Landscape,orientation and celestial phenomena on the ‘Coast of Death’ of NW Iberia

Issues of visibility for our case-study area

Again, this has been discussed and evidenced in our first paper on dolmens of this region. There we explained the possible visual influence of the surrounding vegetation of the dolmens. ²² We demonstrated that those areas of Galicia without mountains or sierras, like Costa da Morte, were likely primarily covered with mixed deciduous forest, dominated by oak (Quercus) with a light under-story of shrubs and herbs during the fifth to fourth millennium BC. ²³ As well as dolmens possibly being erected within these open, deciduous forests, there is evidence for the reduction of forest and the creation of pockets of grasslands at the beginning of the first half of the fifth millennia BC, ²⁴ coinciding with the earliest dates we have for dolmens in A Coruña region, like Forno dos Mouros 5a (Ua 20009 5635 ± 50 BP, 4590–4350 cal BC, 95.4%). Added to these lines of evidence is the idea that there were also natural open aspects within such forests for about six months of the year due to their deciduous nature (approximately late Autumn-Late Spring), suggesting that the horizons were likely visible for at least half the year, and possibly more.

Horizon profile construction

The primary aim of the software Horizon is to combine topographical data with atmospheric and astronomical calculations to produce accurate two and three dimensional landscape information to aid archaeoastronomical surveys and data analysis. ²⁶ This includes the ability to create 2D horizon profiles and rendered 3D landscape panoramas for each site, along with an ascii data file containing calculations of the distance and direction between the point of origin or monument to paths of astronomical phenomena like the sun on the longest day and the shortest day along with all points on the horizon and the altitude of each of these horizon points measured every 0.01 of a degree. Apart from the location, Horizon allows you to choose epochs (dates) and astronomical phenomena. We have chosen to map the Moon at the times of its most extreme rising and setting points in its cycle, as well as the solstitial rising and setting of the Sun, and the equinoctial points, onto the landscapes.

Monument orientations

We obtained the list of the directions of the dolmens’ corridor or entrance axes (not all tombs are made with corridors) from Vilas-Estévez’s postgraduate work. All confirmed dolmens had their orientations measured in the field from magnetic north (azimuth) and corrected to True North readings by Vilas-Estévez. ²⁷ Unlike Hoskin et al., he used different methods of measurement in accordance with the dolmen’s original construction or remains. Method A applies to those with a corridor, there he took the middle point of the backstone and the middle point of the entrance corridor to create and measure the axis line. Method B, in dolmens without a backstone, he took the middle point of the chamber and the middle point of the corridor. And method ‘D’, where dolmens without a corridor (original construction or through ruination), were measured from the middle point of the backstone and the middle point of the chamber’s entrance. These measurement approaches can produce differing results when applied to the same dolmen. For instance, for the Dolmen Rego da Murta (Alvaiazere Portugal), Vilas-Estévez found a 7.5° range between D and B, with A being 2.5° from Axis D and 5° from Axis B. ²⁸ An alignment, therefore, is made up of the defining physical characteristics from which we can measure an orientation, like ‘the middle point of the backstone and the middle point of the chamber’s entrance’.

Furthermore, if we consider using Method A as the axis of symmetry for dolmen Dombate in Costa de Morte, the size of its corridor and width indicates that the window of observation of the corridor entrance as seen from the back stone is 7¾°. ²⁹ This viewing position along the axis of symmetry, was considered for the possible purpose of illuminating effects within the dolmen by González-García et al., ³⁰ but it can also direct the viewer at the back stone to a particular view of the outside world. Nevertheless, we saw above that González-García et al., explain that, depending on the orientation and shape of the megalithic structure, natural light could penetrate the dark inner areas, creating different illumination effects. ³¹ So, whilst entrances appear wide they are also restrictive and the window of observation (looking out) becomes the window of illumination (‘looking’ inwards), where particular areas of the interior of the megalith could be illuminated on several days, whilst other parts may remain in shadows and darkness.

This discussion leads us to consider a possible plausible level of accuracy for dolmen orientations of ±4° if the purpose might include illumination of the back stone(s), for instance. We note that none of these dolmens have central orthostats along their axes, like some of these in southern Iberia. If such illumination is not the (only) purpose, such as a clear alignment with an astronomical body as dictated by a ritual obligation or constraint, then the integrity and potency of such an orientation is likely connected to its precision. Here orientations with an accuracy within ±1° might be expected. However, given the state of some of the dolmens today and possible error margins of equipment use and so forth, as discussed in González-García et al., ³² orientations with an accuracy within ±2° are required for it to be categorised as one that is ‘constrained’, for this paper. Therefore, orientations will be categorised as ‘illumination’ orientations or alignments only where they are greater than ±2° and less than or equal to ±4° from an astronomical target, to distinguish them from those possibly containing stricter orientation constraints.

Data for 2D and 3D Horizon creations

For this project, we use a be-spoke version of the software Horizon to create our horizon profiles. It was created to use 25 and 5 m LiDAR data made freely available by the Galician Regional Government. ³³ These data had to be converted and tested for use with this version of Horizon by the software engineer and designer of the program, A.G.K. Smith (University of Adelaide ³⁴ ).

The creation of the 2D and 3D landscape panoramas was done by merging different elevation data sets: 5 m data extending out from the viewing point to 2.5 km, after which it is replaced with 25 m data to speed up the processing time. Then 90 m STRM data are used to fill all the areas not covered by the LiDAR data. These 3D landscape panoramas are essentially 360°-viewsheds created from the eye-height of 1.5 m above local ground level. As Horizon also has the capacity to add a variety of astronomical paths to the 3D landscape panorama output, we used this to create sky-landscape panoramas with paths of the Sun and the Moon. In particular, those paths connected to their extreme rising and setting points on the horizon (i.e. the Sun’s paths closest to the solstice and the Moon’s closest to the time of the major standstill), along with those closest to the equinox for the Sun and to the minor standstill for the Moon. We then added the dolmen orientations as a layer to these created 3D sky-landscape panoramas, to investigate any possible coincident occurrences of landscape shape, astronomical bodies and orientations. We made a note of the number of degrees in azimuth that each orientation was from the possible target along the horizon for each monument. Note that at any point going forward when we use the terms singularly/simply, like equinox or major standstill, we mean the visible rising (or setting) phenomenon that occurs closest to these times. We are not advocating that Neolithic people could calculate these exact astronomical times.

The polar plots

We used the ascii data of our horizon profiles created by Horizon to make our ‘God’s-eye’ viewsheds in the form of polar plots and layered onto these the relevant monument orientation. This was done using the software IDL. We created two types of polar plots. The simpler version included the entire viewshed up until the furthest point seen from each monument location. Thus the outer circle of these polar plots will equate to a different distance for each monument. However, for the second polar plots, the outer circles all equate to the same distance of 10 km and allow us to see the shape (volume in terms of distance and direction) of the viewshed closer to the monuments more clearly. The second polar plots also include the shape of 2D horizon-line derived from the DTM (Horizon software, see above), which represents, of course, the farthest points seen.

Analyses and outcomes

2D Horizon profiles and solar ranges

By examining the individual dolmen 2D horizon profiles of our 17 sites we can now confirm that there are fairly symmetrical topographical features for the majority of sites which centre on cardinal east and west and clearly drop off towards the north and south (Figure 4). There are 8/17 sites like this and further 7/17 sites that centre on cardinal east or west, this means that 15/17 sites are included in these two categories. Also, as indicated above, for the majority of sites, the eastern horizon is statistically shown to be lower than the western for all of the confirmed dolmens in Costa de Morte (n = 31 ³⁵ ). These topographic features were often visible in the field (see Figure 5a – Pedra Cuberta).

OPEN IN VIEWER

Figure 4.

Six 2D Horizon Profiles illustrating the ridges or higher profiles that usually appear along the horizons in the east and west for the majority of our sites in Costa da Morte: From top to bottom – Cuberta, Dombate, Pedre da Lebre, Pedra Moura do Monte Carnio, Pedra da Arca de Regoelle, Casota de Freans. For the first three of these sites we have added the angled paths of the Sun for each specific location layered on top and we can see how the eastern and western ridges coincide with the ranges for the solar sunrise and sunset, respectively. Here the general pattern is for the solstitial points to appear close to the edges of the higher horizon elevations. In the NE and NW is the path of the Sun closest to the SS, in the east, the path of the Sun closest to the Equinox; in the SE and SW the path of the Sun closest to the WS. n.b. Lebre does not have a measurable orientation and is added purely as another example of this horizon pattern found in Costa da Morte.

OPEN IN VIEWER

Figure 5.

(a) In the centre we can see the megalithic chamber of Pedra Cuberta facing the eastern ridge. (b) Here we see a 3D landscape panorama extract. The top left of edge of the ridge coincides with summer solstice sunrise (2nd slanted line from the left); the centre of the ridge coincides with the sunrise closest to the time of the equinox (slanted line on the far right; see Figure 10e for full panorama model and 10a for the figure legend). The black vertical line as it crosses the horizon represents the line of the orientation of the tomb entrance where it reaches the horizon. Photographer Gail Higginbottom.

In archaeoastronomical terms, we can also confirm that for the sites with eastern and western ridges these tend to coincide with the ranges for the solar sunrise and sunset, respectively. In fact, the solstitial points appear close to the flanks or edges of the highest horizon elevations in these directions and the central position of these eastern and western ridges tends to coincide with cardinal east and cardinal west – near the equinoctial points of the Sun (Figure 4). Here we can clearly see that both the landscape and the celestial phenomena emphasise each other.

Polar plot viewsheds and individual orientations

The simpler polar plots of the complete viewsheds seem to indicate that the orientation of nearly every tomb is not directed towards the most distant horizons (Figure 6). So views towards the most distant landscape are not the priority from the chamber, corridor or the front of the monument in line with the orientation, with only 2/17 having clearly distant horizons: Fornela dos Mouros and Pedra de Arca, Regoelle.

OPEN IN VIEWER

Figure 6.

Simple viewshed polar plots for every single monument that has an orientation. The viewsheds have been ordered from the more distant northern views first then moving to an easterly focus to south, then west. The orientations have been layered onto these plots as an arrow. N, E, S, W represent the cardinal points: for example, N: north; SS: the approximate position of the Sun at the summer solstice; WS: the approximate position of the Sun at the winter solstice; all assuming a flat horizon (altitude = 0°). These solar azimuthal positions would be seen to change as viewed from site to site, dependent upon the horizon altitude.

Also, overall, most of the horizon surrounding each dolmen tends to be relatively close. This was something we also noticed in the field, but couldn’t initially confirm as the horizons were often blocked by closely grown modern forestry plantations. Using horizon distance plots, we found that 13/17 sites have more than half the horizon less than 3 km away for a length or range of 185°–320° out of the entire 360°^-horizon, with their most distant horizons ranging from 17 to 52 km.

The polar plots in Figure 6 are an invaluable tool for the comparative conceptualisation of these overall distance patterns surrounding the dolmens. However, the maximum distance of each of these simple polar plots constrains and designates the outer rim of the plot such that we lose detailed visual information of the behaviour of the closer horizons. For this reason, we also created a second form of viewshed polar plot that somewhat disregards the full distance information of the most distant horizons for each site, but indicates the shape of the horizon profile related to altitude (curved line on the outer part of each plot). In this way, we have a better sense of horizon distance connected to horizon altitude and more detailed comparative distance information of the closer horizons (Figure 7a–c). We can also observe potential patterns between sites more readily, which, if they appear to exist, can be confirmed by a close study of the 2D profiles and their accompanying ASCII data (see Figure 8 as an example). For the ease of comparison, these polar plot viewsheds have been ordered according to orientation groups of 10° (i.e. orientations of the dolmens). The first panel groups those with orientations between 95° and 105° that could be closer to a nearly equinoctial orientation (Figure 7a). The second panel includes dolmens with orientations between 105° and 115° that may not be directly connected to the equinox nor the solstice (Figure 7b) and the final panel those with orientations greater than 115° that would be closer to the solstice (Figure 7c).

OPEN IN VIEWER

Figure 7.

(a–c) Shows three panels of the second type of polar plot viewsheds with orientations. The undulating outer circles indicate the shape of the horizon profile related to altitude and the dashed lines in the SE quadrants represent the orientations of the chambers or corridors of the dolmens. Whilst the orientations are linked to the horizon profile shape information, the SS and WS mark the approximate position of the Sun at the winter solstice assuming a flat horizon (altitude = 0°). These solar azimuthal positions would be seen to change as viewed from site to site, dependent upon the horizon altitude.

OPEN IN VIEWER

Figure 8.

Full 2D graphic of horizon information for the site of Dombate. Top section denotes profile shape defined by azimuth and altitude of the horizon. Bottom section denotes the concomitant horizon distances for each point on the horizon profile.

In the first panel (Figure 7a), it is interesting to notice that all dolmens seem to face a relatively close eastern horizon, with the exception of Mina das Folias. Therefore, the general pattern observed for the simple polar plots seems to occur here, with the majority of sites having orientations towards closer horizons rather than those that are the most distant. Whilst 4/5 of these eastern horizons associated with orientations are less than c.1.2 km away, most of these are not the closest horizon ranges to the monuments.

Moving onto the second panel (Figure 7b), there seems to be a transition from close eastern horizons to those that are more distant, the majority of which are further than those found in the west. In this panel, four of the six dolmens have orientations towards, or very close to, a break in the horizon distance: this is clearly seen in Recesindes where the orientation of the dolmen coincides with a sharp jump in the distance profile. The western horizon is overall closer than the eastern for all of these sites.

The qualitative information observed in the second panel is reinforced in the third group (Figure 7c), with 3/6 monument orientations right on the cusp of a dramatic shift in horizon distance, from close horizons to clearly further distanced ones. Using Pedra de Arca, Malpica as an example, this dolmen was found to face a horizon that is only 900 m distant, with a dramatic shift just c. 0.2° to east of this point of c. 6.5 km. Four out of six (4/6) have closer horizons in the west.

Noticeably, only one tomb (Fornela dos Mouros) is clearly aligned along the direction of the immediate slope in the downward direction of the surrounding land, and it was facing a very distant and open horizon (Figure 9) whilst the rest are positioned across the local down slope.

OPEN IN VIEWER

Figure 9.

Fornela dos Mouros facing the downhill slope. Photographer Gail Higginbottom.

Viewing individual orientation targets on the 3D landscape panoramas

Before we layered the individual orientations of the dolmens onto the 3D landscapes of each dolmen, we examined the latter. Upon these were already layered the paths of Moon and Sun in the sky until they met the horizon at particular times in their cycle. It will be remembered that these were the Moon paths closest to the times of the major standstill and minor standstill and the Sun paths connected to the time of the solstitial rising and setting and those closest to the time of the equinox. As well as further illustrating the findings of the 2D landscapes above, that ridges occur centring on cardinal east and/or west at every site, our new observations revealed that the most extreme rising and setting points of the Moon in the east and in the west flank these ridge tops, either very close to the edges or on the slopes of these ridges, but sitting out further than the Sun (Figure 10: also Figure 5b). In this way, the Moon’s travels are marked out by the landscape, as are the Sun’s.

OPEN IN VIEWER

Figure 10.

(a–q) 3D landscape panoramas of sites with orientation measurements taken from chambers or corridors (towards the east) indicated by vertical lines on the right, the orientations in the opposite direction indicated by the vertical lines to the left. The first image contains the labels of the paths of the Sun and the Moon at significant times in the cycles of these bodies in order to understand the images that follow. Dombate has two orientations as there are two dolmens at the same location with different orientations. Software and 3D landscape created by A.G.K. Smith. Based upon the Ordnance Survey 1:50,000 Landform PANORAMA map with permission of the Controller of Her Majesty’s Stationery Office © Crown.

By checking the azimuth of each orientation and examining its position on the horizon models, we clarified the possible astronomical targets of these orientations. The azimuths for the entire horizon are marked out on each model (Figure 10a–q; where Figure 10a labels the paths of Sun and Moon seen on all of the Figure 10 images).

Orientation targets: Looking out to the east

It is to be remembered that all of these orientations sit within a specific clustered arc running approximately E to SE on the horizon (see Figure 2 above) (Figure 10; Table 2). On examination of the 3D landscape panoramas, we can confirm that four dolmens (4/17) are aligned to within 4° of the event of the rising Sun at the WS, with the great tomb of Dombate being within 1° of accuracy (Figure 10; Table 2). However, the eastern orientation targets of two of these tombs, Pedra da Arca de Malpica and Caxadas (Figure 10o and q; Table 2) sit between the Sun at the WS and one of the foremost Moon targets. ³⁶ Most of these orientations sit to the north of the winter rising Sun at the solstice, and one to the south. ³⁷

OPEN IN VIEWER

Table 2.

Lists of sites with eastern and western orientations and their concomitant topographical positions along the horizons. The topographical information was gained using the 5 m, 3D panoramic landscapes. MinLSs: Rising Moon at the minor standstill in the south; MinLSn: Setting Moon at the minor standstill in the north.

Name of dolmen	Most close horizon (west = Grp 1; east Grp 2)	Azimuth of dolmen orientation	Possible target time	Degrees from phenomena in azimuth	Opposite Azimuth of dolmen orientation	Possible target	Degrees from phenomena in azimuth
Arca da Piosa	East	95	Equinox	3	275	Equinox	5
Prado de Mina – (Mina de Folias)	East	94	Equinox	2	274	Equinox	3
Dolmen Cabaleiros	East	92.5	Equinox	1	272.5	Equinox	2.5
Pedra Cuberta	West	98	Equinox	3.8	278	Mid-point SS set & Equ	16.1 and 18 either side c. 1° from midpoint
Parxubeira	West	97	Equinox	4	277	Mid-point Equ and MinLSn	13 and 12.5 either side
Casota de Freans	West	110	Mid-point WS and Equ	16.2 and 17.1 either side 0. 5° from midpoint	290	MinLSn	1.5 (5 fr SS)
Dombate Antigo	West	110	Mid-point WS and Equ	15.8 and 16.9 either side 0.5° from midpoint	290	MinLSn	1.0
Mina de Recesindes	West	115	MinLSs	4	295	MinLSn	0.05
Casa Dos Mouros – Pedra Arca (Regoelle)	West	112	MinLSs	4.6	292	MinLSn	Dead-on
Anta da Gándara de Baiñas 1	West	117	MinLSs	Dead-on	297	Midpoint?	MinLS (3.8) SS set (4.3)
Pedra Moura Carmeo (Carnio)	West	117	MinLSs	1.5	297	MinLS set	Between MinLS (2) SS set (5)
Aquela Banda	East	124	MinLSs	Dead-on	304	SS set	0.02
Caxadas	West	124	WS rise	Between MinLS (4) WS (2)	304	SS set	1.0
Rabós M1	West	120	MinLSs	3	300	SS set	0.05
Dombate New	West	126	WS rise	1	306	MajLS	Dead-on
Pedra da Arca (Malpica)	East	122.5	WS rise	Between MinLS (3) WS (2.2)	302.5	SS set	1.5
Fornela dos Mouros (Aprazadoiro)	West	128	WS rise	Between WS (2.6) MajLS (5.4)	308	N/A	X

Our landscape model assessments show that there are a further five out of 17 sites (5/17 ³⁸ ) within 4° of the MinLS, with two of these four (2/4) precisely on target (Table 2). These, along with the previous two sites, gives a total of seven with some possible focus on the MinLS.

There are also five monuments within 4° of the Sun rise that occurs closest to the time of the equinox, ranging from 1° to 4° away from the exact location on the horizon. ³⁹ Finally, we have two sites whose orientation sits almost exactly mid-point between the Sun rise closest to the equinox and the WS sunrise, with both orientations 0.5° away from the midpoint. Early farming almanacs and cultural celebrations mark these solar days. ⁴⁰ Interestingly, both sites are in quite different geographical locations. In sum, we have 7/17 constrained orientations (3 lunar, 4 solar) that are within 2° of a possible target as well as two almost dead-on mid-solar points (2/17), so potentially 9/17 constrained targets. ⁴¹ If we set a limit of 95% significance, we can calculate the probability to obtain this number out of chance alignment. ⁴² For the 2° accuracy, and five possible targets (equinox, midway between solstice and equinox, MinLS, Solstice and MajLS), we have that p = 0.02 and we can exclude the chance alignment at the 95% level. What is of some fascination, is that one of the constrained solar orientations from the first group comes from a monument that also has a lunar illumination orientation, because the orientation sits between the Sun at the winter solstice and the Moon at the time of the Minor standstill (Caxadas). It is possible, that there was a second tomb designed in the same manner, Pedra da Arca Malpica. For whilst this tomb’s orientation fell within range of our apriori defintion of illuminated alignments and not constrained (2° accuracy), we note that, as well as being within 3° of the MinLS, it was within 2.2° of the WS. For the purposes of consistency, however, we must categorise this site’s orientation as including two illumination targets. Overall, then, we have 7 illumination orientations, but where two of these orientations have two illumination targets each, for Fornela dos Mouros also has one of each of the solar and lunar targets. For the 16/17 orientations within the 4° limit we obtain a probability of 2.23 10⁻⁵ and again we can exclude chance alignment beyond our significance level. ⁴³

Targets in the west: The other side

Considering an idea of Belmonte, that researchers should contemplate investigating landscape features that may indeed be in the opposite direction of the corridor or tomb opening, such as mountain peaks that are essentially siting ‘behind’ the tomb, ⁴⁴ we consider the idea of astronomical targets setting in the opposite directions of our eastern orientations (Table 2; Figures 2 and 10). If we consider orientations within the 4° window, as we did above, we find that we have in total 15/17 orientations hitting a target compared with 16/17 for the orientations facing outwards from the monument towards the east. So, nothing yet distinguishes these east versus west orientations (Figure 10; Table 2).

However, we can see that, on examination of the 3D landscape panoramas, we can confirm that four dolmens (4/17) are aligned to within 2° of the event of the rising Sun at the SS (Figure 10; Table 2). These are more accurately aligned than those of the WS towards the east and thus they do not sit centrally between the SS and its flanking the settings. There are also two equinotically aligned settings within 4° (2/17). Many are in one way or another linked to the position of the Moon’s setting closest to the time of the minor standstill (6/17) and one precisely on the Moon’s setting at the time of the major standstill (1/17). This number is similar to the orientations towards the east, except again the orientations appear far more accurately placed than those found in the east, with 5/6 minor standstill orientations being within 2° or less. There are also two clear midpoint settings, one at Parxubeira and Pedra Cuberta (2/17). Most of these orientations sit to the south of the summer setting Sun within about 26° of this setting point.

Most notably, in the west we have 10/17 constrained orientations (6 lunar, 4 solar) that are within 2° of target as well as two constrained mid-targets (2/17), one solar (Pedra Cuberta), and one likely lunar (Parxubeira). ⁴⁵ So essentially, there are 12/17 constrained orientations. The probability of chance alignment for the dolmen orientations as a group is now 2.83 10⁻⁵ excluding that possibility.

We also have three separate ‘illumination’ orientations (3/17), two solar settings closest to the time of the Equinox (2/17; ±4°; Cabaleiros, Mina Folias) and one lunar setting closest to the time of the Moon at the minor standstill (1/17; Gándara). This makes 15 separate orientations out of 17 within the 4° window, and now the probability, p = 0.0002, again excludes the possibility of chance alignments beyond our set limit.

Target discussion

Clearly there is a variety of orientations in both directions, but significantly they cluster within just a few groups which have narrow ranges in azimuth and are thus tightly contained, especially given that there is a choice of the entire horizon measuring 360° (Figure 2).

Whilst not statistically compared, the fact that there is a higher number of constrained orientations which ‘hit’ targets towards the west with a greater degree of precision than towards the east is intriguing and there are certainly more illumination targets (not orientations), as defined by us, in the east that in the west (8:3).Whilst the specific difference in number of constrained targets – west:east – is not great, the number of orientations less than 1° is 8:4, or 6:2 if we do not count the midpoints. There is also a possibility that some alignments may well be created for both solar and lunar orientations, sometimes where one is constrained and the other is more likely for illumination.

Interestingly, in our initial examination at least, tomb form cannot be linked to a specific astronomical focus. So, whilst it is possible that the three great tombs of Dombate, Pedra da Arca Malpica and Caxadas have a solar reference in the SE (WS), two of these include a possible lunar focus (Table 2). Other architecturally similar tombs also seem to lack a similar definitive astronomical focus, for instance: Casa Dos Mouros (Pedra Arca, Regoelle), the orientation of which which is unknown but within 5° of the MinLS. Also, the Sun at the astronomical Equinox has been considered an unlikely target for prehistoric cultures north of the equatorial regions, ⁴⁶ so in the cases of Pedra Cuberta (4° south of the Sun at the equinox) and Arca da Piosa (2.8° south the equinox) we could argue for other ‘equinoxes’ such as the observable horizon midpoint between solstices. However, we do note that Pedra Cuberta and Parxubeira have close structural similarities and both face the same direction. ⁴⁷ It is possible that differences in tomb structure are hard to perceive and a much closer study is intended to relate form to other dolmen qualities along with orientations in these monuments.

Despite these uncertainties at this point, we have more of an idea of what people could see from the dolmens. Importantly, given the high number of constrained target hits in the west, we argue that this direction might have been as relevant as the east. We have also addressed important methodological issues. Yet we still haven’t examined the horizon shapes in detail and how these details interact with the locations and the orientations of the dolmens.

Horizon distance and dolmen targets: Some observations

Here we must remember that in Figure 10a–q, the 3D landscape panoramas have been arranged into the dominant two groups of relative horizon profile patterns: those with the closer horizons in the west than the east first (Group 1, n = 12) and then those with closer horizons in the east than the west (Group 2, n = 5), following.

There are some striking observations for the Group 1 sites when we compare their eastern and closer western orientations. We have 5 unique constrained astronomical targets for 5/12 orientations (including midpoints) and 5 unique illumination target for a further five sites (5/12). We also have a single orientation that potentially includes two targets (MinLS and the WS). Here Caxadas possesses a solar constrained and lunar illumination. One site has no clear target. In total then, considering constrained versus illumination, the ratio is 6:6 for 12 orientations with 11/12 orientations having at least one target.

Looking west, on the other hand, including the midpoints, we have 10 unique constrained astronomical targets for 10 separate orientations. We also have a separate single orientation that potentially includes two illumination targets (Gándara, MinLS and the WS) and a further orientation that has no clear target (it is set beyond the paths of the Sun and Moon). In total, then, considering constrained versus illumination, the ration is 10:2 for 12 orientations with 11/12 orientations having at least one target. Group 1. We also saw that seven of these 10 constrained orientations were within 1° or less of a target. So we calculated whether or not the probability of our 12 orientations having seven such precise hits onto one of the possible seven major astronomical targets was by chance (namely, the Sun at the equinox, the two solstices, or the Moon at the time of one the four standstills – no midpoints were included). We found that that the likelihood that these seven hits were due to chance was 0.0096, indicating that their appearance was unlikely due to chance factors. As the numbers are rather low, we understand that we should take care in interpreting the statistical result of Group 1 and use it rather as a possible guide or indicator, for the outcome might change if in the future new data can be added to the calculations.

Three out of the five dolmens in Group 2 looking to their closest horizon in the east show orientations closer to the time of the equinox. One other dolmen is possibly compatible with WS (Pedra da Arca, Malpica), certainly in terms of illumination, and the fifth, Aquela Banda, sits dead on the horizon at the time of the Moon setting closest to the minor standstill. Still looking in this direction we observe that these monuments are made up of three (3) constrained and three (3) illumination orientations, as predefined above, where Pedra da Arca, Malpica is seen to contain two potential illumination targets. ⁴⁸

Towards west, two out of five (2/5) have the wider orientations generally towards the time of the equinox (Mina de Folias and Cabaleiros) and two have constrained orientations towards the Sun setting at the SS (2/5, Pedra da Arca Malpica and Aquela Banda). As with its eastern orientation, it is possible that Pedra da Arca was constructed for a constrained and an illumination effect towards the west, for it is also 4.15° along the horizon from the Moon setting at the time of the major standstill, 0.15°, just out of the predefined illumination window.

All of the information on orientations so far suggests to us two most interesting observations. Firstly, in the direction of the closer horizon, there is a possible tendency for a more accurate alignment. So, for Group 1 towards the west, we have 10/12 constrained orientations and for Group 2, towards the east there are 3/5. Secondly, looking at all orientations towards the east, we find 8/17 include an illumination alignment, but facing the west, there are only 3/17 (possibly four if we consider Pedra da Arca). It is tantalising to suggest that this difference might be due to the lack of an entrance or chamber to illuminate on the western side of the dolmen. In other words, it might be a secondary-order data effect suggesting that illumination might have been more important for the chambers of these particular dolmens.

Interpretive and concluding remarks

In our first paper focusing on these dolmens of Costa da Morte, we found that there was a statistically significant difference between the average random and observed horizons in terms of horizon altitude (angular height) combined with horizon direction, as well as horizon distance connected to horizon direction. We also noted, that on examining the average observed horizons there appeared to be striking topographical features centred upon cardinal east and/or west and that these fairly symmetrical features appeared to contain the area for the solar sunrise range. ⁴⁹

Having investigated these possibilities further, we can confirm that these observed topographical features are not just an effect of the mean horizon profiles but actually occur at individual sites. The same is true for the following (i) the Sun’s path closest to the equinox often sits at the very centre of these ridges, (ii) the solstitial paths tend to be towards the inner edges of the tops of these ridges and (iii) the most extreme rising and setting points of the Moon flank the edges or sit on the slopes of these same ridge tops, sitting out further than the Sun (Figure 10). Thus, all the solar and lunar risings and settings are contained along such ridges. However, how the dynamics between these bodies and the horizon were understood by the prehistoric people who created the dolmens of Costa da Morte is still beyond our reach. Are one of these things the controlling entity or force of the events that take place here? Is it the Moon that contains the landscape or are the ridges of the landscapes controlling where these bodies fall?

At this point, we argue that orientations in both directions of the tomb’s axis may well be deliberate. The likely significance of this is underscored by the fact that, despite every tomb facing the eastern part of the horizon, there is no unique choice regarding which type of horizon should be closer – the eastern or western, and, whichever is the closer horizon, there is a tendency to have the more accurate astronomical orientations in these directions, as demonstrated in this paper. It is more likely that each orientation has a particular role and that there is no simple conclusion for the behaviour or beliefs attached to the erection of monuments. Despite these things, we have found some plausible motives related to the different behavioural foci of east and west.

Firstly, from our earlier work, we found that at least two major visual effects resulted from, or are coupled with the topographical horizon choices of the dolmen builders, namely that the horizons in the east are shown to be lower in altitude than those in the west, and statistically lower in the east than the random profiles from the same area. ⁵⁰ These outcomes are also now observable on the individual horizon profiles of the dolmens. The visual effects are: (i) the lower horizons in the east make it possible for the ‘rays’ of the rising Sun or Moon to enter the tomb and such horizons will ensure that the Sun or the Moon can be seen from inside the dolmen entrance or even the chamber, as we see at Dombate in Costa de Morte and the southern dolmens of discussed above – Alberite, Huerta Montero and Menga ⁵¹ , and (ii), with higher horizons in the west, a large dolmen or a future mound ⁵² would not block the views of the horizon or the astronomical bodies setting upon it, behind the dolmen. These effects further support the idea of deliberately chosen horizon profile qualities. Interestingly, Rodríguez Casal 1998 makes the observation that for Pedra Cuberta, the stones were ‘dragged several hundred metres down the eastern slope of a valley, across and halfway up the western slope, apparently so that the tomb might look easterly (and towards an acceptably low horizon)’, ⁵³ offering further and differing support for the deliberate choice of horizon types in particular directions.

Secondly, the relevance of dolmen orientations towards the west is possibly also linked to their architecture, use and decoration, for this is the direction of the dolmens’ back stones. Excavation has shown, for the superlative dolmens especially, that the backstone can be the largest, and the capstone usually second in size to this (e.g. Dombate ⁵⁴ ), and regardless of age, the backstone is most often the largest appearing stone of the orthostats (e.g. Forno dos Mouros ⁵⁵ ). In addition, concerning to orthostat artworks, the backstone or back area tends to contain the majority of painted or incised works within Galician dolmens. Indeed, the ‘stretched hide’ or pele esticada figure that appears in Galicia and Portugal occurs in this part of the tomb, such as that found at Orca dos Juncais. This example also has an antlered deer (cervid) to the top right and a branching motif (another cervid?) above it. ⁵⁶ The pele esticada is considered to be a figurative representation of the hide of an animal. Though interestingly, the same shape found at Coto dos Mouros (Rodeiro) could be interpreted as an anthropomorph. ⁵⁷ Pertinently, Oliveira Jorge ⁵⁸ tells us that in terms of decoration and activity, ‘the most sacred zone of the chamber was the area near the backstone’ and consecutive areas at the bottom of the chamber. Therefore, the backstone area is defined by the archaeological remains and megalithic art as the most sacred of areas within the tomb. ⁵⁹ From this, it is clear there are a number and variety of visual phenomena both within and outside the tomb. We must also recall here that in terms of accuracy, the considered astronomical targets towards the west (through the backstone) seem to outweigh those towards the east.

Together, all these observations suggest that east and west are both significant but likely considered somewhat differently. They possibly have different meanings and this is likely why they are connected somewhat differently to the sky. Using the Galician tomb of Dombate as an archetypical example, the west is possibly the realm of the ancestors, of the dead, the chosen area within the entire structure in which their bodies are laid out with emblematic objects, close to the main arcane ‘lithographs’. These ‘lithographs’ are seen bathed by the light of the Sun or the Moon from the east for a short time at least once a year, warming or welcoming the ancestors into the world of the living at this time. Thus, looking east out of the dolmen with the ancestors, the light of the world and times of the cosmos can be seen. Those standing outside with the Statuettes can turn from east to west on the same or a different day, and see the significant Moon or Sun set on the higher horizons behind the dolmen, setting into the realm of the ancestors. And when the people have left, the Statuettes remain in position, on the side of the living, as witnesses of the ancestors and the cosmic realms, binding all these things together.

Since the methods we have applied are new, and thus the resultant outcomes are different to previous works, it makes straightforward, detailed comparisons with other dolmen groups along the European Atlantic Façade difficult at this point. However, we draw your attention to the discussion in the Introduction, where we demonstrated that lighting was clearly considered and applied elsewhere in dolmens and later tholos in southern Spain. Most particularly, there are chambered tombs which are structured to carefully manipulate the entry of both Sun and Moon light into particular areas of the tomb at certain times of the year, and we can say that in terms of astronomical targets at least, that Costa da Morte shares these characteristics. We know, too, that similar alignments are said to occur in Portugal. Upon examination, Hoskin et al.’s 1998 data showed that, for Portugal as a whole, most of the dolmen orientations fell within the same range as the dolmens of Costa da Morte. ⁶⁰ It was also clear, though, that individual local regions may or may not show the same pattern. Thus in the Minho region, on the border with Galicia, no dolmen yet uncovered on the Sierra has a chamber entrance that faces or is near the winter solstice, plus there are more than 10 orientations just north of east overall in Portugal. ⁶¹

We conclude for Costa da Morte that, all together, the landscape, astronomical and architectural elements presented and discussed here, form a raft of complex characteristics used to create the ‘ideal’ dolmen location. Further research is required to uncover why there is a greater number of illumination alignments towards the east than those in the west, whilst there are more accurate alignments towards the closer horizons (whether east or west). Do lower horizons in the east, which ensure light entering at sunrise, mean you don’t have to be so careful with your alignments, as suggested above? Is it possible that the target of such orientations includes other bodies also and/or we have dual alignments as suggested by some dolmens which contain two possible targets – one constrained and one illumination, or even two of the latter? Further, in terms of illumination effects, what exactly are the relationships between corridors and chambers in Costa da Morte? Do they all allow the light to hit the back stone like in Dombate, or are there some where light only hits the side walls or chamber orthostats? Rodríguez Casal 1998 draws our attention to the fact that in NNE Galicia, in Lugo, there are three tombs whose chamber orientation (based upon the backstone) and corridor orientations are quite different). ⁶² Excavation may reveal similar structures in Costa da Morte that can not yet be seen (the far end of the corridors, for instance). Thus much more work on the connections between internal dolmen structures, the placement of material culture within and around dolmens and illumination is also a priority for us. Farther in the future, we look forward to uncovering rituals that may have taken place in the vicinity of such monuments.