blanks are characteristic,
including typical Levallois flake products; trapezoidal, rectangular, and ovoid shapes
are represented in broadly similar percentages, and dorsal scar patterns are mostly
radial or unidirectional. Layer 9 yielded 112 tools and debitage exceeding 2.5 cm;
made from the same raw materials, this assemblage differs from layer 8 in that ovoid
shapes and radial scar patterns clearly dominate. The number of tools is small in
both layers, and they are mostly notches, denticulates, and irregularly retouched
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
283
Figure 6. Views of Gruta da Oliveira during the excavation. Left: The Access Corridor in 2008 (unit 18 is
under excavation; note the large boulder, resting on layer 15, in front of the archaeologist). Middle: The
27-S Chamber seen from the Access Corridor, in 2006; the cross section PR18-W (Figure 9) is visible on
the right. Right: Excavations in the Side Passage, 1998.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 283

flakes. From these attributes, the lithics from layers 8–9 compare well with those from
other Late Mousterian contexts of Portugal, for example, Foz do Enxarrique and
Figueira Brava (Raposo, 1995).
The assemblages from the underlying units are still under study, but we can note
that in layers 10–14, backed microliths, associated with a few prismatic and pyram-
idal blade and bladelet cores, are present among otherwise characteristically Middle
Paleolithic assemblages based on discoidal, Levallois, and Kombewa flake produc-
tion schemes. These layers probably record an industrial phenomenon akin to that
reported by Maíllo, Cabrera, and Bernaldo de Quirós (2004) for Mousterian levels
20–21 of El Castillo and 11–12 of Morin, which are dated to the same time range. In
underlying layers 15–18, no such Upper Paleolithic–type stone tools exist, but a few
cleavers, as well as a couple of small, flattish handaxes, have been recovered; flake
blanks are also, on average, significantly larger, and there is an increased use of flint,
chert, and chert-like raw materials.
Where the faunal assemblages are concerned, the sequence is fairly homoge-
neous: Red deer is present throughout, while ibex, horse, aurochs, rhino, and tortoise,
represented in varying proportions in layers 10–18, were not recorded in layers 8–9.
Large carnivores are rare if not altogether absent from most units. The presence of
hyenas, in particular, is attested by coprolites, abundant in some units, where they
occur spatially clustered; however, despite the large number of excavated faunal
remains (tens of thousands), skeletal parts of the hyenas themselves are so far
restricted to a single tooth. Carnivore damage is apparent in at least one—a tibial
diaphysis—of the published Neanderthal remains, which also include a phalanx, a
distal humerus, and a proximal ulna (Trinkaus, Maki, & Zilhão, 2007).
Chronometric Framework
In order to obtain a chronology cross-certified by different, independent tech-
niques, dating work at Gruta da Oliveira currently involves thermoluminescence
(TL) on burnt flints, U-Th on animal bones and speleothems, and radiocarbon on
wood charcoal and burnt bone. Available results, so far, concern only this latter tech-
nique, and are presented in Table I, together with the U-Th dates originally obtained
for the Mousterian Cone.
Excluding the two clearly anomalous dates and that on Holocene charcoal from
a small pocket of disturbed sediments intruding into the upper part of layer 8 in the
outward end of the Access Corridor, the calibrated radiocarbon dates are plotted in
Figure 7. A pattern of increasing age with depth is apparent down to layer 11, after
which a plateau is reached. We interpret this plateau as indicating the limit of appli-
cability of the method in the site’s particular geochemical setting. Considering the
large standard deviations of samples OxA-8672 and GrA-22024, as well as the broad
similarity in lithic technology and typology across layers 9–14, both results are prob-
ably valid, with stratigraphy implying that the true age of layer 14 must fall at the lower
end of the interval for GrA-22024. The accumulation of this part of the deposit can
thus be constrained to ca. 42–50 ka.
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
284
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 284

Table I. Gruta da Oliveira: Current radiometric chronology.
Layer Sample Number Material Method Lab # Result B.P. ⌬13C Age cal B.P. Observations
8 P15-A13 Charcoal (Olea sp.) AMS 14C GrA-29384 6055 ⫾45 ⫺24.6 – Intrusive (sample collected
from a burrow)
8 R16-24 Burnt bone AMS 14C GrA-10200 31900 ⫾200 ⫺22.2 35760⫾280 Alkaline fraction dated
8 R15-38 Burnt bone AMS 14C OxA-8671 32740 ⫾420 ⫺21.9 37100⫾830
9 R17-7 Burnt bone AMS 14C GrA-9760 38390 ⫾480 ⫺21.7 42610⫾390 Alkaline fraction dated
9 R17-16/18 Burnt bone AMS 14C Beta-111967 40420 ⫾1220 ⫺22.1 44080⫾990
11 U18-99 Burnt bone AMS 14C OxA-8672 42900 ⫾1200 ⫺20.8 46070⫾1330
11 U17-106 Burnt bone AMS 14C OxA-9379 ⬎25850 ⫾550 ⫺22.9 – Very low collagen content
12 O17-70/72 Charcoal (Pinus sp.) AMS 14C GrA-24408 26940/⫹270/⫺250 ⫺24.4 – Anomalous; minimum age
13 L21-21 Charcoal (Erica sp.) AMS 14C GrA-24410 39540/⫹490/⫺410 ⫺26.0 43310⫾490
14 M19-A12 Charcoal (Pinus sylvestris) AMS 14C GrA-22024 42800/⫹2300/⫺1800 ⫺24.0 46350⫾2050
14 M19-A13 Charcoal (Pinus sylvestris) AMS 14C OxA-13137 27850 ⫾550 ⫺23.7 – Anomalous; minimum age
14 O17-185/O17-A18 Charcoal (Pinus sylvestris) AMS 14C Beta-183537 40900 ⫾1100 ⫺23.2 44380⫾960
15 O17-199 Charcoal (Pinus sylvestris) AMS 14C GrA-24407 37520/⫹380/⫺330 ⫺24.2 42160⫾340
18 O17-489 Burnt bone AMS 14C GrA-29385 37120/⫹380/⫺330 ⫺23.0 41960⫾330
Mousterian Cone Equus sp. (tooth enamel) U-Th SMU-247E1 – – 70250⫾9000
Mousterian Cone Equus sp.(tooth enamel) U-Th SMU-308-247E2 – – 53000/⫹5600/⫺5300 Second measurement
of SMU-247
Age B.P. of radiocarbon results calculated with CalPal 2007. U-Th results after Zilhão and McKinney (1995). All errors are 1-sigma.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 285

The hiatus picked up by radiocarbon between layers 8 (ca. 35–38 ka) and 9
(ca. 42–45 ka) is real, as it corresponds to a discontinuous flowstone observed at
different points during the excavation of the Access Corridor; laterally, instead of a
crystalline crust, induration of underlying layer 9 also occurred, and, in squares
PR14–15, this interface was represented by a several-cm-thick, all-bone-no-sediment
microfaunal breccia. The agreement in the pairs of dates obtained for each of these
layers, on samples of the same kind (burnt bone with d13C values consistent with
collagen) and processed at three different laboratories, further strengthens the reli-
ability of the results, particularly where the young age of layer 8 is concerned.
The Mousterian Cone is situated ca. 2 m below the surface of layer 19. Its U-Th age
of ca. 62.5 ka (average of two statistically identical measurements for a single sam-
ple) is consistent with the radiocarbon results for layer 14. Together, these chrono-
logical data constrain the antiquity of layers 15–19 to the ca. 50–60 ka interval.
METHODS
The first stage of this study was to describe and draw the cross sections exposed
during excavation. Some (see Figure 5) were described in detail, taking into account
the sedimentological, pedological, stratigraphic, and archaeological features of the
recognized geoarchaeological field units (as defined in Angelucci, 2002).
Samples for micromorphological observation were collected during the 2006 and
2007 field seasons. Due to the good cohesion of the sediment, undisturbed samples
were drawn by simple extraction, wrapped in paper, marked, and oven-dried at 60°C
until a constant weight was attained. Thin sections were prepared at the Servizi per
la Geologia laboratory (Piombino, Italy), in the following stages: impregnation with
a mixture of resin, styrene, and hardener; curing; cutting into cm-thick slabs; and final
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
286
Figure 7. Gruta da Oliveira: Plot of calibrated radiocarbon dates. Gray bars: Burnt bone samples. Black
bars: Charcoal samples.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 286

preparation of 25-mm sections, measuring 95 mm ⫻55mm (e.g., Figure 14; only one
slide measures 55 mm by 45 mm).
Thin sections were observed under a polarizing microscope at magnifications
between 25⫻and 1000⫻, using plane-polarized light (PPL), cross-polarized light
(XPL), and oblique incident light (OIL), the last for observation in standard
light conditions and for primary fluorescence. Fluorescence observation was per-
formed using three different wideband filter combinations: ultraviolet, blue (super
wideband), and blue, with excitation filters respectively between 330–335, 420–480,
and 400–440 nm, and corresponding suppression filters at 420, 520, and 475 nm.
Images were captured through a digital camera for polarizing microscopy. Thin-sec-
tion description follows the guidelines proposed by Bullock et al. (1985) and Stoops
(2003), with additions from Courty, Goldberg, and Macphail (1989) and Goldberg
and Macphail (2006).
RESULTS
Stratigraphic analysis of the Gruta da Oliveira fill shows that the deposit’s geom-
etry strongly adapts to cave morphology, with irregular bedding and significant lat-
eral variations (in the quantity and size of stones, the grain size and color of fine
material, the intensity and type of carbonation, and even the presence/absence of
some units). Other factors, such as the quantity of roots or the presence of areas
softened or decarbonated by biological activity, are related to the near-surface con-
ditions of some parts of the cave, namely the Access Corridor.
Three main stratigraphic complexes form the site succession:
1. The collapse sealing the cave (layers 1 to 6), which filled the Exterior and
Access Corridor areas.
2. The Middle Paleolithic archaeological sequence (layers 7 to 19, so far), which
is formed of fine (silty loam to silty sand), mostly reddish-brown sediment,
more or less regularly alternating with flowstones or carbonate crusts.
3. A remnant of fluvial sediment, older than the archaeological sequence, pre-
served in the karren-like depressions of the 27-S Chamber.
The Archaeological Sequence: Field Characteristics and Micromorphology
Geometry and Stratigraphic Layout
The archaeological sequence of Oliveira is sealed by the deposit that filled the
cave entrance, composed of material deriving from roof collapse and the slope.
The top of the sequence is marked by a well-cemented continuous carbonate flow-
stone, with laminar structure and a maximum thickness of 13 cm, recognizable across
the whole cave. Due to lateral variations, this flowstone can be superimposed to dif-
ferent units, for example, layer 7 in the Access Corridor but layer 13 in the innermost
part of the 27-S Chamber.
Observing the sediments’ geometry, it is clear that inputs came from two distinct
directions, the Access Corridor and the Side Passage, which determines spatial con-
figurations that vary between the different loci. The distribution of large boulders
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
287
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 287

fallen from the roof, particularly at the interfaces between layers 19/18 and 15/14, also
influenced the presence and distribution of both the layers and their content in the
three main sectors of the cave. On average, in the Access Corridor, layers dip NW,
having been washed in directly from the slope (Figure 8); in this position, the strat-
ification is slightly disturbed by post-depositional processes, particularly root action
and faunal activity along the walls and the edges of large boulders. The layers in the
Side Passage (units 8 to 13bis, which rest on bedrock) dip SE with a low angle. In
the 27-S Chamber, the stratification is broadly horizontal and represented by layers
8 to 16, the latter resting on bedrock (Figures 9, 10).
Field Characteristics
In the archaeological deposit, most layers display similar characteristics, both in
the field and under the microscope: They are mainly silty, with colors ranging from
reddish brown to light reddish brown (see Table II for Munsell notations), often mas-
sive, and separated among them by flowstones or carbonate crusts. All layers contain
variable quantities of éboulis, that is, heterometric limestone fragments, usually
more common near the cave walls and deriving from the degradation of the encasing
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
288
Figure 8. Gruta da Oliveira: Topographic profile and schematic cross section of the Access Corridor
prior to the 2008 excavation of grid units NR/15–16 down to the interface between layers 18 and 19 (see
Figure 5 for location).
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 288

GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
289
Figure 10. Gruta da Oliveira: The cross section PR18-W (see Figure 5 for location; this cross section
was drawn before its final cleaning for preservation purposes; thus, slight discrepancies exist between
this picture and the drawing in Figure 9).
Figure 9. Gruta da Oliveira: Drawing of the cross sections N20-W, O19-N, and PR18-W (see Figure 5 for
location).
Key: (a) stones; (b) coprolites; (c) lithic artifacts, flint; (d) lithic artifacts, quartzite; (e) bones; (f) micro-
morphological samples; (g) burrows; (h) carbonation; (i) unit names.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 289

ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
290
Table II. Gruta da Oliveira: Field characteristics of excavated units.
Unit Main Characteristics Lower Boundary
1–6 Surface horizons and boulders sealing cave entrance
k/7 Well-cemented carb. flowstone with laminar structure
7 Calcareous breccia with sandy loam matrix
8 Clayey silty loam; 4YR4/4; common limestone frs. with surface Sharp, wavy
patina; poor to moderate carb. cementation and common carb.
nodules
k/9 Two superimposed carb. crusts with laminar structure
9-up Silty loam; 5YR4/3 with 6YR4/2 mottles; common limestone frs. Sharp, wavy
with surface patina, scarce calcareous mm-sized fraction,
occasional calcite crystals; moderate carb. cementation
9-low Clast-supported calcareous breccia with silty loam matrix; Clear
5YR4/4 [Access Corridor] to 4YR4/5 [27-S Chamber]; strong
[Access Corridor] to moderate [27-S Chamber] carb.
cementation
k/10 Discontinuous carb. cementation, at places degraded carb.
crust
10 Clayey silty loam; 5YR4/4; few limestone frs., calcareous Clear to sharp [Side
mm-sized fraction and calcite crystals; common coprolites Passage]
and digested bones; moderate carb. cementation
k/11 Well-developed continuous carb. flowstone [Side Passage]
11 Clayey silty loam; 6YR4/4 [Side Passage] to 6YR5/5 Clear, with large limestone
[27-S Chamber]; few to common limestone fr. (including frost frs.
slabs), scarce mm-sized fraction and speleothem frs.; patchy
carb. cementation and few carb. pendants; few calcified roots
[Access Corridor]
k/12 Discontinuous, moderately cemented carb. crust, degraded
at places
12 Silty loam; 6YR4/4; few to common limestone fr. (including Clear, dipping SW [Side
frost slabs), scarce mm-sized calcareous fraction and Passage], with limestone
speleothem frs.; carb. pendants and few calcified roots [Access frs.
Corridor]; organic spots in the upper part [27-S Chamber]
k/13 Carb. crust
13 Silty loam; 6YR4/4; common limestone and speleothem frs., Clear, dipping SW [Side
scarce mm-sized fraction formed of limestone frs. and bone frs.; Passage]
platy elements dip SW; patchy carb. cementation, few carb.
pendants and calcified roots; darker areas (6YR3/3) with
organic matter, ash, charcoal, and phosphate
k/14 Carb. crust
14 Silty loam; 6YR4/4 [Access Corridor] to 5YR4/4 [27-S Chamber]; Clear linear
very few limestone frs.; poorly developed granular structure
[Access Corridor]; discontinuous carb. cementation, few
calcified roots
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 290

bedrock (e.g., Figures 9, 10), as shown by their lithology, shape, and the common pres-
ence of endolithic patinas on one or two sides. The main characteristics of the layers
excavated so far are found in Table II.
Micromorphological Features
Given their relative homogeneity, the thin sections from the archaeological deposit
of Gruta da Oliveira can be discussed as a whole, particularly for constituents,
microstructure, and pedofeatures (Tables III, IV). Several coarse components are
ubiquitous and cluster into four main groups:
• Siliciclastic silt and sand fraction (SIL). All layers contain silt and sand frac-
tions formed of noncarbonate elements ranging from very fine silt to coarse
sand, with occasional larger grains (around 2–3 mm). Their shape is varied:
Subangular grains are dominant but all degrees of roundness, even angular and
well-rounded (Figure 11a), are present. This fraction includes the following
components: monocrystalline quartz, sometimes with a “dirty” aspect due to
the presence of inclusions or bubbles, with varied extinction patterns (some-
times strongly wavy), representing between one-half and two-thirds of the
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
291
Table II. (Continued)
14-base Silty loam; 5YR4/4; common limestone frs. and speleothem frs. Clear linear
Boulders [Access Corridor]
k/15 Degraded carb. crust
15 Mm-thick sets, dipping SW, of clayey silty loam (on average); Clear
6YR4/4 [Access Corridor] to 5YR4/4 [27-S Chamber]; few
[27-S Chamber] to common [Access Corridor] limestone frs.
(including frost slabs); degraded carb. crust between sets;
open-work among larger stones
k/16 Thin continuous carb. crust, at places with laminar structure
16 Silt; 5YR4/5 [Access Corridor] to 4/3 [27-S Chamber]; few small Clear linear
limestone frs.; moderate carb. cementation
17 Silty sand; 6YR4.5/5; common limestone frs. and few Clear linear
speleothem frs.; weak lamination
k/18 Well-cemented carb. crust
18 Silty loam; 6YR4.5/6; common limestone and speleothem frs.; Sharp, erosive [Access
weak lamination Corridor]
Boulders [Access Corridor]
k/19 Carb. crust
19 Sandy silt; 6YR4.5/6; scarce limestone frs. and speleothem Clear linear
frs.; weak carb. cementation (on top)
Key: carb.: calcium carbonate; fr(s).: fragment(s). Square brackets indicate the location inside the cave.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 291

Table III. Gruta da Oliveira: Main micromorphological characteristics (I), microstructure and components.
Unit Aggregation Porosity Coarse Components Fine Material
11 Microgranular (masked by Few channels and Few SIL; absent LST; few CRB; very few ABC (phosphate frs., Reddish brown, speckled
carbonate precipitation), with chambers some botrioidal; coprolites); very few clay aggregates
crumb and channel areas
12 Granular, with microgranular Few vughs; very few Few SIL and LST; common CRB (weathered calcite common in Varied (see text)
areas channels and chambers this class); common ABC (phosphate and organic frs.; bones
with varied thermoalteration or weathering)
13 Subangular blocky to crumb, Common chambers; few Few SIL; common LST; few CRB; common ABC (bones, often Reddish brown, speckled,
weakly developed channels, planes and vughs small or digested; few phosphate frs.); few opaques (often with dark brown, dotted
rounded) areas
13 carb. Channel and chamber, with Common channels and Common SIL and LST; few CRB; common ABC (bones, some Reddish brown, speckled,
microgranular (masked by chambers; very few vughs thermoaltered; phosphate frs.); very few lithic artefacts and frs. with dark brown dotted
carbonate precipitation) areas of thermoaltered sediment areas
14 Channel, with microgranular Common channels; few Frequent SIL (quartz dominant, micas absent); common LST Reddish to grayish brown,
and chamber areas complex packing voids (often size ⬎1 mm); few CRB; common ABC (bones, with varied speckled to dotted
thermoalteration, often in clusters and digested; phosphate frs.);
very few opaques; some ash aggregates
14b Blocky subangular, weakly Few planes and channels; Frequent SIL; few LST and CRB (weathered calcite common in Reddish brown to brown-
developed, with channel and very few chambers this class); common ABC (particularly bones with varied size, ish red, speckled
chamber areas thermoalteration, and weathering); very few opaques
(often rounded)
15 Granular to subangular blocky, Few channels, planes, and Dominant SIL; few LST, CRB and ABC (bone frs. frequent in this Reddish brown, speckled
with microgranular areas chambers class); some frs. of reworked alluvial sediment and phosphate frs.
18 Subangular blocky to crumb, Few channels and planes Dominant SIL; LST and CRB almost absent (see text); few ABC Reddish brown, speckled
with channel and (bones, often digested, and phosphate frs.); very few opaques
microgranular areas
80 Single grain Packing voids (variable), few SIL; clay pellets (see text) Reddish brown, speckled
vughs and vesicles
FEAT Subangular blocky, weakly Few planes, channels and Well sorted fine to med. silt (see text) Dark reddish brown,
developed chambers dotted
Key: SIL, LST, CRB, ABC: see text; FEAT: archaeological feature in sq. KL/20–22, interface of layers 14b and 15.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 292

Table IV. Gruta da Oliveira: Main micromorphological characteristics (II), groundmass and pedofeatures.
Unit c/f RlDP b-Fabric Pedofeatures and Sedimentary Features
11 Open porphyric Crystallitic Common sparite/microsparite infillings; few loose discontinuous biogenic infillings; few dense silty clay
coatings around grains
12 Gefuric to enaulic Crystallitic Diffuse cementation by micrite/microsparite; few phosphate nodules; very few loose discontinuous biogenic
infilling; very few irregular impure clay coatings in voids
13 Open to close porphyric Undifferentiated Common sparite/microsparite infillings and coatings in voids; few loose discontinuous biogenic infillings; few
Fe-Mn coatings and silty clay coatings around limestone frs.; few dense silty clay coatings around grains
13 carb. Open to close porphyric Undifferentiated with Common microsparite infillings, coatings, and hypocoatings, and very few nodules; common loose discontinuous
crystallitic parts biogenic infills; few dense silty clay coatings, particularly around bones and phosphate frs.; few phosphate
nodules
14 Open to close porphyric Undifferentiated with Common micrite/microsparite hypocoatings and nodules, and few intercalations; few dense continuous
crystallitic parts biogenic infillings; few dense silty clay coatings around grains; very few phosphate nodules; very few Fe-Mn
coatings around grains
14b Open to close porphyric Granostriated, poorly Common loose continuous and discontinuous biogenic infillings; few dense silty clay coatings around grains;
developed very few fabric infillings (a.k.a. passage features) in channel; few micrite hypocoatings in channels; very few
micrite hypocoatings and coatings on grains; very few micrite concentrations/nodules (degraded crust); very few
phosphate “rinds” on limestone frs.
15 Close porphyric with Undifferentiated Few micrite hypocoatings (on grains), coatings, and infillings (in pores); very few fabric infillings (a.k.a. passage
gefuric parts features) in channel; very few dense silty clay coatings around grains
18 Open to close porphyric Undifferentiated Common loose discontinuous and few loose continuous biogenic infillings; few carbonate (sparite ⫹micrite)
nodules; few micrite coatings and hypocoatings, irregulary scattered; poor bedding
80 (See text) — Few micrite coatings in voids and nodules; few Fe-Mn nodules; very few biogenic infillings
FEAT Open porphyric Undifferentiated Few micrite coatings in voids; few Fe-Mn small irregular nodules (mottling)
(c/f 30/70)
Key: c/f RlDP: coarse/fine related distribution pattern; FEAT: archaeological feature in sq. KL/20–22, interface of layers 14b and 15.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 293

fraction; feldspars, common to scarce in this fraction, unweathered to slightly
weathered (e.g., Figure 13g); polycrystalline quartz, scarce, mainly formed of
medium-sized quartzite, with strong wavy extinction; occasional fine-grained
chert grains, mafic minerals, and micas.
• Limestone fragments (LST). Different types of limestone, with varied shape
and size, usually large (from 1 mm to several cm), occur. The limestone types
found are: fine micritic limestone, organic, very dark gray in PPL (e.g., Figures
11b, 11c), with fine microsparite veins or small nodules, or even common
small subcircular to ovoid microsparite concentrations deriving from the
recrystallization of fossils; grainstone formed of peloids/oncoids with sparry
cement; pelmicrite with poorly recognizable peloids/oncoids embedded in
organic dark micrite.
• Other carbonate components (CRB) are also found, including: calcite crystals,
usually large (often mm-sized; e.g., Figures 11e, 11f), monocrystalline or
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
294
Figure 11. Gruta da Oliveira: Micrographs of the archaeological sequence: (a) well-rounded grain of per-
titic feldspar, unit 18, XPL; (b) limestone fragment with amorphous phosphate internal hypocoating, unit
14, PPL; (c) as (b) but XPL; (d) as (b) but fluorescence UV; (e) calcite crystal with phosphate replacement
along rim, unit 14b, PPL; (f) as (e) but XPL; (g) fragment of burnt bone, unit 14b, PPL; (h) as (g) but XPL.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 294

twinned, showing varied weathering patterns (from unweathered crystals to
crystals with alteration along the rim and others with evidence of reaction
with phosphates); reworked fragments of speleothems or carbonate crusts,
formed of spar with elongated or fibrous crystals.
• Anthropogenic and biogenic components (ABC) are systematically present, and
include: bones, abundant in some thin sections, showing a wide range of ther-
mal alteration (Figures 11g, 11h, 12a); lithic artifacts, occasional but almost always
present (Figure 12b); angular to subangular fragments of phosphatic composi-
tion, mostly amorphous and with undifferentiated fabric, ranging from hundred-
mm to mm-size (Figures 12c, 12d, 12e).
No pattern of orientation or distribution of the coarse constituents was observed
in any thin section.
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
295
Figure 12. Gruta da Oliveira: Micrographs of the archaeological sequence: (a) as Figure 10h but fluo-
rescence UV; (b) flint artifact, unit 13, XPL; (c) fragment of amorphous phosphate, unit 11, PPL; (d) as
(c) but XPL; (e) as (c) but fluorescence UV; (f) overall aspect of unit 11; note the large phosphatized frag-
ment, the silty clay external hypocoating around the rounded feldspar grain, the former granular microstruc-
ture, and microsparite filling pore space, unit 12, PPL; (g) as (f) but XPL; (h) granular microstructure
masked by microsparite and micrite precipitation, unit 13, XPL.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 295

The characteristics of the fine material often vary in the individual thin sections.
The coarse/fine (c/f) ratio (c/f limit fixed at 2mm) and the c/f–related distribution
pattern may also show some heterogeneity; b-fabric ranges from undifferentiated
to crystallitic, and only in one case (layer 14b) was a weakly developed granostriated
b-fabric noticed.
Pedality is, on average, poorly developed and only a weakly developed subangu-
lar blocky aggregation was observed in some samples (Table II), sometimes masked
by post-depositional carbonate precipitation (e.g., Figures 12f, 12g, 12h), while apedal
microstructures prevail.
Pedofeatures also show similarities across the thin sections observed. Those
related to carbonate accumulation are widespread, and most layers exhibit various
forms of accumulation of micrite or microsparite as nodules, coatings on voids or
around grains, and hypocoatings in voids. In the upper units (layers 11 to 13),
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
296
Figure 13. Gruta da Oliveira: Micrographs of the archaeological sequence: (a) well-crystallized carbon-
ate flowstone with fibrous fabric and detritic intercalations, unit k/7, XPL; (b) carbonate flowstone with
microcrystalline fabric, degraded by phosphate impregnation and mechanical disturbance on top, unit k/12,
PPL; (c) as (b) but XPL; (d) as (b) but fluorescence UV; (e) phosphate nodule, unit 12, XPL; (f) as (e) but
fluorescence UV; (g) feldspar grain with silty clay external hypocoating, unit 11, XPL, with condenser;
(h) bedding in unit 18, PPL.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 296

carbonation is particularly intense, with microsparite accumulation entirely infilling
the preexisting porosity and leading to the cementation of the sediment (Figures
12f, 12g, 12h). Carbonate accumulation is also represented by the widespread pres-
ence of carbonate crusts or flowstones that are often found at the boundaries between
units. Under the microscope, well-developed flowstones show variable microstruc-
tures. Unit k/7, the main flowstone sealing the archaeological sequence, is formed
of well-crystallized sparry calcite with fibrous microstructure (Frisia et al., 2000)
and contains occasional detrital grains (mainly silt-sized quartz; see Figure 13a),
while other flowstones exhibit microcrystalline fabrics (Figures 13b, 13c).
Some of the flowstones appeared in the field as weakly cemented and almost
powdery when dry. Microscopic observation shows that their alteration is related to
a chemical interaction between carbonates and phosphates, with original carbonate
components being impregnated and replaced by amorphous phosphate (Figures 13b,
13c, 13d). Phosphate accumulations—well detected by fluorescence—are wide-
spread throughout the deposit. Nodules of amorphous phosphates, with diffuse
boundaries and a moderate degree of impregnation, were observed in the upper
units of the sequence (particularly in layers 12, 13, and 14; Figures 13e, 13f), as also
were phosphatic hypocoatings or coatings around limestone fragments (Figure 11d)
and calcite crystals, which often show reaction rims related to the interaction between
calcium carbonate and phosphates (Figures 11e, 11f).
Other widespread pedofeatures are dense silty clay external hypocoatings or coat-
ings around grains, particularly around larger grains (Figures 12f or 13g), bone frag-
ments, or calcite crystals, which resemble “rolling pedofeatures” (Boschian, 1997).
Finally, pedofeatures denoting moderate bioturbation processes, such as biogenic
infillings, are present in all units, as are those resulting from moderate oxidoreduc-
tion processes, which were observed in some samples (see Table IV for details). For
the micromorphological features of the individual layers, see Tables III and IV.
The Fluvial Remnant: Field and Micromorphological Characteristics
In the 27-S Chamber, a fluvial remnant, conventionally named layer 80, was
detected under the archaeological sequence, filling depressions in the bedrock. In the
field, this fluvial sediment was formed of lenticular layers of well-sorted dark gray
sand with intercalations of yellowish red (4YR4/6) silt. The sand layers show cross-
bedding and scours with erosive boundaries cutting finer beds (well visible in thin
section; see Figure 14a).
Under the microscope, it is clearly visible that the bedding results from variations
in the grain size, rounding, and packing of the coarse components. The layers show
similar composition: dominant monocrystalline quartz, often with wavy extinction,
common feldspars and polycristalline quartz, scarce mica and occasional amphi-
boles, pyroxenes and opaques (i.e., the SIL complex of components; see above),
with varied shapes, from subangular to rounded (Figure 14b). In many beds, the sec-
ond most common constituents are rounded clay pellets (Figure 14c), their size rang-
ing from very fine to medium sand, formed of various types of speckled reddish-
brown clay. Some of the clay pellets are reworked clay coatings, or papules (Brewer,
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
297
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 297

1976). The quantity of fine material is also variable: the sand layers’ c/f ratio ranges
from 75/25 to 90/10, with c/f–related distribution patterns varying, accordingly, from
close porphyric to chitonic or gefuric; the finer layers exhibit a c/f ratio of about
50/50, open porphyric c/f–related distribution pattern, and undifferentiated b-fabric
(see Tables III and IV for details). No anthropogenic or biogenic inputs were observed
in the fluvial sediment.
The KL/20–22 Hearth
The hearth found in squares KL/20–22 at the interface between layers 14b and 15
was recognized due to its structure, reddening, and the distribution of artifacts, char-
coal, and burnt bone in and around it (Figure 15). Under the microscope, the reddened
belt is formed of well-sorted, fine to medium silt material (the main components are
quartz, feldspars, micas, and chert), embedded in a dense, brownish-red clay (see
Tables III and IV for other characteristics). Pedofeatures are represented by
micrite coatings in channels and planes and some mottling given by iron-manganese
oxide. No human inputs were detected, except for some clusters of fragments of amor-
phous organic matter. The characteristics of this sample are obviously distinct from
others in the archaeological sequence but its interpretation is difficult, as the stronger
reddish color of the fine material is the only clear evidence of thermal alteration.
FORMATION PROCESSES AT GRUTA DA OLIVEIRA
The Gruta da Oliveira succession is complex, as a result of the cave’s peculiar
configuration and context. The sediments forming the archaeological sequence show
significant lateral variability, derive from multiple inputs and different sedimentary
processes, and were affected by post-depositional processes, both natural and
anthropic.
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
298
Figure 14. Gruta da Oliveira: The alluvial remnant: (a) thin section OLV 7-10 (size 95mm ⫻55 mm);
(b) micrograph of the laminated alluvial sand, XPL; (c) detail of one lamina formed of clay pellets, PPL.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 298

Sedimentary Sources
Field and microscopic evidence indicates that no significant changes of sedi-
mentary sources occurred throughout the accumulation of the archaeological
sequence. Four main groups of coarse components are systematically present in all
the archaeological layers.
The ultimate origin of the siliciclastic fraction (SIL) are the Plio-Pleistocene sur-
face formations fed by metamorphic terrains, widespread in central Portugal (Ribeiro
et al., 1979), as indicated by the common presence of quartz grains with strong wavy
extinction and of quartzite fragments. The immediate sources of these components,
however, most probably are the soil and surface sediments covering the Serra d’Aire
massif, as well as the fluvial sediment already present in the cave when the archae-
ological sequence started to accumulate. Bearing in mind the labyrinthine nature of
the network, it should also be noted that this kind of material is present in passages
and chambers located higher up in the Almonda escarpment, from where it could have
been remobilized, by gravitational or alluvial processes, eventually to contribute to
the build-up of the Oliveira deposit. At the same time, the presence of clay pellets and
papules detected both in the fluvial sediment and in some layers of the archaeolog-
ical sequence point to the existence of inputs coming from the erosion of Tertiary ter-
rigenous sediments and the soils developed from them. The complex origin of this
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
299
Figure 15. Gruta da Oliveira: The hearth found in squares KM/20–22 at the interface between layers 14b
and 15. Top left, dense scatter of lithics and bones above the burned surface. Top right, the burned sur-
face exposed. Bottom left, the hearth in the L20–21 S profile (note the fluvial remnant at the base of the
sequence on the left side of the view). Bottom right, detail of the hearth in profile view.
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 299

SIL fraction, related to distinct phases of the cave infilling process and necessarily
implying that the sediments must contain inherited quartz grains, thus has broader
implications, namely as a cautionary tale in the use of luminescence techniques for
the dating of cave deposits.
Limestone fragments (LST, or éboulis) are present throughout the archaeological
sequence. Frost slabs were detected in units 11, 12, and 15, indicating that, during
their formation, shattering by discontinuous frost was active at the cave entrance and
that climate conditions were indicatively cold and humid (Woodward & Goldberg,
2001; Ferrier, 2002). For the rest of the units, the origin of limestone fragments most
probably resides in “normal” processes of mechanical degradation of the roof and
walls of the cave, as indicated by their heterometry and varied shape, as well as by
the common presence of endolithic alteration surfaces. No other features relating to
frost action were observed in the deposit, suggesting a relatively mild climate dur-
ing its formation.
The other carbonate components (CRB) include calcite crystals and fragments of
speleothems or carbonate crusts, which may derive from short-distance transport in
the cave and originate in exposed surfaces or in preexisting carbonate concretions.
Their presence indicates phases of dissolution or erosion inside the cave.
As a fourth group, we have considered components that are clearly anthropogenic
and biogenic (ABC). On average, lithic artifacts are scarce (in micromorphological
terms), while bones are common and even frequent in some samples. The origin of
bones is both anthropogenic and biogenic, as revealed by the presence of both ther-
moaltered bones and clusters of digested bones, strongly reduced in size. Some
bones display evidence of rolling, caused by reworking as sedimentary clasts during
the accumulation, or by biological activity. Reworked phosphate fragments were
detected in almost all units, indicating that phosphatic inputs have always been sig-
nificant inside the cave. The accumulation of guano is still active today, and the
homogeneous phosphatic pieces presumably derive in part from the incorporation
of this material in the sediment. In other cases, the clasts are coprolite fragments
resulting from the degradation of the excrements of hyenas and other carnivores
using the cave during phases of human abandonment.
Sedimentary Dynamics
Several characteristics suggest that the dynamics involved in the accumulation of
the sedimentary matrix of the Oliveira sequence relate to the movement of slope-
washed material or displaced through shafts and galleries of the inner karst. Most
of the microfacies observed in thin section contain soil aggregates and mineral grains
arranged in a relatively loose fabric, often masked by carbonate accumulation, with
the juxtaposition of distinct components and without recognizable preferential ori-
entation or distribution patterns, giving a chaotic aspect to the sediment. B-fabric is
almost always undifferentiated and no preferential clay domains are visible. Some
of the constituents display dense silty clay coatings or external hypocoatings resem-
bling rolling pedofeatures that may be related to depositional mechanisms of
the mud flow or debris flow type (Boschian, 1997; Ferrier, 2002). In the lower part
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
300
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 300

of the archaeological sequence (units 15 to 18), weak bedding is recognized both in
the field and under the microscope (Figure 13h).
All these features suggest that mass movements acted under wet conditions with
variable, intermediate to high, sediment concentration (Bertran & Texier, 1999).
Runoff or overland flow are the dominant processes responsible for the formation
of the lower sequence (units 15 to 18), while debris flow was mainly involved in the
genesis of the upper units, at least for layers 11 to 14. This change of sedimentary
dynamics is due to the reconfiguration of the cave entrance because of the fall, over
the interface between layers 15 and 14, of large boulders that obstructed the Access
Corridor (see Figures 6, 8). Thus, from then on and until layer 8 times, when sedi-
mentation inside the cave reached the height of the boulder, the build-up of the
sequence (layers 14 to 9) resulted from slope-wash through the inner karst (other parts
of the cave itself, or fissures communicating with the surface or with other cave fills
higher up in the system) or from spill-over across the mass of boulders obstructing
the entrance.
Another phase of major roof collapse resulting in the accumulation of boulders
with volumes in the order of cubic meters is documented at the interface between
layers 19 and 18. The interpretation of these events is not straightforward: It indicates
phases of decompression of the rock mass which might depend on climatic factors
(for instance, frost slabs were detected in unit 15), seismic events (which cannot be
excluded, considering the tectonic situation of the region), or a rapid retreat of the
escarpment related to regional morphodynamic evolution.
To these main processes, one should add the inputs of éboulis from the roof and
walls. It cannot be excluded, however, that some limestone fragments were brought
into the cave from the outside.
The dynamics that led to the sedimentation of the fluvial remnant are rather distinct.
Its characteristics are typical of sediments accumulated by running water through
tractive mechanisms of varying energy, with a few intercalated episodes of decanta-
tion. This sediment, still undated, indicates that at the time of deposition, the 27-S
Chamber was already open and under phreatic conditions in a phase when water was
flowing out of the karst via the cave entrance, that is, when the Oliveira passages were
an active spring—thus much earlier than the deposition of the archaeological sequence.
Syn- and Post-Depositional Processes
Most of the samples display pedofeatures mainly related to carbonation, phos-
phatization, oxidoreduction (slight), and bioturbation. The accumulation of sec-
ondary carbonates was detected in all samples and is one of the major processes that
acted on the archaeological sediments. The percolation and flow of carbonate-
saturated water into the cave was more or less continuous during the time interval
corresponding to the accumulation of the strata, as demonstrated by the almost
systematic presence of sparitic flowstones between layers, formed during non-
depositional hiatuses and often used as stratigraphic markers that, in excavation, are
of much assistance in the distinction of the different layers. At least ten phases of
carbonate-saturated vadose flow were detected in the archaeological sequence,
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
301
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 301

corresponding to the tops of layers 19, 18, 16, 15, 14, 13, 12, 11, 10, and 9, to which
we must also add the flowstone (k/7) sealing the archaeological sequence (Table II).
The fibrous and microcrystalline fabrics observed in the flowstones sampled for
micromorphology indicate formation by the input of water in low-supersaturation con-
ditions (Frisia et al., 2000), with significant variation of discharge rates (as in unit
k/12), or under a constant flow of water (as in unit k/7).
Microsparite cementation also affects the bulk of the units immediately underlying
the k/7 flowstone (e.g., layers 11 to 14 in the 27-S Chamber), indicating waterlogging
by carbonate water, probably synchronous to the formation of that flowstone. The
lower part of the sequence (layers 15 and 18, as well as the basal fluvial sediment), on
the contrary, is affected by micrite accumulation, which derives from carbonate pre-
cipitation in a near-surface vadose context, sometimes by means of biological media-
tions; as a result, calcified roots and micritic hypocoatings are common in layers 15–18
of the Access Corridor. These pedofeatures indicate a constant presence of carbonate-
rich water percolating from the entrance and the chimneys during the accumulation
of the deposits, thus suggesting a humid microclimatic context inside the cave.
The accumulation of phosphates was observed in almost all samples. Phosphate
inputs in the cave are related to the human occupation and to biogenic contribu-
tions, often recognizable directly in the field. Phosphatic pedofeatures are present
in various forms: moderately impregnated nodules, only detectable through obser-
vation in autofluorescence (e.g., Figures 13e, 13f); amorphous coatings or hypocoat-
ings, particularly around limestone fragments (Figures 11b, 11c, 11d); dissolution
and replacement rims around calcite crystals (Figures 11e, 11f); and impregnations
on carbonate crusts or flowstones (Figures 13b, 13c, 13d), which are responsible
for the “degradation” of these crusts to almost powder, often observed in the field.
Since the remobilization and precipitation of phosphate are considered as near-
surface processes (Karkanas et al., 2000), we have to infer that phosphatic inputs to
exposed surfaces occurred continuously during the accumulation of the sequence
and that the input rate was fairly high. Phosphate components laid down over the sur-
face were later incorporated in the sediment by bioturbation or other processes
(such as trampling or puddling) and partly remobilized by the water circulating in
the sediment—another indication that, during the formation of the archaeological
sequence, the microclimate of the cave interior was benign and wet. The abundance
of water in the cave’s sediments is further proven by the occasional but widespread
presence of amorphous ferromanganesian features (see Table IV), indicating slight
hydromorphism due to periodic waterlogging.
Bioturbation, even if present, does not seem to have significantly affected the
archaeological layers. Biogenic infillings were often observed, as well as occasional
fabric (“passage”) pedofeatures, which become common along the wall and around
boulders in the Access Corridor.
Anthropic Impacts
Human inputs were detected in all samples (except for the fluvial sediment, layer
80) and are clearly recognizable in thin section, even if their relationships with
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
302
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 302

respect to the sedimentary matrix are varied. Besides the constituents clustered in
the group named ABC (see above), anthropogenic microfacies—including in situ
archaeological features—were detected in the 27-S Chamber and the Side Passage,
the sectors of the cave where human occupation took place during the accumulation
of layers 14b to 11.
The micromorphological features of the hearth found in squares KL/20–22 (see
above and Figure 15) resemble those of the dense active zone of a residual structured
microfacies (Gé et al., 1993), with the underlying unit 15 being the corresponding pas-
sive zone. This feature may therefore be interpreted as a prepared surface that was
subject to slight thermal impact and trampling. The preservation of this archaeo-
logical feature is mainly due to its position in a sort of “cul-de-sac” at the bottom of
the cave, within one of the Karren-like depressions of the bedrock, and to its formation
during a phase when the cave entrance was obstructed by large boulders.
DISCUSSION
The complex, 9-m-thick deposit currently exposed at Gruta da Oliveira is a good
example of the multiple processes that may affect a cave during all phases of its
evolution, from speleogenesis to ultimate filling. In the following, we summarize the
main points concerning site formation, stratigraphic integrity, dating, and paleonvi-
ronmental reconstruction. Ongoing stone tool refitting studies, as well as U-series and
TL dating, to be published elsewhere, will provide additional detail.
Formation Processes and Stratigraphic Integrity
The cave originated as a result of the activity of a former spring of the Almonda,
today flowing out some 40 m below. A remnant of fluvial sediment, contextual to
this phase of speleogenesis but still undated, is preserved in the 27-S Chamber. The
archaeological sequence is composed of sedimentary facies that are typical of cave
entrances, where autochthonous and allochthonous inputs easily mingle, in a con-
text dominated by slope or gravitational dynamics (e.g., Ferrier, 2002). The deposit
results from several episodes of sediment slope-wash, with hiatuses marked by the
precipitation of carbonates over the exposed surface of the sediment, generating
the flowstones observed between layers.
The base of the succession (the Mousterian Cone) is ca. 2 m beyond the present
bottom of the excavation, which, in a 2006 test trench, reached the base of layer 19.
A major unconformity separates the top of layer 19 from the overlying sequence.
Above that unconformity and up to layer 15 (i.e., during the ca. 50–60 ka time span),
the deposit was mainly formed by the action of surface waters through runoff, with
slope processes leading to the reworking of the archaeological record, originally
accumulated at the mouth of the cave but syn- or post-depositionally washed into the
Access Corridor. Close-range stone tool refits observed during excavation indicate
that these surface processes did not significantly affect the original stratigraphic
configuration of the displaced remains.
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
303
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 303

After the deposition of layer 15, the fall of large boulders at the entrance redrew
the inner space of the cave, creating a sort of depression at the back and inward
of the Access Corridor. Layers 9–14 accumulated in that depression, with the chang-
ing topography of the available space (dictated by the process of filling itself, in
combination with the irregular morphology of the cave walls) explaining their vari-
ation in thickness and horizontal extent. The site’s microtopography also explains
the shifting emplacement of human occupations. As shown by find distribution and
refitting patterns, corroborated by the presence of features and confirmed by micro-
morphology, humans at this time used the inner part of the site, located behind the
boulders obstructing the Access Corridor. The occupied area extended into the
27-S Chamber in layers 13–15 and into the Side Passage in layers 9–12. Sediments were
laid down through repeated pulses, with non-sedimentation hiatuses marked by
the percolation of carbonate-saturated seepage water leading to the formation of the
flowstones and carbonate crusts that separate the archaeological layers. A direct
link with the Exterior area was reestablished with the build-up of layer 8, ca. 38–35 ka,
at a time when the interior areas used during the accumulation of layers 9–14 had
become too shallow for humans, who then settled only at the entrance.
The geoarchaeological data also show that, in specific areas (such as the Access
Corridor), the Gruta da Oliveira sequence is partly affected by minor syn- and post-
depositional processes, as well as by penecontemporaneous disturbance (namely
hyena scavenging). These processes explain the vertical migration of individual
finds, over distances that rarely exceed 25 cm, evidenced by ongoing refitting work.
Given the fast depositional rates and attendant thickness of the different units, the
impact of such displacements on the stratigraphic integrity of the archaeological
assemblages is, however, relatively minor. Thus, although not fully corresponding to
sealed sediment packages containing archaeological collections dating to a time
interval of restricted duration, the stratigraphic units recognized during excavation
come sufficiently close to that ideal condition.
Dating
The ca. 35–38 ka dates for layer 8 of Gruta da Oliveira postdate by some five mil-
lennia the latest Mousterian of Iberian regions to the north of the Cantabro-Pyrenean
range, where the Mousterian ends before ca. 42 ka (Zilhão, 2006). Conceivably, this
difference could simply reflect incomplete sample decontamination, with the layer
8 results being minimum ages only.
Indeed, improvements in pretreatment have led to the production of more precise,
and generally older, results (sometimes by several millennia) for bone samples that
had been AMS-dated prior to 2005 (Higham, Jacobi, & Bronk Ramsey, 2006). For the
most part, however, these redatings concern samples closer to the limits of
the method (i.e., older than ca. 40 ka), and from sites where the non-collagen con-
taminant fraction arguably relates to modern or Holocene soil formation. In such
circumstances, even very small amounts of contamination suffice to produce a major
impact. For instance, in the case of the dates for layer 8 of Oliveira, a rejuvenation
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
304
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 304

of 4000 years could result from the presence in the samples of as little as 1% of con-
taminants with an age of 5000 radiocarbon years. Moreover, those Oliveira dates are
on burnt bone and, because of unsolved difficulties in the chemical characterisa-
tion of such samples, burnt bone dates tend to be treated as minimum ages anyway.
The samples from layers 8 and 9, however, come from variably indurated deposits
immediately underlying the k/7 flowstone and, as all other faunal and lithic remains
from these layers, were themselves covered by thin carbonate coatings. As discussed
above, this process of microsparite cementation must be either synchronous to the
formation of the layers themselves or related to the deposition of the k/7 flowstone.
The implication is that any contamination by younger carbon must have occurred no
later than the time of formation of that flowstone, which, pending results from ongo-
ing U-Th dating, can be constrained to late Oxygen Isotope State (OIS)-3 or early
OIS-2 times because (1) the k/7 flowstone was in turn overlain by the brecciated col-
lapse that completely filled up the Access Corridor and (2) the few finds made in that
collapse were of fossilized Pleistocene bone and of lithic artifacts of Lower or Middle
Paleolithic affinities.
Using a ball-park estimate of ca. 30,000 radiocarbon years for the age of the cemen-
tation process of unit k/7, 49% of non-collagen, contaminating carbon of that age
would have to remain in the samples from layer 8 for results rejuvenated by 4000 years
to be obtained. Such a poor level of decontamination is extremely unlikely even for
burnt bone samples that, as in this case, underwent the routine pretreatments used
in dating labs through the last decade. Moreover, if such a level of contamination per-
tained in the Oliveira samples, it would have been impossible to obtain the Beta-
111967 result for level 9 (40,420 ⫾1220 14C yr B.P.; 44,080 ⫾990 cal yr B.P.; Table I);
even for a sample of infinite age, a radiocarbon age measurement of 40,420 is impos-
sible if ⬎27.5% of the extracted carbon is only 30,000 radiocarbon years old.
If the true age of level 8 were in fact of some 36,740 14C years (i.e., no later than
the Protoaurignacian), and if a (still quite high) percentage of 27.5% of 30,000 14C yr
B.P. contaminants existed in the samples from level 8, the resulting rejuvenation
would be of no more than a couple of millennia, that is, insufficient to produce the
32,740 14C yr B.P. result obtained for OxA-8671. A corollary of this conclusion is that,
even in such an extreme scenario, the radiocarbon evidence still implies an age
younger than 35,000 14C yr B.P. for the dated burnt bones from level 8.
The likelihood that sediment-derived carbon is present in burnt bone samples
can also be assessed through their d13C values. The ranges reported by Stuiver and
Polach (1977) are ⫺22 to ⫺27‰ for wood and charcoal, and ⫺18 to ⫺22‰ for bone
collagen, with five out of the six burnt bone samples for levels 8–11 of Oliveira falling
within the latter interval. Moreover, the two GrA results for levels 8 and 9, despite
having been obtained on the alkaline fraction, which contains the potential con-
taminants, are statistically identical to those reported by OxA and Beta for the cor-
responding levels. This evidence implies that any potential contaminants remaining
in the five samples with acceptable d13C values are indeed of broadly the same age
as the dated bone, as inferred from the geological context. By the same token, treat-
ment of the sixth result (OxA-9379 for level 11) as a minimum age is justified by its
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
305
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 305

anomalous d13C value (⫺23‰), which falls in the range of plant material, indicating
incomplete decontamination (and the same applies to the level 18 burnt bone result,
for which an identical d13C determination was obtained).
Since, in cave deposits, burnt bone is anthropogenic by definition, the results for
layer 8 of Oliveira date human behavior and, if the burning is accidental, one that may
well relate to hearths that actually postdate the bone (even if only marginally). The
stone tools in layer 8 are Middle Paleolithic, and there are no overlying Upper
Paleolithic layers from where the dated samples could have been downward dis-
placed. In short, by their nature, geological context, and chemical composition, the
ca. 35–38 ka burnt bone dates for layer 8 do indicate a persistence of the Middle
Paleolithic in Portugal well into the time range of the Protoaurignacian and
the Aurignacian I of Iberian regions situated to the North of the Ebro Frontier.
Paleoenvironmental Reconstruction
Throughout OIS-2, the North Atlantic polar front oscillated between the latitude
of Galicia, during cold pulses, and that of the Algarve, during warm ones. A similar
pattern can be inferred for OIS-3, as indicated by the presence of ice-rafted debris
(denoting so-called Heinrich Events) in marine cores off the coast of the Iberian
Peninsula, as far south as 37°N (Baas et al., 1997; Hemming, 2004).
Such a pattern of alternating and markedly contrasted sea surface conditions
would have exposed the climates and environments of littoral central Portugal to
cyclical, rapid, and rather dramatic shifts. The pattern of rhythmic succession of
episodes of sedimentation build-up alternating with episodes of flowstone forma-
tion observed at Gruta da Oliveira is in all likelihood a reflection of these changes
in the sea surface temperatures and corresponding impact on adjacent land. Given
currently available chronometric controls and the number of episodes of flowstone
formation already recognized in the sequence of Oliveira layers 8–19, it is possible
that these episodes coincide with the interstadial phases of individual D/O cycles
as defined in Greenland ice cores. In particular, the major sedimentation hiatus
between layers 8 and 9 probably corresponds to GIS-8, whose onset is marked by a
significant spike in the d18O curve and, according to current chronological models,
was rather long lasting (Shackleton et al., 2004).
The possibility of establishing one-to-one correlations between sedimentary units
and individual episodes of the detailed curve of global climate change drawn from
finer resolution archives is exceptional for cave sites. Certainly, in Iberian regions
to the south of the Ebro drainage, no record of similar resolution has so far been pub-
lished, and those sequences that span the later part of the Middle Paleolithic (e.g.,
Beneito, Bajondillo, Gorham’s Cave, Sima de las Palomas; Zilhão, 2006; Walker
et al., 2008) are, if not insufficiently documented, truncated by erosional processes
that prevent direct comparison with Oliveira.
An important implication of the sediment build-up pattern seen at Gruta da Oliveira
is that any remains of human or animal uses of the site abandoned in the cave dur-
ing warm-phase sedimentation hiatuses are likely to have become incorporated in
preexisting deposits through ordinary trampling and bioturbation processes. As a
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
306
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 306

result, the overall stratigraphic integrity of the different units should not be taken as
an indication that their content reflects a homogeneous climatic–environmental sig-
nal. On the contrary, it is to be expected that each stratigraphic unit recognized dur-
ing excavation and analysis will contain material that accumulated in the cave
throughout the entire duration of one or more D/O cycles: syn-depositionally, for
material indicative of colder conditions, when the reduction in vegetation cover
would have favored sediment slope-wash; post-depositionally, for material indicative
of milder conditions, when an expanded vegetation cover would have fixed exte-
rior soil, and any remains of animal or human uses of the stabilized cave floor would
penetrate into at least the upper part of the subsurface sediments accumulated in the
preceding build-up phase.
This finding has broader implications, and is a cautionary tale to bear in mind
when interpreting the paleoenvironmental archives from areas where even minor
oscillations of global climate would have generated fast and dramatic biome responses,
as in those parts of northern and central Europe situated at the edge of OIS-3 ice
caps. As the Gruta da Oliveira example illustrates, the “mixed” faunal assemblages
without modern analogue long recognized to be a feature of cave sites in those parts
of the world (Van Andel & Davies, 2003) may well be, in fact, a reflection not of coeval
environments but of the properties of the sediment accumulation processes.
CONCLUSION
The overall stratigraphic integrity of the Gruta da Oliveira deposit is of particu-
lar relevance because of the site’s broader implications for the debate on the delayed
survival of the Middle Paleolithic, and Neanderthals, in southern and western Iberia
(Zilhão, 2000, 2006; Finlayson et al., 2006; Vaquero, 2006; Zilhão & Pettitt, 2006). In
such a context, the radiocarbon dates for layer 8 provide sufficient evidence of
that survival until ca. 35–38 ka, that is, into the time range occupied by
the Protoaurignacian and the Aurignacian I in regions situated to the North of the
Cantabro-Pyrenean mountain range. In Greenland ice cores, this period of
Neanderthal persistence coincides with a rather long interval of high d18O values—
the GIS-8 interstadial. The resolution of the site’s sequence promises to yield detailed
comparison with global and regional archives of Upper Pleistocene climate change,
but sound correlation requires additional chronometric evidence and will be the
object of future research.
Over the years, the ongoing excavation of the Gruta da Oliveira has been supported by grants from dif-
ferent agencies and benefited from the logistical support of several regional and national entities:
Associação Arqueológica do Algarve, Câmara Municipal de Torres Novas, CRIVARQUE Lda.,
Junta Nacional de Investigação Científica e Tecnológica, Instituto Português do Património Arquitectónico
e Arqueológico, Parque Natural das Serras de Aire e Candeeiros, Sociedade Torrejana de Espeleologia e
Arqueologia, RENOVA S.A. Since 2006, fieldwork has been carried out in the framework of PALEOAL-
MONDA III, a research project financed by an associated grant of the Instituto Português de Arqueologia.
Thoughtful reviews by an associate editor, one anonymous referee, and Mike Morley are gratefully
acknowledged. We also thank Dr. António Monge Soares for his help in the interpretation of radiocarbon
dating results.
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
307
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 307

REFERENCES
Angelucci, D.E. (2002). The geoarcheological context. In J. Zilhão & E. Trinkaus (Eds.), Portrait of the artist
as a child: The Gravettian human skeleton from the Abrigo do Lagar Velho (pp. 58–91). Trabalhos de
Arqueologia, 22. Lisboa: Instituto Português de Arqueologia.
Baas, J.H., Mienert, J., Abrantes, F., & Prins, M.A. (1997). Late Quaternary sedimentation on the Portuguese
continental margin: Climate-related processes and products. Palaeogeography, Palaeoclimatology,
Palaeoecology, 130, 1–23.
Bertran, P., & Texier, J.-P. (1999). Facies and microfacies of slope deposits. Catena, 35, 99–121.
Boschian, G. (1997). Sedimentology and soil micromorphology of the Late Pleistocene and Early Holocene
deposits of Grotta dell’Edera (Trieste karst, NE Italy). Geoarchaeology, 12, 1–23.
Brewer, R. (1976). Fabric and mineral analysis of soils, 2nd ed. New York: Robert E. Krieger.
Bullock, P.N., Fedoroff, N., Jongerius, G.J., Stoops, G., Tursina, T., & Babel, U. (Eds.). (1985). Handbook
for soil thin section description. Wolverhampton, U.K.: Waine Research Publishers.
Cabral, J. (1995). Neotectónica em Portugal continental. Memórias do Instituto Geológico e Mineiro, 31.
Lisboa: Instituto Geológico e Mineiro.
Chabai, V.P., Sitlivy, V.I., & Marks, A.E. (2000–2001). Lower Paleolithic industry of Brecha das Lascas,
level 7 (Portugal). Préhistoire Européenne, 16–17, 17–41.
Courty, M.-A., Goldberg, P., & Macphail, R.I. (1989). Soils and micromorphology in archaeology. Cambridge,
U.K.: Cambridge University Press
Ferrier, C. (2002). Les dépôts d’entrée de grotte. In J.C. Miskovsky (Ed.), Geólogie de la préhistoire:
Méthodes, techniques, aplications (pp. 189–206). Paris: Association pour l’étude de l’environnement
géologique de la préhistoire.
Finlayson, C., Giles, F., Rodríguez-Vidal, J., Fa, D.A., Gutierrez, J.M., Santiago, A., Finlayson, G., Allué, E.,
Baena, J., Cáceres, I., Carrión, J.S., Fernández- Jalvo, Y., Gleed-Owen, C.P., Jimenez, F.J., López, P., López-
Sáez, J.A., Riquelme, J.A., Sánchez, A., Giles, F., Brown, K., Fuentes, N., Valarino, C.A., Villalpando, A.,
Stringer, C.B., Martínez, F., & Sakamoto, T. (2006). Late survival of Neanderthals at the southernmost
extreme of Europe. Nature, 443, 850–853.
Frisia, S., Borsato, A., Fairchild, I.J., & McDermott, F. (2000). Calcite fabrics, growth mechanisms, and envi-
ronments of formation in speleothems from the Italian Alps and southwestern Ireland. Journal of
Sedimentary Research, 70, 1183–1196.
Gé, T., Courty, M.-A., Matthews, W., & Wattez, J. (1993). Sedimentary formation processes of occupation
surfaces. In P. Goldberg, P.T. Nash, & M.D. Petraglia (Eds.), Formation processes in archaeological con-
text (pp. 149–164). Monographs of World Archaeology, 17.
Goldberg, P., & Macphail, R.I. (2006). Practical and theoretical geoarchaeology. Oxford: Blackwell.
Hemming, S.R. (2004). Heinrich events: Massive Late Pleistocene detritus layers of the North Atlantic
and their global climate imprint. Review of Geophysics, 42, 143.
Higham, T.F.G., Jacobi, R.M., & Bronk Ramsey, C. (2006). AMS radiocarbon dating of ancient bone using
ultrafiltration. Radiocarbon, 48, 179–195.
Jöris, O., Álvarez, E., & Weninger, B. (2003). Radiocarbon evidence of the Middle to Upper Paleolithic
transition in southwestern Europe. Trabajos de Prehistoria, 60, 15–38.
Karkanas, P., Bar-Yosef, O., Goldberg, P., & Weiner, S. (2000). Diagenesis in prehistoric caves: The use of
minerals that form in situ to assess the completeness of the archaeological record. Journal of
Archaeological Science, 27, 915–929.
Maíllo, J.M., Cabrera, V., & Bernaldo de Quirós, F. (2004). Le débitage lamellaire dans le Moustérien final
de Cantabrie (Espagne): Le cas de El Castillo et Cueva Morin. L’Anthropologie, 108, 367–393.
Manuppella, G., Balaçó Moreira, J.C., Graça e Costa, J.R., & Crispim, J.A. (1985). Calcários e dolomitos
no Maciço Calcário Estremenho. Estudos Notas e Trabalhos, 27, 3–48.
Marks, A., Monigal, K., & Zilhão, J. (2001). The lithic assemblages of the Late Mousterian at Gruta da
Oliveira, Almonda, Portugal. In J. Zilhão, T. Aubry, & A.F. Carvalho (Eds.), Les premiers hommes mod-
ernes de la Péninsule Ibérique. Actes du Colloque de la Comission VIII de l’UISPP, Vila Nova de Foz
Côa, Octobre 1998 (pp. 145–154). Trabalhos de Arqueologia, 17. Lisboa: Instituto Português de
Arqueologia.
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
308
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 308

Marks, A.E., Brugal, J.-P., Chabai, V.P., Monigal, K., Goldberg, P., Hockett, B., Peman, E., Elorza, M., &
Mallol, C. (2002). Le gisement Pléistocène Moyen de Galeria Pesada (Estrémadure, Portugal): Premiers
résultats. Paléo, 14, 77–99.
Martins, A. (1949). Maciço calcário estremenho: Contribução para um estudo de geografia física. Coimbra:
Universidade de Coimbra.
Paço, A., Vaultier, M., & Zbyszewski, G. (1947). Gruta da Nascente do Rio Almonda. Trabalhos de
Antropologia e Etnologia, 11, 171–187.
Raposo, L. (1995). Ambientes, territorios y subsistencia en el Paleolítico Medio de Portugal. Complutum,
6, 57–77.
Ribeiro, A., Antunes, M.T., Ferreira, M.P., Rocha, R.B., Soares, A.F., Zbyszewski, G., Moitinho de Almeida,
F., Carvalho, D., & Monteiro, J.H. (1979). Introduction à la géologie générale du Portugal. Lisboa:
Serviços Geológicos de Portugal.
Ribeiro, O. (1970). Génèse et diversité des montagnes portugaises. Argumenta Geographica, 12, 214–224.
Shackleton, N.J., Fairbanks, R.G., Chiu, T., & Parrenin, F. (2004). Absolute calibration of the Greenland
time scale: Implications for Antarctic time scales and for ⌬14C. Quaternary Science Reviews, 23,
1513–1522.
Stoops, G. (2003). Guidelines for analysis and description of soil and regolith thin sections. Madison, WI:
Soil Science Society of America.
Stuiver, M., & Polach, H.A. (1977). Discussion: Reporting of 14C data. Radiocarbon, 19, 355–363.
Trinkaus, E., Maki, J., & Zilhão, J. (2007). Middle Paleolithic human remains from the Gruta da Oliveira
(Torres Novas), Portugal. American Journal of Physical Anthropology, 134, 263–273.
Trinkaus, E., Marks, A.E., Brugal, J.-P., Bailey, S.E., Rink, W.J., & Richter, D. (2003). Later Middle Pleistocene
human remains from the Almonda karstic system, Torres Novas, Portugal. Journal of Human Evolution,
45, 219–226.
Van Andel, T.H., & Davies, W. (Eds.). (2003). Neanderthals and modern humans in the European land-
scape during the last glaciation. Cambridge, U.K.: McDonald Institute for Archaeological Research.
Vaquero, M. (2006). El tránsito Paleolítico Medio/Superior en la Península Ibérica y la frontera del Ebro:
Comentario a Zilhao (2006). Pyrenae, 37, 107–129.
Walker M.J., Gibert, J., López, M.V., Vincent Lombarda, A., Pérez-Pérez, A., Zapata, J., Ortega, J., Higham,
T., Pike, A., Schwenninger, J.L., Zilhão, J., & Trinkaus, E. (2008). Late Neandertals in southeastern
Iberia: Sima de las Palomas del Cabezo Gordo, Murcia, Spain. Proceedings of the National Academy
of Science, Early Edition, www.pnas.org/cgi/doi/10.1073/pnas.0811213106.
Weninger, B., & Jöris, O. (2004). Glacial radiocarbon calibration: The CalPal program. In T. Higham, C. Bronk
Ramsey, & C. Owen (Eds.), Radiocarbon and archaeology. Fourth International Symposium, Oxford,
2002.
Woodward, J.C., & Goldberg, P. (2001). The sedimentary records in Mediterranean rockshelters and caves:
Archives of environmental change. Geoarchaeology, 16, 327–354.
Zbyszewski, G., Manuppella, G., & Veiga Ferreira, O. (1971). Carta geológica de Portugal na escala de
1/50.000. Folha 27-C—Torres Novas. Lisboa: Serviços Geológicos de Portugal.
Zilhão, J. (1993). Le passage du Paléolithique moyen au Paléolithique supérieur dans le Portugal. In
V. Cabrera, (Ed.), El origen del hombre moderno en el suroeste de Europa (pp. 127–145). Madrid:
Universidad Nacional de Educación a Distancia.
Zilhão, J. (1997). O Paleolítico Superior da Estremadura portuguesa, 2 vols. Lisboa: Colibri.
Zilhão, J. (2000). The Ebro frontier: A model for the late extinction of Iberian Neanderthals. In C. Stringer,
R.N.E. Barton, & C. Finlayson (Eds.), Neanderthals on the edge: 150th anniversary conference of the
Forbes’ Quarry discovery, Gibraltar (pp. 111–121). Oxford: Oxbow Books.
Zilhão, J. (2001). Radiocarbon evidence for maritime pioneer colonization at the origins of farming in
west Mediterranean Europe. Proceedings of the National Academy of Sciences, 98, 14180–14185.
Zilhão, J. (2006). Chronostratigraphy of the Middle-to-Upper Paleolithic transition in the Iberian Peninsula.
Pyrenae, 37, 7–84.
Zilhão, J., Maurício, J., & Souto, P. (1991). A arqueologia da Gruta do Almonda: Resultados das escav-
ações de 1988–89. In Actas das IV Jornadas Arqueológicas (pp. 161–171). Lisboa: Associação dos
Arqueólogos Portugueses.
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
STRATIGRAPHYAND FORMATION PROCESSES
309
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 309

Zilhão, J., Maurício, J., & Souto, P. (1993). Jazidas arqueológicas do sistema cársico da nascente do
Almonda. Nova Augusta, 7, 35–54.
Zilhão, J., & McKinney, C. (1995). Uranium–thorium dating of Lower and Middle Paleolithic sites in the
Almonda karstic system (Torres Novas, Portugal). In Actas da 3aReunião do Quaternário Ibérico
(pp. 513–516). Coimbra: Universidade de Coimbra.
Zilhão, J., & Pettitt, P. (2006). On the new dates for Gorham’s Cave and the late survival of Iberian
Neanderthals. Before Farming, 2006/3, article 3.
Received 8 September 2008
Accepted for publication 15 January 2009
Scientific editing by Panagiotis Karkanas
ANGELUCCI AND ZILHÃO
GEOARCHAEOLOGY: AN INTERNATIONAL JOURNAL, VOL. 24, NO. 3
310
GEA243_04_20267.qxd 4/3/09 4:37 PM Page 310