di Andrea Orlando
The ancient Greeks built hundreds of magnificent temples over the course of several centuries, from the 7th century BC onward . Leaving aside regional and chronological distinctions in the lay-out and in the column orders, these sacred buildings were always based on the same conception: an imposing rectangular construction adorned with columns on the fac¸ ade. Although in many cases the presence of columned porticoes on all sides made the view of the structure enjoyable from all directions, the main principle always remained the same: a Greek temple was meant to occupy a natural place with an obviously man-made feature, and it was to be admired from the outside only.
Admission was reserved to priests and to the privileged few, and public rites were celebrated outside, in front of the temple, which in many cases was equipped with an altar and a plateia (religious occasions included festivals, processions and long rituals). The interior of the temple was, strictly speaking, the home of the god. The god’s domestic welfare (hence, the beauty and decorum of the building, correct insertion in the landscape, regular giving of offerings) wasfundamental to assure benevolence and protection to the community.
The cult image, located in the central place of the temple, was
in many cases an out-and-out masterpiece, like the famous ivoryand-gold
statues of Zeus at Olympia and of Athena in the Parthenon
The positioning of Greek temples has been the subject of interesting scholarly research. For instance, a connection between the terrain on which the temple is erected and a related deity has been suggested by Retallak . The relationship with the landscape as a whole was first suggested by Scully . His work pioneered research on the Archaeology of the Landscape, pursuing the idea that landscape and temples formed an architectural unit that was characterized in accordance with the specific god worshipped.
Occasionally, the choice of the terrain and/or of the landscape was exploited to achieve a specific orientation, while in other cases the temple architecture in itself created the sacred landscape . In any case, the matter of understanding the specific orientation of the Greek temples deserves to be dealt with on its own.
The orientation of a Greek temple is preferably defined as the direction of the main axis from inside looking out, which is the direction in which the statue of the god was in principle looking, as well as being the direction along which the sun would illuminate the fac¸ ade, which, as we have seen, was the scene for rites and celebrations taking place outside the temple. The majority of these monuments face the eastern horizon, mostly within the arc of the rising sun [5–8]. Recent research, however, has shown that east-ern orientation is not the universal key to Greek temples, as was previously believed [9–13].
What appears to be a simpler situation occurs in the case of the Greek temples of Sicily. The orientation of the temples of Sicily demonstrates indeed a very clear pattern [14,15]. It has been deter-mined that 38 out of 41 measured temples are oriented within the arc of the rising sun (Fig. 1). This sample is virtually exhaustive for all but one (studied in this paper) of the existing monuments, and clearly we have no need of any statistical analysis to conclude that orientation within the arc of the rising sun was intentional. However, in a way, we are only at the beginning. As it happens, there is no specific concentration of data, for instance, around the solstices or the equinoxes, or other dates for that matter – so how was the alignment chosen? Was it the day of foundation of the temple, or the day of the feast of the god, or what? Perhaps there was a tradition passed down from the original town of provenance? So far, attempts to gain more insight into this problem – for instance, by investigating possible groupings for patron deities – have not been successful. Matters are complicated by the fact that the calendars in use in Greek towns were lunisolar, so that alignments based on feast days would not have been calendrically effective in relation to the timing of the rituals carried out annually in front of the temple (presumably at dawn).
The orientation also appears somewhat unusual when one looks at comparable families of monuments, for instance the Italic temples (temples of the peoples inhabiting continental Italy before the Roman conquest, like the Samnites) which are oriented to the sun ascending in the sky, and the Etruscan temples, which are mostly oriented to the sun ascending or descending in the sky, that is, between the winter solstice sunrise and the winter solstice sun-set [16,17]. Gonzalez-Garcia and Belmonte  have argued that while Greek temples in the same region may differ among them-selves with regard to orientation, nonetheless the orientations as a whole correlate better with those of temples from other, culturally related regions, e.g. the temples in Sicily compared with those of South Italy.
Finally, it should also be noted that solar orientations cannot be distinguished from – eventually – stellar orientations occurring at the same declination. Thus all Greek temples oriented to the rising sun also happen to be broadly oriented towards the constellations in which the sun was rising at the epoch of construction, and can on occasion be accurately oriented to specific stars of such con-stellations as well as to other stars that had the same declination.
A possible, specific interest by the builders in this kind of stellar target must be investigated separately case by case (see e.g. ).
Motivated by such a variety of open questions, we decided to approach the problem by performing a complete survey and analysis of a homogeneous and particularly significant set of temples: the world famous UNESCO site of the Valley of the Temples of Agrigento, ancient Akragas. Our results are actually quite unexpected and show that a variety of factors, not all of them astronomical, influenced the Akragantine architects.
- The Valley of the Temples
Akragas – today’s Agrigento – was one of the most important Greek colonies in Sicily, founded in 582 BC by settlers from the nearby Gela and from Rhodes. The site lies on a huge plateau, naturally protected from the north by the Athena Rock and the Girgenti Hills, and from the south by a long rib-hill, bounded on either side by the rivers Akragas and Hypsas, confluent to the south in a single water’s course, at the mouth of which the port was constructed.
From the very beginning, under the tyranny of Phalaris (570–554 BC), the city was characterized by a regular urban layout, dominated by the Acropolis on the Athena Rock and bordered by the rib-hill which started to house monumental sanctuaries; in the central area were dwellings and public buildings in accordance with an orthogonal grid layout; the necropolis was located outside the city walls. In the last decades of the sixth century BC, Akragas was surrounded by massive walls 12 km long, with 9 gates. The colony reached fame and power under the tyrant Theron (488–471 BC), who defeated the Carthaginians at Himera in 480 BC, and during the years of the democracy (471–406 BC) established by the philosopher Empedocles. It is in this period that the extraordinary series of Doric temples, today comprised in the UNESCO archaeological site called the Valley of the Temples, was built.
There are as many as ten temples in the complex. We list them in the following using the traditional names  which are of current use in all publications on the site; it is however fundamental to recall that only the temples of Heracles, Jupiter and Demeter (and the sanctuary of Aesculapius, which is extra moenia) really belong to these deities:
- temple of Olympic Jupiter (Fig. 2): this is the largest Doric tem-ple in the western Mediterranean. The temple was, however, left unfinished and later collapsed, perhaps due to an earthquake. It is built on a huge stand (56.30 113.45 meters) and was reached through a crepidoma of five steps. The most relevant architectural peculiarity comprises the stone giants (the Telamones), which were probably placed in each intercolumniation. The temple was founded to commemorate the Battle of Himera (480 BC), won by Akragas and Syracuse against the Carthaginians, and it is men-tioned by both Diodorus and Polybius;
- temple of the Dioscurides (Fig. 3): the Doric temple was hexastyle with 13 columns on the long sides (13.40 31 meters); today only a reconstructed corner can be seen. The attribution is completely unknown;
- temple L (Fig. 4): adjacent to the temple of the Dioscurides, it was left unfinished but the foundations excavated in the rock are clearly visible, together with blocks at the northeast corner, and
numerous drums of columns scattered in the building area. The attribution is completely unknown;
- temple of Concordia (Fig. 5): this is the best preserved temple in Agrigento. The temple stands on a pedestal with a crawl space inside that corrects the natural inclination of the ground. It is hexastyle with a peristyle of 6 13 columns on a stylobate of 39.42 16.92 meters, which is accessed by a crepidoma of four steps. Exceptionally preserved are the architraves with frieze of triglyphs and metopes, the geison, and, on the facades, the tym-panum. Probably almost contemporary with the temple of Juno, with which it shares many characteristics, its attribution is com-pletely unknown;
- temple of Heracles (Fig. 6): this is probably the oldest of the Doric temples of Akragas, dating to the end of the 6th century BC. It stands on a mighty pedestal, has a peristyle of 6 15 columns (stylobate of 67.04 25.28 meters), which is accessed by three steps. The attribution is based on a passage by Cicero who mentions a temple dedicated to the hero and placed not far from the agora, which should likely be in this area; · temple of Aesculapius (Fig. 7): this temple is located outside the city walls, south of the Hill of the Temples, on a substrate of sedi-mentation clay. The Doric temple has a portico in antis (with two columns between the doors), and size of 21.70 10.70 meters. It was comprised in a vast sanctuary easily accessible from the sea. A sanctuary to Aesculapius is mentioned by Polybius who placed it outside the walls of the city, so the attribution looks certain;
- temple of Vulcan (Fig. 8): the temple is peripteral, Doric, hex-astyle, with 13 columns on the long sides. Few remnants are preserved of the building: two stretches of crepidoma with four steps and two columns. The temple was built extending an archaic precinct. The attribution is completely unknown;
- temple of Juno (Fig. 9): this temple is located at the eastern end of the hill on a high base. Probably dating to the middle of the 5th century BC, it is hexastyle with a peristyle of 6 13 columns; the stylobate measures 38.15 16.90 meters while the crepidoma is of four steps. The attribution to Juno (Hera) is purely hypothetical;
- temple of Demeter and Persephone (Fig. 10): the temple building was incorporated in the Middle Ages into the Norman church of San Biagio. It was a Doric temple with the cella preceded by a portico in antis. Attribution is firmly based on archaeological material.
At least two further temples existed in Akragas, but not in the Valley. They were located on the Acropolis, which corresponds to the centre of the modern town. One, the so-called Temple of Athena, is buried and partly visible below the Church of St. Mary of the Greeks. The remains of the second lie almost certainly below the medieval Cathedral nearby.
- The orientations of the temples: survey methods and
Results The orientations of the Akragas temples are dealt with in  and, more recently, in [13,14]. As we shall see, however, not all the published data are reliable, and not all the temples have been considered (further details below).We have,therefore, re-measured all
the temples with a high-precision optical theodolite during fieldwork which lasted one week, from 1 to 8 August 2015 (Table 1).
The geology of Agrigento is quite complex , and the upper town is subject to submergence and landslides, largely due to modern man-made factors. The geological situation of the Valley appears to be more stable, but in any case long-term geological activity (which might affect orientations) is not expected to have occurred, although dramatic events, such as earthquakes, cannot be excluded . Agrigento belongs to an active seismic area, which was nonetheless – as far as we can know on the basis of historical records – only marginally affected by the great earthquakes
registered in Sicily; the alleged earthquake sometimes invoked as the cause of the collapse of the temple of Jupiter is not historically proven. So we tend to exclude that the present orientation of the temples might have been affected by intervening natural causes.
Almost all temples have been visited and measured twice on different days. North was calibrated at each measure using a longdistance GPS measure from the theodolite station to a recognizable feature (a corner of a skyscraper) of the modern town of Agrigento (at distances of about 2.5 km).
The nominal accuracy of the transit instrument that we used
is optimal, below 1 . However, this accuracy refers to the direction, which is beingmeasured. Errors can arise because one has to be sure thatthe direction measured really is the azimuth ofthe temple, and in some cases the corners cannot be individuated precisely, while in others the sides are partly covered by huge amounts of sand and/or trees. In all measures we always used one or two poles set in the far corner and in the middle of the side, optically aligned with the instrument (Fig. 11). Allowing ∼20 cm of error in the choice of the measured direction, we actually estimate that the error of our measures can be reasonably assumed to be ±15
. This value is anyhow below the error that the builders might reasonably have committed in tracing their alignments, especially if they used the sun.
Of course we also accurately measured the visible horizon
– defined as the visible heightfrom the centre ofthe entrance to the temple – for each temple. Again, the nominal accuracy is below 1 but it should be noted that in the case of the temples located on the central terrace the measure was possible only with great difficulty, owing to intervening modern features, so that the calculated declinations for such temples are only approximate. However, as we shall see, it also appears thatthese temples were not astronomically oriented anyway (Section 4).
The consistency of all measures was cross-checked with the
readings of a professional SUUNTO compass-clinometer; the compass readings were corrected for magnetic declination using the
calculator kindly provided by NOAA (http://www.ngdc.noaa.gov/ geomag-web/#declination) (magnetic anomalies are not to be expected at Akragas). Our corrected compass readings are fully consistent with transit readings, as the latter all fall within the nominal error band of the corresponding compass readings (±½◦).
Finally, we performed a cross check of the data also using
Google Earth. The compass overlay tool of the program was used for azimuths and the elevation profile tool was used for horizon readings; again, all the data are consistent within ±½◦.
The results of the campaign are given in Table 1. As a comparison with the literature, it can be noticed that our results differ, in some cases considerably, from those of Salt obtained with a magnetic compass. It is difficult to explain this discrepancy since – as
mentioned – we also performed compass readings, and they turned out to be consistent with our transit data. However, Salt mentions the presence of steel fences, which obliged him to perform some of his measures on maps. There are fences in Akragas, encircling many of the temples and restricting access to the public; of course they have no effect on transit readings but they disturb magnetic readings if taken from outside. Our magnetic readings do not suffer from this problem because we were granted full access to the inside perimeter and cellas of all the temples.
Another set of published data is that by Aveni and Romano who worked with a theodolite. In this case it can be noticed that our results in azimuths agree relatively closely. On the other hand, our results in declinations differ considerably. We calculated declinations using the program GETDEC kindly provided by Clive Ruggles, which takes into account atmospheric effects, butthis is not enough to explain the difference with the Aveni and Romano data. It is unfortunately impossible to fully compare the results because these two authors did not give horizon heights; however, a few trials that we have made with their data seem to show that they simply assumed a flat horizon for alltemples, something thatis definitively untrue.
For each temple, we report azimuth, horizon, declination, GPS position, the first solar dates and the bright stars, which eventually correspond to such declination in the century of construction.
4.1. The temples of the central terrace: Jupiter, Dioscourides, Temple L Akragas was planned on the basis on an orthogonal street grid plan in the Greek style, with ‘meridian’ roads (plateai) crossed at right angles by longitudinal streets (stenopoi) (Fig. 12). We have accurately measured the street grid plan; our result is that the grid is orthogonal with a very good accuracy, and the stenopoi are oriented at 78◦ 15 . This orientation is probably topographical as
it is roughly orthogonal to the slope of the Akragas hill. In particular, a stenopos crossed the hill longitudinally heading towards the central sacred area, which houses the circular sanctuary of the chthonic deities, the temple of the Dioscourides, temple L and, to the left of the road, the temple of Jupiter. There is therefore little (if any) doubt that one of the largest temples of the Greek world, the Akragas temple of Jupiter – azimuth 78◦ 30 – was orientated topographically in accordance with the street grid (curiously, the declination associated to the temple is compatible with the star Alpha Arietis, but Aries was possibly related to Jupiter only much later, and with the identification with the Egyptian God Amun).
The same topographical criterion holds for Temple L – azimuth 77◦ 54 – which fronted the road directly, occupying the horizon of any person descending the hill. This did not occur for the nearby temple of the Dioscourides, which formed instead a sort of scenography for the space fronting Temple L. It is probably for this reason that the Dioscourides temple was skewed clockwise, with an azimuth of 82◦ 54 . However, this temple shares the same declination as the Juno temple, whose orientation certainly has nothing to do with the topography of the town (see discussion below). It may therefore be that both orientations have the same origin (for a comparison with the topography of nearby Selinunte see ).
4.2. Equinoctially oriented temples: Concordia and Heracles Among the three temples located on the Akragas ‘rib’, the two westernmost ones, Concordia and Heracles, are oriented cardinally, with the front to due east, with very good precision (½◦ within our accuracy of 15, so the maximal error committed by the builders certainly did not exceed ¾◦).
It is of course tempting to think that these temples might have been orientated to the rising sun at the equinoxes. A non-trivial horizon raises the declination (from zero) for each one of them, but for Concordia and Heracles this deviation is very small(declinations respectively 0◦ 46, 0◦ 34 ) giving dates very close to the equinoxes, as we also verified directly on site with the rising of the sun in alignment with the temple on 24 March 2015 (Fig. 13).
4.3. The sanctuary of Aesculapius The sanctuary of Aesculapius, located right below the rib outside the city walls, is oriented to the cardinal points with an astonishing precision: our measure furnishes 89◦ 57 . The temple is definitively not, however, oriented to the rising sun on the days of the equinoxes, since the non-trivial horizon raises the declination to 2◦ 25 and the solar dates are later than the spring equinox (and correspondingly earlier than the autumn equinox). On the other hand, the temple orientation to the cardinal points is so accurate that it leads us to believe that the main aim of the orientation was not to orient towards the rising sun on specific days of the years.
We propose here instead the idea that it was deliberately oriented cardinally. Astronomy was there, of course, in the determination of true east (or, more likely, of true north) and symbolism was certainly there as well, with the choice of orienting a square sacred building with the sides along the cardinal directions. However, the builders – who were of course aware that the sun rises at true east on the equinoxes only if the horizon is flat – were not interested in this last phenomenon.
4.4. The temple of Vulcan
The Temple of Vulcan (azimuth 87◦ 05
), in spite of its relatively large deviation from true east, is probably to be ascribed to the family of temples orientated along this direction. In fact the topographical setting of the temple is quite special: the temple sits on a separate hill with the narrow Colimbetra valley passing behind. The temple contains – and is built over – an archaic building oriented along the line of maximal slope of the hill ∼80◦, while the project of the enlarged monument was skewed towards the cardinal directions as much as possible with respectto the existing topographical constraint.
4.5. The Temple of Juno
This magnificent temple sits at the easternmost end of the rib, and is sustained by an enormous artificial terrace. The attribution to Juno is only traditional, as no archaeological material supporting
any attribution has ever been found there.
To begin the study of this temple, we verified accurately and without any possible doubt the following fact: placing the very same building with a cardinal orientation was possible without any geological or topographical obstruction and over the very same terrace.
Therefore theremust be a different reason why this temple is not Oriented equinoctially, as the two which follow along the rib – forming an internationally renowned, spectacular sequence – instead are, as discussed above. The azimuth of the temple is 82◦ 24
. We propose here the possibility that this temple was oriented to the stars.
In the 5th century BC, the front of the temple very neatly faces the region of the sky where a relatively faint constellation Delphinus (the Dolphin) was rising (its brightest star is of magnitude 3.8).
As is obvious, it is impossible to speak about the azimuth of rising of an entire constellation, so this assertion needs a detailed explanation.
Delphinus occupies a small portion of the sky, which can be
individuated by a small ‘quadrilateral’ offour stars. Due to theirhigh magnitude, we must of course consider these stars at their minimal visible height due to extinction, which according to Thom’s rule (visible height = magnitude in degrees) is not less than about 4◦. In the Akragas sky of, say, 450 BC these four stars were visible in the region of azimuths between 80◦ 45 and 83◦ 15 , with the unique other relevant star of the constellation, Epsilon, at 86◦. We have therefore a very good match with the Juno temple in the century of its construction.
Owing to the lack of archaeological material, the attribution of the temple is completely unknown. The orientation to Delphinus that we are tentatively proposing could make sense especially if a dedication to Apollo – a god who is currently absent from any other temple in Akragas – could be suggested by future research or excavations. Delphinus is indeed one of the constellations connected to
Apollonian mythology. According to a recent proposal, it was even ,used as a marker for the season of the pilgrimage to Delphi , although the faintness of the constellation casts doubts on its practical calendrical use  and other, brighter constellations, Lyra and Cygnus, were more tightly connected with this God .
4.6. The Temple of Demeter and Persephone (San Biagio)
The temple is of Doric style, and belongs to the final phase of the Archaic period (480–470 BC). It is very well preserved because 8 R. Hannah et al. / Journal of Cultural Heritage 25 (2017) 1–9 in the Middle Ages the building was transformed into a church dedicated to San Biagio. The facade of the church points to the north-west, but it is very likely that the facade was obtained by opening an entrance in the back wall of the cella of the temple, which was therefore originally fronting south-east. However, of the portico with two columns that the Greek temple should have had, only the (so-called honeycomb) foundation sections can be seen. The stone foundation platform of the whole building and a part of the original stone walls are preserved, accurately built in isodomic masonry of huge rectangular blocks. Inside the church, excavations have revealed a cistern belonging to the Greek phase, located close to the north-west corner and therefore inside the cella of the temple, a quite unusual feature.
There is no available space for an altar in front of the temple, since the terrain slopes down abruptly on to the escarpment of the city walls.Arelatively large esplanade is instead present onthe back (thatis, in front ofthe church). This esplanade is contemporary with the temple and was obtained artificially through the construction of huge retaining walls on the south side and an accurate excavation and leveling of the rock on the north side. The area was accessed from the town through a large road partly excavated in the rock as well, which is still perfectly visible today.
During the excavations of 1925 a votive deposit was found,
formed by a large amount of objects. In particular,there were many fragments of two female busts of terracotta, one of which could be reconstructed in its integrity and was identified as Persephone. As a result of this discovery, the temple is attributed to the Eleusinian divinities. The attribution to Demeter and Persephone is also confirmed by the presence of two small circular altars: one, with a diameter of 2.53 meters, is filled in and the other, with a diameter of 2.70 meters, has a central well (bothros), which was found filled with ritual offerings, i.e. broken kernoi, or ritual vessels of Demeter.
These altars are located in the ‘corridor’ formed between the rock cut to the north and the side of the temple.
The front of the temple could be accessed directly by a stairway, which crosses the town’s walls through a postern and leads to a strange building located outside the walls. It is a protohellenic (7th century BC) sanctuary probably dedicated to chthonic deities, whose architectural elements are integrated with the natural features of the site, as is often found in the holy shrines of the gods
of the earth (for example at Eleusis, in Lykosoura or in Enna). The sanctuary consists of a rectangular building up on the cliff (below the temple of Demeter and Persephone), on which there are two, communicating hypogea, which were filled with votive offerings.
A third gallery was used as an aqueduct to supply water collected from a nearby source in the basin of the building, which was therefore a sort of fountain-sanctuary.
The temple of Demeter and Persephone was measured by Nissen ; for reasons we do not know, it has passed unnoticed in more recent works. Nissen was a very conscientious scholar and his data are usually reliable, but not in this case. In fact, we discovered that the azimuth given by Nissen, which is close to that of the rising sun at the winter solstice, is in defect by as much as 4 degrees. As a consequence, the azimuth of the temple falls below the arc of the rising sun. To our surprise, then, we discovered that the temple is the only one in the Valley whose declination (–27◦ 32
) is not in the solar range,1 and adds to the very few Sicilian temples whose declination has this property.
It can be noticed that the building could have been built in the same place but skewed some 4 degrees towards the east, in order to align with winter solstice sunrise, without any practical problem.
1 In fact it is the unique non-solar temple in the whole of Akragas, since we managed to measure the azimuth of the basement of the Athena temple on the Acropolis to be ∼110◦ (a solar azimuth). The horizon there is not measurable.
Even more, although the horizon in front of the temple is flat, a rock curtain located immediately to the left (east) looking from the entrance was leftin situ (thehuge excavationofthe terrace ends just nearby), and this curtain even obscures the midwinter sun at rising.
Perhaps the fact that the temple is not aligned with the rising sun was actually discovered by Nissen, but seemed so unnatural to him that he decided the measurements were wrong? We shall probably never know. In any case,the – clearly deliberate – orientation of the temple cries out for an explanation.
The first thing to be noticed is that the declination of the temple is enclosed between the declinations of the sun at winter solstice and that of the moon at the major southern standstill, and fits the Venus minimal declination. We were therefore at first intrigued by the idea that the temple could be aligned to Venus. However, although Venus can attain – in principle – its maximal and minimal declinations both as the morning and as the evening star, the morning star has never had a declination significantly greater than that of the sun at the solstices in the last four millennia or so. This fact is already well known from studies on Venus alignments in the Mayan world , but in any case we verified it independently – using the software StarryNight Pro 7 – in all the 8-year Venus cycles of the 5th century BC.
We then re-analysed the orientation taking into account the
possibility of a lunar alignment. Lunar declinations are affected by parallax by about ½◦ (see e.g. Ruggles ): the front of the temple – again using the program GETDEC – yields a lunar declination –26◦ 56
. This value looks too far from the major lunar standstill
(at ∼–29◦) to be considered significant per se, although–as is well known–lunar standstills are very difficult to be determined precisely and, as we shall now see, we do propose that the temple was oriented to the Moon. The reason stems from the fact that it is very likely that processional rites were carried outin this place. The rites involved both the fountain-sanctuary and the temple uphill, and we can imagine nocturnal processions coming up from the sanctuary and reaching the temple, in front of which, however, there is not – and there never has been – enough space to house worshippers.
It is therefore conceivable that the people, after the ascent, crossed the corridor between the north side of the temple and the hill (perhaps throwing votive offerings in the bothros) and gathered in the vast esplanade located at the back of the temple (recall that this esplanade has been constructed artificially, and with huge effort, at the same time as the building). So motivated, we visited the temple magain and measured accurately the horizon to the north-west from the facade ofthe medieval church and therefore from the back side of the temple. This horizon is very striking, since it is occupied by the hill where the acropolis of Akragas once stood. In particular, the tower of the medieval cathedral, almost certainly built over the main temple on the acropolis, is clearly visible directly in front of the temple. Measuring the horizon we took into account an estimated average for the height of the modern buildings, and the results in declination are +28◦ 44 uncorrected by parallax, with a lunar declination +29◦ 22 . This result is impressively close to the maximal lunar declination which in the 5th century BC (due to the slight variation of the obliquity) was around 28◦ 50 .
As mentioned above, precise lunar extremes azimuths are very difficult to individuate. However, the full moon near to the winter solstice in the years close to the standstill attains an azimuth which is always very close to the maximal one . All in all, then, we propose a lunar connection for the complex of San Biagio, a fact which would fit well with the religious connections of the water cult in the Mediterranean basin.
Our results show that the orientation of the Greek temples of Sicily could be affected by different considerations, so that there R. Hannah et al. / Journal of Cultural Heritage 25 (2017) 1–9 9 is no ‘golden rule’ to explain it. In particular, the long sought ‘day = god’ rule appears to be rather a superimposed idea which is probably due to a misleading parallel with (some) Christian churches. At least in the case of the temples of Akragas, a variety of different reasons played a role.
First of all, the pre-existing urban layout influenced the orientation ofthe temples ofthe centralterrace, including the huge project ofthe temple of Jupiter, which has the same azimuth ofthe stenopos leading to the central terrace. The resulting declination also points to the stars of the constellation Aries, but a relation of this asterism with Jupiter may occur only later; actually, practically all the temples do point also to relevant stars at the time of their construction, but a stellar explanation seems to be (tentatively) feasible only for the Juno temple, which was intentionally skewed to the north of east withrespectto the easterly, equinoctial orientationofthe other two temples ofthe samehill,themagnificentConcordia Temple and the Temple of Heracles.
Secondly, morphological aspects of the terrain could be as
important as symbolical ones, as very probably occurred for the temple of Vulcan. On the other hand, symbolic reasons could be more difficult to incorporate than expected: in particular, our data clearly show that the sanctuary of Aesculapius was deliberately oriented to the cardinal points – with the front to true east – irrespectively ofthehorizonandtherefore independently fromthedate when the sun would have risen in alignment with it.
Finally, there are cases in which orientation was clearly and
deliberately sought in spite of any topographical difficulty: besides the above-mentioned temple of Juno, this holds for the temple of Demeter. The orientation of the latter – which was essentially unpublished before the present study – appears to be understandable if framed in a setting of other buildings: the ‘sacred fountain’ located outside the city walls but connected through a postern, and the huge artificial esplanade to the back of the temple. Considering the ‘anomalous’ (non-solar) orientation of the temple within its (natural and human-built) landscape allowed us to propose a lunar explanation for it.
The present work has been developed within the convention
for scientific cooperation existing between the ‘Ente Parco Valle dei Templi’ and the FDS Laboratory for the Communication of Science at the Politecnico di Milano. The authors wish to thank Giuseppe Parrello, Direttore di Ente Parco, and Dr. Giuseppe Presti, Responsabile di Ente Parco, for their continual cooperation and support.
References A.W. Lawrence, Greek Architecture, 5th ed., Yale University Press, London and New Haven, 1996.  G.J. Retallack, Rocks, views, soils and plants at the temples of ancient Greece, Antiquity 82 (2008) 640–657.  V. Scully, The Earth, the Temple, and the Gods: Greek Sacred Architecture, rev. ed., Yale University Press, New Haven, 1979.  B. Castro, I. Liritzis, A. Nyquist, Oracular functioning and architecture of five ancient Apollo temples through archaeastronomy: novel approach and interpretation, Nexus Netw. J. 18 (2) (2015) 373–395.  H. Nissen, Das Templum: Antiquarische Untersuchungen, Weidemann, Berlin, 1869.  F.C. Penrose, On the Orientation of Greek Temples, Being the Results of Some
Observations Taken in Greece and Sicily in the Month of May, 1898, Proc. Royal
Soc. Lond. 65 (1899) 370–375. R. Koldewey, O. Puchstein, Die griechischen Tempel in Unteritalien und Sicilien, A. Asher & Co, Berlin, 1899.  W.B. Dinsmoor, Archaeology and Astronomy, Proc. Am. Philos. Soc. 80 (1939) 95–173.  E. Boutsikas, Placing Greek Temples: an archaeoastronomical study of the orientation of ancient Greek religious structures, Archaeoastronomy 21 (2009) 4–19.  E. Boutsikas, C. Ruggles, Temples, Stars, and Ritual Landscapes: the Potential, for Archaeoastronomy in Ancient Greece, Am. J. Archaeol. 115 (2011) 55–68.  I. Liritzis, B. Castro, Delphi and Cosmovision: Apollo’s Absence at the Land of
the Hyperboreans and the Time for Consulting the Oracle, J. Astron. Hist. Herit. 16 (2013) 184–206. I. Liritzis, H. Vassiliou, Astronomical orientations of ancient temples at Rhodes and Attica with a tentative interpretation, Mediterr. Archaeol. Archaeometry 2 (2002) 69–79.  A.M. Salt, E. Boutsikas,KnowingWhen to Consultthe Oracle at Delphi,Antiquity 79 (305) (2005) 564–572.  A.F. Aveni, G. Romano, Temple Orientations in Magna Graecia and Sicily, J. Hist. Astron. 31 (2000) S51–S57.  A.M. Salt, The astronomical orientation of ancient Greek temples, Plos One 4
(11) (2009) 1–5. A.F. Aveni, G. Romano, Orientation and Etruscan Ritual, Antiquity 68 (1994) 545–563.  G. Bagnasco Gianni, S. Bortolotto, G. Magli, Astronomy and Etruscan Ritual: The Case of the Aradella Regina in Tarquinia, Nexus Netw. J. 15 (2013) 445–455.  A.C. Gonzalez-Garcia, J.A. Belmonte, Sacred architecture orientation across the Mediterranean: A comparative statistical analysis, Mediterr. Archaeol. Archaeometry 14 (2) (2014) 95–113.  P. Marconi, Agrigento, Vallecchi Editore, Firenze, 1929.  V. Cotecchia, F. Fiorillo, R. Pagliarulo, A. Reina, Caratteri geologici della Valle
dei Templi (Agrigento), in: Proceedings of I Convegno Gruppo Nazionale di Geologia Applicata ‘La città fragile’. Taormina, 1995, pp. 10–18. V. Cotecchia, F. Fiorillo, L. Monterisi, R. Pagliarulo, Slope instability in the Valley of Temples, Agrigento (Sicily), Giornale di Geologia Applicata 1 (2005) 91–101.  R. Hannah, G. Magli, A. Orlando, The role of urban topography in the orientation of Greek temples: the cases of Akragas and Selinunte, Mediterr. Archaeol. Archaeometry 16 (4) (2016) 213–217.  I. Sprajc, Venus orientations in ancient Mesoamerican architecture, in: C.L.N. Ruggles (Ed.), Archaeoastronomy in the 1990s, Group D Publications, Loughborough, 1993.  C.L.N. Ruggles, Astronomy In Prehistoric Britain And Ireland, Yale University Press, New Haven & London, 1999.