ArticlePDF Available

An analysis of astronomical alignments of Greek Sicilian Temples

Authors:

Abstract

In the eighth century BC something peculiar seems to happen on Sicily. The archaeological record starts to show the arrival of Greek material culture. By the fifth century BC the island is effectively 'Hellenised' and ancient historians record the political and military action of poleis, Greek city-states. Each polis has traditionally been seen as the offshoot of a city elsewhere. Genealogies of cities ultimately end in cities found in the cities of the Peloponnese and the Aegean. The 'Greek' identity of the Sicilian cities is part of a wider debate on the concept of Identity in the ancient world. This paper considers if there is a contribution archaeoastronomers can make to such discussions by considering the alignments of Greek temples. Greek religion was intimately related to notions of civic identity and what it meant to be 'Greek'. I propose a method of studying small samples of temples, which combines both alignment analysis and historical context. Therefore it may be possible that a study of the temples may yield useful information about collective identities. However, as this method shows, the more ambiguous the cultural data the less certain any astronomical patterns may be. Comment: Archaeoastronomy/History of Astronomy 25 pages
1
An analysis of astronomical alignments of Greek Sicilian Temples
Alun Salt, The Centre for Interdisciplinary Science, University of Leicester,
e: alun.m.salt@gmail.com t: @alun
Abstract
In the eighth century BC something peculiar seems to happen on Sicily. The archaeological record
starts to show the arrival of Greek material culture. By the fifth century BC the island is effectively
'Hellenised' and ancient historians record the political and military action of poleis, Greek city-
states. Each polis has traditionally been seen as the offshoot of a city elsewhere. Genealogies of
cities ultimately end in cities found in the cities of the Peloponnese and the Aegean. The 'Greek'
identity of the Sicilian cities is part of a wider debate on the concept of Identity in the ancient
world. This paper considers if there is a contribution archaeoastronomers can make to such
discussions by considering the alignments of Greek temples. Greek religion was intimately related
to notions of civic identity and what it meant to be 'Greek'. I propose a method of studying small
samples of temples, which combines both alignment analysis and historical context. Therefore it
may be possible that a study of the temples may yield useful information about collective
identities. However, as this method shows, the more ambiguous the cultural data the less certain
any astronomical patterns may be.
Binomial Distribution and the temples of Greek Sicily
This paper develops a method initially published by Salt (2009) using data published in the same
paper. The data set comprises 41 orientations of 44 temples dating from the archaic and classical
periods found in ten poleis around Sicily. Some temples have been combined as one data point.
One example would be the Temple A/B at Himera. The Temple B overlies the Temple A, being built
shortly after the smaller temple (Bonacasa, 1970). The reason for combining the two temples is
that the alignment of Temple B seems to be defined by the alignment of Temple A. Therefore the
two temples may not be independently aligned. A response to this approach could be to argue that
the Greeks were perfectly capable of building over older temples with differing alignments as the
ruins at Gela show (Orlandini, 1968). Therefore a reason the two temples share an alignment is
because the alignment is important rather than a happy accident of architecture. I have chosen to
continue the practice of combining of possibly co-planned temples. Temples A and O and F and G
may be pairs of temples planned along a shared axis, and so may not have been independently
aligned. When showing an intentionally astronomical motivation for the orientation of a temple,
2
this method will yield more conservative results, which is desirable when presenting results to a
sceptical reader.
The method in the 2009 paper was an extremely simple application of the Binomial Distribution to
the answer the question of whether or not Greek temples pointed 'east'. The advantages of this
method are its simplicity, its emphasis on the importance of cultural information in defining a test
probability and its applicability to small samples. The output is related to the size of the sample
and so, unlike some statistical tests, smaller samples will produce less emphatic results. The
binomial distribution is given by the formula

P(k;n;p)n
k


 

pk(1 p)41k
where n is the sample size, p is the probability an event will occur and k is the number of
occurrences. This is not a friendly formula. The key feature of the binomial distribution can be
calculated by two considerably simpler formulae. The peak of the distribution can be calculated by
the formula
peak = np
where n is the sample size and p is the probability between 0 and 1. The standard deviation of the
distribution, is calculated by the formula
= (np(1-p))
Therefore the expected probability of temples facing the eastern half of the horizon in sample size
of is 41x0.5, or 20.5. The actual result, 40, lies six standard deviations from the peak value.
There is no formula for significance itself. This is for the reader to judge. Typically social scientists
will attempt to demonstrate an effect to 2, roughly a 95% confidence interval. It should be noted
that these effects are supported by a variety of cultural factors, which give the experimenter
confidence that the test is meaningful. In contrast Schaeffer (2006a, p. 29) notes in astronomical
research that 3 claims are wrong 50% of the time and that a target of 4 or 5 would be
preferable. The reason for the difference is that astronomical data has no social context and so we
cannot always be sure we are comparing culturally coherent samples. Expressing results in terms
of standard deviations away from the predicted result allows the reader to determine whether or
not a claim demonstrates significance.
3
In the case of the 2009 paper, the results vary according to what definition of 'east' is used. If 'east'
means the range of horizon marked by the rising sun the number of temples facing east, 38, is 8.3
away from the expected result of 6.8 temples. The 2009 paper goes no further than analysing the
orientations of Sicilian temples as a whole. Is it possible to isolate sub-samples and see if
significant distributions of alignments occur at local scales?
I shall start by considered the temples of Akragas. Is there something specifically Akragan about
the choices of temple alignments? If the inhabitants of Akragas had a preference for one
alignment, 90º to take an example, then I could examine the temples at Akragas to see if they were
more likely to face 90º than typical Greek Sicilian temples. This sadly is not the case. The temples
face many directions, but they could still be clustered in atypical targets. Working out if this
creates an unusual skew in the temple dataset is not straightforward.
What are the odds that a temple would face the sort of direction that the Akragans chose to
orientate their temples to? Of the 41 data points in the Sicilian set, 9 obviously face this kind of
direction the nine temples found at Akragas. Additionally, the temple of Zeus Melikhios at
Selinous faces 80º, which is the orientation of the temples of Zeus Olympios and Temple I at
Akragas. So the probability of an Akragan alignment would appear to be 10/41 or 0.244. Are there
more Akragan alignments than we would expect in the Sicilian data set? We would expect
41x0.244 temples in a set of 41, or 10. This method clearly does not work. All it shows is that the
temples that face the direction of Akragan temples are the temples that face those directions.
The reason this method cannot work is that when we are calculating a value for p we are not
calculating the probability that a typical Greek Sicilian temple is faces in an Akragan direction. We
are calculating the probability of a set which includes typical Greek Sicilian temples and the
temples from Akragas which may or may not be similar to typical temples. To calculate the
probability that typical Greek Sicilian temples face an Akragan direction it necessary to exclude
the Akragan temples from the initial calculation.
The alignments of the temples at Akragas give us the list of directions an Akragan temple would
face: 80º (twice), 81º 82º, 87º (twice), 90º (twice), 110º. If this reflects a cycle of ritual behaviours
then this is not a continuous range, but rather fragmentary. Therefore it is a matter of which
ordinary Greek Sicilian temples face these events.
4
I shall now borrow an assertion from McCluskey who had a similar problem analysing the
alignments of medieval churches in England. McCluskey (2006, p. 415) has argued that churches
aligned to specific saints would not be intentionally orientated to avoid certain azimuths. Likewise
I shall use the assumption that the alignments of Sicilian temples may indicate a preference for
certain alignments, but not away from specific orientations. Therefore if I examine any specific
sub-group of temples, the remainder of temples from the data set form a usable comparison
sample set of typical Greek Sicilian temples.
Of these remaining 32 temples just one of them, the temple of Zeus Melikhios at Selinous, faces in
one of these directions. Therefore the probability that a typical Greek Sicilian temple will face in
an Akragan direction is 1 in 32. In terms of the formulae above n is 32 and p is 0.03 with the
standard deviation of a binomial distribution being 0.98.
The sample set of typical Greek Sicilian temples cannot be directly compared to the data set I
observed. My set has 41 temples, the set of typical temples is too small. To increase the size of the
typical set to 41 temples for comparison I must multiply n by 41/32 or 1.28. Now for a set of 41
typical Greek Sicilian temples we would expect np or 41x0.03 temples to face the same direction
as an Akragan temple. The standard deviation of a binomial distribution would now be 1.11.
Obviously the data set including the Akragan temples will have more alignments. The question is,
are there so many more Akragan temples that the number of alignments is significantly more?
There are 10 Akragan alignments in the data set, 8.72 temples more than the expected value. This
is 8.72/1.11 or 7.83 standard deviations away from the expected value. If you are satisfied with
claims 5 or higher, this would be a significant result. If you prefer claims to be 10 or higher then
there is nothing special about temples at Akragas.
There would be a few problems with this approach.
Firstly there is the matter of the accuracy of the data. The data is recorded to an accuracy of ±1º.
Therefore any temple which is has an alignment a degree different to any temple at Akragas
should also be considered a match. This produces many more matches. The Temple B at Gela faces
an azimuth of 111º, while the Temple E at Akragas faces 110º. Allowing for errors, these temples
may be facing the same direction. Similarly, the Heroon and Temple Sud-Est at Megara Hyblaea
and the temple of Demeter Malaphoros at Selinous all have orientations within a degree of
temples at Akragas. Therefore the probability of a random Greek Sicilian temple having an
Akragas-like alignment is in fact 5/32. If we apply this value of p to a sample of 41 typical Greek
5
Sicilian temples we would expect to have 6.4 Akragan alignments with a standard deviation of
2.32. This means that the observed result of 14 temples having Akragas-like alignments in Sicily is
only 3.27. The probability of this happening is less than one in a hundred and so I would argue is
worth discussing, but nothing like as striking as the first figure.
A second objection is that I am also assuming that the Greeks themselves built temples to an
accuracy of one degree. This is not an assumption the reader may share. Henriksson and Blomberg
(2000, p. 304) have argued that the accuracy of astronomy in the Phaenomena of Aratus is two and
a half degrees. Aratus was writing during the 3rd century BC, which would make a claim for 1º
accuracy in earlier periods unlikely. While I do not agree with Henriksson and Blomberg for
reasons I shall give below, I can accept that the reader may. Therefore it would be helpful if I could
calculate what the probability is that temples would be thought of as facing the same direction if
this level of accuracy were used.
If we count any temple with an alignment within two degrees of an Akragas alignment as a match,
then the Temple Ouest at Megara Hyblaea and the Athenaion at Syracuse also match temples at
Akragas. The result of 16 temples is now only 2.66 from the expected result. At three degrees
there are three further matches, the temple at Camarina, the Temple Sud a Colonnade at Megara
Hyblaea, and Naxos's Temple C. Significance is now reduced to 2.08.
Even this may be too accurate for some readers, who will note that the Greeks had a lunar
calendar and the same conceptual day of the year cover quite a range of solar declinations.
Therefore the results below are calculated for assumed accuracies between 1º and 10º for solar
azimuths.
Finally there is the matter that this method always produces a positive result. This test alone
cannot define a sample as inherently significant. It can provide a guide as to whether or not a
sample is worth investigating further and for comparing differences between samples. A sample
that produces a staggeringly high result might seem purely statistically significant, but this
significance would still be due to the value of p and this figure must be justified with reference to
some cultural information.
The calculations for Akragas produce the following results.
6
Presumed Accuracy
10º
Number of matches
5
7
10
14
15
20
22
23
25
25
Distance of result from expected
value ()
3.27
2.66
2.08
1.59
1.50
1.09
0.95
0.88
0.74
0.74
Similar tables and graphs have been produced for the azimuths of the other Greek poleis with
three temples or more.
0
0.5
1
1.5
2
2.5
3
3.5
1
2
3
4
6
7
9
10
Significance σ
Accuracy (º)
7
Analysis of Greek temple alignments by location
Gela
Presumed Accuracy
10º
Number of matches
3
3
3
4
4
4
4
5
5
5
Distance of result from expected
value ()
1.60
1.60
1.60
1.37
1.37
1.37
1.37
1.20
1.20
1.20
Himera
Presumed Accuracy
10º
Number of matches
2
2
2
2
3
6
6
6
8
9
Distance of result from expected
value ()
2.61
2.61
2.61
2.61
2.10
1.58
1.58
1.58
1.19
1.10
0
0.5
1
1.5
2
1
3
4
6
7
8
9
10
Significance σ
Accuracy (º)
0
0.5
1
1.5
2
2.5
3
2
3
5
6
7
8
10
Significance σ
Accuracy ± º
8
Megara Hyblaea
Presumed Accuracy
10º
Number of matches
13
13
17
20
23
23
23
24
24
24
Distance of result from expected
value ()
1.22
1.22
0.96
0.81
0.68
0.68
0.68
0.63
0.63
0.63
Naxos
Presumed Accuracy
10º
Number of matches
2
3
4
5
5
9
9
9
9
11
Distance of result from expected
value ()
2.00
1.61
1.38
1.21
1.21
0.85
0.85
0.85
0.85
0.75
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1
2
4
5
7
8
10
Significance σ
Accuracy ± º
0
0.5
1
1.5
2
2.5
1
3
4
6
7
8
10
Axis Title
Accuracy ± º
9
Selinous
Presumed Accuracy
10º
Number of matches
6
7
9
14
16
18
19
19
22
22
Distance of result from expected
value ()
2.82
2.56
2.17
1.54
1.36
1.20
1.12
1.12
0.91
0.91
Syracuse
Presumed Accuracy
10º
Number of matches
3
10
10
12
14
15
16
17
18
20
Distance of result from expected
value ()
1.54
0.76
0.76
0.67
0.59
0.56
0.53
0.50
0.48
0.43
0
0.5
1
1.5
2
2.5
3
1
2
3
5
6
7
8
10
Significance σ
Accuracy ± º
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
1
2
3
4
7
8
9
10
Significance σ
Accuracy ± º
10
Some common features occur. The most obvious is that the quoted significance is positive in all
cases. It should be emphasised that any figure below 2 cannot be said to be significant. The
positive values are a result of the target alignments being known parts of the total data set.
Another factor is that significance reduces as presumed accuracy lessens. This is also not
surprising. This would be a reiteration of the widely accepted fact that the more certain you can be
about an alignment the more likely it is you can say something about it. It is also for this reason
that the lack of significance in any of the small samples, Gela, Himera and Naxos should expected.
The sample sizes are simply not large enough to say anything statistically significant.
However there are striking differences between some samples. Some locations would seem to
have distinctive patterns of alignments. Selinous would seem to have a distinctive pattern of
alignment, if you accept that the Greeks planned to one degree of accuracy. Selinous is a site where
this claim is particularly plausible as many of the temples are aligned with the street grid. This
alignment is coherent over a distance of more than a than a kilometre. The early date of Temple E
on site in the late seventh century BC (Gullini, 1985) would suggest this gridding was conceived by
this period at the latest. The lack of any significant pattern at Megara Hyblaea would also seem to
suggest that this is a genuinely distinctive alignment. However, the adherence to the grid plan
could be consistent with a topographical rather than astronomical motivation. Himera too has
what might be a significant preference for some alignments, but at this site too the temples are
aligned to within a degree of the street grid, which is aligned to 67º.
It is the alignment of temples at Akragas that most strongly suggests that astronomy may have
been a factor in the orientation of Greek temples. As with other sites there is a general preference
for east. This need not be due to observational astronomy. Greek religion practice did not occur in
the temples but rather at the altar, which was almost always in front of the temple (Yavis, 1949, p.
56). Therefore religious ritual would be performed in the open. There are therefore good
mundane reasons why east would be the favoured side. Greek ritual was performed early in the
morning. An altar in the west could be in the temple’s shadow before sunrise. Not only would light
arrive, but so too would the warmth. This would not just warm the people but the soil where they
stood. An easterly orientation would also aid the evaporation of morning dew from the interior of
the temple and Vitruvius (4.1) gives this reason for orientating libraries to the east. Looking at the
locations of the temples at Akragas would suggest that this was not a primary factor in the
orientation of the temples.
11
The temples of Akragas are of a similar or later date to those found at Selinous (De Miro, 1994;
Marconi, 1929) and, broadly speaking, can be arranged into three groups. One cluster sits on top
of the Acropolis overlooking the city to the south. The temples here, Athena Lindioi, Zeus
Atabyrion and a temple of Demeter, that was not part of the survey all, face what was probably a
distant horizon to the southeast. The buildings of modern Agrigento make it hard to be certain. A
second cluster set of temples sit on the ridge that marks the south of the city and these seem to be
arranged to face distant horizons rather than the next easterly temple along the ridge. The final
group is the temple between these two regions of high ground.
The temple of Zeus Olympios stands next to the ridge and its alignment is skewed to the north to
avoid facing the ridge. Further to the west lie the temples in the sanctuary of the chthonic deities.
These all look up the valley to the distant horizon. Motivations regarding heat and light would be
little affected by local topography. A rise of 4º would bring a horizon closer, but would only delay
the arrival of the sun by around a quarter of an hour. A sun rising over a near horizon is just as
warm as a sun over a far horizon. What is altered by a close horizon is the apparent position of the
sun. The sun may not appear at the same time for all people present at an event. A distant horizon
would mean that all people could share a collective observing experience. If the arrival of the Sun
is part of a ritual then this removes a source of friction over the appropriate time to perform an
action.
The results from Megara Hyblaea may seem to refute this argument. The results from Megara
Hyblaea are only 1.22 away at most in comparison to 3.27 at Akragas, but Megara Hyblaea has a
very different topography. It lies on the east coast of Sicily and from the city all that lies to the east
is sea. Therefore the citizens had no topographical restrictions if they wished to point a temple at a
distant sunrise.
It is possible that the orientations could be to star-rises rather than the Sun. As almost all the
temples face a span of orientations within the range of the rising Sun, it is less problematic to
propose that it was a solar event that was being observed.
Genealogies of colonies
Trümpy (1997) has extensively researched the relationships between the calendars of the Greek
cities. They were usually named after a major festival in the month and show similarities between
related cities. The relationship between cities tracing their origins back to Corinth has led Freeth
12
et al. (2008) to suggest that the Antikythera Mechanism is possibly a Syracusan device based on
the similarity of the parapegma, or calendar, on the device to that of Tauromenion, which was a
daughter city of Syracuse. Religion in ancient Greece was less about a personal relationship with a
God, and more about the act of being a citizen. Citizenship was usually required to take part in the
religious events of a city and part of the duties in helping the community maintain an amicable
relationship with the gods. Therefore there is a sound reason why calendars related to religious
events should be also guides to political allegiances. The citizens of a colony would have argued
their ancestors where originally citizens of the mother city and so a similarity of the mechanics of
appeasing the gods can be expected. However, this requires textual evidence. Most calendars from
Greek cities are fragmentary if they exist at all. Is there a pattern in the alignment of Greek temples
that would also connect genealogies of cities?
For Sicily there are three prominent genealogies in the dataset. The easiest genealogy to identify is
Megara Hyblaea. The city name comes in part from Megara in the homeland. It is said that the
inhabitants of Megara Hyblaea sent for an oikist, a founder, for their new city in the west of Sicily,
Selinous (Thucydides 6.4.4). These two are analysed below as the Megarian colonies. Another
early settlement in Sicily was Syracuse. Syracuse was said to have been founded from Corinth
(Thucydides 6.3.2). In turn Syracuse founded Akrai, Helorus, Casmenae and Camarina. There are
no obvious temples at Casmenae, which is why it doesn’t appear in the data set. Syracuse, Akrai
and Helorus are analysed below as the Corinthian colonies. Camarina is excluded from this set, as
the temple would be better examined as a Rhodian construction for the reason below.
In the traditional histories Rhodes was comparatively late arrival to Sicily, founding Gela in 688 BC
with the aid of Cretans. Gela is said to have founded Akragas around 580 BC (Thucydides 6.4.4). It
also settled Camarina twice after the Syracusans had abandoned it (Pelagatti, 1966). The temple at
Camarina dates from the mid-fifth century after the Geloan settlement. While the city grid and
early fortifications may be Syracusan in origin, the temple would appear to be a product of Geloan
settlers and so the temples of Akragas, Gela and Camarina are collected as the Rhodian lineage
below.
13
The Corinthian temples
Presumed Accuracy
10º
Number of matches
5
12
12
14
17
19
20
21
23
26
Distance of result from expected
value ()
1.94
1.10
1.10
0.98
0.83
0.74
0.70
0.66
0.59
0.48
The Megarian temples
Presumed Accuracy
10º
Number of matches
13
13
16
17
20
20
22
22
22
22
Distance of result from expected
value ()
2.44
2.44
1.88
1.70
1.19
1.19
0.80
0.80
0.80
0.80
0
0.5
1
1.5
2
2.5
1
2
4
6
7
8
9
10
Significance σ
Accuracy ± º
0
0.5
1
1.5
2
2.5
3
1
2
3
5
6
8
9
10
Significance σ
Accuracy ± º
14
The Rhodian temples
Presumed Accuracy
10º
Number of matches
6
8
9
12
13
18
18
29
21
21
Distance of result from expected
value ()
3.89
3.21
2.95
2.34
2.18
1.51
1.51
1.28
1.17
1.17
The above results are somewhat mixed. There is clearly not a statistical case for common
Corinthian alignments in Sicily. The results are not significant, no matter what degree of accuracy
you argue for in alignment. Superficially the Megarian colonies would appear to be distinctive. If
an intention of high-accuracy is assumed, and the grid at Selinous would suggest it could, then
there seem to be distinctively Megarian alignments. However, this is almost certainly the result of
Selinous having a distinctive pattern of alignment with so many temples facing 96º. The
orientations at Selinous were 2.82 away from the expected result, and when Megara Hyblaea’s
temples are included this value drops.
In contrast the Rhodian colonies do seem to be distinctive if intentionally high-accuracy in their
orientation can be assumed. Not only are they distinctive, but also collectively this value is higher
than their individual analyses. This would suggest they are less distinctive when examine at a polis
level as they are then compared with sister cities with which they share deliberate architectural
similarities. The result is not so emphatic that the relationship is unquestionably real, but it is high
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
6
7
8
9
10
Significance σ
Accuracy ± º
15
enough to suggest that further investigation is required. One way of tackling this would be to
examine the temples when grouped by their patron deity.
Analysis of temples by patron deity
Herodotus argued that the Greek were unified by a common tongue, blood, religion and customs.
It may be hard to strongly argue for a common religion. There was certainly a generic Greek
religion, which local cults tended to be based upon. There were some Panhellenic sanctuaries,
most notably at Delphi and Olympia. However because of the relationship between religious
activity and political belonging in a polis, the gods worshipped in a polis were usually local forms
of gods. These could be subtly different or grossly different. Zeus Olympios at Akragas was a
celestial god. Zeus Melikhios at Selinous was probably a chthonic god, as in the form Melikhios was
associated with Demeter and her quest to recover Persephone from the underworld. Therefore
simply analysing temples by god may not be a sensible idea. The data exists to analyse temples to
Aphrodite, Zeus, Demeter and Athena. However, even if an apparently significant pattern were to
be found, the reader could question whether or not the temples are a culturally meaningful
assemblage. For example are temples to Aphrodite Ourania in her heavenly aspect really
comparable to temples of Aphrodite Pandemos? Further the lack of information about these
temples means that in some cases we cannot be sure which aspects of the god are being
celebrated. This problem is particularly striking in the case of the temples of Athena, which
provide the best historical evidence. The temples of Athena are so puzzling they will be examined
last as they require discussion at some length.
16
Aphrodite
Presumed Accuracy
10º
Number of matches
4
4
6
8
9
9
10
10
11
12
Distance of result from expected
value ()
1.79
1.79
1.42
1.19
1.10
1.10
1.03
1.03
0.96
0.90
Zeus
Presumed Accuracy
10º
Number of matches
2
3
5
8
8
9
17
20
22
25
Distance of result from expected
value ()
1.99
1.60
1.20
0.91
0.91
0.84
0.52
0.44
0.40
0.34
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
6
7
8
9
10
Significance σ
Accuracy ± º
0
0.5
1
1.5
2
2.5
1
2
4
5
6
7
8
10
Significance σ
Accuracy ± º
17
Demeter
Presumed Accuracy
10º
Number of matches
4
4
10
15
17
17
21
23
24
25
Distance of result from expected
value ()
1.37
1.37
0.78
0.58
0.52
0.52
0.42
0.38
0.36
0.34
The temples at Herakleia Minoa complicate analysis of the Aphrodite temples. There are two
temple-like buildings there. Diodorus Siculus tells us one is a temple of Aphrodite and the other is
the tomb of Minos, who was said to have been murdered there while pursuing Daedalus.
Unfortunately it is not possible to tell which is which. Given the preference for solar alignments it
would seem reasonable to assume it is the most northerly orientated of the temples that is to the
goddess, and that the second temple is a mistakenly identified tomb. This is not certain though. I
have counted the northern temple as the temple of Aphrodite as it produces the less impressive
results. The sample sizes for the three gods are small, only three or four temples and the aspects of
the gods uncertain. Their alignments might have an astronomical significance, but the lack of data
means this could only be demonstrated with substantial cultural evidence.
The temples of Athena
There are far more temples of Athena to analyse. Three in Gela, and another one in each of
Akragas, Camarina, Himera and Syracuse.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1
2
3
4
5
6
7
8
9
10
Significance σ
Accuracy ± º
18
Presumed Accuracy
10º
Number of matches
5
8
9
18
21
23
24
24
28
31
Distance of result from expected
value ()
2.63
1.97
1.82
1.03
0.86
0.76
0.71
0.71
0.51
0.34
The temples of Athena might have a distinctive orientation if a high-accuracy is assumed.
However, the results above would seem to be less impressive than the results for Rhodian
colonies. As so many of the temples in this sample are Rhodian a better explanation would be one
that combined both the Rhodian and religious elements of the temples.
As mentioned above deities are not uniform and have local aspects. In the case of Akragas, the
temple of Athena is to Athena Lindioi, which sits overlooking the city. At the two other Rhodian
cities, the temples of Athena sit on the high ground overlooking the cities. Examining solely the
Rhodian Athena temples gives the result below.
0
0.5
1
1.5
2
2.5
3
6
7
8
9
10
Significance σ
Accuracy ± º
19
Presumed Accuracy
10º
Number of matches
2
2
2
3
3
3
3
4
4
4
Distance of result from expected
value ()
3.22
3.22
3.22
2.59
2.59
2.59
2.59
2.21
2.21
2.21
The figures are not as unusual as for the Rhodian lineage results at high accuracy, but they are
consistently above 2 right across to 10º. Temples of Athena at Rhodian colonies do seem to be
pointing in a direction where few other temples face. Ideally I would now reveal that the temple of
Athena at Lindos, Rhodes, also points to around 117º. Sadly a survey by the Danish Archaeological
Institute has the Hellenistic temple pointing to 32º (Dyggve, 1960, p. 135) or 37.5º or 216º
(Liritzis & Vassiliou, 2003, p. 95), which is too far to the north to face the rising sun.
Another objection which can be raised is that this result could be due to the three iterations of the
Temple of Athena on the acropolis of Gela. It can be argued that the shape of the Acropolis is partly
responsible for the orientation of the Geloan temples, and rather than five temples, the sample size
should be three. The temple at Camarina would be evidence against such restrictions. The temple
of Athena currently stands slightly short of the peak of the hill overlooking the ancient town. The
temple is more or less at the top of the hill, so the rise in slope around the temple is slight, and may
be due to geomorphological processes since antiquity rather than fully indicative ancient
topography, but it does show that there was likely to have been a freedom of orientation at Gela,
and the reason the successive temples face a similar direction is that, for whatever reason, that
direction was the to place they wanted their temple to face.
0
0.5
1
1.5
2
2.5
3
3.5
1
2
3
4
5
6
7
8
9
10
Significance σ
Accuracy ± º
20
Not only do the Rhodian temples form a coherent cultural unit, but there is are good grounds for
examining the non-Rhodian temples as part of a different cultural unit environment. It can be
shown that the temples of Athena at Himera and Syracuse were built in a very different historical
context to the Rhodian temples at the Himera and Syracuse constructions can be linked to an
historical event. Both temples were built after the Battle of Himera in 480BC.
The post-Himera temples
480BC marks a watershed in the development of a Greek identity. In was in this year that Xerxes
invaded the Greek mainland. The Greeks collectively fought against a barbarian (non-Greek)
enemy for their independence. Fighting alongside the other cities cemented a notion of
brotherhood. This raises difficulties for a Sicilian Greek identity, as the major cities of Greek Sicily
were not fighting the Persians. In 480BC they were fighting the Carthaginians in Sicily. Victory
over the Carthaginians was won at Himera, though the date is uncertain. Herodotus says the
victory occurred the same day as the defeat of the Persian navy at Salamis (Herod. 7.166).
Diodorus Siculus puts it at the same time as the battle of Thermopylae (Diod. 11.24.1). These two
events were around a month apart. Feeney (2007, pp. 43-46, 51) has argued that the confusion has
a narrative purpose. History in this period was not so much a matter of chronology, but a tale of
how things came to be. Time has a narrative function, and by fighting the Carthaginians at the
same time as the mainland Greeks were fighting the Persians, the Sicilian Greeks could show how
they too were defending Greek homes from the barbarian. Following the battle of Himera Gelon,
the ruler of Syracuse, dedicated a temple to Athena in her form of Athena Nike, goddess of Victory.
He returned to Syracuse and on his arrival dedicated a temple to Athena there too.
So far, above, the idea, which has been tested, is that astronomy was primarily used in the practice
of Greek religion at temple sites. The preference for distant horizons would suggest an interest in
observations yet, apart from the temples of Athena at the Rhodian cities of Sicily, there is no
strong evidence for astronomical practice at temples. Another suggestion is that astronomy was
used in the construction of temples. If there was a general interest in an easterly orientation, it
may be enough that the temple faced sunrise on the day that the plan was laid out. At most sites
this would be impossible to test. The dating of most temples in Sicily is only to within a decade or
two on stylistic grounds. However, the start of temple construction following the battle of Himera
can be dated to the year.
21
In the case of the temple of Athena Nike at Himera the orientation is to 71º azimuth (Aveni &
Romano, 2000, p. S54)(Marconi, 1931) and a declination of 15º. This means the temple would face
sunrise on the 14th of August1. It is not certain if this would be plausibly close to the date of the
battle of Thermopylae. Labarbe (1959) dates the battle to as early as the 31st of July. Dascalakis
(1962, pp. 140-169) and Hignett (1963, pp. 448-449) both date the event to the 29th of August.
Sacks (1976) rejects this date in favour of the 19th of September. This is due to an eclipse on the
2nd of October, which would suggest he is using a Julian calendar. The proposed eclipse is not
strong evidence for the date of the battle as it was only visible as a partial eclipse from Greece, and
is unlikely to have noticeably dimmed the sky (Mosshammer, 1981). The lack of concordance
means that it would be futile to attempt to date the temple with respect to the Greek mainland, but
it idea that it faces sunrise on the day its foundation could still be tested against the temple at
Syracuse.
The Athenaion at Syracuse faces 92º and, if the horizon is flat, a declination of around -2º. It is
hard to be certain as the ruins lie beneath the modern city (Orsi, 1919). This would be consistent
with a date of around the 28th of September, which, in 480BC was also the day of a new moon and
so possibly the start of a new Syracusan month. This date is be consistent with Gelon leaving
Himera and arriving back home and founding a new temple. The return of a victorious army from
Himera around 250km by foot through central Sicily, or a by longer but almost certainly faster trip
by ship around the coast, would be feasible within this timescale. This may be plausible, but not
convincing. The alignment of the temple of Himera to the north of east made a more southerly
orientation quite likely anyway.
There is another post-Himera temple we can examine. The temple of Olympian Zeus was built by
Theron of Akragas following victory at the battle (Diod. 11.26.2, 13.82.1-4). Akragas was closer to
Himera, 110km through central Sicily. If Theron dedicated his temple on his arrival, then it should
have an alignment between that of the temple of Nike at Himera and the Athenaion at Syracuse.
The temple of Olympian Zeus faces an azimuth of 80º and a declination of +10º, which is sunrise
around the 29th of August in the Gregorian calendar, and also the day of a New Moon in 480 BC.
This allows around ten days to cross the terrain of central Sicily, if Theron’s army was still at
Himera when Gelon was dedicating his temple.
1 This is a Gregorian date. The date in the Julian calendar is around 19th of August, which some
planetarium software uses for BC dates.
22
The argument above cannot be used to date the battle of Himera, or other battles purely by
astronomy. It does however give some circumstantial evidence in favour of the earlier dates and
also is consistent with the alignment of temples towards the rising Sun when they are being built.
This act of using astronomy in laying out a temple may also be consistent with the preference for
distant horizons.
A possible use of astronomy at Greek temples
If you use some simple geometry then laying out a rectangle, which would, form the base for a
Greek temple is a trivial matter. One person could lay out an arbitrary baseline from which the
rest of the gridding for a building is derived. Laying out a temple is not merely an architectural
problem; it is also a religious matter. The value of an offering to the gods would be increased by
the effort expended in making it. The building of a temple would also be a communal matter. A
method, which allowed more people to work on a project, might be inefficient from an engineering
point of view, but part of the dedication of the temple would be the process of making the temple.
A base line could be derived from laying out a line using ranging rods, rope and shadow from the
Sun. If the horizon is distant then the area being worked upon is lit when the shadows are longest,
and multiple work groups could lay out parallel lines by working at the same time. Building a
temple is undoubtedly a religious event, which would explain why the alignment was to the east at
the morning, rather than away from the west as the sun set. Aligning the temple as the sun set
would be the more elegant solution from an engineering point of view. It would allow the shadows
to be tracked as the sun fell in the sky, but it would be at the wrong time. If the alignment was a
matter of laying out parallel lines for a grid to mark out the temple plan, then the fact that the
alignment may not have exactly hit the sunrise point may be of no great concern. Such an
approach to alignment would also explain why Selinous and Himera have topographically
orientated grids; there was a better or more convenient local marker for alignment to ‘east’ than
the Sun.
This does leave the alignment of the Rhodian Athena temples as a puzzle. It is possible that these
temples were dedicated on a specific day of the year related to the cult of Athena Lindioi. In this
case the shared alignments would be indicative of a shared religious cult. Comparative data is
clearly needed but as the alignments of the post-Himera Athena temples show, the comparisons
would have to be Athena in the same aspect as her cult in the Rhodian cities. It may be that the
23
cultural evidence enabling the identification of similar cults elsewhere in the Mediterranean no
longer exists.
Conclusions
As argued in Salt (2009), the Greek temples of Sicily do seem to be orientated with respect to the
Sun, but the precise relationship between temple alignment and the Sun remains uncertain. The
alignments of temples with respect to the city grids of Selinous and Himera would indicate that
high-accuracy alignment was practiced by the Greeks when building temples. The topographical
orientations of the temples of Akragas also supports the idea of an interest in observation of
distant horizons at these sites, and the results from the other Greek sites are consistent with this.
However, only at the poleis tracing their ancestry back to Rhodes is there any hint of inherited
practice in the alignment of temples. Paradoxically this may mean that the survey shows both that
the alignment of temples towards the sun was emphatically not the result of chance and that it
was of little importance to Greek religious practice. One way out of this problem is to argue that
astronomy was an important part of the process of building the temple and it is this process that
leaves its mark in the alignment of temples. The alignments of the post-Himera temples are
consistent, if not conclusive, with this proposal.
Greek temples are an attractive feature to examine for archaeoastronomy. They often have an
unambiguous axis of alignment and frequently have an attribution allowing for the development
of detailed hypotheses. There will doubtless be further studies of such temples. However, there is
room for a reassessment of where religious activity took place in relation to the temple. The altars
are almost always east of the temple, which would suggest that religious events took place on the
sunny side of the temple. This may be true of Greek temples in regions where easterly orientations
are less emphatic and it should be possible to re-examine data to see if practice occurred on the
sunny side elsewhere.
The term ‘sunny side’ is intentionally vague, and a stronger definition will require a greater
discussion of the practice of Greek astronomy. The method used above, based on the Binomial
distribution can be adapted for this by adjusting the value of p in light of cultural evidence. Indeed
the method shows that cultural information can massively transform the confidence we can have
in our interpretation of results. The Binomial distribution also allows for more flexibility in
debating the meaning of proposed alignments. While I believe that alignment to within a degree is
possible, the discussion is far from closed. By presenting the results in a way that allows many
24
variations in terms of both presumed accuracy and what a significant result would be the reader is
not compelled to accept all my proposals to make use of my findings.
Bibliography
Aveni, A., & Romano, G. (2000). Temple Orientations in Magna Graecia and Scily.
Archaeoastronomy: Supplement to the Journal for the History of Astronomy , 25 (Supplement to Vol.
31), S51-S57.
Bonacasa, N. (1970). L'Area Sacra. In A. Adriani, N. Bonacasa, A. Di Stefano, E. Joly, M. Manni
Pirano, G. Schmeidt, et al., Himera I, Campagne di scavo 1963-1965 (pp. 51-253). Rome: L’Erma di
Bretschneider.
Dascalakis, A. (1962). Problèmes historiques autour de la bataille des Thermopyles. Paris: De
Boccard.
De Miro, E. (1994). La Valle dei Templi. Palermo, Italy: Sellerio editore.
Dyggve, E. (1960). Lindos: Fouilles de l'Acropole 1902-1914 et 1952 (III,I Le Sanctuaire d'Athana
Lindia rt l'Architecture Lindienne). Berlin: Walter de Gruyter & Cie.
Feeney, D. (2007). Caesar's Calendar: Ancient Time and the Beginnings of History (Sather Classical
Lectures). Berkeley, California, USA: University of California Press.
Freeth, T., Jones, A., Steele, J., & Bitsakis, Y. (2008). Calendars with Olympiad display and eclipse
prediction on the Antikythera Mechanism. Nature, 454 (7204), 614-617 DOI:
10.1038/nature07130
Gullini, G. (1985). L'Architettura. In G. Pugliese Carratelli, Sikanie: Storia e civiltà della Sicilia greca
(pp. 415-491). Milan: Garzanti.
Henriksson, G., & Blomberg, M. (2000). New Arguments for the Minoan Origin of the Stellar
Positions in Aratos' Phaenomena. In C. Esteban, & J. Belmonte, Astronomy and Cultural Diversity:
Oxford VI and SEAC 99 (pp. 31-34). Tenerife: Museo del la Ciencia y el Cosmos.
Hignett, C. (1963). Xerxes' Invasion of Greece. Oxford: Clarendon Press.
Labarbe, J. (1959). Un Témoignage capital de Polyen sur la bataille de Thermopyles. Bulletin de
Correspondence Hellénique , 78, 1-22.
Liritzis, I., & Vassiliou, H. (2003). Archaeoastronomical Orientation of Seven Significant Ancient
Hellenic Temples. Archaeoastronomy: the journal of Astronomy in Culture , 17, 94-100.
25
Marconi, P. (1929). Agrigento. Topografia e arte. Florence, Italy: Vallecchi Editore.
Marconi, P. (1931). Himera: Lo Scavo del Tempio della Vittoria e del Temenos. Rome: Società Magna
Grecia.
McCluskey, S. (2006). The Orientations of Medieval Churches: A Methodlogical case Study. In T.
Bostwick, & B. Bates, Viewing the Sky Through Past and Present Cultures: Selected Papers from the
Oxford VII International Conference on Archaeoastronomy (pp. 409-420). Phoenix: City of Phoenix
Parks and Recreation Department.
Mosshammer, A. (1981). Thales' Eclipse. ransactions of the American Philological Association , 111,
145-155.
Orlandini, P. (1968). Gela - Topografia dei santuari e Documenazione Archeologica dei Culti.
Rivista dell’Istituto Nazionale d’Archeologica e Storia dell’Arte , XV, 22-66.
Orsi, P. (1919). Gli scavi intorno all'Athenaion di Siracusa negli anni 1912-1917. Monumenti
Antichi , XXV, 353-762.
Pelagatti, P. (1966). Camarina (Ragusa). Scavi nell'area urbana. Bollettino d'Arte , 51, 95-96.
Sacks, K. (1976). Herodotus and the dating of the battle of Thermopylae. Classical Quarterly , 26
(2), 232-248.
Salt, A. (2009). The Astronomical Orientation of Ancient Greek Temples. PLoS ONE , 4 (11), e7903.
DOI: 10.1371/journal.pone.0007903
Schaefer, B. (2006a). Keynote Address: Case Studies of the Three Most Famous Claimed
Archaeoastronomical Alignments in North America. In T. Bostwick, & B. Bates, Viewing the Sky
Through Past and Present Cultures: Selected Papers from the Oxford VII International Conference on
Archaeoastronomy (pp. 27-56). Phoenix: City of Phoenix Parks and Recreation Department.
Trümpy, C. (1997). Untersuchungen zu den altgriechischen Monatsnamen und Monatsfolgen.
Heidelberg: Universitätsverlag C. Winter.
Yavis, C. (1949). Greek Altars: Origins and Typology. St. Louis, Missouri, USA: St. Louis University
Press.
... For many of them were typical religious constructions oriented on astronomically important directions or at least to cardinal directions. For example, this show dedicated to the study of astronomical regularities of spatial organization: of Egyptian temples and pyramids [1,2,3], of building complexes of ancient Egypt [4,5], of temples of Sicily [6], of Roman temples and cities [7,8], of ancient Alexandria [9]. ...
Article
Full-text available
The article presents research data using arhaeoastronomical methods of Levinsadovka settlement sacrificial complex in Northern Black Sea coast. In this work, new method of accounting terrain elevations using topographic maps was developed in arhaeoastronomical studies. Calculations of azimuths of the sun and moon rise and set in the astronomically significant moments of the year were made and astronomical regularities in the organization of the sacrificial complex identified. Related to the moon directions, as the most dedicated, identified. Revealed that the stone tools and fragments, used as sacrifices, mark the direction to the northern minor standstill moonrise. A similar situation was found on two archaeological sites: in Bezymennoye II settlement South sanctuary and in Pustynka settlement religious building. Drawing on ethnographic and folklore data, it is concluded about semiotic connection of stone tools, found in these religious constructions, with the "thunderstones" and meteorites.ut semiotic connection of stone tools, found in these religious constructions, with the "thunderstones" and meteorites.
... Efforts to explain the orientation of Greek temples can be dated back to the works of Nissen (1873Nissen ( , 1885Nissen ( , 1887 and Penrose (1892Penrose ( , 1893Penrose ( , 1897Penrose ( , 1899. Since then, many scholars (Dinsmoor (1939), Herbert (1984), Mickelson (1999), Aveni & Romano (2000), Liritzis & Vassiliou (2003, 2005, 2006, Boutsikas (2005Boutsikas ( , 2007aBoutsikas ( , 2007b, Pantazis, (2008), and Salt (2009Salt ( , 2010, among others) have approached the problem with direct onsite measurements of azimuths and by contributing statistical results and interpretations mainly intended to associate the orientation with the rising or setting, on the local horizon, of celestial bodies on the day of a deity festival or the foundation date, or on the dates of important historical events. These interpretations are mostly convincing, their consistency depending on the knowledge of rituals and their associated celebration dates. ...
Article
Full-text available
I have measured orientations of the diagonals of Greek temples using a larger sample than before (Ranieri 2010), amounting to 200 temples, including several Treasures and Megarons. In addition, for each temple, I have used the best plan among those obtainable from the archives of archaeological journals, archaeological reports, and books. This approach has allowed more accurate and reliable geometrical analyses. The results show 57 temples with an unequivocally cardinally oriented main axis. Among the remaining 143 temples, 103 (72% of 143) have their diagonal cardinally oriented, with a standard deviation of about +/- 2.5 degrees. For the other 40 temples (28% of 143), the cardinal orientation can be attributed to the diagonals of half of the rectangle of the temple. The results point to a generalized topographical practice of establishing, as the first activity prior to construction, a cardinally oriented line on the terrain.
Article
Full-text available
Despite its appearing to be a simple question to answer, there has been no consensus as to whether or not the alignments of ancient Greek temples reflect astronomical intentions. Here I present the results of a survey of archaic and classical Greek temples in Sicily and compare them with temples in Greece. Using a binomial test I show strong evidence that there is a preference for solar orientations. I then speculate that differences in alignment patterns between Sicily and Greece reflect differing pressures in the expression of ethnic identity.
Article
The original version of this paper was presented orally in the Herodotus Seminar at the University of California, Berkeley, offered in winter 1972, by Professor W. Kendrick Pritchett and benefited much by his comments. As I also acquired my reading knowledge of Herodotus in his undergraduati course of the previous year, it is perhaps especially fitting that I dedicate this article to him on the occasion of his formal retirement. It is hoped that Professor Pritchett wil enjoy continued years of good health and of interest in all things Greek.
Article
Previous research on the Antikythera Mechanism established a highly complex ancient Greek geared mechanism with front and back output dials. The upper back dial is a 19-year calendar, based on the Metonic cycle, arranged as a five-turn spiral. The lower back dial is a Saros eclipse-prediction dial, arranged as a four-turn spiral of 223 lunar months, with glyphs indicating eclipse predictions. Here we add surprising findings concerning these back dials. Though no month names on the Metonic calendar were previously known, we have now identified all 12 months, which are unexpectedly of Corinthian origin. The Corinthian colonies of northwestern Greece or Syracuse in Sicily are leading contenders-the latter suggesting a heritage going back to Archimedes. Calendars with excluded days to regulate month lengths, described in a first century bc source, have hitherto been dismissed as implausible. We demonstrate their existence in the Antikythera calendar, and in the process establish why the Metonic dial has five turns. The upper subsidiary dial is not a 76-year Callippic dial as previously thought, but follows the four-year cycle of the Olympiad and its associated Panhellenic Games. Newly identified index letters in each glyph on the Saros dial show that a previous reconstruction needs modification. We explore models for generating the unusual glyph distribution, and show how the eclipse times appear to be contradictory. We explain the four turns of the Saros dial in terms of the full moon cycle and the Exeligmos dial as indicating a necessary correction to the predicted eclipse times. The new results on the Metonic calendar, Olympiad dial and eclipse prediction link the cycles of human institutions with the celestial cycles embedded in the Mechanism's gearwork.