-
4 ARCHAEOASTRONOMY 20072008 by the University of Texas Press,
P.O. Box 7819, Austin, TX 78713-7819
Efrosyni Boutsikas is a Lecturer of Classical Archaeology at the
University of Kent and presently holds a Visiting Fellowship at the
University
of Leicester. She received a B.Sc. in Archaeological Science
from the University of Sheffield and an M.A. in Archaeology from
the University
of Leicester. Boutsikas completed her Ph.D. (University of
Leicester) on astronomy and ancient Greek cult in 2007. Between
2006 and 2008
she was an osteological and archaeological supervisor for the
University of Leicesters Archaeological Services (ULAS), while
between 2007
and 2008 she worked as a university teacher in Ancient History
and Archaeology (University of Leicester).
EFROSYNI BOUTSIKAS
Abstract
This paper revisits the generally accepted view that the normal
orientation of ancient Greek temples is toward the east through a
general analysis of 107 Greek temple orientations col-lected by the
author. The paper also attempts to establish whether there existed
a general principle that related to speciic astronomical
observations and could have determined the orientation of Greek
temples. The analysis applies archaeoastronomical methodology in
investigating orientation patterns of Greek temples from the
Geometric to the Hellenistic periods in Greece. These irst results
show that the Sun does not seem to have played as decisive a role
in the orientation of temples as currently thought. Instead, there
appears to be a much larger variation than accounted for at present
that cannot be simply explained by the concept of the predominance
of eastern orien-tations. It is concluded that all-encompassing
interpretations do not appear to apply in Greek religion and cult
practices and that the study of Greek cult needs to account for
local variations, traditions, and landscapes.
Resumen
Al analizar los alineamientos de 107 templos griegos la autora
del presente artculo somete a nuevo examen la idea, comnmente
acep-tada, de que los templos en Grecia Antigua se orientaban
normalmente hacia el Este. Este artculo tambin trata de veriicar si
existi algn principio general relacionado con las ob-servaciones
astronmicas especicas y si este principio pudo determinar la
orientacin de los templos griegos. El estudio de los patrones de
orientacin de los templos griegos construidos en Grecia entre el
periodo geomtrico hasta el periodo helenstico emplea la metodologa
arqueoastronmica. Los primeros resultados demuestran que el
movimiento del sol no pa-rece jugar el papel tan determinante en la
elabo-racin de las orientaciones de los templos como se ha pensado
hasta ahora. En cambio, parece que la variacin de orientaciones es
mucho ms grande de lo que se supona y ello no puede explicarse por
el hecho de que simplemente predominan las orientaciones hacia el
este. En conclusin, las interpretaciones que pretenden explicar la
totalidad de orientaciones, no se aplican a los estudios de la
religin griega y de las prcticas cultuales, por lo tanto, cualquier
estudio de los cultos griegos tiene que tomar en cuenta las
variaciones, tradiciones y paisajes locales.
Placing Greek Temples: An Archaeoastronomi-cal Study of the
Orientation of Ancient Greek Religious Structures
-
VOLUME XXI 20072008 5
In 1939 William Bell Dinsmoor published his study on the
principles behind ancient Greek temple orien-tations. His treatment
and conclusions in this paper brought together earlier research
that had been car-ried out by Francis C. Penrose in the 1890s and
by Heinrich Nissen published between 1869 and 1906. Dinsmoors
general conclusion on the orientation of Greek temples followed
that of his predecessors, who argued in favor of the predominant
eastern orienta-tion. He claimed that 73 percent of Greek temples
were oriented within 60 of due east (1939:115116), and therefore
the placing of Greek temples was dic-tated by the need to face the
rising or setting Sun. This result was derived from plotting
Nissens temple ori-entations (published in 1906) in a graph in an
attempt to examine the presence of trends. Eighty years since
Dinsmoors paper, the orientations of Greek temples have been
shoe-horned in such a way that the presence of a much broader
variation of orientationswhich is in fact the caseis commonly
overlooked in favor of the idea of the predominance of an eastern
orienta-tion, which remains a point of reference for modern
scholars (Beyer 1990; Mikalson 2005:20; Scully 1979:44, 151). Prior
to Dinsmoors publication Nis-sen and Penrose had argued that
temples were aligned to sunrise on the day of the gods major
festival (Nis-sen 1873:527528; Penrose 1893:380). The eastern
orientation of Greek temples was explained as the result of
Egyptian inluence (Nissen 1906:249).
The study presented here intends to offer a much needed
structured and rigorous approach through the discipline of
archaeoastronomy as prescribed by Aveni (2002), McCluskey (1982,
2004), and Ruggles (1984, 1999, 2000a, 2000b). These scholars have
pioneered methods of archaeoastronomical research, leading to new
directions with regard to the contribu-tion of archaeoastronomy to
the reconstruction of past societies and practices (Ghezzi and
Ruggles 2007), wherever possible in conjunction with ancient
writ-ten sources (McCluskey 2006; Vail and Aveni 2004). This paper
challenges for the irst time the argument that Greek temples had a
predominantly eastern ori-entation, raising as a result serious
doubts about the assumed role of the Sun in the orientation of many
Greek temples. The study presents a general analy-sis of the
orientation of 107 Greek temples from the
Greek mainland and the islands of the Aegean (Figure 1)
collected by the author and covering a time period from 900 to 200
B.C. (Table 1). The analysis that fol-lows tests the existing ideas
on the general orientation of Greek temples andthrough a
quantitative assess-ment of the distribution of the
orientationspresents new data in order to test current
understanding of the role and function of the orientation of Greek
temples. It demonstrates that Greek religious structures were
placed over a far wider range than can be simply explained by a
solar orientation.
Sample Description
The dataset of this study includes some of the most important
and representative sites of the periods during which they were
constructed and some of the earliest self-standing religious
structures found in Greece from around 900 B.C. (e.g., Apollo
Thermios, excluding the megara, the function of which has not been
irmly established to this date). The region cov-ered by this study
includes the area covered by the modern Greek state (Figure 1)
rather than the world of Hellenic city-states as a whole, which
extended from the western Mediterranean to the Black Sea. In the
selection procedure of temples to be surveyed, no deities or types
of sites have intentionally been given greater emphasis. This study
includes the vast majority of religious sites that could be
measured within the study area. All religious structures for which
permission was given and whose preservation was suficient have been
surveyed (including those of foreign deities).
The geographical area covered by the sample presented here
includes the Greek mainland and the Aegean islands of Aigina,
Delos, Kos, Naxos, Poros, Rhodes, Samos, and Tenos. The dataset
includes different types of sites, including temples located in
organically grown settlements that demonstrate the continuity of a
cult over several successive temples constructed in the same
location. Settlements that developed organically are important to
this study, as they allow the examination of patterns of
continu-ity and, more importantly, observations of changes in the
orientation between successive structures. In some cases as many as
four reconstructions of the same temple have been measured (e.g.,
the temples
-
6 ARCHAEOASTRONOMY
ID Location Site Building Azimuth Altitude Declination
1 Acheron Oracle of the dead Main sanctuary 4 3 53 212 Acheron
Oracle of the dead Palace of Hades & Persephone 4 0 50 13
Aegina Sanctuary of Aphaia Temple of Aphaia 67 1.5 18 354
Amphipolis Sanctuary of Attis Temple of Attis 101 4 -5 565
Amphipolis Thesmophorio Thesmophorio-Nymphaion 165 11 -36 286 Argos
Heraion Old Temple of Hera 118 3 -19 127 Argos Heraion New Temple
of Hera 119 3 -19 568 Athens Acropolis Parthenon 77 2 11 79 Athens
Acropolis Temple of Athena Polias 85 3.5 5 4810 Athens Acropolis
Erechtheion 353 3 54 1511 Athens Agora Metroon 102 4.5 -712 Athens
Agora Temple of Apollo Patroos 97 4.5 -2 5913 Athens Agora Temple
of Zeus & Athena Phatria 99 4.5 -4 3314 Athens Agora
Hephaisteion 104 5 -8 615 Athens South slope Old Temple of Dionysos
75 3 13 2116 Athens South slope New Temple of Dionysos 75 4 14 0017
Bassae Sanctuary of Apollo Temple of Apollo 4 14 62 118 Calydon
Ancient Calydon Temple of Apollo 129 1 -29 419 Calydon Ancient
Calydon Heroon 180 0.5 -51 3620 Calydon Ancient Calydon Temple of
Artemis 122 3 -22 3421 Corinth Agora Temple of Apollo 77 3 12 122
Delos Sanctuary of Apollo Letoon 186 1 -51 4023 Delos Sanctuary of
Apollo Artemisio 108 3 -12 3724 Delos Sanctuary of Apollo Temple G
347 2 52 1925 Delos Sanctuary of Apollo Poros Temple of Apollo 265
0.5 -4 1126 Delos Sanctuary of Apollo Temple of Apollo (Athenians)
263 0.5 -5 2327 Delos Sanctuary of Apollo Great Temple of Apollo
264 0.5 -4 5928 Delos Sanctuary of Apollo Dodekatheo 97 3.5 -3 3329
Delos Sanctuary of Foreign Gods Heraion 172 7 -45 830 Delos
Sanctuary of Foreign Gods Serapeion C 178 2 -50 5231 Delos
Sanctuary of Foreign Gods Temple of Isis 268 0 -1 4532 Delos
Sanctuary of Foreign Gods Serapeion A 297 2 22 2433 Delos Sanctuary
of Mount Kythnos Temple of Zeus 286 0 12 17 Hypsistos Mount
Kythnos34 Delos Sanctuary of Mount Kythnos Sanctuary of Artemis
Locheia, 85 0 3 37 Hercules-Baal Zeboul, gods of Askalon35 Delos
Sanctuary of Mount Kythnos Sanctuary of Agathe Tyche 266 0 -3 3236
Delos Theatre district Aphrodision 170 9 -42 4537 Delphi Sanctuary
of Apollo Old Temple of Athena Pronaia 177 7 -44 3438 Delphi
Sanctuary of Apollo Temple of Apollo 49 27 47 4939 Delphi Sanctuary
of Apollo Old Temple of Apollo 49 27 47 4940 Delphi Sanctuary of
Apollo Temple of Athena Pronaia 190 8 -42 4241 Dion Sanctuary of
Demeter Temple A 64 0 19 242 Dion Sanctuary of Demeter Temple 1 70
0 14 37
Table 1. List of the structures included in the dataset of this
study
-
VOLUME XXI 20072008 7
ID Location Site Building Azimuth Altitude Declination
43 Dion Sanctuary of Demeter Temple B 78 0 8 3744 Dion Sanctuary
of Demeter Temple 2 71 0 13 5245 Dion Sanctuary of Demeter Small
temple with offering table 61 0 21 1246 Dion Sanctuary of Egyptian
Gods Temple of Isis 162 1 -46 747 Dion Sanctuary of Egyptian Gods
Temple of Hypolympia Aphrodite 68 0 16 648 Dion Temple of Zeus
Temple of Zeus Hypsistos 150 1.5 -40 3249 Dodona Oracle of Zeus
Temple of Aphrodite 116 8 -14 1550 Dodona Oracle of Zeus Temple of
Themis 129 7 -23 5151 Dodona Oracle of Zeus Temple of Zeus (hiera
oikia) 125 7.5 -20 5052 Dodona Oracle of Zeus New Temple of Dione
110 8 -9 5653 Dodona Oracle of Zeus Old Temple of Dione 176 12 -38
2354 Dodona Oracle of Zeus Temple of Hercules 158 3.5 -42 3555
Eleusis Sanctuary of Demeter & Kore Megaron 111 2 -15 2756
Eleusis Sanctuary of Demeter & Kore Telestirio-Solonion 115 2
-18 2957 Eleusis Sanctuary of Demeter & Kore
Telestirio-Peisistratid 115 2 -18 2958 Eleusis Sanctuary of Demeter
& Kore Ploutoneion 103 2 -9 1859 Gortyn Asklepieion Temple of
Asklepios 108 20 -1 1260 Isthmia Sanctuary of Poseidon Old Temple
of Poseidon 98 0 -6 3561 Isthmia Sanctuary of Poseidon New Temple
of Poseidon 97 1 -5 462 Kos Asklepieion Large Temple of Asklepios
25 1 46 4763 Kos Asklepieion Prostyle Ionic Temple of 114 2 -18 13
Asklepios64 Lebadeia Temple of Zeus Temple of Zeus Vassileus 64 0
2065 Mantineia Agora Temple of Hera 93 8 2 3266 Mantineia Agora
Podareion 86 8 8 367 Megalopolis Agora Temple of Zeus Soter 101 4.5
-5 5868 Messene Asklepieion Temple of Asklepios 115 11 -12 1169
Messene Asklepieion Temple of Artemis 129 11 -21 5670 Messene
Asklepieion Artemision 115 11 -12 1171 Messene Asklepieion Oikos
Asklepeiou & Paidon 215 1 -40 3072 Naxos City Temple of Apollo
Portara 140 0 -38 673 Naxos Sanctuary of Dionysos Old Temple of
Dionysos 203 4 -43 4674 Naxos Sanctuary of Dionysos Temple of
Dionysos 202 4 -44 1175 Naxos Sagri Temple of Demeter 213 0 -42
3076 Nemea Sanctuary of Zeus Temple of Zeus 75 7 16 877 Nemea
Sanctuary of Zeus Old Temple of Zeus 75 7 16 878 Olympia Sanctuary
of Zeus Temple of Zeus 83 3 7 2879 Olympia Sanctuary of Zeus
Heraion 87 2 3 3980 Olympia Sanctuary of Zeus Pelopeion 208 3 -42
881 Pella Thesmophorio Thesmophorio 267 2 -1 182 Pella Thesmophorio
Thesmophorio 84 1 4 4783 Perachora Heraion Temple of Hera Akraia 93
12 5 484 Poros Sanctuary of Poseidon Temple of Poseidon 68 2 18
1385 Pylos Nestors Palace Hiero-Oplostasio 147 2 -37 1886 Pylos
Nestors Palace Queens Hall SW entrance 220 0 -35 27
-
8 ARCHAEOASTRONOMY
ID Location Site Building Azimuth Altitude Declination
87 Pylos Nestors Palace Megaron 147 3 -36 23 88 Rhodes City of
Rhodes Temple of Aphrodite 93 0 -3 589 Rhodes Ialyssos Temple of
Athena Polias & 184 0 -53 57 Zeus Polieos90 Rhodes Kameiros
Temple of Pythian Apollo 357 0.5 53 3691 Rhodes Lindos, Acropolis
Temple of Lindia Athena 34 0 41 2192 Samos Heraion Rhoecus Temple
79 0.5 8 4793 Samos Heraion Hekatombedon II 79 0.5 8 4794 Samos
Heraion Greater Temple of Hera 79 0.5 8 2395 Samos Heraion
Hekatombedon I 77 0.5 9 5796 Sikyon Acropolis & Agora Temple of
Artemis or Apollo 95 2 -2 4997 Sounio Sanctuary of Poseidon Temple
of Poseidon 105 1 -11 3798 Sounio Sanctuary of Poseidon Great
Temple of Athena 98 1 -6 799 Sounio Sanctuary of Poseidon Small
Temple of Athena 103 1 -10 3100 Sparta Sanctuary of Artemis Orthia
Temple of Artemis Orthia 100 4 -6 16101 Tegea Temple of Athena Alea
Temple of Athena 87 5 5 24102 Tenos Sanctuary of Poseidon Building
B 194 0 -50 47 & Amphitite103 Thermum Ancient Thermum Temple of
Apollo 191 5 -45 26104 Thermum Ancient Thermum Megaron A 194 5 -44
44105 Thermum Ancient Thermum Megaron B 196 4 -45 9106 Tiryns
Palace Temple of Hera 180 2 -50 42107 Tiryns Palace Megaron 180 2
-50 42
Table 1. (Cont.)
of Hera in Samos). Sites on coasts (e.g., Perachora), in plains
(e.g., Messene, Athens), and on hilltops or mountains (e.g., the
Menelaion near Sparta and the temple of Apollo at Bassae) are also
included in the dataset. The sample contains not only temples that
belonged to settlements of various sizes but also those with access
to a number of different resources: some have limited local trade
routes, while others were cosmopolitan trade centers and therefore
subject to a variety of cultural inluences. The study also includes
temple measurements from sanctuaries located out-side and on the
boundaries of urban centers (e.g., the Thesmophorion-Nymphaion in
Amphipolis) as well as temples independent of the control of a
certain city (e.g., the sanctuary of Apollo in Delphi). Wherever
possible, cities that were planned from the outset and followed
town-planning concepts and principles
before they were laid out have been included in the sample
(e.g., Rhodes).
Field Methodology
The measurements comprising this study were col-lected using a
magnetic compass and clinometer over four ield seasons. A compass,
duly corrected for magnetic declination, will only determine the
direc-tion relative to true north to an accuracy of around one
degree. Taking into account the highest level of astronomical
precision that the ancient Greeks would have been capable of
measuring, this level of accuracy is considered adequate. Local
magnetic anomalies were tested in two ways. Minor anomalies were
tested by several measurements taken along each of the long walls
of rectangular structures and from either end of the wall. Great
magnetic anomalies that could have
-
VOLUME XXI 20072008 9
FIGURE 1. Map of ancient Greece showing the sites included in
this study. 107 measurements of temple orientations were collected
from 42 sites. The map shows 40 sites. The two sites missing are
located in Athens. The point for Athens cov-ers, therefore, three
sites: the Acropolis, the south slope, and the Agora. Outline map
created by R. A. LaFleur and Tom Elliott. Copyright 20002001,
Ancient World Mapping Center, http://www.unc.edu/awmc.
affected a large geographical area were examined by studying the
geology of sites prior to their survey.
The structures of this study were all of rectangular shape. To
determine their orientation the magnetic bearing was recorded along
each of the long walls from either end. In those cases where only
half of the structure survived, the long and the short walls were
measured from either end. This repetition of mea-surements was
necessary to ensure the most accurate readings of the temples
orientation. In addition to measuring the magnetic orientation of
each structure, horizon proiles were also recorded for the horizon
surrounding each structure. The horizon proiles were measured using
a compass and a clinometer, and these measurements involved the
combination of the
magnetic orientation of each point and the altitude of the
horizon on that orientation. These measurements were repeated until
the entire horizon proile was recorded, and all measurements were
taken from the center of the temples entrance.
I have attempted to ensure that data collection was as inclusive
as possible. And although decisions had to be made about what sites
would be included, the decision to include temples was mostly
driven by factors of site preservation and accessibility during
ieldwork.Data Reduction
This study improves the methods of analysis applied to the
orientations compared with previous studies
-
10 ARCHAEOASTRONOMY
by accounting for the height of the local horizon (altitude),
refraction, and atmospheric extinction. The temple orientations
have been converted to dec-linations using the command-driven DOS
program GETDEC created by Clive Ruggles
(http://www.le.ac.uk/archaeology/rug/aa/progs/decpak.html), which
makes corrections for atmospheric refrac-tion and extinction (Table
1). In order to obtain the declination of a structure, GETDEC
requires the structures latitude, the magnetic orientation, the
horizons altitude, and the magnetic correction. This means that
each declination obtained is speciic to the particular horizon and
location. The term declination in this sense, used when discussing
the orientation of a structure, needs to be explained. Declination
is the angular distance between a celestial object and the
celestial equator, whether to the north or the south; it is the
celestial counterpart of terrestrial latitude. As a result, a
structure as such cannot have a declina-tion. This term is employed
throughout this paper in order to denote the exact part of the
celestial sphere toward which the structure is oriented and to
therefore be compared to the celestial objects with the same or
similar declination or celestial latitude. In the present context
the declination is more informative than the azimuth (bearing of
magnetic compass) of a structure. This is because by using
declination we instantly account for extinction, latitude, and the
alti-tude of the local horizon aligned with the structures
entrance. In addition, the use of declination enables a direct
comparison between the orientation of a struc-ture and the position
of a speciic celestial object, or a position on the horizon.
The declinations of horizon points indicate which celestial
bodies rise and set there and (once preces-sion, refraction,
atmospheric extinction, and proper motion are allowed for) which
ones would have risen or set there at any given era in the past.
Furthermore, by obtaining declinations for speciic points along a
horizon (horizon proiles), we can calculate the declination of any
point on the horizon proile and hence reconstruct the celestial
bodies visible at that particular horizon at different times. The
orientations were plotted in the form of cumulative frequency
dis-tribution (curvigram). Each declination shown in the following
graph is represented by a computed curve.
The peak of the curve is the deduced declination. The curve of
each declination is centered on the median of all the measurements
of the structures orientation and with a standard deviation
determined by combining the standard deviation of those
measurements with the uncertainty in the magnetic declination.
These curves allow us, therefore, to investigate the patterns of
emerging distribution, with the added advantage of avoiding the
display of a false accuracy (given the limited precision of the
instrument used) that a simple point in the place of the curve
would have offered.
The Orientation of Greek Temples
Graph 1 shows the distribution of the deduced temple
declinations. Three general groups of orientations are depicted in
the graph. The largest group of measure-ments points broadly east
and west, spanning the declination range -30 to +23 with distinct
borders at the northern and southern ends. The vast majority of
this group falls within the solar range (Graph 2, high-lighted
section). Within this group there is a particular concentration of
declinations between -8 and +8. This concentration, if interpreted
in terms of sunrise or sunset, represents a range of dates falling
roughly within one month of the equinoxes. If we were to argue that
the position of the rising or setting Sun on the horizon at the
time of the equinoxes was used as a factor in orienting some Greek
religious structures, we would expect that the distribution of such
a group of declinations would show an accumulation of data at the
time of the actual equinox (declination 0). As shown in Graph 1,
the dataset includes no structures oriented between 0 and 2, only
two structures have declination -1, and one structure declination
-2. This very distinct absence of data in the range of the Sun
rising at the actual equinox may signify that the con-centration of
data around the equinoxes, although empirically real, could be an
example of unintended astronomical alignment by those who
constructed it (Ruggles 2000b:152).
The declinations falling within the eastwest group comprise 65.3
percent of the total amount of data (70 measurements). Of the 70
measurements belonging to this group, eight face toward the west:
the Poros temple of Apollo, the temple of the Athenians and Great
temple of Apollo, the temples of Zeus Hypsistos
-
VOLUME XXI 20072008 11
and Agathe Tyche on Mount Kythnos, the temple of Isis and the
Serapeion A in the Sanctuary of Foreign Gods, all from the island
of Delos, and the west entrance of the Thesmophorion in Pella. This
result deduces that the eastern declinations are therefore 62,
comprising 58 percent of the total sample collected for this study.
This result conirms earlier indications that a large number of
Greek temples face toward the east. However, the eastern
orientations of this study comprise a considerably smaller part of
the total
data than earlier conclusions: Dinsmoor argued that 73 percent
of Greek temples were oriented within 60 of due east (1939:115116).
The present sample indicates that the eastern-facing temples are
not as predominant as previously thought, and in addition, the
distribution of the orientations shows a much greater variation
that cannot be ignored or explained by the movement of the Sun.
Graph 1 shows the presence of a second group of data formed
toward the southern part of the sky,
GRAPH 1. The distribution of the orientations of 107 Greek
temples from 900 to 200 B.C. The Y axis shows the temple count. The
graph includes adjustments for standard deviation. Southern
declinations are between -60 and -40 (to the left). Western and
eastern declinations overlap in the center, and northern ones are
between +40 and +70 (to the right).
GRAPH 2. Reproduction of the distribution of data, displaying
the range of declinations visited by the Sun during its annual
movement (-24 to +24) (highlighted section). In the highlighted
area both eastern and western declinations are included. This group
comprises 58 percent of the total sample facing toward eastern
declinations and 7.4 percent of the total sample facing toward
western declinations.
-
12 ARCHAEOASTRONOMY
ranging between declinations -55 and -34. This group comprises
25.2 percent of the total sample (total number of measurements 27).
As neither the Sun nor the Moon visit these declinations, if the
orientation of these structures was related to astro-nomical
observations, this could only involve stellar observations. The
constellations rising and setting in the declinations covered by
this group are Centaurus ( or for the Greeks), Lupus (for the
Greeks a wineskin from which the Centaur was about to drink, or the
Therion [], meaning wild animal), Ara (the Greeks called it
Thytrion or Thysiasterion [ or ], mean-ing altar), Vela (for the
ancient Greeks the sail of the constellation of Argo [Argo Navis]),
the southern part of Sagittarius (for the Greeks Toxeutes or
Toxotes [ or ], meaning archer), Phoe-nix (the Egyptian Bennu,
possibly named Phoenix by the Greeks), and the southern part of
Eridanus ( or in Greek, the latter meaning river). Although perhaps
unintended, a concentra-tion of data is observed between
declinations -42 to -46 of 13 structures. These structures do not
indicate a preference with regard to a speciic deity,
chrono-logical period, or geographic location. This subgroup
includes hero cults (Pelopeion in Olympia and the temple of
Herakles in Dodona), other chthonic cults like the two temples of
Athena Pronaia in Delphi and the temple of Demeter in Naxos,
temples constructed
over Mycenaean megara (Apollo in Thermon [three structures] and
Dionysos in Naxos), a temple dedi-cated to a foreign deity (temple
of Isis in Dion), as well as the Heraion (two temples) and the
Aphrodi-sion in Delos.
Finally, a small cluster of data is observed in the northern
declinations (+40 to +68) representing 8.4 percent of the total
sample (nine measurements). This cluster includes only cults of
Apollo and chthonic cults. The Apollo temples falling in this group
are those in Delphi, Bassae, in Kameiros, Rhodes, and temple in
Delos. Although these form a signiicant part of the surveyed
temples dedicated to Apollo, it should be noted that the remaining
surveyed temples of Apollo are oriented toward different parts of
the horizon (Graph 3). The temple of Apollo in Corinth and that of
Apollo Patroos in the Athenian Agora face the east; the temple of
Apollo Erethymios in The-ologos, in Rhodes, is oriented to the
northeast; that of Calydon is toward the southeast; and the temples
of Apollo in Naxos and in Thermon face south. The northern
orientation of the Delian temple , although of much earlier date,
can be contrasted to the other Apollo temples on the island, all of
which are oriented toward the west. It is possible that Apollo
being the only ouranic deity represented in this northern group
could be a deliberate choice, but further investigation is needed
in order to examine possible reasons behind such a choice. Such a
study would need to contain
GRAPH 3. Declinations of twelve temples dedicated to Apollo.
-
VOLUME XXI 20072008 13
an in-depth analysis of each Apollo cult, the material culture,
and the local horizon and landscape. A study of the Delphic temple
of Apollo has indicated that its orientation may have been
connected to stellar observations, and, more speciically, it seems
possible that the orientation of the temple, the operation of the
Delphic oracle, and the presence of Apollo in Delphi for a certain
number of months may have been related to the movement of the
constellation of Delphinus (Salt and Boutsikas 2005).
This group of northern orientations includes the following hero
cults: the Doric temple of Asklepios in Kos, the north porch of the
Erechtheion in Athens, and the axis of the oracle of the dead in
Acheron, which also includes the underground palace of Hades and
Persephone. Although the evolution of the cult of Asklepios from a
mortal physician to a Thessalian hero, to a chthonic oracular demon
to a Panhellenic Apollonian deity with mantic character is complex
(Compton 2002:320321), in Kos his cult developed to an important
state cult, retaining, however, its chthonic character. The temples
of Asklepios in Kos have different orientations, but they all face
the altar (from different directions). With regard to the
Erech-theion, a recent study of the structure and the north porch
indicates that the north porch and the west cella were of greater
cultic signiicance to the east porch and cella and that this
northern orientation may have been deliberate and associated with
the movement of the constellation of Draco (Boutsikas 2007).
As is apparent from Graphs 1 and 2 and Table 1, the dataset
presented here displays no preference toward the cardinal points.
The largest number of data accu-mulation toward a cardinal point is
that facing east, with, however, only ive structures facing within
3 of due east (just under 4.7 percent of the dataset). Three
structures of the examined sample face due south (within 3), two in
Tyrins and one in Calydon, and only one due north (Rhodes) and due
west (Pella). The analysis of the sample demonstrates a
distribution of orientations that is much wider than the range in
the horizon visited by the Sun (Graph 2). It is evident that the
movement of the Sun alone is not suficient to explain the
orientation of Greek temples.
In examining the possibility of lunar associations, Graph 4
shows that the lunar rising or setting points in
the horizon do not seem to have been associated with the
orientation of the temples either. The Moons path along the horizon
is similar to that of the Sun, but it moves a little farther north
and south (shaded darker in Graph 4). As such, it appears dificult
to determine whether the orientations of the eastwest group could
be associated with the movement of the Moon or that of the Sun.
However, if the former were the case, we would expect to ind
measurements falling also within the part of the horizon that is
only visited by the Moon: declinations -24 to -30 and +24 to +28
(extending on either side of the solar range). Graph 4 shows
explicitly that only one structure (the temple of Apollo in
Calydon) is oriented within the space between the end of the solar
range and the southern and northern major lunar limits.
The data have also been divided into chronological periods in
order to investigate whether a practice of deliberate general
orientation of Greek temples was introduced at a speciic period or
whether, if present, it declined after a certain time. In the vast
majority of re-ligious sites we encounter continuity in the
construc-tion of religious buildings; the destruction of temples
from natural disasters (e.g., the temple of Apollo at Delphi,
destroyed in 373 B.C. by an earthquake) or by human action (e.g.,
the destruction of the temple of Poseidon at Sounion by the
Persians) was followed by their replacement with new structures.
The new temples were built either adjacent to or on top of the old
foundations, always dedicated to the same deity. As ritual practice
changes on a slow timescale even in cases of rapid social change,
the chances of identify-ing trends are greater, as they may be
sustained long enough to be picked up by the archaeological record
(Ruggles 2000b:163).
The investigation of changes in orientation as a result of the
precession of the equinoxes between successive building phases
cannot be examined at this stage. In order to do so it is necessary
to determine the celestial body toward which the structure was
aligned, but such a conclusion needs to be determined through the
examination of archaeological and liter-ary evidence rather than by
using the orientation of subsequent structures in order to ix on a
celestial body that simply shares the orientation. In the case of
Greece there is no single celestial body that could
-
14 ARCHAEOASTRONOMY
have determined the orientation of all or the majority of
temples.
The declinations from this study were split into subgroups by
chronological period as determined by archaeological inds:
Geometric (900700 B.C.), Archaic (700480 B.C.), Classical (480330
B.C.), and Hellenistic (330 B.C.A.D. 14). The results of this
analysis produce graphs that in terms of their dis-tribution
patterns are similar to those of Graph 1. The two largest
chronological groups were for the Archaic and Classical periods
(Graphs 5 and 6, respectively). The distribution of the data from
the Classical pe-
riod (Graph 6) is representative of those generated for the
other periods also. As demonstrated also in these two
representative graphs, this analysis shows no visible shift between
the consecutive periods. The graphs generated by the division of
the data into the aforementioned chronological periods depict the
same three clusters of data that have been discussed previously
(eastwest, northsouth).
A preliminary study of the sites included in this study
indicates a frequent shift of orientation between earlier and later
structures. The dataset includes, among other cases, four sites
with four successive
GRAPH 4. Reproduction of the distribution of the dataset, with
the annual path of the Moon shaded darker (-30 to +28),
superimposed on the solar declination range.
GRAPH 5. Distribution of sample dating to the Archaic period
(700480 B.C.) (30 structures).
-
VOLUME XXI 20072008 15
reconstructions of the same temple (e.g., the Heraion of Samos
and the temples of Dionysos in Sagri, Naxos), six sites with three
successive reconstruc-tions (e.g., the temples of Apollo and
Artemis on Delos), and nineteen sites with two reconstructions
(e.g., the temples of Dionysos in Athens, the temples of Poseidon
in Isthmia, and the temples of Demeter in Dion). In a number of
cases two or more successive temples with different orientations
fall in the same chronological subgroup (e.g., the two temples of
Poseidon at Isthmia and the two temples of Asklepios in Kos). The
general scheme of chronological periods, as given above, rests on
identiied changes in tech-nology, the architectural development of
structures, and changes in pottery and art. It becomes apparent
that the boundaries of these periods are not directly applicable to
a study that investigates successive religious structures.
Graph 7 shows the changes in the temple orienta-tions grouped
according to successive structures. In the majority of the cases
(18 out of 28) there is an observed change in orientation between
successive temples. It is intriguing that in 17 cases out of 18 the
change in orientation occurs between the irst temple and the
second. Only in one case (the temple of Athena Pronaia in Delphi)
do the irst and second structures have the same orientation with a
change occurring in the third. The chronological division analysis
and that of the orientation of consecutive structures makes
apparent the need for examining sites with continuity in the
construction of religious structures individu-ally and within their
religious context, regardless of modern views about the time frame
of chronological periods.
The general distribution of temple orientations reveals clusters
of data that may or may not be de-liberately placed by the groups
who built them. For more conclusive arguments on either the
dismissal of the possibility that Greek temples were
astro-nomically oriented or, alternatively, in support of a case
for deliberate astronomical orientation, further investigation of
possible reasons and principles be-hind potential deliberate
placing of temples would be necessary. The following section
discusses such possibilities.
Discussion
Previous research by Dinsmoor, Penrose, and Nissen focused on
the signiicance of the Sun in the orienta-tion of Greek temples. To
this day this idea has been offered as the explanation for the
general principles behind the orientation of temples. In doing so,
how-ever, we overlook a very large body of data that falls outside
positions in the horizon that are visited by the Sun. Dinsmoors
ideas have persisted for years without any attempt at veriication
or testing by other researchers who have used his results. This
study forms the irst systematic collection and analysis of
GRAPH 6. Distribution of sample dating to the Classical period
(480330 B.C.) (27 structures).
-
16 ARCHAEOASTRONOMY
Greek temple orientations in more than a century. This study
takes a irst step toward a systematic ap-proach by focusing on a
geographically smaller area that has, however, been surveyed more
thoroughly than before. The present dataset does not include
temples from Asia Minor, Italy, and Sicily, as earlier researchers
attempted. I believe that these areas need to be surveyed just as
thoroughly and to be examined independently before we can attempt
to put forward an all-encompassing model and interpretation of
Greek temple orientations.
This paper provides hard evidence in order to dem-onstrate that
care should be taken when making gen-eral statements about the
direction of Greek temples, statements that unavoidably bear weight
in what we perceive as determining factors for this orientation.
The data presented here suggest that the Sun alone was not the
all-encompassing phenomenon deter-mining the placement of the vast
majority of Greek religious structures. In fact, this appears to be
a gross oversimpliication of a much more complex and more
interesting pattern of temple orientation and religious practice.
The general analysis shows that 58 percent of the temple
orientations falls within the points on the horizon that the rising
Sun visits in a year and 7.3
percent within the points of the setting Sun. A total of 34.7
percent of the sample falls outside the solar range. This also
indicates that we need to explore other ideas about temple
orientation and that Panhel-lenic trends appear unlikely to explain
this pattern. Had the Sun been the predominant factor determining
orientation, we would expect temples to be oriented within the
solar range alone or at the very least to ind only a few exceptions
to this rule. The absence of measurements between the solar range
limits and the major lunar limits (shaded darker in Graph 4), with
the exception of one measurement, does not support a lunar
explanation either. The Moon revisits positions in the horizon
monthly. Exceptions to this are those declinations close to the
major lunar limits that are visited annually (shaded darker in
Graph 4). The prob-lems of using the Moon as a marker have been
noted by ancient writers (Aristophanes Clouds 615626) and by modern
researchers (Hannah 2005a:4750; Ruggles 1999:6063), as has the
incompatibility of the calendars of the different Greek city-states
(Hannah 2005a:48; Thucydides 5.19.1), and no fur-ther discussion is
necessary here.
The general analysis presented here is understand-ably limited:
it can enlighten insofar as it indicates
GRAPH 7. Changes in orientation between successive structures.
Orientation measurements from 29 cults (of 72 successive
structures).
-
VOLUME XXI 20072008 17
patterns in the material record but tends to ignore the rich
variety and diversity of symbolism that was almost certainly
perceived in the celestial and terres-trial environment by a
particular culture (Ruggles and Saunders 1993:16). Second, in
addition to the problems encompassed in the concept of objective
data, a general visual analysis (like the one presented here)
eliminates the human factor in the depiction of trends and the
creation of these trends as the result of social processes that
cannot be subject to prediction or universal laws (Ruggles and
Saunders 1993:17). Although it is acknowledged that the meaning and
role of the night sky is neither self-evident nor com-mon between
peoples and is, instead, subject to so-cial processes and use
(Saunders 1991:13), because of the volume of data presented in this
paper, only an analysis of orientation patterns can be presented
here.
New Directions
Epigraphic, literary, and archaeological evidence attest that
several minor games, competitions, and celebrations were held in
Greek sanctuaries. Usually there was one major festival that was
considered the largest and most important, held in honor of the
deity to which the sanctuary and the main temple within it were
dedicated. This festival would usually take place on a set day in
the year, most commonly annually or, in the case of major
Panhellenic sanctuaries, every two or four years, with minor
celebrations on the same day in the other years. It was important
to ensure that festivals were held on the correct day and that the
calendar did not move out of season. Lunar calendars make such a
requirement dificult. The Greeks were well aware that the lunar
cycle (approximately 29.5 days) does not it into a year comprised
of 365 days. They compensated for this by intercalating an extra
month approximately every three years. Each polis had its own
calendar, with different month names and intercalation times. In
addition, although the new months would always start with the
sighting of the new Moon, this was determined by local
observa-tions, was far from ixed, and was subject to manipula-tion
(Aristophanes Clouds 1134; Trmpy 1997:1, 5). Those festivals that
attracted participants from across Greece demanded a more
Panhellenic timekeeping
method in order for other cities to know that the time for a
certain festival was arriving.
If we suppose for a moment that temples pointed toward a part of
the horizon in which a certain astro-nomical phenomenon was
observed or predicted at the time when the annual festival was to
be held, this phenomenon had to be annual, like the religious
fes-tivals, and connected either to stellar (i.e., the heliacal
rising or setting of stars, apparent acronychal rising, apparent
cosmical setting) or to solar observations (i.e., the point on the
horizon where the Sun rises on a speciic day in the year). As the
solar explanation can be eliminated at least for the data falling
outside the solar range, we may examine the possibility of stellar
associations. Homeric references (circa 750 B.C.) to such stellar
observations (Iliad 18.483489, 22.2631), Hesiods Works and Days
(383384, 609611) (circa 700 B.C.), and the use of parapegmata from
at least the ifth century B.C. (Hannah 2005b) testify that
alternative timekeeping methods to the lunisolar cal-endar were
known and practiced by the Greeks since the Geometric period. These
methods were thus avail-able in those cases when precise
timekeeping was of the essence, such as the performance of
agricultural activities. In the religious sphere we know that the
gods had to receive their sacriices at the correct time every year
(Aristophanes Clouds 615626). The use of star calendars for
religious purposes is much easier to demonstrate during and after
the Classical period. Astronomical observations based on the
fourth-cen-tury paragegma of Eudoxos are displayed in an Egyp-tian
papyrus from Hibeh, a festival calendar dating to 300 B.C. that
recorded astronomical movements of interest to the religious
authorities, assisting in the keeping of the festival celebrations:
in time with the agricultural seasons to which the cults were
attached (Hannah 2005a:62). Parapegmata may have been used
throughout the Greek city-states in order to assist with the timing
of the religious festivals (in addition to other functions). The
example of the Pythais in Athens (the religious procession that the
Athenians sent to Delphi every year) demonstrates clearly that
watching the skies for a sign (in the case of the Pythais a
meteorological sign) before commencing a religious procession was a
reality in ancient Athens, at least from the second century B.C.
(Dillon 1997:24,
-
18 ARCHAEOASTRONOMY
234n118). Rising and setting stars span the entire range of
declinations. The plethora of stars in the night sky means there is
a strong risk of identifying totally spurious correlations between
structure orientations and stellar bodies. Thus, it is essential
that appropriate criteria are employed in order to avoid random and
ungrounded associations. For a convincing case to be made, a study
of the orientation of a structure must draw upon epigraphic,
historical, mythological, and archaeological evidence when
considering possible correlations. The simple association of
stellar bodies to a structure that is purely based on the
structures orientation is no longer suficient.
Preliminary results from the oracle of Apollo in Delphi (Salt
and Boutsikas 2005), the sanctuary of Artemis Orthia in Sparta
(Boutsikas 2008), and the Erechtheion (Boutsikas 2007:119145)
suggest that there may be a connection between the timing of a
religious activity and a stellar event visible in the part of the
horizon toward which the main temple in the sanctuary was oriented.
The temples of Artemis Orthia in Sparta and Apollo in Delphi may
well have been oriented toward the heliacal rising of a particular
star or constellation, and in the case of the Erechtheion the
associated cult rites seem to be tightly timed at the most
signiicant phases of the culmination of a constellation associated
with the myths surround-ing the structure and the Acropolis. The
association between the deity and the speciic constellation is
demonstrated in all three cases by mythology, by the connection
between the movement of the constella-tion and the timing of the
annual festival, by ancient historical records, by archaeological
inds, and by the foundation myth of the cult. Such a network of
inter-locking relationships is hardly surprising: throughout
archaeological and anthropological research we learn about the
enmeshing of landscapes and places with meanings and symbolism and
the necessity of hu-man actions to maintain the cosmic balance
(Ruggles 1999:120121). Greek religious practice and cult in its
early stages prior to the development of temples were performed in
the open air. This implies that normally cult practices and ritual
preceded temple construction. Further studies will establish
whether temple orientation is in fact strongly contextual and
largely determined by local rather than regional
trends in cult practice, in other words, whether the
construction and orientation of a temple were unique and
historically situated within the particular group that built
it.
Acknowledgments
This project would not have been possible without the
cooperation of the following Greek Ephorates of Classical and
Prehistoric Antiquities who have kindly given me permission to
survey the archaeological sites included in this study: , , , , , ,
, , , , , , , , , , , . I am also very grateful to the British
School at Athens for awarding me the Richard Bradford Mc-Connell
Fund for Landscape Studies in 2004, which funded the survey of the
majority of the sites in the Aegean islands, and to Professor Ilias
Mariolakos for his help with questions of a geological nature.
Finally, but by no means least, I am indebted to Professor Robert
Hannah, Professor Graham Shipley, and Pro-fessor Clive Ruggles for
their feedback and valuable comments on earlier drafts of this
paper.
ReferencesAveni, Anthony F.2002 Conversing with the Planets: How
Science and Myth
Invented the Cosmos. Revised edition. University Press of
Colorado, Boulder.
Beyer, A. 1990 Die Orientierung griechischer Tempel. Zur
Bezie-
hung von Kultbild und Tr. In Licht und Architektur, edited by
W.-D. Heilmeyer and W. Hoepfner, pp. 19. Ernst Wasmuth,
Tbingen.
Boutsikas, Efrosyni 2007 Astronomy and Ancient Greek Cult: An
Application
of Archaeoastronomy to Greek Religious Architec-ture,
Cosmologies and Landscapes. Unpublished Ph.D. dissertation, School
of Archaeology and Ancient History, University of Leicester.
2008 The Cult of Artemis Orthia in Greece: A Case of
Astronomical Observations? In Lights and Shadows in Cultural
Astronomy. Proceedings of the SEAC 2005, Isili, Sardinia, 28 June3
July, edited by Mauro P. Zedda and Juan Belmonte, pp. 197205.
Associazione Archeoila Sarda, Isili.
Compton, Michael T.2002 The Association of Hygieia with
Asklepios in
Graeco-Roman Asklepieion Medicine. Journal of the History of
Medicine and Allied Sciences 57:312329.
-
VOLUME XXI 20072008 19
Dillon, Matthew 1997 Pilgrims and Pilgrimage in Ancient Greece.
Rout-ledge, London.Dinsmoor, William Bell1939 Archaeology and
Astronomy. Proceedings of the
American Philosophical Society 80:95173.Ghezzi, Ivan, and Clive
Ruggles2007 Chankillo: A 2300-Year-Old Solar Observatory in
Coastal Peru. Science 315:12391243.Hannah, Robert2005a Greek and
Roman Calendars. Constructions of Time
in the Classical World. Duckworth, London.2005b Euctemons
Parapgma. In Science and Mathemat-ics in Ancient Greek Culture,
edited by C. J. Tuplin and T. E. Rihll, pp. 112132. Oxford
University Press, Oxford.
McCluskey, Stephen C.1982 Archaeoastronomy, Ethnoastronomy, and
the His-
tory of Science. Annals of the New York Academy of Sciences
385:342351.
2004 Astronomy, Time and Churches in the Early Middle Ages. In
Villards Legacy: Studies in Medieval Technol-ogy, Science and Art
in Memory of Jean Gimpel, edited by M. Zenner, pp. 197210. Ashgate,
Aldershot.2006 Medieval Liturgical Calendar, Sacred Space, and the
Orientation of Churches. In Time and Astronomy in Past Cultures,
edited by A. Soltysiak, pp. 139148. Institute of Archaeology,
University of Warsaw, Torun and Warsaw.
Mikalson, Jon D.2005 Ancient Greek Religion. Blackwell
Publishing, Ox-
ford.Nissen, Heinrich1873 ber Tempel-Orientierung. Erster
Artikel. Rhein-
isches Museum 28:513557.1906 Orientation. Studien zur Geschichte
der Religion.
Weidmann, Berlin.Penrose, Francis C.1893 On the results of an
Examination of the Orientations
of a Number of Greek Temples with a View to Con-nect these
Angles with the Amplitudes of Certain Stars at the Time the Temples
were Founded, and an Endeavour to Derive therefrom the Dates of
their Foundation by Consideration of the Changes
Produced upon the Right Ascension and Declination of the Stars
by the Precession of the Equinoxes. Philosophical Transactions of
the Royal Society of
London, series A, 184:805834.Ruggles, Clive L. N. 1984
Megalithic Astronomy: A New Archaeological and
Statistical Study of 300 Western Scottish Sites. BAR British
Series 123. British Archaeological Reports, Oxford.
1999 Astronomy in Prehistoric Britain and Ireland. Yale
University Press, New Haven and London.2000a Ancient
AstronomiesAncient Worlds. Journal
for the History of Astronomy: Archaeoastronomy
Supplement 25:S6576.2000b The General and the Speciic: Dealing
with Cultural Diversity. Archaeoastronomy: The Journal of
As-tronomy in Culture 15:151177.Ruggles, Clive L. N., and Nick J.
Saunders
1993 The Study of Cultural Astronomy. In Astronomies and
Cultures. Papers Derived from the Third Ox-
ford International Symposium on Archaeoastrono-
my, St. Andrews, United Kingdom, September 1990, edited by Clive
L. N. Ruggles and Nick J. Saunders, pp. 131. University Press of
Colorado, Boulder.
Salt, Alun, and Efrosyni Boutsikas2005 Knowing When to Consult
the Oracle at Delphi.
Antiquity 79:564572.Saunders, Nick J. 1991 The Jaguars of
Culture: Symbolising Humanity in
Pre-Columbian and Amerindian Societies. Unpub-lished Ph.D.
dissertation, Department of Archaeol-ogy, Southampton
University.
Scully, Vincent1979 The Earth, the Temple and the Gods. Greek
Sacred
Architecture. Revised edition. Yale University Press, New Haven
and London.Trmpy, Catherine1997 Untersuchungen zu den
altgriechischen Monats-
namen und Monatsfolgen. Universittsverlag C. Winter,
Heidelberg.
Vail, Gabrielle, and Anthony F. Aveni 2004 The Madrid Codex: New
Approaches to Under-
standing an Ancient Maya Manuscript. University Press of
Colorado, Boulder.