The Pena d'Água rock-shelter (Torres Novas, Portugal): two distinct life ways within Neolithic sequence.

[page-n-1]
Del neolític a l’edat del bronze en el Mediterrani occidental.
Estudis en homenatge a Bernat Martí Oliver.
TV SIP 119, València, 2016, p. 211-223.

The Pena d’Água rock-shelter (Torres Novas, Portugal):
two distinct life ways within a Neolithic sequence
António FAustino CArvAlho

abstract

Among testimonies of other occupations, the Pena d’Água Rock-shelter revealed two “stratigraphic blocks” (i.e. sets of layers)
reporting, respectively, to the Early and Middle Neolithic, with preservation of organic matter (charcoal and faunal assemblages).
The older “stratigraphic block” is characterized by decorated pottery, use of local raw materials, and the balanced exploitation of a
large spectrum of animals (both domestic and wild), whereas the younger is characterized by plain pottery, long-distance exchange
of raw materials, and specialization in the hunting of cervids and herding of sheep/goat (itinerant pastoralism?). This cultural and
subsistence shift coincides with increasing aridity and the emergence of megalithism in Portugal, but sound correlations between
these phenomena are still to be made.
keywords:

Portugal, Neolithic, site formation process, economy, palaeo-environment.

resumen

El abrigo de Pena d’Água (Torres Novas, Portugal): dos modos de vida distintos en una secuencia neolítica. Entre los testimonios
de otras ocupaciones, el abrigo de Pena d’Água ha revelado dos “bloques estratigráficos” (es decir, conjuntos de capas) referidos,
respectivamente, al Neolítico antiguo y al Neolítico medio, con preservación de materia orgánica (restos de carbón y de fauna).
El “bloque estratigráfico” más antiguo se caracteriza por la cerámica decorada, el uso de materias primas locales y la explotación
equilibrada de un amplio espectro de animales (domésticos y salvajes), mientras que el más reciente viene determinado por la
cerámica lisa, el intercambio de materias primas a larga distancia y la especialización en la caza de cérvidos y el pastoreo de
ovicápridos (¿pastoreo itinerante?). Este cambio cultural y de subsistencia coincide con el aumento de la aridez y la aparición del
megalitismo en Portugal, si bien la correlación entre ambos fenómenos precisaría confirmarse.
palabras clave:

Portugal, Neolítico, proceso de formación de yacimientos, economía, paleoambiente.

1. INTRODUCTION
Promoted by the Servicio de Investigación Prehistórica, the resumption of the systematic study of the important Cova de l’Or
(Alicante, Spain) took place from 1975 onwards under the direction of B. Martí, who studied the remains devoid of a secure
context (Martí, 1977) and began excavations with stratigraphic
control in the cave’s Sector J (Martí et al., 1980). At this very
same time, Portuguese Prehistory was experiencing a major
methodological turning point regarding cave archaeology. As
widely acknowledged, after a promising beginning in the mid19th century, a decline in the quality of excavation methodologies would characterize most of the following century. It was
only in the 1970–80s that a renewed focus on stratigraphy,
recording of particular contexts (human-made structures, funerary practices, etc.), the introduction of so-called “ancillary
disciplines” (sedimentology, zooarchaeology, radiocarbon dating, etc.), and the thorough sieving of sediments, would become common procedures. Guilaine and Ferreira’s (1970) paper

on Early Neolithic pottery production in Portugal constitutes an
excellent example of the state of the art of cave research before this turning point: pottery chronologies were established
through stylistic comparisons with cave sequences from Spain
and France rather than by the provenance contexts of the vessels
themselves.
An evaluation and discussion of the changing methodologies and techniques put in practice in karst archaeology in
Portugal during the 1970–80s is beyond the scope of this text.
However, the work by B. Martí on the Iberian Neolithic and
cave archaeology impacted Portuguese research perhaps more
than usually perceived. Not only a new, sound stratigraphic sequence for the Neolithic in the peninsula was provided by the
Cova de l’Or sequence (Martí et al., 1980, 1987)—that would
soon become crucial as a comparison framework for the homologous evidence from Portugal—but also a critical perspective
on the understanding of cave deposits was introduced in the debate by the often-cited paper in collaboration with J. Fortea on
the beginnings of the Neolithic in Mediterranean Spain (Fortea
211

[page-n-2]
A. F. Carvalho

and Martí, 1984/85). These contributions would be developed
by J. Zilhão (1992, 1993) in his reinterpretation of several cave
sequences in the Iberian Peninsula and the Languedoc, which
would revolutionise the whole picture of the transition to farming in the western Mediterranean regions. This methodological turning point inspired my own methodological options in the
excavation, among other sites, of the long and rather complex
stratigraphic sequence discovered at the Pena d’Água Rockshelter (Torres Novas, Portugal).
The aim of this text is, thus, to present Pena d’Água by focusing on the current understanding of the processes underlying
the formation of its sedimentary deposit and, regarding the Neolithic period as recorded at the site, by noting the main changes
observable in both cultural (mainly through pottery production)
and economic (raw material and animal exploitation strategies)
behaviour. As will be shown, two main moments (corresponding to two “stratigraphic blocks”) within the Neolithic sequence were recognized and may bear relevant consequences for the
understanding of the Early and Middle Neolithic periods in the
southern regions of Portugal. This cultural record is here tentatively framed in the available palaeo-environmental data, either
locally obtained or derived from larger-scale proxies.

2. DISCOVERY AND DESCRIPTION
When, in December 1991, a local team of speleologists from the
Sociedade Torrejana de Espeleologia e Arqueologia (STEA)
used a backhoe to open the access to a seasonal spring through
the sedimentary deposit that was obstructing it—in a place
symptomatically named Pena d’Água (meaning literally “scarp
of the water”)—numerous fragments of pottery, knapped stone
and animal bones emerged from the removed earth. Immediately halted, this operation nonetheless enabled the discovery of
a rock-shelter in this sector of the Arrife, a local term derived
from the Arabic ar-rîff (“coast”, “scarp”) to name the several
kilometre-long fault escarpment that separates the Limestone
Massif of Estremadura from the Tagus Basin (Fig. 1).
Indeed, the inspection of the deposit profiles by J. Zilhão
and STEA members were able to confirm the presence of strata
with human occupation down to a depth of around 5 metres. The
scarp wall is very abrupt here—around 30 m in height, culminating at ca. 180 metres a.s.l. in its upper summits—whereas the
area in front of the Arrife forms a steep slope due to the accumulation of huge boulders collapsed from the rock-shelter’s roof.
The deposit reveals itself through a very prominent topography
visible at the foot of the limestone scarp (Fig. 1), reaching 125–
130 metres a.s.l. It has an estimated length of around 70 metres,
with a NE–SW orientation, only affected by the opening of the
access at its easternmost end.
At a regional scale, it should be emphasised that there are,
along the foot of the Arrife, successive sedimentary deposits
whose archaeological interest is evident. These, however, lie
under thick vegetation cover (shrubs, olive and oak trees, some
of them centennial) and, most probably, collapsed rock-shelters,
as in the case of Pena d’Água. Only very intrusive actions may
be able to identify similar sites and evaluate their archaeological
potential. This was what happened at Pena d’Água, permitting
archaeological excavations to be carried out during eight short
field seasons between 1992 and 2000.
212

In a first stage (1992–1995), a 2×3m test pit was excavated
in squares L29-30 (Fig. 2). In 1997, this was extended to an
adjacent area of 3×7m, corresponding to the I-K/25-30 squares,
which allowed the observation that the tip of the deposit was
affected by hydrologic processes caused by the local spring,
resulting in the thinning of the strata and their truncation by
erosive channels in the lower layers. Given these limitations,
the excavation had to be focused on the rectangle represented
by squares I-K/29-30 in the last, third stage of excavations at the
site (1998–2000). Overall, the excavation is very limited in area
and only further excavations, in the upper sector of the deposit,
would permit the recovery of sounder evidence related to the
Neolithic occupation of the site.
The above work resulted in several publications. After a
first modelling of the Neolithic sequence in the regional framework (Zilhão and Carvalho, 1996), a set of studies was
published in the 1998 volume of the Revista Portuguesa de
Arqueologia, where detailed accounts of the 1992–1995 fieldwork, site formation processes and human occupations were
made (Carvalho, 1998a) along with studies on anthracology
(Figueiral, 1998), insectivores and rodents (Póvoas, 1998),
and zooarchaeology of larger mammals (Valente, 1998).
More recently, a geo-archaeological analysis of the excavated
deposit has also been carried out (Simões, 2012) and further
zooarchaeological studies of Early (Carvalho, Valente and
Haws, 2004) and Middle Neolithic (Luís, Correia and Fernandes, n.d.) assemblages from the 1998–2000 seasons have
been published.
The Early Neolithic has been the occupation phase to which
a larger number of studies have been devoted: alongside a more
complete approach (Carvalho, 2008a), lithic techno-typological
and use-wear analyses (Carvalho, 1998b; Carvalho and Gibaja, 2005; Gibaja and Carvalho, 2005) and ceramic provenance
studies (Masucci and Carvalho, 2015) have also been carried
out. Other occupations were also published: Medieval and/or
Modern potsherds from Layer A (Ferreira, 1998), Iron Age pottery and radiocarbon determinations from Layer B (Carvalho,
2008b), and the Epipalaeolithic occupation from Layer F (Pereira and Carvalho, 2015). A brief synthesis of the site is available
in Spanish (Carvalho, 2012: 193–196).

3. STRATIGRAPHY AND SITE FORMATION
PROCESSES
Immediately after its discovery, the upper stratigraphic unit,
formed by huge limestone boulders, was observed to be the
result of the rock-shelter’s collapse. All the underlying units
with remains of human occupation were sealed under it. At
first, this conditioned the excavation methodology (Fig. 3):
picks and shovels (and explosives whenever necessary) had to
be used to remove this deposit (thereafter, Layer A) and reach
archaeologically-rich layers. Sediments were then excavated
with trowels and systematically dry-sieved using a 3mm mesh
screen. Bulk samples of unsieved sediments (10 litres per artificial level and unit square) were collected for flotation (presently in course) in order to recover very small-sized elements
(microfauna, seeds, etc.). Strata were subdivided in 5 or 10cm
thick arbitrary levels and materials were given 3D coordinates
as exhaustively as possible.

[page-n-3]
The Pena d’Água rock-shelter (Torres Novas, Portugal): two distinct life ways within a Neolithic sequence

Fig. 1. Location of the Pena d’Água Rock-shelter. A: location in Portuguese Estremadura; B: Google image of the Arrife, with
the limestone plateau on the right and the plains of the Tagus Valley on the left; C: photo of the Arrife with indication of the
excavated sector (note the rising topography to the left of the arrow due to the rock-shelter deposit beneath the vegetation cover).

This methodology was able to identify nine main stratigraphic units (Fig. 4), designated, from top to bottom, Layers
A to F (for a short description of each layer and its respective
cultural assignment, see Table 1). Unfortunately, bones preserved insufficient collagen for AMS dating (despite the systematic attempts) and, therefore bulk charcoal samples (mostly of
long-lived species) were the only usable type of sample, which
implies severe limitations regarding their correlation with human occupational events, the only exceptions being samples of
charcoal exhumed from hearths in Layers B and Eb-top (Table
2). However, for the reconstitution of site formation processes
the available determinations may be of some help (see below).

Indeed, what seems at first glance to be a simple, straightforward stratigraphic sequence is, however, the result of a rather
complex interaction of multiple phenomena. Six main phases
in the formation history of the Pena d’Água Rock-shelter deposit can be preliminarily proposed based on stratigraphic observations during excavation, geo-archaeological analyses and
various proxies of palaeoenvironmental nature, either local, regional or even global (Table 3).
The earliest, Phase 1, corresponds to the formation of layer F,
which took place under a very distinctive condition: according to
field observations (Carvalho, 1998a) and sedimentological analysis (Simões, 2012), its accumulation was due to the circulation
213

[page-n-4]
A. F. Carvalho

Fig. 2. Excavation plan of the Pena d’Água at top of layer B.
Marked squares refer to the excavated area: L-N/29-30 in 1992–
1995 and I-K/29-30 in 1998–2000. The blank area on the upper
right corresponds to the access opened in 1991.

of water from nearby springs, probably through the remobilizing
of the Miocene substratum. Technologically and typologically, its
artefacts are of Epipalaeolithic age. This was confirmed by a radiocarbon result, whose calibration makes it broadly coeval with
the 8.2 kyr cold event (or Bond 5) which caused major changes
in human settlement and mobility in the Estremadura and lower
Tagus region (e.g., Pereira and Carvalho, 2015).

After a first episode of roof collapse in an indeterminate
moment in time (Phase 2), apparently associated with a sedimentary hiatus, a constant accumulation of sediments with an
argillaceous component started to take place (Phase 3). Spring
activity resulted in the truncation of layers by channels and in
more or less severe—depending to the channels’ topography
and depth—disturbance of the archaeological horizons. The
whole formation of layers Eb to Ea, dated in the Early Neolithic, occurred under such environmental conditions at the local
scale. As pointed out above, these limitations constrained the
excavation to Squares I-K/29-30 in the last stage of excavations
(1998–2000).
Phase 4 is particularly visible in profile due to a major change in stratification (Fig. 4): after a horizontal, slow accumulation
of sediments in the previous phase (evidenced by a horizontal
level of 30–40 cm large blocks on top of Layer Ea), Layers Db
to C, dated to the Middle Neolithic, were accumulated according to a SW–NE inclination (evidenced by a second, sloping
level of blocks of the same size). This changing sedimentation
angle was due to still unknown reasons. It is also associated
with a faster rate of sedimentation, reduction in its argillaceous
component and the presence of iron oxides in Layers Db and
Da, which may be related to in situ post-depositional alterations
associated to soil-forming processes and increasing aridity conditions (Simões, 2012). There is no evidence for spring activity
in this phase, which together with the precipitation of iron oxides is therefore congruent with the inferred aridity.
After the formation of Layer B, a long period of time with
no significant sedimentation must have taken place at Pena
d’Água, thus giving place to Phase 5. This sedimentary hiatus
and the continuous human occupation at the site resulted in an
archaeological palimpsest (Late Neolithic, Iron Age, Roman),
attested by mixed material culture items and disparate radiocarbon determinations (Carvalho, 1998a, 2008b). This means

Fig. 3. Evolution of the excavation works at the Pena d’Água Rock-shelter. A: before the beginning of the excavations (1992); B: during
the removal of boulders from layer A (1992); C: general overview of the excavated area in 1997 (note the boulders of the collapsed roof).

214

[page-n-5]
The Pena d’Água rock-shelter (Torres Novas, Portugal): two distinct life ways within a Neolithic sequence

Fig. 4. Stratigraphic profile of the
Pena d’Água (layer A removed). Note
the inclination of the upper layers (C
and Da) while the lower ones (Ea to F)
show a horizontal stratification.

Table 1. Summary of the stratigraphy and archaeology of the Pena d’Água Rock-shelter.
Layer

Stratigraphy

Archaeology

A

Big boulders (>2 tons) from collapsed roof and loose sediments of miscellaneous
colours with penetrating tree and shrub roots.

Scattered Medieval and
Modern potsherds.

B

Medium-sized clasts (10-15 cm) in a sandy-argillaceous matrix of greyish to brownish / reddish sediments away or closer to the shelter’s wall, respectively. In outer
squares (rows I–J) part of the matrix was slope-washed.

Palimpsest of Late Neolithic,
Iron Age and Roman occupations.

C

Small-sized clasts (<10 cm) in a sandy-argillaceous matrix of greyish to brownish/
reddish sediments away or closer to the shelter’s wall, respectively. In outer squares
(rows I–J) part of the matrix was slope-washed.

Middle-to-Late Neolithic
transition with Iron Age
intrusions.

Da

Scattered small-sized clasts (<10 cm), mostly in it top level, in abundant, more compacted sedimentary matrix of a sandy-argilleous sediments, of homogeneously greenish-to-brownish colours. Slope-wash phenomena not recorded. An alignment of
large blocks (30-40 cm) indicates an south–north inclination of the layer,

Middle Neolithic with Iron
Age intrusions.

Db

Scattered small-sized clasts (<10 cm), mostly in it top level, in abundant, more compacted sedimentary matrix of a sandy-argillaceous sediments, of homogeneously
greenish-to-brownish colours. An alignment of large blocks (30-40 cm) in the lower
level indicates a horizontal stratification.

Initial Middle Neolithic.

Ea

Loose sandy-argillaceous sediments with small-sized clasts (<10 cm), some weathered, with brownish colours.

Evolved Early Neolithic

Eb-top

Loose sandy-argillaceous sediments with small-sized clasts (<10 cm). Numerous
micro-fragments of charcoal induce a more greyish tonality in otherwise brownish
sediments.

Evolved Early Neolithic

Eb-bottom

Loose sandy-argillaceous sediments of brownish colours with small-sized clasts (<10
cm).

Cardial Neolithic

F

Yellowish (light-greenish when wet) coarse sands with small-sized clasts (<10 cm),
very dense and compacted. There are also larger blocks (40–60 cm) and fragments
of limestone tuff. It lies on top of a local Miocene substratum of very coarse sands
with pebbles.

Epipaleolithic

215

[page-n-6]
A. F. Carvalho

Table 2. Radiocarbon determinations for Pena d’Água Rock-shelter.
Sample provenance (1)

Lab code (2)

Sample type / species

C result (BP) Calibration BC (3)

ICEN-965 (St.)

bulk charcoal (Olea?) 2000±50

14

Calibration BP (3)

Roman
N29.B.2 [hearth]

161–132 BC (3.5%)
118 BC–87 AD (90.3%)
105–121 AD (1.6%)

2110–2081 (3.5%)
2067–1863 (90.3%)
1845–1830 (1.6%)

757–679 (21.0%)
672–402 (74.4%)
756–679 (19.0%)
671–396 (76.4%)

2706–2628 (21.0%)
2621–2351 (74.4%)
2705–2628 (19.0%)
2620–2345 (76.4%)

Iron Age
K29.Da.1 + K30.Da.1

Wk-9742 (AMS) bulk charcoal (Olea?) 2428±56

I30.Da.5

Wk-9215 (AMS) bulk charcoal (4)

2410±59

Beta-137945
(AMS)
Sac-1822 (St.)
ICEN-1147 (St.)

bulk charcoal (5)

4250±50 (6)

bulk charcoal (Olea?) 3430±60 (6)
bulk charcoal (Olea?) 5180±240

—
4522–3515 (95.0%)
3412–3405 (0.1%)
3399–3384 (0.3%)

—
6471–5464 (95.0%)
5361–5354 (0.1%)
5348–5333 (0.3%)

M29.Ea.2

ICEN-1148 (St.)

bulk charcoal (Olea?) 5170±200

K29.Ea.3.48 + K29.Ea.3.50

Wk-9743 (St.)

bulk charcoal
(Olea europaea)

4448–4416 (0.8%)
4404–3631 (93.9%)
3562–3536 (0.7%)
4998–4459 (95.4%)

6397–6365 (0.8%)
6353–5580 (93.9%)
5511–5485 (0.7%)
6947–6408 (95.4%)

4791–4580 (94.4%)
4567–4559 (1.0%)
4763–4759 (0.3%)
4728–4457 (95.1%)
—

6740–6529 (94.4%)
6516–6508 (1.0%)
6712–6708 (0.3%)
6677–6406 (95.1%)
—

5623–5011 (95.4%)
5777–5610 (91.8%)
5592–5563 (3.6%)
—

7572–6960 (95.4%)
7726–7559 (91.8%)
7541–7512 (3.6%)
—

Initial Middle Neolithic
N30.Db.4 + N29.Db.4
N29.Db.4
L29.Db4

—

—

Evolved Early Neolithic

5856±114

Evolved Early Neolithic
K29.Eb-t.1 [hearth]

Wk-16418
single charcoal
(AMS)
(Olea europaea)
K29.Eb-t.1.65 + K30.Eb.1.43 Wk-9744 (AMS) bulk charcoal
(Olea europaea)
L29.Eb-t.4.116
OxA (AMS)
single bone
(Ovis aries)

5831±40
5753±62
(7)

Cardial Neolithic
N29.Eb-b.10
K29.Eb-b.3
M29.Eb-b.11.154
N29.Eb-b.11.177

ICEN-1146 (St.) bulk charcoal (Olea?)
Wk-9214 (AMS) bulk charcoal
(Olea europaea)
OxA (AMS)
single bone
(Sus scrofa)
Wk (AMS)
single bone
(Ovis aries)

6390±150
6775±60

Wk-9213 (St.)

7370±110

(7)
(7)

—

—

Epipalaeolithic
K29.F.1

bulk charcoal
(Quercus suber)

6436–6034 (95.4%)

(1) Respectively: square, layer (t = top; b = bottom), artificial level, 3D coordination number.
(2) St.: standard determinations; AMS: accelerator determinations.
(3) Calibration with IntCal13 (Reimer et al., 2013) using version 4.2 of the OxCal program (Bronk-Ramsey, 2009).
(4) Bulk sample of short-lived plant species (Leguminosae, Arbutus unedo and Rosaceae or Ericaceae).
(5) Bulk sample of short-lived plant species (Leguminosae).
(6) Aberrant result (due to the mixture of charcoal of different ages).
(7) Abandoned due to lack of collagen.

216

8385–7983 (95.4%)

[page-n-7]
The Pena d’Água rock-shelter (Torres Novas, Portugal): two distinct life ways within a Neolithic sequence

Table 3. Preliminary phasing and associated events related to the formation of the deposit at the Pena d’Água Rock-shelter (1).
Phase Description

Local palaeoenvironment

Events (local, regional or global)

1

Deposition of layer F (by redeposition of the
Miocene substratum?).

Spring activity. Weathering of the
limestone wall. Oleo-Lentiscetum
association.

8.2 kyr event (~ 6.2 cal BC).

2

Limestone boulders collapsed, marking the layer
F–E interface (with sedimentary hiatus?).

Spring activity truncating top of
layer F.

1st stratigraphic discontinuity
(partial roof collapse).

3

Deposition of layers Eb and Ea by horizontal
accumulation (ratio: 1 cm / 32 yrs) of small-sized
sediments with argillaceous component.

Spring activity with formation
of erosional channels. OleoLentiscetum association.

Hearth dated to 4791–4580 cal
BC.

4

Deposition of layers Db, Da and C with increasing
accumulation pace (ratio: 1 cm / 2.5 yrs) of smallsized sediments with minor argillaceous component
(due to deforestation and/or increasing aridity).

Spring shutdown. Sedimentation
change indicating absence
of water. Oleo-Lentiscetum
association and degraded maquis.

2nd stratigraphic discontinuity
(strata inclination). 5.9 kyr event
(~ 3.9 cal BC).

5

Deposition of layer B followed by sedimentary
hiatus (archaeological palimpsest) with formation
of limestone tufa along the wall.

Forest growth?Presence of vine
and possible domesticated olive
tree.

3rd stratigraphic discontinuity
(sedimentary hiatus). Hearth
dated to 118 cal BC–87 cal AD.

6

Deposition of layer A, sealing the underlying deposit.

Definitive roof collapse.

60 cal BC earthquake.

(1) After stratigraphic, botanical and geo-archaeological studies by Carvalho (1998a, p. 52, updated), Figueiral (1998) and Simões (2012), respectively.

that the top of Layer B was the deposit surface for around
three millennia, from the end of the Neolithic to Roman times.
Limestone tufa deposits formed at the contact between Layer
B and the rock-shelter wall reinforce this conclusion. This sedimentary hiatus remains to be adequately explained; because
of reforestation due to human abandonment of the region in
the 4th–3rd millennia BC transition, as initially put forward
(Carvalho, 1998a).
Phase 6 corresponds to the final collapse of the rock-shelter
roof. Given the impressive size and number of the collapsed
boulders, only a major event could be responsible for such a
dramatic and sudden change in local topography. If one considers the radiocarbon result obtained from the Roman hearth
in Layer B: 118 cal BC–87 cal AD, at 90.3% probability (Table 2), the event may have been the catastrophic earthquake
(M=8.5) and tsunami that occurred around 60 cal BC in coastal Portugal and Galicia (Baptista and Miranda, 2009), with
an estimated impact similar to the Lisbon event of AD 1755
(M=8.5 ± 0.3).

4. THE EARLY AND MIDDLE NEOLITHIC:
MAJOR TRENDS IN MATERIAL CULTURE, LITHIC
ACQUISITION AND ANIMAL EXPLOITATION
STRATEGIES
Phases 3 and 4 in the site formation process (Table 3) cover, respectively, the Early and Middle Neolithic occupations recorded
at the Pena d’Água. Each period is represented by a “block” of
three stratigraphically and/or culturally independent, successive
layers (Table 1): Eb-bottom, Eb-top and Ea in the former period,
and Db, Da and C in the latter (Figs. 4 and 5).
In the following sections the most relevant aspects of their
material culture, lithic raw material acquisition and strategies of
animal exploitation will be described. It should be noted that the

two “stratigraphic blocks” are still unequally studied: whereas a
full inventory and analysis has been made of all the Early Neolithic contents, the Middle Neolithic material culture items are
still under study. However, a quantitative approach to pottery
and lithics in presented below for the first time.

4.1. early neolithic
A full study of the Early Neolithic “stratigraphic block” is provided by Carvalho (2008a: 56–62). Regarding material culture, the most relevant aspect is the presence of decorated pots.
If the Minimum Number of Vessels (hereafter MNV) is taken
into consideration, decorated specimens show decreasing percentages over time: 59% (7 out of 12 vessels), 41% (6 out
of 19) and 32% (9 out of 22), from Layer Eb-bottom to Ea,
which is a well-marked pattern (Fig. 5). Some other trends
at this level are observable. Cardial sherds were found in all
Early Neolithic layers but only form a coherent, systematic assemblage in Layer Eb-bottom (Fig. 6), where, despite the low
absolute numbers involved (MNV=2 plus one loose sherd),
they represent a qualitatively distinct production within a very
homogenous assemblage—channelled and corded pots represent the remaining decorated types—that match the stylistic variability of the earliest pottery productions in Portugal
(Carvalho, 2011). This was one of the findings that led to the
conclusion we were indeed facing a different archaeological
horizon at Pena d’Água. A higher stylistic diversity occurs in
Layers Eb-top and Ea, with varying percentages of impressions, incisions, “boquique”, “false acacia leaf” impressions,
and bowls with a incised line below the rim—or “sulco sob
o bordo”, the Portuguese term for this type-fossil that marks
the Early-to-Middle Neolithic transition in Southern Portugal
(e.g., Silva, 1987)—in Layer Ea. Recipient sizes are usually
medium or small, and of simple shapes (hemispherical, spherical) together with some short necked pots.
217

[page-n-8]
A. F. Carvalho

Fig. 5. Stratigraphic variation of decorated versus plain pottery and of main mammal species (from left to right, sheep/goat, cattle, red
deer and wild boar) and photo of profile.

Fig. 6. Examples of Early and Middle Neolithic pottery types from Pena d’Água. 1–2: cardial and impressed rimsherds from layer Ebbottom (Cardial Neolithic); 3–4: impressed and “boquique” sherds from layer Ea (Evolved Early Neolithic); 5–8: rimsherds decorated with
an incised line below the rim from layer Db (Initial Middle Neolithic).

218

[page-n-9]
The Pena d’Água rock-shelter (Torres Novas, Portugal): two distinct life ways within a Neolithic sequence

Until recently, it was commonly accepted that chert sources
in the Arrife would be restricted to its NE (near Ourém) and
SW (Rio Maior) ends (e.g., Zilhão, 1997), with the exception of
some minor sources of poor quality recorded elsewhere in the
area, “[...] at Paredinhas and at the Arrife summit between the
source of the Almonda River and Moitas Vendas” (Carvalho,
2003: 144; Portuguese original). Systematic geological surveys,
however, confirmed that chert sources do exist in the massif
(Aubry et al., 2014) and, in particular, in the area around Pena
d’Água (T. Pereira, pers. inf.). For this reason chert has commonly been considered a non-local, regional resource, whereas
quartzite and quartz are locally-available raw materials obtainable in the surrounding Quaternary terraces of the Tagus Valley (Carvalho, 1998b, 2008a). However, whatever its specific
geographical provenance, an internal trend can be observed in
the use of chert within the Early Neolithic (Table 4): from 51%
in Layer Eb-bottom, it decreases to 21% and 14% throughout
the succeeding, upper levels in favour of quartz and, especially,
quartzite. As put forward elsewhere, “[...] the most parsimonious
interpretation of this shift seems to be the existence of changes
in lithic resource procurement strategies in the passage from the

6th to the 5th millennia BC” (Carvalho, 2008a: 59; Portuguese
original), probably related with changes in human mobility. The
relative abundance of chert in the bottom of Layer Eb, associated with the cardial pottery, was also considered to be testimony
of a distinct archaeological occupation. The prevailing trend
during the period will be the preponderance of quartzite (Table
4), a fact also observable in most post-Cardial sites in the Arrife
region (Carvalho, 2008a). Lithic tools, in all raw materials, consist mostly of side-retouched or notched blanks—flakes in particular—while truncations, perforators or composite tools are
less common. Geometric microliths are dominated by segments,
from the Cardial onwards, and these represent around 10% of
the total chert tools.
Regarding animal exploitation strategies, there is no clear
diachronic trend within the Early Neolithic “stratigraphic
block” (Fig. 5). The only possible exception may be the absence
of cattle (Bos taurus) and/or aurochs (Bos primigenius) in the
Cardial occupation, but this is likely an artefact of the available
sample and the taphonomic conditions of Layer Eb-bottom. Indeed, the high total number of remains from the Cardial occupation sharply contrasts with the Number of Identified Specimens

Table 4. Lithic inventory of knapped stone raw materials in Pena d’Água Rock-shelter (1).
Layers

Quartzite

%

Quartz

%

Total

Eb-bottom

0.455

51

0.320

36

0.122

13

0.897

100

0.658

21

1.875

51

0.620

20

3.153

100

Ea

0.493

14

1.925

57

0.983

29

3.401

100

Db

0.396

3

9.271

73

2.962

24

12.629

100

Da

0.205

4

3.355

71

1.155

25

4.715

100

C

Middle Neolithic

%

Eb-top

Early Neolithic

Chert

%

0.050

<0

2.530

66

1.240

33

3.820

100

(1) Weight (in kg) and relative percentage of each raw material.

Table 5. Zooarchaeology of the Pena d’Água Rock-shelter (1).
Early Neolithic (2)
Species

Layer Eb-bottom

Layer Eb-top

Middle Neolithic (3)
Layer Ea

Layer Db

Layer Da

Layer C

1

Sheep (Ovis aries)

1

Goat (Capra hircus)
Sheep / goat (Ovis aries / Capra hircus)

2

11

3

Red deer (Cervus elaphus)

1

1

8

Undetermined cervids

1

5

5

Cattle / aurochs (Bos sp.)

7

5

Aurochs (Bos primigenius)

24

9

3

1

73

27

1

18

1

1

Wild boar (Sus cf. scrofa)

1

4

8

Rabbit (Oryctolagus cuniculus)

1

3

3

133

Cattle (Bos taurus)

Total NISP
Total Number of Remains (TNR)

4

6

34

33

160

83

50

155

118

118

260

108

145

(1) Birds and carnivores not included.
(2) Values based on the analyses by Valente (1998) and Carvalho, Valente and Haws (2004).
(3) Layer Db analysed by Valente (1998) and Luís, Correia and Fernandes (n.d.); layers Da and C analysed by Valente (1998).

219

[page-n-10]
A. F. Carvalho

(hereafter, NISP): 6 out of 155! This is further indirect evidence
for the destructive impact of the nearby springs during the formation of Layers Ea and Eb.
From a general viewpoint, the most striking aspect is the
very broad spectrum of species represented in the Early Neolithic at Pena d’Água, as evidenced in Table 5 (sheep, goat, red
deer, cattle, aurochs, swine and rabbit), which gains further
relevance in view of the relatively small NISP exhumed from
Layers Eb-bottom to Ea (NISP=73). Taken all three layers together, the domestic specimens are clearly outnumbered by wild
species (n=19 vs. n=35, rabbits excluded), but in this count swine have all been classed as wild boar (Sus scrofa), not domestic
pig (Sus domesticus), which is an assumption far from consensual in Portuguese Neolithic zooarchaeology (Valente and Carvalho, 2014). For example, at the Caldeirão Cave, in the near
Nabão Valley, Davis (2002) has reconsidered Rowley-Conwy’s
(1992) observations and concluded that most—if not all—swine
remains in this cave would belong the domestic species instead.

4.2. middle neolithic
Despite the relatively well preserved record observed in the
1992–1995 seasons in Layers C to Da in Squares L-N/29-30,
the 1998–2000 seasons exposed severe post-depositional disturbances affecting the same layers in Squares I-K/29-30 (Fig.
2). Intrusive Iron Age remains constituted here the majority of
both ecofacts (charcoal and fauna) and material culture items,
limiting the study of the Middle Neolithic to Layer Db only.
Plain vessels overwhelmingly dominate the pottery assemblage (Fig. 5). According to the MNV, they represent 100% in
Layers C and Da, with 29 and 56 individualized vessels, respectively. In Layer Db decorated vessels reach 9% (6 out of 71
vessels) and, among these, four correspond to the transitional
Early-to-Middle Neolithic type of bowls with an incised line below the rim (Fig. 6). If loose decorated sherds (those that could
not be securely associated with any individualized vessel) are
counted, the above scenario remains unchanged: 0.09% in Layer C (1 out of 1020 sherds); 0.4% in Layer Da (4 out of 911);
and 2.9% (17 out of 581). This is a very typical trait of so-called
“dolmenic potteries” in the southern regions of Portugal, where
decoration is rare during the Middle and Late Neolithic. With
the exception of a couple of carinated fragments from Layer
C, pots are of simple shapes (hemispheric and spherical) and
small sizes.
As evident in Table 5, quartzite abundantly dominates the
lithic assemblages. It already represented 50–60% of all exploited raw materials in the Early Neolithic but now reaches
about 70%, while chert decreases dramatically to 3%, 4% and
less than 0% in Layers Db, Da and C, respectively, testifying a
major change in the acquisition of this raw material. The type
of relation this trend has with the recently-acquired notion that
chert sources exist in the area (see above) is still to be clarified
(exploitation shutdown due to small nodule size? Oblivious of
its existence?). However, the fact that blades/bladelets represent 50% and 44% of the chert debitage in Layers C (3 out of
6 blanks) and Da (11 out of 25), respectively, indicates a rather
important technological change in the later stages of the Middle
Neolithic, now focused on the circulation of elongated blanks
rather than of cores and/or nodules. Indeed, both in Layer Db
and in the Early Neolithic the obtained ratios are quite distinct:
220

22% (33 blades/bladelets out of 148 blanks) in Layer Db, 31%
(29 out of 93) in Layer Ea, 23% (39 out of 169) in Layer Eb-top
and 22% (15 out of 67) in Layer Eb-bottom. Tool types remain
similar to those that characterize the Early Neolithic, with the
predominance of side-retouched or notched flakes. Only the microlithic assemblage suffers some changes: its total percentage
decreases and a balanced segment/trapezium composition of the
toolkits emerges.
As in pottery decoration, the zooarchaeological contrast
between the two Neolithic “stratigraphic blocks” could not be
sharper (Valente, 1998; Luís, Correia and Fernandes, n.d.). If
rabbits, birds and carnivores are excluded, in Layers Db to C
there are only cervids—exclusively red deer (Cervus elaphus)
whenever species is defined—alongside domestic sheep and/or
goat (Table 5; Fig. 5). Species that were exploited in the previous Early Neolithic occupations—such as swine and bovids,
either domestic or wild—seemed to have been excluded from
the regular animal exploitation strategies during the Middle
Neolithic. Given the species in question, these new strategies of
animal exploitation correspond to a rather more specialized and
mobile pastoralism in the framework of which hunting practices—focusing on red deer only—also took place. Different
NISP values between the Early and Middle Neolithic “stratigraphic blocks”, favouring the latter (NISP=83 vs. NISP=66,
rabbits excluded), is the best testimony that this reduction in
the species spectrum is not biased by sampling and must reflect
instead the past reality.

5. DISCUSSION AND CONCLUSIONS
A chronological framework based on a solid batch of radiocarbon determinations is surely the major limitation to the study
of the Neolithic at Pena d’Água. As shown in Table 2, only determination Wk-16418 obtained from charcoal collected from
a hearth in Layer Eb-top is relatively reliable for cultural inferences. The Early-to-Middle Neolithic boundary at the site is
rather blurred due to the similar results obtained by determinations ICEN-1148 from Layer Ea and ICEN-1147 from Layer Db
(Table 2), as evident from their plotting in Fig. 7. It can only be
concluded that the boundary may be situated some time around
the passage from the 5th to 4th millennium cal BC.
Palaeobotanic data from the Pena d’Água sequence seems to
denote variations in vegetation cover over time, subtle but likely
reflecting changing palaeoenvironmental conditions at broader
scales. In fact, the Neolithic vegetation is overall dominated by
wild olive tree (Olea europaea) in all layers, with ca. 93% and
81% of total taxa in layers Ea+Eb and Da+Db, respectively.
However, while in the Early Neolithic “[...] the presence of Olea
and Pistacia lentiscus suggests the Oleo-Lentiscutum association, with olive tree and mastic forming the arboreal and the
shrubby strata, respectively”, during the following period “[t]he
presence of arbutus, heather, rockrose and leguminosae seem to
be testimony of areas characterized by a degraded maquis (brushwood)” (Figueiral, 1998: 75; Portuguese original). Likewise, a preliminary analysis of the microfauna (Póvoas, 1998),
although devoid of quantitative data, showed that house mouse
(Mus musculus) and Algerian mouse (Mus spretus)—which are
indicative of farming activities—along with wood mouse (Apodemus sp.) were the dominant genera in all Neolithic layers but

[page-n-11]
The Pena d’Água rock-shelter (Torres Novas, Portugal): two distinct life ways within a Neolithic sequence

Fig. 7. Plotting of radiocarbon dates from Pena d’Água (Epipalaeolithic and Neolithic only) with GISP2 curve and correlation with
climatic events 8.2 and 5.9 (after Simões, 2012: fig. 9, adapted).

pine vole (Microtus (Terricola) sp.), a species usually inhabiting
moist habitats with thick vegetation cover, decreases in representativeness from the bottom to the top of the sequence.
Both types of biological evidence, coupled with results from
geo-archaeological analyses (Simões, 2012), converge on the
conclusion that changing bioclimatic conditions towards deforestation and aridity coincided with cultural changes. Put in
other words, the Early and the Middle Neolithic people settled
at Pena d’Água witnessed different environmental conditions
and seem to have experienced distinct life ways.
Indeed, in the former period, included in the Phase 3 of
the site formation history (Table 3), locally obtained resources clearly predominate. Even in the case of chert, which was
thought to have been brought to the site from sources located in
the northern or southern ends of the limestone massif (ca. 25 km
and 80 km, respectively), it is possible today to hypothesize its
local acquisition. Recent petrographic and geochemical analyses of sherd samples showed that Early Neolithic pottery was
also made with local clays and tempers (Masucci and Carvalho,
n.d.). Non-local raw materials are restricted to two schist flakes
from Layers Ea and Eb-top, suggesting episodic exchange or
travels to geological formations located 30–40 km to the northeast. Finally, the presence of cattle—and of pig, if the swine

remains are classed as domestic in the future (Valente and Carvalho, 2014)—indicates a more geographically restricted scale
of herding practices in the Early Neolithic.
The above data does not necessarily mean we are dealing
with fully sedentary human groups at Pena d’Água during this
time period; indeed some degree of mobility can be inferred
from small pottery sizes, expedient knapped stone exploitation
patterns, ephemeral site structures, etc., as observed not only
at this rock-shelter but also at other sites in the Arrife region
(Carvalho, 2003, 2008a). The open question is whether this is
a complete depiction of the whole settlement system or if there
are permanent settlements still to be found, for example in riverine locations near the Tagus or its tributaries where more fertile
soils exist and a fully developed farming economy would have
occurred, as documented elsewhere in Iberia.
In the Middle Neolithic, corresponding to the deposit’s
formation Phase 4 (Table 3), although characterized by the
use of local knapped raw materials (quartzite and quartz), the
chert assemblage reveals a new type of debitage economy: the
circulation of elongated blanks (blades and bladelets) tends
to replace the local knapping of chert cores and/or nodules.
This new trend in chert exploitation and use is likely associated with the acquisition of non-local raw materials, such as
221

[page-n-12]
A. F. Carvalho

schist (which is found in the form of knapped flakes), granite
(a grindingstone from Layer C) and amphibolite (one fragmented polished axe from Layer C and flakes from recycling
polished stone tools in Layers Db to C). The latter rock type
implies long-distance movements or exchange networks, since the nearest sources are located in the Hesperian Massif,
some 100km to the east. Zooarchaeological data indicate what
seems to be “economic specialization” in the herding of sheep/
goat—therefore suggesting itinerant pastoralism practices—
associated with red deer hunting.
In sum, there is a major change towards greater human mobility during the Middle Neolithic, a hypothesis perhaps more
probable than an increasing flow of goods in existing exchange
networks. If this is confirmed by future research to be a regional phenomenon—and not the biased effect of site function—it
bears far-reaching consequences in two aspects. First, it apparently parallels the model obtained from the study of the coeval Bom Santo Cave necropolis where a Neolithic community exploited a ca. 100km-long territory and practised itinerant
sheep/goat pastoralism as detected at Pena d’Água (Carvalho,
2014). This may mean that Middle Neolithic societies in Estremadura and neighbouring sectors of the Alentejo province were
organized as evidenced by both sites. Second, this moment in
time—the 5th-to-4th millennium transition onwards—corresponds to the onset and development of megalithism in central
and southern Portugal, which raises a series of questions regarding socioeconomic features of the human communities and
the palaeoenvironmental conditions (Fig. 7)—namely the role
played in this scenario by the 5.9 kyr climate event (Bond et
al., 1997)—underlying the advent of the above cultural phenomenon.
ACKNOWLEDGEMENTS
To the STEA members—mainly Pedro Souto and João
Maurício—for all the support during the several fieldwork
seasons at the Pena d’Água Rock-shelter, to Telmo Pereira
for providing unpublished information on the chert sources
of the Arrife, to Carlos Duarte Simões for comments on the
sedimentological analyses of the deposit, and to Joaquim
Juan-Cabanilles for the Spanish translation of the Abstract.
REFERENCES
AUBRY, T.; MANGADO, J. and MATIAS, H. (2014): “Matériasprimas das ferramentas em pedra lascada da Pré-História do centro e nordeste de Portugal”. In P. Dinis, A. Gomes, and S. Monteiro-Rodrigues (eds.): Proveniência de materiais geológicos:
abordagens sobre o Quaternário de Portugal. Porto: Associação
Portuguesa de Estudo do Quaternário, p. 165–192.
BAPTISTA, M.A. and MIRANDA, J.M. (2009): “Revision of the
Portuguese catalog of tsunamis”. Natural Hazards and Earth
System Sciences, 9, p. 25–42.
BOND, G.; SHOWERS, W.; CHESEBY, M.; LOTTI, R.; ALMASI,
P.; deMENOCAL, P.; PRIORI, P.; CULLEN, H.; HAJDAS, I.
and BONANI, G. (1997): “A pervasive millennial-scale cycle
in North Atlantic Holocene and glacial climates”. Science, 2768,
p. 1257–1266.
BRONK-RAMSEY, C. (2009): “Bayesian analysis of radiocarbon
dates”. Radiocarbon, 51 (1), p. 337–360.

222

CARVALHO, A.F. (1998a): “Abrigo da Pena d’Água (Rexaldia,
Torres Novas): resultados das campanhas de sondagem (19921997)”. Revista Portuguesa de Arqueologia, 1 (2), p. 39–72.
CARVALHO, A.F. (1998b): Talhe da pedra no Neolítico antigo do
Maciço Calcário das Serras d’Aire e Candeeiros (Estremadura Portuguesa). Um primeiro modelo tecnológico e tipológico.
Lisboa: Colibri.
CARVALHO, A.F. (2003): “O Neolítico antigo no Arrife da Serra
d’Aire. Um case study da neolitização da Média e Alta Estremadura
Portuguesa”. In V.S. Gonçalves (ed.): Muita gente, poucas antas?
Origens, espaços e contextos do Megalitismo. Lisboa: Instituto Português de Arqueologia (Trabalhos de Arqueologia; 25), p. 135–154.
CARVALHO, A.F. (2008a): A neolitização do Portugal meridional.
Os exemplos do Maciço Calcário Estremenho e do Algarve
ocidental. Faro: Universidade do Algarve (Promontoria
Monográfica; 12).
CARVALHO, A.F. (2008b): “Cerâmica estampilhada do Abrigo
da Pena d’Água (Torres Novas): contexto, cronologia e breve
enquadramento regional”. In J.P. Bernardes (ed.): Sic memorat.
Estudos em homenagem a Teresa Júdice Gamito. Faro: Universidade do Algarve, p. 53–68.
CARVALHO, A.F. (2011): “Produção cerâmica no início do Neolítico de Portugal”. In J. Bernabeu, M.A. Rojo, and L. Molina
(eds.): Las primeras producciones cerámicas: el VI milenio cal
AC en la Península Ibérica. València: Universitat de València
(Saguntum Extra; 12), p. 237–250.
CARVALHO, A.F. (2012): “Portugal”. In M.A Rojo, R. Garrido,
and I. García (eds.): El Neolítico en la Península Ibérica y su
contexto europeo. Madrid: Cátedra, p. 175–212.
CARVALHO, A.F. (2014): “Bom Santo Cave in context. A
preliminary contribution to the study of the first megalith
builders of Southern Portugal”. In A.F. Carvalho (ed.): Bom
Santo Cave (Lisbon) and the Middle Neolithic Societies of
Southern Portugal. Faro: Universidade do Algarve (Promontoria
Monográfica; 17), p. 209–230.
CARVALHO, A.F. and GIBAJA, J.F. (2005): “Talhe da pedra no
Neolítico antigo do Maciço Calcário Estremenho (Portugal):
matérias-primas, tecnologia e análise funcional”. III Congreso
del Neolítico en la Península Ibérica. Santander: Universidad
de Cantabria, p. 373–382.
CARVALHO, A.F.; VALENTE, M.J. and HAWS, J.A. (2004):
“Faunas mamalógicas do Neolítico antigo do Maciço Calcário
Estremenho: análise preliminar de dados recentes”. Promontoria, 2, p. 143–156.
DAVIS, S.J.M. (2002): “The mammals and birds from the Gruta
do Caldeirão, Portugal”. Revista Portuguesa de Arqueologia, 5
(2), p. 29–98.
FERREIRA, M.N. (1998): “As cerâmicas medievais / modernas do
Abrigo da Pena d’Água (Torres Novas)”. Revista Portuguesa de
Arqueologia, 1 (2), p. 97–106.
FIGUEIRAL, I. (1998): “O Abrigo da Pena d’Água (Torres Novas):
a contribuição da antracologia”. Revista Portuguesa de Arqueologia, 1 (2), p. 73–80.
FORTEA, J. and MARTÍ, B. (1984/85): “Consideraciones sobre los
inicios del Neolítico en el Mediterráneo español”. Zephyrus, 2728, p. 167–199.
GIBAJA, J.F. and CARVALHO, A.F. (2005): “Reflexiones en torno
a los útiles tallados en cuarcita: el caso de algunos asentamientos
del Neolítico antiguo del Macizo Calcáreo Extremeño (Portugal)”. Zephyrus, 58, p. 183–194.
GUILAINE, J. and FERREIRA, O.V. (1970): “Le Néolithique ancien au Portugal”. Bulletin de la Société Préhistorique Française, 67 (1), p. 304–322.

[page-n-13]
The Pena d’Água rock-shelter (Torres Novas, Portugal): two distinct life ways within a Neolithic sequence

LUÍS, S.; CORREIA, F. and FERNANDES, P.V. (n.d.): “Middle
Neolithic zooarchaeology at the Pena d’Água Rock-shelter (Portuguese Estremadura)”. VI Jornadas de Jóvenes en Investigación
Arqueológica. Barcelona: Universitat de Barcelona; in press.
MARTÍ, B. (1977): Cova de l’Or (Beniarrés, Alicante), vol. I.
Valencia: Servicio de Investigación Prehistórica (Serie de Trabajos Varios; 51).
MARTÍ, B.; FORTEA, J.; BERNABEU, J.; PÉREZ, M.; ACUÑA,
J.D.; ROBLES, F. and GALLART, M.D. (1987): “El Neolítico
antiguo en la zona oriental de la Península Ibérica”. In J. Guilaine, J.-L. Roudil, and J.-L. Vernet (dir.): Premières Communautés Paysannes en Méditerranée Occidentale. Paris: Centre
National de la Recherche Scientifique, p. 607–619.
MARTÍ, B.; PASCUAL, V.; GALLART, M.D.; LÓPEZ, P.;
PÉREZ, M.; ACUÑA, J.D. and ROBLES, F. (1980): Cova de
l’Or (Beniarrés, Alicante), vol. II. Valencia: Servicio de Investigación Prehistórica (Serie de Trabajos Varios; 65).
MASUCCI, M.A. and CARVALHO, A.F. (2015): “Ceramic
technology and resource use during the Neolithic in Central–
Southern Portugal”. Archaeometry. Doi: 10.1111/arcm.12206.
PEREIRA, T. and CARVALHO, A.F. (2015): “Abrupt technological change at the 8.2 ky cal BP climatic event in Central
Portugal. The Epipaleolithic occupation of the Pena d’Água
Rock-shelter”. Comptes Rendus Palevol. Doi: 10.1016/j.
crpv.2015.04.003.
PÓVOAS, L. (1998): “Faunas de micromamíferos do Abrigo da
Pena d’Água (Torres Novas) e seu significado paleoecológico:
considerações preliminares”. Revista Portuguesa de Arqueologia, 1 (2), p. 81–84.
REIMER, P.J.; BARD, E.; BAYLISS,A.; BECK, J.W.; BLACKWELL,
P.G.; BRONK-RAMSEY, C.; BUCK, C.E.; CHENG, H.;
EDWARDS, R.L.; FRIEDRICH, M.; GROOTES, P.M.;
GUILDERSON, T.P.; HAFLIDASON, H.; HAJDAS, I.; HATTÉ,
C.; HEATON, T.J.; HOFFMAN, D.L.; HOGG, A.G.; HUGHEN,
K.A.; KAISER, K.F.; KROMER, B.; MANNING, S.W.;

NIU, M.; REIMER, R.W.; RICHARDS. D.A.; SCOTT, E.M.;
SOUTHON, J.R.; STAFF, R.A.; TURNEY, C.S.M. and VAN
DER PLICHT, J. (2013): “IntCal13 and Marine13 radiocarbon
age calibration curves, 0-50,000 years cal BP”. Radiocarbon, 55
(4), p. 1869–1887.
ROWLEY-CONWY, P. (1992): “The Early Neolithic bones from
Gruta do Caldeirão”. In J. Zilhão (ed.): Gruta do Caldeirão.
O Neolítico antigo. Lisboa: Instituto Português do Património
Arquitectónico e Arqueológico (Trabalhos de Arqueologia; 6),
p. 231–257.
SILVA, C.T. (1987): “Megalitismo do Alentejo Ocidental e do Sul
do Baixo Alentejo (Portugal)”. El Megalitismo en la Peninsula
Ibérica. Madrid: Ministerio de Cultura, p. 85–93.
SIMÕES, C.D. (2012): “Estudio geoarqueológico del Abrigo de
Pena d’Água (Torres Novas, Portugal) en el contexto de la neolitización de la vertiente atlántica meridional peninsular”. Estudios de Cuaternario, 2, p. 49–69.
VALENTE, M.J. (1998): “Análise preliminar da fauna mamalógica
do Abrigo da Pena d’Água (Torres Novas). Campanhas de 19921994”. Revista Portuguesa de Arqueologia, 1 (2), p. 85–96.
VALENTE, M.J. and CARVALHO, A.F. (2014): “Zooarchaeology
in the Neolithic and Chalcolithic of Southern Portugal”. Environmental Archaeology, 19 (3), p. 226–240.
ZILHÃO, J. (1992): Gruta do Caldeirão. O Neolítico antigo.
Lisboa: Instituto Português do Património Arquitectónico e
Arqueológico (Trabalhos de Arqueologia; 6).
ZILHÃO, J. (1993): “The spread of agro-pastoral economies across
Mediterranean Europe: a view from the Far West”. Journal of
Mediterranean Archaeology, 6 (1), p. 5–63.
ZILHÃO, J. (1997): O Paleolítico Superior da Estremadura Portuguesa, 2 vols. Lisboa: Colibri.
ZILHÃO, J. and CARVALHO, A.F. (1996): “O Neolítico do Maciço
Calcário Estremenho: crono-estratigrafia e povoamento”. I
Congrés del Neolític a la Península Ibèrica, 2. Gavà: Museo de
Gavà (Rubricatum; 1), p. 659–672.

223

[page-n-14]