The Pass at Thermopylae, Greece

Board of Trustees, Boston University
The Pass at Thermopylae, Greece
Author(s): John C. Kraft, George Rapp, Jr., George J. Szemler, Christos Tziavos, Edward W.
Kase
Reviewed work(s):
Source: Journal of Field Archaeology, Vol. 14, No. 2 (Summer, 1987), pp. 181-198
Published by: Boston University
Stable URL: http://www.jstor.org/stable/530139 .
Accessed: 18/01/2012 15:06
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Boston University and Board of Trustees, Boston University are collaborating with JSTOR to digitize, preserve
and extend access to Journal of Field Archaeology.
http://www.jstor.org
The Pass at Thermopylae,Greece
JohnC. Kraft
Universityof Delaware
Newark,Delaware
GeorgeRapp, Jr.
Universityof Minnesota
Duluth,Minnesota
George J. Szemler
LoyolaUniversityof Chicago
Chicago,Illinois
ChristosTziavos
NationalCentrefor MarineResearch
Athens,Greece
EdwardW. Kase
LoyolaUniversityof Chicago
Chicago,Illinois
Conflicts among historians over the battle at Thermopylae in 480 B.C. tend to
center around supposed inconsistencies between ancient sources, particularly
Herodotus, and the modern topography of the region. The area, however, is
one of extensive tectonic activity, fluctuations in sea level, and sediment deposition. Any attempt to reconstruct ancient events on the basis of modern
topography alone is therefore bound to be misleading. Precise paleogeographic reconstruction would require a large-scale drilling program. This
study involved the drilling and analysis of seven core holes in sediments
infilling the Gulf of Malia. Results clearly demonstrate a Holocene epoch
marine incursion to the far west of the Malian embayment and subsequent
considerable variation over time in the physiography at Thermopylae. We
have reconstructed the shoreline for ca. 480 B.C. and examined variations in
the morphology of the middle gate with respect to the width of the pass.
Geological and geomorphic evidence suggest that the pass at Thermopylae
was closed for long portions of recorded history and may therefore have had
less relative importance as a route into southern Greece.
Introduction:Historical Studies
The heroicdefense of Greeceby Atheniansand Spartans againstmultiplePersianinvasionsin the early 5th
centuryB.C. has loomed large in studiesof the originof
Westerncivilization, with Herodotus'work on the invasion of Xerxes serving as a focus of scholarlyattention. Extensive efforts have been made to reconstruct
details of these importanthistoricalevents. One major
event in the conflict was the stand againstthe Persians
by the Spartansand allies underLeonidasat the narrow
coastal"pass"betweenthe mountainsand the sea (Gulf
of Malia) at Thermopylae(a Greekword meaning"hot
gates,"in referenceto the thermalspringsthatissue there
from the base of the mountains).Xerxes' route from
northernGreeceto Attica(Athens)supposedlyfollowed
the coast becausehe intendedto keep in touch with his
fleet andbecauseGreeceis mountainousandwouldhave
affordeddifficultinlandpassage for such a large army.
Readersnot familiarwith the details of Herodotus'accountof the battleareadvisedto review it in the original
182
The Pass at Thermopylae/Kraft, Rapp, Szemler, Tziavos, and Kase
Figure 1. Location of the passes at Thermopylae on the Malian Gulf, eastern
Greece, Aegean Sea.
BULARI
I~TURKEY
YUGOSLAVIA.a
ALOAMIA
~
GREEC
:i
I
SE
~
~
ri-
~~?
OIIA
s:~
t~lf
sa
~
AEGEAN
:rSEA
A
QUATERNARY:::?;
~
PR-UTRNR
K
50l')O
!00
Figure 2. A geomorphic map of the area
of modern Thermopylae showing the line
of drill cores used in constructing crosssection, Figure 7. The relationship of the
west, middle, and east gates, and the alluvial fan of the deep ravine emerging from
Mt. Kallidromon, near the ancient site of
Anthela, are shown.
~goo
LEGEND
. on
!
o_K ILOMETERSR
TP- 1
1,;
CONTOURS
a eACKSWAMP
-- m
HOLE
$DRILL
LLAL
WoLMARSH
*
J~L
HOT SPRING
o
WES
ALN
AL
SHOREE480B.C.UL
1L
GATE
EAST GATE
1-WG-1-'
M
..-880
4TP-
-KOL
NOS
KALLIDROMON
or in one of the countlessancientand modem summaries.
Historianstell us that the battle of Thermopylaein
480 B.C. (and many other battles)was fought there because of the narrownessof the passagebetweenthe sea
andthe steep slope of Mt. Kallidromon(FIGS.
1, 2) (Her.
7. 198-201, 211-213; Paus. 10.20). Further,we know
thatat least one Greekship was able to remainin contact
with the Spartanforces at Thermopylaeuntil the end of
the battle, suggesting water depths greater than 1 m
immediatelyoffshore (Her. 8. 21.1). Thus the pass at
Thermopylaewas usable in 480 B.C.Historianssuggest
that the pass was approximately20 m wide or wider at
the middle gate (see FIG. 2) but much narrowerat the
west gate andeast gate (Leake 1836; Gell 1819). "Gate"
refersto each of three narrowpoints in the pass where
Journal of Field Archaeology/Vol. 14, 1987
low cliffs of the mountainsjutted out very close to the
sea.
Historianshave regardedthe pass along the sea at
Thermopylaeas the most logical and relatively easy
route from northernGreece south to Attica. Because
Herodotus(and other ancienthistorianswho discussed
the battle at Thermopylae)wrote nearly a half century
or moreafterthe battleof 480 B.C.,lively disputeshave
continuedconcerningthe location where the Spartans
madetheirstand,the widthof the pass, andotheraspects
of the geographyof the battle. Attemptsby modemrn
scholarsto do topographicreconstructionsof the battle
areahave been completelycompromisedby the infilling
of the Gulf of Malia by sedimentsfrom the Sperchios
River. Sedimentaryprocesses have completelyobliterated the ancientlandscapeof the Malianplain. It is not
our intentionin this paperto review and critiqueall of
the pointsat issue in the geographyof the 480 B.C.battle
at Thermopylaebut ratherto set out new informationon
the paleogeographicconstraintsof the battle based on
our three-dimensionalgeological reconstructionof the
ancientphysiography.
A battlesimilarto the one in 480 B.C.was fought in
191 B.C.between forces under Antiochus the Syrian,
with Aetolianallies, and the Romanswith PhilipV and
Macedonianallies. In this case, Antiochusentrenched
his army at the east gate while the main Roman army
underM'. Acilius Glabrio'campedwithinthe pass near
the hot springsat the middlegate (Livy 36. 15.3; 16.5).
Althoughthe Romans outnumberedAntiochus' forces
approximatelytwo to one, Antiochuswas able to hold
themoff initiallyat the east gate. The Romans,however,
made a direct flanking movement along the cliff-like
slopes of Mt. Kallidromon, and M. Porcius Cato
droppeddown on the flank of Antiochus'army (a feat
probablynot deemed possible for even an integrated,
disciplinedregiment)with approximately2000 legionnaires and routed them so completely that only 500,
includingAntiochus, escaped. Over several millennia,
the Gauls, the Huns, the Bulgarians,and otherpeoples
invadedGreece via the Sperchios-Malianplain.2Each
time, when opposed in the pass at Thermopylae,the
invaders made a flanking movement and the Greeks
retreated.MacKay(1963:241-255) recordsa largenum1. "M'." is an abbreviation for the praenomen "Manius";"M.", however, stands for "Marcus."
2. The battle of 480 B.C.vs. the Persians; 279 B.C. vs. Brennus and
the Gauls; 207 B.C.Philip vs. the Aetolians; 191 B.C. saw Antiochus
of Syria vs. the Romans; A.C. 395 Alaric (unopposed); 6th century
A.C. fortifications by Justinian; A.C. 1204 Boniface of Monferrat (unopposed); and even the Germans in 1941 (by this time, however, the
pass was no longer narrow). It is probable that the pass at Thermopylae was never successfully defended.
183
ber of walls constructedin many areas on the NWside
of Mt. Kallidromon,presumablyto preventsuch flanking movements.Perhapsno one ever successfully defended the pass at Thermopylae when a flanking
movementwas made. Both the Persiansand the Gauls
turnedback and enteredsouthernGreecevia the pass at
Vardhatesbetween the TrachinianCliffs and Mt. Oeta
to the west of the Sperchiosplain. It is possible that
Greekstrategythroughoutancienttimes was not to win
battles at Thermopylae,but ratherto turnthe flanksof
armies and peoples intendingto enter Greece through
the centralcorridorinto the Valley of Doris. Pritchett
(1985: 190-216) presentsa summaryof militaryactions
relatedto the pass at Thermopylae,somewhatin contrast
to our own thoughts.
Concerningthe ancienttopographyandthe geography
of the aforementionedbattles MacKay (1963) has observed:"If, afteronly 150 yearswe have some difficulty
following the journeysof English travelers,after 2000
years we must be extremely hesitant in rejecting the
accountsof an ancientsourcesimplybecausewe cannot
immediatelydeterminehis itinerary."MacKayis alluding to the largenumberof scholarsin the last 250 years
who have arguedover detailsof the ancienttopography
at Thermopylae.Much of the argumentis based on
statementsof Herodotusthat are consideredby many
scholarsto be factual. These scholars assume that the
Persianarmy was camped on the Malian plain at the
base of the NWedge of Mt. Kallidromon.Herodotus
claims that the Persian flanking unit (the Immortals)
startedout by crossing the Asopus River and thence
began its climb. A crucialstatementis that the Persian
Immortals,guidedby Ephialtes,hadthe "Oetaeanmountainson theirrighthandandthe Trachinianon theirleft"
when startingtheirclimb aroundMt. Kallidromon(Her.
7. 216, 217). A look at present geomorphology (FIGS.3,
4) makes the problemclear. Because the Asopus River
is at the foot of Mt. Kallidromonand in ancienttimes
may have coursedbetweenpartof the Persianencampment and the mountain,a dilemmais created.The Asopus River must be crossed and yet a path with the
mountainsof the Trachinianson the left andthe Oetaeans
on the rightmustbe accountedfor. These two conditions
are not necessarilycontradictoryas to the flankingroute
of the Persian Immortals (MacKay 1963; Pritchett 1962,
1965, 1982a-d). Further, Herodotus noted that the Asopus River flowed past Anthela, near the west gate (only
possible a minimum of 3 km wNW of Anthela townsite)
(Her. 7. 200).
We know from various historic sources that ships were
able to approach Greek positions at the middle gate.
Indeed, two centuries after 480 B.C., archers on ships
fired arrows at the attacking Gauls (Paus. 10. 21.4). A
184
The Pass at ThermopylaelKraft, Rapp, Szemler, Tziavos, and Kase
Figure3. An indexmapto the topographyof the Gulfof Maliaandthe SperchiosRiverdeltaplainshowinglines of
cross-sectionanddrillsites (Stahl,Aust,
andDounas1974;Austet al. 1980, 1984;
Tziavos1977, 1978).
t. t
?!M
rOTHXtYS
OTHR
Y
DRILLHOLESTHISSTUDY
a DRILL
HOLES
SN
NIGMIR & YEB
contour
,
N
314
3T74
20 m
STYLI$
c'
LAMIA
Interval
v/cvt
GULF OF MALIA
L
01%j
,,
OS
1 2 3 4 5
KILOMETER
MOLOS
O
McO
SMt.
K
THERMOPYLAE
MON
ALLIDRO
C:2
~
t,
r
11;kl^
Figure4. A topographiccontourmapof
theSWcomerof the Malianplainincluding the areaof Mt. Oeta, the Trachinian
Cliffs, andthe presentlocationsof the
Asopus,Xirias(Melas),andGorgopotamos (Dyras)Rivers.Shorelinesfor480
LEGEND
\
.Qh
oJmL
%
1
•CONTOURS
DRPLLHOLE
20m
4m
U
B.c. and 4500 B.P.are shown. Alluvial fans
andthepassat Vardhatesareemphasized
by the topographiccontours.Startof Persianflankingmovementas describedby
Herodotus(8. 21.1) andWallace(1980) is
shown.Recently,a drainageditchfrom
springsat thefoot of the TrachinianCliffs
flowingpastdrillsite KR-1has beenrenamed"Melas"afterthe ancientnameof
theXiriasRiver1-2 km to the north.D =
DehmaGap.
SKILOMETERS
0ATO
VARDHATES
s-s
MODERN
TRACHIS/
HERACLEIA
KR-1
AVARDHA
PERI8AN
FLANK
AEHITA
IANOPEA
MT
Journal of Field Archaeology/Vol. 14, 1987
centurylater, the groundbetween fortificationsand the
sea was an impassablestretchof swamp and quicksand
(Livy 36. 18.4). Yet we also know thatduringLeonidas'
last desperatefight, many Persiansdrownedin the nearshore watersof 480 B.C.(Her. 7. 223.4). At the end of
the battle, an Athenianwarshipunder Abronichusleft
the beach nearKolonos and returnedto the allied Greek
fleet at the northend of Euboea. On the other hand, in
279 B.C.,in the battle with the Gauls, only with great
difficulty could the Athenian fleet rescue the trapped
Greek defendersof the middle gate. From 480 B.C.to
279 B.C., either local relative sea level had dropped
slightly or else shallow marinemuds of the Sperchios
River pro-delta formed muddy offshore shoals that
affordednaval vessels only poor access to the middle
gate.
Figure5 is a photographtakenfrom Kolonos looking
towardthe middlegate, showing a large expanseof the
travertinefan. It is importantto realizethatthis view in
480 B.C. would have been looking downhill approximately20 additionalm underthe presentland surfaceto
the actualbattle site, at a slight wideningin the pass of
approximately20 m between the cliff and the watersof
the MalianGulf at the middle gate. Thus, modem observationscan be extremelydeceiving in the interpretation of ancienthistoricevents and geographies.
A similar problemoccurs in consideringthe site of
the city of Anthela (FIG.2), which is located in an indentationor cul de sac to the SEof the west gate. The
ancientvillage was located at the easternmostpoint of
the territoryof the AmphictyonicLeague. One might
questionwhy it was placedin its presentposition. Study
of possibleeustaticfluctuationsof sea level (FIG.6) shows
thatat certaintimes, e.g., early-middlethirdmillennium
B.P., Anthela would have overlooked the sea with an
extremelynarrowbeach, if not completelycut off from
the middle gate. Possibly Anthela was located in this
site because it was a naturalgeological and geomorphologic limit of access to the territoryof the Pylaean
Amphictyony(Kase and Szemler 1983).
Xerxes may have watchedthe battlefromthe low spur
of Mt. Kallidromonabove and to the west of Anthela.
As Herodotusstates: "Duringthese onsets the King (it
is said) thrice sprangup in fear for his army from the
thronewherehe set to view them"(Her. 7. 212.1). The
hill west of Anthela is the only place from which the
battlecould be watchedby a Persianking who presumably requireda large numberof troops to protecthim.
No other geomorphicfeature of 480 B.C.fits these requirements.
A large wide pass at Vardhates (FIGS. 3, 4) extends
throughthe so-called Dhema Gap into the Valley of
185
Doris fromwhich easy passagemay be made southward
via Chaeroneaand Thebes or via the narrowdefile at
Graviathroughthe mountainsto Amphissaand thence
to the Gulf of Corinth(Kase 1973; Kase and Szemler
1983). P. W. Wallace(1980: 14-23) analyzedthe strategic problemsof flankingthe Greek army entrenched
in the pass at Thermopylaein 480 B.C.He notedthatthe
Figure5. Themiddlegateat modemThermopylae,phototakenfrom
the topof KolonosHill lookingwest to the battlefieldof 480 B.c. Up to
20 m of travertineandothersedimentoverliesthe surfaceof the battlefield.The shorelineof 480 B.C.,likewiseburied,was situatedin the
centerof thephotograph.
YEARS BEFORE PRESENT
12
11
10
D
8
7
a
x 1000
4
5
3
2
1
0
PRESENT SEA LEVEL
-10
c
-'-
-20
IL
SEA LEVEL CURVES
-so
BAY OF NAVARINO
GULF OF MESSENIA
-40
PLAIN OF ARGOS
ANATOLIA (0.
EROL)
-50
Figure6. Localrelativesea-levelcurvesfromsomeof the grabenlike embaymentsin Peloponnese,as contrastedwitheustaticor absobasedon evidencefromthe nearbycoasts
lutesea-levelinterpretation
of Anatoliadeterminedby ProfessorO. Erolandsynthesizedin part
froma globalcurveconstructedby Fairbridge.The sea-levelcurves
fromthevariousgrabensaredifferentbecausethe tectonicsaredifferent. Thereis considerablecontroversyregardingthe eustaticsea-level
curve,butthe curvecorrelateswell withsimilarcurvesestablished
elsewherein the world(Fairbridge1961, 1981). Subsidenceratesfor
theGulfof Maliagrabenarenot knownbecauseof minimaldataon
relativesea-levelpositions.
186
The Pass at Thermopylae/Kraft, Rapp, Szemler, Tziavos, and Kase
valley at Vardhates(near the DhemaiGap) was wide
enough for an army to march throughto the higher
elevations, thence turn southward,join the Anapoea
path, descend to the east of the pass at Thermopylae,
and outflankthe Greeksentrenchednearthe hill of Kolonos west of the middle gate. Pritchett,however, has
locatedthe flankingmovementof the PersianImmortals
underHydarnesin a much more direct route:from the
headof the Malianplain at the foot of Mt. Kallidromon
up the edge of Mt. Kallidromon'swesternend, thence
to the Anapoeatrail and down to the Gulf of Malia to
outflankthe Greeks.
MacKay(1963: 241-255) has also studiedthe problem
of the battle of Thermopylaeand the path of the Immortals in great detail. Both he and Pritchett(1962,
1965, 1982a-d)have little questionthatthe shortestpossibleroutewas up the Nwflankof Mt. Kallidromonfrom
the Malianplain, west of Dhamasta.As Pritchettnoted,
it makesno sense for an armyand campwithinsight of
its objectiveto make a lengthycircularroutenorthward,
westward,then south, then east, almost doubling the
distanceof the mostdirectroutearoundMt. Kallidromon
to the middlegate. On the otherhand, a heavily-armed
force of more than 5000 men climbing a mountainat
nightfor a frontalassaultwould surelyhave been heard
by the defending Phocian outposts at the top of Mt.
Kallidromon.Wallace'sargumentwas basedon the fact
that, at the time, the countrysidewas open forest and
the path at Vardhateswas wide enough that an army
could easily march along this route. Wallace (1980)
madethe trip with a flashlightin one night and arrived
at dawn at the top of Mt. Kallidromonheadingdown
the Anapoeapathto the rearof the Greekpositionin the
same amountof time recordedby Herodotus.Most recently, E. W. Kase and G. J. Szemler(1982: 357) have
shownthatthe access throughthe DhemaGap was used
by Xerxes on his way to Athens, and not (as generally
assumed)the access throughThermopylae.
Background:Geological Considerations
The Gulf of Malia is a deep marine embayment
formed as a graben, a rapidly subsidingblock of the
earth's crust (FIG.3). It is flanked on the north by Mt.
Othrysand on the south by Mt. Kallidromon(geologically-upthrown blocks of the earth's crust). To the NW
lies the broad valley of the Sperchios River, to the west
Mt. Oeta, and to the sw the Trachinian Cliffs. It is an
area that has been extremely active tectonically throughout the Quarternary Period (Maratos 1972). With the
waning of the last major ice age more than 10,000 years
ago, sea level began to rise and enter the Gulf of Malia.
Evidence of local relative sea-level rise obtained in our
geological studies shows that the Gulf of Malia and areas
of the presentSperchiosRiverfloodplainanddeltawere
almost completelycoveredby the sea at approximately
4000 B.C.At that time, the river fans and deltas of the
Asopus and SperchiosRivers and other minor streams
flowinginto the Gulf of Maliabeganto infillthe western
end and alluvial fans began to infill the flanks of the
Gulf of Malia.
A curve establishedby Erol (Kraft,Kayan, and Erol
1980) suggests that relative sea level was somewhat
higherthanthatof the presentday for much of the past
6000 years in the Aegean area (FIG.6). Local tectonic
evidence and studies in grabenselsewhere suggest that
the grabenor basin of the Gulf of Malia continuedto
subside throughoutthe Holocene epoch (past 10,000+
years) (Kraft,Rapp, and Aschenbrenner1980). Studies
in the grabenof the Gulf of Malia and sea-level data
from grabenselsewherein Greece show that local relative sea level has variedsomewhatwhile the sediments
in the grabenscontinuedto subside(Kraft,Aschenbrenner, and Kayan 1980).
Approximately14,000 years ago, nearthe end of the
latest glaciation, sea level was approximately100 m
below its presentlevel. Withthe waningof the last major
glaciation, world sea level began to rise to its present
elevation and, accordingto many authors, above its
presentlevel (Fairbridge1961, 1981). Fromthe onset to
about6000 years ago, at the time of the "ClimaticOptimum," sea level rose rapidly, inundatingthe valley
systems of the Mediterranean.There is disagreement
aboutsea levels since 6000 B.P.One school of thought
holds thatsea level rose slightlyabove presentlevel and
then fluctuatedabove andbelow presentelevationsto its
currentposition. Anotherschool of thoughtstates that
the rate of rise droppedsharplyabout 6000 years ago
and furtherslowed down about3000 to 2000 yearsago.
Figure6 presentsseveralsea-levelcurves.Thosefrom
the Bay of Navarino, Gulf of Messenia, and Gulf of
Argosarelocal relativesea-levelcurvesconstructedwith
informationfromradiocarbondateson sedimentsin subsidinggrabens.Therefore,althoughthese curvescan be
accuratefor theirlocal setting, they cannotbe projected
with accuracyto other embaymentswith differenttectonic settings.They aremerelyexamplesof whata local
relativesea-level curvein the Malianembaymentmight
look like without added eustatic sea-level fluctuation.
The eustatic curve constructed by Professor Erol (University of Ankara) is based on data obtained along the
shorelines of Aegean Anatolia added to a synthesis of
Fairbridge's curves. This curve represents the school of
thought asserting that sea level rose rapidly to above its
present level about 6000 B.P. and has since fluctuated
above and below present sea level (Fairbridge 1961,
1981). It is probable that all curves shown in Figure 6
Journal of Field Archaeology/Vol. 14, 1987
arecorrectfor limitedregions.Thus, a local relativesealevel curve derivedfor the Malianembaymentmust be
a compositeof tectonicsubsidencepluseustaticsea-level
rise and fall.
BecauseErol'ssea-levelrise curve(Kraft,Kayan,and
Erol 1980) was based on data from the Aegean coastal
zone, we use his curve in this paperto calculateancient
sea levels. Using Erol's eustaticsea-level curve we can
interpretthe complex geological events that occurred
aroundThermopylae.At the time of the battle in 480
B.C.,sea level was slightly above its presentlevel, thus
deepeningthe offshore watersand narrowingthe width
of the pass at the west, middle, and east gates. Radiocarbondata and stratigraphicinterpretationsfrom the
cross-sectionat Thermopylae(FIG.7) arecompatiblewith
Erol'ssea-levelinterpretation.
Marinesandsextendvery
close to the subsurfaceextensionof the grabenmargin
and the steep cliffs of Mt. Kallidromon(480 B.C.,see
dashedline in FIG.7). Marinewaterswere close to shore
4500 years B.P.and the marinestrataextendto within 1
m of presentsea level (see TABLE
1). A date of 2150 or
2300 B.P.,as shown in core TP-2, shows the positionof
a swampor marshat that time. These data suggestthat
the pass at Thermopylaeat the time of the battlein 480
B.C.would have been extremelynarrow.Yet at the time
of the battle with the Gauls several centurieslater, the
pass was wider, with shalloweroffshore waters. Fluctuationsin width of the pass at Thermopylaeover the
past five millenniaare common, as expected in an unstablestructuralconfigurationalong the flankof a major
graben.
The rates of tectonic subsidencein the Malian Gulf
are unknown.The tectonic events probablyare shortterm, discontinuous,and markedby earthquakes.Many
minor earthquakeshave been recordedthroughouthistoryon the Malianplainincludingat least one inundation
by a tsunami,or majorsea wave (Pritchett1965). Figure
8 is a cross-sectionto the west of the pass of Thermopylae (see line of sectionB, FIG.3). Thesedataarebased
on drill holes made by Stahl, Aust, and Dounas (1974)
and by Aust et al. (1980; 1984) in groundwateror hydrologicstudiesof the subsurfacesedimentsof the Malian plain and SperchiosValley. They clearly identified
up to 300 m of valley-grabensedimentinfill. Theirdeterminationof three gross units (lower mud, middle
sand, and uppermud)basedon more than 18 drill holes
includes a stratigraphicrecord of transgressionand
regressionof the sea throughoutlaterQuaternarytimes.
The sedimentaryrecordof fluvial and marinedeposits
in the Gulf of Malia and the SperchiosValley graben
may extendto much greaterdepths (approximately800
m). The sedimentaryrecordmay possibly includea stratigraphicrecordof the entireQuaternaryand latest Tertiary(Zamaniand Maroukian1980).
AlthoughAust and Stahl and othersclarifiedartesian
relationshipsin the western Malian Basin, they also
showed that single sedimentaryenvironmentallithosomes such as strandlines,paralic sands, distributary
sands, alluvialfans, deltaic organicmuds, marineprodelta muds, backswampfloodplaindeposits, etc. could
be preciselydelineatedgiven a sufficientnumberof drill
cores, radioisotopicdating,etc. (Sahl, Aust, andDounas
A
SOUTH
+40
A'
Mt. KALLIDROMON
UP TO 1300m
NORTH
THERMOPYLAE
MIDDLE GATE
TP-4
+30
+20
0
(EARLY 19th CENTURY)
_/
SEA
SPERCHIOS RIVER
(OLD CHANNEL)
O•
7~740
B.P..'
I
LEVEL
FLOODPLAIN
150or2300
B.P
MARINE SANDS
Figure7. A geologicalcross-sectionat
the middlegateof the passat Thermopylae. Drill-coresectionsprovideinterpretationof depositionalenvironmentsfor
subsurfacesedimentsandradiocarbon
datingof materialsestablishesa time
frameforthe depositionof theseenvironments.Caremustbe used, however(i.e.,
coreTP-3shows a veryold dateoverlying
a youngdate). Spuriousdatesarecommon in streambed sequencesas materials
can be retransported
anddeposited.The
dashedline indicateslandsurfaceandsea
bottom,480 B.C. (See line of sectionin
FIG. 2.)
w -20
MLEGENDR
-30
o
GRAVEL
COARSE
COARSE
TO VERY
SAND
* MEDIUM SAND
*FINE TO VERY
FINE SAND
0
-40
0
250
500
-
750
METERS
SILT
CLAY
MICROFAUNA
TRAVERTINE
a
187
MARSH
1000
1250
188
The Pass at ThermopylaelKraft, Rapp, Szemler, Tziavos, and Kase
Table 1. Radiocarbondatafrom drill-coresamples.
Lab
Sample
#
TP-2
Elevationof
Samplein
m*
4.5 to 3.75
TP-2
-1.2 to -1.5
TP-2
-9.25 to -9.6
TP-3
-1.3 to -1.8
TP-3
-2.2
TP-4
-8.0
KR-1
WG-1
+25
-
Material
Dated
Grassroot
Grassroot
systemin gray
mud
Pelecypodand
gastropod
shells
Totalorganics
in darkgrayblack mud
Woodfragmentof
tree
Marinegrasses
and/oralgae
Wood
Total organics
Inferred
Depositional
Environment
Coastal
swampmarsh
Coastal
swampmarsh
Shallow
marine
-
1
14C
90 ? 9
Age in YearsB.P.
5568
5730
half-life
half-life
755.? 80
778
MASCA
CorrectedDatet
A.C. 1220-1200
232 ? 8
2120 ? 80
2184
200 or 350 B.C.
419 ? 10
4360 ? 140
4491
3150 B.C.
Very shallow
marine
478 ? 7
5220 ? 110
5377
4040-4060 B.C.
Very shallow
marine
56 - 9
460 ? 80
474
A.C. 1420
Very shallow
marine
24
3970 ? 330
4089
2585 B.C.
Alluvium
Pro-delta
-
240 ? 80
3760 ? 100
247
3872
A.C. 1703
2180 B.C.
to meansea level.
*Relative
tSee Ralph,Michael,andHan1973.
1974). Thus, precise paleogeographies could be reconstructed with a large-scale drilling program. We have
shown that the upper 10-20 m of sediment are of midHolocene epoch and younger (FIG. 7; TABLE 1) and thus
represent Helladic and younger depositional events, all
of great import to our hypotheses and interpretations.
Information available to us clearly demonstrates a
Holocene-epoch marine incursion to the far west of the
Malian embayment. The precise margin of the marine
environments along the surrounding cliffs is indefinite,
however. We have therefore used our interpretive section
at the middle gate (FIG.7) as a guide to maximum possible strandlinepositions from Figure 8 and as illustrated
in Figures 2 and 4. Figure 8 is of particularinterest since
it shows clearly that the marine embayment of the Gulf
of Malia in Helladic times extended west of the presentday delta-plain towns of Anthili and Moschochorion.
Indeed, it is probable that with the initial peak flooding
of the graben, approximately 6000 years ago, the entire
embayment was flooded to all flanking cliffs.
Erosion of the highlands brought sediment to stream
valleys where alluvial systems transportedthe materials
to the low-lying coastal plains. In some of the deltas that
were formed, the land protruded far into the sea. In the
sheltered Gulf of Malia, at the delta of the Acheloos
River to the west of Mesolongi in Acarnania, and at the
classic deltas of the Thermaic Gulf where the ancient
Macedonian port of Pella has now been abandoned, lateHolocene deltaic deposits have pushed coastlines far
seaward. Other importantchanges have been wrought in
areas where the sediments from smaller streams were
redistributedby wave action and resultant barrier accretion plains. An example of this type of morphological
modification occurred in the area of the "five rivers" and
Pamisos River in sw Greece (Kraft, Rapp, and Aschenbrenner 1975). Summaries of some of the processes
involved have been published by Kraft, Aschenbrenner,
and Rapp (1977); Kraft, Rapp, and Aschenbrenner
(1975); Kraft, Kayan, and Erol (1980); and Kraft,
Aschenbrenner, and Kayan (1980).
Tziavos (1977, 1978) studied the sediment distribution
and micropaleontology of the modem sediments of the
Sperchios delta. Zamani and Maroukian (1979) outlined
the changes in coastal geography of the various distributaries of the Sperchios River after the great southern
extension of the bird's-foot distributary bypassed the
southern shoreline at Thermopylae, destroying the geomorphic landform of a "pass" that had existed for the
previous four to six millennia. Precise delineation of
shoreline positions, coastal and deltaic environments,
and general geomorphology of the plains of the Sperchios and tributaryvalleys prior to A.C. 1810-1830, can
only be accomplished by the application of scientific
methodologies supportedby archaeological evidence and
Journalof Field Archaeology/Vol.14, 1987 189
topographiesat Thermopylae,we drilled and analyzed
NOfTH
the materialsfrom seven core holes in the sediments
?A?
XODIT
????
2, 7, 9).
infillingthe Gulf of Malia (FIGS,
*????
D
( ????
5-2
8DPERCHIO8
***???
7
Surface
and
*
geological studies and subgeomorphic
+20
MOSCHOCHORIONRIVER
SOUTH
+40
?,,
.?....o..
DHAMABSTA
BD-1
.:.:::
....
<ft
>,z
surface geological results show the natureof the pass
between
the mountainsand the sea over the past 10,000
.HOLOCENE
FLO
SEA 0
From
the core drillingwe were able to identify
PLAIN
LEVEL.
I•D
lr I i/ r????? years.
.?p. ii
the
HOLOCENE
depositionalenvironmentof the various sediments
?.:.e..:
SHALLOW So..
that
form the coastalplain at presentday Thermopylae.
-2o
In
the
firstmillenniumB.C.,the east andwest gates were
-201?6
very narrow,faced by steep cliffs. The middlegate was
a wider (about20 m) sandy travertineshoreline(many
Lu
- m1
.- L.:
36
2
:
battleswere foughthere) betweenthe sea and very high
precipitouscliffs (see note 2). Evidence from our drill
cores numberedWG-1, TP-1, TP-2, TP-3, and TP-4
12
14
13
9
7
10
11
8
5
6
4
0
3
1
2
(FIG.2), andfrommapsby Leake(1836) andGell (1819),
KILOMETERS
LEGEND
suggests that the narrow pass ceased to exist at the
beginningof the 19thcentury.Massive sedimentationof
the SperchiosRiver delta has bypassed Thermopylae,
creatinga delta-plain.Nevertheless,transportation
along
Figure8. A geologicalcross-sectionacrossthewesternMalianplain
3
into
the pass was still poor in the second decadeof the 19th
FIG.
for
line
of
Shallow
marine
sedimentation
the
(see
section).
farwesternpartof theMalianplainis provenby themarinesediments
century(Gell 1819). Our line of drill cores was made
inthecores.
to the cliffs of Mt. Kallidromonfrom the
perpendicular
area of the hot springsand Kolonos at the middle gate
historicalevidencederivedfromthe examinationof narnorthwardacrossthe southernside of the Malianplain.
rativehistoricalrecords.
They show thatfrom4000 B.C.to approximatelyRoman
times the southernshorelineof the MalianGulf lay very
close (20 to 100 m) to the flankof Mt. Kallidromonat
New Data: Geological Field Studies
the middlegate (FIGS.
2, 7). Thereafter,the widthof the
In orderto determinethe details of the sequenceof
pass fluctuateduntil the beginningof the 19th century,
when a distributaryof the SperchiosRiver infilled the
geological environmentsand to reconstructthe ancient
shorelineareaat Thermopylae.
D
D'
SOUTH-WEST
Antiquemaps and sketchescan vary from precise to
NORTH-EAST
60
TRACHINIANCLIFFS
useless and/ormisleadingsourcesof data. We were forUP TO 700m
tunatethatLeakepublisheda map of the studyareaand
50
left good notes regardingrapidchanges in landformsat
Thermopylaein the early 19thcentury.His observations
are proven, because the recent (19th-century)southerly
_
of the SperchiosRiver can be directlyand
distributary
FAN
FINEALLUVIAL
favorablycomparedwith its extant dry channeljust 1
0
km NEof the "gates"of Thermopylae(FIG.10).
25
0.50
1
1.25
1 50
SEA
FKILOMETERSOODPLAIN
Pouqueville(1835) left us with what is potentiallya
LEGE
great treasure-a sketch of the landscape along the
southernshore of the Gulf of Malia ca. 1800 based on
his travelsin Greecefrom 1798 to 1801, which he first
TONVERY
.go
publishedin 1805. Pouqueville'swritingsregardingtopographyat Thermopylaeare of little use. His sketch
"Gr:ce-Thermopyles"
(FIG.11) publishedin 18353in a
of Greece, however, includesa perspectivediahistory
Figure9. A geologicalcross-sectionfromthe steepTrachinianCliffs
gram of the passes at Thermopylaethat shows a landatRakhita(a BronzeAge site-possibly partof ancientTrachis)at the
E -$
OANTOILI.
.
0
'o-
tr
-
0.
:.
u-
•.
.%
...8
0.............
4
*
:
"
O
PEBBLE
81LT
-
COARSE
GRANULE
CLAY
$AND
40o
MICROFAUNA
(7HELLS
MEDIUM
$AND
J-kPLANT
FINE $AND
FRAGMENTS
RAKHITA
LEVEL0
SEA
-10
-20
P LMARINE
GRAVEL
PROBABLE
SILT
*FINE .9PLANT
FRAGMENTS
ND
KILOMETERS
headof the Malianplain.Core-holeKR-1suggeststhat20 m of sedimentsin thealluvialfansandfloodplainsof theXirias(Melas)and
AsopusRiverscoalesceover shallowmarinesediments.
3. Althoughthe firstinitialsof Pouquevillearenot given in thiswork,
we assumethe authorto be the same as Pouqueville1805.
190
The Pass at ThermopylaelKraft, Rapp, Szemler, Tziavos, and Kase
Atovwr
C)ORTA
(74
SOF
AND
TMx
of-:
"
Pi'
___
1~1
;?
1A1w
FTRAQHIIIIA
wKatavothr
otn
04
AGPUS
ls
v
/P
saltAo*
0
-SALTPOND
THRMPY
SATT
Awd
0
r4R
mw
r
ZARA&
1
ILI
(
l:.
elTATUT
W:
. ..
4.*6
~O
Aqa
AW
h:
v6
Otb
r
WLES
I..w46,
A.W
`i
i E*
a
\
F
P.:YU
.
r avrt
E-
dUL
?LEAKE
"Al i
o
~~~L;
NALIAC G(rL
*i.
MILESii
:STATUTE
.
...
......
.................
. ......
Figure10. W. M. Leake's"Planof Thermopylaeandthe AdjacentCountry"(1836). By thetimeof Leake'stravels,the
hadprogradedseapass at Thermopylaehadwidenedconsiderablyandthedeltaof the SperchiosRiverandits tributaries
wardup to 8 mi. A = ancientcoastlineas interpreted
by Leake;B = middlegateof Thermopylae.
scape sketchednearthe time of Leake's firsttravels(ca.
i802).
Figure 11 is a view to the east from the cliffs at
Trachis. While the vertical and distance (depth) perspectivesaredistorted,the view clearlyshows the southern Sperchiosdistributaryin the foregroundadjacentto
the west gate, and in the distant center, the passes at
Thermopylaesurroundedby a low talus slope and sharp
cliffs mergingwith the sea (Gulf of Malia). From geological and geomorphicevidence, the sketch (FIG.11) is
much as we would predictthe passes to have appeared
shortlybefore or at the time of Leake's first visit. Ap-
Journalof Field Archaeology/Vol.14, 1987 191
"
ban
iZ~
~=;;u
'"`L"s~?~;PL~pi0~~ _~--_~Z_~**CL)^;W...^R
Figure 11. A sketchlookingeastwardfromthecliffs at Trachis/Heracleia to the cliffs at Thermopylae,ca. 1798-1802. Althoughpublished
to the lower
by Pouquevillein 1835, the SperchiosRiverdistributary
left andthe openwatersof the Gulfof Malia(centerandleft) against
the cliffs andalluvialfansof the passesat ThermopylaeclearlypredateLeake'smap(FIG.
as publishedin 1836. Althoughverticalexo10)
aggerationis great,the sketchshowsThermopylaemuchas it appearedfor manymillennia.Problemsof large-scaletrafficthroughthe
"passes"aresomewhatevidentat the firstcliff (nearthe west gate).
Duringan on-sitevisit, this view was identifiedby SzemlerandKase
as thatdescribedin Pausanias10.22.1.
proximately20 years later, on his second visit, Leake
includedhis map (FIG.10), which shows a dramatically
differentThermopylaeby-passedby the SperchiosRiver
delta. Further,this appearanceof Thermopylae(minus
the forefrontwith distributary)is what the passes at
Thermopylaewould have been like from 1800 back to
the ClimaticOptimumof 5,000-6,000 B.P.Pouqueville
(or his artist)shows what he saw at Thermopylaeat the
very end of the eighteenthcentury.
Two excellent geologic-topographicmaps of the area
have greatlyaidedus in our studyof ratesof alluviation
in the Gulf of Malia (MaliakosKolpos) (Marinoset al.
1963; 1967). The presentGulf of Malia is greaterthan
20 m deep. We have no way of knowingthe total thickness of the Quaternarysedimentsthathavepartlyinfilled
the gulf. In none of our borings have we encountered
the base of the Ho;,%Lcne-agesediment. More than 20
boringsby the Greek National Institutefor Geological
and Mining Research/IGMEhave penetratedup to 200
m on the lower floodplainof the SperchiosRiver and
the Sperchiosdelta plain.4These sedimentsincludedeltaic distributarysands, shallow marinepro-deltamuds,
levee and backswampsands and muds, floodplainsediments, and alluvial-fandebris. Correlationof sand and
mud sedimentologicunits is possible in this artesian
and
interpretation
system. Precisetime-linestratigraphic
4. The Holoceneinfill of the MalianGulf is at least70 m thick, south
of Lamia.
arealdelineationof these Quaternarysedimentaryenvironments,however, is not yet possible, althoughPleistocene sedimentsof alluvialfans can be observedalong
the flanksof the highlandssurroundingthe graben.
Figure 7 is a cross-section showing the subsurface
stratathat have accumulatedat the middle gate. The
locationof this cross-sectionis noted on Figure2 (A to
A'). Our sediment cores were used to determinethe
environmentsof depositionby sedimentologicalandmicropaleontologicalanalysis. In those fortunate cases
where organic remainswere encountered,radiocarbon
dates were obtainedto place the sequence of environments into a Holocenetime frame(TABLE
1).
Figure 8 shows subsurfacerelationshipsacross the
western Malian plain. The location of this section is
shownon Figure3 as B to B'. Figure9 is a cross-section
from the TrachinianCliffs towardthe sea. The location
of this sectioncan be seen on Figure4 as D to D'. Figure
12 illustratesthe subsurfacesection of the easterntip of
the moderndelta of the SperchiosRiver. Its location is
shown in Figure3 as extendingfrom C to C'.
Figure2 is a mapof the pass at Thermopylaeshowing
the steepcliffs of Mt. Kallidromonnearthe middlegate.
The alluvial fans of several valleys that emerge to the
east and west of Kolonos and from the deep ravinenear
Anthelaarealso shown. It is importantto note, however,
that these fans were much smaller at the time of the
battleof Thermopylaein 480 B.C.The mergingcontours
shown on Figure 2 indicate that the slopes at the east
gate and the west gate are extremelyhigh and cliff-like.
Duringthe drillingof core-holeKR-1 at the western
end of the Malianplain near the TrachinianCliffs, we
C
C'
NORTH-NORTHWEST
SOUTH-SOUTHEAST
+60
PANOURGIA I
PANOURGIA
1
+40
GULF OF MALIA
+20
SPE RCHIOSOS
DELTA
..
SEA OFN
LEVEL
V
**
iiii
SHALLOW
MARINE
ANDDELTAIC
-20
YELLOW
-40
:
•
a
LEGEND
-40
w
, PEBBLE
GRAY
-60
0
GRANUL
GRAVEL
SILT
iii
CLAY
i
SCOARSE
CONGLOMKILOMETERSERATE
0
I
2
3
4
5
KILOMETERS
6
7
8
9
10
II
12
13
Figure12. A geologicalcross-sectionof the extremeeasterntip of the
moderndeltaof the SperchiosRiveracrossthe Gulfof Maliain the vicinityof Agios Trias.The SperchiosRiveris still rapidlyadvancing
eastward,infillingthe relativelydeepGulfof Malia.(See FIG.3for line
of section.)
192
The Pass at Thermopylae/Kraft, Rapp, Szemler, Tziavos, and Kase
encountereda thick section of probableparalicmarine
sands underlyingfloodplainand alluvial-fansediments.
A small numberof red, coarse sand-size grains were
observed among the gray sand grains dominantin the
entire section. Out of curiosity, one sand grain was
collected and sent to Dr. Diana Kamilli for analysis.
This showed the specimen to be a medium-fired,oxidized ceramic fragmentcontaining biotite-quartz-feldspar with a hematite-stainedmatrix. Although this
occurrenceof ceramicmaterialin the marinesection is
withouta precise date, it does indicatethatpeople were
using potterywhen the watersof the MalianGulf were
at the westernmostend of the Malian plain. This is
compatible with our geological interpretationsof the
westernend of the MalianGulf in the EarlyBronzeAge.
Cross-sectionFigure 9 shows the relationshipof the
BronzeAge site at Rakhitaat the foot of the Trachinian
Cliffs of ancientTrachisto the low floodplainbetween
the Xirias(ancientMelas) andAsopusRivers. Core-hole
KR-i was drilledin the low-lying areabetweenthe two
riversalong a new drainageditch to which the old term
"Melas River" has now been applied. Figure 9 shows
that 3 m of sedimentationoccurredover the last 247
years. Morethan20 m of floodplainand streamdeposits
have formed in this area overlying a probablecoastal/
marinesand unit. Presence of the paralic sand unit in
core-holeKR-1 shows that the ancestralmarineMalian
embaymentextendedclose to the TrachinianCliffs (FIGS.
4, 13),possibly precisely to the foot of the cliffs.
The varieddepositionalenvironmentsidentifiedfrom
our surface studies include alluvial fans, river-channel
deposits, the naturallevees and channelsof bird's-foot
deltaic extensions, floodplainor overbankalluvial de-
Figure13. The westernend of the floodplainof the Gulfof Malia.
TheTrachinianCliffs areto the left, withthe passof Vardhates(near
theDhemaGap)in the centerandMt. Oetato the right.Photograph
takenfromthe low floodplainbetweenthe alluvialfansof theAsopus
Riverandthe Xirias(Melas)River,lookingwest.
posits, salt-marshswamp deposits, and shallow-marine
sandy silts and muds. In addition, in the pass at Thermopylae, particularlyat the middle gate, a large travertinefan is buildingupwardand seaward.
Discussion: Reconstruction of the Ancient
Landscape
The physiographyat Thermopylaehas variedconsiderablythroughtime. Most of this variationis the result
of changes in local relative sea level and sedimentdepositionalinfill of the west end of the Malian graben,
includingmajoralluvial fan, travertinefan, and deltaic
progradation.
A reconstructedshorelinefor approximately480 B.C.,
the time of the battleof Thermopylae,is shown in Figures 2, 4, and 14. Equallyimportant,the reconstructed
Bronze Age shorelinefor approximately4500 years before presentis shown in Figures4 and 14. Any Bronze
Age sites along the westernend of the Malian embayment probablywere coastal sites. Particularly,any of
these sites that might be associatedwith the passes of
the Gorgopotamos(Dyras) River or the Xirias (Melas)
River might be consideredto have been port sites for
of materialsinlandvia passes to the Valley
transshipment
of Doris and thenceinto southernGreece. When the sea
extended to the westernmostlimits of the Malian embayment, the pass at Thermopylaewould have been a
cliff into the sea at the west, middle, and east gates and,
therefore,impassableto large-volumevehiculartraffic.
From 5000 to 3000 years B.P.many small towns may
have lain along the foot of Mt. Oeta along the ancient
shoreline,butthese would now be buriedunderas much
as 20 to 30 m of sediment and are thus essentially
undiscoverableby archaeologists(Simpson 1981: 81).5
Those few sites thatwere higherup on the flanksof the
hills or buriedin the earlierportionsof alluvialfans take
on a new importancein termsof the trafficfromnorthern
to southernGreece (see Rakhita,FIG.9).
In the firstdecadesof the 19th century,Leake(1836)
studiedthe areaof the head of the MalianGulf and the
pass at Thermopylae.He produceda map (FIG.10)that
is extremelyrevealingin termsof the fluvialand marine
processesthathadcontinuedfromthe time of the ancient
Greeksto his time. Leake's map shows his estimateof
the locationof classical Greece's ancientcoastline. His
interpretationis obviously schematic, although he at5. R. H. Simpson(1981: 81) reported"threeMyceneansherdson
the lowest terracesof the site of ancientHeraclea"(see FIG.4) but no
signs on the citadel above. This raises the possibility that Helladic
shorelinesextendedfurtherwestward,as proposedin this paper.Unless verifiedby additionalevidence, however, the sherdscould as
well have been introducedduringa later occupancyor become emplacedas pebble-sizeclasts in transitacrosslowerprehistoricTrachis.
Journalof Field Archaeology/Vol.14, 1987 193
Figure 14. Variationsin modemandancientshorelinesof the Gulfof MaliaFrom
SperchiosRiverdelta-floodplain.
in this
variousmapsandas reconstructed
paper(Tziavos1977, 1978;Bittneret al.
537 1
1880).
J
SHORELINE
•d
__
•
,4500
HORELINE 1950
B.P.
8.4,"
GULF OF MALIA
• •
SHORELINE
1876
HO
,
,
•".,-"
SHORELINE
1972
E
doLE
SHORELIE
.)\
CONOURINTRV.A80
*R0
1
2
3
4
5
-1800
d
;C
-THEROPYLE
1 ON
TOQ
,o
KILOMETERS
?117
20
COTORINERA
temptsto becomemorespecificat the middlegate where
he shows a deep indentationof the coastlinetowardthe
cliffs of Kallidromon,east of Anthelaandjust west of
the hill of Kolonos. Clearly, the SperchiosRiver was
the majorstreamcarryingthe greatestamountof sediment into the head of the Malianembayment.Its tributarieswere, in decreasingorderof importanceof volume
of sediment,the Asopus River, the Dyras River (Gorgopotamos),and the Melas River (Xirias).It is interesting to note thatthe pathof the SperchiosRiverdiverges
southeasterlyimmediatelyupon emergingonto the Malian plainandthen flows almostdirectlytowardthe west
gate, thenceparallelto the cliffs of Thermopylae,to its
distributarieseast of Thermopylae. (This abandoned
channelis still preciselyidentifiablein surfacemorphology.) This is a most unlikelypathfor a majorsedimentbearingriverflowinginto an embayment,however.One
would expect the river to flow into the center of the
embaymentalthough,as its levees rose, it would break
throughand flow to other paths. To have such a long
extension of the river flowing along the flanks of the
embaymentas opposedto the middle suggeststhe possibilityof humanintervention.Leakelabels a majorarea
in the centerof his map as "rice grounds"and notes a
salt works near the shoreline. The possibilitymust be
consideredthat humansalteredthe course of the Sperchios River to its maximumsoutherlyposition in the
later 18thor very early 19thcenturyin orderto createa
large, flat, well-irrigatedarea for the growing of rice.
Naturalphenomenacould have createda similarsitua-
tion, but it is not as likely. Leake also notes the great
changes in delta morphologybetween his first (1802)
and, some 20 years later, second visits.
The dynamicsof geological change in the mountain
rangesprobablyhave been fairly slow to effect change
in the hard mountainrock. On the other hand, in the
fluvial systems of the SperchiosRiver, its floodplain,
andits tributaries(the Asopus,the Melas, andthe Dyras)
anddepositionof sedimentsinto the Gulf of Malia, quite
the oppositeis true. Precisioncontourmaps of the Asopus River, the Xirias (Melas) River, and the Gorgopotamos (Dyras) River clearly show large alluvial fans.
These deltoid intrusionsinto the ancientGulf of Malia
variedgreatlyin geographicextentover the past 5,00010,000 years (FIG.8). At times the alluvial fans were
low-lying and small, at othertimes large and expanded,
as at present.The Asopus River has a 1300 fan spread
to the NWof its presentposition(FIG.4). On the average,
the river flowed 300 to the northof its presentposition.
Thus it is possible, as notedby Green(1970: 126), that
a large portionof the Persianarmy"pitchedcamp near
Trachisbetweenthe Sperchiosand the Asopus Rivers,"
althoughit is questionablewhetherthey occupied Anthela (on the other side of the Asopus River) and a low
ridge of Mt. Kallidromonat the same time. We must
also rememberthat the flankingmovementof the Persians in 480 B.C.refersonly to a regimentof 5,000 men
(the Immortals)as opposed to approximately300,000
men and women in the Persianarmy,if we can believe
Herodotus'numbers(How and Wells 1928). But it is
194
The Pass at ThermopylaelKraft,Rapp, Szemler, Tziavos, and Kase
quite clear that 10,000 to 20,000 Persianscould have
been campedon the rightbankof the AsopusRiver fan
and plain between the river and Mt. Kallidromoneven
if the AsopusRiverof 480 B.C.flowedalongthe extreme
channelon the fan closest to Mt.
southeasterly-possible
Kallidromon.The Persianencampmentsurfaceprobably
lies underthe presentMalianplain, buriedat least 15 to
20 m beneathpartof the presentAsopus alluvialfan by
alluviumof the tributaryrivers and floodplainsof the
SperchiosRiver.
A similarsituationexists at the "pass"at Thermopylae. MacKay(1963) noted that Marinatosdug a trench
8 m deep at the middle gate and was still in sediments
of Byzantineage. Ourgeological analysisof the middle
gate areaclearlyshows thatup to 20 m of travertineand
otherforms of sedimentoverlie the actualbattle site at
themiddlegate(FIG.7).
Figure4 is a topographicmap showingMt. Oeta and
the TrachinianCliffs. At the TrachinianCliffs, a precipitous cliff dropsoff sharplyto the Malianplain. Nearby
and to the NW,the hills and cliffs of Mt. Oeta are still
high but not precipitous.The XiriasRiver(MelasRiver)
and the GorgopotamosRiver (DyrasRiver)carriedsedimentsinto the head of the ancestralGulf of Malia.
It is, therefore,reasonableto identifya large areaof
the Malian plain at the foot of Mt. Kallidromonwith
partof the Persiancamp betweenthe AsopusRiver and
Mt. Kallidromon.Thus Herodotus'statementthat the
Immortalsfirst crossed the Asopus River and then entered a pass with the TrachinianCliffs on the left and
Mt. Oeta on the right is a reasonableaccountand accurategeomorphicdescriptionof the situationin 480
B.C.This would lend supportto the hypothesis(Wallace
1980) that the pass at Vardhateswas that used by the
Immortals.The argumentregardingthe path of the Immortalswill continueinto the futureas it has ragedfor
the past250 years. It is absolutelycritical,however,that
any furtherdiscussionof this matterconsiderthe paleotopographyof the westernend of the Malianplain, including the variouspositions of the Asopus and Xirias
(Melas) Rivers, and the fact that the plain was much
lower (ca. 20 m), with its river channels constantly
changing.
Neither currently-visiblenor abandonedchannelsof
the Asopus River observedby anyone within the past
500 yearshave any knownrelationshipto loci of Asopus
Riverchannelsof morethan2000 yearsago. The "modem" alluvialfan-floodplainlies manym abovesediments
deposited2500 years ago as the Asopus River fan expandedupward,north,andeast acrossthe Malianplain.
The AsopusRiverchannelof 480 B.c. couldhave flowed
in any directionand azimuthover a 1300 are into the
Malianplain. The river of thattime would have flowed
across a much lower slope alluvial fan-riverplain than
that of today. Precise geological determinationof the
location of the river channel(or indeed multiplechannels) is impracticalfor two reasons. First, a very large
numberof drill holes is requiredfor accuratestudy of
the third dimension (subsurface)morphologyof any
time-depositionalsurface. Second, one is unlikely to
encounterorganicmaterialssuitablefor radiocarbondating for each channel and stratumof fluvial sediments
relatedto each of the manypaleo-channelsof the ancestral Asopus River. The Asopus River channel of 480
B.C.could have been anywherebetween the Trachinian
Cliffs andthe northernflankof Kallidromon,morelikely
in the middlethree-fifthsof the river fan.
Therehas been an extremelysharp, "cliff-like"drop
into the sea at the west gate for at least the last 6000
years.Passageto southernGreecewas possiblevia either
the very narrow"wagon-track"
pathalong the sea or by
over
the
climbing
relativelylow-lying (between100 and
250 m in elevation)ridge-likeextension of Mt. Kallidromonin the vicinity of Anthela,thence downwardto
the seashoreby the middle gate, which is borderedby
very high cliffs, thenceby the east gate which againwas
"cliff-like"into deeperwaterswith a very narrowpass.
The topographysouthof the east gate or along the flank
of Mt. Kallidromonat the east gate was of relatively
low slope andcouldbe bypassed(FIG.2). Thus, in studying possibilitiesof passage from northernto southern
Greecevia the pass at Thermopylae,one must focus on
the middlegate along the ancientstrandline.
Exceptionsof lower sea level, with accompaniedwidening of the pass at Thermopylae,possiblyoccurredca.
300 B.C.to A.C.1100 and 1700 B.C.to 1300 B.C.(Livy
36. 16.5).6 Thusgeologicalevidenceshowsthatthe pass
at Thermopylaeshouldbe eliminatedas a majoraccess
route to southernGreece throughoutthe greaterpartof
pre-Romanhistoric and prehistorictimes back to Late
Neolithic-EarlyBronzeAge time.
6. The impacton historicalevents of sea-level fluctuationalong a
precipitouscliff, narrowstrandline,and adjacentsea must have been
profound.At the time of the 480 B.C.battle, sea level was relatively
high and the middle gate narrow.On the other hand, duringthe
invasionof the Gauls in 279 B.C., relativesea level was lower, and
historiansrecordthe problemsof the Athenianfleet in evacuatingthe
surroundedGreekarmydue to the offshoremuddy shoals. By 191
B.C. the middlegate had becomewide enoughto serve as the Roman
armystagingareawhile Antiochusopted to defend the higherridge
at the east gate. At times of peak sea levels we should anticipate
waves againstrockycliffs with no significantpassagepossible;i.e.,
the pass at Thermopylaewould be "closed."Analysis of sea-level
fluctuationas usedhereinis basedon a curvederivedfromthe eastern
Aegean, locally-derivedgeologicalevidencefroma subsidinggraben,
and compatiblehistoricalstatements.Withthis caveatone can reconstructcoastalpaleogeography.
Journalof Field Archaeology/Vol.14, 1987 195
The slope of the topographyabove the west and east
gates is not cliff-like and might have been flanked.The
Spartanleader, Leonidas, may have chosen the middle
gate for his defensive position based on the vicinity of
the low hill, Kolonos, and a wall supposedlybuilt by
thePhokians.In effect, he adopteda positionfromwhich
his troopscould attractthe Persiansinto a slightlywider
portionof the pass and then retreatbehind protective
positionsflankedby the precipitouscliffs of Mt. Kallidromon, Kolonos, and the sea (Gulf of Malia of 480
years. The argumentas to how the PersianImmortals
and othersmade a flankingmovementvia the Anopoea
pathmustconsiderthe probabilitythatthe morphologies
of the areasof the Persianencampmentat the head of
the MalianGulf and the east gate at Thermopylaewere
much differentthan those at present. The differences
may have been so great as to place at least partof the
Persianencampmenton the oppositeside of the Asopus
River from that assumed at present. If so, then the
requirementsthat Herodotus' path of the Immortals
startedwith their crossing the Asopus River, with the
B.C.).
The morphologyof the middle gate varies from perimountainsof the Oetaeanson the right and the Trachiods when waves lapped against the cliffs of Kallidronian mountainson the left, might easily be met by the
as
at
to
6000
4000
and
much
of
the
third
mon,
yearsB.P.
simpleexplanationthatthe AsopusRiverbuiltits alluvial
millenniumB.P.,to the currentplain of 5-km width. In
fan and flowed in channelsover a 1300 spreadof river
all probability,the pass was not more than 20-30 m
flow.
wide during the 480 B.C.battle. During the battle of
Similarly,the reasonfor Anthela'slocation in an indentationbetween a deeper ravine and the west gate
Thermopylaein 191 B.C.,it is possible thatrelativesea
level had begun to dropand the pass at the middlegate
must be reexamined.The presentmorphologyhas little
wouldhave been wider. This may have been the reason
to do with the originalsiting of Anthela.This important
that Antiochusdecided to base his defense at the east
site of the AmphictyonicLeagueof the thirdmillennium
B.P.was located in a highly defensible position, with
gate, forcing the Romansto a disadvantageas they attacked out of the narrowmiddle gate towardthe east
deep watersagainstcliffs at the west gate and the pass
at Thermopylaeprobablyclosed, with waterslappingat
gate. It is furtherpossible that Antiochus, wishing to
avoid the many flankingpossibilities (e.g., the actual
the very foot of Kallidromon'srocky cliffs.
disasterthatbefell him when Catoattackedhis flankand
Finally, we shouldraise the issue of the relativeimdestroyedhis army), fearedentrapmentif he positioned
portanceof the pass at Thermopylae.Whenthe pass was
his armyat the middlegate.
open to trafficit was an easy routeinto southernGreece
In additionto the slow but steady changes due to
(Szemler 1981). Geological and geomorphicevidence,
erosionanddeposition,catastrophictectoniceventshave
however, suggests that the pass was closed for great
andwill continueto affect the grabenembaymentof the
portionsof the last 5000 yearsand even when open was
Malian Gulf. Pritchett(1965) noted a referenceto a
frequentlyvery narrowand marshy.Recentresearchby
tsunamiengulfingthe headof the MalianGulf in the 6th
the Phokis-DorisExpedition of Loyola University of
centuryA.C.Pritchettquotes Finlay (1877: 255): "The
Chicago (EdwardKase, Director) has shown a more
watersof the MaliaeGulf retiredsuddenly,and left the
favoredroute into southernGreece to the west of the
shores of Thermopylaedry; but the sea, suddenlyreAsopusGorge, at DhemaGap near Vardhates.Here an
turningwith violence, swept up the valley of the Spereasily accessibleand wide pathwayled to the Valley of
chios, and carriedaway the inhabitants."
Doris andthencesouthto Atticaor to the Gulf of Corinth
via the pass at Graviaand Amphissa. This route was
Conclusions
alwaysopen, in the geomorphologicalsense, for the past
We have shown that the Gulf of Malia, particularly 5000 years. It was fortifiedin the Gap and at Kastro
the westernpart includingthe Malian plain, is highly
Orias. It is noteworthythat the armies of both Xerxes
and Brennusthe Gaul fought majorbattles in the pass
dynamicand ever-changingin geomorphicform due to
tectonic, catastrophic,and erosional-depositional-sedi- at Thermopylaeand then, having attainedvictory, they
mentologicalprocesses.We havefurthershownthatvery
turnedwest (backwards)and entered southernGreece
high rates of deposition occurredat the head of the
via the DhemaGap. Perhapstheirfear of a Greekarmy
Malianplainandthatup to 20 m of sedimentoverliethe
on theirrearflankoutweighedtheirdesire for access to
site of the battlefieldat Thermopylaeat the middlegate.
the south via Thermopylae.Certainlyfrom the Greek
Similarvolumesof sedimentoverlie the land surfaceof
point of view this makes more strategic sense for rethe Malian plain upon which the Persiansencamped.
peatedlyforcingbattleat Thermopylae.
Therefore,any conclusions based on present geomorThe geological interpretationswe have presented
phologyare suspectunless one takes into accountall of
shouldbe majorfactorswhen consideringthose historic
the physical processes occurringduring the past 2500
and prehistoricevents that occurredat the head of the
196
The Pass at Thermopylae/Kraft, Rapp, Szemler, Tziavos, and Kase
MalianGulf over the past 5000 years. Both catastrophic
andslowergeologicalprocessesoccurredleadingto massive changes in morphologyat the "passes"of Thermopylae and at the head of the Malian plain. The
shorelinehas progradedeastwardinto the Gulf of Malia
at least 15 km within the last 4500 years (FIG.14). Because the alluviumthat formed the Malian plain is so
extensive and so thick, we cannot hope to know the
specifics of any ancient humanworks underthe plain.
We can only state that changes in morphologyand in
flow directionsof the tributarystreamsandthe Sperchios
Riverwere probablythe normratherthanthe exception
throughoutthe last five millennia.
We will end with a brief commenton the natureof
evidence and the philosophy of our approachto the
solution of interdisciplinaryproblems. Historiansand
some archaeologistsapproachproblemsof ancientlandscapesfroma narrativehistoricalline of reasoning.Clasof others
sicists constantlychallengethe interpretations
and attemptto make a "best fit," sometimeswith antipodal answers. Scientists, geologists in particular,and
some archaeologistsprimarilyuse empiricaldata. Sciof facentific methodologycombinesthe interpretation
tual information,hypothesisbuilding, and deductionin
orderto reacha conclusion(sometimesmultipleworking
hypotheses).In this paperwe promotethe integrationof
a multiplicityof disciplinesinto a commonapproachto
the solution of problemsthat addressinterpretationof
ancientlandformsand historicalevents.
Acknowledgments
We wish to acknowledge the Instituteof Oceanographic and Fisheries Research (I.O.K.A.E.) Athens,
Dr. ConstantineVamvakas,then GeneralDirector;the
Phokis-DorisExpedition of the Loyola University of
Chicago;7the ArchaeometryLaboratoryof the University of Minnesotaat Duluth;and staff of the Department
of Geology, University of Delaware, for their participationin this research.
Drillingequipmentand personnelfor a portionof this
study were providedby the NationalInstitutefor Geological and Mining Research (N.I.G.M.I.R.) Athens,
Greece. Mr. John Papadopolous,driller,kindly helped
us completethe drill-corestudies in 1979. We also acknowledgethe help of N.I.G.M.I.R. and the Serviceof
LandDevelopment(Y.E.B.) of Greecein supplyingdetailed drill logs of deeper borings across the lower alluvial plain of the SperchiosRiver. Drs. M. Buzas and
7. An expandedversionof this study,includinggeomorphicanalysis
of the "GreatIsthmusCorridor,"will be publishedas a partof the
finalpublicationsof the Phokis-DorisExpeditionby the LoyolaUniversityof Chicago.
R. Foresterof the SmithsonianInstitutionandDr. F. M.
Swainof the Departmentof Geology, Universityof Delaware,kindlyhelpedwiththe microfaunalidentifications
used for environmentaldifferentiation.Dr. Horst Aust,
Bundesanstaltffir Geowissenschaftenund Rohrstoffe,
Hanover, Federal Republic of Germany, engaged in
lengthy correspondenceand discussion regardingthe
of the artesiansystems in the
lithologic interpretations
and
its
delta floodplain.Very helpful
SperchiosValley
commentsby ProfessorsWilliamFarrandandTj. H. van
Andel on an earlierdraftof this paperwere responsible
for majorimprovements,and JudithD. Holz carefully
edited the final draft. Concepts and rationalefor the
study are detailed in Kraft (1972). Dr. John Morgan,
Departmentof Physics at the Universityof Delaware,
providedhelp regardingthe classical writings as they
bearon this work. And last, we thankMr. TeranceVidal
for his perseverancein trackingdown the source for a
diagramby M. Pouqueville,a matterthathad eludedus
for manyyears.
John C. Kraft is the H. Fletcher Brown Professor of
Geology and Marine Geology at the University of Delaware. His research includes the application of geological principles to the interpretation of coastal
landform changes as related to sea-level fluctuations
and paleogeographic reconstructions of coastal environmental settings. Most of his research concerns the
coastal zones of the U.S. mid-Atlantic region, Hawaii,
Greece, and Turkey.Mailing address: Department of
Geology, University of Delaware, Newark, DE 19716.
George Rapp, Jr., is Professor of Geology and Archaeology, Director of the Archaeometry Laboratory,
and Dean of the College of Science and Engineering,
University of Minnesota, Duluth. He is also a professor at the Center for Ancient Studies, University of
Minnesota, Minneapolis. He has done extensive work
in geoarchaeology in Greece, Cyprus, Turkey, Israel,
and North Africa.
G. J. Szemler is Professor of Ancient Greek and Roman History, Loyola University of Chicago. He has
published books and various articles on ancient Roman
priests, politics, and religion. As Associate Director of
the Phokis-Doris Expedition, he is a co-editor of the
historical section in the Expedition's final publications.
Christos Tziavos is a geologist and micropaleontologist with the Ministry of Industry, Energy, and Technology at the National Centre for Marine Research at
Aghios Kosmas, Hellinikon, in Athens, Greece. His research includes micropaleontology and geology of the
coastal zone and marine environments of Greece.
Journal of Field Archaeology/Vol. 14, 1987
Edward W. Kase is Associate Professor of Ancient
History at Loyola University of Chicago. Since 1974
he has directed the Loyola University Phokis-Doris Expedition, which has been exploring and excavating in
the Isthmus Corridor between Itea and Vardhates in
west-central Greece.
Aust, H., A. Athanassios,H. Flate, R. Miittfeld,N. Kakavas,
W. Stahl, A. Stavrou,N. Tassios, and K. Trippler
1980 "Groundwaterin the Sperkhios Basin, Central
Greece, Results of Hydrogeologicaland Geophysical Investigations in 1970-75," Geologisches
JahrbuchReihe C, Heft 4: 1-26 and map.
Aust, H., N. Kakavas,N. Tassios, and K. Trippler
1984 "BoreholeInvestigationsof the AquiferSystem in
the EasternSperkhiosValley, Greece,"Mitteilungen aus dem GeologischePalaeontologischenInstitutder UniversitaetHamburgHeft 56: 175-184.
Bittner,A., L. Burgerstein,F. Calvert,Fr. Heger, V. Hilber,
M. Neumayr,and Fr. Teller
1880 KiistenliindernDes GriechischenArchipels. Vienna: Kaiserlich-K6niglichen
Hof- und Staatsdrukerei.
Fairbridge,R. W.
1961 "EustaticChanges in Sea Level," Physics and
Chemistryof the Earth4: 99-185.
1981 "HoloceneSea-LevelOscillations,"Striae 14: 131139.
Finlay,G.
1877 Historyof Greece I. Oxford:ClarendonPress.
Gell, Sir W.
1819 The Itineraryof Greece (map). London:Rodwell
and Martin.
Green,P.
1970 Xerxesat Salamis. New York:Praeger.
Herodotus
1927 Herodotihistoriae,ed. C. Hude, 3rd edn. Oxford:
ClarendonPress.
How, W. W., and J. Wells
1928
A Commentaryon Herodotus.Oxford:Clarendon
Press.
Kase, E. W.
1973 "MyceneanRoadsin Phokis,"AmericanJournalof
Archaeology 77: 74-77.
Kase, E. W., and G. J. Szemler
1982 "Xerxes'MarchthroughPhokis,"Klio 64: 357.
1983 "The AmphiktyonicLeague and the First Sacred
War,"in Jan6s Harmatta,ed., Proceedingsof the
VIIthCongressof the InternationalFederationof
the Societiesof ClassicalStudies1. Budapest:Akad6miai Kiad6, 107-116.
Kraft,J. C.
1972 A Reconnaissanceof the Geologyof SandyCoastal
Areasof EasternGreeceand thePeloponnese.College of MarineStudiesTechnicalReport9. Newark:
University of Delaware.
197
Kraft,J. C., S. E. Aschenbrenner,and i. Kayan
1980 "LateHoloceneCoastalChangesandResultantDestruction or Burial of Archaeological Sites in
Greeceand Turkey,"in M. L. Schwartz,ed., Proceedings of the Commissionon the Coastal Environment,Field Symposium,Coastal Archaeology
Session, 24th International Geographical Congress, Shimoda, Japan. Bellingham:Bureau for
FacultyResearch,WesternWashingtonUniversity,
13-21.
Kraft,J. C., S. E. Aschenbrenner,and G. Rapp, Jr.
1977 "PaleogeographicReconstructionsof Coastal Aegean ArchaeologicalSites,"Science 195:941-947.
Kraft,J. C., I. Kayan,and 0. Erol
1980 "GeomorphicReconstructionsin the Environsof
AncientTroy,"Science 209: 776-782.
Kraft,J. C., G. Rapp, Jr., and S. E. Aschenbrenner
1975 "Late Holocene Paleogeographyof the Coastal
Plain of the Gulf of Messenia, Greece, and Its
Relationships to Archaeological Settings and
Coastal Change," Geological Society of America
86: 1191-1208.
1980 "Late Holocene PaleogeographicReconstructions
in the Areaof the Bay of Navarino:SandyPylos,"
Journalof ArchaeologicalScience 7: 187-210.
Leake, W. M.
1836 Travelsin NorthernGreece II. Amsterdam:Adolf
and Hakkert.
Livy
1914 T. Livi ab urbe condita, eds. R. S. Conway and
C. F. Walters.Oxford:ClarendonPress.
MacKay,P. A.
1963 "Procopius'De Aedificis and the Topographyof
Thermopylae,"AmericanJournal of Archaeology
67: 241-255.
Maratos, G. N.
1972
Geology of Greece. Athens: Institute for Geology
and SubsurfaceResearch.
Marinos, G., C. Anastopoulos,N. Maratos,B. Melidonis,
and B. Andronopoulis
1963 Geological Map of Greece, Stylis Quadrangle
1:50,000. Athens:Institutefor Geology and SubsurfaceResearch.
Marinos,G., C. Anastopoulos,N. Maratos,B. Melidonis,B.
Andronopoulis,J. Bornovas,G. Katsikatsos,N. Maragouldakis, and N. Lalekhos
1967 Geological Map of Greece, Lamia Quadrangle
1:50,000. Athens: Institute for Geology and Sub-
surfaceResearch.
Pausanias
1959 PausaniaeGraeciaedescriptio,ed. F. Spiro,reprint
of 1903 edn. Stuttgart:Teubner.
Pouqueville,F. C. H. L.
1805 Voyageen Morde, Constantinopleet Albania, et
t
dansplusiersautresparties, de l'EmpireOthomon,
pendant les annies 1798, 1799, 1800, et 1801.
1835
Paris: Chez Gabon et Compe.
Grdce.Paris:FirmanDidot Frbres.
198
The Pass at ThermopylaelKraft, Rapp, Szemler, Tziavos, and Kase
Pritchett,W. K.
1962 "NewLighton Thermopylai,"
AmericanJournalof
1965
Archaeology 62: 203-213.
"The Battle of Thermopylai in 191 B.C.," Studies
in Ancient Greek Topography Part I. Berkeley: Uni-
versityof CaliforniaPress.
1982a "Appendixon Some RecentCritiquesof the Veracity of Herodotus," in Studies in Ancient Greek To-
pographyPartIV (Passes). Berkeley:Universityof
CaliforniaPress, 234-285.
1982b "Herodotusand his Critics on Thermopylai,"in
Studies in Ancient Greek Topography Part IV
(Passes). Berkeley:Universityof CaliforniaPress,
176-210.
1982c "Passesfrom Thermopylaito Elateia,"in Studies
in Ancient Greek Topography Part IV (Passes).
Berkeley:Universityof CaliforniaPress, 123-175.
1982d "Routeof the PersiansafterThermopylai,"in Studies in Ancient Greek Topography Part IV (Passes).
1985
Berkeley:Universityof CaliforniaPress, 211-233.
"In Defense of the ThermopylaiPass," in Studies
in Ancient Greek Topography Part V. Berkeley:
Universityof CaliforniaPress, 190-216.
Ralph,E. K., H. N. Michael, and M. C. Han
1973 "RadiocarbonDates and Reality," Universityof
Pennsylvania Museum of Applied Science and Center for Archaeology Newsletter 9: 1-20.
Simpson,R. H.
1981 MyceneanGreece. ParkRidge, NJ: Noyes Press.
Stahl, W., H. Aust, and A. Dounas
1974 "Originof Artesian and ThermalWaters Determinedby Oxygen, Hydrogen,and CarbonIsotope
Analysisof WaterSamplesfrom SperchiosValley,
Greece," Isotope Techniques in Groundwater Hy-
drology (InternationalAtomic Energy Agency) 1:
317-339.
Szemler,G. J.
1981 "Thermopylaeand the Great IsthmusCorridorin
Greek and Roman History," Proceedings, 96th Annual Meeting of the American Historical Association, Session 105. Ann Arbor: University of
Michigan,unpaginated.
Tziavos, C. C.
1977 "Sedimentology,Ecology, and Paleogeographyof
the SperchiosValley and MaliakosGulf, Greece,"
unpublishedM.S. thesis, Universityof Delaware,
Newark.
andOstracoda
of RecentForaminifera
1978 "Distribution
fromthe SperchiosRiverDeltaandMaliakosGulf,
Greece," Thalassographica 2: 58-81.
Wallace,P. W.
1980 "The Anopaia Path at Thermopylai,"American
Journal of Archaeology 84: 14-23.
Zamani,A., and H. Maroukian
1979 "A MorphologicalStudy of an Old Delta of the
SperchiosRiver," in-G. A. Kallerges, ed., Proceedings of the Colloquium on the Geology of the
1980
AegeanRegion. Athens:Instituteof Geologicaland
MiningResearch,417-424.
"DeltaicSedimentationof the SperchiosRiver in
Historical Times," Annales Geologiques des Pays
Helleniques 30: 430-440.