Abstract
The work presented in this article is part of a wide-ranging and ambitious project, started few years ago, to study the role of astronomy in Roman urban layout. In particular, the main aim is to check whether Roman cities present astronomical patterns in their orientations. The project emerged from ideas on how to properly orientate the main streets of a Roman town, as attested in a number of ancient texts and later discussions led by contemporary scholars. We present here the final conclusions of a particular study developed in the Iberian Peninsula (Roman Hispania), where the urbanism that we tend to characterize as properly Roman flourished during both the Republic and the Empire. The sample analysed includes 81 measurements of Roman urban entities spread throughout the Iberian Peninsula and is the largest dataset obtained in a specific region so far. Our results present suggestive orientation patterns that seem to point towards an astronomical intentionality.
… gave four reasons to account for the formation in men’s mind of their ideas of the gods … And the fourth and most potent cause of the belief … was the uniform motion and the revolution of the heavens, and the varied groupings and ordered beauty of the sun, moon and stars, the very sight of which was in itself enough to prove that these things are not mere effect of chance.
Introduction
One of the characteristic elements of the Roman society is the incorporation of normativity in measuring units or in the time reckoning, something that was also present in the division and planning of the territory confiscated. These practices played a key role in the expansionist project since it was essential to reorganize the territory after a new conquest. The foundation and refoundation of towns included important decisions about the spatial organization of the new urbs. For example, having a rationalized space division was very helpful for calculating the taxes that should be collected for each land lot. 1 In addition, the design of towns was subject to empirical guidelines involving environmental factors, such as providing the healthiest conditions for the inhabitants. However, symbolic aspects were also implicit, for example, in the necessity of performing a sacred ritual prior to the foundation of a new settlement, in the election of the correct location, 2 and in the use of the orthogonality in the arrangement of the main streets of a town.
The present paper focuses on the Iberian Peninsula – Roman Hispania – and is actually an updation and improvement of a previous one.
3
Interest in the urban development in this area is indisputable since during the Roman presence over more than five centuries – from the Roman landing in Emporiae in 218
In the same way as in previous works, our main aim is first to find if a statistically significant sample of the orientation of Roman towns in this area shows signs of following any patterns. This in itself would be an interesting result, since a purely practical motive, such an avoiding the winds or following the local topography, should not lead to significant concentrations in orientations over a large region. If such concentrations are found, we would then seek to determine whether these could have an astronomical explanation. In particular, results will be compared with solar positions, at dawn or sunset, in order to detect whether they can be explained by important dates in the Roman calendar, as well as by further astronomical events that might be potentially relevant at that time and within the Roman worldview.
The sample analysed here consists of on-site measurements of 81 Roman urban structures, comprising a statistically significant dataset, as well as being the largest set so far of orientations of Roman sites measured in a specific region. Through the statistical analysis of this dataset, it is thus possible to obtain sufficiently reliable results to enable us to infer whether astronomy really was embedded in Roman city planning.
Finally, an archaeoastronomical approach to Roman urbanism in Hispania can contribute towards gaining a holistic understanding of the role of the city within a particular historical context. Through astronomical orientations of the streets, Romans could have commemorated relevant historical or religious events as well as express ideas about their spatial or temporal order.
Roman urbanism: halfway between symbolism and pragmatism
The Roman Empire would have been impossible without cities. This is because of their major role in land control and exploitation as well as in the production of Roman citizens, who maintained the Urbs hegemony over conquered lands. Nonetheless, the role of a city as a Roman advertisement was not merely strategic but also symbolic. The creation of a territorium – or a culturized and socialized space – 5 implied control over lands and enforced social hierarchies that favoured the status quo of Roman elites. Furthermore, new settlements normally incorporated characteristic features of Rome itself, highlighting its authority as they acted as small replicas of it. 6 Besides this, religion would have played a major role in this endeavour by converging with politics in the control of natural forces. For example, from their founding acts, cities were sacred land but on a global scale. And, most notably when Roman expansion was at its peak, Romans endeavoured the control of time and space by monitoring cosmic motions. This was in order to create a time reckoning system and space division subject to government will. 7
Execution of a new urban project
To undertake the ambitious urbanizing endeavour, it was necessary to possess, together with an enormous manpower, an ensemble of technicians – such as engineers, professional land surveyors (agrimensores), and architects – to design, arrange, and mark out a site prior to the development of infrastructures. Some of them were the curator operis, who coordinated the project, the architectus or mensor who defined the plan, and the gromatici or agrimensores, who laid out the divisions, but also measured irregular plots, understood the complexities of the land law, or even became judges in land disputes by the Late Empire. 8
Several treatises have survived from many of those experts, most of them grouped together in the so-called Corpus Agrimensorum in which they laid down a set of guidelines about how to proceed when starting a new land division project. A significant number of those texts date from the Age of Augustus, when it was really practical to have common technical criteria to carry out the Emperor’s vast programme of expansion. 9 Settlements are normally easy to identify by land laid out in regular plans with orthogonally consistent orientations based on two main axes: decumanus and cardo maximus, which ideally run east-west and north-south, respectively (Figure 1(a)). These streets might have had a cosmological meaning that was printed in the land by the topographers. 10 Furthermore, in the opinion of some scholars, the treatises of agrimensura could have suffered a process of secularization that masked more cosmological principles. 11 The forum should have been placed at the intersection of both main axes. It was the political, commercial, and administrative centre of the town, where the most important public buildings, such as the basilica or the curia, were located (Figure 1(b)). 12

Reconstruction of the orthogonal plan (a) and forum square (b) of Baelo Claudia, a Roman town in Bolonia, Cádiz (Spain). Image (a) courtesy of the museum at the archaeological ensemble of Baelo Claudia.
But, regarding the orientation of the streets, there is no general agreement among authors. Surveyors normally had to cope with a wide variety of topographical settings that were central in the execution of a new project. In this sense, when examining a site, one should consider these obstacles, as well as other geographical peculiarities such as the weather or water supply. However, and as previously mentioned, new towns were thought to be a new sacred land, and thus a symbolic orientation could reinforce the cosmological meaning of such sacredness and if possible it was also sought and implemented. In this sense, Varro states that Romans inherited from the Etruscans the interest in agrimensura and possibly also in orientating their towns according to the cardinal directions. Nevertheless, it is clear that in many occasions such rule was not followed.
Frontinus and Hyginus Gromaticus 13 mention a number of different methods, some of them non-astronomical, others astronomical, such as using a gnomon or the direct observation of the sun given the suitability of a decumanus following the path of the sun or the moon. Vituvius stressed as a priority the avoidance of the main winds to provide the healthiest conditions. 14 However, he also stated that architects should master astronomy to achieve these practical aims. 15
While in some cases, the texts mention direct observation of the sun, it must be stressed that such was not always needed. There is an alternative technique that Romans might have applied to attain a particular orientation: the uaratio. Attested by Marcus Iulius Nypsus
16
in the second century
Foundation as a sacred act
An unavoidable dimension of Roman urbanism was the requirement to perform a foundational ritual, which additionally supports the idea of the existence of an underlying symbolism in the directions of the Roman urban features. This ceremony was supposedly based on that conducted by Romulus in the mythical foundation of Rome and was described by many ancient authors, such as Varro 18 or Plutarch, 19 and by several contemporary works. 20 According to this tradition, new towns should have been founded item conditae ut Romae. The various interpretations of this rite make a single re-creation of it virtually impossible, but different versions of the rite do arise and these share a common background based on well-defined steps during the performance. This ritual was led by an eclectic group of people that included at least an augur – one who inspected the skies and the fly of birds for auspicious signs (auspicia ex avibus) – the founding magister and an agrimensor. It was the latter who allegedly set the direction that the urban grid should follow, possibly by using, among other surveying instruments, a groma. 21 This was a widely used instrument that helped to trace long lines at right angles in the terrain (see Figure 2).

Reconstruction for the groma held by a Roman land surveyor at tracing orthogonal lines. Image courtesy of the Archaeological Museum of Carmona (Sevilla, Spain).
Previous works
The first papers that explored the orientation of Roman towns tried to elucidate whether there existed any connection between the spatial design of cities and their date of foundation. 22 A few years later, Le Gall 23 stated that Roman towns do not present any preferred orientation pattern. But, more recently, several publications on the orientation of Roman towns 24 and military camps 25 have argued for a coincidence of the orientations of the city grids or forums with sunrise or sunset at particular moments of the Roman calendar, when geographically restricted areas are considered.
We must also bear in mind that most of the Roman expansion occurred at the peak of Hellenism, characterized by the mixture of elements that combined the concept of Cosmos, the ruler as a divine entity, and the effective display of such elements in some of their architectural manifestations. Previous studies point towards a motivation of this kind behind the orientations of towns and monuments from the Hellenistic period: for example, the city grid at Alexandria 26 and the Mausoleum of Antiochos I at Nemrud Dag. 27
In fact, our hypothesis is that the Romans would have embodied in their civic plans certain aspects of their conception of the cosmos, along with state propaganda, in order to extol a particular Emperor or dynasty. That is why we find convenient to attend to orientations, which could contain underpinning astronomical configurations that may help us grasp either political or religious ideas by their expressions through the semantization of the space by the introduction of symbolism in the created urban landscape.
Data sample and methodology
This research is framed within the multidisciplinary field of archaeoastronomy. More specifically, the research involves the statistical analysis of a dataset of orientations that are related to positional astronomy and the landscape surrounding the sites. Based on these concepts, the present work is also carried out from a landscape archaeological perspective since measurements are interpreted with regard to a given settlement’s position and its configuration in connection with the visible space, which includes the sky. The aim of studying all these data statistically was to extract orientation patterns and test whether these were in general isotropically distributed throughout the horizon or whether, on the contrary, they fitted particular orientation trends linked to astronomy.
The sample analysed here is the most complete set so far of orientations of Roman settlements studied in a specific region. The dataset consists of the azimuths and altitudes of the horizon of the main urban structures of 76 Roman sites spread throughout the Iberian Peninsula. These sites constitute approximately 75 percent of the identified Roman towns in Hispania at the moment, as derived from a list of toponyms resulting from various ancient sources. However, not all the localized settlements are visible today; rather, many remain unexcavated. There are also more than 30 attested but still non-located sites. 28 We can therefore state that we have a reasonable number of data to test and enable the identification of trends in the orientation of Roman towns.
Figure 3 shows a map with the location of the settlements included in the sample. The measured structures depend on the state of preservation of the remains in each case. Our primary focus has been the main streets of a town – cardo and decumanus maximus – and/or the forum sides. In several cases, the absence of well-identified remains of those elements prompt us to measure secondary roads (called decumani and cardine) or even ancient buildings that follow the same direction of the leading streets. In all cases, we have undertaken more than five measurements at different structures in each site to reduce the uncertainty in our measurements.

Location of the 76 Roman settlements whose orientations are included in the data sample. Five of these sites comprise two different urban layouts. The boundaries between adjoining provinces in Augustan times are roughly indicated, as well as the status of settlements.
The sample is at the same time homogeneous (since all settlements included were Roman at some point) and heterogeneous if one focuses on details. For example, lifestyle in the Iberian Peninsula differs among inhabitants from the inner regions and those from Mediterranean cities, who were in contact with more urbanized eastern cultures through trade prior to the Roman expansion. The dataset also contains settlements of diverse status – as oppida, municipia, or coloniae, depending on the terms of foundation and the type of inhabiting population, among other factors – founded during either the Republic or the Empire, and foundations ex-novo as well as re-foundations over pre-existing settlements.
All data were acquired in situ during a number of fieldwork campaigns conducted by the members of our research group. The instruments used were precision compasses and clinometers in order to measure azimuths and altitudes of the horizon; it should be noted that altitude is often missing in formerly published papers. These instruments introduce nominal accuracies of ¼° and ½°, respectively. Azimuths needed to be corrected for magnetic declination and this magnitude was obtained in most cases from a recent geomagnetic model, WMM2012 or updated ones. 29 Repeated measurements of the same structures were always taken and in most cases data were acquired by more than one instrument. In consequence, the final values are the mean of the several individual measurements at each site.
In order to minimize the intrinsic errors, the architectural elements measured were, as far as possible, sufficiently large and in a relatively good preservation state. In general, we find it appropriate to estimate the maximum azimuth error at about ½°, although it sometimes depends on the degree of preservation. In those cases in which horizons were blocked and the altitudes could not be measured with a clinometer, HeyWhatsThat 30 was used. This is an online panorama generator that computes horizon profiles based on NASA SRTM (Shuttle Radar Topography Mission) images. The adequate precision of this tool has been estimated in a previous study. 31 Measurements with HeyWhatsThat are indicated with an asterisk in Table 1, and we introduced a B (blocked) in the altitude columns when resolution of the terrain model did not allow to drawing the horizon properly.
Orientation of the 81 structures of 76 Roman sites included in the sample.
Measurements of Italica and Corduba are separated but count as one settlements. If the Roman name is known or surmised, it is given in normal typeface; otherwise, it is given in italics. The Roman province to which towns belonged to is highlighted near the site name: (B) stands for Baetica, (L) for Lusitania and (T) for Tarraconensis. Columns contain measurements of azimuth (A) and altitude of the horizon (h) of the main urban structures in each site, their latitude (φ) and the resultant declination towards eastern (δe) and western (δw) horizons. Altitudes with an asterisk indicate that those values were obtained from a reconstruction of the horizon in HeyWhatsThat. When horizon was blocked (B) declinations were computed considering h = 0° and those values are indicated with?
Attending to an astronomical quantity (in this case the corresponding geocentric declination of these orientations, which is independent of the geographic coordinates and the local topography) becomes virtually imperative owing to the astronomical nature of this analysis. By considering the shape of the horizon – and thus the landscape – it is possible to identify directly which cosmic events take place in the observed portion of the horizon. 32
Table 1 includes the 81 orientations of the 76 settlements of the sample. In some cases, there are more than one orientation per site, normally due to a change in the legal status or a territorial reorganization that involved expansionist processes causing the construction of further urban layouts. Italica and Corduba are good examples where the consequences of those events can be appreciated. They have both a Republican and an Imperial sector, which do not follow the same orientation, and are well distinguished parts of a single town. Sites such as these form excellent testbeds in which to observe possible links between orientation and particular political stages.
Azimuth and declination values are represented in histograms calculated using a density distribution with an Epanechnikov kernel with passbands of 1° and 1½°, respectively (see Figures 4 and 5 for azimuth histograms and Figures 6 to 8 for declination, respectively). These values were chosen considering the sample size and the intrinsic errors of the data such that fine details can be appreciated while avoiding the oversmoothing of distributions. In Figures 4 and 5, the data are normalized by the mean relative frequency so the value of peaks indicates their distance above the mean.

Top: the eight divisions of the horizons through which the prevailing winds blow, according to Vitruvius. Where Vitruvius’ precepts were followed, the orientation of towns should avoid the central areas of each sector. Short solid lines in the top figure indicate cardinal points and solstitial azimuths at latitude of 40°. Bottom: the azimuth histogram for the data sample scaled using normalized relative frequency for the entire horizon. Vertical solid lines indicate the limits of the wind sectors shown in long solid lines in the top picture. Red colour solid lines in top and bottom figure indicate the same wind sector, and the shorter blue line the direction of the wind that should run within.

Azimuth histogram for the data sample normalized by the mean relative frequency. Vertical solid lines indicate solstitial azimuths and vertical dotted lines those of lunar standstills at latitude of 40° (mean latitude for the Iberian Peninsula).

Declination histogram towards the east (top) and west (bottom). The dark grey distribution is that of our sample while the light grey area shows declination given by a dataset of the same size as our sample, considering a flat horizon, a latitude of 40°, and homogeneously distributed within the decumanus sector; azimuth within [45°,135°] and [225°, 315°] for the eastern and western horizon, respectively. Vertical solid lines indicate solstitial declinations and vertical dashed lines those of the lunar standstills in all histograms.

Declination histogram of the data sample (dark grey) compared to a homogeneous dataset of declinations distributed over one solar year (light grey). Vertical solid lines indicate solstitial declinations and vertical dashed lines those of the lunar standstills in all histograms.

Declination histogram towards the eastern horizon compared and normalized by the mean standard deviation of a set of 100 random samples of declination, computed with azimuths ranging from 45° to 135°, altitude of the horizon 0° and latitude 40.5° (see text for further details).
The histograms in Figure 6 show declination distributions towards both the eastern and western horizons that are compared with a homogeneous distribution in declination of the same size as the sample. The aim was to test whether our data were either randomly or homogeneously distributed. Since large random samples look like homogeneous ones, we can assume we are comparing it with a random distribution if we consider our sample as large enough. To achieve this goal, we have computed the declination of the same number of azimuths as our dataset homogeneously spread within the azimuth ranges of the decumanus – [45°, 135°] and [225°, 315°] for the eastern and western horizons, respectively – for a theoretical flat horizon without atmospheric refraction and a latitude of 40.5° (an average value for the Iberian Peninsula). Figure 9 shows an example of a solar orientation of a Roman city, as predicted by our data at Segobriga.

Rising sun at Segobriga on 10 August, corresponding to a solar declination of 15° approximately. The same configuration would occur at the beginning of May. Picture courtesy of Rosario Cebrian.
Figure 7 shows data distributions compared with that of the declination of the sun during one solar year in order to test whether or not there was particular interest in specific dates.
Finally, a statistical test was done to discard the null hypothesis that the declination distribution of this sample was compatible with a random one (Figure 8). To achieve that, we have compared the dataset of declinations towards the eastern horizon with a homogeneous distribution like that in Figure 6. Our sample and the homogeneous distribution were also compared with 100 random samples of the same size as that of our data, extracted from a set of 10,000 numbers within an azimuth range from 45° to 135°, considering altitude of the horizon 0° and latitude 40.5°. The sigma value was computed for every random sample regarding to the homogeneous distribution, and we used the mean value of that sigma to normalize the data distribution and obtain Figure 8. Finally, a Kolmogorov–Smirnov test was conducted. Through this test, we compare two samples, the data and a reference distribution, in order to see if both could be drawn from the same parent distribution. This comparison was done for each of the 100 random samples and the declination distribution of the towns studied, and in all of them, it is possible to discard that both datasets come from the same parent population. That is, in all the cases, the probability is less than 0.05 so we can state that the data differ statistically from a random set of declinations with a confidence level above 95 percent.
Orientation of Roman towns
In order to test other hypotheses derived from ancient writings, we found it appropriate to check whether Vitruvius’ guidelines, outlined in his treatise De Architectura, were normally fulfilled by a majority of towns since this work is one of the main sources on Roman architecture. As mentioned above, in his first book, Vitruvius stressed the advisability of architects mastering astronomy; he also described how to properly orientate an urban grid in order to avoid the main winds by using a gnomon. 33 According to Vitruvius, cities should be oriented towards the limits of each wind sector (black solid lines in Figure 4) in order to avoid the main winds therefore blowing close to the centre of the sectors. In Figure 4, it is noticeable how some lines separating wind sectors agree or fall close to local azimuth minima, which is at odds with Vitruvius precepts. Although there are four maxima near those limits, there are absolute minima in the remaining four. Furthermore, there are few peaks in the central areas as well, where winds were expected to run through. In this scenario, we cannot affirm that wind avoidance was the main criteria at orienting a city, although it might have influenced in the city layout in some cases and not for all the winds. A first test with a smaller sample was conducted in a previous paper on Roman towns in Hispania 34 with similar results.
The second step, and the real aim of this project, was to check whether Roman cities presented statistically significant patterns in their orientations. Then we can try to identify if such patterns are astronomical as suggested according to the previously referenced requirements of Frontinus and Hyginus Gromaticus. 35 Even though these authors pointed, among others, to a likely astronomical intentionality, they did not clarify which specific solar or lunar positions needed to be followed, so our purpose is to first test it and then identify and interpret these.
Figure 5 shows how azimuths are distributed within a sector of the horizon of 90° wide centred on a cardinal point. This pattern is repeated in the remaining three sectors, assuming the orthogonality of the urban layout in Roman cities. Given this orthogonality and if the orientation of towns does not follow any general criteria, one should expect that the values tend to vary little from those of a homogeneous distribution in each sector. Figure 5 shows that, although the values are spread over the entire azimuth range, they seem not to be uniformly distributed within it but tend to cluster around specific values.
Figure 6 also stresses that the null hypothesis for homogeneously distributed orientations seems not to be fulfilled. In fact, there exist a limited number of maxima from which a few of them are statistically significant either towards the east or the west (see below). Furthermore, in the case of the east direction, we observe that the three main maxima are on or above 3σ level when the data are compared to random samples (Figure 8). The same occurs when our data are compared with the declination distribution of the sun along one solar year (Figure 7). The divergence among both distributions is evident in histograms at Figure 7, on either the eastern or western horizon. In that particular instance we could reject the idea that orientations were randomly chosen or that they are randomly distributed during the year in the case that orientations had a solar intentionality, instead of deliberately selecting particular days.
Are Roman towns astronomically orientated?
A clear result of our analysis is that the orientation of Roman towns in Hispania cannot be explained by a single cause alone. Our results do seem to indicate that Vitruvius’ precepts were not fulfilled as a general rule although it may work in some cases. One possibility could be that some winds were more important than others. In addition, it seems reasonable to think that surveyors might have attended to local conditions instead of a theoretical model in order to avoid wind running through the streets. Furthermore, even though the azimuths show a non-uniform distribution but few clusters (Figure 5), the most noteworthy results are those derived from the declination analysis.
In the declination profile towards the east, six maxima are observed (Figures 6 and 7) and three of these have a high significance level (Figure 8). These three maxima towards the eastern horizon are more evident in Figure 8, where after the comparison with 100 random samples the value of those peaks do coincide or is above a 3σ level. That is, these values are statistically significant with a confidence level equal or above 99 percent. Furthermore, it has been proved that the declination values of the sample and 100 random samples differ significantly, with a confidence level equal or higher than 95 percent, by following a Kolmogorov–Smirnov test in each case. This fact suggests that the dataset distribution is not random and that those three mean values may agree with certain orientation trends followed at Roman times.
In principle, we do not have any reason to impose a preference to choose eastern or western directions for each town. It is known that a given direction towards east will produce a symmetric orientation with respect to the equinox towards west. This means that if one has an eastern orientation towards winter solstice sunrise, the orientation towards west would be the summer solstice sunset, or if we have an orientation towards a declination of 17° east, the symmetric in the west would be an orientation towards sunset at −17°. This, of course, happens if we have a flat horizon. Once the horizon is not flat, as is often the case, the symmetry is broken, but it may come to our aid by helping us discern the intended direction. For a group of orientation that we suspect to have a cultural meaning, but we do not know if the eastern or western peak should be favoured, the sharper one, that is, with a larger amplitude and smaller deviation, should be preferred, as even considering the topographic conditions it seems to fit better with a given astronomical target. 36 In Figure 6, by comparing the peaks towards east and west, we find that in general, the peaks towards east seem rather sharper than those towards west. In other words, if an astronomical direction was sought and taking into account the local topography, we can discern which direction was preferred by comparing the symmetric peaks towards east and west. The maximum at 15° may have been chosen towards the sunrise, something that looks clear in Figure 8 where its significance is above 3σ.
Regarding the three main declination maxima towards the eastern horizon (Figure 8), a particularly remarkable one is that towards sunrise at the Winter Solstice (δ ≈ −23.5°). This value has a 3σ level in Figure 8 and this orientation has been detected in further Roman settlements in other regions of the Empire, 37 as well as in several remains from previous cultures across the Mediterranean. 38 Moreover, from an astronomical perspective, solstices have traditionally been considered as temporal tipping points by most cultures throughout history in the sense that they represent transitions among seasons and they usually involve some kind of festivities. The Saturnalia is one of the best known Roman festivals and was held over a few days before the Winter Solstice in honour to Saturn, a clearly Roman god with agrarian characteristics and later assimilated to the Greek Kronos. 39 These days were a time of general jollity, and it seems it was commemorated during the Republic but continued to be enjoyed throughout Imperial times. 40 Another interesting point is that Capricorn and the Winter Solstice were incorporated in Augustan propaganda as a metaphorical representation of his rise to power; reflection of the political shifts that were taking place. The epoch of the solar cycle in which days start to become longer would have symbolized the transition from the darker days of the Republic to a brighter era ruled by himself as Princeps. In this sense, the solstitial sun might have represented some kind of divine element endorsing and enhancing the Emperor’s power on Earth. 41
The second salient outcome is a cluster around δ ≈ 15°, which is clearly above 3σ when is compared with several random samples (Figure 8). There are two hypotheses which could acceptably explain this result. The first, and most obvious one, is that the decumani of those cities looked towards the rising sun on the day of the mythical foundation of Rome, on 21st April. Although the strict declination of the sun on that day is approximately 11½°, the dispersion of the results due to intrinsic error in the data means that it is reasonable not to discard this possibility but rather to support it as a plausible solution since that orientation might commemorate this central event. This orientation has been previously found in the Republican temple at Carthago Nova 42 and in the Pantheon in Rome, where a hierophany takes place on this very same day. 43
An alternative interpretation evokes pre-Roman traditions when considering orientations facing the sun on either 1st May or the middle of August (when solar declination is roughly 15°). These dates have been traditionally regarded as Celtic mid-season feasts and it has recently been suggested that they were integrated into the Julian solar calendar in Gaul at some point during Julio-Claudian dynasty. 44 In a recent investigation into the alleged “Celtic calendar,” it has been proposed that a horizon calendar may have been used in areas where Celtic speakers have been identified in order to coordinate different events. 45 These could be found in the orientation of towns or further structures that may be linked with these mid-quarter festivals.
This explanation sounds reasonable for a number of cities in the sample that look towards sunrise or sunset during mid-season feasts. It applies mainly to those cities in the Iberian north-west – a region traditionally considered to have been inhabited by Celtic cultures in pre-Roman times – as well as cities in the Celtiberian area (central Iberia) that contained Celtic theonyms and toponyms in their names; such as Segobriga (see Figure 9). At this point, it becomes advisable to allude to the heated controversy concerning the issue of Celticism that has given rise to long-lasting debates among scholars in part owing to the political use of this concept by northwestern Celtic nationalism in Spain during the nineteenth and twentieth centuries. In fact, this conception of the Celts still remains as the prevailing notion of the term in many academic discourses 46 and this topic has divided currents of thoughts into two juxtaposed positions that Sims-Williams defined as “Celtomania” and “Celtoscepticism.” 47 From the point of view of an hermeneutic project that avoids nationalist interpretations of the term, we find our results to be significant ones since they can attest further forms of manifestation of a similar cultural background; let us call it Celtic.
In this way, these shared astronomical patterns could be expressions of that common indigenous background in addition to the evidences listed by Marco García Quintela of the Celtic presence in the northwestern Iberian Peninsula 48 and further results obtained in previous archaeoastronomical researches. 49 Moreover, they can even represent potential records of how the expansionist process and cultural relationships played out at any given time and place by sometimes imposing Roman frames and sometimes incorporating local ones. A likely integration of pre-Roman practices and beliefs into Roman urbanism has already been manifested in previous works in Hispania, 50 Gaul, 51 and in other parts of the Roman Empire; in particular, the Near East and North Africa. 52
The third of these most significant maxima is out of the solar range (δ ≈ ±26°), but lies within the lunar range, towards both the eastern and western horizons. At this point, it is difficult to give a single and decisive explanation but there are few possible interpretations that could acceptably support the astronomical intentionality of these orientations.
The maximum coincides approximately with δ ≈ ±26°, which necessarily raises a presumable “Venusian” explanation although some of these orientations could be potentially lunar or even solstitial considering the error margin. 53 As derived from the declination histograms and from the above discussions, the “vespertine Venusian” orientation hypothesis would be easier to prove since the maximum towards the west clusters to the southern extreme position of Venus as the Evening star. Nevertheless, the greatest northern position as Morning star should not be dismissed either.
Even though this may sound remarkably speculative, by using concepts introduced by Juan A. Belmonte, 54 it would not be a wild speculation since we are aware of the uncertainties and it is not, of course, by any means a final interpretation. Overall, we consider it is worthwhile to explore this possible scenario at each town to see if this can be a plausible explanation for them.
Interestingly, cardinal or equinoctial orientations are rather seldom found in our sample for Hispania, even though this is what people may assume at first to be the general trend in Roman towns. However, there are three exceptions in Basti, Gerunda, and Libisosa, all located in former Iberian areas where cardinal and equinoctial patterns have been identified by Cesar Esteban, among other authors, in Iberian settlements and sacred places. 55
Conclusion
As shown by our analysis, the data seem to be non-randomly distributed and point towards the existence of some kind of deliberateness in the orientations. As an interpretative act with regard to what motivated the choice of a given orientation in preference to others, we understand that practical reasons may have prevailed at the time of choosing the correct location but the orientation seems to be more closely related to symbolic reasons. As a consequence, the maximum values in declination appear to be connected to sunrise on relevant days of the Roman calendar. A plausible presence of indigenous traditions and beliefs embodied in Roman spatial layouts is inferred from a number of cities orientated towards solar positions on days in which Celtic mid-seasonal feasts were held. It is especially visible in the significance of the maximum at declination 15° in Figure 8. This necessarily leads us to reconsider the terms of the cultural interactions during the Roman expansion, leaving aside the term Romanization as traditionally assumed. 56
Hence, these results act as a backdrop to the implementation of specific technical or ideological standards in land planning, and seem also to reveal signs of the integration of the sky into the creation of landscapes. By this hypothesis, a decumanus maximus orientated according to a specific solar position would have reinforced the grandeur of a place by incorporating the calendar, their ritual, and their symbolism within its spatial design. Making use of a prosopopoeia, the sun could have run through a Via Principalis illuminating it progressively and thus producing a hierophany by rising as the supreme announcer of a memorial date.
According to all the evidences listed along this article, it is reasonable to put aside positions that explicitly deny the presence of astronomical features in the Roman urban layout. Indeed, we strongly support that it arises the need to a more holistic approach to the landscape configuration that inevitably must include the sky in order to get a global understanding of past societies. As stated by Clive Ruggles, 57 “we cannot ignore the sky, any more that we can concentrate upon it to the exclusion of everything else.”
Footnotes
Acknowledgements
We wish to express our sincere gratitude to all those who joined us and participated in the various fieldwork campaigns. We particularly acknowledge Marco V. García Quintela, Lourdes Costa-Ferrer, Margarita Sanz de Lara, José R. Belmonte Avilés, Sixto Giménez Benítez, Dionisia Antón Correas, Javier Alegre García, Alberto García Aparicio, and Jaime Rodríguez Álvarez for their great support. We would like to thank those specialists who kindly offered to guide us and opened the doors of archaeological sites, making our labour easier by their clarifications: Javier Andreu, Rosa María Armendáriz, José Bonnet Casciaro, Alejandro Caballero, Eduardo Cabrera, Marcelo Castro, Rosario Cebrián, Romana Erice, Francisco Escudero, Jose M. Martínez Torrecillas, Salvador Montañés, José Miguel Noguera Celdrán, Josep Ros, Pilar Sáez de Albéniz, Iván Sánchez, José Luis Serrano, and Encarnación Zamora (our apologize if we miss someone).
Funding
This work is partially financed within the framework of the projects Orientatio ad sidera III (AYA2011-26759) and Orientatio ad sidera IV (AYA2011-66787-P) of Spain MINECO. A. C. González-García is a Ramon y Cajal Fellow of MINECO.
Notes on contributors
Andrea Rodríguez-Antón holds a doctorate in astrophysics (Instituto de Astrofísica de Canarias, Spain) and is pursuing research in the archaeoastronomy of the Roman world under the supervision of Drs Juan Antonio Belmonte (IAC) and A. César González García (Incipit-CSIC). Her preliminary studies have concerned the orientation of Roman cities and fortresses in different regions of the Roman Empire and she has produced a number of papers on the Iberian Peninsula, Britain, Jordan, North Africa, and the South of France.
Antonio César González-García holds a doctorate in astrophysics (Groningen) and researches cultural astronomy at the Institute of Heritage Sciences in Santiago de Compostela (Spain). President of the European Society for Astronomy in Culture since 2014, his research interests include the modelling of the possible astronomical orientation of classical cultures and megaliths, possible astronomical and landscape relations of Iron Age sanctuaries, and the study of the orientation of Roman cities.
Juan Antonio Belmonte is an astronomer at the Instituto de Astrofísica de Canarias, where he researches both exoplanets and cultural astronomy. Over the past decade, he has been investigating the astronomical traditions of ancient civilizations, notably Egypt, Rome, and the Middle East. He is the co-editor (with Mosalam Shaltout) of In Search of Cosmic Order: Selected Essays on Egyptian Archaeoastronomy (SCA, Cairo, 2009) and the author of Pirámides, templos y estrellas: astronomía y arqueología en el Egipto antiguo (Crítica, Barcelona, 2012).
