PIA Volume 22 (2012), 99-114 DOI: http://dx.doi.org/10.5334/pia.

405

Landscape and Astronomy in Megalithic

Portugal: the Carregal do Sal Nucleus and
Star Mountain Range
Fabio Silva*

Central Portugal, delimited by the Douro river to the north and the Mondego
to the south, is the second densest region of megalithic monuments in the
country. The Neolithic archaeological record indicates seasonal transhumance
between higher pastures in the summer and lower grounds in the winter. The
monuments are found in lower ground and it has been suggested that they
were built during the winter occupation of their surroundings. The astronomi-
cal orientation of their entrances lends further support to this hypothesis.
A recent survey of the orientation of the chambers and corridors of these
dolmens, conducted by the author, found good agreement with prior surveys,
but also demonstrated that other interpretations are possible. This paper
presents an update on the survey, including extra sites surveyed in the spring
of 2011, as well as the GIS confirmation of all horizon altitudes that couldn’t
be empirically measured. The megalithic nucleus of Carregal do Sal, on the
Mondego valley, is then looked at in more detail. It is found that there is a
preference for the orientation of dolmens towards Star Mountain Range in-
line with the topographic arguments of landscape archaeology. In addition, it
was found that the topography also marks the rise of particular red stars,
Betelgeuse and Aldebaran, during the period of megalithic building, at the
onset of spring marking the transition from low ground to the high pastures.
This hypothesis finds further support from toponymic folktales that explain
the origin of the name of the mountain range.

The Mondego platform of central Portugal usage since the fifth millennium BC (Cruz
is bordered by the Central Massif to the 2001: 297). Shortly after this early Neolithic
southeast, the Marginal Massif to the west occupation, megalithic dolmens and tumuli
and northwest, and the Douro basin to the began to punctuate the surrounding land-
north (see Fig. 1). The area covers most of the scape and are especially concentrated around
drainage basins of the upper Mondego, the the most important rivers of the region.
Vouga and part of the Douro rivers. The Cen- Passage graves, dolmens and other mega-
tral Massif, especially Star Mountain Range, liths throughout the Atlantic façade of
offers good pastures during the spring and Europe are located and oriented according to
summer seasons and has evidence of human the landscape (Tilley 1994), but also towards
particular astronomical events (Ruggles
1999). In the 1980s Michael Hoskin surveyed
*
UCL Institute of Archaeology dolmens in this region of Portugal and con-
31-34 Gordon Square, London WC1H 0PY
United Kingdom
cluded them to be loosely oriented towards
[email protected] sunrise/climb at the moment they were
100 Landscape and Astronomy in Megalithic Portugal

built (Hoskin et al. 1998; Hoskin 2001). The basins would most likely be now submerged
current and on-going research project was or destroyed by the rising water levels (Cruz
designed to expand on the measurements 2001: 296).
of Hoskin and reassess the data using up-to- Zilhao proposed a Maritime Pioneer
date methodological tools (Silva 2010). Colonisation model for the arrival of farm-
ing, and the Neolithic, in Portugal (2001,
The Early Neolithic of the Mondego 2003). In his model, agricultural communi-
Platform ties were formed by pioneer colonists that
Evidence for the occupation of the Mondego leap-frogged along the coast of the Mediter-
platform prior to the Neolithic is scarce. To ranean. This would have occurred at least
the northeast, in the Coa valley, famous for between central Italy and Portugal, and is
its Upper Palaeolithic open-air rock art sites, supported by similarities in ornaments and
a few microliths, possibly Mesolithic, were pottery decoration between the two extreme
discovered underneath levels with ceramics regions, the littoral placement of settle-
and polished stone (Rodrigues 2000). These ments, as well as radiocarbon dating indicat-
levels have been radiocarbon dated to about ing a swift movement (Zilhao 2001: 14184).
6400-6100 BC1. In Oliveira do Bairro, in the Mesolithic communities thrived for about
Vouga basin, and close to where the coastline 500 years after the first Neolithic societies
would have been at the onset of the Holo- appear in the archaeological record, as the
cene, microblade finds have been attributed latter mostly occupied territories previously
to the Mesolithic, at about 6500-5500 BC unoccupied by Mesolithic hunter-gatherers.
(Silva 2000). Based on these finds Cruz con- Discoveries of the past twenty years,
siders that the Mondego platform was not especially of Neolithic settlements in the
devoid of Mesolithic occupation but that Mondego basin, have allowed a picture of
most settlements in the Vouga and Mondego local life during the fifth millennium BC, the

Fig. 1: Map of Portugal, with distribution of megalithic monuments and the region under
study (inset). The enlarged map of the Mondego platform shows the location of all
surveyed dolmens, coloured by river basin: green for Vouga, dark blue and red for
Mondego, light blue for Paiva, pink for Torto and yellow for Coa.
Landscape and Astronomy in Megalithic Portugal 101

Small communities sustained themselves

by small game hunting, the gathering of
acorn and other winter fruits, and transhu-
mant pastoralism. The raising of ovicaprids
- sheep and goat, introduced to the region
by the first Neolithic settlers - suggests that
spring and summer were spent on high
ground pastures and winter on lower ground
(Senna-Martinez et al. 1997: 664, Cruz 2001:
Fig. 2: Dolmen da Orca with extant cover- 313). According to Senna-Martinez and
stones and cairn. Ventura, later developments can be inter-
preted as the consolidation and intensifica-
tion of this seasonal economy.

The Monumentalization of the Landscape

Radiocarbon dates indicate that the first
megaliths were built roughly a thousand
years after the first Neolithic communities
showed up in the region (Senna-Martinez
and Ventura 2008a: 317-50). Senna-Mar-
tinez and Ventura (2008a: 333) suggest a
division between Middle and Late Neolithic
monuments. The first phase (4,000-3,500
BC) would be characterized by tumuli with
a small polygonal chamber and a short cor-
ridor or none at all. Megaliths of the second
phase would have been built during the Late
Neolithic (3,500-3,000 BC). These are gener-
ally bigger and possess developed corridors
with heights that are different to those of
Fig. 3: Painted chamber orthostat of Dolmen the chambers, usually composed of 7 or 9
de Antelas (top), and carved face of orthostats, and include more complex and
the backstone of Dolmen 2 do Chão differentiated scenic spaces.
Redondo (bottom). Cruz, however, notes that it is difficult to
establish a chronology for the small monu-
Early Neolithic, to emerge (Senna-Martinez ments and that, given the vast geographical
and Ventura 2008a: 318-9). Pottery frag- regions considered, it is possible that mega-
ments attributed to this period show char- liths of both sizes were being constructed
acteristics belonging to the traditions of the and used contemporaneously by different
Early Neolithic from Estremadura, to the communities (2001: 302). The general trend
southwest, and Andalusia, to the southeast. towards more and more complex monu-
Based on this, Senna-Martinez and Ventura ments is nevertheless in-line with arguments
(2008b: 80) suggest the possibility of a dual put forward for other megalithic groups in
origin. However, considering the scarce but Iberia (Cruz 2001: 302). On the other hand,
increasing evidence for pre-Neolithic occu- “there is no evidence for the construction of
pation of the region, continuity cannot be dolmens” after 3,700 BC, even though there
discounted and the pottery evidence might is evidence for the continued use of some of
simply indicate the existence of trade. them throughout the Late Neolithic. Cruz
102 Landscape and Astronomy in Megalithic Portugal

contends that in this period forms of inhu- variety of depositions and other artifactual
mation other than burial in megalithic struc- finds, as well as the evidence for “condemna-
tures were being practiced but details are still tion structures” that permanently closed-off
sketchy (2001: 307-8). access to the dolmens (Cruz 2001: 290-1).
Throughout the vast region and time There is evidence for rituals involving fire in
period under consideration the depositional these ‘condemnation’procedures, as well as
assemblages are quite limited. Microliths, a care in the deposition of the stones that
blades, polished stone axes and beads are closed off the entrance, so that they would
universally present, whereas ceramic is nota- blend in with the surrounding cairn.
ble for its absence, even though it shows up These monuments can be seen as “true
in other contexts (Cruz 2001: 305). Arrow- anchors in the landscape for populations that,
heads are also found in the younger and on the other hand, have a high seasonal mobil-
more complex monuments, of about 4,000 ity” (Senna-Martinez and Ventura 2008b: 82).
BC, once again in-line with the introduction In this way Senna-Martinez and colleagues fit
of this artifact in the Iberian northwest. the monuments into their economic model:
The Late Neolithic is thus not so much the necropolis, by marking the territory,
characterized by new architectural styles would legitimize its winter occupation. Settle-
but by the presence of new depositional ments in the Mondego basin appear in close
artifacts (Senna-Martinez 1994), suggesting proximity to the dolmens and suggest an
an increase in complexity of the funerary autumn and winter occupation, as evidenced
rites. New elements include novel pottery by the presence of central hearths and earth
styles, projectile points, blade sickles and ovens where acorns were roasted.
flint daggers (Senna-Martinez 1994: 18), The archaeoastronomical survey of the
which are also found in known settlements orientations of the dolmens conducted by
that have been radiocarbon dated to the Hoskin lends further credence to this sea-
same period. sonal model (Senna-Martinez et al. 1997).
Some tumuli have complex structures in The measured orientations were inter-
front of them, which have been interpreted as preted as solar alignments to either sunrise
“scenic spaces”. The monuments themselves or sun-climb (that is, alignments to the sun
can be seen as temples, with different spaces shortly after sunrise) (Hoskin 2001). For the
having different access levels: the chamber Mondego dolmens, which exhibit a prefer-
and corridor being of limited access and visi- ence for southeasterly orientations, and thus
bility from the outside, and the tumulus itself for autumn/winter sunrises, this seems to fit
and surrounding space being a more public the hypotheses of the archaeologists. How-
space (Cruz 2001: 314). This division might ever, the dolmens in the other basins of the
indicate different degrees of participation in region exhibit a preference for slightly north-
rituals: the inside chamber being accessible of-east directions, which can only fit an early
only by a few who, afterwards, would come spring/late summer sun. Outlier orienta-
out into the atrium which could have been a tions that cannot be explained by this gen-
sort of amphiteatre. Judging by the number eralized solar interpretation abound and this
of extant cases of megalithic art in the cham- prompted the author to start a more compre-
ber’s orthostats, of both the painted and the hensive survey of the region.
carved variety (see fig. 3), it stands to reason
that they might also have played a role in Archaeoastronomical Survey
whatever rituals were enacted there. Methods
Some of the monuments (both of the sim-
ple and complex types) had a relatively short During the spring of 2010, thirty-one dol-
lifespan judging by the small quantity and mens of the Mondego platform were surveyed
for the orientation of their entrances and/or
Landscape and Astronomy in Megalithic Portugal 103

corridors. Five more sites were surveyed in The main issue with measuring the ori-
2011, bringing the total to 36 monuments. All entation of a prehistoric monument, such
measurements were made using a compass, as an Iberian dolmen, is how to define the
clinometer and a GPS unit, and local magnetic line from which one intends to measure
anomalies and magnetic declinations were the azimuth. The dolmens have been built
estimated on-site and, wherever possible, vali- using rough stones and the geometrical
dated afterwards. An in-depth discussion of straight-line perfection of later historical
the employed methodology can be found in peoples didn’t seem to much bother these
Appendix A of Silva (2010). Here only the nec- megalithic builders. The chambers are not
essary elements will be reiterated. perfect geometrical figures, and the cor-
ridors can bend slightly, close or open up.
Finding the orientation to measure is not as
straightforward as, for example, measuring
the orientation of the base of the Great Pyr-
amid. In the end, choices and assumptions
have to be made. Hoskin chose to measure
the line defined by the middle point of the
backstone (the stone opposite the chamber
entrance) and the mid-point of the corridor
(2001: 12). This survey purposefully decided
to define this line differently in order to test
for deviation due to differences in assump-
tions, as well as to go beyond this approach
by measuring a maximum ‘window of vis-
ibility’ (fig. 4).
The ‘average orientation’ was measured by
marking the middle of the chamber entrance
and the middle of the corridor entrance,
with surveyor’s rods carefully leveled (circles
in fig. 4). These were then used as backsight
and foresight for the compass measurements
(orange arrow). For dolmens without cor-
ridor the middle of the backstone, and the
middle of the chamber entrance were taken.
The ‘window of visibility’ is defined as
the maximum extent of the horizon which,
given the dolmen’s corridor and entrance
geometry, can be seen from the chamber.
This amounts to a range of azimuth bearings
between a minimum and a maximum value
defined by the diagonals of the corridor and
chamber entrances (blue arrows in fig. 4).
The first orientation, the orange arrow, is
thus merely an ‘averaged’ or ‘most likely’ ori-
entation, but not necessarily the most mean-
ingful one, as will be shown below.
Fig. 4: Orientations measured, in orange For archaeoastronomical purposes, the
and blue, on any given dolmen with ‘horizon altitude’ is as important as the azi-
a corridor.
104 Landscape and Astronomy in Megalithic Portugal

muth, as the azimuth where an astronomical

event occurs changes for different horizon
altitudes. For instance, in figure 5, the star
rises along the black dotted line, as indicated
by the yellow arrow. Because of local topogra-
phy, the star is seen to rise further south than
it would if there was no hill (the point where
the black dotted line crosses 0º of altitude
is where the star would have risen in such a
situation). Once the orientation one wants to
measure has been determined, a clinometer
can be used to measure the horizon altitude Fig. 5: A star rising behind a hill and all the
along that line to the nearest half-degree or important measures: the azimuth (red
less, but only if the horizon is visible. arrow), measured from true north,
the horizon altitude (blue arrow), and
Data the declination (black arrow), which
With the azimuths and horizon altitudes it is is calculated from the other two.
convenient and standard in archaeoastronomy
to convert these measurements to a single measured by both surveys deviating from
value, the declination. The declination is the each other by 10-12º or less. This illustrates
equivalent of latitude on the celestial sphere: the case put forward above: different defi-
it is an angular measure of how far from the nitions for the line of “average” orientation
celestial equator the point being considered can yield variations of up to 12º in azimuth.
is, measured along a line that is perpendicu- This, however, corresponds to only a 1-3º
lar to this equator (see fig. 5). Objects lying on variation in declination and means one
the celestial equator will have zero degrees can consolidate both surveys by including
of declination, whereas those at the celestial Hoskin’s measurements for those dolmens
poles will have ninety (positive for north, neg- that weren’t included in the present survey
ative for south). This simplifies the search for (Silva 2010).
astronomical events along those orientations Differences in architectural style and depo-
as well as allowing the comparison of monu- sitional assemblages suggest that the regions
ments in different geographical regions, as of central-north and central-south Portugal,
the conversion takes care of the differences in defined by the Vouga, Paiva, Torto and Coa
latitude between the sites.2 rivers and the Mondego river basin respec-
The following table shows the reference tively, were inhabited by different people
number (internal to this survey) and name of (Ventura pers. com.). As noted elsewhere
the dolmens, as well as the river basin that (Silva 2010, in press), this combined data-
they are located, true azimuth and horizon set continues to validate this regional split.
altitudes for the “average” orientation as well This is better visualized by looking at the
as the minimum and maximum azimuths for orientation histograms for the two regional
the window of visibility. Some of the horizon groups. Fig. 6 shows the declination histo-
altitude values have been corrected by using grams obtained by making each declination
a Digital Elevation Model (see further below). the mean of a stochastic Gaussian curve
with standard deviation of 2.5º and add-
Analysis ing them up together. This permits ranges
as well as preferences in orientations to be
The survey finds good agreement with pre-
visualized as peaks. Also illustrated as verti-
vious measurements by Hoskin (Silva 2010),
cal dashed lines, are the declinations of the
with the azimuths of the twenty dolmens
Landscape and Astronomy in Megalithic Portugal 105

Name Basin Az Alt Dec Win Max Win Min

1 Orca da Malhada do Cambarinho Vouga 111.75 4.9 -12.962 122.5 102.25

2 Dolmen da Lapa de Meruje Vouga 88.5 7.5 6.010 92 76.75
3 Dolmen de Arca Vouga 77.75 4.7 12.349
4 Orca dos Juncais Paiva 76 2.5 12.206 83.75 69.25
5 Orca de Pendilhe Paiva 126 2 -24.909
6 Orca do Picoto do Vasco Paiva 110.25 0 -15.165 116.75 98.5
7 Orca das Castonairas Paiva 107.25 5.4 -9.313 125.25 89.75
8 Orca dos Merouços Paiva 87 1.7 3.380
9 Dolmen 2 do Chão Redondo Vouga 66.5 0 17.606 85.5 46.75
10 Anta de Capela de Mouros Vouga 99.5 1.2 -6.400 114 91.25
11 Anta da Cerqueira Vouga 70.75 2 15.801 89.25 59.5
12 Dolmen do Carapito Mondego 91 3.7 1.658
13 Anta de Cortiçõ Mondego 119.25 1.6 -20.626 130.5 113.25
14 Anta da Matança Mondego 97.5 2.2 -4.239
15 Anta da Cunha Baixa Mondego 109.25 3.2 -12.342 121.25 101
16 Anta da Orca Mondego 109.25 6.3 -10.248 147.5 75.25
17 Anta da Pêra de Moço Coa 86.5 5.2 6.036 110.25 61
18 Lameiro de Cima I Torto 87.25 -0.5 1.749 96.5 79
19 Lameiro de Cima II Torto 81.25 -0.5 6.269 101.75 75
20 Dolmen do Carvalhal Torto 67 3.8 19.733 98.25 41.5
21 Dolmen/Capela da Sra do Monte Torto 87.5 3.1 3.920 105.75 69
22 Dolmen do Sangrino Torto 89.75 3.4 2.419
23 Orca das Pramelas Mondego 108.5 4 -11.264
24 Orca de St Tisco Mondego 118 2 -19.533 141.5 93
25 Orca do Outeiro do Rato Mondego 95.75 2 -3.069 124.25 74.25
26 Orca do Santo Mondego 102 2 -7.786 112 85
27 Dolmen da Orca Mondego 109.75 2.7 -13.081 117.5 98.5
28 Orca 1 do Ameal Mondego * * *
29 Orca 2 do Ameal Mondego * * *
30 Orca da Palheira Mondego 94 3 -1.092
31 Dolmen de Antelas Vouga 90.5 0.3 -0.183
32 Orca de Porto Lamoso Paiva 97 1.35 -4.402 115.5 86
33 Anta do Turgal Torto 147 5.4 -34.612
34 Dolmen 1 da Lapinha Torto 104 5 -7.160
35 Dolmen de Areita Torto 89 3.1 2.789 113.25 73.25

Table 1: Corpus Mensurarum of all surveyed dolmens. The window of visibility (win max and
win min) was not possible to obtain for dolmens without a corridor. Those marked
with an * have been surveyed, but there are lingering doubts with regards to their
entrances, so they have been excluded from this analysis.
106 Landscape and Astronomy in Megalithic Portugal

solar extremes (winter and summer solstices, no clearly defined lunar or solar event.3 This
WS and SS respectively), as well as the lunar peak is also visible in fig. 6, even though it is
extremes (southern and northern, minor and not as prominent as it is for this nucleus.
major lunar standstills, smLS, nmLS, sMLS
and nMLS). The horizon at Carregal do Sal
The distribution of orientations of dol- Throughout the survey the horizon of some
mens in the second group is considerably dolmens, in Carregal do Sal and elsewhere,
broad, ranging from -35º to 33º. Against this could not be surveyed because of tall vegeta-
seemingly random baseline there is a well- tion, mostly pine trees, which covered the
defined, statistically significant peak around view. In Silva (2010) empirical guesstimates
declination 4º. This, as previously argued by were indicated and used for the calculation
the author (Silva in press, Silva and Pimenta
2012), corresponds to the expected distribu-
tion of the Autumn Full Moon (green dashed
line). This is a particular Full Moon, occur-
ring close to the Autumnal Equinox, which
is empirically marked by the crossing-over of
the sun- and moonrise positions (Silva and
Pimenta 2012).
The situation is, however, not as clear-cut
for the Mondego dolmens: although there
seems to be a preference for negative dec-
linations, there isn’t a unique well-defined
peak. This could be due to the presence of
several peaks close to each other, one of
which could be a Spring Full Moon (about
-4º) and another one of the lunar extremes,
the southern minor standstill or the winter
solstice sunrise (Silva, in press), with a possi-
ble third around -10º of declination. In order
to shed some further light on the orienta-
tion of these dolmens the second half of this
paper focuses on a single nucleus of Neo-
lithic activity in the Mondego valley.
Fig. 6: Declination histograms for the con-
The Carregal do Sal Megalithic Nucleus solidated dataset. Dolmens of the
The nucleus of Carregal do Sal (the blue Mondego basin are shown above, and
markers in Fig.1) is located on the northern of other basins below.
bank of the Mondego valley, a mere kilome-
tre and a half from the riverbed. It is com- of declination but using digital elevation
posed of eight of the measured dolmens (refs data one can recreate vegetation-free hori-
23-30), as well as several as of yet unsurveyed zons for all sites and confirm or correct them.
smaller ones and other tumuli. A Digital Elevation Model (DEM for short)
The distribution of the orientation of is a computer model composed of a raster of
these dolmens, similarly to the whole of the cells where each has the value of the mean
Mondego valley, exhibits a range of negative elevation across the area defined by that cell
declinations, with a slight preference for (Connoly and Lake 2006: 27). There are sev-
negative 10-12º, a value for which there is eral DEM models around, using elevation data
Landscape and Astronomy in Megalithic Portugal 107

Ameal 2 (refs 28 and 29) are small dolmens

without corridors, and there are still open
questions regarding the orientation of their
entrances (if they ever had them: Silva 2010,
Ventura pers. com.), whereas the window of
visibility of Orca das Pramelas (ref.23) could
not be surveyed at the time. This leaves only
four dolmens: Orca de Santo Tisco, Orca do
Outeiro do Rato, Orca do Santo and Dolmen
da Orca (refs 24-27). The entrances of all
of these are oriented towards a particular
mountain range in the SE direction: Serra da
Fig. 7: Declination histogram for dolmens in Estrela (Star Mountain Range), which con-
Carregal do Sal. tains the highest peak in continental Portu-
gal. The figure below shows one such exam-
acquired and treated differently. One such ple, that of Orca de Santo Tisco (see also fig. 9
case is the SRTM or Shuttle Radar Topography below for all four dolmens).
Mission, which comprises elevations deter- Figure 8 shows a 360º panorama of the
mined roughly every 100 feet north-south horizon around the dolmen, as recreated by
and east-west for the US and every 300 feet HeyWhatsThat. The different colours repre-
elsewhere, obtained via radar interferometry sent different distances of the topographi-
by the Space Shuttle program during February cal features (green is closest, then blue, dark
2000 (US Geological Survey 2010). magenta and finally purple), whereas the
With this data one can use a GIS software inverted red triangles mark known peaks.
package to create a virtual reconstruction The cardinal directions are also marked. The
of the horizon profile at any location and/ vertical scale has been exaggerated 8.3 times
or calculate the horizon altitude on a given to facilitate the identification of topographic
direction. HeyWhatsThat is a free, online, features and vertical lines mark the “average”
utility that uses SRTM data to do just that orientation (in brown) as well as the meas-
(HeyWhatsThat Team 2012), requiring only ured window of visibility (in black). To facili-
the latitude and longitude of the archaeolog- tate visualization, the areas of the horizon
ical site. It is also fully integrated with Google that can’t be seen from within the dolmen’s
Maps, so one can export one’s horizon pro- chamber have been greyed out.
files and viewsheds. The brown line marks the centre point
Using HeyWhatsThat, most of the horizon of the corridor window (not always, see the
altitudes that were possible to measure in the appendix to Silva (2010) for a discussion),
field have been validated to within a degree. but the window permits a much wider range
Guesstimates have now been replaced by the of the horizon to be seen from within the
virtually generated horizon altitudes (see chamber. This information is not accurately
table 1 above), and have been used to cal- encoded by a histogram of the “averaged”
culate the declination histograms shown in orientations, such as those of figures 6 and 7.
Figs 6 and 7. The importance of topographic features
By looking at the horizon profiles of the for Mesolithic and Neolithic people has been
eight surveyed dolmens of Carregal do Sal it highlighted by Tilley (1994). He argued that, in
quickly became clear that only Orca da Pal- the Mesolithic ‘known, named and significant
heira (ref.30) didn’t have a distant horizon - places [were] linked by paths of movement
the view was obstructed by local topography. to which populations repeatedly returned
Out of the remaining seven, Ameal 1 and during their seasonal activity rounds’ (1994:
108 Landscape and Astronomy in Megalithic Portugal

Fig. 8: Horizon profile of Orca de Santo Tisco showing the measured “average” orientation
(brown line), as well as the maximum window of visibility (non-shaded area).

202). In the Neolithic this dynamic would 0.45 and 0.87 respectively)4 and both are col-
have changed, as the megaliths became the oured red even to the naked eye.
anchors of the seasonal movements. Impor- Axial precession, also known as precession of
tant landscape points ‘became captured in the equinoxes, is a wobble of the Earth’s axis
the orientation of morphological features of against the celestial sphere, over very long
the monuments and their placement in the periods of time. For an Earth-bound observer
landscape’ (1994: 202-3). it would look like the background of stars, as
Besides having potentially ritual and a whole, is moving very slowly. But the stars
funerary functions, the dolmens of the themselves also have an intrinsic movement,
Mondego valley also marked the winter ter- called proper motion, which is different for
ritories of their builders (Cruz 2001: 311, different stars. These two factors conspire to
Senna-Martinez and Ventura 2008b: 82), move the stars around over periods of thou-
whereas Star Mountain Range, in the hori- sands of years so that today they do not rise
zon, marked their summer pastures (Senna- and set where they rose and set six thousand
Martinez and Ventura 2008a: 332). It seems years ago. Table 2 below shows the declina-
that both man-made and natural landscape tion of the two stars under consideration at
markers were anchoring the seasonal move- three points of the megalithic phase5.
ment. The orientation of the Carregal do Figure 9 shows a section of the horizons
Sal dolmens towards Star Mountain Range, for the four identified dolmens of the Carre-
then, makes sense from a landscape/spatial gal do Sal nucleus, along with the window of
perspective. visibility (unshaded areas) and “average” ori-
This, however, does not preclude the pos- entation (vertical brown line). The location of
sibility of an astronomical event marking the rise of Aldebaran (in red) and Betelgeuse
the seasonal movement temporally. Consid- (in blue) is also indicated by the coloured
ering the lack of well-defined solar or lunar shaded areas, the right-hand side limit mark-
explanations for the peak of the distribution ing the rise position in 4,300 BC, and the left-
one can explore whether, at the time of the hand side in 3,500 BC.
megalith builders, any bright stars would rise Both Betelgeuse and Aldebaran, in the
in that range. period 4,300-3,500 BC would thus be seen
to rise from within the chambers of the four
Aldebaran rising dolmens, even though only a glimpse of
The only bright stars that, around 4,000 BC, Betelgeuse would be possible from Orca do
rose in the neighbourhood of the mysteri- Santo, as the star would have risen very close
ous -11º peak are Betelgeuse and Aldebaran, to the right-hand side of the window.6
in the modern constellations of Orion, the Even though these stars, whenever visible,
Hunter, and Taurus, the Bull, respectively. would always be seen to rise in the same
Both stars are the brightest in their respec- place on the horizon throughout the year,
tive constellations (apparent magnitudes of stars that lie so close to the celestial equator
go through a period in which they are not
Landscape and Astronomy in Megalithic Portugal 109

Year BC Declination Heliacal Rising around

Betelgeuse 4,300 -14º 38’ 21 May
4,000 -12º 58’ 23 May
3,500 -10º 13’ 29 May
Aldebaran 4,300 -12º 18’ 18 Apr
4,000 -10º 39’ 21 Apr
3,500 -7º 51’ 27 Apr

Table 2: Declination and approximate date for the heliacal rising of Aldebaran and Betel-
geuse during the megalithic phase.

visible in the night-sky. Considering Senna- The origin of the name of Star Mountain
Martinez and Ventura’s seasonal model, it One can now consider whether other sources
would be interesting to look at the seasonal- support the possibility that the mountain
ity of their ‘heliacal rising’, their first appear- range’s name is linked to this very star, and
ance after this period of invisibility. whether the name may thus be as remote as
In the epoch under consideration, the the Neolithic itself (if not older). Local eth-
heliacal rising of Betelgeuse would have nography and folklore are filled with myths
occurred between 21-29th May, or therea- and stories that explain the origin of the
bouts, whereas Aldebaran heliacally rose name of Star Mountain Range. There are sev-
around 18-27th April (see table 2).7 The preci- eral recorded stories by locals on this topic,
sion of these dates needs to be taken with a some of which have been recorded by profes-
pinch of salt, as visibility, climatic and other sional folklorists (CEAO 2006). The simplest
observational criteria might not permit the version is the one currently on the website of
star to be seen so close to the horizon on a the Covilhã City Council (C M Covilha 2012),
given day. Generally, however, one can say one of the five municipalities that surround
that Aldebaran reappeared in the eastern sky the mountain range. It goes something like
in late April/early May, whereas Betelgeuse this (author’s translation):
did so in late May/early June.
The seasonal model states that, in the spring People say that the name Star Mountain
and summer, the dolmen builders would take Range was given in the olden days by a
their ovicaprids to the high pastures of Star shepherd living in an unknown place in
Mountain Range to take advantage of its grazing the Mondego valley. He spent his nights
grounds. If the dolmen-builders were observ- contemplating a star that was so bright
ing the heliacal rise of these stars, as is here that it illuminated the top of a nearby
proposed, they could have used it as a tempo- mountain range. One day he decided
ral marker for their transhumant movement to to take his faithful dog and follow the
high pastures. The period of invisibility of the scintillating light that attracted him so
stars would provide the perfect amount of time much. After climbing for many days
to make preparations for whatever rituals were they reached the peak. Impressed by
going to be enacted in the dolmens, as well as the luminosity of the star the shepherd
for the move to higher grounds. Although both told his dog: “to this place that seems
identified stars can fit this picture, it can be to be favoured by the celestial objects
argued that Aldebaran, with its earlier heliacal [astros] I will name Serra da Estrela [Star
rising, just a month after the Vernal Equinox, fits Mountain Range], and you that accom-
this picture better than Betelgeuse. panied me I shall give the same name.
110 Landscape and Astronomy in Megalithic Portugal

Fig. 9: Sections of the horizon profiles of all four dolmens in Carregal do Sal with a distant
horizon in order of reference number. The horizon outside the windows of visibility is
greyed out. Also marked is the variation in rise position of Aldebaran (red) and Betel-
geuse (blue) during the megalithic phase.
Landscape and Astronomy in Megalithic Portugal 111

The latter part refers to the Serra da Estrela Conclusions

dog, also known as the Portuguese Shepherd Although not as monumental as Stonehenge,
and one of the oldest breeds in Portugal. Avebury or Newgrange, the Iberian dolmens
There are several other versions of this story, still demonstrate some level of cultural unity
some of which include other elements and across vast areas, as well as regional diversity
mythemes8 but the central plotline remains in both architectural elements and in the
the same. All versions remark how unique orientation of their chambers and corridors.
this star is, either by mentioning that the star This survey has focused on an area of roughly
illuminated the top of the mountain, just 10,800 km2 and, together with previous sur-
as in the version above, or by, at the end of veys, included fifty-eight dolmens. Within
the story, describing it as still today shining this area, at least two regional groups can
“more”, or “differently”, than the other stars. be distinguished and this is supported by
The collection and analysis of Portuguese landscape phenomenology and archaeoas-
folktales was neglected for a long time but tronomical arguments.
this has been increasingly addressed in the Judging by the analysis of the necropolis of
past decades. Notable projects include the Carregal do Sal, the dolmens of the Mondego
online “Archive of Portuguese Legends” valley show a preference for locations from
(CEAO 2006) by the Centro de Estudos Ataíde which Star Mountain Range, containing the
Oliveira (University of Algarve), which col- highest peak in continental Portugal and
lects at least three different versions of this the very source of the Mondego river, can be
tale. The interpretation of myth and folklore seen. Outside of Carregal do Sal there is at
has a long tradition in academia, and several least one other dolmen whose horizon pro-
schools of thought have pursued different file analysis yields similar conclusions (Anta
interpretative paths.9 da Orca, ref.16).
The tale above, however, fits nicely with From within the chambers of all these dol-
the hypothesis put forth in this paper. The mens it would not only be possible to see Star
main hero of the tale is a shepherd living in Mountain Range in the horizon, it would also
the Mondego valley, from where he would be possible to see the stars Betelgeuse and
see a particular star atop a mountain range. Aldebaran rising above it. These two stars are
This triggered him to “follow” this star, tak- not only very bright they are also both red.
ing his dog, a shepherd dog, with him. The Around 4,000 BC, when the megaliths were
comparison with the Neolithic people of the built, these stars would disappear from the
Mondego basin is very suggestive. Firstly they night-sky at the end of February/beginning
also inhabited the river valley - the megalithic of March, not to be seen for two lunar cycles
nucleus of Carregal do Sal. Secondly, the evi- (two and a half in the case of Betelgeuse),
dence suggests that they too, like the hero of until they would reappear in the eastern sky
the tale, were transhumant pastoralists. And at dawn before sunrise, just as spring was
thirdly, archaeologists believe that these peo- starting to reinvigorate the landscape. This
ple would take their ovicaprids to graze on was when the Neolithic communities of the
the high fields of Star Mountain. In the topo- Mondego valley would “follow this star” and
nymical tale the transhumantic movement of transition to the high pastures of the moun-
the shepherd is triggered by the appearance tain range it illuminates. One can’t be sure
of this star over the mountain range. In the whether it was Betelgeuse or Aldebaran that
case of the megalithic builders Betelgeuse or, was being targeted, or even both, but the tim-
perhaps more likely, Aldebaran could have ing of the heliacal rising of the latter, closer
served the same purpose, and the dolmens to the Vernal Equinox, would make it better
would then have served as spatial and tem- suited to be used as a seasonal marker. The
poral markers for this important event. presence of the Hyades star cluster around
112 Landscape and Astronomy in Megalithic Portugal

Aldebaran, giving it a “different shine”, would ers might have intended the orientation of
further support this, based on a (semi-)literal the chambers to simply point out, indicate or
reading of the toponymical folktales about direct one’s gaze towards a broad area of the
Star Mountain Range. horizon, and an ‘obvious’ topographical fea-
Unfortunately this effect is no longer ture, within that horizon range, would mark
observable today as, due to axial precession an important astronomical event. Measuring
and proper motion, Aldebaran and Betel- several of these ‘averaged’ lines and plot-
geuse have both migrated further north, ting a histogram of these orientations is not
to positive declinations (+16.5º and +7.5º enough: one needs to shed modern, math-
respectively). They now rise north of east and ematical and geometrical assumptions and
not in alignment with Star Mountain Range immerse oneself in the landscape itself, as
as seen from Carregal do Sal. Tilley proposed (1994).
The use of the heliacal rising of a star as Both points have already been made by
a marker for ritual, a new season and the Clive Ruggles (1999: 156-63), based on his
new year is not without historical and ethno- experience surveying prehistoric stone cir-
graphic precedent. For instance, the ancient cles, rows and mounds in the British Isles.
Egyptians timed their ritual calendar by the This certainly seems to be the case for the
heliacal rise of Sirius (Schaefer 2000, Steele Carregal do Sal nucleus as well, and thus
2007) and, in the ethnographic present, might very well be extendable to the breadth
South American Indians mark the helia- of Megalithic Europe.
cal rise of the Pleiades (Lévi-Strauss 1986). A third point relates to the interpretation
As is the case with the dolmen builders of and encoding of information in folktales and
the Mondego valley, the heliacal rises of the myths. While the stellar hypothesis argued
chosen stars coincide with important events here is completely independent of the inter-
in the subsistence economies of these peo- pretation of the legends that explain the
ples: in Egypt it coincided with the flooding origin of the name of the mountain range,
of the Nile, rebooting the agricultural cycle; this possibility certainly adds strength to
in South America it coincides with the shift the argument. Tilley said that the “build-
from wet to dry season. ing of the monuments prevented the ritual
Three important epistemological points and mythological significance of particular
have also been highlighted in this work. places being lost and forgotten” (1994: 204).
Firstly, that landscape archaeology and If the reading of the toponymical myth made
archaeoastronomy are not incompatible, here is valid, it means that the reverse can
but complementary: topographical features also be true: the folktale preserved the link
could have been used as markers for astro- between star and mountain range that was
nomical events. Secondly, that the empha- only visible from across the Mondego val-
sis on methodological precision, which has ley where the Carregal do Sal dolmens lie.
characterized megalithic archaeoastronomy, Together, the mountain range and the dol-
should be alleviated in favour of horizon sur- mens anchored a seasonal movement that
veys with a special eye for topography and is, in a reduced form, still practiced by local
phenomenology. Megalithic monuments, shepherds today.
such as the studied dolmens, are imprecise
by their very nature. Measuring an “averaged” Acknowledgments
straight line to as much precision as one can The author would like to thank Roslyn Frank,
(for instance by using a theodolite) does not Fernando Pimenta and the anonymous ref-
counter the fact that one cannot be sure eree for careful reading of the manuscript,
whether one is measuring the ‘right’ line (or as well as insightful discussions on the topic.
whether such a line even exists). The build-
Landscape and Astronomy in Megalithic Portugal 113

Notes 8 Examples include the personification of the

1 Dates throughout the paper are in cal BC. star, the death of the dog from exhaustion,
2 The same astronomical events, say sunrise before they reach the peak and the exist-
on the winter solstice, occur at different az- ence of a king that seeks to buy the stellar
imuths depending on latitude and horizon woman from the shepherd, to no avail.
altitude, but always at the same declina- 9 Examples include, but are not limited to,
tion. Monuments at different latitudes and the Junguian, Proppian and Lévi-Strauss-
landscapes can then be oriented towards ian schools (e.g. Segal 1998, Propp 1968,
different azimuth bearings, but still be Lévi-Strauss 1986).
aligned to the same astronomical events.
References
Declination allows one to test this.
3 One possible interpretation for this peak Connoly, J and Lake, M 2006 Geogra phical
could be sunrise on specific days of the Information Systems in Archaeology. Cam-
year, namely around the 20th February and bridge: Cambridge University Press.
the 22nd October. These dates, particularly CEAO 2006 Archive of Portuguese Legends
the February one, seem to be too early to [online], http://www.lendarium.org
mark the spring transition. Covilha, C M 2012 Parque Natural [online],
4 The apparent magnitude is a measure of the http://www.cm-covilha.pt/simples/?f=2612
brightness of a celestial object as seen by an [accessed 2th August 2012]
observer on Earth. The brighter the object Cruz, D J 2001 O Alto Paiva: megaltismo,
appears, the lower the value of its magni- diversidade tumular e práticas rituais
tude. durante a Pré-história Recente. PhD Disser-
5 Because different software packages use tation. Coimbra: Universidade de Coimbra.
different algorithms to calculate the posi- HeyWhatsThat Team 2012 HeyWhatsThat
tions of stars in the past the values given in Testbed [online], http://www.heywhat-
table 2 are the mean of the values of dec- sthat.com/main-0904.html
lination, to nearest arc-minute, between Hoskin, M et al. 1998 Studies in Iberian Archae-
three software packages: Starry Night Pro oastronomy: (5) Orientations of Megalithic
Plus, Stellarium and Cartes du Ciel, the last Tombs of Northern and Western Iberia. Jour-
two of which are freely available online. nal for the History of Astronomy, 29, S39-88.
6 The window of visibility of Orca do Outeiro Hoskin, M 2001 Tombs, Temples and their
do Rato seemed to exclude the possibility Orientations, Bognor Regis: Ocarina Books.
of observing the rise of any of these stars Lévi-Strauss, C 1986 The Raw and The
as they’d have risen south of the southern- Cooked – Introduction to a Science of
most limit of the window. However, the Mythology 1. Middlesex: Penguin Books
corridor of this particular dolmen bends [1964].
10º towards the south halfway through Propp, V 1968 The Morphology of the Folktale
(Silva, 2010), creating a new window of (2nd ed). Austin: University of Texas Press.
visibility that looks directly at the highest Purrington, R D 1988 Heliacal Rising and
topographical feature on the horizon and Setting: Quantitiative Aspects. Journal for
to the rise of Aldebaran and Betelgeuse. the History of Astronomy, 19, S72-84.
7 These were calculated by the author using Rodrigues, S M 2000 A estação neolítica do
an algorithm developed in Matlab, based Prazo (Freixo de Numião – Norte de Por-
on the works of Schaefer (1987, 1997, 2000) tugal) no context do Neolítico Antigo do
and Purrington (1988). The dates are given Noroeste Peninsular. Algumas consider-
in the modern calendar, i.e. April 21st cor- ações preliminaries. In: Actas do 3o Con-
responds to the thirty-first day after the Ver- gresso de Arqueologia Peninsular (vol.3).
nal Equinox( which occurs on March 21st). Porto: ADECAP, 149-180.
114 Landscape and Astronomy in Megalithic Portugal

Ruggles, C 1999 Astronomy in Prehistoric Senna-Martinez, J C and Ventura, J M Q

Britain and Ireland. New Haven and Lon- 2008b Neolitização e Megalitismo na Plata-
don: Yale University Press. forma do Mondego: Algumas Reflexões
Schaefer, B E 1987 Heliacal Rise Phenom- sobre a Transição Neolítico Antigo/Neolítico
ena. Journal for the History of Astronomy, Médio. In: Actas do IV Congreso del Neolítico
18, S19-33. en la Península Ibérica, Alicante, 77-84.
Schaefer, B E 1997 Heliacal Rising: Defini- Silva, F A P 2000 A estação Epipaleolítica-
tions, Calculations, and Some Specific Cases. Mesolítica do Cabeço Branco (Portinho,
The Archaeoastronomy and Ethnoastronomy Oiã, Oliveira do Bairro, Aveiro). Arqueolo-
News, 25. gia, 25, Porto, 79-88.
Schaefer, B E 2000 The Heliacal Rise of Sirius Silva, F 2010 Cosmology and the Neolithic:
and Ancient Egyptian Chronology. Journal A New Survey of Neolithic Dolmens in
for the History of Astronomy, 31, 149-55. Central Portugal. Journal of Cosmology, 9,
Segal, R A 1998 Jung on Mythology. London: 3094-3106.
Routledge. Silva, F (in press) Equinoctial Full Moon Models
Senna-Martinez, J C 1994 Megalitismo, and Non-Gaussianity: Portuguese Dolmens
habitat e sociedades: a bacia do médio e as a Test Case. In: Rappenglueck, M, Rappen-
do alto Mondego no conjunto da Beira glueck B and Campion, N (eds.) Astronomy
Alta (c. 5200-3000BP). Estudos Pré-históri- and Power. Oxford: Archaeopress.
cos, 2, Viseu, 15-29. Silva, F and Pimenta, F 2012 The Crossover
Senna-Martinez, J C, López-Plaza, M S of the Sun and the Moon. Journal for the
and Hoskin, M 1997 Territorio, ideología History of Astronomy, 43(2), 191-208.
y cultura material en el megalitismo de la Steele, J M 2007 Calendars and Years:
plataforma del Mondego (Centro de Portu- Astronomy and Time in the Ancient Near
gal). In: O Neolítico Atlántico e as Orixes do East. Oxford: Osbow Books.
Megalitismo. Actas del Coloquio Internac- Tilley, C 1994 A Phenomenology of Land-
ional (Santiago de Compostela, 1-6 de Abril scape. Oxford: Berg Publishers.
de 1996). Santiago de Compostela: Univer- US Geological Survey 2010 Shuttle Radar Topog-
sidade de Santiago de Compostela, 657-76. raphy Mission [online], http://srtm.usgs.gov
Senna-Martinez, J C and Ventura, J M Q Ventura, J M Q 1998 A Necrópole Megalítica
2008a Do Mundo das Sombras ao Mundo do Ameal, no Contexto do Megalitismo da
dos Vivos: Octávio da Veiga Ferreira e o Meg- Plataforma do Mondego. MA dissertation.
alitismo da Beira Alta, Meio Século Depois. Lisboa: Faculdade de Letras da Universi-
In: Homenagem a Octávio da Veiga Ferreira - dade de Lisboa.
Estudos Arqueológicos de Oeiras, 16, 317-50.