On Ship Construction in Antiquity

I& VI HELLENIC INSTITUTE
FOR THE PRESERVATION

OF NAUTICAL TRADITION
1
I
LAMIA 1996
proceedings
edited by
Harry Tzalas
ATHENS 2001
HELLENIC INSTITUTE FOR THE PRESERVATION
OF NAUTICAL TRADITION
6thINTERNATIONAL SYMPOSIUM
ON SHIP CONSTRUCTION
IN ANTIQUITY
LAMIA 1996
PROCEEDINGS
ATHENS 2001
6thINTERNATIONAL SYMPOSIUM
ON SHIP CONSTRUCTION
IN ANTIQUITY
TPOnlZ Vl TROPIS VI
LAMIA, 28,29,30 AUGUST 1996
proceedings
edited by Harry Tzalas
.
Xopqyoi TOU Eupnoaiou Sponsors of the Symposium
Ynoupycio I l o h ~ ~ ~ a pMinistry
ou of Culture
. .
Eupondi~fi'Evouq European Union
0 Afipoq Aap~Cov The Municipality of Lamia
The 6mSymposium on Ship Construction in Antiquity was organized by:
The Hellenic lnstitute for the Preservation of Nautical Tradition,
with the support of:
The 14th Ephorate of Prehistoric and Classical Antiquities of Lamia,
The European Union.
European Commission, action to safeguard and enhance the value of the
European cultural heritage, contributed with its financial support.
Published by the Hellenic Institute for the Preservation of Nautical Tradition

Editor: Harry E. Tzalas
Text preparation: Katerina Poulis
Proof reading: Peter G. Calligas, Charalambos Kritzas, John Phillipson
and Christine Ayoub.
Electronic processing & Design: Angelo Yfantis
Printed by: Editions SCHEMA & CHROMA, 574 00 Sindos, Thessaloniki.
02001 Hellenic Institute for the Preservation of Nautical Tradition
and the individual authors.
The editor would like to thank Peter G. Calligas, Charalambos Kritzas, John Phillipson
and Christine Ayouh for their valuable assistance in the preparation of this edition.
All manuscripts and correspondence should be addressed to: Harry E. Tzalas, 94 Skra str.,
Kallithea, Athens 176 73,tel: 9514291,fax: 9564388,e-mail: hmarine@hol.gr
Cover: General plan of the two Greek shipwrecks of

place Jules-Verne. Drawing by M. Rival.
From Dr. Patrice Pomey's contribution "Les epaves grecques archalques
du Vle S. av. J.-C. de Marseille: epaves Jules-Verne 7 et 9 et Cesar 1.
Used with the kind permission of the author.
THE INTERNATIONAL
ORGANIZING COMMITTEE
President: Harry E. Tzalas
Vice-president: Lucien Basch
Secretary General: Nikos Lianos
Treasurer: A. I. Tzamtzis
Members: David Blackman
Fanouria Dakoronia
Honor Frost
Charalambos Kritzas
Dimitris Papavasiliou
THE EXECUTIVE COMMITTEE

President: Harry E. Tzalas
Vice-president: Elisha Linder
Secretary General: Yiannis Vichos
Treasurer: Stelios Kokios
Members: E. Kakavoyiannis
Periklis Koukis
Nikos Lianos
TABLE OF CONTENTS
Editor's notes
Address of the President of the Organizing Committee
Abd-el Maguid, M. Les Fouilles recentes du Phare d'Alexandrie.
Augerinou, G. NCa ~onoypacpl~a o-rolx~iayla TO apxaio A~pavl
TOU AAKLVOOU. (in Greek).
[New topographical evidence for the ancient
port of Alkinoos, Corfu.] A summary.
Artzy, Michal The Medinet Habu boat depictions: can we
trust Ramses Ill?
Auffray, Danielle Un site maritime proto-archaique dans I'ile de
Paros. A summary. (see also editor's notes on
p. 13).
Avissar, Miriam The representation of two merchant ships on a
Late Roman mosaic floor in Lod (Lydda),lsrael.
Basch, Lucien La voile latine, son origine, son evolution et
ses parentes arabes.
Blue, Lucy An assessment of maritime conditions, coastal
aspects and the suitability of selected second
millennium BC anchorages of the eastern
Mediterranean, in the provision of shelter.
A summary.
Bockius, Ronald A Roman depiction of a war ship equipped
with two catapults?
Bonino, Marco Further steps in the study of the Nemi Ships:
architecture and clues for their reconstruction.
Bouyia, Polyxeni Z~uypa-ranhoiov. (in Greek). [Bridges of
boats.]
Brandon, Chris Jean-Pierre and Anne Joncheray's excavation
of two small Roman coastal crafts on the south
of France. A summary.
Christidis, Vassilios Fireproofing of war machines, ships and
garments.
Christopoulos, M. Ships and trips in the Odyssey.
Coates, John Planking tenons in ancient Mediterranean ships
built shell first.
Dakoronia, Fanouria Further finds from Kynos. A summary (see
also editor's notes on p. 13).
Detorakl, Marina Informations sur la construction navale dans
des documents de la 1'" periode byzantine.
A summary.
Friedman, Zaraza Ship iconography on 'black-and-white'
mosaics of the 1*-3rdcent. AD.
Frost, Honor Anchor look-alikes.
Gianfrotta, Piero The Argo ship in Rome. A summary.
Gillmer, C. Thomas Ships of the 121hDynasty, Egyptian Kingdom,
and their relation to the 1 7 century B.C.E
Aegean ships.
Guillerm, Alain Eperons a bec et eperons trilames.
Gunsenin, Nergis Byzantine shipwrecks discovered around the
Marmara islands (Prokonnessos): points of
departure and probable destinations.
A summary.
Hockman, Olaf The Kynos sea-fighters: exception or rule?
Hornig, Karin Underwater finds of ship and boat models.
Jung, Reinhard Ta nhoia q q Apu66vaq - Z~&q&tq yta &vav
~pcmjpa an6 TOV Kamava. (in Greek).
[The ships of Amydon - Some thoughts on a
crater from Kastanas.]
Kahanov, Yaacov The Byzantine shipwreck (Tantura A) in the
Tantura Lagoon, Israel. Hull construction report.
Kanda-Kitsiou, K. 'Evaq ve60ol~oq ~ p q p TOU
a YMa'i~ou
Atpavtou ~ q apxaiaq
q K ~ p ~ u p a(in
q . Greek).
[A shipshed part of the port of Hylaikos of
ancient Corfu.]
Kapitan, Gerhard Pyramidal and other pierced stone - What
passed through the transverse hole?
Karovic, Gordana Notes about the Roman navigation in the
middle Danube area. A summary.
Kashtan, Nadav The Ship as reality and symbol: how it was
perceived in Hellenistic and Roman Palestine.
Koniordos, V. and Preliminary report on early post Byzantine ship's
Pelekanidou, E. Graffiti found in a cistern south of the
Trigonion tower, Thessaloniki.
Kourkoumelis, D. The Pyramidal stone-anchors: The case of the
wreck of Antidragonera - Kythera. A summary.
(see also editor's notes on p. 13).
Kourtis, Apostolos To nAoio q q Orjpaq: Mta 6Mrl EK~o)(T~ yta TOV
T ~ O K ~ ~O T ~ U K E U ~Ka
TOU ~L q
~ ~TlCJtp0qTa
TOU npupvaiou E~POAOU TOU. (in Greek).
[The ship of Thera - A different interpretation
on the construction and utility of the stern
appendage.]
Koutsouflakis, G. 8.Longboats and tuna fishing in Early Cycladic
Period: a suggestion.
Lambrou-Phillipson,.Suddensealevel changes: causes and
C. & Phillipson, J. consequences. First historical and
archaeological evidence for ice-sheet
decoupling events?
Lehmann, L. Th. Remarks on the hypozoma.
Linder, Elisha Mobility of craftsmen among Greek and
Phoenician shipwrights. A working hypothesis.
A summary.
Lolos, Yannos G. Kaphereus and Kyme: Late Bronze Age
shipwrecks off Eubea. A summary.
Marangou, Christina More evidence about neolithic inland craft
(Dispilio, Lake Kastoria).
Meijer, Fik. A shipbuilding scene on an unpublished relief.
tMorrison, John Identification of rituals and the function of the
vessels of Akrotiri, Thera. A summary. (see also
editor's notes on p. 13).
Murray, William M. The use of catapults in Hellenistic naval
warfare. A summary.
Pomey, Patrice Les epaves grecques archaiques du VIesiecle
av. J.-C. de Marseille: epave Jules-Verne 7 et 9
et Cesar 1.
Pulak, Cemal The Uluburun shipwreck - An update.
A summary. (see also editor's notes on p. 13).
Raban, Avner The enigma of the angular sailing ships in the
Red Sea since the 4'" millennium BCE.
Reinders, Reinder The Coastal landscape between Thermopylae
and Demetrias from a maritime point of view.
Riccardi, Edoardo The Olbia-Sardinia-wreck of the Siciliano.
Rouskas, Yiannis I l p o C k u q Kal TEXVLKE~~ a ~ a o ~ e u f i q
TOU ~ a m o p ~ a v o '~apap~ou'.
u (in Greek).
[Origin and construction techniques of the lake
Kastoria 'karavi'.]
Sleeswyk, A. W. The Lineage of the Triacontor. A verifiable
hypothesis.
Soueref, C. and 0 1 0&oiq q Zapoepa~11q Kai rl Baaaoa.
Mitta, D. (in Greek).
[The gods of Samothrace and the sea.]
Steffy, Richard J. A Mediterranean ship construction database;
dating and classifying shipwrecks by their
hull remains.
Tilley, Alec F. The Numbers in the names of ancient
warships: Some proposed compromise.
Tzahos, Evangelos A trireme on a funerary lekythos.
Tzalas, Harry, E. Two new representations of ancient ships from
Attica.
Wachsmann, S. The INAJCMSjoint expedition to Tantura
Lagoon, Israel: Report on the 1994-1995
seasons of excavations.
Ward, Cheryl Watercraft for heavy transport in ancient Egypt.
A summary.
Wedde, Michael On the role of multi-functional hybrid hulls in
the construction of a narrative of early Greek
ship architecture.
Witt, Richard Ship's music in the ancient world. A summary.
EDITOR'S NOTES
Mrs. Danielle Auffray's communication at the 6m Conference: "Un site maritime

proto-archaique dans I'ile de Paros" was completed with her presentation made at
the 7m Symposium held in Pylos, during August 1999. The text to be published in
TROPIS VII covers both contributions.
Dr. Fanouria Dakoronia made a presentation at the 6m Symposium with the title
"Further finds from Kynos". A communication with the title "New finds from Pthiotis"
was presented at the 7m Conference of 1999. A unified text will be published in
TROPIS VII.
Dr. Dimitris Kourkoumelis' abstract on the "Pyramidal stone-anchors found at

Antidragonera, Kythera" is included in this volume. TROPIS VII will contain an
updated communication on the same topic.
The late professor John Morrison made a verbal presentation referring to the
identification of rituals and the function of the various vessels pictured on the wall
painting of Akrotiri, Thera. The abstract is published in this volume.
An update of the work on the Uluburun Shipwreck was presented at the 7m

Conference held at Pylos by Professor Cemal Pulak. In consequence, his
contribution "The Uluburun Shipwreck" made at the 6mSymposium of Lamia is not
inserted in TROPIS VI. An up-to-date text will be published in TROPIS VII.
Mrs. Goulielma-Kyriaki Avgerinou, Dr. Lucy Blue, Mr. Chris Brandon, Dr. Marina
Detoraki, Prof. Pierro Gianfrotta, Prof. Nergis Gunsenin, Mrs. Gordana Karovic,
Prof. Yannos Lolos, Dr. Elisha Linder, Prof. William Murray, Dr. Cheryl Ward, Prof.
Richard Witt made verbal presentations and as no written text was sent to the editor
we are publishing only the relative abstracts.
OBITUARY
The publication of this volume was nearly complete, when we were

saddened by the announcement of Professor Morrison's death.
John Morrison was an assiduous participant in our symposia. From the first
Symposium held in Piraeus in 1985 to the fifth in Lamia of 1995, this great
scholar made important contributions to all our meetings.
An authority in the field of marine archaeology and oared ships in particular,

two generations of researchers in the field have gained invaluable
knowledge from Professor Morrison's work.
I was honored with his friendship since we first met in April 1983 at the
advisory meeting for the construction of a replica of the 5'h century B.C.
Greek Trireme, at the National Maritime Museum of Greenwich. Since then,
on several occasions I benefited from his wise advice.
John Morrison lived to see his dream, the construction of the Athenian
Trireme, come true. He had the satisfaction of sitting on the trierarch seat at
the stern of <<Olympias,,and sail the waters of Salamis.
We will all miss the soft-spoken, gentle and smiling scholar who marked
marine archaeology for over six decades. When in August 2002 the sessions
of the 8thSymposium on Ship Construction in Antiquity start, John Morrison
will be remembered with affection.
Professor John Morrison and the editor of the Tropis Series, in front of the
Lenormant Marble at the ~ c r o ~ o lMuseum,
is in 1983.
KEIMENO XAIPETIIMOY TOY ADDRESS OF THE PRESIDENT
IlPOEAPOY T H I OPrANnTlKHI OF THE ORGANIZING
EIllTPOllHI K. XAPH TZAAA rIA COMMITTEE MR. HARRY TZALAS
TO 6' N M I l O I I O FOR THE 6mSYMPOSIUM
I&paoptch~a-re,K& Ileptcpsp~tapxa, Your Reverence, Mr. District

KE Nopapxa, KEArjpapx~, Kupi~q Governor, Mr. Prefect, Mr. Mayor,
Kat Kuptot ~ m p o a o n o~t o v Representatives of the local,
~ O ~ ~ T L~KTW~ V~ ,T L U T
TO~T~KC~V, ~ political,
K ~ V military authorities, Dear
apxhv, AyanrlToi ouvs6pot. Colleagues,
Kupi~q Kat Kuptot, Ladies and Gentlemen,
'Orav anocpaotoa nptv Cva wove va When a year ago I decided to

~ ~ O T E ~ V UTqV
U OpyavoTtKfi propose to the Organizing
Ent~ponfiTOU Iupnooiou Apxaiaq Committee of the Symposium on
NauIV'lyt~fiq, ~ q nohq
v ~ q Aapiaq
q Ship Construction in Antiquity, that
oav ~ o n o6t~cayoyrjq~ o vEpya- the 6th Conference be held in the
otwv ~ q 6qqq 6 1 0 p y a v ~ o q q TO , town of Lamia, I was aware that the
C ~ a v ayvwpi<ov~aqOTI wpaia town had no particular relation to
nohq ~ w vAaptCov 6 ~ CXEL v c~xopt- the sea, the ship, or nautical
axdaq p~ ~ q Bahaooa,
v TO tradition.
nhoio, q v a m t ~ rnapa6oq.
j
However, seamanship for the
H vau~oouvqopwq ~ w EAhrjvov v Greeks is not limited to our islands
~ E nEpl~pi<ETal
V m0Uq K ~ T O ~ K O Uand ~ shores. The nautical tradition is
TWV w p h v Kat TWV napahthv paq. deeply rooted in the soul of all the
H vau~oauvq~ i v a paetat pt<wpCvq Greeks including those living in the
oav Cwota m q v quxrj oAov ~ o vhinterland.
EMrjvwv a ~ o p aKat amhv nou <o6v
q v ~v6oxcbpa. Lamia was selected for two reasons:
The first is the unusual
H Aapia ~ n ~ h c yytar l 660 Aoyouq: o concentration of Mycenaean ship
?plhToq &xEl OX&q pE TQV RpUTO-
depictions in the excavation of Dr.
cpaM o u y ~ C v r p w qnapamaosov Fanouria Dakoronia at the site of
nhoiov TOV ~ u K T ) v ~ ~ K C ; ))@OVOV
V OE
Kynos - Livanates, belonging to the
avaa~acpfi~ q qAp. Oavoupiaq jurisdiction of the 14thEphorate of
A ~ K o ~mqv ~ vBCaq L ~ Kuvoq, ~ q qAntiquities of Lamia.
Kotvoq~aq Atpavcrrhv, nou unays- I wanted in this manner, that those
Tat o ~ q v6 t ~ a t o 6 0 o i a~ q 14qq
q who study the ancient
Ecpopeiaq Apxatonj~ovAapiaq. Mediterranean ship, pay tribute to
'HB&Aa p~ TOV ~ p o n oam6 va npj-
ooupe oAot epeiq nou pehE~apeTO the hard work of this untiring
apxaio nhoio, ~ t cplho~tpeq q npo- archaeologist who had the good
onaeeteq ~ q aq~ o u p a m q qa u ~ j q fortune to excavate this unique
ap~atoAoyou,nou E U T U X ~ U Eva hoard of concentrated knowledge
a v a o ~ a q e ta u ~ o vTOV p o v a 6 t ~ o on the prehistoric ships of our
eqoaupo yvhoewv yta Ta npoiaro- ancestors and worked hard so that
pt~a nhoia ~ w vnpoyovwv paq Kat this precious information be made
poxBqae yta va cpeaoouv au-rdq ot promptly available to scholars. This
I K o A u T ~ ~ nhqpocpopieq
E~ m o u q was my way, this is our way, of
pehEq~Cq.'H~avo 6 t ~ o qpou T ~ O - saying: Thank you Fanouria
noq, o 61~oqpaq ~ponoq,va noupe Dakoronia for the precious ships
oAot pa<i: aavoupia Aa~ophvtaoe that you have brought and still
e u ~ a p t o ~ o u pyta e Ta n o h u ~ t p a continue to bring to light.
~ a p a p t anou cpCpvstq Kat ouvexi-
<etq va cpCpvetq aro cphq. The second reason was the
eagerness of the municipal
0 6ehepoq Aoyoq j ~ a vq npoeu- authorities of the town, headed by
pia TUV 6 q p o ~ t ~ hapxhv v ~ q qthe Mayor Mr. George Delis and
noAqq pe ent~ecpaAjqTOV Ajpapxo Deputy Mayor Mr. Dimitris
KO r t h p y o N T E A ~ Kat TOV Papavassiliou, to host our
A v ~ t 6 j p a p x o KO A q p j ~ p q international conference, offering
flanapaotAeiou, va cptAo~evjoouv generously jointly with the 14'h
TO AteevCq Iu~(noato paq, npoocpC- Ephorate of Prehistoric and
povraq yewato5wpa p a g pc ~ q v Classical Antiquities of Lamia the
14q Ecpopeia f l p o ' i o r o p t ~ h vKat pleasant conference hall and all the
KAaoot~hvApxaloT?)Tov TO opaio arrangements. From my position as
~ U V E ~ ~ K L ~~ vKr pO0 ,Kat TqV 0Aq President of the Organizing and
cptho~evia.An6 ~ q OCoq v TOU Executive Committees I express my
n p 0 ~ 6 p 0 Tqq ~ Opya~COTl~fiqsincere appreciation.
E n l ~ p o n j qEK(PP~<W TLq E ~ ~ L K ~ L V E ~ ~
euxaptarieq. It was decided that this Symposium
would be dedicated to the memory
Eixape anocpaoioet e66 Kat noAl1 of Melina Mercouri, that great lady
~ a t p oTO ~ u p n o o t oa u ~ o va eivat whom Greece had the luck to have
acptepwpCvo urqv p v j p q ~ q qas Minister of Culture for many
MeAivaq Mep~oOpq,~ q pcyaAqq q years. We were not, however, aware
EMqvibaq nou eu-rlj~~)oe o ~onoq that our plenary session would
va dxet yta noMa ~ p o v t aYnoupyo coincide with the consecration of
TOU f l o h t ~ ~ o p o u AEV. yvwpi<ape the first monument dedicated to
Melina Mercouri, at the gate of the
opwq o ~ Ba t ouvCntrrre q Cvapcq
TWV E ~ Y ~ U TOU L ~ Vouv~6pioupE Ta
historical Castle of Lamia. I am glad
ano~aAurr@p~a TOU ay6ApaToq q q for this coincidence.
o ~ q v eioo6o TOU L U T O ~ L K O L ~
Kampou q q Aapiaq. Xaipopal yLa I had the advantage of being for
q v oupmwq. many years a collaborator of Melina
Mercouri on matters of nautical
Eixa T V &tClLfl&TLKfi T U Va~ UUVEp- archaeology; we owe her a lot. We
y a m 6 yLa noMa ~ p o v t ape q v aei- owe her the establishment of the
p v q o ~ qMeAiva M e p ~ o u p q .Av International Symposia on Ship
m a 0 6 crrov x6po ~ q apxatoho- q Construction in Antiquity. We owe
yiaq TOU nhoiou, epeiq q q xpwma- her the great exhibition of 1985
pe napa noMa. Tqq ~pwo-rapeTO <<Greeceand the Sea),. Without her
~ E K ~ V ~TOU ~ C ~I E U ~ OTWV U A L E ~ V ~ V
support the two programmes of
Iupnooiwv Apxaiaq Naunqyl~fiq, experimental archaeology:
~ q xpwmape
q ~ q pey6Aq
v C~0eo-q <<KYRENIA11)) and the <<Papyrella),
TOU 1985 <<H Ehha6a KaL q would not have been realized. It is
@6Aaooa),.Xwpiq q v oupnapama- to her enthusiasm that we owe the
uq ~ q ~q E 0a V eixav yivei Ta npo- fact that the Athenian Trireme was
ypappa~an e ~ p a p a ~ ~apxa~oho- ~rjq built in Piraeus by Greek shiprights
yiaq yla q V K ~ T ~ U K ETOU U ~<<KYPH-
~ in cooperation with our English
NElA 11)) KaL ~ q <<llanup&hhaq~~.
q friends of the Trireme Trust. I could
ZTOV~ v 0 0 u o ~ a o p~o q Kat q m q v mention numerous nautical
ayanq q q yLa q v BiAaoaa, ocpei- exhibitions, publications, events that
k ~ KaLaTO YEYOVO~ ~ OTLr) A 0 ~ ) ~ a i ~ fi
were organized by the Hellenic
Tpurjpqq Cytvs m o v nsipata an6 Institute for the Preservation of
'EM~VES p a m d p o ~ qKaL T C I C ~ ~ E ~ ENautical Tradition after 1981 with the
m a E M ~ v Lvspa K ~ pe q v ouvepya- support of Melina. Melina's love for
oia ~ w Ayyhwvv cpihov paq TOU the sea greatly benefited the
TRIREME TRUST. @a~ o p o u o a va research for the ancient ship in
plhfi~ yla ~ noMdq V ~ U T L K &E ~K ~ C - Greece. The least we can do is to
UELS, E K ~ O C J E L ~ KaL ~ ~ 6 q h 6 0 ~ n OLUq honor her memory.
Cylvav an6 TO E M ~ V L KI V O ~LTOL~TO
npomaoiaq NamL~rjqn a p a 6 0 q q For the future of the Symposia I will
an6 TO 1981 Kai pzra pe q v ap&p~- have the opportunity to speak
T oupnapamaoq TOU during the closing session,
Y ~ T o u ~ ~ E~ O U . p&h&Tq presenting concrete propositions.
~ Ao LuT L U ~ O H
yLa TO apxaio nhoio P y r j ~ eK E ~ ~ LThat - these encounters are useful for
o@vq an6 q v ayarfq q q Mehivaq the enrichment of our knowledge of
yLa q 06Aaooa. flq ~ A a ~ m - ravri- o the ancient and medieval ship is
6 ~ p 0 Ep&iq 0 f i p ~ p aT L ~ O U Tq ~ E witnessed by the ever increasing
pvrjpq ~ q q . r ~ TO a ~ C A A O V~ w v number of contributors.
Iupnooiwv paq 0a pou 6 0 0 ~qi
e u ~ a ~ pva i a p~hrjowm q v KaTahq- I welcome the distinguished
W p t a auve6pia, ~avovraqa u y ~ e - scholars who have come to Lamia
~ptpdveqnpo~aaetq.'OTLOL auva- from the five continents, with the aim
vnjaetq paq dxouv anopei xpljatpeq of enriching our knowledge on the
yla TOV E ~ ~ O U T L O ~ OY V ~ O E ~ Vancient ship.
TWV
paq ycpo an6 TO apxaio Kat peaat-
WVLKO ~ECYOVELQKO nhoio, ~ i v ayEy0-
t I welcome our longtime friends -
voq nou TO pap-rupei q oho Kat several are participating for the sixth
aucavopeq O U ~ ~ E T O Xentmq T~ po- continuous time - but also the new
vwv pe aqpavrt~dqava~otvhoetq. contributors. I welcome participants
Odhw va K ~ ~ U O O ~ ~ ~ 1 6 ~from
o uW q
~ U - distant Australia, Egypt, Israel,
K0ljq E ~ ~ L O T ~ ~ n~O O U V6E x~ 0 ~Cyprus,
~ Turkey, Bulgaria, Serbia,
auy~ev-rpweeioljpepa u r q Aapia Austria, Germany, Italy, France,
an6 T L ndvre
~ qneipouq yta va nhq- Belgium, Holland, Great Britain, the
pocpopljaouv Kat va nhqpocpopq- United States and Canada. I
806v yta OTLvdo unapxe! ~ E T L pe K ~ welcome the ever-increasing
Ta nhoia q q Meaoyeiou, aunjq q q presence of our Greek colleagues.
~ o t ~ i 6 anohl~taphv.
q Kahoaopi<w The love for the ship is uniting us
~ouq nahtouq paq cpihouq - noMoi beyondseasandoceans.
oupp~~dxouv yta 6q auvsM cpopa -
a M a Kal ~ o u vdouq
q nou rjpeav Via Before concluding I have to refer to
n p h ~ q Kahoaopi<w
. ~ouq auvd- the absence of a good friend. Prof.
Bpouq an6 q p a ~ p i wAumpahia,
j Octavio Lixa Filgueiras is no more
q v Aiyumo, TO lapaoh, q v Kunpo, among us. We will always
T ~ T , Bouhyapia, ~q remember him, gentle, soft-spoken,
V o u p ~ i a ~q
Ieppia, q v Aumpia, q reppavia, transmitting, the extended
T ~ I-rahia,
V TV TaMia, TO BdAyto, knowledge he had amassed on
T ~ V Ohhav6ia, ~ q Ayyhia, v T L ~Portuguese and other
Hvwpdveq flohi-reieq, TOV Kava6a. Mediterranean vessels.
Kahwaopi<w ~ o u q 'EMqveq auva-
6dAcpouq, q 6Ao aucavopevq I express my thanks to all the
napouaia ~ o u qpaq xaponoiei. H members of the Organizing and
ayanq yLa TO nhoio paq E V ~ V E L Executive Committees, stressing the
ndpa an6 T L 8aAaooeq
~ Kat ~ o u q assistance of Dr. Fanouria
OKE~VOU~.
Dakoronia and Mr. Charalambos
Kritzas. Charalambos Kritzas'
n p o ~ o uKAEiaw npdnet va avacpdpw assistance has been continuous
q v anouaia evoq ~ a h o upaq cpihou, during the last two years of
TOU Kae. Octavio Lixa Filgueiras
preparation and valuable in
nou 6ev eivat nta avapeaa paq. Qa particular for the preparation of
TOV 8upopame navra npao, y h u ~ o -
TROPIS Ill, IV and the Vih volume
that is in preparation.
I must not forget to mention the

assistance of the Symposium
secretary, Miss Katherine Poulis,
Euxap~crrhohouq ~ o u qouva6~A- who silently worked very hard, as
cpouq ~ q q0 p y a v w ~ i ~ T ) q KaL well as the scientific and
EKTEAEO-~LK~)~ En~~ponrjq K ~ V O V T ~ ~administrative personnel of the 14th
c e x w p ~ o ~ rpveia
j yLa ~ q vKa Ephorate of Lamia. My sincere
Qavoupia Aa~ophvLa,Kai TOV KO thanks to all of you.
XapaAapno K p l ~ < a . rka TOV
XapClhapno K ~ I T <npdne~
~ va ~ o v i -
ow OTLq Porj0e~aTOU r j ~ a vouvefiq
~ a e d h qTq 6 t a p K ~ t aTqq ~ L E T O U ~
npoe~o~padaq TOU Iupnooiou, 161-
ai~epau q v C~6o~t-l ~ w nv p a m i ~ h v
TOU TROPIS Ill, IV, KaL Thpa 6 0 ~ -
AEUOU~Epa<iTOV 50 ~ o p oAev . np&-
nsl va 5exaow q v ypappa~CaTOU
Iupnooiou ~ q v Ai6a K a ~ e p i v a
IlouAT) nou aeopupa E K ~ V EnoMa,
q oh0 TO E ~ L O T ~ ~ O V L K OKal
d n ~ KaL
~LOLK~TLn KO p o o w n ~ ~~ oq 14qq
q
Ecpopeiaq A p x a ~ o n j ~ o Aapiaq.
v Zaq
euxapmh 6Aouq and paeouq Kap-
6taq.
LES FOUILLES RECENTES DU PHARE D'ALEXANDRIE
Le Phare d'Alexandrie, ce bâtiment gigantesque, provoqua dès sa

construction l'admiration des anciens et fut rapidement rangé au nombre
des sept merveilles du monde. II a tiré son appellation de l'île de Pharos sur
laquelle il fut érigé et a donné son nom à tous les autres phares.
II s'agit d'un édifice ayant servi la vie maritime durant seize siècles.' C'est
le seul monument, excepté les temples - comme celui de Karnak -, qui a été
en service aussi longtemps, et c'est sans doute en partie pour cela qu'il reste
présent dans notre pensée et dans notre imaginaire.
Le Phare est souvent cité dans les textes anciens et contemporains. On

en trouve de nombreuses représentations sur des décors peints ou sculptés
(dessins, mosaïques, monnaies... ); des lampes prennent sa forme, etc ...
Tous ces documents ont aidé H. Thiersch à donner une reconstitution
graphique de l'architecture du Phare (fig.
Un siècle après sa destruction, le sultan Qaitbay construisait sa citadelle

sur le même emplacement, sur la pointe est de l'ancienne Ile de Pharos (fig.
2)3.Il se servit des pierres du Phare pour construire ce fort.
De nombreux auteurs ont écrit sur le Phare. On trouve des descriptions

de l'édifice dans les récits des voyageurs anciens et arabes. II n'y a pas eu
d'interruption dans la mention du Phare drpuis son érection jusqu'à nos
jours. Même après sa destruction, De Vaujanf a signalé la présence de
vestiges sous l'eau à côté de la citadelle.
Avec l'invention du scaphandre autonome, le Phare d'Alexandrie

redevint un sujet d'actualité. Kamel Abou el Saadats, un des pionniers de la
plongée sous-marine en Egypte avait attiré l'attention sur ces ruines. En
novembre 1962, il avait, avec I'aide de la marine égyptienne, sorti des eaux
une statue colossale dite d'Isis Pharia6.Cette statue se trouve maintenant au
Musée Maritime National d'Alexandrie. Le musée abrite aussi ses autres
découvertes qui ont été renflouées en 1987 : une base colossale et une
double couronne de la basse et haute Egypte (fig. 3)'.
En 1968, l'Unesco a envoyé la spécialiste d'archéologie sous-marine

Honor Frost et le géologue Vladimir Nesteroff pour effectuer le relevé
cartographique des pièces les plus importantes du site. Avec I'aide de
MOHAMED MUSTAPHA ABD-EL MAGUID TROPIS VI
Kamel Abou el Saadat, H. Frost a établi le premier plan du site et a localisé

17 pièces architecturales antiques. C'est en 1975 que l'archéologue a décrit
le site et publié les photos et les croquis des pièces observées sous la mer
(fig. 4)" Depuis cette date les recherches se sont succédées sur ce site avec
l'intervention d'amateurs ou d'aventuriersg.
En 1993, le site se trouvait menacé par la construction d'un mur de béton

qui était censé protéger la citadelle des tempêtes hivernales. Les autorités
égyptiennes ont alors demandé à Jean-Yves Empereur, directeur du Centre
d'Etudes Alexandrines (C.E.A.), de procéder à une fouille d'urgence sur le
site. Une intervention de 6 semaines a été engagée en octobre et novembre
1994. La fouille était placée sous la direction de J.-Y. Empereur et financée
entièrement par l'Institut Français d'Archéologie Orientale (I.F.A.O.).
L'équipe comprenait une quinzaine de plongeurs (archéologues,
dessinateurs, photographes, topographes et plongeurs professionnels)
ainsi que des archéologues du Supreme Council of Antiquities (S.C.A.), le
Conseil Suprême des Antiquités. Le but de cette campagne était de mesurer
l'étendue du site et d'entreprendre un relevé topographique et
photographique avant de mettre à terre quelques-unes des pièces choisies
par les autorités égyptiennes (fig. 5).
Ce site immergé, de 6 à 8 m de profondeur, couvre 2,25 ha et comporte

des milliers de pièces architecturales (colonnes, bases, chapiteaux,
architraves). La plupart sont en granite d'Assouan et certaines de ces pièces
atteignent des tailles colossales et des poids considérables. Ces éléments
architecturaux appartiennent à l'époque gréco-romaine mais aussi à la
période pharaonique, par exemple les colonnes papyriformes (fig. 6).
La sculpture n'est pas absente du site : on trouve des statues colossales

hellénistiques de style pharaonique, et un certain nombre de sphinx portant
les noms des pharaons ; des reliefs représentent des pharaons et le dieu
Ptah (fig. 7).
Une seule inscription grecque a été découverte. Elle se trouvait sur un

fragment de marbre portant les traces de cinq lettres en bronze. En
revanche, on a une abondance d'inscriptions en hiéroglyphes (fig. 8).
Les résultats très importants révélés par la prospection ont encouragé,

d'une part, le S.C.A. à interrompre la construction du mur de béton et,
d'autre part, le C.E.A. et I'I.F.A.0 à demander la mise en place de fouilles
systématiques.
LES FOUILLES RECENTES DU PHARE D' ALEXANDRIE
Dans les années 1995 et 1996, il y a eu quatre saisons de fouilles :

chacune d'elles représentait deux mois de travail sur le site (au printemps et
en automne) et regroupait 20 plongeurs scientifiques et 10 plongeurs
professionnels. Ces fouilles ont été financées par la société Gédéon, la
fondation ELF, la fondation EDF et la société Leica.
Après la prospection du site, il a été décidé de cartographier en détail

tous les blocs présents. Pour ce faire, plusieurs moyens ont été utilisés. Par
mer calme, les blocs ont été topographiés, soit à I'aide d'un théodolite placé
sur le rivage, soit à I'aide d'un GPS (fig. 9)''. Lorsque l'état de la mer ne
permettait pas ces opérations, la triangulation sous l'eau a été utilisée.
Chaque point relevé a été enregistré sur ordinateur pour permettre
l'établissement d'une carte précise du site. Chaque bloc a été gratté à I'aide
de racloirs puis mesuré, dessiné, photographié et mis en fiches descriptives.
Pour les blocs présentant un intérêt architectural ou épigraphique, des
dessins détaillés complètent les banques de données. L'emploi d'une
caméra numérique a permis un travail d'image directement sur ordinateur
(fig. 10).
Comme on l'a vu précédemment, le site était constitué de plus de 2000

blocs. Trente-quatre d'entre eux ont été choisis pour être mis à terre pendant
l'automne 1995 (du 4 au 21 octobre).
A la reprise du travail le 8 avril 1996, deux blocs de plus ont été renfloués
devant le président français Jacques Chirac, des ministres égyptiens et
français et le Gouverneur d'Alexandrie. Les blocs renfloués, statues, blocs
inscrits et décorés, etc., devaient être représentatifs du site sous-marin (fig.
11)".
Tous ces blocs ont fait l'objet d'un traitement de dessalinisation par
bains successifs renouvelés pendant une période de six mois (fig. 12).
Ensuite, le nettoyage mécanique a commencé pour enlever les concrétions
marines (fig. 13). Les blocs en quartzite ont été consolidés par le silicate
d'éthyle (wacker OH). Ces travaux ont été effectués par les chimistes du
laboratoire de conservation et restauration du S.C.A. à Kôm el Dick et les
blocs ont été placés dans un musée en plein air tout proche de leur
laboratoire (à 20 m du Théâtre Romain) de façon à ce que ces restaurateurs
puissent y réintervenir régulièrement (fig. 14).
D'où viennent ces ruines ? C'est la principale question qui se pose.

D'après les sources, on peut leur attribuer deux provenances : une partie
des blocs aurait appartenu au Phare et aux constructions qui se trouvaient
sur l'île détruite à la suite de tremblements de terre consécutifs ; pour ce qui
est des autres blocs, l'auteur arabe El-ldrissi raconte que pour empêcher les
Croisés d'entrer dans le port, le vizir nubien du sultan Saladin a fait jeter
dans la mer en 1167 de notre ère des blocs antiques situés près de la
colonne de Pompée12.
L'abondance des monuments égyptiens sur le site s'explique par le désir

des Ptolémées de décorer leur ville d'éléments empruntés à des sanctuaires
vénérables de I'Egypte pharaonique. En l'occurrence, ces objets semblent
provenir du site d'Héliopolis que Strabon en 25 av. J.-C. décrit déjà comme
un champ de ruines13.
Au fur et à mesure de la fouille l'idée que certains vestiges du site

auraient appartenu au Phare antique s'est peu à peu confirmée. En effet,
après avoir établi un plan du site, il s'est avéré qu'un grand nombre de blocs
de plus de 20 tonnes et de plus de 5 mètres de long présentaient un
alignement. L'un de ces blocs s'était fracturé en deux parties (no 1009 et
1010). Ces deux observations permettent de supposer que ces éléments
faisaient partie d'un édifice gigantesque qui s'est effondré (fig. 15).
II est clair que ces énormes blocs ne font pas partie de ceux qui avait été
déplacés par le vizir du sultan Saladin. En effet, le plus lourd pèse 75 tonnes
et il aurait été laborieux avec les techniques existant à l'époque médiévale
de déplacer un tel poids.
La majorité des blocs du site sont en granite. II est probable qu'ils ont
servi dans la construction du Phare. Même si Strabon décrit ce dernier
comme un bâtiment blanc, je propose d'imaginer qu'un plâtre ou un enduit
blanc recouvraient ces pierres en granite (ce qui est très fréquent dans les
constructions égyptiennes).
Après toutes ces constatations, une question se pose en Egypte : est-il

nécessaire de mettre à terre les blocs submergés ; pourquoi ne pas créer un
parc sous-marin ? Le renflouage des blocs offre le temps, la possibilité de
procéder à des recollages et à des reconstitutions (le cas du sphinx n o
1008a+ sa tête 1008b et les fragments de l'obélisque n o 2001, 2026 a et b).
Le nettoyage des concrétions et la dessalinisation des pièces permet de
compléter la lecture des inscriptions déchiffrées sous l'eau et même d'en
découvrir de nouvelles. Ce travail a pu aider par exemple à dater plus
justement l'obélisque 2001, 2026 a et b : daté sous I'eau de l'époque de

Ramsès II, il s'avère aujourd'hui, après restauration, appartenir à son père
Séti 1"'. L'exposition de l'objet à la lumière naturelle permet en outre de noter
de nouveaux éléments et de compléter notre lecture des inscriptions
(comme pour le sphinx 2002 sur l'épaule et le flanc gauche). De la même
façon on arrive aujourd'hui à mieux interpréter les décors en relief et en
creux. Ces blocs ont ainsi fait l'objet de nouveaux dessins. Enfin, la mise à
terre valorise les recherches effectuées jusqu'à présent, apporte une
connaissance sur l'histoire d'Alexandrie et révèle au public des chefs-
d'œuvre jusqu'alors inaccessibles.
Les travaux ne sont pas terminés, il nous faut de trois à quatre saisons
pour effectuer la cartographie et accomplir les études architecturales. Les
études géophysiques sont nécessaires pour connaître le niveau de I'eau à
l'époque antique. Les archéologues ont l'espoir d'obtenir le permis du
démontage du mur de béton pour accomplir leurs travaux, d'autant plus que
les pièces distinguées se sont trouvées entre les blocs modernes, comme la
tête n o 1999 (fig. 16) et les bases des statues, ou à côté du mur, comme la
statue colossale et le buste. En même temps, les études des experts se
poursuivent à terre pour avoir une interprétation complète du site.
Mohamed Mustapha Abd-El Maguid

Departement d'Archéologie Sub-aquatique
Avenue El Geish, Stanley
Alexandrie
NOTES
1. Le Phare fut commencé sous le règne de Ptolémée I vers 290 avant J.-C. et terminé une
dizaine d'années plus tard sous le règne de Ptolémée II. Aux environs de 700 après J.-C.
sa lanterne est tombée, il fut restauré en 880 et 910, mais le tremblement de terre de I'an
1100 a fait tombé l'étage octogonal. Enfin, il fut complètement détruit par le tremblement
de terre de I'an 1303.
2. THIERSCH, H., Pharos Antike Islam und Occident, Leipzig und Berlin, 1909.
3. C'est l'une des 14 citadelles fondées par Qaitbay sur la Méditerranée pour protéger
I'Egypte.
4. DE VAUJANY, H., Description de I'Egypte, Alexandrie et la Basse Egypte, partie ii, Paris,
1885, p. 40.
5. Une médaille commémorative a été remise à son frère, ainsi qu'à Honor FROST, lors d'une
cérémonie officielle au Consulat Général de France à Alexandrie, le 9 avril 1997.
6. Sur son plan, H. Frost a placé côte à côte les deux statues colossales, suivant les
indications de Kamel Abou el-Saadat. Elles se trouvaient en contrebas de leurs bases
jumelles.
7. Le colonel Mahmoud Sami m'a informé que ces deux pièces se trouvaient à 5 m à l'ouest
du colosse.
8. FROST H., <<ThePharos Site, Alexandria, Egypb,, IJNA 4 (1975), pp. 126-130.
9. En 1979 deux journalistes italiens ont plongé sur le site et publié un article dans IL MONDO
SUMERSO, -11 Faro di Alessandria,,, 1980, pp. 48-52.
Dans la même année des Américains ont plongé aussi sur le site, cherchant le tombeau
d'Alexandre. SCHWARTZ, S.A., Le Projet d'Alexandrie, New York, 1985, pp. 254-5, 267-
284, 290-1, 294-297.
10. L'utilisation d'un GPS différentiel a permis de réduire l'erreur de l'ordre de 1 cm.
11. GRIMAL, N., *Travaux de I'IFAO en 1996,,, BIFAO 96 (1996), pp. 563-570. On corrigera
l'identification erronée, dans ce rapport provisoire, d'un sphinx en grès : il n'appartient pas
à Ramsès II, comme on a pu le croire sous l'eau, mais en fait au pharaon Merenptah.
12. ROWE, A. <BhortReport on Excavation of the G-O-Roman at Pompey's Site,>, BSAA 35
(1942), pp. 132-133.
13. La plupart des monuments inscrits mentionnent des divinités ou un toponyme : il s'agit
toujours d'Héliopolis et de ses dieux (par exemple, les âmes d'Héliopolis, le gouverneur
d'Héliopolis, etc...).
Fig. 1
Fig. 2
MOHAMED MUSTAPHA ABD-EL MAGUIQ TROPlS VI
Fig. 4b
MOHAMED MUSTAPHA ABD-EL MAGWlD TROPES VI
BLOCS RELEVES EN 1994

(30 blocs) j
I
t5
.
-------- - - -.
"
\
',
w
Fig. 10
BLOCS RENFLOUES EN 1995 ET 1996
Fig.
MOHAMED MUSTAPHA ABD-EL MAGUID TROPlS VI
Fig. 12
CARTE DES BLOCS PAR POIDS

_* .
.-f
. ...
..--.
,- -
, ,
> -
. ,, 'Fort Qai-tbay
. 5
- -
. --
--
m
.
:
'
;*.
.=.;
s .-*
. :
.,
;'
a.,
..
u
I
L
I Old
h
51
k
( l x -
Fig. 13
THE MEDINET HABU BOAT DEPICTIONS:
CAN WE TRUST RAMSES Ill?
One of the most cited pieces of art in our world of the ancient Near
Eastern Mediterranean, is the depiction of the naval battle between Ramses
Ill and the "Sea Peoples," which is dated to the first part of the 12* century
BCE. It has become the basis of the data used by historians, archaeologists,
those interested in the Biblical narrative who concentrate on the problems
associated with the Philistines whose first appearance in the local scene is
in this context. This seems to be the earliest depiction of a maritime battle
scene and nautical archaeologists and boat specialists have diagnosed the
details of the vessels in quantitative manner as if the engravers presented
realistic and minute details of this work of art. The question as to the
authenticity of the ship iconography in this scene is the topic of this paper.
Instead of analyzing the details of the ships themselves, data from Papyrus
Harris I was used for the analysis, as is data from newly discovered ship
representations which can be dated to the period.
Among other sources on Ramses Ill we have the naval scene depicted
on the walls of Medinet Habu and Papyrus Harris I which is the written
account of the event, dated to the end of Ramses Ill's reign or shortly
thereafter. The written data is often quoted when discussions concerning the
enigmatic "Sea Peoples" arise. These are the people to whom all the worries
of the end of the 13thcentury BC and the first years of the 12* seem to be
attributed. If we believe these exaggerated and obviously sensational reports
we have to admit that the group called by us the "Sea Peoples" certainly got
"bad press" from Ramses Ill and his scribes.
But can we trust this report? How much of the report was the boasting of
a king who had had his share of problems (Sandars 1978)? Enough is
known about Ramses Ill and his 30 or so years of reign to perceive that he
was beset with serious problems, both externally and internally (Lesko
1992). Not surprisingly, scholars working on the sections dealing with his
northern campaign suggest that parts of the reliefs are probably copies of
earlier materials and earlier Pharaohs (Lesko 1992: 152-153). Well before
Ramses Ill's reign, Egyptian Pharaohs faced problems caused by peoples
from the north and west, including ones who came via the sea. In one
account, dated almost 100 hundred years earlier, Ramses II confronted the
Shardan, as he recounts it, on their warboats (Artzy 1987: 28). Merneptah
met a massive attack of the Libyans and their allies (Lesko 1992). Even
MICHAL ARTZY TROPIS VI
Ramses Ill, according to his own account, faced several attacks of which the
most famous is the one commemorated in the naval battle scene.
Close inspection of the Harris Papyrus reveals that the written account
does not necessarily agree with the scene depicted at Medinet Habu. While
the scene introduces a battle in which the multiple Egyptian boats are all of
one type, the account acquaints the reader with three types of boats said to
have surprised the enemy. The three were the br, the mns and the aha. The
boat depicted in the scene is probably the aha, the war boat, a term already
used by Ramses II in his description of the naval attack by the Shardan
which he so valiantly repelled (Artzy 1988: 184). Likewise, the artist
describes diverse groups of surprised adversaries with their different
attributes, but only one type of a boat, a strange occurrence considering the
supposed dissimilar origins of these invaders. One possible explanation is
that different boats participated in the battle, but only one was chosen to be
represented in the scene. This fact had little to do with the reality of the
battle, but more with an artistic expediency. Could it have been that these
same Shardan, vanquished by the forces of Ramses II (see above), became
the shipwrights who instructed the Egyptian in the art of building the small,
fast and maneuverable aha depicted almost a century later in the Medinet
Habu Naval Scene of Ramses Ill.Thus the "state of the art" boat, the aha, the
war boat, was the one chosen to represent the mighty Egyptian navy in the
scene. Yet another example of the problem associated with using this
account as an historical record is the mention of the fall of the Hittites, of
which, unfortunately, this is the sole report. We are informed that those who
attacked Egypt had already sacked various other important states, among
them the Hittites and Carchemish (Sandars 1978: 119). Of course Rarnses
boasts that these same enemies who had been so successful with the
Hittites were trounced by his own troops. The destruction of Hattusha, the
Hittite capital situated well in the Anatolian plateau, we must remember, is
unlikely to have been carried out by ships. To the Egyptians, the destruction
of a site in coastal Cilicia (a Hittite province) could have meant the whole of
Hatti. Possible signs of Cilicia reverting to its natural maritime associates,
which included Cyprus and the Dodecanese, appear just at that time or a bit
earlier, at the later part of the 13mcentury BC.2The "Sea Peoples" did not
have to exchange their boats for pack animals in order to attack and destroy
it. A much better choice for the final destruction were the semi-nomadic
Kashka (Bittel 1983) who had already caused havoc, time and again, in
Hattusha and its countryside. The reliability of the Ramesside account has
also come into serious question as revealed in its report of the fall of
Carchemish, which has been shown to have continued to exist under the
direct rule of a Hittitle family well after 1200 BC (Hawkins 1988: 102-103). It
is a curious statement, after all, as Carchemish is situated in inland Syria well
beyond the coast and pirates were not its natural enemies. It is also possible
that the report of the fall of Ugarit prior to the naval battle is not to be
completely trusted. Although there is no doubt that Ugarit did fall in the first
part of the 12thcentury BC and was not re-settled (Yon 1992: 11I), the fate
of contemporary coastal sites, such as Byblos, Tyre and Sidon, was
completely different (Caubet 1992: 128-130). Thus a complete destruction of
the coastal sites, as is often presented, was not necessarily the case; on the
contrary, it is probably not the case at all (Caubet 1992: 128-130). Whatever
the date of Ugarit's fall, which is now being debated again, the trade
continued with changes of patterns well into the 12'hcentury BC.
The Ramesside information, a long term counterintelligence, has misled

scholars for many years. Naturally, a written account of such a contemporary
narrator could not be completely discounted, but one has to consider the
circumstances of the period, Ramses' position and his scribe's or his
scribes' ability to accurately relay the news. Thus, in the same way as we use
other boastful scenes and reports of the Egyptian Pharaohs with caution, we
should view Ramses' report in the same manner. As an historical record, it
is problematic.
We will not dwell on the exact date of the battle, its locale or the events
immediately preceding or following it, as this is a separate study. The thrust
of this paper concerns mainly the period which preceded the events
described by the Egyptian scribes by scores of years, and which, we feel,
contributed greatly to the "Crisis Years". The balance of power at the
Levantine coast towards the end of the 14thand most of the 13thcenturies BC
was dominated by two main entities, namely the Egyptians and the Hittites
in Anatolia. There were also other centers, such as Ugarit or Cyprus (or parts
of it) for that matter, which might not have been equal to the two super
powers in military ability, but certainly were economic powers to be
reckoned with. The competition for the markets necessitated management
such as in the construction and maintenance of the sea-going vessels, the
upkeep of open routes and anchorages and the availability of required
merchandise. It also demanded enough manpower, inhabitants of the
economic entities or emissaries procurable at all times for all tasks and
assignments. The upkeep of maritime routes and their outlets was an
arduous chore. It kept the mariners and merchants away from their homes
for very long periods of time (sometimes years), not to mention the grave
dangers associated with maritime travel, even in the few navigable months
MICHAL ARTZY TROPlS VI
of the year. Even the task of the building of sea-going vessels, which we
think of as being of foremost importance was, at times, consigned to others
(Lambrou-Phillipson 1993: 170). Yon mentions the estimates of population
as being between 6,000 and 8,000 urban inhabitants (Yon 1992: n. 2) and
mentions an estimate of Liverani of 10,000 at the end of the Late Bronze Age.
Heltzer (1976) estimates the rural population as being no more than 25,000.
We have to bear these numbers in mind when we consider the trade
networks in which Ugarit was active. Ugarit either paid for or hired others to
fulfil tasks which they did not wish to carry out. These included military
mercenaries for guard duties (Heltzer 1983: 13) as well as most probably
hirelings who partook in their flourishing trade. We meet such maritime
mercenaries already in the el-Amarna letters in a group named: Mi-Shi.3
These people of whom Rib Addi of Byblos complains to the Pharaoh in the
letters, were actually hired mariners from the general area who acted in the
sea around Byblos and Amurru, modern Lebanon, as a form of coast guards
for the Egyptian overlords. It is hard to imagine that Egyptians carried out the
task. It is more likely that as a local, Syrian fringe group the Mi-Shi people
were hired to keep the interests of the Egyptians. But, when the pay was
greater elsewhere, even when offered by the enemies of the Egyptians or
their allies, it was not hard for the Mi-Shi to play the market for all its worth
and take another side. The Egyptians overlords were, after all, far away,
physically and mentally. Interestingly, a boat model found in Byblos looks
much like a Medinet Habu "Sea Peoples" boat without the birds' heads
(Basch 1987: 67). Who were the people who produced this model and for
what use? We obviously do not know. But the general type of boat seems to
have been known in the area of coastal Syria.
But returning to the Ramses Ill maritime battle record: In the past I have
mentioned incised boats which were found in the Carmel Ridge, in close
proximity to the site of Tel Nami. The site was settled in the 13'hcentury and
possibly the first years of the 12mcentury BC. The natural setting, of the
crevice-like opening of the Me'arot River in the ridge, ca. 3.5km from Nami
and the coast, might have been used as a benchmark for the mariners. That
area was also quite conducive for a road inland to Megiddo and eventually
the modern state of Jordan (Artzy 1994 and 1997). Nami was found to be
rich in material goods, luxuries and much bronze, which was right for a spot
connecting sea to desert.
The incised boats are of different types. The most usual shape is the kind
we refer to as the AkkoIKition type with a "fan" (fig. 1). The best
representative of the type are the ones from the Akko altar (fig. 2). The Akko
altar is dated to the end of the 13", beginning of the 12" century BC
stratigraphically. The most impressive boat in the area (fig. 3), as far as size,
depth of incision and the positioning on the rocks which we know so far, is
no doubt, the boat of an Aegean type, well documented from Gazi, Tragana
and Dramesi in Beotia (Basch 1987: 142-147; 1994: 20-21). This type of boat
has also recently been found in records from Teneida, in the Dakhla Oasis
in the Western Delta of Egypt and published recently (Basch 1994). The
Teneida example has more details, which include men holding small ships
(Fig. 4). Basch has already shown that the people standing in it are now, we
know from the Egyptian records, considered to be Libyans. The appearance
of this type of boat in the Western Delta, dating more or less to the Ramses
II and Ill period, 13m- beginning of the 12thcentury BC, comes at a time when
fortresses were constructed to keep the coasts and western borders of
Egypt safe (Habachi 1980). The concurrent appearance of this type of vessel
on the Carmel Ridge should not simply be taken as a coincidence. The small
boat models4held by the mariners on the Teneida boat is of a familiar type,
an outward inclined stem with an animal, probably a bird's head on it.
There might be some jest in the presentation of these men. Their

emphasized nakedness combines with a most elaborate headdress, which
must have been very cumbersome, if indeed they wore one while active.
There may be some exaggeration a little "artistic freedom" in the rendering of
the scenes carried out by the artists involved with the production of this
monument and the one depicted at Medinet Habu. The geographical
position of the scene should be considered. One of the possibilities is that it
was found so far inland because these people were settled quite a distance
from the sea by Ramses II in order to keep them away from mischief and this
would be their tribute to their former glory. This would obviously be much
like the settlement of the "Cilician" pirates by Pompey to keep them out of
harm's way in the first century BC. A completely different explanation is one
proposed by Basch (1994).
A third type of boats appearing on the cliffs of the Carmel Ridge are
similar to the boat models, if that is what they are, held by the men in the
Teneida boat. These are boats with an animal head on their prow, facing
outward (fig. 5).'Curiously, all the examples of these boats we have noticed
thus far, at least the ones in which both protomes are discernable, never
appear with two animal heads. In one case the "head" could be interpreted
as that of a bird, possibly a duck. In the Medinet Habu record the adversary
boats are all represented with two duck protomes. We feel that the artist of
the scene in Medinet Habu may have taken a fancy to such decorations and
decided, with poetic license, to use it in his composition. We can not

disregard the possibility that such boats may have been sighted or reported
by a contemporary and the artist thought it a good eye catcher for the
glorious account of the battle.
Thus, in reconsidering the Medinet Habu Naval Scene, the boats of the
Egyptian adversaries, the ones referred to as "Sea Peoples" join some other
erroneous data which we, until recently, have accepted at face value. The
proportions of the boats, the mariners and their attire should also be used
with di~cretion.~ The artistic representation thus joins the written record
which is not necessarily correct. We should remember that propaganda,
especially that dished out by the State, or in our case, a besieged Pharaoh
of the Pdmillennium, should be read with great care before being used as a
historical record.
Michal Artzy
University of Haifa
31905 Haifa
Israel
NOTES
1. Sandars' book on the Sea Peoples is still a good source for laymen and scholars. Although
we do not agree on several crucial points presented in the study, its usefulness cannot be
overestimated. She says: 'The language has been called "poetical" but is more justly
described as "bombastic". It is a murky substitute for straightforward historical narrative, but
that is something the ancient world never set out to give" (Sandars 1978: 117). It probably
could not have been expressed in a clearer way.
2. Indeed, ceramics which have been identified as Mycenaean IIIC1, or Late Helladic IIIC1,
(associated with the "Sea Peoples" destruction) have been found on the Cilician coast of
Southern Anatolia, but not surprisingly, not in the central part. Although there is little imported
Mycenaean ware associated with the H i l e Imperial levels at Kazanli, Mersin and Tarsus, the
appearance of an Aegean type of pottery in a LBllb level is not necessarily due to invasion.
(Sherratt and Crouwell 1987). Mycenaean-type pottery could well have been produced in
Cyprus or in Eastern Greece, both of which have been natural trade counterparts being
situated on similar maritime networks. When Hittite control weakened, the reversion was a
natural development.
3. Lambdin pointed out that already in the Ebeling glossary of Knudson's edition of the Amarna
letters (Knutdzon 1915: 1550) the identity is questionable. Lambdin proposed the name Mi-
Shi which is to be equated with the Egyptian word msh' "army, troops". They are mentioned
in at least 5 texts: EA 101:4,33; 105:27; 108:38; 110:48, 52 and 126:63). Save-Soderbergh
(1946:60) still called them Mi-Um. It is very likely that these people were hired for their task
as a form of coast guard. It is hard to believe that any real Egyptians carried out that chore
along the Byblos and Amurru coast, south of Ugarit. Altman has already shown the good
relations between the family of Abdi-Ashirta and the Mi-Shi people (Altman 1977:s). He
proposes that their relations with the avowed enemy of Rib-Addi of Byblos, who was,
according to his protestations, a devoted servant of the Egyptian king (for another view on
the letters of Rib Addi see: Liverani 1973), were understandable in view of the corruption
rampant in the Egyptian camp. We would like to propose that the answer lies in the nature of
the employment of this group. As a local, Syrian fringe group the Mi-Shi people were hired to
keep the interests of the Egyptians. But, when the pay was greater elsewhere, even when
offered by the enemies of the Egyptians or their allies, it was not hard for the Mi-Shi to play
the market for all its worth and take another side. This explains their ambivalent relationship
with Rib-Addi and eventually the treatment of Abdi-Ashirta himself.
4. We are not sure that they are models. If the mariners serving as coast guards are pirates,
these vessels might signify their pride in the booty captured by them. The appearance of
these men suggests prowess and strength.
5. C. Lambrou-Phillipson (1996) has shown quite convincingly that the Thera Ships are likely not
accurate representations of vessels and that it is problematic to use them in a quantitative and
diagnostic manner.
BIBLIOGRAPHY
Altman, A. 1977 The Fate of Abdi Ashirta. UF 9:l-10.

Artzy, M. 1987 On Boats and Sea Peoples, Bulletin of the American School of Oriental Research
266:75-85.1988 Development of WarIFighting Boats of the 2* millennium BC in the
Eastern Mediterranean. Report of the Department of Antiquities of Cyprus, pp. 181-186.
1991 Conical Cups and Pumice, Aegean Cult at Tel Nami, Israel. Pp. 203-206 in
Thalassa, the Prehistoric Aegean and the Sea, Aegeum VII, eds. R. Laffineur and L.
Basch, Li&ge: University of Li&ge. 1994 Incense, Camels and Collared Rim Jars: Desert
Trade Routes and Maritime Outlets in the 2nd Millennium, Oxford Journal of
Archaeology 13:121-47.
1995 Nami: A Second Millennium International Maritime Trading Center in the
Mediterranean. Pp. 17-41 in Recent Discoveries in Israel: A View to the West, ed. S.
Giten, New York: Archaeological Institute of America Colloquia and Conference
Papers. 1999 Carved Ship Graffiti - and Ancient Ritual? in Tropis V, ed. H. Tzalas, 21-
29.
Forthcoming: Trade, Boats, Routes and "Nomads of the Sea" in Mediterranean Peoples
in Transition: Thirteenth to Early Tenth Century BCE, eds. E. Stern, A. Mazar and S. Gitin,
Jerusalem.
Basch, L.1987 Le Musee imaginaire de la marine antique. Athens: Hellenic lnstitute for
the Preservation of Nautical Tradition.
1994 Un navire grec en Egypte a I'epoque d'ulysse. Neptunia 195:19-26.
Basch, L. and Artzy, M.
1986 Ship Graffiti at Kition. Pp. 322-344 in Excavations in Kition V, appendix in V .
Karageorghis and Martha Demas, Nicosia: Department of Antiquities.
Bittel, K. 1983 Ship archaeollogische Situation in Kleinasienum 1200v. Chr. und warend der
nachfolgendenvie Jahrhunderte. Pp. 25-50 in Griechenland, die ~ g a i und s die Levante
wahrend der "Dark Ages". Symposium Zweltl 1980, ed. S. Deger-Jalkotzy. SBWien 418.
Coubet, A. 1992 Reoccuptation of the Syrian Coast after the Destruction of the "Crisis Years".
Pp. 123-131 in The Crisis Years: The lr" Century, eds. W. Ward and Martha S.
Joukowsky, lowa: KendallIPark Publishing. Giiterbock, H.G. 1992 Survival of the Hittite
Dynasty. Pp. 53-55 in The Crisis Years: The 1.P Century, eds, W. Ward and Martha S.
Joukowsky, lowa: KendallIParks Publishing.
Habachi, L. 1980 The Military Posts of Ramses II on the Coastal Road and the Western Part of
the Delta. Bulletin de I'lnstitut Fraqais d'hcheologie Orientale 80: 12-30.
Hawkins, J.D. 1988 Kuzi Teshub and the "Great Kings of Karkamis". Anatolian Studies 38:99-108.
Heltzer, M. 1977 The Metal Trade of Ugarit and the Problem of Transportation of Commercial
Goods. Iraq 39:203-211.
1983 The Serdana in Ugarit, Israel Oriental Society 9:9-16.
Knudzon, J.A. 1915 Die El-Amarna Tafeln, Leipzig.
Lambdin, T.O. 1953 The Misi-People of the Byblian Amarna Letters. Journal of Cunieform
Studies 7:75-80.
Lambrou-Phillipson,Connie 1993 Ugarit: a Late Bronze Age Thalasocracy? The Evidence of the
Textual Sources. Orientalia 62: 163-170.
1996 The Reliability of Ships' Iconography: The Theran Miniature Marine Fresco as an
Example. Pp. 351-365 in Tropis W ,ed. H. Tzalas, Athens: Hellenic Institute for the
Preservation of Nautical Tradition.
Lehmann, G.A. 1979 Die Sikalaju - ein neues Zeugnis zu den "Seevolker"Heerfhrten. UF 11:481-
494.
Lesko, L.H. 1992 Egypt in the 12"' Century BC. Pp. 151-156 in The Crisis Years: The 12mCentury,
eds. W. Ward and Martha S. Joukowsky, lowa: KendallIPark Publishing.
Linder, E. 1981 Ugarit: A Canaanite Thalassocracy? Pp. 31-42 in Ugarit in Retrospect, ed. G.D.
Young, Winona Lake: Eisenbrauns.
Liverani, M. 1987 The Collapse of the Near Eastern Regional System at the End of the Bronze
Age in Case of Syria. Pp. 66-73 in Centre and Periphery in the Ancient World, eds. M.
Rowlands, M. Larsen and K. Kristiansen, Cambridge: Cambridge University Press.
Neve, P. 1989 Bogazkoy-Hattusha. New Results of excavations in the Upper City. Anatolica 16:-
90 7-19.
Oren, E.D. 1987 The "Ways of Horus" in North Sinai. Pp. 69-120 in Egypt, Israel, Sinai -
Archaeological and Historical Relationships in the Biblical Period, ed. A.F. Rainey, Tel
Aviv University.
Otten, H. 1983 Die letzte Phase des hethitischen grossreiches nach den Texten. Pp. 13-21 in
Griechenland, die ~ g a i sund die Levante wahrend der "Dark Ages". Symposium ZwetH
1980, ed. S. Deger-Jalkotzy. SBWien 418.
Parkinson, R. and Schonfield, Louise 1993 Akhenaten's Army? Egyptian Archaeology 3:34-36.
Redford, D.B. 1992 Egypt, Canaan, and Israel in Ancient Times, Princeton: Princeton University
Press.
Sandars, N.K. 1978 The Sea Peoples. London: Thames and Hudson.
Sherratt, Susan and Crouwel, J.H. 1985 Mycenaean Pottery from Cilicia in Oxford, Oxford
Journal of Archaeology 6:325-252.
Singer, 1. 1988a Mereneptah's Campaign to Canaan and the Egyptian Occupation of the
Southern Costal Plain of Palestine in the Ramesside Period. Bulletin of the American
Oriental Society 269: 1- 10.
1988b The Origin of the Sea Peoples and Their Settlement on the Coast of Canaan. Pp.
239-250 in Society and Economy in the Eastern Mediterranean, eds. M. Heltzer and E.
Lepinski, Leuven: Orientalia Lovaniensia analecta 23.
Yon, M.1992 The End of the Kingdom of Ugarit. Pp. 111-122 in The Crisis Years: The IPCentury,
eds. W. Ward and Marth S. Joukiwsky, lowa: KendallIPark Publishing.
LIST OF FIGURES
Fig. 1 AkkoIKition .Fan,, type boat from Carmel Ridge.

Fig. 2 AkkoIKition <<Fan.type boats on altar from Akko.
Fig. 3 <<Aegean* type boat from Carmel Ridge.
Fig. 4 Teneida boat (drawing by L. Basch).
Fig. 5 .Bird,, type boat from Carmel Ridge.
THE MEPINIT HABU BOAT DEPICTIONS:
Fig. 1: AkkoJKition Fan type boat from Carmel Ridge
Fig. 2:AkkdKition Far; type boats on alte from Akko
43
Fig. 3
Fig. 5
UN SITE MARITIME PROTO-ARCHAIQUE
DANS L' ILE DE PAROS
Nous avions eu 1' occasion lors d'un précédent symposium d' évoquer
rapidement un site situé à Drios , dans 1' ile de Paros, présentant de longues
entailles creusées dans la pierre le long du littoral que nous presentions
conune de possibles cales à bateaux. Mais la difficulté était que ce site
maritime n' était pas associé a un site archéologique datable. Depuis des
recherches ont été entreprises dans 1' ile (D. Schilardi avec la Société
archéologique de Grèce) ainsi qu' un livre de synthèse sur Paros archaïque
(Danièle Berranger, Université de Clermont Ferrand). Ils ont tous deux
remarqué , outre Drios , 1' existence de plusieurs sites du même type, mais
legèrement ennoyés, situés sur la commune de Naoussa, et associés à des
sites archéologiques terrestres datés du Xème siècle. Nous avions de notre
côté repéré et travaillé sur ces sites que nous présenterons.
Il s'agit de plusieurs ensembles de tranchées taillées dans le roc, de 80-
90 cm de large et pour la plupart d'environ 40m de long, espacées entre
elles d' un mètre environ (avec 1' exception d' un ensemble particulier où des
stries de même largeur s' étendent sur 160m de long associées à des trous
rectangulaires placés sur une ligne parallèle à ces stries)...
Daniele Auffray
Laboratoire d' Histoire Maritime
C. N. K .S.
THE REPRESENTATION OF TWO MERCHANT SHIPS
ON A LATE ROMAN MOSAIC FLOOR IN LOD (LYDDA) ISRAEL
A salvage excavation initiated in April 1996 on Hehaluz Street in Lod led

to the discovery of a mosaic floor. Lod - the Roman Diospolis was a
prosperous town in the Late Roman period, and it seems the upper class
dwellings were situated in this part of the town.
The mosaic floor is 9m wide and more than 17m long. It is of exceptional
quality and in an excellent state of preservation. Discovered about one meter
below the surface, the mosaic floor was covered by debris, which contained
many vividly painted fresco fragments, some fairly large and well preserved.
The numerous pottery fragments found in this debris included imported
amphorae of the third-fourth centuries CE and coins, most of which date
from the third century CE and no later than the fourth century CE. The
mosaic floor probably adorned the reception hall of a Roman villa.
The north and west walls of this hall can be traced along the boundaries
of the mosaic floor; they were terre pisee walls based on a rubble
foundation, and covered by a thick layer of plaster. Part of the base of a white
washed terre pisee wall on the west side of the floor has survived. A fragment
of a massive stonewall on the northeast side cannot be dated with certainty
and its architectural context is unclear. The wall may have contained a
doorway or opening onto the hall. A step added later damaged the east
edge of the mosaic floor; the step was carelessly paved with coarse white
tesserae. This sole repair suggests that the building existed for only a short
time. The rapid crumbling of the terre pisee walls had sealed the floor, thus
preserving the mosaics in their original condition. The southern end of the
floor has not been preserved, due to modern road works, but the rubble
foundation of the southern terre pisee wall has been discovered.
The mosaic floor consists of two colorful rectangular carpets - the

northern and the southern - and a broad transverse band separating them,
all set against a white background. The north carpet is also set into a white
frame containing stylized flowers and bordered by a thin black line. This
carpet is especially large and comprises three panels enclosed in frames of
geometric designs. The north panel consists of nine hexagonal medallions
and two half medallions, all enclosed in a guilloche band and surrounded by
a wave pattern. The medallions contain representations of fish and birds, as
MIRIAM AVISSAR TROPlS VI
well as wild beasts in hunting scenes; some are well known motifs and are
paralleled on North African mosaics of the 3rd-4m
centuries.' A tiger attacking
an onager appears on a mosaic from the "House of the Dionysiac
Procession" at El Jem, North Africa, dated to the second c e n t ~ r y The
.~
northeast medallion, which depicts a basket full of fish, is an exception. A
mosaic with a similar motive comes from a Roman house near Sousse,
T~nisia.~
The square middle panel, which is surrounded by a broad braided band,

contains a twelve-sided polygon, consisting of a central hexagonal design,
surrounded by squares and triangles that are delineated by double guilloche
bands, surrounded by a row of stepped pyramids. In each group of
triangular panels fish and two birds are represented. Two of the square
panels show wild beasts and their prey. The eastern square shows a North
African antelope caught by a panther in a quite unparalleled manner. A
pastoral scene, a hare eating grapes and a hunting dog with unusual leash,
is depicted in the western square and a crater flanked by two felines in the
south square. The central hexagon contains a lion and a lioness confronting
each other from two mountain tops, which are separated by a body of water
in which a mythological sea creature (ketos) frolics. A group of animals - an
elephant, a giraffe, a rhinoceros, a tiger and a water buffalo - is depicted at
the foot of the mountains. A pair of dolphins flanking Neptune's trident
occupy each corner of the square. The elephant is portrayed with a
crisscross pattern to indicate pachydermy, the creature's wrinkled hide.
Elephants with net patterns are known from North Africa, Spain and the
mosaic of the piazza Armerina.4 The artist probably has never seen an
elephant for the animal's hindquarters are wrongly depicted.
A marine scene is shown in the south panel: fish, a dolphin and shells
are scattered in abundance around two merchant ships - one sailing west,
the other at anchor, with folded sails and its mast lowered to the stem-post,
facing east. The scene also includes four unidentified triangular objects,
which seem to be lobster pots. This part of the mosaic floor was damaged
by a cesspit dug in the Ottoman period and obliterated the larger part of one
ship. The two merchant ships seem to be of the same type. The length of the
complete ship is 93cm, its height from keel to gunwale 23cm, the overall
height of the ship 83cm, the length of the mast 60cm, the width of mast at its
base 6cm, mid-ship 23cm, the steering oar length 47cm and the width of the
yard 50cm. It has a simple rounded prow with a curved pointed stem-post
adorned with a square device in red, decorated with circle and cross in
yellow. The ship has an overhanging stern gallery and a goose-headed
sternpost, which is facing the bow. The darker lines on the hull probably
indicate waling pieces. The gunwale is painted red. There are no bulwarks
and the steering oars seem to have been operated by the helmsman from
inside the deckhouse, which is a feature so far not found on representations
of Roman merchant ships. The deckhouse of the ship at anchor seems to be
larger. The depiction of the rigging is quite elaborate. Black lines indicate the
stripes of cloth sewn together to form the square, main sail and the triangular
topsail. The flag on top of the mast is correctly depicted. The second ship is
of the same type but seems to be larger. The owner of this Roman villa might
have been in the shipping business and represented the vessel on the
mosaic for protection as well as to record his po~sessions.~
Marine scenes, where the sea creatures are represented together with
mythological figures, such as Oceanus and Thetys, are known from Antioch
and Jeru~alem.~ But as one would expect Orpheus residing in the central
emblem instead of a Ketos, the marine scene, as a matter of fact, the whole
floor is void of any human figures, a phenomenon, which so far can not be
explained.
A damaged panel with a decorative frame adjoining the north carpet on

the east depicts a tripod table bearing a crater flanked by amphorae. This
panel, whose orientation differs from that of the other panels, may have
faced an entrance in the east wall.
A narrow band, separating the two main carpets, depicts a crater flanked
by two peacocks; tendrils with leaves issue from the crater and birds are
shown among the leaves. Glass tesserae were used in profusion in this
band. Sandwich gold-glass has been used to decorate the crater.
The south carpet is divided into two panels. The north panel, which is
enclosed in a double frame with rhomboids and a braided motive, depicts
birds perched on branches. This panel was damaged when a water pipe was
laid in recent times. The south panel has been only partly preserved. It is
enclosed in a frame composed of a row of diamonds, a braided band and a
bead and reel motive, all set in a plain black frame. Glass tesserae were used
in this carpet as well. This panel shows an intertwining design of guilloche
bands forming hexagonal medallions inhabiting birds, fish and various
animals including two mythological beasts. Glass tesserae were used in this
carpet as well.
The finds, such as coins and pottery as well as stylistic elements point to
MIRIAM AVISSAR TROPIS VI
the end of the 3rdor the beginning of the 4* century CE as the suggested
date for mosaic floor. The floor seems to be the product of a local workshop,
which used patterns from Antioch as well as from North Africa.
Miriam Avissar
Israel Antiquities Authority
P.O. Box 586
Jerusalem 91004
Israel
NOTES
1. K.M.D. Dunbanin, The Mosaics of Roman North Africa, Oxford 1978: PI. XXVIII: 72-73.
2. D. Parrish, A mosaic of a lion attacking an onager, Karthago 21 (1987), Figs. 1-7.
3. S. Aurigemma, Italy in Africa, Rome 1960: PI. 172.
4. C. Salvatore, Mosaics of Villa "Eerculia" in Piazza Armerina - Morgantina: 56-57.
5. Dunbanin 1978: 127 (see above n. 1).
6. L. Roussin, East meets West: the mosaics of the villa of Ein Yael (Jerusalem), in R. Ling
(ed.), Fifth International Colloquium on Ancient Mosaics, Ann Arbor 1995: 36-39, Figs. 7, 10-
11.
LIST OF FIGURES
Fig. 1 General view of the mosaic floor, looking south.

Fig. 2 Northern panel of north carpet.
Fig. 3 The central panel of the north carpet.
Fig. 4 The panel east of the north carpet.
Fig. 5 The marine scene.
Fig. 6 The two merchant ships.
Fig. 7 Detail of band separating the two carpets.
Fig. 8 The bird carpet-the northern panel of the south carpet.
Fig. 9 The southern panel of the south carpet.
ON A U T E ROMAN MOSAIC FLOOR IN LOD (LYDDA) ISRAEL
. ,
Fig. 1
Fig. 2
51
MIRIAM AVISSAR TROPlS VI
I
:,
.
.
.
fig. 3
Fig. 4
52
THE REPRESENTATION QF W O MERCHANT SHIPS
Fig. 5
Fig. 6
53
MIRIAM AVISSAR TROPIS VI
Fig. 7
Fig. 8
Fig. 9
LA VOILE LATINE, SON ORIGINE, SON EVOLUTION
ET SES PARENTES ARABES
On doit à R. Burlet une heureuse définition de la voile latine

<<méditerranéenne,,,c'est-à-dire la seule vraie voile latine : <<il s'agit d'une
voile longitudinale, triangulaire, s'inscrivant dans un plan parallèle à l'axe du
navire, et disposant d'un nombre important de degrés de liberté -
puisqu'elle pivote autour du point qui relie la <<vergue>,au mât. Elle présente
l'avantage des voiles longitudinales de remonter au vent de façon
convenable, et possède des caractéristiques aérodynamiques très
favorables... Cette voile avec son antenne rigide agit en aile dressée
verticalement...>>(Burlet, 1988 : 11-12).
Les opinions relatives à l'origine de cette voile, si importante pour

l'histoire de la navigation en Méditerranée, sont divergentes. Un débat
sérieux s'engage en 1926, lorsque H.H. Brindley écrit prudemment que <<the
Arabian and Mediterranean lateen, whether they evolved separately, or
whether the latter was copied directly from the former (l'hypothèse inverse
n'est pas envisagée), have presumably the square-sail as a common
ancestom (Brindley, 1926 : 17). On observe que les deux types de voiles
sont considérés comme équivalents, ce qui sera le plus souvent le cas au
cours des controverses ultérieures.
En 1932, G.S. Laird Clowes n'hésite pas à affirmer : <<Thetriangular

lateen sail arrived in the Mediterranean in the wake of the conquering Arabs,,
(Laird Clowes, 1932 : 12), sans donner de raisons particulières justifiant son
opinion.
J. Poujade, en 1946, se montre l'avocat ardent de l'origine arabe de la

voile latine (Poujade, 1946 : 140 et S.) ; il écrit notamment : <<Aveceux (les
Arabes), la voile arabe a été introduite dans les ports européens et après leur
départ elle est restée parce que les marins, contraints et forcés de servir
leurs maîtres musulmans, se sont rendus compte de la supériorité de la
nouvelle voile qui <<pinçait>> mieux le vent et permettait d'aller plus vite,,
(Poujade, 1946 : 151). C'est en 1946 aussi que P. Paris publie une étude
pratiquement exhaustive de toutes les sources, tant écrites
qu'iconographiques disponibles à cette époque, susceptibles de fournir la
LUCIEN BASCH TROPlS VI
solution du problème, et cela tant du point de vue arabe qu'européen ;

hésitant, il ne conclut que sur un point d'interrogation (Paris, 1946 ; cette
étude demeure indispensable par la richesse de ses informations).
P.H. Dolley, en 1949, estime qu'il est impossible d'attribuer l'introduction

de la voile latine en Méditerranée 40 the desert nomads who composed the
offensive forces of Islam,, et que les candidats les plus probables à la
paternité de cette invention doivent être cherchés en Egypte ou en Syrie
(Dolley, 1949 : 55).
R. Leb. Bowen, après avoir écrit : <<theevidence seems to indicate that

the Mediterranean lateen was evolved from the Arab lateen,, (Bowen, 1953 :
192), puis : <<there is a chance that the Arabs themselves may have modified
the dipping lug to the triangular lateen. (Bowen, 1956 : 241), conclut : <<since
it is evident that the Arabs have never had the true triangular sail, it is evident
that they could not have introduced it in the Mediterranean. However, I do
not think that anyone doubts that the Arabs spread the triangular lateen sail
westward in the Mediterraneam (Bowen, 1959, 1 : 185 et II : 305, point 9).
Bowen est le premier auteur qui accorde une importance très significative à
la différence entre la voile latine purement triangulaire et la voile arabe
cequasi-triangulaire».On verra qu'il aurait pu pousser sa curiosité plus loin.
Dans un livre important publié en 1963, G.F. Hourani raisonne de la

manière suivante : .the earliest evidence of lateens in the Mediterranean is
in Greek Byzantine manuscripts of the late ninth century. In antiquity... only
the square sail was found in this sea. Thus we are led to suspect that the
lateen came from the Mediterranean in the wake of the Arab expansion,,
(Hourani, 1963 : 103). Hourani ajoute que dans la littérature arabe des IX-t
X" S., une voile vue à grande distance est parfois comparée à la nageoire
d'une baleine, ce qui, selon lui, serait l'indice de l'existence d'une latine
(Hourani, ibid.) ; l'auteur oublie que le sommet angulaire d'une voile peut
être celui d'une voile quadrangulaire. Hourani précise enfin (ibid.) : <<before
that, we have no evidence at all, but at least it can be said that there is no
trace of a square sail anywhere among the Arabs., ce qui n'est qu'à peu
près exact.
R.W. Unger, en 1980, émet une hypothèse originale : «The Arabs may
have got this different type of sail (la voile latine) from the lndian Ocean
where it was introduced by the Romans. The Arabs then diffused the sail
through the Mediterranean,, (Unger, 1980 : 69, n. 20). On se demande
pourquoi les Romains auraient réservé l'exclusivité de la voile latine à
ETSESPARENTESARABES
l'océan Indien.
Dans ce débat, l'iconographie a joué un rôle considérable : on peut voir

une voile latine sur une illustration du psautier dit Kludhoff, du IX" S. (Basch,
1991 : p. 14, fig. 3 A et B) et sur certaines illustrations du manuscrit des
Sermons de Saint Grégoire de Nazianze, datant de 880 (Ms grec no 510 de
la Bibliothèque Nationale, fol. 3 rOet 367 vO; Omont, 1929 : pl. XX et LI1 ;
Antoniadis-Bibicou, 1966, pl. 2 et 3 ; Hourani, 1963 : pl. 5 et 6)'. De la date
de ces représentations, d'aucuns ont été tentés de croire qu'elle constituait
le terminus post quem de l'existence de la voile latine en Méditerranée, et
Hourani, p. ex., en a tiré les conclusions qu'on a vues. En fait, ce terminus
post quem est chancelant : il existe certaines raisons de croire que ces
illustrations pourraient avoir été inspirées de documents plus anciens2et,
surtout, il est hautement improbable que l'illustrateur ait fait appel à des
types de navires au gréement révolutionnaire pour l'époque : ce gréement,
en 880, devait être déjà ((typique., c'est-à-dire déjà largement répandu.
Ces illustrations ont perdu une partie de leur importance depuis la

publication, en 1983, d'une peinture murale décorant une cellule
monastique des Kellia, à 80 km environ d'Alexandrie, datée d'entre 600 et
630 et représentant, sans la moindre équivoque, une voile latine (fig. 1). J'ai
commenté ce document en 1991 (Basch, 1991) et n'insisterai ici que sur
deux points :
1. Le ((quadrillage,>de la voile atteste une parenté certaine avec la voile

carrée ((quadrillée,, de l'époque romaine (très nombreux exemples :
Basch, 1987 : p. 460 et S.). Son origine méditerranéenne est évidente.
II est peu vraisemblable qu'un gréement aussi hautement spécialisé (je

pense aux manœuvres courantes très caractéristiques du gréement latin
(à ce sujet : Burlet, 1988) soit né à la date même de la représentation, ce
qui permet de présumer qu'elle constitue le point final d'une évolution, qui
dut être longue, en Méditerranée orientale. Rappelons la date de l'arrivée
des Arabes en Egypte : 639.
Avant de tenter de retracer les étapes de cette évolution, il convient de

prendre en considération quelques idées fondamentales en matière de
voilure mises en lumière par F. Beaudouin : toutes les structures de
gréement peuvent être ramenées à deux structures de base - ou bien une
surface souple, la voile, est fixée à une perche horizontale (la vergue) hissée
au sommet d'une perche verticale (le mât) (le gréement en T), ou bien cette
LUCIEN BASCH TROPIS VI
surface souple est soutenue par deux ou trois de ses angles au moyen de
perches : gréement en V, technologiquement moins élaboré que le premier
(Beaudouin, 1975 : 8). Le gréement en V fut très largement utilisé par les
marins de la Crète minoenne (Basch, 1987 : 107-114) et il est certainement
à l'origine de la voile à livarde (Basch, 1987 : 119, fig. 218 ; 478, fig. 1078,
1079 ; 479, fig. 1081-1083). Le gréement et T, lui aussi représenté dans la
Crète minoenne (Basch, 1987 : 98 et S.), apte à propulser de grands navires,
devint la voile-type des marines grecque, phénicienne et romaine, reléguant
dans l'ombre le gréement en V3.
La voile latine a-t-elle existé dans l'univers gréco-romain ? A cet égard, il

convient d'abord d'examiner deux textes séparés par cinq ou six siècles, qui
ont été cités plusieurs fois dans ce contexte.
1. Aristote, Mechanica, 851 b : <<Pourquoi (les marins), après avoir navigué

par un vent favorable, lorsqu'ils veulent poursuivre leur route bien que le
vent soit devenu défavorable, carguent-ils la partie de la voile proche du
timonier, mais, serrant le vent de près, laissent-ils la partie de la voile
proche de la proue déferlée et amurée ? La raison en est que le gouvernail
ne peut produire aucun effet contre le vent lorsqu'il est violent, mais qu'il
le peut lorsqu'il est faible, et voilà pourquoi (les marins) réduisent (la partie
postérieure de la voile). Le vent propulse le navire en avant et l'action du
gouvernail transforme le vent en brise favorable en utilisant la mer comme
pivot de levier. En même temps, les marins combattent le vent en
penchant leur corps contre lui>>*.
2. Achille Tatius, Leucippé et Clitophon (Alexandrie, IV" S. ap. J.-C.), 3. 1. 1-

3.
ce... après un beau temps serein, les ténèbres se répandirent soudain et

la lumière du jour s'évanouit. Un vent de mer se leva, par le bas, contre
l'avant du navire, et le pilote donna ordre de tourner la vergue. Les matelots
s'activèrent pour la tourner, d'un côté serrant la voile vers le haut sur la
vergue, de toutes leurs forces, car le vent dont l'attaque se faisait plus
violente contrariait leurs efforts, de I'autre côté, conservant une partie de la
surface avant (souligné par moi) contre laquelle le vent pourrait venir à leur
aide pour la manœuvre. Alors, le bateau, penché d'un côté, s'incline et
s'élève dans les airs de I'autre TOUSles passagers se ruent sur un
bord pour rétablir l'équilibre, puis, à la suite d'une saute de vent, sur I'autre ;
la tempête augmentant de fureur, le navire finit par se briser sur des récifs..
ETSESPARENTESARABES
Le premier texte a été examiné par J.S. Morrison (1968 : 312-3), L.

Casson (1971 : 276-7) et J. Rougé (1979 : 275-6), le second par L. Casson
et J. Rougé (loc. cit.). Ces trois auteurs ont en commun de voir dans ces
passages une illustration d'une manière commune d'allure au plus près
avec une voile carrée. L. Casson, qui illustre, à sa manière, le stade final de
la manœuvre décrite par Aristote (fig. 2), écrit qu'une voile carguée et
orientée de cette façon, 4s not unlike a lateen and may possibly have
sparked the invention of this all-important invention,, (Casson, 1971 : 277).
Pour J. Rougé, dans la description de la manœuvre racontée par Achille
Tatius, «il n'est pas question de mettre le navire à la cape... mais tout
simplement d'utiliser au mieux un vent contraire pour continuer sa route., et
le même auteur estime que le romancier a entendu signifier : <<quele vent
vienne à son aide pour louvoyer,,. J. Rougé ajoute encore : .C'est avec juste
raison que L. Casson... a montré comment la manœuvre consiste à donner
à la voile apiquée vers l'avant une position à peu près parallèle... à I'axe du
navire et à ne garder à l'avant qu'un minimum de toile" ainsi la voilure tend-
elle à prendre une apparence triangulaire, voisine de la voilure a la trina
(latine),,. <<Latine>>
? En ce cas, <<latineinversée,,, ce qui n'a guère de sens.
Les textes d'Aristote et d'Achille Tatius, effectivement apparentés, me
paissent devoir être expliqués d'une manière très différente de celle
proposée par le regretté Jean Rougé.
En premier lieu : que signifie, dans le texte d'Aristote : «décider de

poursuivre sa route en passant d'un vent favorable à un vent défavorable,, ?
Un vent <<favorable>>
(oupioq) est tout simplement, dans la Grèce antique, un
vent qui permet une heureuse traversée7, quelle que soit l'allure du navire,
du vent arrière au bon près, en passant par toutes les allures intermédiaires.
Le vent <<défavorable>>
est celui qui risque de mettre le navire en danger.
Toujours dans le même texte, il est clair que le brusque changement de

vent oblige l'équipage à modifier son gréement de façon importante, s'il veut
poursuivre sa route (il pouvait aussi choisir la fuite - au sens technique du
terme : cette manœuvre était connue, Aristophane (Les Cavaliers, v. 432-3)
y fait très clairement allusion). Nous ignorons quel gréement était établi
lorsque le vent était <<favorable>,, mais il est sûr que lorsque le vent est
devenu «défavorable., la vergue est amenée autant que possible dans I'axe
du navire et que celui-ci serre le vent de très près, une ralingue de chute bien
raidie sur la proue et la partie de la voile située entre le mât et la poupe
carguée. Le navire gîte au point d'exiger une manœuvre de rappel de la part
de l'équipage. Selon L. Casson (mais non dans la lecture de J.S. Morrison)
la vergue aurait immédiatement été apiquée, non pas volontairement, mais
par un effet mécanique. Ni le texte d'Aristote, ni celui d'Achille Tatius ne

mentionnent cet apiquage, mais on peut admettre que l'action du vent sur la
partie déployée de la voile, de même que les poids asymétriques des parties
de la voile que la vergue avait à supporter, avaient pour effet d'entraîner une
légère inclinaison de la vergue vers la proue, mais non au point très exagéré
qu'imagine graphiquement L. Casson (1971, fig. 188 c, ici : fig. 2 c).
Personne ne contestera que le navire, dans la description d'Aristote,

navigue au plus près, et même, très probablement, au plus près serré. Or il
est tout aussi vrai que le centre de voilure a été amené volontairement vers
l'avant, sûrement en avant du centre de dérive. Que se serait-il passé si la
partie carguée de la voile avait été la partie proche de la proue ? Le navire
aurait loffé au point d'engager et de chavirer. J.S. Morrison l'avait bien
compris lorsqu'il écrivait : «The area of canvas abaft the mast is reduced so
that the tendency of the ship to luff, i.e. come up into the wind, is lessened,,
(Morrison, 1968 : 312).
Dans le texte d'Achille Tatius plus encore que dans celui d'Aristote, il est
évident que le vent violent qui se lève brusquement et frappe le navire de
face ne permet aucune allure portante. L'équipage, ici aussi, tente d'adopter
une allure de plus près accompagnée par une avancée du centre de voilure
pour corriger la marche du navire risquant à tout moment de devenir
dangereusement ardent. Dans les deux cas, les marins ont recours à une
méthode qu'ils connaissent bien, celle de mise à la cape courante :
<<manœuvre qui a pour but de mettre un navire en mesure de supporter un
mauvais temps avec le moins de voile possible, de perdre peu sous le
rapport de la route8et de recevoir le choc des lames de la manière la moins
désavantageuse,, (Bonnefoux et Pâris, 1847 : 157). Amener le centre vélique
vers l'avant était le seul moyen d'adopter la cape courante sur un navire à
un seul mât à voile carrée unique, mais il l'était aussi souvent sur des trois-
mâts à gréement bien plus complexe, tel qu'il est figuré sur des ex-voto de
la fin du XVIeet du début du XVIIeS. (fig. 3 et 4) et ultérieurement (fig. 5 et 6)'.
L'interprétation proposée ici des manœuvres décrites par Aristote et

Achille Tatius n'ont certes pas pour but de diminuer leur intérêt, au
contraire : ce sont, à ma connaissance, les seuls textes décrivant une cape
courante dans l'Antiquité. Toutefois, il convient de les écarter d'une
recherche relative à l'origine de la voile latine.
Contrairement aux manœuvres qui viennent d'être décrites, I'apiquage

volontaire d'une vergue et de sa voile n'était pas sans avantages sous
ETSESPARENTESARABES
certaines allures : <sur les galères antiques gréées d'une voile carrée, il
suffisait de I'apiquer pour reporter le centre de voilure sur l'arrière du centre
de dérive afin que le navire devint ardent et remonte facilement au vent),
(Fourquin, 1991 : 431), c'est-à-dire une manœuvre très exactement inverse
de celles que nous venons de voir. Toutefois, il ne suffit pas d'apiquer la
vergue, encore faut-il, surtout sur un navire d'une certaine importance,
maintenir I'apiquage et le régler, au moyen d'un cordage particulier.
Ce cordage, sur la voile latine développée, existe évidemment : il est

frappé sur l'extrémité basse de l'antenne (vergue), proche de la proue, et la
maintient apiquée : la cargue d'avant (ou cargue devant). Or une cargue
d'avant est bel et bien clairement tracée sur trois graffiti de Délos,
représentant un navire marchand (Basch, 1987 : 373, fig. 7 et 9 ; 374, fig. 10)
(fig. 7) ; dus très probablement à la même main, ils sont tracés l'un au-
dessus de l'autre et sont de peu postérieurs soit aux dommages subis par
Délos en 88 av. J.-C., soit à la catastrophe que I'île subit en 69 av. J.-C.1° II
n'est pas douteux que ces navires sont encore pourvus d'une voile carrée,
mais comme ils étaient en mesure de reporter en arrière, pour une durée
plus ou moins longue, un maximum de toile, ils annoncent la voile latine, et
je qualifierais volontiers ce gréement de proto-latin.
Arrien (début de IIeS. ap. J.-C.) rapporte, dans son Périple du Pont-Euxin
qu'Achille apparut à des navigateurs au large de I'île Fidonisi, près des
bouches du Danube E n i TOU imoü ij 6n'â~pouTOU ~ É p o q(Périple, 34 M =
23, 1, 11 - édition de A.G. Roos, Leipzig : Teubner, 1958), c'est-à-dire : <<sur
le mât, au sommet de la vergue,). Corazzini (1907 : 205), en se fondant sur
l'emploi du singulier a ~ p o qconclut
, qu'Arrien pense à une voile triangulaire.
P. Paris (1946 : 78-9) n'estime pas le raisonnement probant, à juste titre : il
peut s'agir tout simplement de la vergue d'une voile carrée apiquée, c'est-à-
dire d'un gréement <<prote-latin)>".
L. Casson a cru pouvoir identifier trois exemples iconographiques de

voile latine antique :
1. La stèle funéraire d'Alexandre de Milet, au Musée Archéologique National

d'Athènes, du II9S. de notre ère (Casson, 1966 : fig. 4 ; 1971 : fig. 181).
Cette voile, de forme quadrangulaire, ne ressemble en rien à une voile
latine méditerranéenne mais tout au plus, en raison de sa longue chute
avant, à une voile au tiers, voile dont on ignore tout dans l'Antiquité (le
relief d'Alexandre de Milet en serait le seul exemple), qui l'a très
,
probablement ignorée12. Je pense, comme F. Moll (1929 : 21), que la
courbure très prononcée de la vergue (son apiquage ne pose aucun

problème) répond, en l'épousant, à l'arcade qui la surmonte : il n'y a ici
qu'une solution pratique à un problème d'esthétique (I'horor vacul].
2. Un graffito incisé sur un fragment de tuile de Thasos représente un navire

doté d'une voile latine «pure>,(Casson, 1971 : fig. 181, d'après Bon, A.-M.
et Bon, A., 1957 : no 2274). Le 23 mai 1967, M. A. Bon m'écrivait que ce
graffito, incisé après cuisson de son support, <<neprésentait aucun indice
qui permette de l'attribuer à une époque précise,,. M. Bon ajoutait qu'il
ignorait le lieu de la trouvaille et soulignait que <<laville de Thasos est
encore très vivante à l'époque byzantine jusqu'au VII" siècle. (cf. Basch,
1971).
3. Une mosaïque du IVeS. de notre ère au Musée Correr de Venise (Casson,

1971 : 244 et fig. 182 ; ici : fig. 8). Un examen personnel de ce document
me fait soupçonner une forte probabilité de restaurations postérieures au
IV" S. ; de toute façon, comme le navire est représenté tournant sa poupe
vers le spectateur, le mât devrait être visible, sur le fond de la voile, alors
qu'on ne distingue aucune trace de ce mât : on ne peut s'appuyer sur un
document aussi douteux. Enfin, l'<<artiste>> a peut-être voulu représenter
une voile carrée apiquée, ou copier maladroitement une telle image.
L. Casson a invoqué une source littéraire, une lettre de 404 de l'évêque

Synesius (Epist. 4) décrivant un voyage vers la Cyrénaïque au départ
d'Alexandrie, au cours duquel, lors d'une tempête, la voile <<habituelle>> fut
remplacée par une autre, c<presumablysmaller. (Casson, 1966 : 49, 1971 :
268-9), ce qui est, en effet, la méthode utilisée par gros temps sur les galères
à voiles latines de Méditerranée. Si le <<presumably smallerm fait problème, le
texte de Synesius mérite néanmoins tout notre intérêt.
Le départ de Synesius d'Alexandrie doit attirer l'attention sur I'Egypte. Sa

marine fut avant tout une batellerie du Nil, ce qui n'interdit nullement l'esprit
inventif. F. Beaudouin a démontré que dès la XVIIIe dynastie au plus tard
I'apiquage des vergues devait être pratiqué sur les navires d'Hatshepsout
figurés dans son temple funéraire à Deir el Bahari, le but de cette pratique
étant de reporter sur l'arrière le centre de voilure (Beaudouin 1976-77 : 53).
Le même auteur proposait la reconstitution de gréement d'un bateau
appartenant à la flottille de Toutankhamon avec les vergues inclinées à 45'
(Beaudouin 1966-67 : 54, fig. 5) ; cette proposition théorique a été
brillamment confirmée par la découverte dans la collection Stéphane Cattaui
d'un talatat d'époque amarnienne provenant d'Hermopolis et représentant
ETSESPARENTESARABES
une barque à vergue unique apiquée à 45' (Vinson, 1993 : 135, fig. 2 a et b ;
Vinson, 1994 : 42, fig. 29).
Dans l'hypogée no 2 d'Anfouchy, à Alexandrie, datant probablement du

IIIeS. av. J.-C., figure une série de graffiti, ou plutôt de dipinti, les dessins
n'étant pas incisés mais tracés au charbon de bois, probablement au le'S.
av. J.-C. Parmi ces dipinti figurent deux navires portant un éperon à trois
lames, caractéristiques de l'époque hellénistique. Un troisième dipinto, tracé
de la même manière, montre une barque pourvue d'une voile latine <<pure,>
(Basch, 1987 : 480, fig. 1084 ; Basch, 1989 : 331, fig. 8) ; s'il est
contemporain des deux précédents, comme la technique de traçage porte à
la croire, il s'agirait de la première représentation d'une voile latine (fig. 9). A
Alexandrie encore, un dipinto datant probablement du VIeS.,tracé sur le mur
d'une habitation, montre un navire dont la voile est très probablement latine :
la vergue est très apiquée, mais le dessin de l'extrémité inférieure de la voile
n'est pas parfaitement clair (Rodziewicz, 1984 ; 221, fig. 250 ; Basch, 1993 :
51, fig. 23). Le navire des Kellia (fig. 1) se situe donc en fin d'une évolution
qui fut très longue en Egypte, où les jalons ne manquent pas. Faut-il en
conclure que c'est à Alexandrie qu'est née la voile latine ? Le texte d'Aristote
et les gréements -prote-latins,, de Délos empêchent toute certitude à cet
égard, mais il est certain que le grand port d'Alexandrie était l'endroit idéal
pour constituer un laboratoire d'expériences navales nouvelles.
En tout cas, après la découverte du navire des Kellia, il n'est plus

étonnant de lire que Bélisaire, en 533, <ditenduire de vermillon les voiles des
trois navires conducteurs de sa flotte sur un tiers environ à partir de l'angle
du haut,, (Procope, Guerre des Vandales, 1, 13, 3). J. Sottas fut le premier à
y voir la preuve de l'existence de la voile latine du VIeS. (Sottas, 1939), mais
P. Paris observa que ces voiles (qui ne pouvaient être des voiles carrées)
pouvaient être des livardes ou des voiles au tiers et que certaines voiles au
tiers de l'Adriatique sont encore teintes de rouge dans l'angle supérieur
(Paris, 1946 : 79). Bowen (1956 : 242) fit la même observation.
La voile à livarde pour un grand navire de guerre est pratiquement

inconcevable. L'argument des voiles au tiers en Adriatique est devenu sans
valeur : M. Bonino a démontré que jusqu'en 1550 le gréement, dans le Delta
du Po, était latin (Bonino, 1978 A), ce que confirme d'ailleurs l'iconographie :
la célèbre vue perspective de Venise de Jacopo de Barbari (1500), qui
montre des dizaines de navires venus de diverses parties de la
Méditerranée, ne représente que des navires à gréement carré ou latin :
aucune voile au tiers.
Celle-ci n'a supplanté la voile latine en Adratique qu'au XVIIeou au XVIIIe

S. (Bonino, 1978 B : 99 ; Marzari, 1984 : 24 ; Marzari, 1988 : 31). La voile au
tiers est, en Europe, une voile tardive (cf. n. 12) et il est à présent certain que
les navires de Bélisaire étaient équipés d'une voile latine. II serait étonnant
que les navires amiraux aient été pourvus d'un gréement différent de celui
du reste de la flotte (92 navires). On peut conclure qu'en 533 la voile latine
était devenue la voile normale des dromons légers - ce qui ne signifie pas
que la voile carrée ait été complètement éliminée de la Méditerranée.
Avant de quitter ici la voile latine méditerranéenne (ou : européenne, afin

d'inclure le Portugal), il est indispensable de souligner l'une de ses
caractéristiques : sur aucune représentation on ne voit la cargue d'avant,
frappée sur l'extrémité inférieure de l'antenne, aboutir ailleurs qu'à
l'extrémité avant du pont (pour le Portugal - je songe aux voiles latines qui
arrivèrent en Inde en 1498 - nous disposons pour la période 1500-1550 des
images suivantes : Barata, 1989, 1 : 21, 230, 241, 242, 266 ; II : 40, 41) ;
jamais I'amure n'est fixée au bout d'une espèce de beaupré dépassant le
sommet de l'étrave, du moins avant le XVIIe s13. II n'y a là rien que de très
normal : dès les gréements (<prote-latins. de Délos, I'apiquage de la vergue
est réglé par une cargue devant aboutissant à l'intérieur de la coque.
Le domaine arabe
La documentation dont nous disposons au sujet de l'histoire du

gréement dans le domaine arabe est bien maigre, alors que le navire fut
toujours important en Arabie : le Coran le mentionne à douze reprises
comme un don de Dieu (11, 164; X, 18; XIV, 32; XVI, 14; XVII, 66; XXII, 65;
XXX,46; XXXI,31 ; XXXV, 1 2 ; XL, 8 0 ; XLIII, 12; XLV, 12).
Si la voile latine méditerranéenne dérive certainement de la voile carrée,

en est-il de même dans le domaine arabe (que je définis ici comme
l'ensemble formé par la mer Rouge, le golfe Persique, l'océan Indien et la
côte orientale d'Afrique) ? On a vu ci-dessus (p. 56) que Hourani estimait
que les Arabes n'avaient jamais connus la voile carrée. Cette assertion doit
être sérieusement nuancée :
1. De nombreux navires romains à voile carrée en route pour I'lnde ont

sûrement été vus par des Arabes, certains ont probablement dû être
capturés. Mais copiés ? La question reste ouverte.
ETSESPARENTESARABES
2. Un navire de la côte orientale d'Afrique, souvent utilisé par les Arabes,

navire <<cousu>> de type archaïque, était propulsé uniquement par une
voile carrée le mtepe (Hornell, 1941 ; Prins, 1965).
3. Les Blemmyes ont-ils exercé la piraterie en mer Rouge ? Selon E.H.

Warmington, les Blemmyes dévastaient la Haute Egypte tandis que les
Himyarites et les Axumites, au IV" S., contrôlaient le commerce maritime
en mer Rouge (Warmington 1928 : 138). Une telle «division du travail,,
était-elle aussi nette ? H. A. Winkler a découvert plusieurs gravures
rupestres de navires Blemmyes dans le désert oriental d'Egypte : tous ont
une voile carrée (Winkler, 1938 : 16 et pl. 111, 3 et 4). Les navires de ces
pillards opéraient-ils sur le Nil ou en mer Rouge ? Le contrôle du Nil
semble avoir été plus facile à exercer par les Romains que celui de la mer
Rouge ; s'il en est ainsi, la voile carrée devait être utilisée en mer Rouge
par les Blemmyes qui, certes, ne sont pas des Arabes, mais qui auraient
pu les inspirer.
4. Un manuscrit de Mardin (Irak), daté de 113415, conservé au Musée de

Topkapi (Istanbul), représente certainement deux navires à voile carrée
(Nicolle, 1989 : 173 et 174, fig. 14 a et b).
5. On verra plus loin que de petites voiles carrées étaient encore utilisées sur
des embarcations primitives en 1950.
II ne semble toutefois pas que l'origine de la voile quadrangulaire arabe

doive être cherchée dans cette direction. Le monde arabe était demeuré, au
cours du XIXeS., remarquablement conservateur : on y rencontrait aussi bien
des monoxyles et des radeaux de mer que de grands navires cousus allant
en Inde. Heureusement, il s'est trouvé un observateur particulièrement
éclairé pour décrire les bateaux arabes qu'il avait minutieusement examinés
en 1837 et 1838 et qu'il publia en 1841 : l'amiral F.-E. Pâris.
e
Le type le plus primitif au point de vue technologique observé par
l'amiral Pâris est assurément celui d'un petit bateau de pêche de Moka
(Pâris, 1841 : pl. 2 ; ici fig. 10) ; il s'agit d'un bateau à gréement à perches.
La voile, longitudinale, a son point d'amure fixé à l'extrême avant. En dépit
de son mode de construction fort sommaire, cette .pirogue», écrit l'amiral
Pâris, <<tientassez bien le plus près», résultat plutôt remarquable, car les
navires grecs de l'époque classique ne l'auraient pas obtenu. II est vrai que
la très difficile navigation côtière en mer Rouge exigeait un dispositif
permettant de naviguer au plus près. Ces deux caractéristiques, voile

quadrangulaire longitudinale et point d'amure à l'extrême avant, vont rester
constantes pendant plusieurs siècles sur les types de navires arabes bien
plus évolués technologiquement. La nécessité de porter le point d'amure
aussi en avant que possible a conduit les constructeurs arabes, soucieux
d'augmenter la surface de la voile, à fixer le point d'amure, non au sommet
de la proue, mais à l'extrémité d'un boute-hors dépassant largement l'étrave.
d'amure, écrit I'amiral Pâris, est fixée à une perche percée au bout, qui,
poussée en avant, lui fait dépasser l'étrave... ; cette pièce de bois sert à
orienter la voile dans toutes les directions, et elle se fixe sur les barrots par
des attaches ; cette méthode est utilisée par tous les navires et tous les
bateaux de pêche arabes,, (Pâris, 1841 : 12 ; les passages soulignés l'ont
été par moi). Cette perche s'appelle en arabe (avec des variations locales)
le dastûr (Tibbets, 1981 : verso de la p. de couverture, no 7 et p. 524 : «part
of the boat resembling a bowsprib, ; Johnstone et Muir, 1964 : 309 ; Grosset-
Grange, 1993 : 45). Le dastûr est déjà présent sur un bateau de pêche,
assemblé par ligatures, de Mascate, d'une construction archaïque (fig. 11)
(Pâris, 1841 : pl. 8), mais il ne s'agit pas de sa forme la plus primitive. R. Leb.
Bowen Jr. a décrit une expérience personnelle à Mukalla, au Yémen, en
février 1950 : des huris (soit des monoxyles aux flancs surélevés d'un
bordage portant un ou deux hommes, soit des bateaux cousus portant deux
ou trois hommes) partaient pour la pêche à la pagaie et uniquement à la
voile au retour. Cette voile était de forme carrée ; au vent arrière, la vergue
pouvait être suspendue au mât soit à sa moitié, soit à un tiers ou à un quart ;
dans les trois cas, l'un des angles inférieurs de la voile était fixé au plat-bord,
l'autre l'étant à l'extrémité d'une pagaie qui s'élançait au dehors et pouvait
être réglée en direction. Mais pour remonter au vent, la pagaie était dirigée
vers l'avant (fig. 12) (Bowen, 1952 : 217). 11 est plus que vraisemblable que
le dastûr est né de cette façon.
Le dastûr a survécu jusqu'à une époque récente dans le golfe Persique

(Oman, 1979 : 133 ; Hawkins, 1977 : 126 ; Howarth, 1977 : 53), sur la côte
orientale d'Afrique (Hawkins, 1977 : 33 ; Jewell, 1969 : x, xvi, 26) et à
Madagascar (Waller, 1965 : 8-9, où l'on voit le dastûr s'étendant jusqu'à la
base du mât, auquel il est solidement fixé).
Si le dastûr a été, jusqu'à une date récente, une caractéristique

constante des navires arabes, la forme de la voile a été variable, tout en
demeurant, jusqu'à une date tout aussi récente, quadrangulaire. Sur le type
le plus archaïque de <<gréementen T, relevé par l'amiral Pâris, le beden
seyad de Mascate (fig. 4), l'envergure égale les 819es de la bordure et la
ETSESPARENTESARABES
chute avant les 6110es de la chute arrière ; il s'agit presque de la forme d'une
voile carrée (qui, en fait, est souvent trapézoïdale), mais gréée
longitudinalement.
Dans son traité de navigation Kitab al-Fawa'id fi usul al-bahr wa'l-qawa 'id
(1488), l'illustre pilote arabe Ahmad ben Majid al-Nadji décrit la forme de la
voile en la comparant au quadrilatère formé par les quatre étoiles de la
constellation de Pégase, deux au nord, deux au sud, ces dernières étant
plus écartées les unes des autres que les premières dans la proportion de
10 : 13 113, alors que la proportion de la chute avant (iawsh) à la chute
arrière (daman) est de 3 : 4 (Tibbets, 1981 : 52, 115, 116) : rien de commun
avec la voile latine méditerranéenne qui, à cette date, est triangulaire depuis
neuf siècles au moins. Par chance, un document figuré datant d'environ 30
ans après la description d'Ibn Majid montre la justesse des proportions qu'il
avait indiquées : on les retrouve sur de nombreuses représentations de
navires arabes figurés dans l'«Atlas Miller,,, réalisé à Lisbonne en 1519
(aujourd'hui à la Bibliothèque Nationale) (De la Roncière et Mollat du
Jourdin, 1984 ; pl. 30 à 32 ; Nicolle, 1989 : 179) (ici, un exemple, fig. 13). Sur
aucune de ces représentations, l'auteur de l'Atlas n'a manqué de montrer le
dastûr. De même, il démontre sa connaissance des grands voiliers arabes
en mettant en évidence la très haute étrave, qui a subsisté jusqu'à une date
très récente (Hawkins, 1977 : 66 ; Oman, 1979 : 126, 127). Le cartographe a
également représenté sur la carte du Brésil (De la Roncière et Mollat du
Jourdin, 1984 : pl. 33, en haut à droite) deux caravelles portugaises à trois
voiles latines : le contraste avec la voilure des navires arabes du même atlas
est tel qu'on est en droit de se demander comment la question de la
diffusion de ce gréement d'est en ouest ou inversement a pu se poser.
Gaspar Correa, qui vit I'lnde 16 ans après sa découverte par Vasco de
Gama, a décrit dans ses Lendas da lndia les navires arabes que les
Portugais ont vus pour la première fois à Angediva, île située au sud de Goa.
Correa confirme Ibn Majid quant à la forme de la voile : la chute arrière de la
voile est, écrit-il, d'un tiers plus longue que la chute avant ; il ajoute qu'un
tiers de la vergue est en avant du mât, les deux autre tiers en arrière. II
précise qu'il n'existe qu'une seule écoute et que I'amure est attachée à
l'extrémité d'une perche presque aussi longue que le mât (<<huma entena,
quasi tamanha como O mastro») - c'est évidemment le dastûr - au moyen
de laquelle les Arabes tirent la voile fort en avant et naviguent à la bouline
(«corn que aponto muito pola bolina...,,). La mention de la bouline est, ici,
surprenante, puisque la bouline ne concerne que la voile carrée, alors que
Correa décrit une voile axiale (il donne même une bonne description du
gambiage de la voile arabe), mais moins qu'il n'y paraît si I'on définit la
bouline comme «une manœuvre courante de la voile carrée, frappée sur la
ralingue de chute, au vent, et servant à la porter au vent, jouant en somme
le rôle du transfilage au mât d'une voile aurique,, (Merrien, 1963 : 103)14.Ne
perdons pas de vue que le beaupré avait initialement pour seule fonction de
porter les boulines aussi en avant que possible (Basch, 1987 : 476, fig. 1074
et 477, fig. 1075), fonction qui n'est pas éloignée de celle du dastûr.
Ce qu'a voulu dire Correa en parlant de bouliner à propos de la tension

vers l'avant d'une voile longitudinale se comprend encore mieux si I'on
considère un navire de pêche existant vers 1918 à Port Saïd et à Suez
(Moore, 1925 : 95, fig. 99) (fig. 14) : il s'agit d'une voile arabe sans dastûr,
mais une amure, que I'on pourrait appeler une bouline, tire la voile vers
I'avant.
La forme de la voile arabe est longtemps demeurée voisine de celle

qu'avaient décrite Ibn Majid et Gaspar Correa : on la voit encore sur une vue
de Mascate publiée en 1680 par Olfert Dapper dans sa Beschrijving van Asië
(Oman, 1979 : 68). A partir du XIXeS., la forme de la voile va se rapprocher
de plus en plus de la forme triangulaire, la chute avant devenant minime
(Johnstone et Muir, 1964 : 326) pour finir par disparaître complètement,
comme en témoigne la photographie d'un sambouk prise près d'Aden en
1967 (Hawkins, 1977 : 64).
En conclusion : il est rigoureusement impossible que la voile latine

méditerranéenne ait été introduite par l'invasion arabe au VII" S., non
seulement parce que cette voile était déjà répandue en Méditerranée
orientale, mais surtout parce la voile arabe dite latine était encore, au XVeS.
très différente. Chacune de ces deux voiles sont le produit d'évolutions
locales qui ont fini par converger complètement au moment où la propulsion
par le moteur les a éliminées toutes deux.
Quelques cas particuliers.
1. Les graffiti de Bet Shearim (Israël).
Le graffito de la fig. 15 (qui date probablement du 2" S. ap. J.-C.) a fait

l'objet d'une étude de M. Pliner, architecte naval, qui a cru pouvoir en
déduire suffisamment de données pour établir les plans d'un modèle, qui
figure dans les collections du Musée Maritime de Haifa (Pliner, 1966 : pl. V,
ETSESPARENTESARABES
2). M. Pliner considérait que la voile constituait un stade de la <slow

transition from square sails to the more effective trapezoidal type,, (Pliner,
1966 : 25). En 1979, W. Mondfeld estimait même que cette voile constituait
une étape vers la voile trapézoïdale arabe (Mondfeld, 1979 : 9-10). En fait, il
s'agit d'un navire relevant d'une architecture romaine courante, dont la
vergue est légèrement inclinée. Pour qui a vu des photographies de la
navigation du Kyrenia II,une inclinaison est familière et n'implique nullement
une voile trapézoïdale.
Par ailleurs, M. Pliner avait bien compris que pour régler cette
inclinaison, il était bon de disposer de balancines, et son modèle est doté
d'un système de balancines très sophistiqué. Or le graffito ne montre
aucune trace de balancine. Ce document doit donc être écarté de toute
discussion sur l'évolution de la voile carrée vers une autre forme de
gréement.
Un autre graffito de Bet Shearim, incisé à l'entrée d'un complexe de

tombes du V" S. de notre ère (fig. 16) pourrait être invoqué dans les
controverses relatives à l'origine de la voile latine. En effet, au-dessus de la
coque sont représentés un vase ainsi qu'une figure plus ou moins
triangulaire (fig. 16 b, ABC). A. Ben Eli a émis une hypothèse selon laquelle
le vase symboliserait le transport d'huile et de vin de Palestine vers la
diaspora juive, afin de respecter certaines prescriptions de la Mishna et du
Talmud (Ben Eli, 1969-71 : 89). A ma connaissance, la figure ABC n'a pas
encore fait l'objet de commentaires publiés, mais il est évident que la
tentation de la considérer comme une voile latine est compréhensible.
Toutefois, les raisons de résister à cette tentation sont les plus fortes :
- II n'est pas douteux que l'extrémité droite représente une proue droite
et inclinée et l'extrémité gauche une poupe recourbée très classiquement.
Le triangle ABC serait une voile latine posée à l'envers sur la coque.
- Les traits représentant le vase et la figure ABC sont nettement plus

minces et incisés moins profondément que les autres traits : il n'est pas
certain que tous les traits soient dus à la même main ; or les incisions qui
représentent le mât et les quatre cordages qui descendent de son sommet
sont particulièrement profondes ; on voit mal pourquoi la <<voile,>,élément
essentiel du gréement, n'aurait pas été gravée aussi profondément.
- Les courbures prononcées au voisinage de B (l'une, convexe, au-

dessus, l'autre, concave, au-dessous) sont, dans le cas d'une voile latine,
aussi inexplicables que le trait D-B ; en revanche, la figure pourrait s'identifier

avec une palme courbée par le vent, dont le trait D-B serait la nervure
centrale ; la palme est connue <<universellementcomme un symbole de
victoire, d'ascension, de régénérescence et d'immortalité» (Chevalier et
Gheerbrant, 1982 : 724). Or le graffito a été trouvé dans un contexte
funéraire.
2. Le graffito d'el-Auja.
La ville appelée en arabe d'el-Auja ou el-Aujeh el-Hafir (puits tortueux),

actuellement Nitsana (fleur en bouton, en hébreu), en Israël, à 50 km à vol
d'oiseau d'el-Arish, près de la frontière avec I'Egypte, possède les vestiges
d'une église byzantine du V" S. sur laquelle était gravé un graffito
représentant un navire à gréement latin méditerranéen, relevé par E.H.
Palmer (1871 : pl. face à p. 29), aujourd'hui disparu (fig. 17). L. White écrivait
en 1940 qu'il s'agissait probablement de la plus ancienne représentation
d'une voile latine (White, 1940 : 145), ce qui n'est qu'une possibilité. Le
terminus post quem étant assuré, quel peut être le terminus ante quem ? Un
indice important est l'espèce de plate-forme (de guet ?) en forme de petite
cage, à l'avant : on retrouve une plate-forme analogue sur un relief de
Narbonne d'époque impériale (Espérandieu, 1907 : 422, no 687) (fig. 18).
Certes, on rencontre une plate-forme semblable sur un relief de la Tour de
Pise (XIIIeS.) (Landstrom, 1963 : fig. 227), mais il me paraît peu probable
qu'un graffito ait encore été tracé au Moyen Age sur une église dévastée au
plus tard quelques années après la conquête arabe de la région en 634 ;
aussi suis-je tenté de croire que le graffito ne peut guère être postérieur à
cette date, s'il ne lui est pas antérieur.
3. La fresque d'Eboda.
Avdat, en Israël (en arabe : Abda ou Abden) est l'antique Eboda, cité
fondée au IIeS. av. J.-C. par les Nabatéens. Située au croisement des routes
de Petra, d'Eilat et de Gaza, elle fut un centre commercial important, qui
déclina lorsque les Romains construisirent la route Eilat - Damas. Elle connut
une nouvelle prospérité à partir du règne de Dioclétien qui atteignit son
zénith vers 550 ; au cours des années 618-620, l'invasion perse y mit fin.
Conquise en 634 par les Arabes, elle fut abandonnée peu après.
La citadelle d'Eboda était entourée d'hypogées, dont l'un, exploré en

ETSES PARENTESARABES
février 1904, révéla une paroi revêtue d'un ((enduitépais (d')une teinte crème
sur laquelle s'enlèvent, très claires, les figures tracées en rouge brique à peu
près exclusivement au trait avec quelques bavures de la couleur et de teinte
pleine,, (Jaussen, Savignac, Vincent, 1905 : 81 - la fresque, longue
d'environ 3 m., est reproduite, d'après une aquarelle, aux pl. VI et VII). II n'est
pas sûr que la fresque <<raconte>> une histoire cohérente, mais la grande
majorité de ses éléments évoquent le désert et l'activité caravanière :
chameaux bâtés, palmiers, gazelle, cavalier, lévrier, guépard (?), étoiles et
constellation (?), puits (?) ainsi qu'un buffle. Deux navires allant en sens
inverse y sont représentés ; chacun d'eux est remorqué (son epholkion ou
epholkis ; cf. Casson, 1971 : 248, n. 93), mais néanmoins eux-mêmes sous
voile (même phénomène sur le graffito du navire Europa, à Pompéi : Basch,
1987 : 469, fig. 1051 ; aussi : Tsaravopoulos, A., 1996 : 502, fig. 502) pour
un epholkion remorqué, mais sans voile).
Tous ces bâtiments sont pourvus d'une voile latine méditerranéenne :

l'extrémité inférieure de l'antenne reste à l'intérieur du navire et on ne voit
aucune trace d'un dastûr. Les deux grands navires, représentés en section
longitudinale, sont pontés, et sous le pont quatre rameurs s'activent. La
proue et la poupe de ces navires sont surmontés d'un personnage ; trois
d'entre eux font un geste de salut. Le navire de droite a une tête d'oiseau
comme figure de proue.
II est impossible de dater avec précision ce précieux document,

aujourd'hui perdu, mais il est difficile de le situer en dehors de l'époque de
prospérité d'Eboda, due au carrefour de routes commerciales régulièrement
actives, au milieu du VIeS. II ne serait pas étonnant que ces navires aient eu
Gaza pour port d'attache. Certes, il ne peut être exclu que des navires
semblables aient été construits à Eilat, mais dans ce cas, leur gréement n'a
eu aucune influence dans le domaine arabe.
4. Le graffito de Malaga (fig. 20).
Un graffito provenant de Malaga, aujourd'hui au Musée Naval de Madrid,

représente une galère propulsée par une très grande voile latine ; dans une
histoire de la marine espagnole, Enrique Manara Regueyro l'attribue à
l'époque romaine (Regueyero et al., 1981 : 15). Les autorités du Musée
Naval, que j'ai interrogées, ne m'ont répondu que de manière vague. R.C.
Anderson (1962 : pl. 8 B et p. 53) I'attribue, à raison me semble-il, à une main
arabe de la fin XIVeS. : une voile latine aussi développée me paraît dépasser
les possibilités technologiques de l'époque romaine.
Retour à la Méditerranée, après le VIIeS.
L'iconographie, du XIe au XIIIe S. ne montre en Méditerranée que des

voiles latines, comme si la voile carrée en avait été complètement chas~ée'~.
Le gréement antique aurait-il été si médiocre ? Sûrement pas.
Les Romains, avec leur voile carrée dont la surface pouvait être, à partir
du pont, diminuée à volonté grâce à son système d'anneaux de cargue
disposés sur la face avant de la toile (Poujade, 1946 : 129 et s ; Casson,
1971 : 275-6 ; Basch, 1987 ; 460), disposaient d'un excellent moyen de
propulsion ; ce système permettait de sélectionner très aisément la partie de
la voile qu'on voulait réduire, ce qui cessera d'être vrai pour la voile carrée
<<moderne,>. Par ailleurs, on n'a pas suffisamment prêté attention au fait que,
sous l'Empire, le rendement de cette voile fut grandement amélioré par
l'usage de la bouline (La Roërie, 1956 : 248-9 ; Basch, 1987 : 477).
Au IIeS. av. J.-C. déjà, Nicandre compare la reptation de la vipère à

cornes à la marche en zig-zag d'un bateau, ce qui implique la connaissance
du louvoyage (Casson, 1971 : 275, n. 19). La faiblesse de cette voilure est
qu'elle ne permettait pas de serrer le vent au plus près. Paul Adam a fort bien
résumé les avantages et les inconvénients du gréement latin ; parmi les
avantages :
- <<Legréement latin permet de serrer le vent mieux que le gréement

carré, la différence étant d'environ un quart (à peu près 11 ce qui est
O),
appréciable, mais est loin de condamner le gréement carré,, (Adam, 1970 :

205).
- Contrôlée du pont, la voile latine <<explique

en partie son utilisation sur
les bateaux étroits (les galères en particulier) où des voiles carrées de
mêmes dimensions auraient été d'un maniement délicat ou impossible,,
(Adam, ibid).
- <<Le gros avantage de la voile latine est qu'en cas de coup dur, il suffisait
de filer l'écoute... la voile se déventait d'autant et, à la limite, se mettait en
drapeau... manœuvre impossible avec une voile carrée,, (Adam, 1970 : 218,
d'après une observation du Commandant Denoix).
ETSESPARENTESARABES
Le grand inconvénient de la voile latine était la nécessité, en cas de

changement d'amures, de devoir gambier, c'est-à-dire d'amener au pont
I'antenne et la voile et de les mettre à nouveau en place de l'autre côté du
mât, manœuvre longue et difficile par bonne brise (Adam, 1970 : 206) ; sur
de grands bâtiments, le poids de I'antenne exige, pour cette manœuvre, un
équipage important. P. Adam conclut justement : <<Leprogrès introduit par
la voile latine était à double sens. D'un côté, il rendait le travail des marins
plus dur, mais de I'autre, il ne demandait pas de compétences trop
spécialisées... l'avantage ne devait pas tellement être sur le rendement mais
bien plutôt sur le degré beaucoup moindre des qualités manœuvrières
exigées de I'équipage. Le patron ordonnait, les matelots donnaient du
muscle. La voile latine simplifiait le travail de I'équipage tout en le rendant
plus dur,, (Adam, 1976 : 562). Cependant, pendant deux siècles au moins,
les avantages de la voile latine l'ont emporté, de loin, sur ses inconvénients.
L'une des raisons de cette situation, d'après B. Kreutz (1976), aurait été le
dérèglement et l'insécurité des voies de navigation, qui favorisaient les
navires naviguant facilement par des vents contraires. Cette question n'a
toutefois pas reçu de réponse absolument satisfaisante.
La fin de la suprématie de la voile latine est très probablement due à

l'adoption, à la fin du XIIIeS. ou au début du XIW, de la voile carrée nordique
<<avecun mât haubanné par des manœuvres dormantes (permettant) un
changement d'amures ne demandant ni réglages délicats, ni travail plus dur
pour I'équipage : simplification du travail sans aggravation de la peine de
l'équipage,, (Adam, 1976 : 562). Mieux encore : la <<nouvellevoile carrée,,
permettait une réduction de I'équipage : F.C. Lane (1934 : 39) notait qu'un
navire vénitien de 240 tonnes, à voile latine, au XIIIe S., nécessitait un
équipage de 50 hommes, alors qu'un navire vénitien de taille comparable, à
voile carrée, aux XIV" et XV" S. n'exigeait que 28 hommes.
Cet aspect de la question pose immédiatement un problème non résolu

à ma connaissance, celui des moyens de subsistance de I'équipage : quelle
était la proportion, à bord des navires <<latins,,,des esclaves ? Et pour les
marins (plus ou moins) libres : quelle était leur solde ? Même des marins aux
«compétences pas trop spécialisées» ont un coût ; le raisonnement, exact
en soi, de Paul Adam suppose une main d'œuvre à faible coût. Est-ce
vérifiable au cours du haut Moyen Age, et ce dans les diverses parties de la
Méditerranée (républiques maritimes d'Italie, Empire byzantin, marines
islamiques, etc.) ? On voit, en tout cas, que l'innovation apportée par
l'apparition de la voile latine, puis sa généralisation, n'ont pas qu'un aspect
purement technique, à présent assez bien connu, mais aussi un aspect
socio-économique dont nous sommes malheureusement en grande partie

ignorants.
Notes additionnelles.
1. Mechanica, 851, b.
Le problème posé par les Mechanica d'Aristote, 851 b, a attiré l'attention

des traducteurs et des commentateurs depuis la Renaissance. On trouvera
un excellent résumé des perplexités engendrées par ce problème dans :
Ludwig Rank, Die Theorie des Segelns in ihrer Enfwicklung (Berlin, Dietrich
Reimer Verlag, 1984). L. Rank, lui-même, commentant l'interprétation de L.
Casson, s'exprime ainsi : <<EineInterpretation, die aus Gründen der
seemannischen Praxis nicht recht überzeugen kann. Es dürfte namlich fast
unmoglich sein, ein Schiff mit derart umgeformte Segel (Rank fait allusion à
la fig. 2) an einen einigermassen kraftig wehenden Wind zu bringen,, (Rank,
1984, 42). On a vu que je partage cette opinion, que je n'ai rencontrée
qu'après avoir rédigé le texte ci-dessus.
2. «Membrure première» et voile latine.
B. Kuentz a observé en 1976 (p. 104-5) que la voile latine semble avoir
été adoptée au cours de la période qui a vu naître la technique de
construction <<membrurepremière,,. Vingt ans plus tard, cette observation,
en raison de découvertes et de travaux récents, est devenue encore plus
pertinente : si la flotte de Bélisaire, en 533, comprenait des navires à voile
latine, Procope décrit l'existence, à Rome en 536, d'un navire construit
((membrurepremière,,, le mavire d'Ede>>(Basch, 1985) et c'est selon cette
même technique que furent construits le navire dont les restes sont connus
sous le nom de <<Saint-Gervais (premier quart du VIIe S. - Jézégou, 1985 A
II>>
et 6) et celui dont les vestiges ont été trouvés à Tantoura (Israël ; ils datent
de la fin du VIeou du début du VIIeS. (Center for Maritime Studies, University
of Haifa, Report n " 22, August 1995 ; communication de Y. Kahanov au cours
du <<SixthInternational Symposium on Ship Construction in Antiquitp
(Lamia, 28-30 août 1996).
En 1985, j'émettais l'hypothèse d'une priorité de la technique

<<membrurepremière,, en Méditerranée occidentale plutôt qu'orientale
(Basch, 1985 : 27) ; depuis la découverte de l'épave de Tantoura, elle
ETSESPARENTESARABES
m'apparaît comme regrettablement prématurée.
L'un des facteurs qui ont favorisé la nouvelle technique de construction

navale est probablement l'augmentation du coût de la main d'œuvre (très
qualifiée en matière de construction <<bordépremier,, ) en raison de la
diminution du travail servile (Steffy, 1994 : 85). Un autre facteur est
probablement la déforestation (Lombard, 1958) : la construction (<bordé
premier,, était extrêmement gourmande de bois.
En revanche, les équipages, supposés plus nombreux sur les navires à

voile latine, ne devenaient-ils pas plus onéreux que ceux des navires à voile
carrée ?
On voit que s'il existe un lien entre les deux innovations, qui semblent
être en effet plus ou moins contemporaines, et non une simple coïncidence,
ce lien est fort difficile à saisir.
Lucien Basch
206, avenue Armand Huysmans
1050 Bruxelles
NOTES
1. Kreutz (1976: 85, n. 24) estime que des voiles latines <<virtually
certain,, figurent dans: Omont,
1929, pl. LX (fol. 452) et XLI (fol. 239) du même manuscrit. On voit, sur ces peintures, une
vergue inclinée à 45" portant une voile carguée; la pl. XLI montre un document très dégradé,
dont la reconstitution semble bien difficile et sur la pl. LX la proue, bien visible, ne paraît
montrer aucune trace de cargue-devant. Un examen des originaux serait utile.
2. La prudence s'impose cependant au sujet du navire du Ms. grec 510, fol. 367: A. Grabar
(1953: 172) écrit que cette <<image.a dû être créée au V' S., ce qui est en effet crédible pour
des raisons de thématique idéologique, mais le thème seul peut remonter au V' S., alors que
le gréement peut avoir été modernisé^ quatre siècles plus tard.
3. Cette ombre est probablement d'origine sociologique: le gréement en V est rarement figuré
parce que la livarde n'a jamais propulsé que de très petits bâtiments qui ne furent que très
tardivement jugés dignes d'être représentés.
4. Traduction personnelle.
Les Mechanica ne sont qu'attribuées à Aristote; le passage 851 b, en particulier, semble être
l'œuvre de Straton de Lampsaque (milieu du III" S. av. J.-C.).
5. Traduction de J.-Ph. Garnaud (Belles Lettres, Paris, 1991), p. 73.
6. On s'attendrait, peut-être, à: <<unmaximum de toile., inais J. Rougé estime que si la partie

postérieure de la vergue ne porte plus de toile du tout, il en subsistait un minimum^ sur la
partie antérieure. Le texte grec ne donnant que aune partie. (pÉpoq) de la toile, il est vain de
discuter du point de savoir si cette partie est un .minimum- ou un *maximum>>:nous
l'ignorerons toujours.
7. Voir, par exemple: Aristophane, Lysistrata, v. 550 et Sophocle, A l a , v. 1083.
8. C'est pourquoi, dans le texte d'Aristote, les marins décidèrent de poursuivre leur route,
malgré un .vent défavorable*, et en dépit de l'importante dérive qu'entraîne la cape
courante. La fuite, manœuvre d'ailleurs dangereuse, les auraient évidemment complètement
déroutés.
9. E. Rieth a souligné justement l'un des aspects particuliers des ex-voto marins: <<... le caractère
exceptionnel de la navigation wotivem fournit un tableau très riche des manœuvres par gros
temps, que ce soit la fuite, la cape sèche, la prise de ris ... (Rieth, 1981: 183). Les fig. 3 et 4,
ici, ne constituent que de simples exemples parmi d'autres de la pertinence de cette
observation. Du seul catalogue (Rubin de Cervin, 1972) d'où sont extraits les fig. 3 et 4, on
peut encore citer, naviguant sous leur seule misaine - et manifestement en péril de mer -
, les trois-mâts des fig. 44, 47, 50 et 52.
Est-il besoin de souligner que, pour des navires au gréement complexe, comme celui des
trois-mâts, il existe, pour prendre la cape, d'autres moyens que celui-ci pour réduire le
gréement? .Cela dépend non seulement de la construction de chaque vaisseau, mais de
bien de circonstances...,, (Encyclopédie méthodique, Marine (1783-1787), p. 256).
10. Ce cordage devait être très important aux yeux de l'auteur de ces graffiti: sur la fig. 7, c, il
est doublé. J'ai eu le privilège de voir ces graffiti, très endommagés, mais encore
reconnaissables, en 1969; en 1993, ils étaient pratiquement détruits.
11. En revanche, je souscris au jugement de P. Paris: ~ J a(Glossaire
l Nautique au mot Latena,
p. 915) cite un certain Messianus, auteur d'une vie de saint Césaire d'Arles (470-542), où il
est question de utrois de ces navires qu'on appelle latins>>. Ce n'est pas un signalement, ce
n'est qu'une épithète, dont on ne sait même pas si elle caractérise une coque ou un
gréement. (Paris, 1946: 79). L'étymologie de <<latine>> pourrait, peut-être, fournir une piste
quant aux lieux où elle s'est répandue. Jal (1848, 11: 915) donne comme origine: <<a la trina>>,
à trois angles. Je suis, à cet égard, aussi sceptique que P. Paris (1946: 80, n. 27); quant à
la supposition de O. Hover (1957: 638), qui songe à une dérivation d'un hypothétique velum
laterale, elle me paraît irrecevable.
12. Curieusement, la voile au tiers, ses origines et ses très nombreuses variétés, tant en Europe
qu'en Asie, n'a jamais fait l'objet que d'observations partielles, alors qu'elle mériterait une
étude d'ensemble.
13. Une curieuse exception: un tel -beaupré. (ou dastûr) figure sur le billet actuel de 200
escudos émis au Portugal et représentant une caravelle de l'époque des découvertes, type
de navire au sujet duquel nous possédons une abondante documentation, dépourvue
d'équivoque: toute espèce de dastûr y est totalement inconnue; ce billet vient à point pour
rappeler combien toute iconographie doit être examinée de manière critique. Aux XVII" et
VXIII" S., probablement afin d'augmenter la surface de la toile et sous l'influence de l'avant
des galères, la cargue devant est frappée, sur de nombreux navires méditerranéens, sur un
-éperon. (ou: uflèclen) en avant de l'étrave sur le chébec, le brigantin, la tartane, la barque
latine et la pinque (Boudriot et Berti, 1987: 12-34).
14. Je remercie M. Pierre-Yves Manguin de m'avoir indiqué, au cours du colloque
<êommunautésMaritimes de l'Océan Indien, IV' S. av. J.-C. - XIV" S. ap. J.-C.. (Maison de
l'orient, Lyon, 30 juin-5 juillet 1996) que les observations d'Ibn Majid et de G. Correa au sujet
de la forme de la voile arabe avaient déjà été commentées par lui (Manguin, 1985: 8), ce que
j'ignorais.
Ce n'est malheureusement qu'envers la mémoire du très regretté Professeur Octavio Lixa
Filgueiras que je puis exprimer ma gratitude pour avoir traduit en français, à mon attention,
les passages de G. Correa cités par Da Fonseca (1934: 150-l), que ma connaissance très
imparfaite du portugais m'interdisait de comprendre.
15. Ce qui ne fut probablement pas le cas; comme le note S. Bellabarba (1988: 235), *il serait
ETSESPARENTESARABES
réellement surprenant que le gréement carré n'ait pas été présewé dans les eaux intérieures,
dans des régions marginales et sur de petits bateaux côtiers.. Mais ces humbles
embarcations n'ont pas eu les honneurs de l'iconographie, ce qui n'est pas sans fausser
notre vision d'ensemble.
REFERENCES BIBLIOGRAPHIQUES
Adam, P., 1970,MApropos des origines de la voile latine,,, dans: Mediterraneo e

Oceano Indiano. Atti del VI Colloquio Internazionale di Storia Marittima, Venise 20-29
septembre 1962, p. 203-9.
Adam, P., 1976, conclusions sur les développements des techniques nautiques
médiévales., Revue d'Histoire économique et sociale, p. 560-7.
Anderson, R.C., 1962,Oared Fighting Ships, Londres: Percival Marshall.
Antoniadis-Bibicou, H.,1966,Etudes d'histoire maritime de Byzance à propos du
thème des Caravisiens*, Paris: S.E.V.P.E.N.
Barata, D. G. P. J., 1989,Estudos de Arqueologia Naval, Lisbonne: Banco de Fomento e Exteior.
Basch, L., 1971,*The Lateener from Thasos., The Mariner's Mirror, 57,p. 329-30.
Basch, L., 1987,Le Musée imaginaire de la marine antique, Athènes: Institut hellénique pour
la préservation de la tradition nautique.
Basch, L., 1989,<<TheWay to the Lateen Sail-, The Mariner's Mirror, 75,p. 328-332.
Basch, L., 1991 A, *Un navire marchand byzantin à Corinthe., Neptunia (Association des Amis
des Musées de la Marine, Paris), no 181,p. 14-21.
Basch, L., 1991 B, -La Felouque des Kelliam, Neptunia (Association des Amis des
Musées de la Marine, Paris), no 183,p. 1-10.
Basch, L., 1993 <<Navireset bateaux coptes: état des questions en 1991., Graeco-
Arabica (Athènes), V, p. 23-62.
Beaudouin, F., 1966-67,*Une famille de voiles égyptiennes de l'Antiquité à nos jours.,
L'Ethnographie, N.S., 60-61,p. 47-59.
Beaudouin, F., 1975,Bateaux des côtes de France, Grenoble: Editions des 4 Seigneurs.
Bellabarba, S., 1988, .The Square-Rigged Ship of the Fabrica di Galere Manuscript>,,The
Mariner's Mirror, 73,p. 225-39.
Ben-Eli, A., 1969-71, *Two Ship Engravings from Bet She' Arimm, Sefunim (Bulletin
published by the Israel Maritime League, The National Maritime Museum, Haifa), III: p. 89.
Bon, A.-M. et A. Bon, 1957,Les Timbres amphoriques de Thasos, Etudes thasiennes, IV (Ecole
Française d'Athènes), Paris.
Bonino, M., 1978 A, ~Lateen-RiggedMedieval Ships. New evidence from wrecks in the Po Delta
(Italy) and notes on pictorial and other documents., The International Journal of Nautical
Archaeology and Underwater Exploration, 7,p. 9-28,Londres: Academic Press.
Bonino, M., 1978 B, Archeologia e tradizione navale tra la Romagna e il Po, Ravenne: Longo
Editore.
Bonnefoux, P.M.J. (de) et F.E. Pâris, 1847,Dictionnaire de marine à voiles et à vapeur, Paris, A.
Bertrand.
Boudriot, J. et H. Berti, 1987,Chébecs et bâtiments méditerranéens. Le Requin,
(Collection d'archéologie navale française), Paris: A.N.C.R.E.
Bowen, R. Leb., Jr., 1952, *Primitive Watercraft of Arabian, The American Neptune, 12,
p. 186221.
Bowen, R. Leb., Jr., 1953,.Eastern Sail Affinitieçu, The American Neptune, 13,1: p. 81-117;11:
p. 185-211.
Bowen, R. Leb., Jr., 1956,<<TheEarliest Lateen Sail., The Mariner's Mirror, 42,p. 239-242.
Bowen, R. Leb., Jr., 1959, *The Origins of Fore-and-Ait Rigs*, The American Neptune, 19, 1:
p. 155-199; 11: p. 274-306.
Brindley, H. H., 1926, aEarly Pictures of Lateen Sails., The Mariner's Mirror, 12, p. 9-22.
Burlet, R., 1988, -La Voile latine., Neptunia (Association des Amis des Musées de la Marine,
Paris, no 171, p. 11-21.
Casson, L., 1966, &tudies in Ancient Sails and Riggingm, American Studies in Papyrology
(Essays in honor of C. Brandford Welles), The American Society of Papyrologists, New
Haven, Connecticut, 1, p. 43-58.
Casson, L., 1971, Ships and Seamanship in the Ancient World, Princeton: Princeton
University Press.
Chevalier, J. et A. Gheerbrant, 1982, Dictionnaire des symboles. R. Laffont, Paris.
Corazzini, F., 1907, Vocabulario Nautico, VII, Bologne.
Correa, G., 1858-66, Lendas da lndia (6 vol.), Lisbonne.
Da Fonseca, Qu., 1934, A caravela portuguesa e a prioridade técnica das navegaçoes
henriquinas, Université de Coimbra.
De la Roncière, M. et M. Mollat du Jourdin, 1984, Les Portulans. Cartes marines du XIIP au XVIIP
siècle, Fribourg: Office du Livre.
Dolley, R. H., 1949, -The Rig of Early Medieval Warships*, The Mariner's Mirror, 35, p. 51-5.
Espérandieu, E., 1907, Recueil général des bas-reliefs de la Gaule romaine, 1, Paris: Imprimerie
Nationale.
Fourquin, N., 1991, <<Lexicographie et archéologie navale médiévales*, Medieval Ships and the
Birth of Technological Societies, II, The Mediterranean Area and European Integration,
European Coordination Centre for Research and Documentation in the Social Sciences -
Foundation for International Studies, University of Malta.
Grabar, A., 1953, La Peinture byzantine, Genève: Skira.
Grosset-Grange, H., 1993, Glossaire nautique arabe ancien et moderne de l'océan
Indien, Paris: Editions du C.T.H.S.
Hawkins, C.W., 1977, The Dhow, Lausanne: Edita.
Hover, O., ~ D a sLateinsegel - Velum latinum - Velum laterale*, Anthropos, 52, p. 636-640.
Hornell, J., 1941, «The Sea-Going Mtepe and Dau of the Lamu Archipelago, The
Mariner's Mirror, 29, 54-68.
Hourani, G.F., 1963, Arab Seafaring in the lndian Ocean in Ancient and Early Medieval Times,
Beyrouth: Khayats.
Howarth, D., 1977, Dhows, Londres: Quartet Books.
Jal, A., 1848, Glossaire nautique, Paris: Firmin Didot.
Jaussen, A., R. Savignac, H . Vincent, 1905, -Abdeh - IV. Les Hypogées>,,Revue
Biblique, N.S., 2, p. 74-82.
Jewell, H.A., 1969, Dhows at Mombasa, Nairobi: East African Publishing House.
Jézégou, M.-P., 1985, d'épave II de l'anse Saint-Gervais à Fos-sur-Mer (Bouches-du-Rhône):
un navire du haut Moyen Age construit sur squelette., Tropis 1, Proceedings of the 1"'
lnternational Symposium on Ship Construction in Antiquiiy
(Piraeus 1985), edited by H. Tzalas, p. 139-46.
Jézégou, M.-P., 1985. B, ~Elémentsde construction sur couples observés sur une
épave du haut Moyen Age découverte à Fos-sur-Mer (Bouches du Rhône)>>,
dans: VI Congreso internacional de arqueologia submarina - Cartagena 1982, Madrid.
Kreutz, B.M., 1976, &hips, Shipping and the Implication of Change in the Early
Medieval Mediterranean., Viator, 7, p. 79-109.
Laird Clowes, G.S., 1932, Sailing Ships. Their Histoy and Development as lllustrated by the
Collection of Ship-Models in the Science Museum, 1, Londres: H.M.S.' S t a t i O n e r y
Service.
Lane, F.C., 1934, Venetian Ships and Shipbuilders of the Renaissance, Baltimore.
ETSESPARENTESARABES
La Roërie, 1956, A . Roman Bowline., The Mariner's Mirror, 42, p. 248-9.

Lombard, M., 1958, .Arsenaux et bois de marine dans la Méditerranée musulmane (Vlle - Xle
siècles)~,dans: Le Navire et I'Economie Maritime du Moyen-Age au XVllle siècle
principalement en Méditerranée. Travaux du Deuxième Colloque lnternational d'histoire
maritime (Académie de Marine, Paris, 17 et 18 mai 1957), Paris: S.E.V.P.E.N.
Manguin, P.-Y., 1985, d a t e Mediaeval Asian Shipbuilding in the lndian Ocean. A
Reappraisab, Moyen Orient et Océan Indien - Middle East & lndian Ocean, 212, p. 1-30.
Paris: Société d'Histoire de l'orient.
Marzari, M., 1984, eVecchie barche adriatiche. Bragouo, Bragagna, Tartans>,, Supplément no
10 de la Rivista Marittirna (octobre),Rome.
Marzari, M., 1988, Trabaccoli e pieleghi nella marineria tradizionale delI'Adriatico, Milan: Mursia.
Merrien, J., 1963, Dictionnaire de la mer (2e éd.), Paris: Robert Laffont.
Moll, F., 1929, Das Schiff in der Bildenden Kunst, Bonn: Kurst Schroeder Verlag.
Mondfeld, W., 1979, Die Arabische Dau, Verlag Delius, Klasing & Co, Bielefeld.
Moore, Sir A., 1925, Last Days of Mast and Sail, Oxford: Clarendon Press.
Morrison, J.S. et R.T. Williams, 1968, Greek Oared Ships 900-322 B.C., Cambridge: Universiîy
Press.
Nicolle, D. 1989, &hipping in Islamic Art: Seventh Through Sixteenth Century AD*, The
American Neptune, 49, p. 168-197.
1979, Oman. A Seafaring Nation. (Cet ouvrage a été publié sans nom d'auteur par le
Ministère de l'Information et de la Culture du Sultanat d'Oman - cité ici: Oman, 1979).
Omont, H., 1929, Miniatures des plus anciens manuscrits grecs de la Bibliothèque
Nationale du VI" au XIV siècle, Paris: Honoré Champion.
Palmer, E.H., 1871, <<TheDesert of Tih and the Country of Moab., Palestine Explorafion Fund
Quarterly Statement, N.S., 1, p. 3-80.
Pâris, F.-E., 1841, Essai sur la construction navale des peuples extraeuropéens, Paris: Arthus
Bertrand.
Paris, P., 1946, <<Voilelatine? Voile arabe? Voile mystérieuse., Hespéris: Archives
berbères et bulletin de I'lnstitut des hautes études marocaines, 36, p. 69-96.
Pliner, M., 1966, <<TheSailing Ship of Bet-Shearim,,, Sefunim (Bulletin published by the Maritime
Museum Foundation, Haifa), 1.
Poujade, J., 1946, La Route des Indes et ses navires, Paris: Payot.
Prins, A. H. J., 1965, Sailina from Lamu, Assen: Van Gorcum & Co.
Regueyro, E. M. et al., 1981, El buque en la Armada espanola, Espagne (sans p r é c i s i O n s ) :
SILEX.
Rodziewicz, M., 1984, Les Habitations romaines tardives d'Alexandrie à la lumière des fouilles
polonaises à Kôm el-Dikka (=Alexandrie, III), Varsovie: Editions scientifiques de Pologne.
Rougé, J., 1979, .Romans grecs et navigation: le voyage de Leucippé et Clitophon de
Beyrouth en Egypte., Archaeonautica, 2, 19'78, p. 265-280.
Sottas, J., 1939, .An Early Lateen Sail in the Mediterranean., The Mariner's Mirror, 25,
p. 229-230.
Tibbets, G.R., 1981, Arab Navigation in the lndian Ocean before the Coming of the
Portuguese, (Oriental Translation Fund, N.S. XLII), Londres: The Royal Asiatic
Society of Great Britain and Ireland.
Tsaravopoulos, A., 1996, ~Ship-representationson the Walls of a Cistern in Piraeus., Tropis, A/
- Proceedings of the 4mlnternational Symposium on Ship Construction in Antiquity, Athens
1991, edited by H. Tzalas, p. 493-506.
Unger, R.W., 1980, The Ship in the Medieval Economy, Londres: Croomhelm; Montreal; McGill
- Queen's University Press.
Vinson, S., 1993, .The Earliest Representations of Brailed Sailsm, Journal of the
Amercian Research Center in Egypt, 30, p. 133-150.
Vinson, S., 1994, Egyptian Boats and Ships, Princes Risborough: Shire Publications.
Waller, M., 1965, .Le Boutre de Madagascar., Le Modèle réduit de bateau, no 125, p. 8-10.
Warmington, E.H., 1928, The Commerce between the Roman Empire and lndia,
Cambridge, Cambridge University Press.
White, L., Jr., 1940, ~Technologyand Invention in the Middle Ages=, Speculum, p. 141-159.
Winkler, H.A., 1938, Rock-Drawings of Southern Upper Egypt, 1. Londres: Egypt
Exploration Society, Oxford University Press.
LEGENDES DES ILLUSTRATIONS

1. felouque= des Kellia. Reproduite d'après: L. Basch, 1991 B: 2. Relevé de Madame
Marguerite Rassart-Debergh.
2. Interprétation, par L. Casson (1971, fig. 188), de la manœuvre décrite par Aristote,
Mechanica, 851 b.
3. et 4. Tableaux votifs, datant respectivement de 1596 et de 1604, de la Madonna dell'Arco,
aujourd'hui conservés au Museo Storico Navale de Venise. Les deux trois-mâts naviguent,
par très gros temps, sous leur seule misaine. Figures extraites de Rubin de Cervin, 1972:
57 et 41.
5. Trois-mâts à la cape. Reproduit de l'Encyclopédie méthodique. Marine (1783-1787), pi. 10,
fig. 136.
6. Trois-mâts à la cape, d'après: R. Gruss, Petit dictionnaire de marine, Paris, 1945, pl. XXIV, fig.
133. Dessin de L. Haffner.
7. Graffiti du Quartier du Théâtre, Délos. Peu après 88 ou 69 av. J.-C., Relevés du Commandant
D. Carlini. La flèche indique la cargue-devant.
8. Mosaïque attribuée au IV" S. ap. J.-C. (fragment). Musée Archéologique de Venise. Photo de
I'auteur.
9. Dipinto représentant un petit bâtiment à voile latine. Hypogée no 2, Anfouchy (Alexandrie).
Probablement 1" S. av. J.-C. Dessin de I'auteur.
10. Bateau de pêche de Moka. Reproduit de Bowen, 1952: 216, d'après Pâris, 1841, pl. 2, 2.
11. Beden seyad, bateau de pêche de Mascate. Reproduit de Bowen, 1953: 187, d'après Pâris,
1841, 1841: pl. 8,l.
12. Houris de Mukalla (Yémen), en 1950: A: au vent arrière, la vergue fixée au mât à un tiers de
sa longueur; B: au vent arrière, la vergue fixée au mât par son milieu; C: au plus près,
I'amure est fixée à l'avant d'une pagaie tendue au-delà de la proue, dans l'axe. Cette pagaie
sert également à tendre la voile en A et B. Reproduit de Bowen, 1952: 217, fig. 14.
13. Navire arabe d'après l'=Atlas Miller. (1519). Dessin de I'auteur.
14. Bateau de pêche de Port Saïd et de Suez, vers 1918. Reproduit de Moore, 1925: 95, fig. 99.
15. Graifito de l'église d'el-Auja, d'après Palmer, 1871: pl. face à p. 29.
16. Relief des Musées de Narbonne, d'époque impériale. Dessin de I'auteur.
17. Fresque d'Eboda, reproduite de la Revue Biblique, 1905, pl. VI et VII.
18. Graffito de Malaga, probablement du XIV" S. D'après: Anderson, 1962, pl. 8 B).
19. Graffito de Bet-Shearim (probablement IIeS. ap. J.-C.). D'après: Pliner, 1962, pl. V, 2.
20. Graffito de Bet Shearim, 5" S. ap. J.-C.
A. d'après Ben Eli, 1969, pl. XVII, 2.
B. dessin de I'auteur.
LA VOILE LATINE, SON ORIGINE, SON NOLUTION
ETSESPARENTESARABES
O h C
Fig. 2
. .--,
.t -- - ,S. - ---
t I.
Fig. 4
Fig. 3
81
LUCIEN BASCH TROPE VI
Fig. 6
Fig. 8 Fig. 9
82
ETSESPARENTESARABES
fig. 10 fig. 11
LUCIEN BASCH TROP/S VI
Fig. 16a
Fig.
ETSESPARENTESARABES
ABSTRACT
AN ASSESSMENT OF MARITIME CONDITIONS, COASTAL ASPECTS

AND THE SUITABILITY OF SELECTED SECOND
MILLENNIUM BC ANCHORAGES OF THE EASTERN MEDITERRANEAN
IN THE PROVISION OF SHELTER.
This paper aims to assess the physical criteria, specifically maritime

conditions, that influence the location of second millennium BC harbours of
the eastern Mediterranean. The maritime conditions that prevail during the
summer sailing season are outlined, with particular reference to the winds,
fetch and swell. An assessment of the general aspects of the coastlines of
the eastern Mediterranean is undertaken, based on modified standards first
outlined by McKee (1983: 23-26), that relate to four basic coastal aspects:
the 'lee shore' (McKee Type I ) , 'exposed shore' (McKee Type 2), 'sheltered
shore' (McKee Type 3), and the 'weather shore' (McKee Type 4).
Next, the suitability of selected second millennium BC harbour sites is

analysed with regard to the provision of shelter they afford, by considering
their particular aspects and topographical characteristics in relation to
prevailing summer maritime conditions. Through calculations of the fetch
and the percentage of 'foul winds that prevail from different sectors, an
assessment of the probability of not entering and not departing these
selected harbours is undertaken. By adopting this quantitative
methodological approach it is hoped to establish an objective analysis of the
suitability of particular anchorages in the provision of shelter.
Dr. Lucy Blue

Institute of Archaeology
University of Oxford
36 Beaumont St.
Oxford OX1 2PG, U.K.
A ROMAN DEPICTION OF A WAR SHIP
EQUIPPED WITH TWO CATAPULTS?
In the area of a Roman necropolis situated on the territory of the city of

Mainz, district of Mainz-Weisenau, the archaeological service of Rheinland-
Pfalz under the supervision of G. Rupprecht excavated some years ago the
fragment of an ancient tombstone (fig. 1). The site of the Roman cemetery
has been well known for a long time1. In the beginning, it belonged to a
Roman camp of the early principate where legionaries as well as auxiliary
troops were located. Nearby this fortification, a Roman settlement (canabae)
developed. The necropolis, in which tombstones of Roman soldiers and
civilians have been found2,accompanies an ancient road to a second camp
for two legions nearly four kilometres away in the present city of M a i d .
Only the base of the limestone monument is preserved. Its front had
been ornamented with a flat relief surrounded by a frame of profiles. The
fragment was part of a tomb stele with a square section of nearly 88 by 35
centimetres. Its preserved height is 95 centimetres. The lateral surfaces and
the backside of the stone are roughly smoothed. On the front side the
monument shows two items, above a level (libella), below the side view of a
war ship with a ram of older fashion4and a subsidiary ram. On the fore ship
and on the poop, a unique type of rail is visible, probably with bollards on it
or behind it. The vessel seems to be decked amidships. Its covered
architecture and the mouldings of the stem and sternpost ornaments
indicate a not too small ship of Mediterranean style, obviously without
outriggers5.
The inscription for the deceased must have been depicted on the broken
off upper part of the tombstone. However, the person initially buried beneath
the monument could be interpreted as a workman since the libella shown on
Roman tombstones is an abundant attribute of this profession. Further, the
level as an instrument of precision can also symbolize exacting activities of
the deceased in general6.Although it may be interesting to consider the
various possibilities of explanations for the combination of handicraft or
specialized skills and warship, the lost inscription leaves too many questions
unanswered.
Without doubt, this new representation of a Roman war ship is a

welcome enrichment to the very few art works of this class found near the
border of the Roman provinces of Germania. Nevertheless, attention should
RONALD BOCKIUS TROPlS VI
be paid to a special detail on the Mainz monument: amidships of the vessel,

two exposed devices attract attention (fig. 2). They seem to be oversized if
compared with the proportions of the ship. The dimensions of both objects
differ in beam only insignificantly but more in height. In a typological sense
they seem to be identical. Characterized as the ship's equipment these items
can be nothing else than technical devices or machines since they consist
of a tripod carrying a presumably revolving apparatus. Observing the right-
hand item, the former can be described as a horizontal crossbar resting on
the top of the tripod. On both flanks of this beam, set on its upper brim the
outlines of rectangular objects are visible. Below or behind that strut a
second horizontal beam appears with both ends running into cone-shaped
objects. Between the pair of "rectangles" remains a gap, this being the
centre part of the device. A similar arrangement appears on the left-hand
apparatus but with some different features: only the upper edge of that
horizontal crossbar which carries the "rectangles" occurs. Nevertheless,
another detail is observable: a diagonal strip passes into the right-hand
"rectangle" running from a point above the top of the tripod and not clearly
recognizable, a second strip passing into the left hand "rectangle" of the
device. Further elements may have been painted in colour.
Exactly comparable presentations of the items mentioned seem to be

missing. In principle, we have to deal with objects of volumetric depth whose
three-dimensional character would become clearer only in a perspective
view. The slight difference of both devices are explainable with their
presentation showing them from the front and from the back, presumably
also with differing angles of elevation. A revolving apparatus, which could be
turned abreast to both sides of an ancient war ship, exposed on its main
deck, is hard to interpret as cargo or nautical equipment. This device,
however, resembles the descriptions as well as the few representations of
ancient catapults7, and we have good reason to assume that the
stonemason of the Mainz monument has depicted a kind of ship artillery.
Inasmuch as this impression is true, the tombstone delivers the only
presentation of a war ship armed with catapults. But this fact is not surprising
since representations also of ancient field or fortress artillery are exceedingly
rare (cf. note 7).
Shape, function and proportions of catapults are more or less known by

Hellenistic and Roman descriptions in literature (Philon; Vitruvius; Heron) if
compared with the few ancient presentations of catapults and the lot of
dislocated original parts of both stone- or arrow-throwing machines found on
archaeological sites. Several experimental reconstructions are based on
these recordss.The main elements of such weapons were - as far as visible

from the front - a tripod which carried a framework with two torsion devices
separated by horizontal struts. These frames and struts had been made of
wood sheeted with metal or not but also of iron. Further details are the two
arms put in the torsion ropes and fittings like washers and counter-plates.
The torsion arrangements of only a Roman catapult type were sheltered in
cylindrical casings.
None of the archaeological materials of ancient catapults have been

found on ships with the exception of some pieces of cargo in the wreck of
Mahdia (cf. note 8). However, written sources mention the use of artillery as
a common practice on war ships not only of the Greeks and Carthaginians
but also of the Romansg.So far, we can not be sure if the field and fortress
artillery of Roman warfare had been fully identical in shape and concerning
further outfit with the catapults of the fleets.
The supposed catapults on the tombstone from Mainz are provided with
tripod carriers as some catapults on Trajan's column which are
characterized by cylindrical housings of the torsion devices on both sides of
a horizontal framework (cf. note 7). On the Mainz relief, the gap between the
two "rectangles" on the horizontal strut may determine the position of the
slider upon which rested the projectile. These "rectangles" seem to hide the
upper parts of what looks like hanging plugs but what obviously are
appliances (washers) at the lower ends of the cylindrical housings or metal
tubes of the catapults on Trajan's column (fig. 3). In any case, the flanking
<(rectangles>> may have been protection for the torsion mechanism and/or
defences (shields) for the service crew (fig. 4). Furthermore, the meaning of
the left-hand catapult seems to be clear, too: the diagonal strips, which
discharge into the outer ends of the "rectangular" (here seen from the back),
are parts of the catapult's sinew. Only one detail of the catapults on the
column we are missing on the tombstone relief, that is the second (arched)
strut also mentioned by Heron for the 'cheiroballistra' which are obviously
depicted on Trajan's column and partly preserved as archaeological
remains. With good reason, Marsden called that light catapult an all-weather
engine since the spring frames (campestria)for the sensitive ropes had been
enclosed in metal cylinderslO.
The monument discussed in this paper may had been the tombstone of
a ship architect or of a specialist whose profession was the manufacturing of
Roman navy catapults since the level as an instrument of precision points to
a 'faber' as well as on an 'architectus'". As inscriptions dating to the Pd
RONALD BOCKIUS TROPlS VI
century AD of Roman legionaries which have been found in Mainz, too,

mention them as leaders of the 'navalia' (wharfs or store depots12),he could
have been both a soldier and a craftsman, an engineer or a designer of ships
or ship catapults. If he had been located in the smaller Roman camp of
Mainz-Weisenau in the vicinity of the necropolis, the stone would date not
later than to the 1" century AD but not earlier than to the last decade of the
1" century BC13.Unfortunately, the historical background of the deceased is
missing due to the lost inscription but it is more likely to suppose that the
tombstone presents a Roman war ship of local use (i.e. on the Rhine in
general).
The mentioned features of the vessel shown on the tombstone point to a

ship class with an oar-system of more than one level. Concerning the
measurable proportions of the ship presentation with enlarged prow and aft
ship, amidships remains relatively less spacious for the oarsmen. On the
basis of an interscalmium of some 90 centimetres, the proportions of the
depicted vessel yield a ship of some 30 metres in length if we assume a crew
of only 50 oarsmen amidships in total banked on two levels sub-deck (in the
section between the ends of the rails) or 76 men three-level banked. If the
oarsmen also used the fat parts of the fore- and after body, 70 men on two
levels or nearly 100 rowers on three levels can be accommodated in a 30
metres long hull of the hypothetical developed lines. Following a notice of
Tacitus (hist. V 22,3) who reports upon the flagship of the Roman provincial
fleet of Germania14, a trireme which had been captured by Germans and
hauled up-stream the river Lippe (a small tributary of the Rhine) in the year
70 AD, triremes could operate on narrow and shallow inland waters, too, but
most of the local Roman war ships must have been smaller units. To a
covered and armed vessel with increased fore ship and poop and with
further elements of seagoing men-of-war, the oar system of a two-banked
galley fits better than a vessel worked by single-banked oarsmen. So the
ship at present being built as an experimental model on scale 1:10 in the
Museum fiir Antike Schiffahrt at Mainz was reconstructed as a bireme (fig. 5-
6).
Dr. Ronald Bockius

Museum fur Antike Schiffahrt
Neutorstrasse 2b
D - 55116 Mainz
NOTES
1. G. Rupprecht, Mainz-Weisenau. Die Romer in Rheinland-Pfalz, ed. H. Cuppers, Stuttgart,

1990, p. 470 with further literature. The tombstone hitherto only mentioned by 0. Hockmann,
Jahrbuch des Romisch-GermanischenZentralmuseums Mainz 33, vol.l,1986, p. 390 note 52;
pp. 408-409 note 116.
2. E. Neeb - P.T. Kessler, Die Ausgrabungen auf dem romischen Friedhofe bei Weisenau.
Mainzer Zeitschrift 8-9, 1913-1914, pp. 37; p. 50 fig. 1-5; p. 51 fig. 1-8.
3. G. Rupprecht, Mainz. Die Romer in Rheinland-Pfalz, ed. H. Cuppers, Stuttgart, 1990, p. 459
fig. 374. - Rupprecht loc. cit. (supra note 1). - D. Baatz, Mogontiacum. Neue Untersuchungen
am romischen Legionslager in Mainz. Limesforschungen 4, 1962, pp. 81, 85 map.1. - W.
Selzer, Ein neues Grabdenkmal aus Mainz. Mainzer Zeitschrift 71-72, 1976-1977, p. 230 fig.
1-2. - On the Roman camps of Mainz see also Baatz loc. cit. pp. 61-87 and K.V. Decker - W.
Selzer in: Aufstieg and Niedergang der Romischen We# V, 1 (Prinzipat), Berlin, 1976, pp. 464.
4. On the three-finned ram see L. Casson - E. Linder, The Evolution in Shape of the Ancient
Ram. The Athlit Ram. The NauticalArchaeology Series 3, 1990, pp. 68-71 fig. 5.
5. Like the types illustrated by L. Basch, Le Musee imaginaire de la marine antique, Athens,
1987, p. 425 fig. 913-916 (if not with an outrigger like p. 491 fig. 1120); p. 426 fig. 918; p. 431
fig. 932.
6. G. Zimmer, Romische Berufsdarstellungen. Archaeologische Forschungen 12, 1982, pp. 161
note 201 and 238.
7. Ancient presentations of Hellenistic and Roman artillery: E.W. Marsden, Greek and Roman
Artillery, vol. I: Historical Development, Oxford, 1969, pl. 1-3 and 9-13. - J.W. Crous,
Mitteilungen des Deutschen Arch. Inst., Romische Abteilung, 48, 1933, 77 pl. 2,85; 10,415;
16,690. - On archaeological remains see as a synopsis M.C. Bishop - J.C.N. Coulston, Roman
Military Equipment from the Punic Wars to the Fall of Rome, London, 1993, pp. 30-167 fig. 6,
26-27,44-45,75,98, 120. - On the advance of research D. Baatz in the reprint of E. Schramm,
Die antiken Geschutze der Saalburg, Bad Homburg, 1980, pp. I-XIII.
8. Some literature on the subject given by J.H. Taylor, Parts of Roman Artillery Projectiles from
Qasr Ibrim, Egypt. Saalburg Jahrbuch 47, 1994, pp. 95-96 note 20. - Further finds: Chr.
Boube-Piccot, Elements de catapultes en bronze decouverts en Mauretanie Tingitane.
-
Bulletin Arch6ologique Marocain 17, 1987-1988, pp. 209-217. D. Baatz, Die Katapultteile.
Das Wrack. Der antike Schiffsfund von Mahdia I, ed. G. Hellenkemper Salies. Kataloge des
Rheinischen Landesmuseums Bonn 1, 1994, pp. 701-707.
9. D. Baatz, Katapultteile aus dem Schiffswrack von Mahdia. Arch. Anzeiger 1985, p. 690. -
Marsden loc. cit. (supra note 7) pp. 169-173.
10. E.W. Marsden, GreekandRoman Artillery, vol. II:Technical Treatises, Oxford, 1971, pp. 206-
233 fig. 9-10.
11. Zimmer loc. cit. (supra note 6) 205 mentions the tombstone of Vedennius, an 'architectus
armamentarii'. The symbolic attributes of Vedennius' profession in life are the front view of a
catapult and an angle (norma) this also being an instrument of precision.
12. P. Herz, Zeugnisse romischen Schiffbaus in Mainz. Jahrbuch des Romisch-Germanischen
Zentralmuseums Mainz 32, 1985, pp. 422-426. - Baatz loc. cit. (supra note 3) p. 83 note 191.
13. Decker - Selzer loc. cit. (supra note 3) pp. 464.
14. On the classis Germanica see Ch.G. Starr, The Roman Imperial Navy 31 BC - AD 324, New
York, 1941, pp. 141-151. - M. Redde, Mare Nostrum. Les Infrastructures, le dispositif et
I'histoire de la marine militaire sous I'empire romain, Rome, 1986, pp. 290-298; to Mainz ibid.
pp. 297-298 note 40.
- RONALD BOCKIUS TROPIS VI
TEXTS TO THE FIGURES
1. Roman tombstone in Mainz. Front-view with flat relief (photograph of a copy in the Museum
fur Antike Schiffahrt Mainz).
2. Amidship section of the war ship depiction on the tombstone shown on fig. 1.
3 Trajan's column, Rome. Section of the relief (Cichorius, C W I ) with the presentation of an
arrow-shooting catapult (La Colonna Traiana, ed. S. Settis, Rome, 1988, p. pl.).
4. Sketch of a Roman navy catapult (front- and side-elevation) considering the depictions on
figs. 2-3 and information's given in Heron's cheirobal1istra.- No scale (drawing by R. Bockius).
5. Attempt at a reconstruction. Body plan of the Roman war ship shown on fig. 1.- Scale 1:
(drawing by R. Bockius).
6. Assumed oar-system of the Roman war ship on figs. 1 and 5. Arrangements of a bireme
operated by single-banked oarsmen on two levels (half cross-section amidships and
longitudinal section seen from inside).- Scale l:(drawing by R. Bockius).
Fig. 3
RONALD BOCKIUS TROPIS VI
Fig. 4a
Fig. 4b
Fig. 5
Fig. 6a Fig. 6b
97
FURTHER STEPS IN THE STUDY OF THE NEMl SHIPS:
ARCHITECTURE AND CLUES FOR THEIR RECONSTRUCTION.
Research developed in the recent years (personally from 1972) brought

to an overall revision of the known documents regarding the Nemi ships and
yielded important results, which are summarized here. Main steps of the
research were: the building technique (1972-1980), naval architecture (from
1985) and the architecture of the non-nautical buildings which were
supported by the hulls (from 1995). A wider range of comparisons made it
possible to reach interesting results and to connect them with the culture of
their time, both nautical and religious.
The analysis of the hulls and of their associated buildings is more clear
now that we can follow ideally the building techniques and phases.
Therefore it is useful to divide these artifacts into the three main parts which
can be well recognized: the shell (alveus and related parts), the crossbeams
and inside hull structures system (interamenta) and the civic buildings. Of
course there are connections between them and a final overview is already
possible.
THE SHELL
Its composing parts are the keel (~porr~q,
spina), the posts (m&Tpa~, aro-
AoL),the planking and the wales (<manip, cinct~],sewn together by means
of gomphoi and armoniai. The phases of building the shell are indicated by
the succession of splices, which were surrounded by the wales. After the first
shell was set up, some of the ribs with bottom frames (trabes) were placed,
and then the other splices were developed in alternated phases, with setting
of the ribs with bottom frames. After completion of the shape of the whole
shell (also with the upper wales and the topgallant bulwarks), tt-fe other ribs,
among which those alternated to the former series and without bottom
frames, were placed and finally the shell was caulked, finished with putty and
then covered by wool cloth imbibed with pitch. Finally lead sheets were
nailed on the lower surface of the hull, while the upper part was probably
finished with encaustus wax paint.
The splices composing the shell appear to have been shaped with a
geometrical approach, which allowed the architect to rationalize the intuitive
shape or the coup d'ceil used to build the desired shape. The procedure2can
be recognized already in Egyptian crafts, thanks to the quality of the figures,
found boats and ships. We know the Egyptian way of bending wooden
boards by means of the tension of twisted ropes, which were used in the first
phases of the construction and then as hogging trusses, or the irrro<6pa~a
(tormenta) of classical times, in the finished ships. The resulting shape is an
arc of ellipses or, if two contrasting points are inserted at the ends (like
stretching and lowering the bundles of papyrus at the ends), a central arc of
ellipse and sinusoidal arcs at the ends, or a complete sinusoid.
Rationalization of this curve (Fig. 1) appears to have been developed, from
the ships and boats of Cheops, Sesostris and Kyrenia up to those of Nemi,
to give to the architect the geometric construction of the curves and the
minimum number of elements with which to build the whole shape of the
hull. Seeing the rationalization of the physical shape of a bent board of
wood, we could have the doubt that this geometrical development could
have not been intentional, as it happened with Viking ships.3 But the rational
support of this construction in Roman times is consistent with the use of a
compass, with the construction of the ellipses (like those of amphitheatres)
by means of selected points connected with a curvilineal and with other
figures, like Archimedes' spiral and screw. There can be discussions about
how in detail these curved lines were drawn and kept under control, but it is
highly possible that either generating circumferences (as used in "classical"
geometric constructions), or models or rulers could have given the same
geometrical results.
Then the criteria to shape the shells of the hulls of the Nemi ships
appear:
A - size and weight of the civic buildings to keep afloat;
B - length of the hull, in entire multiples of the Attic or Roman foot (= 295
mm), measured at the base of the main <wo-njp (see table in appendix);
C - breadth at midship, also in entire multiples of the Roman foot (see
table).
Due to the size of the hulls, some criteria were used to correlate these
main measurements and the draught with the loads4and the weights of the
buildings, the results of these criteria tended to give oversized volumes
and structures, like in inland buildings, which appear generally as
oversized.
D - profile and heights (at centre, at the ends and at the "active" or reference
sections), basically also in entire multiples of the Attic or Roman foot (see
table).
The profile was drawn with a rectilinear central part connected to
curves which were developed elliptically and sinusoidally from the
generating circumference of the plan of the first splice. The part above the
floating line in both ships had the typical shape of oared Hellenistic
warships, the profile of the prow of the first ship is well known and can be
reconstructed with a "Vitruvian" procedure; with some correction with
respect to the original reconstruction. In the second ship the bow was the
end taken for the stern during the excavations. The exact nature of this end
is shown by the peculiar narrowing of the main <ocm)p, by the mortice for
the &ppoAovon the keel, the foot for the moAoq and for the fore post, with
their mortices, and by the big block fixing the moAoq, which gives also its
slope (Fig. 3 C). This confirms the relationship with the first ship and gives
more reasons for the presence of the anhmqq.
E - shape of the shell, with side by side succession of the splices
corresponding to the building phases and the division of the whole hull
into three parts: a central one shaped in order to obtain the maximum
volume in agreement to the type of ship, and the narrowing ends.
Reference sections which divided the three parts have been referred to as
"active " frames5(but they can be called also quarti or cao di sesto) and in
correspondence to them the curves of the lines of the splices changed
from elliptical to sinusoidal. Moreover the breadths of the segments of
splices at these sections are the same at both "active" frames, so they can
be set with parallels to the keel. In the first ship these "active" sections
were between ribs 22-26 and 109-112, probably with an oblique alignment
through the various splices. In the second ship they are more exactly
defined and correspond to the spaces between ribs 97-98 and 25-26
(number of ribs is according to G. Ucelli's Le navi di Nemo.
The succession of the splices composing the shell involves a

conceptual growth around the first splice, from the keel to the first wale, the
shape of which was built with a generating circumference, which had the
same radius for both ships. This circumference was used to build also the
line of the profile, as it is clear in Cheop's and Sesostris' Egyptian crafts,
and this splice appears as the conceptual basis around which the
complete shells of both ships were shaped. The other two splices, up to
the main wale at the floating line, were conformed similarly, with
circumferences with the same centre for the first ship, but in the second
ship the centres of the other generating circumferences appear to have
been displaced to an extent which still needs proper identification, but
which has been intuitively recognized also soon after the excavations6.
The flat part of the bottom of the first ship was defined exactly by the
two first splices, while that of the second was wider than the second splice,
but its boundaries were defined by the same geometrical construction,
with change of shape (from elliptical to sinusoidal) in correspondence of
the same pre-fixed "active" sections mentioned before.
Also from the standpoint of the "growth" of the splices to build the
shell, the second ship appears as an enlargement of the first, because the
splices added to the first one (which is exactly as wide as that of the first
ship) have been composed with circles in a fragmented and enlarged
pattern, as though the shape was obtained by adding composable parts,
originating from the shape of the first ship, with the application of enlarging
ratios.
F - cross sections: the same shape was reported along the outer part of the
shell, by adapting it to the geometrical frame built by the plan, profile and
shape of the splices; in Cheop's ship this curve was an arc of
circumference, in Sesostris' boats it was the elliptical development of a
circumference and in the Nemi ships we find the same constructions (Fig.2
and 30). Towards the ends of the hull these master sections were raised
to follow the shape of the profile and of the cinctus, with a movement along
a line which is well reconstructable for the second ship starting from the
"active" section: conceptually this line is similar to the scorer de le seste in
Byzantine and Venetian traditions.'
G - in the construction of the shells, ribs were essential for supporting them,
to avoid their distortion and to distribute evenly the stresses. In the first
ship they were nearer than in the second, but in both ships we observe the
alternation of ribs with and without bottom frame (trabs) as in many ancient
wrecks. The last 8 ribs at both ends of the two ships were without bottom
frames and at the sides of the keel.
The described design of the curves with generating circumferences was

not exclusive in ancient shipbuilding; also other criteria could have been
used, as we see in the warships of Marsala8 (3rd cent. BC) or in the
Comacchio boat9(Augustus' time), but it appears to have been widely used
to plan the other ancient ships and boats mentioned so far (from Egypt, to
Kyrenia, Nemi and Fiumicino). This composition of shapes and the "growth"
of lines around a generating figure recalls a similar "growth", on the base of
golden rectangles, of surfaces in Classical civic architecture.
After Hellenistic times which evidence have we of the design with

generating circumference and elliptical or sinusoidal developments?
Renaissance scholars have found the shaping of the ~ v ~ a oof l qthe Doric
column as feasible with the elliptical development of the circumference
building the modulus of the column. I already mentioned the Fiumicino
boats, the splices of which can be similarly designed. The Venetian treatises
of Byzantine tradition between the 15'hand 16'hcenturies show a wide use of
the partisim or the sinusoidal development of a circumference, obtained with
a ruler called meza luna. Moreover, the same shape was used for all cross
sections of the hull, by displacing a single master frame and lifting it at the
ends according to the narrowing of the hull (scorer de le seste).
Naval architects of the l T hcentury considered the ellipse and the circle
as important tools for shaping the hulls and curvilineals were used with a
flexible wood or steel lists, which gave a composition of elliptical and
sinusoidal curves.1°
In recent Mediterranean building tradition the hulls are shaped with the
garbo (or po6&Ao),which uses the same shape for the ribs of all the central
part of the hull; in some cases an elliptical curve is still obtained in this part
(e.g. the gono in Ligury and Latium or the schifu of Sardinia). In other
traditions (Lake Como, Lake Iseo) the usage of the same shape for all
transversal sections of the hull is still more evident. This means that
Mediterranean boat-building kept for thousands of years basic principles for
shaping the hulls coming from the dawn of its developments.
Constructions and measures could be reconstructed with a precision

higher than 98% on the most critical or distorted parts (in some cases
99.8%). Re-drawing was necessary to correct distortions, which were wider
in the first than in the second ship, mainly on the upper line of the main
<omjp, on the cross sections and on the keel, which was distorted at least
in four sections of its original length. These distortions have been somewhat
simplified in the drawings published in 1940, and further adjustments and
approximations are needed when transferring these surveys on the full scale.
CROSSBEAMS SYSTEM
The shell was the base for the upper buildings: the main keelson was put
over the keel and two other keelsons for each side were put in order to
obtain the widest flat inside room, which was divided into equal parts, to
distribute as evenly as possible the incoming loads. For this reason the
keelsons do not correspond to the wales. At the inside edges of the sides,
stringers were set to support the crossbeams, which did not protrude
outside the planking nor out of the main <wcnTjp. The first ship had 30
crossbeams and the second 34: in both ships their distribution was not on
the whole extensions of the hulls, but it was limited to that supporting the
main buildings and to the anho-rqq. The distances between crossbeams of
the first ship was the same, with the exception of the first two at prow, while
for the second ship the general rule of one crossbeam every three ribs had
many exceptions, according to the distribution of the loads and of the over-
structures. In the second ship the layout of the anhmqq shows exact
geometric criteria: its plan is obtained by a square and an equilateral
triangle, which connect at the main section, and its protrusion out of the
main < w m p is 1/12 of the breadth. Connection of crossbeams with the ribs
was not very strong, or was lacking entirely, as though they were only
leaning against the shell. Moreover, there were no braces, which a Medieval
or Renaissance builder would have never skipped.
Longitudinal stresses were counteracted also by means of longitudinal

stringers nailed against the props and morticed under the crossbeams.
There were a couple of them on the central line for almost the complete
length of the hull and there were single such stringers between the middle
and the sides, also in connection with the distribution of the buildings. On
the first ship they were continuous and symmetric, while in the second ship
the side stringers were in three sections, which followed the layout of the
buildings. Such longitudinal structures are evident in Egyptian models of the
Middle and New Kingdom, as a mean to keep the shape of the shell, by
counteracting at the level of the deck the stresses exerted by the shell, which
tended to move the crossbeams in the longitudinal direction."
If we consider now the shaping criteria of the shell and the distribution of
the crossbeams, we can understand two literary passages, which have been
widely discussed in the past. The Nemi ships show clearly both the criteria
of "growing" parts and those of fixed moduli. The hulls are shaped with
curves "growing" around generating circumferences, but many other
measures are fixed (mainly in entire multiples of the Roman foot) and
typically "Vitruvian": the anhmqq of the second ship, the distances between
props supporting the crossbeams, the number of frames at the ends, the
profile of the first ship, which was built also with scenographic intention and,
of course, the plans and volumes of the civic buildings.
The two ideas of how to give rational shapes and measures are well
shown also by two literary passages: Apollonius of Rhodes ( ' A ~ ~ o v ~ u T L K ~ ,
o 'ApyoOq, ~ a avo-
1, 723-724): OTE n p O ~ o v6puoxouq h - t e 6 6 M ~ ~vqoq i
VEUUL i 5 a ~
<uya p ~ ~ p ~ j u a u 8 a ~ .
This means that the ~ a v o v

was given by the measures of the <uy6, in the
sense, not very explicit, but most probable, of their length. We have seen
how a generating circumference could be obtained from the main breadth,
to draw the lines of the plan and of the profile. In his work Apollonius shows
a deep knowledge of ships, and a lost work ndpt ~ o ~ptfipov

v attributed to
him confirms his familiarity with nautical matters. If he wanted to indicate the
distances instead of the lengths of the Cuya, he would have used different
words.
Vitruvius states the idea of repeated fixed moduli (De Architectura, I, II,
4) :Et iterum.. . navibus interscalmio, quae dinqxuaia dicitur, item ceterorum
operum e membris invenitur symmetriarum ratiocinato: two cubits from one
tholepin and the next.
This distance was not the same in all rowing ships; it depended on how
many oarsmen worked each oar. Moreover, in the second Nemi ship the
distance was not always the same. From the standpoint of the building
technique this distance had little planning importance. It could have been
used only to fix the parts into which to divide the sectors of the generating
circumference, and intuitively they were put regularly in order to distribute
evenly the stresses and to support the benches for the oarsmen. I think that
Vitruvius took a definition of the type of oarship according to the number of
oars (an old term recalling the oars worked by one man each, like on the
~pta~ovrtjp orothe
~ nsv-q~ovrtjpot)as a definition of modular proportions,
which suited more his mentality. Actually, proportions were to some extent
bound to the number of oars, and the usage of moduli was mainly to
memorize rules which allowed the faber to build without detailed drawings.
But Vitruvius and his followers of the Renaissance took only the additive
concepts of the moduli, and made a myth of them: the idea of "growing"
figures was forgotten or not well understood.
So I would suggest that Apollonius refers to the shaping criteria of the
shell, while Vitruvius to that of the subsequent phases: crossbeams and
superstructures.
CIVIC BUILDINGS
The most uncommon feature of the Nemi ship was to have temples,
rooms, peristilium built on them. Their foundations were to some extent
structurally independent of the crossbeams systems, although in some parts
their props and walls were leaning against them and they were connected to
the longitudinal stringers. The supports were made of one or two orders of
keelsons (transversal or longitudinal according to the shape and weight
distributions of the buildings) leaning against the pre-built interamenta
(keelsons and inside planking). The foundations supporting the props for the
skeleton of the buildings were rigidly fixed in the first ship, while in the
second the most important foundations were in double order. Over the main
frame leaning against the keelsons and the inside planking some spacers
have been put, over which the foundation with mortices for the props have
been mounted. Spacers are not nailed, but only put on the frame, thus
having the chance to absorb longitudinal stresses by sliding a little between
the two main horizontal supporting frames. This solution can be called anti-
seismic and shows a particular care to avoid undue stresses exerted on the
hull by the civic buildings.
The masonry of the buildings and the platforms with suspensurae were
made over the level of the main deck, but they were limited to the boundaries
of the buildings. On the first ship there were gangways at their sides and in
the second there were also corridors, but wooden structures were the most
important. Our knowledge of wooden structures of ancient buildings is
limited (Ercolano, the craticii or the military machines mentioned by
Vitruvius), although some interesting comparisons can be done with
Medieval and traditional wooden houses, like those of Bologna of the 14'h
century or those of Tyrol and Germany. Some occasional further
comparisons (like the upper crossbeams of some churches in Venice and in
Murano, or the irons "chains" of the arcades of Bologna) give more clues for
the reconstruction of the ships, with results which were unexpected so far.
The basis for this work is always the quality of the documents produced in
Le navi di Nemi by Guido Ucelli, with some complements given by the
models in the Museo delle navi romane in Nemi (scale 1:5) and in the Museo
storico navale in Venice (scale 1:50).
ASPECT OF THE FIRST SHIP (Fig. 4)

The ribs of the first ship were nearer but crossbeams were more distant
than in the second ship because the ratios of loads were different; raised
parts of the deck were from the middle to the stern end and partially on the
prow area. Buildings were in the centre in a continuous block of successive
ambiences. There was a heavy building in the centre, on a podium of
concrete lightened with suspensurae of terracotta tubes; similar
suspensurae were on the first building sternwards, the other buildings had
only a floor leaning against tiles laid on the deck, all of them were paved with
opus sectile. We have little information about the buildings; architectural
elements are almost absent. We understand that they were multifold, with
rooms, corridors, with mosaics and marbles on the walls. The main central
building recalls the proportions of the Tuscanic temple (divided into tenths
of the width), but its foundations and a small fragment of a Corinthian capital
suggest that it could have been a propyleum, with columns about 4 metres
high. The buildings sternwards are divided into small rooms around a unique
central corridor, to end with two small ambients most probably devoted to
the ships' services (mainly steering). The profile is the well known military
one with architectural and scenographic interpretation. In fact, the ram is just
placed against the end of the keel and of the wales, with no horizontal frame.
There was no Cmho-rqq, inside parts of the sides could not accommodate
oarsmen and the structure supporting the steering devices did not allow for
the presence of oars: the ship was towed.
ASPECT OF THE SECOND SHIP (Fig. 5)

The second ship shows more elements for reconstruction; the military
profile justifies the anhmqq which supported the four steering devices, gave
access to the ship through separate corridors and most probably
accommodated a number of oars on each side, worked by at least two men
each. We could indicate in a first approximation this number as equal to that
of the referred crossbeams (24 excluding the stairs), but space between the
crossbeams (1.80m or more) and the strength required to work the oars of
such a heavy ship suggest that the number of oars should be doubled.
Probably there were oblique benches, but the question is open. The inside
structures were different from those of the first ship (ribs were more distant,
but crossbeams were nearer, with the same even distribution of props
supporting the crossbeams, which divided the space devoted to the
buildings into a sort of modular system), the deck structure was more linear.
The crossbeams supported the main deck; raised walkways were at the
sides and raised partial decks at the ends. What remains of the buildings
reveals their foundations and plans: some props and architectural elements,
three sizes of Corinthian columns, a terracotta frieze, doors and windows,
roof tiles covered by gilded copper, bronze fences. The distribution and the
features of these elements suggest the original presence of a central area
occupied by a peristilium (with proportions 5 x 4 and exact relationships with
the crossbeams system), which ended sternwards with a tetrastyle temple
(with a square naos and a general proportion of 1.21 including the pronaos)
and forewards with a wall with windows surrounded by small Corinthian
twisted columns leaning on trusses. In the middle of the peristilium, made of
six columns on each side, a vertical prop supported a crossbeam, which was
perpendicularly connected to the longitudinal trabeations of the peristilium.
It had the task of absorbing the transversal stresses of that structure and to
divide the space, e.g. with a curtain. The masonry work was made above the
main deck, which in many parts was covered by a thin marble opus sectile
on the closed floors and lithostraton on the open areas. Around the
peristilium there was an upper deck or walkway surrounded by a bronze
fence and it was connected to the parts of deck, at the ends, where there
were other buildings whose nature is still difficult to define, but their sizes
and stairs accesses should have been in line with the buildings described
before and with the maneuvers of the large steering devices.
The distribution and the proportions of the buildings is typical of such

Hellenistic complexes, like, e.g. the temple of lsis in Pompei; transversal
proportions are in even multiples of the total width of the peristilium, which
can be divided into four parts, two of which were taken by the temple. This
is a further clue for the importance of the architectural and the scenographic
aspect built on the nautical part of the ship. The buildings appear as
surrounded by the high sides of the ship, as though the sacred area should
have been seen and recognized only from inside the ship, and the whole
architecture of the buildings shows and confirms a scenographic intention:
the peristilium and the front wall of the building at prow (that with twisted
columns) are like the scene of a theatre.
Such a hull did not need ballast for floating safely, but first approximate
calculation of the uplift in comparison to the loads proved that an additional
350 tons approximately were needed to obtain the correct trim in order to
work the oars correctly. The gravel found on boards was put on purpose in
order to obtain this result.'*
These assumptions are not definite, but always subject to trial and error
verifications. Remaking the drawings, calculations of loads and uplifts and
building a model of the second ship on a scale 1:100, with basically the
same original phases and shaped with the same principles, appear to
confirm the considerations reported in this contribution.
Further steps will be second approximate models or, even, a full size
reconstruction of one of the two ships (I would prefer the second) which was
recently proposed. The elements collected so far make it possible to
propose such scientific reconstructions, which were not possible even to
imagine only a few decades ago.
Dott. Marco Bonino

Via G. Matteotti No4
1-40129 Bologna
NOTES
1. G. Ucelli, Le navi di Nemi, Roma 1940, other editions, 1950, 1970, 1996; works by M.
Bonino:, Una barca costruita dal faber navalis P. Longidieno, in Felix Ravenna, NS, Il-IV, pp
19-54, 1972; La tecnica costruttiva navale antica, esempi e tipi dell'ltalia settentrionale, in
Plinio, i suoiluoghi, il suo tempo, pp 187-226, Como 1984. Notes on the architecture of
some Roman ships, Nemi and Fiumicino, in Tropis I,pp 37-53, Athens 1989 ; Appunti di
tecnica, architettura e cultura navale, in Atti della IV rassegna archeologica subacquea,
1989, pp 113-125, Messina 1991.
2. M. Bonino Un metodo geometric0 nel la conformazione delle imbarcazioni papiriformi
egizie, in Convegno di studi in onore di T. Viola a dieci anni dalla scompasrsa, Torino 1985,
acts forthcoming.
3. R. Steffy, Wooden shipbuilding and the interpretation of shipwrecks, Texas Univ. Press,
1994, pp 100-127.
4. A mention was made to a passage by Hero of Alexandria, De mensuris, Chap. WII,
correlating these concepts, with cargo in Italic modii but still this passage needs revision.
5. L. Basch, Wrecks and the archaeology of ships, in The Int. Journ. of NauticalArchaeology,
N. 1; 1972, pp 1-58.
6. S. Milosa, in Le navi di Nemi, cit, pp 177-178.
7. A. Chiggiato, Contenuti delle architetture navali antiche, in Ateneo Veneto, a. C W I I I , pp
141-211, 1991.
8. A.P. Farrar, The Marsala Punic ship, The shape to be re-lofted, in The Mariner's Mirror,
V. 75, 1989, pp 368-370.
9. M. Bonino, Tecnica costruttiva ed architettura navale, proposte per la ricostruzione, in Fede
Berti (curator) FORNNA MARIS, la nave romana di Comacchio, Nuova Alfa, Bologna 1990,
pp 3542.
10. A. Classon Ralamb, Skepps Bjggery eller Adelig Oeffnings Tionde Tom, Stockholm 1691,
reprint, Malmo 1943, pl Aa, Dassie, L'architecture navale, Paris 1695, PI. 3, 13.
11. M. Bonino, Dalla tecnica all'architettura naale egizia, in Annuario VI (1995-1996),
Associazione Amici e Collaboratori del Museo Egizio di Torino, pp 14-35, Turin 1997.
12. A previous hypothesis was that the gravel was put on board for sinking the ship, but it
proved not to be the case, even if we have some examples of this procedure, like in the
Venetian galley of Lazise (Lake Garda).
La tecnica costruttiva navale antica ci aiuta a leggere i riferimenti per la

conformazione degli scafi delle navi di Nemi. Un approfondimento ed un
rifacimento dei disegni disponibili ha consentito di identificare i metodi forse
usati per il controllo delle linee curve e delle proporzioni delle tre sezioni che
costituivano le navi: il guscio dello scafo, il sistema dei bagli e gli edifici . Gli
spicchi che componevano il guscio, in profilo ed in pianta, erano disegnati
dallo sviluppo ellittico e sinusoidale della circonferenza generatrice costruita
sulla sezione maestra, second0 un metodo che gia troviamo in documenti
egizi e che in parte si 6 tramandato nel Mediterraneo per lungo tempo. Le
sezioni trasversali erano ottenute da un'unica forma ripetuta lungo la
superficie esterna dello scafo; i bagli, con i dormienti e gli altri correnti
longitudianli fornivano il raccordo con le sovrastrutture e gli edifici templari,
diversi per le due navi. Vengono discusse le relazioni di queste tre parti con
i concetti modulari dinamici e fissi e viene data una descrizione di massima
dell'aspetto originario delle due navi, una ricostruzione che ora e possibile
tentare per la seconda nave, che aveva la prua rostrata militare ed era mossa
da remi: della prima conosciamo bene il profilo, ma abbiamo pochi elementi
architettonici sicuri..
CAPTIONS TO FIGURES
1. Elliptical and sinusoidal rationalization of the shape of a bent timber strake.

2. First Nemi ship. Construction of the lines, general planning.
3. Second Nemi ship. Construction of the lines, general planning, details of the prow.
A: profile, B: plan and proportions of the C m h q q , C: structure of the prow, D: main section.
4. First Nemi ship. Prospective view of the reconstructable parts.
5. Second Nemi ship. Prospective view of the first approximation reconstruction presented in
Lamia in 1996. Adjustments of the number of oars and of minor details of the buildings were
necessary since then.
TABLE: Measurements and criteria of proportion; a comparison between

the two ships.
FIRST SHlP SECOND SHlP
Number of ribs 148 118
Distance between ribs 1.55 Roman feet 2.0 Roman feet
Distances between cross beams 6.2 Roman feet 6.0 Roman feet
(ribs beween cross beams) (mainly 4, 2 at the fore end) (mainly 3, also 4 and 2)
Length of straight part of keel 46 Roman feet (?) 46 Roman feet
End of cinctus /first cross beam 12 ribs (stern),l6 ribs (prow) 12 ribs (both ends)
Incompleted ribs at both ends 8 8
Ratio: 112 breadthlheight 6 6.25
Heigth at prow (under the cinctus)
over construction line 7,5 Roman feet (?) 10 Roman feet (?)
Dimensions of slice I (half) 180 x 11 Roman feet 190 x 11 Roman feet
Dimensions of slice II (half) 210 x 24 225 x 25
Flat part of the bottom (half) 180 x 24 200.5 x 30
Main cross sections 66 x 5.5 (shell: 4.5) 80 x 6.4 (shell : 5.5)
'Anhqq +
184.7 (99 49.5~43)x 99
Main building Propyleum: 40 x 50 Temple: 32.5 x 40
Peristilium 50 x 82
Fig. 1
*. f:.~
Fig. 2
Fig. 3
ABSTRACT
JEAN-PIERRE & ANNE JONCHERAY'S EXCAVATION OF TWO SMALL

ROMAN COASTAL CRAFTS ON THE SOUTH COAST OF FRANCE
Jean-Pierre Joncheray, the editor and main contributor to Cahiers

d'Archeologie Subaguatigue, has been excavating Roman shipwreck
remains along the south coast of France over the last 27 years. Anne Lopez,
now his wife, has been working with him since 1989 and together they have
discovered over 5 Roman shipwrecks and excavated 14 which includes:
Dramont 'C' Small coastal craft of the 1". Century BC

Dramont 'I' Ship with a cargo of three large marble blocks
from Teos, 1" Century AD.
Barthelemy '6' Small coastal craft with a cargo of tegula, 1"
Century AD.
Chretienne 'M' 2 wrecks combined one of l"/2ndCentury BC
the other from the 5IhCentury BC.
Heliopolis 'A' 4'h Century AD wreck with a cargo of
amphorae from North Africa
The excavations of the two coastal crafts have revealed interesting

aspects about the life of Roman seamen working along the south coast of
France and are the focus of this paper.
Dramont 'C' was between 12-13m long and had a mixed cargo of
amphorae (Dressel 1B and Lambolia 2), iron bars and pine resin. At least
one of the members of the crew was making cork amphora stoppers on a
stone block at the time of the shipwreck.
Barthelemy '8' was of a similar size, but slightly smaller, and had a cargo
of approximately 100 tegula and imbrices. Although this was only a small
boat it was self contained and obviously home to the crew. A lead <<foculusa>
(an ingenious cooking and water heating device) was located at the stern of
the cargo. Its design is very similar to other examples from Israel and Turkey.
Christopher Brandon
Pringle Brandon
13 Sun Street
London EC2M 2PS.
FIREPROOFING OF WAR MACHINES, SHIPS
AND GARMENTS
The development of offensive weapons and defenses against them -the

perennial cat-and-mouse game - has been going on for centuries, a
conspicuous recent instance being the use of Scud missiles and antimissile
Patriot missiles in the Gulf War.
Incendiary missiles were in use in antiquity and developed rapidly in the

Hellenistic period, and various forms of fire extinguishers were invented to
deal with them. The main extinguisher in ancient times was vinegar (ocoq).
Since vinegar was little better than water, as Pattington has rightly pointed
out, we may assume that <<the vinegar included salty sources, salt left on the
surface of burning wood helping to extinguish it.,,' Other extinguishers in
common use were sand, raw hides, urine (which contains ammonia and
phosphates) and alum.
After the introduction by the Byzantines of Greek fire or *liquid fire,,,

which was perfected by Kalliniko~,~ a more systematic defense against this
very dangerous weapon was developed. Sadly, Chapter 70 of the tenth-
century IuMoyv T a ~ r f ~ cdealing
c ; , ~ with the defenses against Greek fire
and methods of extinguishing it, is missing: only the title (4lQq Bv TO Gypov
K ~ A O ~ ~ E V o6~aesiq
OV nOp, ~ a nQq i 6p6Aq06v SOAolq, T O U T ~ V O L ~ X
awa~~on) has been preser~ed.~
A marginal note informs us that vinegar was used for protection against
Greek fire and the best way to protect a wall was to pour vinegar over it.' It
is equally unfortunate that some important Arabic sources concerning Greek
fire, which would surely have contained information about the methods of
extinguishing it, are also lost.Vortunately there are a few scattered
references to fireproof equipment and fire-extinguishers in the Byzantine
sources and far more in the Arabic sources.
In this paper I shall give a short account of the methods used for
fireproofing war machines and ships and of the fireproof garments worn by
the personnel who used Greek fire, as described by Ibn al-Manqali (Mangli).
VASSlLlOS CHRISTIDES TROPlS VI
War Machines
The Arabic word for fire-throwing machines is mujalladun (<<boundin

leather,,). Raw hides had long been known for their fire-resistant properties:
they were used by Alexander the Great and are also mentioned in Byzantine
sources.Vhe Arabs not only covered their fire-throwing machines with hides
but also coated them with vinegar and talc.' They had used fire-throwers as
early as the time of their invasion of India under Muhammad Ibn Al Qasim in
710-713:' the History of Haider-Rezi tells us that the natives' war elephants
were terrified by such machines bombarding them with missiles called atash
bazi.'
Ships
Efforts were made to protect ships against fire. The Byzantine sources do
not report any such efforts, but the 12'" century author I. Kinnamus reports
that the experienced Venetian navigators used a sort of vinegar (otoq) to
make their ships fireproof. On the other hand the Arabic sources inform us
that warships were covered with raw hides steeped in vinegar, sometimes
called <<acidvinegar>>.l0In addition, ships were smeared with linen rags
moistened with aqaqir (a vague term meaning <<drugs,>, the precise meaning
of which is unclear) from as early as the middle of the eighth century. Talc,
an important fire-extinguishing agent, was also added." Of course, even if
ships were fireproofed they were not immune from destruction, as fire could
by sprayed on the crew members, creating havoc (Fig. 1).
Protective Garments
Firefighters
(Arabic naffatun or zarraqun, Greek atcpova~op~q).
In both the Arab and Byzantine navies the firefighters were an elite
group: in fact, the crews of warships were carefully selected and rigorously
trained." They were called naffatun or zarraqun and a~cpovdrropeqin Arabic
and Greek respectively. According to the Arabic sources, firefighters enjoyed
a position of special distinction among all the crew members. Scattered
references in many Arabic sources inform us that they wore special
garments known as libas a/-Naffatin, smeared with various substances. Of
course such fireproof garments could be worn by other soldiers as well.
FIREPROOFING OF WAR MACHINES,
SHIPS AND GARMENTS
The passage that follows, taken from Al-Ahkam al-Muluqiyah by Ibn al-
Manqali,13 describes the fireproof garments worn by firefighters. It is
published here for the first time in connection with the Greek fire. Ibn al-
Manqali does not mention specifically by whom such fireproof clothes were
used. Arab iconography shows the use of such garments by Arab horsemen
and foot soldiers (See Fig. 2).
Arabic Text
Translation: Coating the cloak which does not burn
A cloak is coated and a man is wrapped in it and is set on fire and the
fire continues but the man suffers nothing from the heat.
One part of smooth, pure talcais taken, one part of Egyptian alumb,one
part of Yemenite alum, one part of ammoniumc, one part of hematited, a
stone from Tur, and one part of gypsum." You pulverize this (mixture) and
soak it in stale urinef for ten days. Then it is pounded with the smooth egg
white and the garment is smeared on the inside and the outside and is left
to dry. Then the man is wrapped in it. Then nap, is heated with sulfur and
goat's fat separately. Then (mixture) matures (cools after its mixing) on the
outside of the garment. Then you set fire to it. And the fire continues to burn
and you add naft continuously, hour after hour, all day. As a result nothing
penetrates inside the cloak and when the man is wrapped in it and he must
protect his face from the flame of the fire.
Arabic Technological Terms
a. Talc frequently appears in the Arabic sources (Biruni, Ibn Baytar, and
others) as the best fireproof substance. Its appellation has been preserved
in modern chemistry. It is <<ahydrated magnesium layer silicate
(phyllosilicate)... Talc is a good insulating material.. See J.G. Liou, <<Talc.
in McGraw-HillEncyclopaedia of Science and Technology 18 (1992), 123-
VASSlLlOS CHRISTIDES TROPlS VI
b. Shab = alum, colorless or white astringent tasting mineral, <<oneof the

most abundant minerals in earth's crust...>>,
in The Merck Index, Rahway,
N.J., 1983, 48. See Ahmad Y. al-Hassan and Donald R. Hill, Islamic
Technology, Cambridge, 1986, 235: <<The alum of Yemen was famous for
its quality... Egypt was also a major production center for alum and for
natron,,.
c. Nushadir = ammonium. <<Ammoniais colorless, gaseous, strongly

alkaline, with a characteristic pungent odor ...>>M. Sittig, Handbook of Toxic
and Hazardous Chemicals and Carcinogens, Pd ed., Princeton, 1985, 71 ff.
Here an ammonium mineral is most probably meant. There are various
types of ammonia compounds, the ammonium phosphate dibasic is used
for <<fireproofingtextiles, paper, wood. ..>,,The Merck Index, op. cit., 561.
d. Shadhinah or hajar al-shadinah = hematite. Hematite (iron trioxide) is used

as pigment and as a polishing powder. <<The most important ore of iron...
It is the major red-coloring agent in rocks and is a common cementing
material for sediments...>>Paul B. Moor, <<Hematite>> in McGraw-Hill
Encyclopaedia of Science and Technology 8 (1992), 382.
e. Jipsin or jibs = gypsum (hydrated calcium sulfate). It is calcined at 190-

200°C to produce plaster of Paris. <<Themost common sulfate mineral...
gypsum deposits act as a seal for many petroleum reservoirs, preventing
the escape of gas and oil...,>Marc. L. Helman and 6. Charlotte Schreiber,
<<Gypsum>> in McGraw-Hill Encyclopaedia of Science and Technology 8
(1992), 264-265.
f. Bau I=urine urea. Urea is excreted in the urine as a major nitrogen

containing end product of protein metabolism. Used as fertilizer, animal
feed, drugs and plastics. M . . . Frequently deposits a precipitate of
phosphates... It consists of approximately 95% water and 5% solids. Solids
include... ammonia (0.3 to 1.0)... magnesium (0.05 to 0.2), phosphorus
(2.0 to 2.5)...>> Taber's Cyclopedic Medical Dictionary, Philadelphia, 1981,
1516-1517. See also R. Straus' article <<Urine>,in McGraw-Hill
Encyclopaedia of Science and Technology, 19 (1992), 125.
g. Naft, often accompanied or replaced by the term nar (=fire), in a broad

sense corresponds to the Greek term bypov nOp or ~ ~ ~ L nOp K O (.liquid
V
or Persian fire,,) but it also means a petroleum substance. See my article
SHIPS AND GARMENTS
-Naft>,in Encyclopaedia of Islam, 2nded. (1991), 884 ff.

I have acquired access to the above-mentioned material through the
help of N. Vardatsikos and C. Canavas, to whom I express my gratitude.
Conclusion
The present paper confirms the author's view, which unfortunately has
been ignored, that it is only possible to understand the various formulae and
uses of Greek fire by studing the Arabic and Byzantine (Greek and Latin)
sources in conjunction.
Scattered references to protectives against Greek fire appear in early

Greek and Latin sources, but little is to be found in the Byzantine sources
since the relevant chapter of the tuMoyfi TQKTLKOV is missing: perhaps it
was censored for security reasons. The Arabic and some Persian sources
not only contain many references to the protectives and extinguishing
agents used on war machines and in ships to combat Greek fire, but also
record the use of special fireproof garments.14A passage from Ibn al-Manqali
with a full description of such a garment is presented here for the first time.
I have pointed out before that Ibn al-Manqali was a practical army officer, not
a theoretical scholar.15 In the short passage on fireproof clothing he
succinctly describes the materials needed for fireproofing garments and
then gives a brief description of one of the many types of Greek fire, which
is prepared separately and then poured over the fireproofed garment.
Although Ibn al-Manqali does not say who actually used the fireproof
garments (libas a/-nafattin), it is reasonable to assume that they were used in
both the army and the navy. Arabic manuscripts contain crude illustrations
of firefighters wearing their protective clothing (Fig. 2). N. Orphanoudakis
confirmed the practicability of Ibn al-Manqali's instructions empirically. I
would like to take this opportunity to thank him for his help.
Obviously the face was the most vulnerable part of the firefighter's body.
That is why the written sources suggest that Greek fire should be aimed
directly at the enemies' faces, a practice that is also attested by artistic
evidence (Fig. 1).
Vassilios Christides
University of loannina
Institute for Advanced Study
VASSlLlOS CHRISTIDES TROPIS VI
NOTES
* This article was mainly written at the Institute for Advanced Study, in Princeton, N.Y.,
during the time I was a visiting member with a Fulbright grant. There I collected an important
part of the material used for the Kuwait project. I would like to thank Professors A. Grabar and
G. Bowersock for their kind invitation. This article is reprinted from the booklet Sailing Ships of
the Mediterranean Sea and the Arabian Gulf (Athens, 1998), edited by Christos Makrypoulias.
1 J.R. Partington, A Histow of Greek Fire and Gunpowder, Cambridge 1960, 5.

2. We have little information about Kallinikos who provided the Byzantines with an advanced
form of Greek fire used by them in the first Arab siege (674-680). He came from Heliopolis
of Syria. According to Salah H. Al-Abidi, nAl Qadhai'f al-Nariyah wa'l Barudiyah al-
Arabiyah., Majjalah Kulliyyah al-Adab 23 (1978), 56, and D. Olster, ~~Theodosius
Grammaticus and the Arab Siege of 674-78*, Byzantinoslavica 56 (1995), 27, the Greek fire
was first invented in Arab-controlled Byzantine territory. Nevertheless, both authors failed
to understand that there were many variations of the Greek fire and Kallinikos perfectioned
an invention which existed before him. See my article ~Naft,~ in EP, 7 (1991), 885.
Unfortunately, there is no comprehensive work written on the Greek fire.
3. A. Dain, ed. li/Mom' Taxrlxdv, Paris, 1938, 114.
4. Ibid.
5. Probably al-Manqali must have used such sources.
6. Leo the Wise, Tactica, ed. J.-P. Migne in Patrologia Graeca 107, 893, C-D-896, A. Leo
reports that the Byzantine wooden towers were protected by hides and other (not
specified) stuff. In the Paschal Chronicle, ed. L. Dindorf, I, Bonn, 1832, 719, 21-22, it is
mentioned that the Avars covered their war machines with hides to protect them from the
fire.
7. See examples in M. Mercier, Le Feu gregeois, Paris, 1952, 56-57.
8. For this expedition see F. Gabrieli, <<Muhammad Ibn Qasim at-Taqafi et la penetrazione
arabe nel Sind>>,in Rendicontidelle Sedute dell' Accademia Nazionale dei Lincei, ser. VIII,
fasc. 7-12, 20 (1965), 345-362.
9. M. Quatremere, ed. and trans., Rashid el Din's Histoire des Mongoles de la Perse, Paris,
1836, 132.
10. 1. Kinnamus, En~ropded. A. Meineke, Bonn, 1836, 283, 1-19.
11. See note 7 above.
12. V. Christides, .Two Parallel Naval Guides of the Tenth Century: Qudama's Document and
Leo Vl's Naumachica: A Study of Byzantine and Moslem Naval Preparedness,,, Graeco-
Arabica 1 (1982), 71.
13. ' Ibn al-Manqali, A/-Ahkam fi Fan al-Qital fir/Bahr al-Mulukiyah wa'l Dawabit an Nammusiyyah,
ed. Abd al-Raim, Cairo, n.d., 124. For this Arab author see my article, dbn al-Manqali
(Mangli) and Leo VI: New Evidence on Arabo-Byzantine Ship Construction and Naval
Warfare., Byzantinoslavica 56 (1995), 83-96.
14. A vaguely described fireproof cloak for horsemen appears in an Arabic treatise published
by C. Cahen, nun traite d'armurerie compose pour Saladin*, BIFAO 12 (1947-8), 147.
15. Christides, ~ I b al-Manqali
n (Mangli)., op. cit. 84-85.
ILLUSTRATIONS
Fig. 1. lllumination from the Skyltzes' manuscript. It shows Greek fire launched directly on the
faces of the enemy crew. (Drawing simplified by A. Babuin).
Fig. 2. lllumination from an Arabic manuscript depicting firefighters protected by fireproof
garments. Oriental lnstitute of Leningrad (St. Petersburg). Courtesy photo.
SHIPS AND GARMENTS
Fig. 1
Fig. 2
SHIPS AND TRIPS IN THE ODYSSEY
Topography and communication in the Odyssey are determined to a

great extent by navigation. An analytical presentation of all the facts that are
contained in the Odyssey regarding ships and their trips is neither intended
nor plausible; besides some of these facts have already been studied. This
study tries to focus on some ships and sea journeys or moves connected
with particular circumstances.
Much of the evidence provided in the Odyssey about the voyages of the
ships is included in the part of the poem where the return of the heroes of
Troy to their hometowns is described. This description is made, as is known,
by Nestor and Menelaus in Pylos and Sparta respectively, where they are
visited by Telemachus on his attempt to gather information about his father.
In Menelaus' narration (whose knowledge also stems from Proteus' words)
the return of Locrian Ajax is included (4.499-511)'. Ajax was shipwrecked on
the Rocks of Gyrae and was lost there together with his ships. These
boulders are placed in the strait between Mykonos and Tenos. In the
southern part of Tenos there was a mountain named Gyras mentionned by
ancient sources2. According to the descriptions in the Odyssey Poseidon
struck the rock where Ajax had taken refuge with his trident and tore it off
from the remaining rocky mass because he had heard the hero boasting
about having been rescued from the sea against the gods'
will.CharacteristicaIly even in this part of the Odyssey the wreck is the result
of an unholy act, the most direct parallel being the punishment inflicted upon
the mates of Odysseus. The true reason behind Ajax's punishment is
Cassandra's rape in the temple of Athena in Troy, an act which justifies the
goddess' rage against Ajax. In the Odyssey Athena's rage is mentionned
without reference to the cause.
Ajax's shipwreck and the area where it is placed is of interest to us for

two reasons. One reason is that it implies a voyage possibly longer than the
others mentionned in the Odyssey for the return from Troy. This route goes
through the Cyclades after following the coasts of Asia Minor and although
it is longer it may also be a safer one, as it offers easier access to many ports.
MENELAOS CHRISTOPOULOS TROPIS VI
The other voyage (3.169)3 leaves the coasts of Asia Minor at the point of
Lesbos and passing above Chios reaches South Euboea (Geraistos); it may
thus ensure a smaller distance, yet it does not evade the danger of a tempest
in the open seas. The second reason why we are interested in Ajax's
shipwreck is the exceptionally concise manner by which the loss of all his
ships is stated (only one verse, 4. 499). From the lliadic <<Catalogueof
Ships. we know that these ships were forty (as far as one can regard the
numbers of the Catalogue as authentic)" This exceptionally brief destruction
of forty ships in the strait between Tenos and Mykonos simply caters to the
needs and priorities of poetic narration. The brevity of this description
possibly reflects the somewhat synoptical presentation of the Lokrian entry
in the Iliad (2. 327-335); yet the realistic element that must be brought to
attention is precisely the very same point at which the destruction is located,
since the existence of gusty winds in the Tenos-Mykonos area is equally
acknowledged by both ancient and recent sailors.
Another interesting route is the one followed by Menelaus, described by

Menelaus himself (4. 351-581) and also by Nestor (3. 276-312) who travelled
with him up to a certain point of the journey. As it has already been
mentionned, this route goes through the northern coasts of Chios, the
Southern coasts of Euboea and the Eastern coasts of Attica until Sounion.
At that point the poet places the death of Menelaus' ship captain, whose
name was Phrontis and whose great skill in governing the ship in times of
rough seas had distinguished him from his peers. Menelaus' ships continue
their journey after Phrontis' burial and reach Cape Maleas, where they
encounter a violent tempest which raises waves as big as mountains. The
tempest separates the ships and takes most of them to Crete where they are
destroyed and five more (among them Menelaus' ship) to Egypt. There are
several problems regarding the precise location of the southern Cretan coast
where the ships became wrecked5.A. Evans6supported the view that it is the
area around the port of Comos (Kwpoq). In this case the reference to the
Cydonians, who inhabit the area around the river Lardanos (3. 2921-293,
probably today's Platanias), most likely serves the narrative needs of
persuasiveness, which demand some known topographic details, even if
these do not exactly pertain to the area of interest, that is the southern shore,
but to the western part of Crete, the vicinity of Chania where, according to
Strabo, the Cydonians lived. I tend to consider that, apart from the needs of
narrative, these topographical details also imply the route of the ships that
reach the southern end of Crete, passing its western end. What remains
interesting in this narration is once again the realistic element of the tempest
which is located in a geographical stigma famous for the winds that arise
there, the Cape of Maleas. As far as one remembers this is the only division
of a fleet in two parts to be attributed solely to the effects of the winds.
Another characteristic detail of the narration is the point where Menelaus tells
of navigating his five ships along the flow of the Nile7.One bears in mind the
fact that it is about the same ships that cross the Aegean and the Libyan
seas and are considered suitable for sailing on river waters.
We may now have to make reference to another journey, one of the most
important in the Odyssey, which at times many have considered to be
unnecessary and useless for the development of the plota. This is the journey
of Telemachus to Pylos and from there to Sparta. The journey this time hides
no particular surprises; the ship reaches Pylos, most lilkely having sailed
through the western Etolian coasts, passing in front of the opening of the
Corinthian Gulf and then following the western coast of the Peloponnese.
Telemachus' trip to Pylos from Sparta takes place on a carriage. This trip on
land takes over a day and requires an overnight stay which occurs,
according to the passage, at Pherae, a place probably situated within the
area of today's KalamataQ.During Telemachus' absence the ship with his
mates remains at Pylos and waits to carry him back to Ithaca. In the
narrative, the trip from lthaca to Pylos lasts slightly over a night; similarly so,
lasts the trip from Pylos to Ithaca. If we count the days of action in the
Odyssey they will amount to forty. Approximately thirty out of these forty days
Telemachus spends in Sparta (even though he claims he is in a hurry to
leave, 4. 594) because, if my calculations are correct, he arrives at Sparta on
the sixth day and leaves the thirty-fifth. The basic reason for this delay of
Telernachus is that arrival of his on lthaca must coincide with Odysseus'
return. The most dangerous point of Telemachus' return is the one where the
ambush of the suitors has been laid, where an attempt on his life is made by
the latter. This point is identified in the text as <<thestrait between lthaca and
the rugged Samos,, (4. 671). The Samos referred to here is probably the
Same of Cephallenia. Telemachus' ship sails by the shores of Triphyllia and
continues its route to the islands named 0oai vqao~(sharp(?) islands) which
Strabo (8. 3. 26) perceived as <<pointed,,(perhaps the southern Echinades,
from the meaning <<sharp,,of the word 000q which refers to the sharpness of
the thorn). The mystery of the identification of these islands remains
unravelled. If I were to risk an interpretation I would say that the interest of
the text is not focused on topography but on Telemachus' imminent death in
the ambush. The word 800q may well mean <<sharp,> in this passage of the
Odyssey as the word 800oa ( 8 0 0 ~means ) sharpened,, also in 9.327. The
<<I
meaning ((sharp,, often occurs in later epic poetry (e.g. Apollonius Rhodius,
Arg. 2. 79: 0ooiq yopcpolq, 3. 1281: 8oOv O ~ O V T ~ V4., 1683: 8ooiq n&AC-
MENELAOS CHRISTOPOULOS TROPIS VI
KEUULV etc). If we accept <<800q> to mean <<sharp>> here, this meaning does
not concern the shape or the name of the islands but their property of
Telemachus' potential <<murderers>>. In this sense neither lthaca nor
Cephallenia are excluded, since between them, on the small island of
Asteris, the suitors have laid their ambush. The argument that the isle of
Daskalio, situated in the strait, does not quite offer the geophysical formation
required for such an ambush, is not strong enough. lthaca and Cephallenia
remain the most probable <<signifies,>of the expression v l j a o ~ o8onu~
~ (15.
299). Thus Athena's advice to Telemachus urging him to keep the ship away
from the islands ( 6 ~ a qvljoov, 15. 33) has the expected meaning since
Same and lthaca were mentioned four verses earlier (15. 29).
Telemachus' journey takes place via a ship which leaves lthaca at night
(in stealth) and arrives at lthaca at night (in stealth). Many are the issues
which corelate this journey of Telemachus with Odysseus journey of return.
The stealthy arrival of both ships and even with such economy on time that
Odysseus and Telernachus meet by chance and in secret at Eumaeus' hut,
is one of these issues. Another issue, in my opinion more significant and not
extensively discussed, is that both ships have foreign owners. The ship that
brings Odysseus to lthaca is, as we know, the ship of the Pheacians who
specialise in such missions and have even aroused Poseidon's wrath
because of this. Telemachus' ship is borrowed by Noemon, an lthacan
whom Athena, transformed as Telemachus, persuades to entrust his ship to
the latter (2. 381-387). The ship is manned with noble youths from Ithaca.
Noemon himself becomes the cause of the suitors' ambush after naively
betraying Telemachus' departure to them (4.630-656). Thus Noemon's
name (= <<intelligent>,)may acquire a particular meaning and I believe that it
is used with a degree of irony, not unknown in the Odyssean contextlo.
The use of borrowed or rented ships is not uncommon in antiquity. Yet

the lack of ships suffered by Telemachus and the destruction of ships in
Odysseus' voyage are generally connected with some peculiarities that
pertain to the ships in the Odyssey. Odysseus' voyages were interpreted in
all eras with indescribable imagination and the scholars having studied the
Odyssey, both ancient and modern, have placed his adventures in just about
every latitude and longitude of the world known then1'. Still one has to focus
on the trips and the places, imaginary or not, mentioned by the poet himself.
It is interesting to observe that in the Odyssey Odysseus returns to the point
of his departure a total of four times12.The second of these times is the return
to the island of Aeolus. I have prieviously had the chance to support that in
the Odyssey Odysseus leaves Troy as leader of a fleet, loses all his ships
(except his own), loses his last ship, constructs a raft, seizes a floating plank
and ends up in the sea naked, having gone through all the stages of
navigation inversely13.There is a strange echo of this idea also in the Aeolus
episode. A skin-bag containing all the winds (except the West Wind) is
offered by Aeolus to Odysseus to make the latter's trip safe. When Odysseus
falls asleep, his companions taking the skin-bag to be full of treasures and
anxious to have their share of those treasures, open the skin-bag and the
winds bring Odysseus' ships back to Aeolus' island. Aeolus' skin-bag is, of
course, offered as a gift, but the issue of the skin-bag on its own belongs to
a particularly enriched whole of anthropological data. I let alone the
Dionysiac aspect and the Marsyas myth, both connected with the skin-bag
but not particularly relevant to the Odyssean context. Still there is another
issue of this theme which should be mentionned here. It is about the inflated
skin-bag used in sailing. First of all this method allows a man to float having
grasped such a vessel and in more advanced form it can lead, as we know,
to the creation of a raft of the type used in Mesopotamia (kelek) and in the
creto-mycenian world. Skin-bags could often be replaced by empty
amphorae. In a Boeotian skyphos of the beginning of the fourth century B.C.
we see Odysseus crossing the sea on two amphoras14. It is not recognised
in this depiction, the direct effect of the Aeolus episode but perhaps a wider
Odyssean idea, determined by the blowing of the wind and the act of
Poseidon, stated implicitly through the trident. In the incident with Aeolus
Odysseus still has all his ships. The next episode in the Odyssey, the
encounter with the Laestrygonians, is the critical point where Odysseus
loses all his ships except his own. In the Aeolus episode the skin-bag seems
to maintain something from its wider semantic field and if it does not become
a vessel of sailing it remains at any rate an aid to sailing.
Where does the issue of the loss of ships lead us? In relation to the
Argonautic myth, the other great - and probably pre-homeric - epic of the
Greek antiquity, we observe that in the Argonautic expedition what is
important is the ship itself as a ground, the ship Argo, which gives its name
to the myth itself. On the contrary, the name of Odysseus' ship is not of the
least importance. No one ever wanted to know what it was. In the Odyssey
the ship is precisely the means, not the aim. Thus another peculiarity which
I think has not been given enough emphasis and which I have tried to
demonstrate through the borrowing of the ship from Noemon to whom
Telemachus resorts and through the issue of the ships Odysseus loses, it is
generally the absence of ships which is noted in many other parts of the
poem (Nausicaa, Pheacians) and which is almost always connected with the
lack of means, inability to move, or lack of power. Odysseus led twelve ships
MENELAOS CHRISTOPOULOS TROPlS VI
to Troy. Eleven of these were destroyed by the Laestrygonians, the twelfth,

which he had boarded, was destroyed after the slaughter of the sacred cows
in Thrinakia. Naturally these ships were warships. Whether they all belonged
to Odysseus or whether as king of the lthacans he steered their ships is part
of the circumstances that define the complex political problem of Ithaca. The
Odyssey seems to imply that the ships belonged to him. There are lthacans
who are owners of ships; whether they are warships or not is not known but
at any rate they are suitable for an ambush. Such lthacans are Noemon, who
has already been mentioned and the suitors, who have no problem in finding
a ship to waylay Telemachus. Among those lthacans who own ships
Telemachus is not included, contrary to what one would expect. The
absence of ships in Odysseus' wealth that is stated constantly and in every
way (together with the diminishment of his wealth by the suitors and the
latter's stated economic power) I believe should be added to the causes that
render the power of Odysseus' family in lthaca problematic and not self-
evident. Far from its purely practical use as a means of transportation at sea,
which I believe to be determinative in the Odyssey, the ship seems to have
already been evaluated and to function effectively as a means of power.
Menelaos Christopoulos
University of Cyprus
NOTES
1. Od.4.499-511
A'iaq pkv p s ~ vquoi
a 6apq 6ohl~qp6~polal.
rupnoiv p ~ npQTav fl00~166w~ bn6Aaao~v
nC~pna~v peyMnot ~ aCic ~ o a w of3aAaocqq.
~
~ avui KEV EK@UYE ~ q p ~ a aCi~ B o p ~ v o
nsp
q 'Aerjvn,
~i p i Ijncp@iaAovEnoq EK6ah& ~ apCy' i aao8~
@fip ' a k ~ ~q E~ Q ~QUYCELV
V pCya Aai~paBaAaocqq.
TOO 6b ~ O U E ~ ~pEyah' ~ O V PKAUEV ( I U ~ ~ ~ U ~ V T O ~ .
~ O T ~ KE' n c ~ ~pialvav
~a thhv x ~ p m16apTjolv
d
ijhaoev rupaiqv nb~pqv,Cmo 6' E q ~ a e v air-njv.
~ aTOi pbv air~o81 psive, TO 6b ~ p l j @ oEpneas
q nov~y,
T@ p' Aiaq TO npOTov C @ E < O ~ E V Op6v ~ aaaeq.
TOV 6' b@6pst~ a n o ~v ~ao vansipova ~ u p a i v o v ~ a .
LSq 6 pbv Pv0' anoAoAev, bnei nisv aApupov I36wp.
2. Some scholiasts of the Odyssey explain the name of these rocks by their shape: sch.Hom.
Od..4. 500: yupna~vnd~palqrrhqoiov Mu~ovou~ f i qvrjaou OUTU Kahoup6valq bnei siot
nspt@~psiq (D). One wonders whether this name reflects the word K u K ~ ~ ~used, E s , as we
know, for all those islands which were thought to be situated in a circle around Delos, cf. sch.
Hom. Od..4. 500: yupai n i ~ p asioi l nspi Tilv Mu~ovovnhqaiov, Mlj~ovoq6t ~ aNacoq i
TQV K U K ~ ~v ~ r j oW~ vVHesychius
. mentions the mount rup& (s.v.): opoq bv Trjvq.
3. Od. 3. 168-183
oqb 66 63 p ~ vat ~ Ka ~Etaveoq M~vbAaoq,
bv AbaBy, ~'EKLXEV 60At~ovnhoov oppaivov~aq,
ij ~ a e l j n ~ xioto
p e ~ v~oipeeanamahobaqq
vljaou Ent Ulupiqq, aljnjv C d aptmbp' EXOVTE~,
ij unbv~pesXioto, nap' l j v ~ p o ~ Mipav~a. v~a
~ T ~ O 66 ~ E~ E VO VQuvat ~dpaq.a b ~ a p6 y' i j p ~ v
6&ic&,~ aqvhyst i ndhayoq pbaov ~ i E66o~avq
T E ~ V E ~ V6, ~ p T&LUfa
a 6n6K K ~ K o T I ~@byotpEv.
T~
G ~ T 6' O Cni Atyirq oSpoq a j p ~ v a tai . 66 p M G K ~
i x B u o ~ v ~KCAEU~OV
a 6tC6papov, Cq 66 r ~ p a ~ m o v
~ V V U X IK~~~T ~ ~ O V no(~Et660~t
T O ' 66 T ~ ~ J ~ o v
n o w C n i pip' E ~ E ~ E nbAayoq
v, pdya p e ~ p j a a v ~ ~ q .
T C T ~ ~ T Oi Vp a p Eqv, o i Cv 'Apyei' vqaq Ciaaq
Tu6&'i6&oE T ~ ~ Atoplj6~0q
OL imo6apoto
imaaav aljrap b y h y ~llljAov6' Exov, ob6C n o t Ea6q
ocpoq, Cnsi 63 npaTa 8 ~ b npoCq~ev q ajvat.
4. 11. 2. 526-535
Ao~pQv~ ' ~ ~ E ~ O V 'O'iAiioq
E U E~Va x i r q
Aiaq,
peiov, 06 TL ~ o a o q YE Oaoq T~Aaphvtoq Aiaq,
Wa nohir p ~ i o ohiyoq v p6v fiv, Anoehpqt,
C y ~ ~6'i nC ~ k ~ a m IlavCMqvaq
o ~ a'Axa~oljq
i
oi KOv6v T' Cvbpovi'Ono~v~a TE KaMiapov TE
Buaaav TE Z K ~ ~TEQK ~Q Aljy~taq ~V Cpa~~tvaq
Taptpqv TE O ~ O V TE L ~Boaypiou
V aptpi pdsepa.
TQ 6'8pa ~ ~ a a a p a ~ pbhatvat ov~a vu~q Enov~o
Ao~pPv,oi vaiouat ndpqv i ~ p u q 'Eu6oiqq.
It has often been brought to attention that, in this description of the Lokrian entry, the poet's
main interest is in Eastern Lokris whose contigeut he tries to strengthen, the Western part of
Lokris not being mentioned at all in the Iliad. Besides, some places usually associated with
Lokris, such as Alos and Alope, are cited in relation to Achilles' jurisdiction (11. 2. 682). Many
scholars believe that several problems in the Catalogue of Ships are due to interpolation partly
caused by some singers' will to reinforce their cities' entries in the Catalogue orland by
contemporary (to Homer) elements introduced to older, traditional entities of the Catalogue. A
typical example where one faces this kind of problems is the Boeotian contingent. The whole
problem is one of the most complicated ones in the study of epic poetry; for a synopsis see G.S.
Kirk, The Iliad: A Commentary, vol. I, Cambridge 1985, pp.168-178 and, in particular (for the
Lokrian entry), pp. 201-205. The Lokrians went to Troy with forty ships (11. 2. 534), as did the
Phocians (whose contigent precedes the Lokrian entry in the Catalogue) and the Abantes
(=Euboeans) whose contigent follows the Lokrian; 2. 534 stating that Ajax' ships were forty is
#a standard ship-number verse* according to Kirk (op.cit. p. 203).
5. Od. 3.286-300
'AM' ~ T 63 E ~ a~ i~ i v ihvo q C n i oivona n o v ~ o v
Cv vqud yAa@upna~ MaAetaov 6poq ainir
~ C O VTOTE, 63 m u y ~ p r j v660v ~bpuonaZEO~
Ctppaaa~o,AtyCov 6' avbpov bn' au~pbvaXEOE,
~ l j p a TE~ aT~O@COVTO nchhp~a,iaa b p ~ a a n .
Evea 6ta~pljtaq~ a p8v q Kplj~n CnbAaaa~v,
fix1 Kbbov~q Evatov 'lap6avou aptpi bCeepa.
E m 1 66 TIC Atafi ainsia TE ~ i 8Aa q n6~pq
MENELAOS CHRISTOPOULOS TROPlS VI
EVOj s
& O ~ T ~L O~ ~ T U V ~p o s d b ~
~OVT(I?
Ev8a NOTOT pCya KOpa n o ~oi~ a t o vpiov beer,
dq @atmov,p l ~ p o q6b Ai9oq pkya K O ~ ano&pyel. '
ai pbv 6p' eve',jA9ov, anouan 6'fjAucav dAe9pov
Bv6psq, a ~ a p vqaq ye n o ~anlha6soo1v
i Eacav
K U ~ ~aThpT ' . Thq ~ E V T E~ C a qK U ~ V O T I ~ C $ ) ~ E ~ O U ~
AiyUmw ~ R C ~ ~ U @Epov OE avspoq TE ~ aii6op. i
6. A. Evans, The Palace of Minos, II, London 1928, p. 86.
7. Od. 4. 576-586
jpoq 6' fiply&v~la @avop06066~T~hoq 'HCi)S,
vqaq ptv napnpo~ovkpOooapev eiq 6Aa 6iav,
Cv Fimoirq ~i9Epea9a~ aimia i vquoiv Eiwq,
av 6 t ~ aa il j ~ o6av~eq
i bni ~Aqiat~aOi<ov.
bcqq 6' b<opevot nohlfiv M a ~ l j m o v Cpe~poiq.
aw 6' eiq Aiyljmo~o~ I L ~ E T Eno~apoioO ~
mqoa vdaq, ~ aEpeca i ~eAq6oaaqbKaTop6aq.
a xoAov aibv Cov~ov,
a u ~ h pCnei ~ a ~ C n a u a9eav
XEO' 'AyapEpvov~~dp6ov,iv' 606emov K A C O ~siq.
T ~ O T&A&uTfiCIaq
T ~ V E O ~ E60oav
~ V , 615 pol O ~ P O V
a9ava~o1,TO^ p' ch~a@hqvEq na~pi6'Enepwav.
For Odysseus' ships sailing on the Nile in Odysseus' false stories see Od. 14. 257-261.
8. Cf. S C Od.
~ 1.93: 6 ~ 0 6~0 0~ eival ~ 1 TqAepaxou fi Cmo6qpia ~ P ~ T~OE V ~iv6uvov
V npo-
cevoooa T@ vEq, ~ E ~ T E P O Enavamao~vV T ~ Vpvll-pwv anelhoOoa, TP~TOV OOK
b@eAoOaaT ~ MT~UIV V TOO na~poq.
For a general survey of the special problems connected with the ~Telemachy~ (=books 1 to 4
of the Odyssey) see S. West, A Commentary on Homer's Odyssey, vol. I, pp. 51-66.
9. Od. 3. 487-490. For the historicity of such an itinerary see W. A. MacDonald, <<Overland
Communications in Greece during LH Ill. in Mycenian Studies, Madison, 1964, pp. 217sqq.
10. It is certainly not by chance that the name of Noemon's father is Opovloq (2. 386), but in this
passage the use of both names is not ironical. For the name Noemon cf 11.5. 678, 23. 612.
11. These theories are depicted on a map illustrating Odysseus' wanderings in H.-H. and A.
Wolf's book : Der Weg des Odysseus, Berlin, 1975 (reproduced in I. O. K a ~ p d f i(ed.),
E M q v i ~MutloAoyia,
~j Athens 1986, vol. 5, p. 253), pl I (courtesy of EkdotikiAthinon).
12. The first time is placed at the beginning of his journey and is mentioned by Nestor who tells
Telemachus that Odysseus left with him from Troy and went as far as Tenedos but returned
to Troy where Agamemnon and the other leaders had remained to offer sacrifices to Athena
(Od. 3. 155-164). The second time is Odysseus' return to the island of Aeolus when the storm
which broke out after the opening of the skin-bag from his mates takes the ships back to
where they had set off (Od. 10. 46-76). The third time is the return to the island of Circe after
Odysseus' visit to Hades (Od. 12. 1-36).Finallythe fourth time is the involuntary return to the
strait of Skylla and Charybdis after the destruction of Odysseus' ship (12. 426-446).
13. [[Le m&t du navire. RBalite et imaginaire en Grece ancienne.., Proceedings of the 3rd
International Symposium on Ship Construction in Antiquity,Tropis Ill, pp. 123-134, .To msvo
q q ZKOMaq Kal o 06uooCap, ApxaioAoyia, 38, 1991, pp. 50-57.
14. Oxford, Ashmolean Museum, 262, pl II (courtesy of EkdotikiAthinon).
PLANKING TENONS IN ANCIENT MEDITERRANEAN SHIPS
BUILT SHELL FIRST
INTRODUCTION
The shell planks of ships of the ancient Mediterranean and of some other
seas were fastened together edge to edge, by stitching, sometimes
reinforced with wood dowels (or pegs) set into the plank edges, or by tenons
often locked in place by pegs through the planks and each end of the
tenons, or by stepping (i.e. joggling) planks over each other to key them
together, or by combinations of these methods. These inter-plank fastenings
were at the heart of the construction of those shell-built ships in antiquity
over a period of more than three millennia and they have by now been
extensively recorded, yet their physical properties and the structural loads
upon them do not seem to have been much studied1.It will be explained in
this paper why sliding between planks tended to occur and had to be
prevented to keep hulls tolerably watertight. Before considering tenons, the
properties of stitching and dowels will be outlined. These are likely to have
limited the sizes of the hulls in which they could have been satisfactorily
employed. The properties of tenons are then discussed, indicating that
tenons would have been necessary in larger and in particular longer ships.
The paper draws some inferences from a simplified theoretical analysis of
tenons and an exploratory experiment carried out on moisture cycling of a
timber suitable for use as tenons, described in the annexes.
PLANK SLIDING
Large, or long and shallow, wooden planked ships have been well
known to be particularly prone to hog, that is, to deform so that their ends
droop. It is known that medieval and later galleys were built with a sag (the
opposite of a hog) in their keels, of as much as 300mm. or so, to anticipate
the hogging which would inevitably develop in service afloat. Hogging in
wooden ships is due to shear deflections of the hull, rather than bending of
the hull as a girder". The shearing forces arise because the ends of hulls
unavoidably have a greater weight than their supporting buoyancy, whereas
amidships, to achieve an overall balance of weight and buoyancy (to satisfy
the demands of Archimedes), buoyancy is greater than weight. The result is
that, over more than a quarter of the hull length from each end, the hull
JOHN COATES TROPlS VI
structure has to carry appreciable vertical shear forces tending to pull the
ends down. With very few relatively modern exceptions, those shear forces
have been carried in wooden hulls almost entirely by the shell, the planking,
and their effect has been to cause planks in the sides of hulls to tend to slide
upon each other, after the manner of the leaves of a leaf spring (Fig. 1).
Those forces have to be borne for so long as the ship is afloat, whether in
still water or among waves. On the crests of waves the shear forces are
intensified; when the ship is in the troughs of waves, they are commonly
reversed and smaller. Besides keeping planks in contact with each other,
preventing sliding between them is the main task of fastenings between
planks forming the sides of wooden hulls, particularly if the hulls are large,
or long and shallow.
Consequences of plank sliding

If the planks in a hull start to slide upon each other to any significant
extent, plank seams, whether filled with driven caulking, with applied
stopping or even with nothing, will soon leak. This must always have been
so, and the trouble would have particularly affected long and shallow ships,
and therefore oared ships, more than "round ships" unless they were large.
The need for ancient warships to "dry out" was probably to enable their
planking to be re-stopped and re-coated to reduce leakage besides
removing fouling and controlling attack by marine borers. To adhere, both
stopping and coating would have needed the surface, but only the surface,
of the planking to be reasonably dry.
STITCHING AND DOWELS
The limits of stitching and dowels in resisting sliding between planks

When sustained for a long time, quite modest loads cause timber and
natural fibres to stretch permanently. This is called "creep", and it occurs at
loads much smaller than those which the material can bear before breaking;
they are also smaller than shorter-lived loads causing the initiation of
permanent stretch. Planks have to be prevented from sliding (or at least
from sliding by an amount to cause unacceptable leakage) under, first, the
sustained shear forces acting between them for the whole time that a ship is
afloat, as well as, second, the larger shorter-lived but repeatedly applied
loads imposed when at sea among waves. The capacity of stitching and
dowels to prevent sliding under these two conditions is therefore central to
understanding the construction of sewn and dowelled ships.
PLANKING TENONS IN ANCIENT
MEDITERRANEAN SHIPS BUILT SHELL FIRST
Natural fibres, like timbers, are made of cellulose combined to varying

degrees with lignin according to species. Cellulose fibres and cells swell
laterally when wet and the effects of this on rope and timber are well known.
Flax, a strong fibre available in the ancient world, breaks at a tensile stress
of about 700 N/mm2,but ropes, on account of the additional stresses caused
by their helical structure, are not so strong as their constituent fibres. A flax
three-stranded cord breaks at a nominal stress of about 140 N/mm2,
stretches permanently at relatively short-lived stresses above about 20
N/mm2 and creeps at long-sustained stresses above about 6 N/mm2.
Timbers behave in much the same general way: oak and yew break at
tensile stresses of about 100 N/mm2,but they creep at stresses above about
5 N/mm2. These figures however vary appreciably with moisture content,
species and conditions of growth; they are therefore only very approximate.
Stitching
Under a nominal tensile stress of 5 N/mm2,a 6mm diameter flax cord, for
example, would carry a tension of 115 N, or 12 kg.f. If, to continue the
example, stitches were simply arranged as in Fig. 2 with the diagonal parts
at 49, the shear force tending to cause the planks to slide, but sustainable
by each stitch without creep, would be 12/42 = 8.5 kg.f. If planks were 25
mm thick and stitches 50mm apart, the plank shear stress sustainable
without creep would be about 8.5/(25 x 50) = 0.007 kg.f./mm2. That
sustainable shear stress could be increased by doubling the stitching or the
cord or increasing the diameter of the cords. Neglecting friction, which could
be large or negligible according to a host of variable factors, the planks
would slide when the stitches were so loaded by about 0.04mm, a
movement which would certainly not distress a seam, nor therefore be likely
to cause leakage.
How big were shear stresses likely to have been in, for instance,
triakontors, pentekontors and triereis? In a manned trieres, it can be stated,
from the calculated particulars of a reconstructed ship of displacement,
manned and equipped, of 50 tonnes, the maximum vertical shear force in
still water is about 9 tonnes force. In ships of a generally similar geometry
and loading distribution but of different size, as the other two types of oared
ships would have been, shear forces would have varied more or less directly
with displacement (or weight). In the trieres, it may also be calculated that a
vertical hull shear force of 9 tonnes force would generate a maximum shear
stress of about 0.048 kg.f/mm2 in the side planking, 7 times the stress
sustainable by the stitching taken as an example above. The trieres therefore
JOHN COATES TROPIS VI
appears to have been much too big a ship to have been fastened only by the
stitching postulated in the example above (or indeed by any feasible
strengthened version of it). If a trieres had been so stitched, each stitch, if
spaced 50mm apart, would have had to sustain without creep a tension of
12 x 0.048 = 82 kg.f.
0.007
What size of geometrically similar oared ship could have been stitched
as in the example, by these indications? As shear force is roughly
proportional to displacement, and as stitch tension would be proportional to
shear force, stitches as in the example and carrying a tension of 12 kg.f.
would appear to be adequate for a ship of displacement of about 50 x
12/82= 7.3 tonnes only. Thus the stitching taken as an example would be
adequate for a ship of about the likely manned displacement of triakontors3,
so, if stitched, they would probably, by this argument, have had somewhat
stronger or doubled stitching which would have been quite feasible. These
figures indicate the order of size, 10 tonnes, of oared ships that could have
been held together successfully by stitching only.
There were, however much larger stitched ships, e.g. the Cheops ship of
the mid-3d millenium BC4 (43m long overall and a loaded displacement
about 100 tonnes), but that ship also (and crucially) had joggled planking by
which one strake in the forward and after body was hooked over the one
below to resist sliding in hog, as well as some tenons, but it must have relied
mainly upon the joggled planking to keep its shape. Round ships, being
shorter and beamier, are less severely stressed than long ships and the
displacement limit for them to rely upon stitching only would be greater. Arab
booms (about 30m long and a loaded displacement in the region of 200
tonnes) have until quite recently relied upon stitching in their seams, with
some dowels which are said to be mainly to align planks while being
wrought and then stitched; they have deep hulls (twice as deep or more
relative to length than several types of oared warships), which would
considerably reduce the maximum shear stresses developed in the
planking.
Dowels, or treenails
Information about the behaviour of dowels or treenails (trennels, or
trunnels) when loaded in shear is hard to find. Brungraber' tested a number
of tenoned joints used in the oak frames of timber house construction. These
were locked by hard dowels and some of his tests tested the dowels'
stiffness in shear. Loads within which the deflection of a dowel was linear
with load was consistent (not very surprisingly) with assuming that the shear
load on the dowel on each side of the joint (or, in the case of a dowel
crossing a seam between two planks, on each side of the seam) was borne
by a length equal to one diameter and with a pressure equal to the
proportional limit of the dowel timber when compressed perpendicular to the
grain. Under increasing loads above that amount, deflections grew, as may
be expected, at an increasing rate, rising to nearly ten times the initial rate,
that is, the rate up to the proportional limit, as the dowel and the sides of the
dowel hole were crushed. This increase has been exploited for millennia by
deliberately misaligning the holes for tenon pegs so that when a conical peg
is driven through, it deforms and draws the tenon more tightly into the
mortice - a practice known as "draw-boring".
Brungraber found that the initial stiffness of a dowel (not draw-bored)

was 1.3 times greater if the dowel was set so that its annual rings were
across the direction of the shear load, than if they were parallel to it.
Brungraber's dowels were of America Red Oak (quercus robur) whose
proportional limit under that loading is about 0.56 kg.f/mm2, so his tests
indicate that a 15mm diameter dowel, of that oak and joining two planks side
by side, has a proportional load limit in shear of 0.56 x 15 x 15 = 126 kg.f.,
equivalent to 10 of the stitches taken as an example above, i.e. such dowels
could be spaced as far as 0.5m apart to have the same sustained shear
carrying capacity as those particular stitches. The deflection (in this case,
slide between the planks) at the proportional limit of the dowels would be
about 0.04mm, about the same as that for the above stitching, but the
capacity of dowels satisfactorily to carry greater sustained shear forces than
stitching is evident, suggesting that they may have been used in oared ships
longer than triakontors, that is if tenons were not used.
The bottom of the small (probably 12 metres long) 6'h century BC

Bon-Porte Ship6was fastened by dowels and stitching, the latter being, in the
bottom at least, wholly perpendicular to the seams and so incapable of
carrying any shear forces between planks, which would therefore have been
carried entirely by the dowels, which were about 9mm in diameter and
150 mm apart. Such dowels may have had a proportional load limit
(following Brungraber) of about 45 kg.f. Planking was 25mm thick, so the
shear stress in the sides of the hull (if the same as the bottom), which could
be carried by the dowels without exceeding the proportional limit of their
timber, would have been about 45/(25x150) =0.012 kg.f/mm2.This is nearly
twice the plank shear stress sustainable by the stitching in the illustrative
example taken above, suggesting a displacement for the Bon-Porte Ship of
the order of 7 tonnes if the above details applied to the sides of the hull. If,
however, as is quite possible, dowels in the vanished sides of the hull were
more closely spaced, a displacement of say 20 tonnes could be consistent
with this construction; if the stitching in the sides were in the style of the
illustrative example, as it could have been, a greater displacement again
would be plausible, because such stitching would carry part of the hull shear
forces across side seams.
The Kyrenia Ship's ceiling included some second-hand planking7which

had been edge-fastened by dowels, also about 150mm apart. That may have
been a commonly used spacing and it could have been adequate for the
very numerous small round ships built over several millennia. It may have
been closer than necessary to carry the hull shear forces in such ships if they
fitted perfectly and were therefore in any locality loaded equally. But such
perfection would rarely, if ever, have been achieved in reality; dowels would
have varied in diameter and the drilled holes would have been out of
alignment; water-tightness would therefore have called for more of them.
Moisture cycling
All wooden hulls enter and leave the water at various times in their lives,
causing changes in the moisture content of their components, particularly in
the parts under water. In timber under restraint, such as dowels fitted into
their holes, this moisture cycling causes loosening, first through crushing
under the restraint of the hole when the dowels are wetted (by inevitable
seepage, if not plain leakage) and then shrinkage when next becoming drier.
The way hammer handles get loose and wooden hulls leak until they "take
up" is well known. The exploratory experiment described at Annex B
demonstrated a loosening of 0.9% in tenon material after crushing, wetting
under constraint and then drying.
In practice, therefore, some looseness will occur in a pegged (or

tenoned) hull no matter how perfectly fitted those fastenings were when the
hull was built. The effect of moisture cycling can be about halved, in the case
of many timbers, by placing the annual rings of dowels (and tenons) across
the direction of the crushing forces, taking advantage of the smaller
shrinkage and expansion of timber with changes in moisture content in the
direction perpendicular to the annual rings (i.e. radially), compared with the
movement tangentially to them. That orientation is also that which
Brungraber found to give the greatest stiffness in dowels under shear up to
the limit of proportionality. It is reasonable to suppose that, by long
experience, boat builders have for long been aware of these effects.
Driving dowels and treenails

It is likely that, as in latter years, dowels and treenails were driven well-
greased and with an interference fit as great as could be conveniently
overcome in driving them home without being crippled or otherwise
damaged. Curtiss gives an interference (i.e. the difference between the
diameters of hole and the slightly larger treenails) of 1.5mm for hardwood
treenails 30 to 40mm in diameter, which is 0.04 to 0.05 of the diameters.
When fastening planking edge-to-edge however, to avoid splitting the plank
in driving the dowels, a smaller proportionate interference for dowels might
be necessary. A tapered lead of reduced diameter at the ends of dowels
would of course be needed, not only to keep the surfaces well greased while
being driven but also to allow for some misalignment between the dowels
across a seam as they are being driven simultaneously into the upper plank
to close it. As has been remarked by Markg, building with dowels and
stitching needed fewer tools and less skill and knowledge than building with
tenons, but stitching would have needed renewing, quite possibly every
year, while tenons would have needed less maintenance. Tenon
construction, it is argued, was therefore more suited to a more specialized
society.
Pentekontors, particularly the longer single-level kind, being more than

twice the displacement (18 tonnes) and one and a half times as long as
triakontors, probably relied upon dowels, or tenons, to carry the hull shear
forces. Tenons date at least as far back as the mid-3rdmillennium BC; the
Cheops Ship4 had them but only in a few places and not for carrying hull
shear forces. The wreck of a cargo ship of about 1400 BC at Cape Uluburun
was built with pegged tenons, which is evidently an ancient technique which
became the standard method of joining planks edge to edge for almost any
purpose in the ancient Mediterranean world.
TENONS
Pegged tenons have the obvious advantage of holding planks together

positively as well as resisting shear between them. Nevertheless there are
cases6 of ships in which stitches were used as well as pegged tenons. As
Markg has said, tenons demand considerable skill to fit accurately,
particularly in ships which have appreciable shear forces in their hulls, to
carry which a tight and simultaneous fit in the fore and aft direction of tenons
in their mortices is necessary. At Annex A is a simplified general analysis of
the forces and pressures acting on this type of joint in resisting shear
between the planks joined.
Advantage of tapering tenons

Tenons found in some wrecks had each half tapered so that the tenon
had the shape of a double trapezium, as shown in Figure 3 ; in others they
have been parallel-sided, i.e. rectangular. Tenons, and their mortices, were
probably tapered to help in obtaining a tight fit in the fore and aft direction.
By marking the faying surfaces of the mortices of the upper plank of a seam
with some form of paint, charcoal or chalk, and offering the upper plank over
the tenons already set in the plank below, those making contact with them
will be marked. Those could then be pared a little where marked, the
mortices marked again and the process repeated until an acceptably
uniform fit and bearing had been obtained when the gap between the planks
was such that, by experience, it could be closed by imposing a degree of
crush upon tenons and mortices depending upon the gap, taper and the
timbers used.
It would probably have been necessary to lock the tenons in place

immediately after closing the seam, by driving the pegs whose purpose in
holding the seam together would have been the more securely performed if
the holes for them had been "draw-bored" with off-sets (or deliberate
misalignment) of 114 to 113 of the diameter before the tenons were
introduced into the finished mortices: to obtain accuracy in draw-boring, it is
likely that some forms of drilling jigs were used. Tapered tenons may be
regarded as evidence of a superior standard of construction, essential in
large warships but quite possibly applied also to the better class of merchant
ship, built, for instance, for a long intended life. Such a process would not be
possible with dowels or rectangular tenons, acceptable in smaller and more
cheaply built ships. Taper of tenons would not appreciably affect the analysis
at Annex A.
As the fit of tenons in the direction of the plank thickness was not
important, the alignment of mortices in bevel need not have been critical,
whereas the alignment of the necessarily drilled and therefore circular holes
for dowels was. This must have been a point of added difficulty in building
with dowels, which did not apply to tenons.
Loosening by moisture cycling

Tenoned plank seams would have been prone to loosening owing to
moisture cycling. Tenons (and dowels), if constrained in tightly-fitting
mortices (or holes in the case of dowels), would be crushed a little when they
expanded owing to being wetted. When subsequently dried out, the crush
would not be recovered as the timber of the tenon shrank. A gap would be
formed, as illustrated by the loosening of the hafts of hammers in hammer
heads. At Annex 6 is a description of a simple exploratory experiment carried
out by the author to verify this behaviour and the extent of the loosening in
one particular set of circumstances.
The fore-and-aft fit of tenons in the underwater part of a hull would loosen
after any prolonged period drying out in shipsheds, as would have been the
case with warships. The only way to minimise that loosening would have
been by choosing for tenons timbers with the smallest expansion and
contraction with moisture content. Fortunately the hardest timbers (needed
in any case for tenons so that the cross-grain crush strength of the tenon
timber should as far as possible match the end-grain crush strength of the
plank timber), contract and expand with moisture less than lighter timbers.
Even so, contractions of up to about 2% could occur radially in such timbers
as Turkey Oak (9.cerris) and Olive (olea hochstetter), which seem to have
been commonly used for tenons. As in dowels, tenons would best have
been cut so that the radial direction of their grain was fore and aft. The
greatest recurring loosening would occur in tenons below the waterline;
those above it would remain drier and therefore permanently tighter, unless
overloaded.
2% contraction in a tenon, say, 45mm wide would cause a gap of nearly
Imm to develop, enough to allow the same amount of slide in the seam. If
the planks were 150mm wide, the planking could shear by an angle of about
1mm/l50mm of a radian = 0.4O, which would give a shear deflection of 10cm
in a length of 15m of ship, enough to interfere with the proper functioning of
the oar system of a trieres. Fortunately, however, the part of the shell
providing the main resistance to shear forces acting on the hull of an oared
ship is largely above water. Underwater tenons would almost certainly not
have remained tight on account of effects of moisture cycling, though they
may have been in a tightly built new oared ship before its first drying out. The
general result would appear to have been that inter-plank fastenings
underwater, while performing the necessary job of holding the seams
together, would not have contributed much to the prevention of shear
deflections of the hull. That would have depended more upon the fastenings
between the planks above water which would remain drier in service.
In the process of "taking up" for the second and subsequent

occasions, underwater seams would certainly be prone to working if the ship
and its joints were subjected to reversals of loading and stress. Working
would soon induce leakage, so whether or not stress reversals occurred is a
matter which would have greatly affected oared ships in particular. These
effects must have been experienced by all those serving in ancient
Mediterranean navies, been well known to them and were probably matters
of everyday interest and concern. However, being everyday and familiar
matters, they would not have been worth recording by historians who were
not, as a rule, interested in technical details.
Aristophanes, however, in describing the preparations for launching a

fleet of triremes for operations1° does mention that "down in the dockyard the
air would be full of the planing of oars, the hammering of dowel-pins ..." One
can certainly understand that many of the pegs locking tenons would be
loose after a ship had been on a slip and under cover for some time, but was
there more to it than that? If the pegs had been draw-bored, there certainly
would have been. Planks when underwater would swell across their breadth,
closing seams more tightly, moving up upon the frames within the hull, and
putting the tenons in tension between the locking pegs and thus loading
those pegs in shear. The pegs would surely deform (and if draw-bored,
deform further) under the heavy force generated by the swelling plank.
After drying out, the seams would open and the pegs would be pulled
the other way by the now shrunk plank. There was therefore every reason,
in commissioning a warship which had been stored in a shipshed, to drive
pegs, generally tapered with the big end inboard, from inboard to fit more
tightly and to pull the seam together a little. If the pegs had been draw-bored,
driving them further would go far towards restoring the tension in tenons
pulling seams together. If pegs were tightened after stopping had been
worked into the seams, and if tightened from the keel upwards, strake by
strake, planking would have been progressively pulled down a little, sliding
back down the frames nearer to their positions as originally built (the
dowelled and clenched spike fastenings of frames would have allowed such
movement fairly easily) and the seams would have been closed on to the
fresh stopping. Protruding ends of pegs would then have been cut off flush
with the planking, the stopping squeezed out of the seams scraped off, and
the bottom recoated to protect it from the ship worm and to reduce frictional
resistance to the ship's motion in the water. Could that have been a more
detailed description of the process behind Aristophanes's remark?
Reduction of stress reversals in service

The working of plank seams would become harder to prevent, or at least
to limit, as ships aged. Well built hulls would have become leaky more slowly
than those built with less closely fitting tenons (particularly those above
water), but another technique could also have been employed with great
effect to reduce the onset of movement in seams and hence leakage. In
building, many shell planks, in being bent upwards at bow and stern, exert
a downward force on the ends of the hull, so there is an in-built hogging
tendency. This can be manipulated to some extent and the hull itself built
with some sagging deformation (that is, with the ends of the hull raised
above their intended height) so that, after the hull has settled, it is about the
intended shape (or a little sagged), but still in a state of self-strain in hog.
Most if not all stress reversals in service could be avoided by the

imposition of hogging shear stresses within the structure greater than, or of
about the same magnitude as, the largest sagging shear stresses likely to be
experienced in service. The Cheops ship4 is one example, in which the
joggled seams would, after settling, carry the hogging generated by the long
above water overhangs of the ends of this ship. The c. 1500 BC 17-metre
North Ferriby boat1'in Britain is another, in which an in-built state of hogging
self-strainwas in all probability generated by forcibly bending the flat bottom
(as could also have been done in the Cheops ship) before building the side
shell and fitting frames. Unfortunately, however, such a state of self-strain,
owing to the tendency of timber to creep as already mentioned, would relax
in time and, with it, its beneficial effects in reducing stress reversals. That
could be another reason why wooden ships often leak more as they age.
Inter-plank fastenings, for so long as they were pre-loaded in this way,

would not have worked and would therefore have remained tighter, unless
and until exposed to opposite and greater loads. It is fortunate that bending
moments and shear forces in hulls in sagging (i.e. in the trough of a wave)
are almost invariably much less than in hogging, so the pre-loading in
hogging needed to prevent stress reversal in service is relatively moderate.
Most wooden ships and boats experience significant hogging when

floating in still water. For example, a trieres in still water carries a hogging
bending moment when fully manned of about 30 tonne-metres, one third of
the likely maximum hogging bending moment, experienced on the crest of
a wave as long as the ship and 1 metre high from trough to crest. The
sagging bending moment in the trough of the same wave is only 15 tonne-
metres, so much stress reversal in a trieres while afloat could have been very
largely avoided if the hull had an in-built hogging moment of that amount.
Stress reversal when launching and slipping

In long and shallow ships, appreciable bending in sagging is
unavoidable in launching and slipping, whether at ship sheds or at beaches.
Stress reversal when a trieres was slipped or launched must have occurred,
because on a slip with a 10' slope the maximum bending moment
experienced, when the keel was just not in contact with the slipway along its
length, as either the bow lifted on launching or it was about to settle on the
slip on being hauled out of the water, must have been about 50 tonne-metres
in sag (assuming the weight to be 25 tonnes). The vertical shear force on the
hull was high at that stage of launch or slipping, because about half the
weight of the hull, say 12 tonnes, was being supported in a concentrated
manner at the after cut-up of the keel. Thus the net maximum shear force on
the hull just forward of the cut-up would have been about 10 tonnes, about
the same as the maximum shear force afloat when on the crest of a wave,
but in the opposite direction. The need for the tenons above water in that
part of the hull, let alone the rest of it, to have a simultaneous tight fit in the
fore and aft direction on both sides of the tenons is made clear by these
considerations. If Kolbe12was right in estimating the average life of triereis to
have been about 20 years, then it strongly suggests that that fit was
generally achieved.
The effect of hypozomata

If stretched horizontally from one end of a hull to the other, hypozomata
would have relieved tensile hogging bending stresses in the upper hull by
imposing a compressive load above the neutral axis of the hull cross-
section, but their effect on shear forces in the hull would have been confined
to that generated by any rise in the neutral axis towards the ends of the hull.
Hypozomata can impose no large sagging shear forces on hulls unless they
are arched over their length, like true hogging trusses.
Adequacy of tenons
Could the necessary shear strength, or more particularly stiffness and
lack of permanent crush, have been obtained in tenons above water in
triereis? As has already been stated, a shear force of 10 tonnes on the hull
would have generated a maximum shear stress in side planking of about
0.048 kg.f/mm2, neglecting, it has to be admitted, the effects of the large
oarports. Referring to Annex A and entering the dimensions and spacing of

the tenons in the trieres reconstruction, as an example, in Equation 6, we
obtain a figure for the maximum pressure on the tenons of 10 N/mm2,about
half the proportional limit in compression across the grain of American Live
Oak (q. virginiana), the material chosen for tenons in the trieres
reconstruction for the sake of longevity. 10 N/mm2is also about 20% of the
maximum compressive stress along the grain of Oregon Pine (pseudotsuga
menziesi~],the plank material, and about the proportional limit for that timber.
The tenon joints in that ship seem therefore to be loaded within the
proportional limit were the sideplanking uninterrupted, but the presence of
oarports, removing about 30% of the side along a line through their centres,
makes their stiff and linear behaviour under maximum hull shear forces
suspect. This is but one demonstration that the trieres was a type of ship
developed to the practical limit of the materials and techniques of the time,
and is consistent with ancient references to triereis becoming on occasions
"heavy in the water", implying that leakage in those ships had become too
great to keep acceptably under control.
This calculation and assessment, it should be emphasised, refers only to

the initial, proportional, behaviour of tenons: it would take loads 5 or 10 times
greater to cause such joints to start to break, so there is no implied threat to
the integrity of the hull. What is implied is that the breadth of tenons, and
probably of the planking too in way of oarports, in triereis may have been
thicker. Tenons in the sides could sensibly have been 20mm thick, not 12,
and side planking 60 mm, or 3 "fingers" instead of 2. Elsewhere in the hull,
planks and tenons could have been satisfactory if of the dimensions found
in the Marsala ship, that sole long oared ship of which informative remains
have so far been found on the seabed. That ship did not have oarports and
was unlikely to have been a warship of the line.
These indications make it probable that the sides of the heavier types of
warships which followed the trieres were quite thick to carry the hull shear
forces. A penteres probably displaced about 100 tonnes or more, while the
bigger polyreis would have grown to about 200 tonnes13.The sides of these
heavier ships could have been 100mm or so thick with tenons of thicknesses
to match. Therein could lie one reason for the larger polyreis to have had two
levels of multi-manned oars, not three, namely to avoid weakening the sides
of their hulls by piercing them for the lower level of oarports near the flexural
neutral axis where shear stresses are maxima. It could also have been a
reason for the suspected development of a Punic type of penteres with the
oarsystem wholly above the hull, probably deckedI3 This problem of shear
stiffness was to return with a vengeance in the large sailing warships of the
early 19* century, each of whose sides were pierced by 60 or so gunports.
CONCLUSION
This paper will have, it is hoped, made it clear that there is yet much to
be found out about the techniques and physical properties of stitching,
dowels and tenons as applied to ancient boat and ship construction. A
programme of experiments is necessary. Until by that means we know more
about the techniques which lay at the core of ancient shipbuilding practices,
we cannot be very sure about our practical understanding of the subject,
either in the light of modern knowledge or as perceived by ancient
shipbuilders, their clients and seamen. If this paper has indicated the need,
to understand the techniques of shell-building, for further and more thorough
studies, including experiments, into the material properties of the inter-plank
fastenings useQ it will have served its purpose.
John Coates
Sabinal
Lucklands Road
Bath, Avon
BA1 4AU, England
ANNEX A
Loads on Shell Plank Tenons and their Pegs in resisting

Shear Stresses in Planking
To transmit shear stress in one plank to the next one, the joining tenons are subjected to
shear forces. These forces generate others necessary to maintain the equilibrium of each tenon.
They are shown in a simplified, but probably not grossly incorrect, form in Figure xx,where P is
the bearing force between a tenon and its mortice, and H and V the components of the bearing
force between the tenon and its peg parallel and perpendicular to the seam respectively. Let S
denote the inter-plank shear force borne by each tenon, here assumed constant in the vicinity
considered. In Figure xx, the edges of tenons and mortices have been drawn straight and parallel
to each other.
The pressure between mortice and edge of tenon, accounting for the force P, is assumed to
extend from the plank edge to the level of the peg, as shown. I t is also assumed that the pressure
rises linearly from zero at peg level towards the plank edge. Let the coefficient of friction between
tenon and mortice be p.
Equilibrium of the plank requires that

V = p P and S = P - H , . . . . . . . . . . . . . . . . . . . .(1)
neglecting forces required for the rotational equilibrium of the plank.
Rotational equilibrium of a tenon, however, requires that
PdB = Hd + pPw, . . . . . . . . . . . . . . . . . . . . (2)
if each tenon is w wide and its pegs are d apart. From (1) and (2),
H=P(lB-pwld) . . . . . . . . . . . . . . . . . . . . (3)
and S = P (W3 + pwld ) . . . . . . . . . . . . . . . . . . . .(4)
If the maximum pressure at plank edge is p-, and the tenon has thickness f ,
than P = p,m dV4,
so that S = p,, df ( ll6 + pwI4d ). . . . . . . . . . . . . . . . . . . . . (5)
S is then the shear force which can be borne by one tenon for a given value of p, which cannot
exceed the cross-grain crushing stress of the timber of the tenon. If the shear deflection, or slide
across the seam is to be restricted to that occurring while p, is less than or equal to the much
smaller proportional limit in compressive stress across the grain of the tenon, p,,,, S will be
reduced accordingly.
The shear stress, q, , in the side shell of the hull, t, thick, may be estimated approximately
by the usual methods of beam theory. The shear force on each tenon, if equally loaded, and they
are r apart centre to centre, is
S=q,t.r,
so the shell shearing stress, q, , can be related to the shell plank and tenon dimensions, tenon
spacing and maximum permitted cross-grain pressure on the tenon, p, (or p,,d
by the equation
q, = p,. df V6 + uw14d 1 . . . . . . . . . . . . . . .(6)
t, r
Shear stress in tenons in the plane of the plank seam will not be uniform, but its mean value will
be q, = st (4 w ) . . . . . . . . . . . . . . . . . (7)
It will be evident, if the maximum shear stress that can be borne by the tenon timber is
a,, , that in the limit, ,
S a,, (f W) , i.e., q, q,, , (4 w)/(t, r) . . . . . . . . . . . (8)
The necessary diameter of pegs may be found by solving for V and H in terms of S or q,, given a
value for the shear load to be camed by the peg timber in double shear without unacceptable
deformation.
ANNEX B
An Exploratory Experiment on the Moisture Cycling

of Timber with and without Restraint across the Grain
The Specimens were of Live American Oak (q.virginiana), 34.0 mm wide and 11.0
mm thick across the grain. They had been kept for ten years in living accommodation
indoors in Britain. The annular rings were a t about 45O to breadth.
The experiment consisted of taking three similar specimens and treating them
respectively thus, measuring their breadth at various times:
No. 1 - Crush in a large steel vice by 0.60 mm (1.8%)and keep it in the vice a t the
same position of the handle bar of the vice.
No. 2 - Crush as No. 1for 27 days. Then immerse in water without restraint for 7
days, followed by drying in air for 26 days.
No. 3 - Crush as No. 1and at the same time immerse in water while restrained as
No. 1 for 7 days. Then remove water and leave the specimen in air but
under the same restraint for 26 days.
Measured breadths of specimens, mm:

Specimen No. 1
Date
27/5/96 Before crushing 34.0 (dry) 33.9 (dry) 34.0 (dry)
27/5/96 After crushing 33.4 (dry) 33.4 (dry) 33.3 (wetbd)
22/6/!X + 27 days 33.3 (dry) 33.3 (wetted 33.3(wet)

and released)
29/6/96 + 7 days 33.2 (dry) 36.6 (taken out 33.5(taken
of water) outdwab-,
released
- m y
to make
measurement)
35.5 (drying) 33.l(drying,
just loose
in vice)
25/7/96 +26 days 33.2 (dry) 35.0 (drying) 33.0(100se
in vice)
25/7/96 33.4 (when released)
Net contraction
after crushing, mm - 1.7 (i.e. expansion) 0.3
Ditto ,% -5.0 0.9
REFERENCES
1. Coates, J.F., (1985) "Some Structural Models for Sewn Boats" in Sewn Plank Boats, Eds. Mc
Grail and Kentley, BAR 276: 9-12. Oxford.
2. Coates, J.F., (1985) "Hogging or Breaking of Frame-built Wooden Ships" in M.M. 71.4:
437-442.
3. Coates, J.F., (1987) "Pentekontors and Triereis compared" in Tropis 11: 115.
4. Lipke, P., (1985) "Retrospective on the Royal Ship of Cheops" in Sewn Plank Boats, Eds.
McGrail and Tentley, BAR 276: 19-34. Oxford.
5. Brungraber, R.L., (1985) Traditional Timber Joinery: A modern Analysis, Stanford University
Ph.D. Thesis.
6. Pomey, Patrice, (1981) "L'Epave de Bon-Porte et les Bateaux cousus de Mtditerranee" in
M.M. 67.3: 225-243.
7. Steify, J.R., (1985) "The Kyrenia Ship: An interim Report on its Hull Construction" in A.J.A.
89: 95.
8. Curtis, W.H., (1919) The Elements of Wood Ship Construction: 178. McGraw Hill, New York.
9. Mark, S.E., (1991) "Odyssey 5. 234-53 and Homeric Ship Construction: A Reappraisal" in
A.J.A. 95: 441-5.
10. Aristophanes, Acharnians: 544-54.
11. Wright, E.V. (1990) The Ferriby Boats. Routledge, London.
12. Kolbe, W., (1901) "Zur athenischen Marineverwaltung", in Mitteilung des deutschen
arch6eologischen Instituts, Athenische Abteilung.
13. Morrison, J.S., (Forthcoming) Greek and Roman Oared Warships, Chapter 7. Oxford
CAPTIONS OF FIGURES
Figure 1. Leaf spring and boat side - sliding movement due to hogging shear forces and
bending
Figure 2. Simple stitching to resist plank sliding due to hogging shear forces
Figure 3. Tapered plank tenons and pegs
Figure 4. A simplified model of forces acting on plank tenons, mortices and pegs in resisting
plank sliding
Leaves of s p r ~ n gs l i c l e on each o t h e r
- - -
___t
---
__t - Fig. 2
-
ABSTRACT IlEPIAHWH
NEA EYPHMATA A n 0 TON KYNO
H a v a ~ o i v w qacpopa ue vCa euprjpcrra an6 TOV K~jvoAipava~hv,o

onoioq TE~VELva xapan-qp~meiwq ~cpa~vopcvo>> apxa~oAoyi~rjqBCqq KaL
avao~acprjqecp' ouov an6 EKE^ EXELnpoCAeei nhqehpa 6elyphwv nhoiwv
Tllq U O T E ~ O E 111 rM ~~LK~~~
nepi660u.
f l p o ~ ~n6Ai
~ ~ yla
a iompa~a pe ana~oviueiqnhoiwv Kai &vaet6hA~0,Ta
onoia av Kai uh<ovrai anoarraupa~i~a paq nap&xouvnohhipeq nhqpocpo-
pieq yLa TO nhoio q q enoxvq ~ o u aMa
q ouy~povwqeyeipouv Kai vCa epw-
TTlpaTiKa.
Ap. Oavoupia A ~ K o ~ ~ v L ~
lA' Ecpopeia np0bT0piKh~KaL K ~ ~ U L
AK~ X~~VL O T ~ ~ T W V
Aapia
FURTHER FINDS FROM KYNOS
The paper concerns new finds from Kynos, Livanates, that tends to be
characterized as a <<phenomenal>> archaeological site and excavation, as it
has produced a multitude of different examples of ships belonging to the
Late Helladic 111 C period.
It concerns once more sherds with representations of ships and a

figurine, which although found in a fragmentary state, provide valuable
information for the ships of that period, but raise also new questions.
Dr. Fanouria Dakoronia

14'hEphorate of Prehistoric and Classical Antiquities
Lamia
RESUME
INFORMATIONS SUR LA CONSTRUCTION NAVALE DANS DES

DOCUMENTS DE LA PREMIERE PERIODE BYZANTINE
Ayant comme point de départ le texte grec du «Martyre de Saint Aréthas

et de ses compagnons. [6" siècle après J.-C.] cette communication porte
essentiellement sur l'enregistrement des éléments de la marine, tels qu'ils
sont répandus dans des textes hagiographiques et historiques de la
première période byzantine. Outre les termes techniques de construction
des navires qui figurent dans les textes (.imoi>>, «~&pa-rapia>>, <<app&va,>,
ÉppoAoi>>, <<auxÉv&p, etc.), les indications textuelles sur le poids des navires
et leur déplacement, ainsi que les renseignements sur le temps nécessaire
pour la construction d'une marine de guerre, présentent un intérêt
particulier. Nous examinons aussi la distinction entre les termes nAoio et
Kapapi telle qu'elle est établie déjà dans les textes. En dernier lieu, les
références précises aux divers types des constructions nous aideront dà
mettre au point l'état de la question [particulièrement à propos du type
«indien>>].Nous devrions préciser que cet essai repose sur les données
fournies par les sources littéraires et leur étude philologique.
Marina Detoraki
Geronimaki 68
71 307 Iraklion, Crete
SHIP ICONOGRAPHY ON "BLACK-AND-WHITE MOSAICS
OF THE 1" - TdCENTURIES AD
Abstract
This paper introduces three sites from the Mediterranean, with ship
iconography depicted on mosaic floors produced in the black-and-white
technique. These mosaics cover the period from the 1" to the 3rdcenturies
AD. The mosaic floors are related to different architectonic structures, such
as private villas, public or private baths, and maritime traders' and shippers'
offices in the Mediterranean. The sites are Migdal, in Israel, Althiburus, in
Tunisia, and Ostia, in Italy (fig.1). The ships depicted on these mosaics will
be described below and also related to the specific names depicted on the
Althiburus floor, as well as mentioned in historical references.
Migdal, on the Kinneret
Migdal is situated on the western coast of the Sea of Galilee (Kinneret),

about 5km north of Tiberias. It is also known as Magdala, the home of Mary-
Magdalena'. Mainly, Migdal is noted as one of the places where many Jews
from Tiberias sought refuge at the outbreak of the Great Revolt in 67 AD. In
The War of the Jews, Josephus wrote about the Revolt which took place on
the shores of the Kinneret, and which was one of the most bloody sea battles
between the Jews and the Romans2. Josephus himself was personally
involved in this war, being the Governor of Galilee during that period. The
sea battle was carried out in fishing boats of different sizes, which were taken
from the local fishermen.
In Strabo's Geography Migdal was known as Taricheae, where the lake

supplied excellent fish for pickling3. The Aramaic name of Taricheae is
Migdal Nunia, i.e. "Tower of FishH4.The name "Taricheae" is indicative of the
source of the local population's livelihood.
Archaeological Remains
The various names attributed to Migdal attest to its maritime character.

Remains of a harbor were discovered near the courtyard belonging to the
Franciscan Monastery. Dr. V. Corbo, from the Franciscan Institute in
Jerusalem, conducted the archaeological excavations carried out between
1971-77. During the excavations were revealed the town-square, some
streets, various buildings and a water system (fig. 2). Also an urban villa was
revealed (Area C) to the northeast of the Water Tower (fig. 2). Most of its
floors were paved with mosaics. The mosaic floor of Room C6 is the subject
of this paper. The room is located in the most central part of the villa (fig. 2).
It forms a square of almost 2.80 x 2.50m. Two layers of mosaic pavements
were found in the room. For an accurate dating of the villa it was necessary
to remove the upper mosaic floor which was of a later period. The lower floor
revealed the mosaic produced in the black-and-white technique. The style of
this mosaic, and the ceramic remains found beneath this floor, established
the dating of the villa as 1" century AD. The mosaic floor consists of two
parts:
1. A short Greek inscription " ~ a06"
i (to you or you too), made of black
tesserae on a white background. The inscription was found on the
southern doorstep5.
2. Two black frames enclose the area of the mosaic with a white strip
dividing them. The entire rectangle measures 1.10 x 1.12m. The
objects depicted within the area are formed with black and a very few
brownish-red tesserae on a white background. They are depicted in
the same orientation. The patterns represent a kantharos, on a half
ring are attached two strigilis and an aryballos. There are some other
objects that have not been identified properly. A unique depiction is a
ship, which is situated close to the lower left corner of the frame (fig.
3). Beneath the stern and slightly to the right is the partially preserved
fish's head. Although the objects are depicted schematically they still
provide us with a general outline and a possible identification.
The Migdal Ship
The ship is depicted from the port side, with the bow pointed and almost
touching the left side of the black frame. The hull, mast, and ropes are
depicted with black tesserae, while the oars and the sail are depicted with
brownish-redstones (fig. 4). The hull has an elongated shape, with a pointed
stem and a rounded stern ended with an inner-turned volute above the aft-
deck (fig. 4). Beneath the bow is the pointed cutwater, which resembles the
shape of the fish head beneath the stern (fig. 3). The stempost extends
almost horizontally above the cutwater, though it is a continuation of the
gunwale. This extension may indicate the bowsprit. Along the same line as
the pointed cutwater is a single brownish-red tessera, probably indicating
the oculus6.A horizontal wide white strip that extends about 213 of the hull's
length may represent the lateral wale or rubbing-strake that reinforces the
outer hull and also supported the oars (fig. 4).
SHIP ICONOGRAPHY ON "BLACK-AND-WHITE"MOSAICS
Above the port gunwale there are four protrusions, probably indicating
the heads of the crew. They appear to be seated behind a screen or fence
attached above the gunwale, or on lower thwarts, facing the stern with their
back turned to the bow. The hull does not show any line of flotation or draft,
nor is a water line indicated.
The ship is rigged with one mast and a yard with the adjacent cordage,
and three oars placed obliquely to the port hull with the looms pointing
towards the stem (fig. 4). The mast is a vertical spar set on the forward third
of the vessel, between the second and third rower. It is depicted with a single
row of black tesserae. The length of the mast (from its tip to the presumed
bilge, ca.1cm above the bottom) is about 213 of the hull's length. The yard
represented by a horizontal spar makes a 90' angle with the masthead and
is parallel to the deck. It is slightly offset, towards the stern. The length of the
yard is equal to the length of the mast. Attached beneath the yard is a line of
brownish-red tesserae. This line is a bit shorter that the yard and represents
the furled sail by means of brails (fig. 4).
Only two distinctive lines can be associated with the standing and
running rig. An angular strip of black tesserae stretches from the starboard
side of the masthead (behind the furled sail) to the tip of the bowsprit. This
line may represent the forestay. From the starboard edge of the right
yardarm, behind the sail, an almost vertical line is hanging down towards the
port gunwale. This line indicates the right brace or sheet (fig. 4).
There are two kinds of oars depicted on the port hull. Only their shafts,
without blades, represent two of them, the left and the middle oars. The third
oar (right-hand) is depicted with a shaft and at its lower end is a blade with
round shoulders and a straight cut end (fig. 4). All the oars are made of
brownish-red tessera. The blade constitutes about 113 of the shaft's length.
The Migdal ship is depicted schematically with no indication of specific

type, other than a merchant vessel or a passenger transporter. It appears to
be shown in 314 view (a side perspective). The highly raised stern, with an
attempt shortening the voluted sternpost, emphasizes it. This vessel may
also portray a fishing boat, which could be rigged with sailing gear for when
the wind conditions were favorable. Inner-volutedstemposts have appeared
in ship iconography since the Th century BCE7. This feature became a
decoration on merchant vessels from the 1"-Pd centuries AD8. It also
appeared in some wargalleysQ.The shape of a bow with the projecting
cutwater was a characteristic feature of merchant vessels since the middle of
the la century AD1'. The closest parallel of the shape of the bow of the Migdal
ship appears in the 1" century BCE graffito of a ship from Delosl1(fig. 5). In
the graffito the bowsprit is supported by an arched stempost. One may
assume that the Migdal ship had a similar device (fig. 4). The forestay of the
ship in the Delos graffito is attached to the aft tip of the bowsprit, while on
the Migdal ship it is attached to the fore tip of the bowsprit.
The depiction of different oars in the Migdal ship was probably meant
either to distinguish the row-oars from the steering-oars, or that the blades
of the left and middle oars are submerged in the water and the right-hand
oar is in the process of maneuvering. Most probably the same number of
oars were set on the starboard hull, though the picture shows the port hull
only. The use of brownish-red tesserae for the oars was probably for the
purpose of distinguishing them from the hull. The heads of the crew above
the gunwale or the fencing, and the angled oars, may indicate that the ship
was rowed in a two-oarslsitlpull technique12,which would indicate the left-
hand sailing of the ship. The oar with the blade may indicate the steering-oar
set on the port quarter and worked by a helmsman seated beneath the
voluted sternpost
The Migdal ship may represent small merchantmen (naves oneraria) with
a crew of five or six men; four rowers, a helmsman and the captain
(kybernetes). To deduce the load capacity of this ship we may rely on two
sources:
1. The results of anthropological studies on skeletons dated to the period
of Josephus 13. These studies showed that the average weight of a
man was 62-67kg. Thus, a vessel with a crew of six men, their own
gear, anchors, and the rigging would indicate a capacity of between
800 to 1.5 ton.
2. The discovery of "Jesus Boat", on the northwest coast of the Sea of
Galilee, at Ginnosar, and excavated in 198614. It was dated to the
period between the end of the Pd century BCE and the first part of the
l Hcentury AD. The Kinneret boat is a fishing craft and the Migdal ship
may represent a small merchant craft, or a passenger transporter that
also could be used for fishing. The Kinneret boat was built in the
traditional Mediterranean fashion, of shell-first with mortise-and-tenon
joints. Hypothetically, if we take the length of the ship in the mosaic
and the Kinneret boat, one can see that the model indicates a reduced
scale of approximately 1:2515.
SHIP ICONOGRAPHY ON "BLACK-AND-WHITE" MOSAICS
Althiburus, in Tunisia
Althiburus (modern Medina) is situated on the central plateau of Tunisia,

about 200km off the northeast coast. The ancient city was built before the
Roman conquest. The location of the city at a strategic point on the trade
route between the Sahara Desert and the Mediterranean shore placed
Althiburus as an important storage center of the products that were shipped
to Rome from the port of Carthage. The environs of Althiburus consisted of
fertile agricultural plains and it has rich deposits of phosphates. During the
reigns of Antoninus Pius (138-161 AD) and Septimius Severus (192-211 AD),
Althiburus was very prosperous, and became an important grain trading
center. The city was abandoned in the Arab conquest, during the 7m century
AD.
The first archaeological remains of Althiburus were revealed in 1895. The

famous mosaic floor with the "Catalogue of Ships" (fig. 7a) was discovered
during this time, in a very large and elaborate villa, the Maison des Muses.
The floors of the entire villa were paved with beautiful mosaics, which
depicted rich and varied maritime scenes. The room where the catalogue
was found has a cross shape (fig. 6). According to the top plan of this room
it was deduced that it formed the frigidarium of a large bath. In 1961, the
entire villa was excavated, and during this period the mosaic with the
Catalogue of Ships was removed and put on display in the Bardo Museum,
Tunist6.The mosaic depicts about 25 different types of vessels, which were
spread over the entire floor without any specific orientation. The ships are
pictured on an olive-green background''. Only 22 vessels are identified by
Greek and Latin inscriptions (fig. 7b). Some of the ships are depicted with
putii, their rigging, and some are laden with cargo (jars, horses, fishing nets).
The putii are not in proportion to the vessels, but appear to be much larger.
The style of the inscriptions and the mosaic work indicates that the villa and
its floors were made during the second half of the 3'dcentury AD.
The largest number of vessels is depicted on the central part of the

mosaic (fig. 7a) which measures 8m in length. The width of the mosaic with
the right arm of the cross measures 5m; the left arm was destroyed. At each
end of the cross are depicted water gods; at the top is Okeanus, at the
bottom a river god and at the right corner the goddess Venus (fig.7a). The
inscriptions associated with the vessels do not necessarily indicate a specific
ship.
Types of Vessels
The vessels depicted in the catalogue can be classified according to the

shape of the hull, the stem-and-stern posts and the rigging.
A. The Hull
1. Very rounded and spoon-shapedwith the stem-and-sternraised almost
vertically (figs. 7bl1, 2, 7).
2. Long and slim hulls with an almost vertical stempost with the forward
projecting pointed cutwater. The rounded stern is ended either with an
inner-turned volute or it is slightly higher than the deck and has a V-
shape (figs. 7bl10, 12, 13, 19, 21, 22, 25).
3. Long hulls with raised stempost curving above the stem and ended
with an inner-turned volute. The stern is rounded and either with
rounded tip or a V-shape (figs. 7bl3, 15, 16, 17).
B. Rigging
The rigging of the vessels comprised of one or two masts and sails with the
adjacent cordage, a pair of steering oars; there is one vessel depicted with
three oars:
1. One steering oar on the quarter, one mast and no sail (fig. 7bll).
Attached to the tip of the masthead is a small flagla.Although only one
steering oar is seen on the starboard quarter, the artist probably
intended to depict two oars on either quarter.
2. One mast and a sail, a pair of oars (figs. 7b15, 11, 12, 13). The vessels
in figs. 7bl5, 12 are depicted with a mast set fore amidships and a fully
open sail. The yard is secured to the mast by lifts. On the fore side of
the sail a checkerboard pattern is visible, representing the brails used
to work the sail. These vessels are also maneuvered by a rower who
is working a pair of oars set on either side of the hull amidships.
The vessels depicted in figs. 7bl11, 13 are rigged with one mast and a
furled sail beneath the yard. The yard is secured to the masthead with
lifts. In fig. 7bll Ione puti works the halyard. The angled left-hand line
that stretches from beneath the sail to the port gunwale may indicate
the left shroud. There is a single oar lain on the starboard gunwale
with its loom behind the putii. The vessel in fig. 7b/13 is occupied by
three putii. One puti is rowing a pair of oars placed on either side
amidships, another one is climbing the ladder on the mast, and the
third figure is holding a hammer in his right hand. The mast is set fore
amidships. The yard is attached to the masthead by lifts. On both
sides of the lower corners of the furled sail a free hanging rope is seen.
This line may indicate the right and left sheets.
3. Two masts and sails (figs. 7b12, 3, 4). These vessels are rigged with a
large main mast and sail, and the second mast is the artemon
(foresail). In figs. 7bl2, 3 both sails are furled. The yard is secured to
the mast by the lifts. At the tip of both left-yardarms is seen the brail or
the sheet. The masts are secured with a series of ropes that represent
the fore-and-back stays. Both vessels are rigged with a pair of oars on
either quarter. The oars have long and wide blades. Most probably
they indicate the rudders.
The vessel in fig. 7bl4 is rigged with two masts and fully open sails.
The main mast is set amidships and a bit higher than the artemon (set
at a slight angle from the main mast). Both yards are secured to the
masthead by a series of lifts. On the fore face of the sails is a
checkerboard pattern, indicating the brails. One puti seems to work
the halyard of the artemon and the loom of the starboard steering oar.
4. There are several vessels rigged only with a pair of oars set on either
side (figs. 7bl7, 8, 9, 10, 14, 15, 16, 17, 20, 21, 22, 23). Some of them
are worked by one puti (figs. 7bl7, 8, 10, 15, 16, 21, 22, 23).
The Hippago ship (fig. 7b/6) is rigged with three oars lain at an angle
on the starboard hull. The looms are pointed to the stem. It is probable
that the vessels had the same number of oars on the port hull as well.
The oars are not worked by anybody. The fishing boat in fig. 7bl19 is
rigged with two oars laid on the starboard rubbing-wale. The looms
are pointed towards the putii who are lifting a fishing net full of fish.
Almost all the vessels are depicted with a long lateral strake beneath the
gunwale (figs. 7bl3, 4, 5, 7, 9, 12, 16, 17, 18), slightly above it, or at the same
level (figs. 7b12, 6, 8, 10, 11, 13, 15, 19, 20, 21, 22, 23, 24). Both ends of
these strakes are protruding from the stem and aft the stern. This element
indicates that the upper strake or the rubbing-strake that was used to
reinforce the outer lateral hull, and protect it when the vessel was anchored
at the quay. This strake also supported the shafts of the oar. There are some
vessels depicted with a small protrusion above the gunwale. This element
may signify the thole-pins used to secure the looms of the oar (figs. 7bl7, 9,
10, 13, 14) or the bitts used for the running and standing rig (figs. 7bl13).
ZARAZA FRIEDMAN TROPlS VI
Ostia, in Italy
Ostia was founded (349 BCE) as a small fort at the mouth of the Tiber
River. It enabled control of the river as well as protection for Rome. Claudius,
who started the work in AD 42, built the first commercial harbor of Ostia. The
building of the harbor was accompanied by the centralization of the corn
administration under imperial controltg.It was completed at about AD 46, as
attested by an inscription that records Claudius' construction of a canal from
the Tiber to the sea, to connect the new harbor with Rome. Another purpose
of the canal was to serve as a second outlet of the river and save Rome from
flooding2'. By AD 62, the Claudian harbor was fully functional, as attested to
by Tacitus2'.The development of the harbor is also associated with services
related to the maritime trade. The storage capacity of corn at Ostia increased
during the 2ndcentury AD in spite of the addition of the new horrea around
Trajan's harbor". The business life of Ostia concentrated around a large
square or pianale behind the theatre, north of the decumanus, and about
half way between the Forum and the ,East Gaten. Offices of various
commercial corporations were set up around the open square surrounded
by colonnades2"known as Pianale delle Corporazioni. A marble pediment
found on the east side of the colonnade bears the inscription "naviculari
Africani". This inscription shows that overseas shippers were present at Ostia
before the Pdcentury AD." Sixty-one rooms open to the colonnade and were
paved with mosaics. A few of the surviving mosaics from these shops
illustrate the occupation of the owner, and subsequent inscriptions indicate
their place of origin. These mosaics are not set at the original level of the
shops. Beneath them were found earlier remains of mosaics which could
reveal the date of the colonnade and the theatre. It is assumed that traders
and shippers who were most important as suppliers of Rome were
concentrated here by imperial authority under Augustus. Since this period
(1" century BCE) and until the 3'dcentury AD, Ostia was one of the main
centers of official controlz6.Traders who were located at Ostia originated
from North Africa, Gaul, Sardinia and the Adriatic. Three inscriptions of the
shippers are accompanied by the traders' naviculatorii et negotiates.
Presumably the negotiates dealt with orders for goods that they would buy
in their home district, and henceforth ship to Ostia2'. Most of the offices in the
corporation square were for foreign and out of town shippers and traders.
There were some groups of workers who served these offices (restiones,
stupatores, codicarii, pelliones) and had their shops in the Corporation
Squarez8.
The Corporation Shops
Following, the group of shippers and traders whose evidence was found
to be in Ostia and on the Tiber will be listed. The presence of an office of the
Spanish and Gallic export tax "statio Antonin(iana) XXXX Galliarum et
Hispaniarum" at Ostia suggests that goods from Gaul and Spain were
coming to the mouth of the Tiber, and then brought to Ostiaa. The nacularii
lignarii indicate that they were the boatmen who transported timber to
Rome30.The list will refer to the merchant vessels that came from different
places in the Mediterranean:
- classis Alexandrinae (from Egypt or in trade with Egypt)
- navicularii maris Hadriatici (from the Adriatic)
- naveculariiNarbonenses (Narbonne in Gaul, France) [fig. 81
- navicul(arii) et negotiantes Karalitani (Cagliari, in Sardinia) [fig. 91
- navicularii Turitanni (Turritani or Turris Libisonis, in Sardinia)31
- navicul(arii) Karthag(iens0 Disuo (Carthage, in Tunisia) [fig. 101
- navicularii Misuaenses (Misya, in North Africa)
- Sabratensium (Sabratha, in North Africa)
- navicularii Syllecti(ni) (North Africa)
- nacularii Gummitani (Gummi, in North Africa)

- nacularii Curbitani (Curubis, in North Africa).
The Ships
The vessels depicted on the mosaic floors of the shops in the

Corporation Square are made with black and white tesserae on a white
background. All the vessels represent merchantmen of different sizes. Some
of them are depicted with a cargo of jars, and others show harbor activities,
such as unloading large seagoing ships (anchored in open sea at the mouth
of the Tiber) into smaller harbor vessels and then carried upriver to the stores
at Ostia. These vessels can be classified according to the shape of their hull
and the rigging:
A. Long hulls with a concave prow, ended with a pointed projecting
cutwater (figs. 8,9, 11). The stern is very round and the sternpost ends
with the figure of a goose-head behind fencing or a railing. The
stempost ends either with a short horizontal fore extension of the
gunwale (fig. 9), with a rounded shallow head (fig. 8), or with an outer-
turned volute (fig. 11). The lateral wales beneath the gunwale are
outlined with a strip of white tesserae. This element may also signify
the rubbing-wale. In fig. 11, a small white triangle is depicted, just
beneath the fore tip of the wale. This decoration indicates the oculus.
B. Rounded hull with two different stems; their sternposts are almost
vertically ended with a short vertical block-shape (figs. 10, 13, 14):
1. Rounded stem ended with a block-shapedstempost (figs. 12, 13).
In fig. 12, there is a trapezoidal frame, which is attached to the tip
of the stempost. It is outlined with a strip of black tesserae and has
an angled forward extension. This frame probably signifies some
kind of bowsprit. Above the port gunwale there are two short
vertical spars. They indicate the bitts used for the fore standing
rigging.
2. Roundedstem, slightly higher than the fore-deck and ending in a V-

shape (fig. 10)
3. Rounded hull with an angled stem, ending with a block-shaped

stempost (fig. 13). The vessel appears to have a flat bottom.
C. The rigging is comprised of one (fig. 10, 12, 14), two (figs. 8, 9) and
three masts and sails (fig. 1I ) , and a pair of steering oars placed on
either quarter. The sails of all the vessels in discussion are fully open.
The masts and sails are seen with the adjacent ropes which make the
standing and running rig. The checkerboard pattern on the fore face
of the sail appears in figs. 9, 11, 12, while in fig. 10 (on both vessels)
it appears on the lee face of the sail. This pattern represents the whole
system of brails used to work the sail. In fig. 8, the sail depicted from
its lee side is made of black tesserae. The brails are not shown. On the
fore face of the sail in fig. 13, there are several longitudinal short
arches depicted with a strip of white stones. They may indicate the
brails. Above the yard (fig. 8) a black triangle is depicted. This pattern
signifies the topsails used on large seagoing ships of the Roman
period.
1. The single mast and sail is set amidships (figs. 10a, b), fore
amidships (fig. 12) and fore close to the bow (fig. 13). The standing
rig is comprised of the forestay (figs. 10a, 13) and the backstay
(figs. 1Oa, 12). In fig. 12, the backstays appear behind the lee face
of the sail. The upper ends seem to be tied to the lifts and their
lower ends are seen beneath the lower edge of the sail, thus
secured to the top of starboard and port aft railing. The running rig
comprised of braces is shown on the fore face of the sail in fig. 13.
The sheets are seen clearly in fig. 12; the lower end of the right
sheet seems to be held by the helmsman set on the aft deck. The
lower end of the left sheet is hidden by the fore starboard corner of
the railing.
2. Vessels with two masts and sails: The main mast is high and
tapered and the large square sail is fully open (figs. 8, 9). In fig. 8
the main mast is seen entirely from the starboard side, while as
seen in fig. 9 it is hidden by the sail and only the lower part is visible
(between the lower edge of the sail and the gunwale). The second
mast and sail make the artemon (foresail). In fig. 8, the mast of the
artemon is stretching at an angle above the stem. It is depicted as
a tapered spar with black tesserae. The sail appears to be furled
beneath the yard and depicted as a wide black strip. In fig. 9, the
artemon mast appears as a very thin and angled spar, stretching
above the bow. The artemon sail is fully open, and seen from its
fore face. It appears as a small rectangle depicted with white
tesserae. On both artemons the running rig is seen as comprised
of sheets or braces (fig. 8), or only sheets (fig. 9).
The standing rigging comprised of fore-and-back stays is clearly

seen on the lee face of the sail in fig. 8. In fig. 9, only the lower ends
of the stays are visible from beneath the lower edge of the main
sail. The ends of these stays run to either sides of the gunwale. The
brails depicted as a checkerboard pattern are seen on the fore face
of the main sail (fig. 9). Above the yard in fig. 8 is depicted a black
triangle. Its base is the length of the yard and the point of the
triangle is attached to the port side of the masthead. This triangle
represents the topsail used on large Roman seagoing vessels3'. It
is probable that the vessel in fig. 9 is depicted with a topsail similar
to the one in the previous illustration, though the short lines seen
between the yard and the lifts may indicate the upper ends of the
brails.
3. Three-masted vessel: Iconographic representations provide us with

different types of ships and their rigging. The representation of
vessels with three masts and sails is very rare. The mosaic
depiction of such a vessel was found in the shop of the shippers
from Syllectum (North Africa) [fig. 1I]. The main mast is set slightly
fore amidships. It is much higher and tapered than the other masts.
The ship is shown from its starboard side and the sails are bunt on
that side, thus partly covering the masts. The artemon is angled
and stretches above the bow. It is slightly shorter than the main
mast. The third mast is the mizzen, set on the aft deck between the
main mast and the stern. It is much shorter than the artemon and
the main one. The yard of all three rigs is supported by the lifts. On
the fore face of the bunt is the checkerboard pattern of brails used
for working the sails. The mizzen sail is depicted with sheets, which
are probably worked by the helmsman set on the aft deck. On the
fore part of the port gunwale there are three short block-like
protrusions. They may represent the bitts used for the standing and
running rig.
The vessels described above are rigged with a pair of steering-oars set
on either quarter. Their looms appear to be supported by the backwards
extension of the wing-like ends of the bulwarks, that ran along the amidships
section outside the gunwale33(figs. 8, 9, 10, 11, 12). The oars are angled to
the quarter with the looms pointing to the stem. The blades of all the oars
appear to be about half the size of the shafts' length. The lower part of the
shafts cross their blades at their mid-point (figs. 8, 10, 11, 12, 13). This
representation indicates that a groove in reversed U or V-shape was cut into
the mid-width of the lower shaft and that the long and thick blade was
inserted perpendicularly into this groove. When this operation was finished,
the blade was secured to the shaft by tree-nails and probably reinforced with
bronze nails. We do not have definite proof of such a construction of large
steering oars. When a ship is wrecked, its rigging is what is first to be
destroyed. We may reconstruct such oars from different iconographic
representations, since no such artifacts have yet been found.
Discussion
The iconography of ships on mosaics pertains to two distinct parts of the

Mediterranean, the east and the west. Geographically, some of the mosaics
are not in close vicinity to the Mediterranean shores. Migdal (ca. 60km from
the sea), though considered as eastern Mediterranean, is located on the
northwestern shore of the Sea of Galilee. The ship depicted on the mosaic
probably represents a vessel of Mediterranean type and origin. One may
assume that the owner of the house was involved in maritime trade not
necessarily only on the Lake. He may have been familiar with Mediterranean
types of vessels. The depiction of the ship model on the mosaic floor
probably was used to demonstrate his knowledge or indicate his
occupation.
The Catalogue of Ships depicted on the mosaic floor of the frigidarium,

in the "Maison des Muses'', at Althiburus, is also located far from the
northern shore of North Africa. This site is considered western
Mediterranean. The depiction on the mosaic indicates that the owner(s) of
the house were familiar with classical literature and art. This is deduced from
the inscriptions and quotations of verses that are associated with the models
of the vessels. The variety of vessels illustrated on the mosaic floor may also
indicate that the owner(s) had a wide knowledge of the vessels sailing in the
western Mediterranean. It is probable that he (or they) had been involved in
maritime trade and owned several ships.
The vessels depicted on the mosaic floor of the shops in Pianale delle
Corporazioni at Ostia emphasize the importance of this harbor, not only as
the largest of the Roman Empire, but as an important link between the great
trade route from east to west. The various vessels illustrated on these
mosaics depict different types of merchantmen that sailed in the
Mediterranean connecting trade centers, mostly in the Pd and centuries
AD.
The vessels portrayed on the mosaics described above represent

merchantmen of different sizes. Sailing gear and steering oar or rudders
mainly operate them. Some of the vessels (of smaller size) are represented
with rowing oar, or sail, or both (figs. 4, 7b/11, 12, 13). Three sails (fig. 11)
propelled the very large freighters. The inscriptions associated with the
vessels on the Althiburus mosaic probably indicate different types that were
known to the Greek and Roman Mediterranean. Some of these names were
identified in the historical writings and were even identified as specific
vessels. The Migdal ship (fig. 4) is a small monokrotos vessel (single-
banked). It may indicate a light craft used as a passenger transporter and
possibly associated with actuaria (akatos, in Greek). In Latin this type refers
to a merchant galley, while in Greek it is a generic name for "boat"". The rig
consisted of a single square sail and the hull was of the pointed cutwater
type35.Actuaria is depicted on the Althiburus Mosaic (fig. 7b113) along with
my~paro (fig.
~ ~7b/13), a type which can also be associated with the Migdal
ship. If the Migdal ship represents a passenger carrier, then it would not
accommodate more than 10-15 people (not including the crew of six). The
load capacity of such a vessel would not be more than 1.5 ton3'.
The vessels on the Althiburus mosaic represent a range of crafts, from

the simplest rowboat-ratis (fig. 7b/14), fishing boat-cydarum (fig. 7b/19), to
the one-masted-corbita (fig. 7/1), actuaria (fig. 7b/13), and the largest types
ZARAZA FRIEDMAN TROPIS VI
rigged with two masts-ponto (fig. 7b/3), and cladivata (fig. 7bl4). The mosaic
also depicts vessels that were used on rivers or along the coast for traffic and
fishing: hippago (fig. 7b/6), slatta (fig. 7b/15), celox (fig. 7b/21), horeia (fig.
7b120). The larger freighters are indicated by corbita (figs. 7bl1, 2) and
amphorae carrier (fig. 7b125). The load capacity of the vessels depicted on
the Althiburus mosaic is estimated to be from 1.5 to 400 tons3'.
The ships depicted on the mosaics in the Corporation Square, at Ostia,

also demonstrate a wide range of seagoing merchant vessels along with
crafts used on the Tiber, and harbor services. The ships with rounded hull
and block-shaped sternpost, and rigged with one mast (figs. 10, 12, 131, can
be associated with an adopted form of corbita (fig. 7bll) or prosumia (fig.
7b/12), without the projecting cutwater. The bigger ships rigged with two
masts (figs. 8, 9),may be associated with ponto (fig. 7bl3) or cladivata (fig.
7bl4). The strip of black and white pattern depicted on the masts of the
vessels in figs. 8 and 10 indicates a rope ladder abaft the mast for getting
aloft. It became a characteristic feature of the Mediterranean sailing
vessels39.The three-masted ship (fig. 11) represents the largest seagoing
freighters, probably more common in the western Mediterranean. This
mosaic was found in the shop belonging to shippers from Syllectum (North
Africa), and it is a rare representation of this type of vessel. The load capacity
of such a ship would be 200-500 tons4'.
Conclusion
Despite the wide geographical location of the sites with ship

iconography on the mosaics described above, they are related to one
another by virtue of the subject matter, and the mosaic making of black-and-
white technique. They also originate from the same period, the 1' (Migdal)
and the 3rdcenturies AD (Althiburus and Ostia). These are some similarities
of the depictions of the vessels, especially Althiburus and Ostia.
The style of black-and-white mosaics developed in the second half of the

1" century AD and was used until the late 3'dcentury. Such mosaics are
mainly found in Italy (Rome, Pompeii, Herculanum) and only very few
examples exist in the Roman provinces. Since the technique of black-and-
white style is much simpler than the polychrome style, the artists developed
a great skill for the use of chiaro-scuro (light and shade). Economic factors
may have dictated the use of the black-and-white style. This technique
requires less planning and is thus faster in production.
The depiction of the Migdal ship, in black-and-white technique, may

indicate that the owner of the house was familiar with the new fashion of
interior design, which developed in the center of the Empire, or that he was
also lead by economic factors. Building his house on the shore of the Sea of
Galilee and wishing to decorate the pavements in the new fashion
encouraged him to use local material from the area. The black stones are
basalt, while the white and brownish-red are different kinds of limestone,
which could have come from around Kinneret, Galilee or the Golan Heights.
The Catalogue of Ships from Althiburus belongs to the period of the 3'*
century, when the Roman economy was at its apex. The African provinces
became the main suppliers of corn, olive oil and garum for Rome. This
change in the economy led to extensive planning and building. The mosaic
floor may indicate that the owner of the house was involved in the business
of shipping products brought from rich inland plains, and shipped to Rome
through the port of Carthage. Despite the fact that the depiction of the water
gods is formed in polychrome tesserae, the vessels were depicted only in
shades of gray on the olive-green background of the sea4'. The style of the
work used for the vessels is very similar to the black-and-whitetechniq~e~~.
The making of this mosaic can be attributed to three factors: 1. The influence
of the new style from the center of the Empire; 2. The use of local material;
3. The cheaper and much faster way of making the mosaic floor.
The black-and-white mosaic in the Corporation Square, at Ostia, may

indicate a unity of the economic factor and the height of fashion for making
such floors. The prosperous economy of the Roman Empire during the first
half of the 2"d centurys led to massive planning and building in Italy,
especially in Rome and Ostia. The fashion of paving the private and public
buildings with mosaics led to the use of the black-and-white style. It required
less planning, and the making of such floors was faster and the setting was
limited to short periods.
Although the mosaics presented in this paper have different

geographical locations, they belong to the same period and are connected
one to another through cultural and economic factors which spread all over
the Mediterranean via maritime connections.
Acknowledgments
This article was made possible with the help and encouragement of
several people. I owe special thanks to my advisor Prof. Michal Artzy, from
the Department of the Maritime Civilizations, University of Haifa. Through the
research of articles and other publications for this work, I came across some
material published in German. I am grateful to Mr. Alex Neber, from the
Zinman Institute of Archaeology, University of Haifa, for helping me with the
translation from German to English. I am thankful to Ms. Nichole Nachshon
for the English editing. I wish to thank Mr. Ezra Marcus, from the Department
of Maritime Civilizations, for giving me some pictures of the mosaics which
he photographed at Ostia.
Zaraza Friedman
Center for Maritime Studies
University of Haifa, Haifa 39105
Israel
NOTES
1. The Synoptic Gospel describes Mary of Magdala as one of the women from the Galilee who
gave financial help and domestic services to Jesus and his disciples. She also was present
at the crucifixion and the burial of Jesus. The Fourth Gospel gives Mary of Magdala pride of
place as the first witness of the resurrection and the risen Christ; Comay, J. & Brownrigg, R.,
1980: Who's Who in the Bible; The Old Testament and The Apocrypha, The New Testament;
Bonanza Books, New York; pp. 299-301.
2. The War of the Jewish, Book iii. 9-10, pp. 74-6.
3. Strabo, Bookxvi. 2.45.
4. Raban, 1988, p. 323.
5. This type,of inscription is the first one to be found in Israel and dated to the 1" century AD.
Such inscriptions were mostly found in private houses in Antioch. They were used as a sign
of protection against the "evil eye"; Corbo, 1978, p. 237.
6. Literally meaning 'eye". It is a device in the form of an eye and sometimes highly stylized. The
decoration was painted on either side of the bow, close to the stem, for reasons of religion
or superstition.
7. Basch, 1987, fig. 871 (ivory fragment from Chios, end of the 7m century BCE), p. 409.
8. Basch, figs. 919, 921, 1081, 1089; Casson, 1974, fig. 177; Casson, 1994, fig. 97.
9. Basch, figs. 971-8, 973-A.
10. Casson, 1971, p. 146; Steffy, 1994, pp. 277-8.
11. Basch, fig. 41, p. 377. The graffito comes from the Maison aux Stucs, in Delos.
12. McGrail & Farrell, 1979, Table 1, p. 157; fig. 6, p. 160.
13. Wachsmann & Steffy, 1990, p. 120.
14. Ibid., pp. 29-47.
15. The measurements of the mosaic and the ship model were taken by the writer in 1996 and
appear in a table used for the MA Thesis, 1999; Table 2.1.1, p. 12.
16. Schmerbeck, 1992, p. 16.
17. The Althiburus mosaic was brought into this paper as a black-and-white example. Recently
I was made aware of a German article about the Althiburus mosaic that was published in
1992. It presents some general information about the entire mosaic, but refers to four types
of vessels depicted on this floor. Although some elements such as the fish and the water
gods associated with the maritime scene are depicted with colored tesserae, the vessels are
illustrated with white and darker shades of gray. The water background is depicted with
olive-green tesserae. Light-colored stones were used for the zigzag strips depicted as the
waves of the sea. Since the first publication (1905) of the mosaic in black and white, there
was no mention of the technique used for its making and the colors that were used.
18. Such flags were attached to the masthead of the flagship in sea combat. In merchantmen it
was used either as a trademark, or to indicate the wind direction, as can be seen on
contemporary yachts.
19. Meiggs, 1973, p.55.
20. Ibid.
21. Tacitus records the loss of 200 vessels within the moles of the harbor, due to a severe storm;
Meiggs, p. 55.
22. Meiggs, p. 280.
23. Ibid, fig. 2 (plan of the site), p. 137.
24. Ibid, p. 283. The colonnades are contemporary with the original building of the theatre,
during the time of Augustus (I* century BCE).
25. Ibid, p. 285.
26. Meiggs, p. 283.
27. Ibid, p. 287.
28. Hermansen, 1981, p. 74.
29. Meiggs, p. 279.
30. Ashby, 1912, p. 179.
31. Houston, 1980, p. 156, note 70.
32. It appears that the topsail was used in iconographic representationon vessels, not later than
the 3" century AD; Casson, 1971, figs. 144, 149, 154.
33. Casson, p. 211.
34. Casson, p. 159.
35. lbid., p. 160.
36. Originally, myoparo was a type of single-banked warship, beamier in proportion with its
length; Torr, 1964, p.118. In the Althiburus mosaic it is depicted as a vessel with a concave
prow ended with a projecting pointed curtwater.
37. See above p. 178, point 1.
38. Casson gives a list of large freighters mentioned in historical writings and also a table of the
wrecks found with their cargo; Casson, pp. 183-4 and pp. 198-90.
39. Casson, p. 240.
40. Ibid., pp. 189-90.
41. See note 17, above.
42. The black-and-white style had great influence on the polychrome mosaic in the organization
of the field of the composition; Clarke, R., J., 1979: Roman Black-and-white Figural Mosaics;
New York University Press; pp. 58-62.
BIBLIOGRAPHY
Abbreviations
Atiqot The Israel Antiquities Authorities
DEGUWA Deutsche Gesellschaft zur Forderung der Underwasserarchaologie e. V.
IJNA The InternationalJournal of Nautical Archaeology
JRS Journal of Roman Studies
LA Liber Annuus
MAAR Memoirs of the American Academy in Rome
Monuments et Memoires Fondation Eugene Piot;
L 'Academic des Inscriptions et Belles-Lettres
Ashby, T., 1912: Recent Discoveries at Ostia; JRS vol. II, pp. 153-194.
Basch, L., 1987: Le Musee imaginaire de la marine antique; Athens.
Casson, L., 1971: Ships and Seamanship in the Ancient World; Princeton University Press,
New Jersey.
Corbo, V., 1978: Piazza e Villa Urbana a Magdala; L A XXVIII, pp. 232-40, pls. 71-6; Franciscan
Printing Press, Jerusalem.
Hermansen, G., 1981: Ostia: Aspects of Roman City Life; University of Alberta Press.
Houston, G., D., 1980: The Administration of Italian Seaports During the First Three centuries
of the Roman Empire; in d'Arms, J., H. & Kopff, E., C. (eds.): The Seaborn Commerce
of Ancient Rome: Studies in Archaeology and History; MAAR vol. XXXVI, Rome;
pp. 157-172.
Josephus, F.: The War of the Jews, vol. 8, Book iii, in Complete Works of Josephus in Ten
Volumes; The World Syndicate Publishing Company.
Glaucker, P., 1905: Un Catalogue Figure de la Batellerie Greco-Romaine, la Mosaique
d'Althiburus; Monuments et Memoires vol. XII, pp. 113-154, Pls. ix-x
McGrail, S. and Farrell, A., 1979: Rowing: aspects of the ethnographic and iconographic
evidence; IJNA 8.2, pp. 155-166
Meiggs, R., 1973: Roman Ostia (2"4ed.); Clarendon Press, Oxford.
Raban, A., 1988: The Boat from Migdal Nunia and the anchorages of the Sea of Galilee from
the time of Jesus; IJNA 17.4, pp. 311-29.
Schmerbeck, U., 1992: Das Schifsmosaik von Althiburus; DEGUWA 2, pp. 16-20.
Torr, C., 1964: Ancient Ships; Argonaut, Inc., Publishers, Chicago.
Wachsmann, S., 1990: The Excavations of an Ancient Boat in the Sea of Galilee (Lake); Atiqot
XIX, Jerusalem.
LIST OF ILLUSTRATIONS
Fig. 1: Sites mentioned in the text; drawing 2. Friedman.

Fig. 2: General plan of the excavated site at Migdal; Corbo, P., V., 1978;
LB XXVII, p. 72.
Fig. 3: Migdal mosaic panel; photo, 2. Friedman.
Fig. 4: Drawing of the Migdal ship, 2. Friedman.
Fig. 5: Ship graffito from Delos; Basch, 1987, fig. 41, p. 377;
computer process, 2. Friedman.
Fig. 6: Plan of the Maison des Muses, Althiburus; Glauckler, P., 1905, fig. 2, p. 123.
Fig. 7a: The Catalogue of Ships, Althiburus; Dunbabin, 1978, fig. 122;
computer process, Z. Friedman.
Fig. 7b: Drawing of the Catalogue of Ships; Casson, 1971, fig. 137.
Fig. 8: Navi Narboninses, Ostia; photo, E. Marcus.
Fig. 9: Navicularii et negotiantes Karalitani, Ostia; photo, E. Marcus.
Fig. 10: Nacularii Karthagiensi, Ostia; photo, E. Marcus.
Fig. 11: Three-masted merchantmen, Ostia; photo, E. Marcus.
Fig. 12: Rounded hull with block-shaped stem-and-stern posts, Ostia; photo, E. Marcus.
Fig. 13: Rounded hull with angled stem, Ostia; photo, E. Marcus.
Fig; 2 dinera1 plan of the excavated site
Fig.3 Migdaf mosaic panel Fig. 4 Drawing of the Migdal ship
193
Fig. 5 Ship Graffito from Delas
Fig. 6 Plan of the Maison des Muses, Althiburus
Fig. 7b Drawing of the Catalogue Ships

SHIP ICONOGRAPHY ON "BMCK-AND-WHITE" MOSAICS
P I ~ H. Navl Narboninses
Fig. 9 Naviculari et negotiantes Karlitani, Ostia
195
ZARAZA FRIEDMAN TROPIS VI
ANCHOR LOOK-ALIKES
INTRODUCTION
Anchors, by definition, are instruments for immobilising a floating object.

But the definition is wide, for, in principle, anything from a battle-ship to a
fishing line can be anchored. Over-interpretation being the besetting sin of
archaeology, it follows that stone anchors lend themselves to it in a big way.
Weight remains the single most important factor in considering ship's
anchors, because of the obvious correlation between anchor-and ship-size.
Some Bronze Age stone anchors weigh as much as a ton, although the half
to quarter-ton range is commoner. With the 20-30kg. weight-range (that is to
say stones light enough for one man to handle) complications set in, for
throughout the world such stone anchors are still used on small craft. In the
Mediterranean, only the poorest fishermen still rely on them entirely,
nevertheless old habits die hard: I recently saw a Sicilian boatman, with
tourists crowding his converted trawler, pull a stone out of his hatch and
substitute it for the iron grapnel he had just lost. So unless there is clear
proof of period, it cannot be assumed that any light-weight stone anchor is
ancient. Publications add to the confusion, because any stone with a hole in
it, from the smallest amulet to the largest counter-weight, can look like an
anchor if it is portrayed without a scale beside it; to crown all this,
archaeologists have to distinguish between stone anchors made for use at
sea and the many look-alikes found on land.
Despite such complications, ancient anchors are well worth studying,

since they yield information unobtainable from any other source. This is
because every anchor lost on the bottom marks the passage of a certain
ship so, providing find-places are recorded and typological comparisons
established, it should be possible to deduce a picture of the sea-lanes of
antiquity; the ownership of ancient fleets and the relative sizes of their
individual vessels. The anchors that divers discover tend to be grouped on
certain offshore shallows where the square-rigged craft of antiquity were
forced to moor (Green, J. 19-), as well as on harbour sites and, of course,
on wrecks.
HONOR FROST TROPlS VI
Identifying stone anchors rests on three basic criteria:

1, on typological matching with specimens on stratified land excavations;
2, on specimens found in situ on wrecks with contents of known period.
3, in regions where anchors have not, as yet, been identified by
comparisons on either land-sites or wrecks, but where great numbers of
them are found undersea (for instance Israel, Turkey and the Bulgarian Black
Sea coast) the recorded aggregate will produce statistically predominant
forms. The significance of any particular form may then be built up through
long-term observations and analyses of characteristics, for instance:
whether the rock from which a specimen is made is local, or foreign to its
find-place.
All the same criteria apply to the much-neglected subject of fishing

tackle, except that tackle gets lost on fishing grounds rather than offshore
moorings. For anchors of both kinds, typological comparisons are, at
present, manly based on major Levantine excavations (in Cyprus, Syria and
Lebanon) where numerous votive specimens have been found in tombs,
shrines and temples, as well as in wells, springs, etc. (Frost, H.: Byblos 1969;
Ugarit, 1969 & 1998; Kition 1985). Such contexts are explained by the
universal salvation-symbolism accorded to anchors (for it is only an anchor's
hold that saves a vessel from being driven to destruction); similarly fishing
tackle symbolises and represents the fertility of the sea.
The seeming lack of votive anchors in major Greek and western

Mediterranean excavations is probably fortuitous, since classical texts
mention such offerings (eg. Apollonius of Rhodes, Argonautica 1,954),while
at the beginning of the 20thcentury the distinguished Greek numismatist, J.
Svoronos, was well aware of the sacred nature of anchors, particularly those
made of stone which he went on to discuss (Svoronos 1914). More recently
P. Gianfrotta (1975), 1 myself (1982) and G. Boetto (1997) have added
instances of stone anchor stocks in post-Bronze Age temples in Greece,
Cyprus and Magna Graecia. However it must be remembered that, before
the popularisation of SCUBA diving, archaeologists seldom understood
what purpose had been served by the pierced, grooved and sausage-
shaped stones they were excavating on land. Even on such a key-site as
Ugarit-Ras Shamra in Syria, its excavator, C.F.A. Schaeffer, for many years
called the anchors " des curieuses steles percees " (Schaeffer 1931).
Similarly, F. Proni (in 1922) sketched a stone stock in an Etruscan tomb,
describing it as " una stele (?) d'arenaria ... "; later, when it reached the
museum at Ferrara it was catalogued as "un oggetto singolare di arenaria",
and only recently was it recognised as part of an anchor (Gianfrotta, P. 1982
ANCHOR LOOK-ALIKES
and Kapitan, G. 1986). By contrast, the present vogue for stone anchors has
brought the new danger of going to the other extreme and indiscriminately
interpreting as anchors the many other oddly shaped stones which had
served quite different purposes on land and undersea.
Anchor look-alikes can be divided into two main categories by size: the
larger pierced stones look like anchors for vessels, and the smaller stones,
in the order of a kilo or less, which weighted down fishing tackle, can look
like loom-weights and similar objects that had other domestic or agricultural
uses. Fishermen's weights take many shapes that can only be learned by
comparison and experience; some look like miniature ship's anchors, but
another very common variety is egg-shaped with a central belt-like groove
cut round it. Such weights are as yet largely unclassified, for although the
importance of the sea to the ancients is apparent, the history of fishing
remains unstudied.
SCUBA diving now makes it possible to compare land and undersea

finds, but in the case of ship's anchors one must be prepared for a few
surprises; the above-mentioned Sicilian boatman was not the first to lose an
anchor and make-do with something else. Accidents have always occurred,
and stone anchors not being particularly efficient, they were and are easily
lost. In antiquity, a crew would substitute any stone object of the right weight,
until such time as they could have a replacement made; in the interim they
might also lose the substitute, as is suggested by the following examples.
OLIVE PRESS WEIGHTS AND OTHER LARGE LOOK-ALIKES
Olive presses feature in most urban excavations (Heltzer, M. 1987,

Hadjisawas, S. 1992) and the counter-weights belonging to them resemble
ship's anchors (Fig. 1). In the town of Ugarit (which is built on the slopes
surrounding the acropolis where the weather god's temple which contains
votive anchors) there are several Late Bronze Age presses containing
asymmetrical counter-weights (Fig. 2) with a projecting shoulder on one
side. This lop-sidedness is unknown among stone anchors, yet a very similar
stone (Fig. 2a) was dredged up with many 'ordinary' stone anchors from the
now silted lagoon surrounding the Phoenician Island-town of Motya in
Sicily*. Motya's Levantine connections may explain the introduction of this
shape of press-weight.
Press-weights are not, however, the only stones that resemble anchors.
Other look-alikes include: door-sockets and hinges (Fig. 3) and even

hammers. At Thera, the discovery of the ship frescoes led to a search for
stone anchors. This was before it was realised that the successive
earthquakes that had toppled the houses had also raised the harbour out of
the sea, so that it is now thought to lie buried on land, mid-way between the
excavation and the present shore-line, while the same cataclysms make it
unlikely that Bronze Age artifacts would remain visible on the surface of the
sea-floor. This was not apprehended when Dr. Marinatos first started
excavating, so it was suggested that two stones might represent anchors.
One of them was big and long-shaped with a central piercing; the other was
small and hammer-shaped. Both are likely to be kinds of hammers; the
larger was probably suspended, then swung into walls made dangerous by
early tremors and in need of demolition.
On excavations that have continued for a century or more (Ugarit, for

instance, was first excavated under Ottoman rule in the lgthcentury before
the present French mission took over) some look-alikes from superficial
levels, like mediaeval torch-holders, mounting-stones, etc. may remain on
site to confuse specialists looking for anchors.
Figure 4 shows pierced-stonetorch-holders in their rightful place on the

front of a mediaeval Venetian house, which Lucien Basch (who took the
photograph) calls << the house of the anchors >,. Islam, as well as the
commerce of the Serenissima, left equivalent buildings all round the shores
of the Eastern Mediterranean where, since shelter is rare, they were built on
top of Bronze Age ports. This may have been stopped by the use of concrete
in founding seaside hotels and high-rise Rats as well as the yacht marinas
that extend out into the sea itself. On land, motor roads have not caused so
much damage. Mounting-blocks and tethering-posts survive along the
ancient trade routes for silk and porcelain crossing the deserts, mountains
and steppes between China and Europe. In one such area of Eastern Turkey
beneath Mount Ararat, not only has a curious, oval shaped rock formation
been claimed as the remains of (< Noah's Ark >>, but also that the Ark's
anchors are still in situ (Fasold, D. 1988). One of them shown on television
took the form of a very large, pierced stone implanted upright in ground
where caravans would have found it useful as a tethering post.
ANCHOR LOOK-ALIKES
FISHING TACKLE AND THE SMALLER ANCHOR LOOK-ALIKES
Again, loom-weights, spindle-whorls, polishers and " anchors for

animals " (FIG. 5), also resemble the various kinds of stones fishermen use
for anchoring their tackle, examples of which remained lost on the sea-floor.
But whereas votive tackle is also found in sacred contexts on land, domestic
and agricultural look-alikes are not. FIG. 5a shows the ubiquitous belted-
stone, ready to anchor tackle, in a modern fishing boat at Byblos where
similar stones have been excavated on Hellenistic levels. In Egypt, stones of
the same shape are frequently portrayed weighing down Pharaonic fishing
nets. In Cyprus, FIG. 6 shows tackle-weights of other shapes. One, which
has remained unfinished, is like a miniature anchor; all were found in
workshops built within the precincts of the Kition Temples. A different form
of votive offering is represented by the three-holed, trapezoidal anchor
suitable for a small boat (FIG. 6a). It came from the bottom of a Bronze Age
well and exemplifies another context where both votive tackle and boat
anchors probably expressed a vessel's need to take on freshwater and the
hope that it would always be found, since along the Levant coast wells and
streams tend to run dry, so water is a precious commodity.
A CORAL FISHING CONTRAPTION
Finally, the large square stones with five piercings, found on the sea-floor
throughout the Mediterranean, are usually classified with anchors if not
always claimed as such. In the strict sense they are neither look-alikes, nor
anchors (because their function is not to immobilise a floating object, but
only to reduce its buoyancy). The square slab of stone with five piercings
(FIG. 7) belonged to an instrument used by early coral fishers, the all-wood
forerunner of the (< St. Andrew's Cross >> (which, being partly metal did not
need much extra weight Fig. 8). Such crosses were current by the beginning
of the 18'hcentury; their use in the Western Mediterranean was then banned
by Spain in 1832 (M. Pernay 1981) because of the indiscriminate destruction
they caused to the already dwindling supply of coral " trees " (or the calcified
remains of polyps, which grow upside down under shelves of rock Figs. 8
and 9). Although Mediterranean tradition, and such dictionaries as
Rodriguez Santamaria (1923) and De la Blanchere (1926), accept the
connection with coral gathering, there has been no archaeological evidence
for either the function or the date of the five-holed stones, because they are
usually found lying in isolation on the bottom where they had had to be
abandoned for some reason, such as getting caught under rocks, so that
their cable had to be cut. After that, their net and wood components
gradually perished. In the process, the central iron pin which had united the
stone with the planks, either became displaced, or rusted away. The recent
discovery, near to one of these stones, of the concreted remains of an iron
pin complete with its ring and a link of chain (and also, for good measure, a
I* century AD amphora sherd embedded in the concretion) produced the
missing evidence. The discovery was made in Sardinian waters by M.
Galasso (1996) who also deduced how the parts were rigged (see Fig. 9).
In conclusion: the symbolism of anchors and tackle is interesting. The .

identification of anchors found undersea (on the basis of comparison with
specimens from dated contexts on land, or on wrecks) constitutes the
unique and valid method of building up a picture of shipping and fisheries in
antiquity. The perils of << anchorology >> do, however, include confusion over
look-alikes, and indulging in fashionable but profitless hypothesising about
how wood and string components fitted into the various piercings; in any
case, these questions are gradually being resolved by discoveries of mud-
buried specimens found complete with their organic components, as well as
by ethnological comparisons.
Honor Frost
31 Welbeck St.
London W.l
REFERENCES
De la Blanchere, H., 1926, La Peche et les Poissons, Dictionnaire General, Paris.

Boetto, G., 1997, e< Ceppi litici asacri~e culti antichi a Metaponto e a Locri >>,
Archeologia Subacquea 11, 51- 64, Rome.
Fasold, D., The Ark of Noah (LCCN 88020437), New York, and the << anchor shown in TV (GB)
programme << Encounters n 13.2.94.
Frost, H., 1969 << The Stone Anchors of Byblos *, Melanges de I'Universite Saint
-
Joseph LXV, offerts a Maurice Dunand, Beirut.
Frost H., 1991, The Stone Anchors of Ugarit m, Ugaritica VI, 235-245, Paris 1969 and
Anchors sacred and profane; the Ugarit-Ras Shamra Stone Anchors Revised and
Compared*, pp.355-410, Ras Shamra-Ougarit VI, Paris.
Frost, H., 1985, The Kition Anchors *, Appendix 1, Excavations at Kition V (part 1 )
pp.281-321 & Plates A-N, Cyprus.
Frost, H., 1982, ee The Birth of the stocked anchor and the maximum size of early
-
s h i ~ sm. DD. 263-273. The Mariner's Mirror 65.2. London.
Galasso, M. iin press 1996) ~invenimentiarcheologici subacquei in Sardegna Sud-
Occidentale >,,Atti del Congresso Nazionale di Archeologia Subacquea, pp. 2-
8, Anzio.
Gianfrotta, P. Le ancore votive di Sostrato di Egina e di Fallo di Crotone m, La parola del
passato, 163, pp.311- 318, Naples, 1975.
Gianfrotta, P., 1982, ec L' ancora di Kutikluna, owero considerazione sulla tomba
ANCHOR LOOK-ALIKES
n. 245 di Valle Trebba ,,,Musei Ferrarese Bollettino Annuale 12, pp.59-62,

Ferrara 1984.
Hadjisawas, S., 1992, Olive Oil Processing in Cyprus, Nicosia.
-
Heltzer, M., 1987, Olive Oil in Antiquity (Conference 1987), Haifa.
Kapitan, G., 1986, Klutikunans anchor and the question: was a stone anchor stock in
the tomb, or a complete stone stocked wooden anchor ? IJNA 15, 2, pp.133-
136, London and New York.
Pernay, M., 1981, u La pbche du corail en Mediterranee m, L 'homme mediterraneen
et la mer, p. 251, Jerba.
Rodrigues Santamaria, B., 1923, Diccionario de Artes de Pesca da Espana y son
Posesiones, Madrid.
Schaeffer, C.F.A., 1931, u Deuxieme campagne * (Minet el Beida), Syria XI/, p. 13, .
Paris.
Svoronos, J., 1914, <e Stylides, ancres hierae, aphlasta, stoloi, embola proembola et
totems marins n, Journal international d'archeologie numismatique, 16, pp.105-110
(anchors), Athens.
NOTE
The waterworn anchors of miscellaneous shape from the Lagoon surrounding Motya must
represent small craft of early but uncertain date. Known as the first of Tyre's colonies in the
West, the Island-town has Levantine Phoenician associations, although its foundation post-
dates the Levantine Bronze Age (but here as elsewhere stone anchors probably overlapped and
co-existed with those of subsequent design). The lagoon was navigable before the fall of the
Island-town. It subsequently silted. The depth of water round the Island now, is little more than
20 cm. so that to allow boats to reach it, channels have to be periodically dredged and in the
course of this work 6m. of Poseidonia rhysomes have been reported. I know of no precise
records of the anchors' discovely.
BIBLIOGRAPHY
L'afte del corallo in Sicilia, ~ U sei procedimentiproduttivi-, Enzo Tartamella, pp.

137-146 (Catalogue of the Exhibition in Museo Regionale Pepolo, Trapani) by
various authors, Palermo 1986.
-
Liverino, B., I1Corallo, p. 12, Naples 1998.
Riccardi, E., A note on Stones with Five Holes ,,, The Mariner's Mirror 82, 2,
pp. 200-203, 1996.
FIGURES
1. A Hellenistic olive press with counter-weights (one rectangular), at

Byblos.
2. A Late Bronze Age olive press (Ugarit); note the asymmetrical shape of the counter-weight.
3. An anchor look-alike sewing as the hinge of a modern gate ((Majorca).
4. Anchor-like torch-holders in a Venetian house (Salizada S. Lio, 12 Sestieri).
Photo Lucien Basch.
HONOR FROST
5. Bovine with a cc belted stone * attached to its hind leg. Prehistoric painting from Fessan (after
A.Castiglione and G. Negro, Fiurne dipietra, no. 8,219, Varese, 1986). 1 am indebted to Mark
Milburn for this reference.
5a. A fishing boat at Byblos; note the belted stone with the tackle (arrowed).
6. Four fishermen's weights found in the workshops that adjoined and served the Temple Area
of Late Bronze Age Kition (Cyprus); the miniature anchor on the right is unfinished. The three-
holed trapezoidal specimen which must have belonged to a small boat (its weight being c.
20kg.), was found by workmen at the bottom of a Bronze Age well, in an area of Kition that is
overbuilt by the present town of Larnaka.
7. An early coral-fishing stone (Haifa Museum). I am indebted to A. Raban for showing it to me.
8. An 18m-centurycoral-fishing St. Andrew's. cross * probably made of metal, after one of the
many contemporary engravings (Salmon, Lo stato presente di tutti Ipopoli, Venice 1740-
1766, pl. XXIV).
9. Reconstruction of an earlier, perhaps Roman, wooden instrument for gathering coral
weighted by a five-holed stone as in Fig. 7 (based on the findings by M. Galasso, see
reference).
-
ANCHOR LOOK-ALIKES
HONOR FROST TROPIS Vl
Fig. 8
ABSTRACT
THE ARGO SHIP IN ROME
In the XIX epigram of Book VII, the poet Martial refers to a relic of the
Argo ship. The association with another reference by Martial (verse XI1 of
epigram I in Book XI) which refers to Jason, now allows us to clarify that this
relic, evidently part of a ship's plank or ceiling, was conserved in the Porticus
Argonautarum identified in one of the long flanks of the Campo Marzio
Saepta in Rome, built by Agrippa in 25 BC.
It is likely that Agrippa himself had the fragment of the Argo ship, which
had been preserved in the temple of Poseidon on the Corinth Isthmus,
brought to Rome in order to connect Octavian and himself ideologically to
the mythical Argonauts, defenders of western civilisation against the eastern
<<barbarians,,.
A thematic analogy can be found at Delphi in the Sycyonian Treasury.
Piero A. Gianfrotta
Universita di Viterbo
Via S. Camillo De Lellia
Viterbo 0 1100
Italy
SHIPS OF THE 12mDYNASTY, EGYPTIAN KINGDOM
AND THEIR RELATION TO 17'" CENTURY BCE AEGEAN SHIPS
The earliest watercraft that we know of today is the Royal Ship of King
Cheops of Egypt. This magnificent boat was built near 2600 BCE: 4600 years
ago. It is a most ancient ship, built more than two thousand years before
those of Classical Greece. It was buried in a ship grave at the foot of the
pyramid of its owner, Pharaoh of Egypt in the Kingdom's 4th Dynasty.
I emphasize the age of this vessel because of some similarities of its

structure to later Egyptian craft in the 12th Dynasty - something about 800
years later. The similarities are basic.
The hull and structure of King Cheops' ship is built of heavy wood
planks. It also has a heavy central planked bottom. There is not one center
plank but three, side by side, forming a central strength core for the ship's
bottom. To quote from Nancy Jenkins who worked with the ship's restorer
Ahmed Youssef Moustafa, the noted Egyptian archaeological authority of
Egypt's Department of Antiquities, "The bottom or central plank was made
up of eight timbers in three sections". These planks were all several inches
thicker than the other planking of the ship and all carefully scarfed together
longitudinally.
This concentration of a stronger bottom as central core emphasized the

builder's awareness of need for a center strength girder. Not only were these
early ship builders aware of bottom strength particularly in a long narrow
hull, they were, from experience, knowledgeable about the forces that cause
a downward bend in the fore and aft ends of the ship. This creep feature is
known as hogging.
The inner structure of this ship of Cheops is also remarkable. It includes

a longitudinal internal truss assembly that extends through the major part of
the hull's length. This internal support is illustrated in a section through the
mid part of the hull (Fig. 1).
I will repeat: this ship was built during or before the 26th century BCE.
THOMAS GILLMER C. TROPlS VI
From that time the builders of ships and boats in Egypt kept the longitudinal
center strength of their hull with the heavy center planks in place and in mind
for the next one thousand years, beyond their middle kingdom (ending in
1660 BCE), and even beyond that. It was the beginning of a fundamental
tradition of shipbuilding.
The conservatism of shipbuilders is well known and it must have begun

early. The existing evidence throughout the Egyptian Dynasties between the
4th and the 12th show little difference in the basic hull structure. However
there are some significant differences in the hull planks and bottom shape.
The longitudinal strength of Cheops' ship depended on the greater
thickness of the several bottom planks laid side by side to form a narrow but
flat bottom from end to end. The longitudinal strength was further supported,
together with the vessel's long arching hull, by an internal truss of vertical
center posts supporting a central deck girder, and each standing on a
transverse frame of the hull shaped to the boat's bottom (Fig. 1). This is a
sophisticated structure.
Nearly 1000 years later we can only look at the shell of the Dahshur boat
of the 12th Dynasty and wonder what improvements have been made in the
state of the art (Fig. 2). There are two significant changes in that gradual
evolution. First the hull does not have a bottom as heavy nor is it flat and
without raking sides. It is almost semicircular. An early round bottom
concept.
This round section is stronger essentially than the flat bottom

configuration of the Cheops ship. The planks are not sewn nor lashed
together but the edges are fastened snugly by butterfly shaped (opposed
wedge) external morticeltenon joints. The cross deck beam-planks are let
through and fastened in the upper strake of side planking. Finally there is
showing plainly, as noted earlier, the existence of an embryonic keel plank.
It is a continuous, end to end, center line plank. There is no other plank that
is continuous or attempting this balanced continuity (Fig. 3). This was the
state of the art in the 12th Dynasty of Egypt.
At this point, I must define the term "state of the art" as I am using it here.
Today's world abuses this terminology badly, however it is a valuable term
when applied to one or more existing cultures in comparable eras, whether
or not we are considering ancient worlds. I believe, as applied here, the state
of the art in Egyptian history can be applied over many centuries as it
developed between the 4th and 12th dynasties before there was any great
SHIPS OF THE 12'" DYNASTY, EGYPTIAN KINGDOM
AND THEIR RELATION TO 17"' CENTURY BCE AEGEAN SHIPS
collapse of cultures. Also, in shipbuilding, as in other technologies, there is

a lateraltransfer of skills and methods where in this case there is, for a given
period, a state of the art between Egypt and the people inhabiting the Aegean
Islands of Crete and the Cyclades, all of the same period (Fig. 4a & 4b). I
believe this because there were no menacing political or geographical
barriers, other than a travelled sea route, between these people who
practiced similar arts or built the same structures. There is also evidence in
Egyptian wall reliefs of these Aegean people transporting their wares.
Methodologies of manufacture transfer between people who sell or trade
their wares very easily. Similar technologies and watercraft are apparent in
more recent history. Example: the ships of England in the 16th century were
not significantly unlike the ships of France or the Netherlands during the
same time, except for regional detail.
More specifically, in our context here, we can see details in Minoan ship
art (Thera ship frescoes - Marinatos, 1970) that show similar features in
ship form and detail with 12th-Dynasty Egyptian ships. Particularly ship No. 1
in upper left of the fresco (Fig. 5) proceeding in procession from left to right,
shows its sail lowered and furled and supported on poles. The multiple
halyards fall from the mast head identically as do the halyards of Egyptian
ships shown in the same Egyptian era, the 12th Dynasty reference.
Further, Egyptian ships of this same period show masts that have lateral
or helical stripes (Fig. 6). These are also present in the ships of the Minoan
Fresco. Whether these stripes on the mast can be functionally identified is
irrelevant at this point. They exist on ships of two different cultures in eras of
the same time (Fig. 5 & 6).
There are other features of similitude between these Minoan period ships
and those in the 12th Dynasty of the Egyptian Kingdoms.
It was these very detailed similarities that attracted my attention some

fifteen years ago, on these ships of Thera and on the fresco itself convincing
me that it was older than had been originally estimated. The date of the
fresco, when first unearthed, was set to agree with the date of the great
eruption about 1450 BCE (Marinatos, 1970) (later corrected). It was my
notion that such conclusion was not necessarily so. It could well have been
earlier but could not, of course, have been any later, having been found in a
house in ruins caused by the cataclysm. However, now that we are aware,
from Arctic snowcap soundings, the volcanic ash of that great eruption
reveals the time with certainty and fortifies the reasoning of shipbuilding
THOMAS GILLMER C. TROPlS VI
"state of the art" determination. The revised date now is extended to 1700
BCE.
Because the profiles of the ships in the Thera fresco make it difficult to
see any of the ship's third dimension, it is the presumed relation to the
Egyptian models and a surviving vessel of the 12th Dynasty that also fortify
some presumed hull shape on the Thera ships. Noting the boat shapes
drawn in perspective, it is evident from the Dahshur Boat and Egyptian tomb
models that the shapes are basically related.
We can note here further similarities between the Thera ships and 12th-
Dynasty Egyptian vessels. Observing these similarities with the Egyptian
ships, the sails are rectilinear with yards at both top and bottom. When furled
the sails and yards are supported on crutch-like poles along the centerline.
Examining the deck layout, there is a shelter "deck house" in the stern for
the presumed ship's commander or personage of leading importance. There
is a "Y"-shaped crotch forward to support the mast when raising or lowering
all to be found on Egyptian ships. The Thera ships are steered with oars on
the after starboard side while the Egyptian boats of the similar period have
long center oars on the elevated stern for steering at the after end of the ship
making for dissimilar steering. But the Egyptian boats are basically river
boats. Steering in river currents and shallows requires a longer oar for center
control with sweeping port and starboard (across the stern). The Thera
rudder-oar steering is required most frequently for sailing a steady course.
There must naturally be some differences that are necessary for local
adaption.
Thomas C. Gillmer
Annapolis
REFERENCES
Jenkins, N. The Boat Beneath The Pyramid, Holt, Rinehart and Winston, New York, 1980.
Landstrom, Bjorn Ships of The Pharoahs, Doubleday & Co., New York, 1970.
Gillmer, T . Identification of Functional Parts of Thera Fresco Ships, TROPlS I l l , Ed. H. Tzalas,
Third Symposium On Ship Construction in Antiquity, Athens, 1989.
Gillmer, T. A History of Working Watercraft of The Western World, McGraw Hill, New York, 1994.
Morrison, J. The Athenian Trireme, Cambridge University Press, Cambridge, 1986.
Marinatos, S. Excavations at Thera VI, 1972, Season, Athens, 1974.
Doumas, Christos The Wall Paintings of Thera, The Thera Foundation, Piraeus, 1992.
SHIPS OF THE 12'" DYNASTY, EGYPTIAN KINGDOM
AND THEIR RELATION TO 17"' CENTURY BCE AEGEAN SHIPS
FIGURES
Fig. 1 Midsection of Cheops' Ship showing vertical stanchion support and stringers support in
section.
Fig. 2 Dahshur Boat of 12mDynasty, King Sesostris, in the Egyptian Collection of The Field
Museum of Natural History, Chicago, USA.
Fig. 3 Planking plan of Sesostris hull. Note center line plank.
Fig. 4a Hull of Sesostris boat in perspective line as compared to Thera ship perspective of likely
hull form.
Fig. 4b Thera Ship Model viewed from same angle.
Fig. 5 Thera ship as it appeared in the well-known fresco. This is a black and white
archaeological reconstruction.
Fig. 6 A 12m-DynastyEgyptian boat as it is simplified, from a procession of similar Egyptian
boats.
Editor's Note: Although the illustrations were mentioned in the text, they were never sent for
publication, notwithstanding repeated requests.
Les Carthaginois, excellents marins mais piètres <<marines,,,préféraient
I'éperonnage à l'abordage. Au contraire, depuis l'invention du <corvus,, par
Caius Duillius, les Romains transformèrent les combats navals en combats
terrestres. Du début à la fin de la première guerre punique, ils furent
vainqueurs grâce à lui, de Myles (260 av. J.-C.) aux Ægates (241 av. J.-C.).
Les éperons trilames étaient utilisés pour le choc latéral qui devint frontal
avec les gros navires hellénistiques (cf. William Murray). Les Puniques s'en
tinrent toujours à leur vieille méthode de I'éperonnage où le bec augmentait
la vitesse du navire et permettait de frapper en-dessous de la ligne de
flottaison, c'est-à-dire sous la préceinte qui structurait I'éperon ennemi (cf.
Michel Reddé).
Cette primauté de I'éperonnage se voit clairement dans le navire

punique de Marsala dont nous avons justement l'avant et I'arrière, avec son
éperon à bec et sa poupe conçue pour éviter d'être éperonnée par l'arrière,
chose la plus commode puisque jusqu'aux dernières galères modernes, en
1708 (cf. Zysberg et Burlet), <<lamanoeuvre d'une galère qui veut attaquer
est de porter sur le derrière du navire qui est le côté le plus faible,,. Le navire
de Marsala, qui fut coulé aux îles Ægates, est donc une navire de guerre
punique fait pour I'éperonnage dans toute sa structure avant et arrière.
On retrouve plus tard le bec dans la flottille du Danube de Trajan sur la

colonne éponyme, notamment sur la trirème de l'empereur. Certes, on
pourrait rétorquer que c'est un appendice ostentatoire mais comme ce sont
des vaisseaux non pontés, leur seule arme est I'éperon contre d'éventuels
marins daces sur le Danube, qui n'avaient sûrement pas de rostre mais
pouvaient les contrer dans la traversée du fleuve.
Plus intéressante encore est la mosaïque de Mâcon également datée du

II9iècle. II ne s'agit en aucun cas d'une représentation de la flottille de la
Saône créée au IV" siècle mais de celle d'une classis sur une mosaïque
faisant pavement dans une villa ayant peut-être appartenu à un navarque en
retraite. Le fait que tous les bateaux représentés soient sous des
<<shipsheds>, très soignés prouve qu'il s'agit d'une vraie flotte. Notons enfin
l'extraordinaire mosaïque de Constantine, du même siècle, représentant un
duel entre un bateau à bec et un bateau trilame, tous deux de même taille.
Selon Michel Reddé, I'éperon à bec et I'éperon trilame ont coexisté au cours
ALAIN GUILLERM TROPlS VI
des âges. II réfute l'idée que le bec soit un éperon <(àun coup,, et qu'il soit
réservé à des bateaux légers. Sa représentation, souvent attestée sur
monnaie et mosaïque à partir du IIe siècle apr. J.-C., montre que les
Romains, n'ayant pas d'adversaire sérieux sur mer, gardèrent leurs
précieuses légions sur le limes au lieu de les embarquer comme «marines>>
et adoptèrent, contre les pirates ou les Barbares, les méthodes
carthaginoises.
Mais ce qui est extraordinaire pour notre propos, outre les

représentations déjà citées, est la fresque du Capitole de 1507 de Jacopo
Ripanda. Elle représente la bataille décisive des Ægates. Probablement
Ripanda possédait des documents aujourd'hui disparus pour peindre cette
fresque. Cela n'a rien d'étonnant, comme nous l'expliquons dans notre
thèse publiée sous le titre .La Pierre et le Vent,, (Arthaud, 1985, p. 45) en
reprenant Lucien Febvre, l'humaniste vénitien Fausto à la même époque
exhuma un manuscrit grec représentant une quinquérème, manuscrit hélas
disparu depuis (Frederic Lane), et entreprit de la faire exécuter. Quant à ses
qualités nautiques, s'il faut en croire un concours organisé par le Sénat, la
quinquérème, malgré ses dimensions, l'emporta sur une galère ordinaire de
l'escadre !
La fresque du Capitole apporte, à mon avis, sur le problème des éperons

à bec, un argument. En effet, le seul navire de guerre antique que nous
ayons retrouvé, le navire punique de Marsala, offre la particularté d'avoir un
éperon à bec, comme nous l'avons déjà vu. Or le bateau de Marsala est plus
gros que la trirème grecque reconstituée (comme nous l'avons exposé in
TROPIS III, Reddé partageant cette opinion indépendamment). II pourrait
donc être une quinquérème en y ajoutant un porte-nage. Le tableau de
Ripanda va curieusement dans notre sens : aux Ægates, deux flottes
homogènes se heurtent, l'une à bec, l'autre trilame. Le malheur veut, pour
que la démonstration soit parfaite, que Ripanda ait inversé les deux flottes.
Mais cela nous semble secondaire, car bien sûr il y a des éléments
fantaisistes comme les Carthaginois affublés d'un turban ! En revanche les
tours de combat et la présence d'un <<corvus~~
(sur une nef...), dans une autre
fresque de cette même salle <<Hannibal»,montrent la solidité de ses
connaissances pour l'époque.
Alain Guillerm
Laboratoire d'Histoire Maritime
Sorbonne/CNRS
EPERONS A BEC ET EPERONS TRILAMES
BIBLIOGRAPHIE
Lucien FEBVRE, Le Problème de l'incroyance au XVP siècle, Albin Michel, 1942 et 1968.
Honor FROST, Lilybaeum, Accademia Nazionale dei Lincei, 1981.
Alain GUILLERM, La Pierre et le Vent, Arthaud, 1985.
La Marine de guerre antique, KronoslSPM, 1993.
The Punic ship of Marsala and the Trireme Olympias, in TROPIS 111, 1995.
Frederic C. LANE, Navires et constructeurs à Venise sous la Renaissance, Centre de
Recherches historiques, 1965.
William MURRAY, Polyereis and the role of the rarn in Hellenistic naval warfare, TROPIS V,
1999.
Michel REDDE, Mare Nostrum, Ecole Française de Rome, 1986.
A. ZYSBERG et R. BURLET, Gloire et misère des galères, Gallimard, 1987.
ILLUSTRATIONS
Figure 1: Villa romaine de la Grange du Bief, vue d'ensemble de la mosaïque.

Figure 2: Même mosaique, détail montrant clairement l'éperon à bec.
Figure 3: Mosaïque de Constantine montrant l'affrontement des deux types d'éperon.
Figure 4: Fresque de Jacopo Ripanda dans la salle Hannibal du Musée Capitolin à Rome, .La
Bataille des îles Ægatesn.
Figure 5: Détail de la même fresque.
AMIN GUlLtERM TROPIS VI
Fig. 1 Fig. 2
Fig. 3
EPERONS A BEC ET EPERONS TRIIAMES
Fig, 4
'"a;?i.
Fig. 5
ABSTRACT
BYZANTINE SHIPWRECKS DISCOVERED AROUND THE MARMARA

ISLANDS (PROKONNESSOS): POINTS OF DEPARTURE AND
PROBABLE DESTINATIONS
Ganos, today's Gazikdy, within the border of Tekirda, situated along the
north shore of the Marmara Sea, is first mentioned by Strabon as a Greek
colony established in the region during the first century BC. Later in the
Byzantine period, from the lothcentury AD onwards, it appears in the written
sources as a monastic center, comparable with Bithynia and Athos. Sailors
passing through the straits used to visit as pilgrims the Ganos Mountains,
which gave their name to the city itself.
Written sources convey, however, very little else about Ganos, especially
about social and economic life there, for which we can turn with benefit to
archaeological evidence. We do not have enough knowledge about the
social and economic life of the population of Ganos except that derived from
some of the historical sources. The archaeological surveys initiated and
conducted since 1991 by the author of the present paper have revealed, for
example, the existence of amphora kilns in the region of Ganos and the
neighbouring village Hora, today's Hosky. Thus we now have evidence of a
vast production of amphoras in Ganos, which must have served as
containers for the wine producted in the region. Bearing in mind that there is
a definite correlation between amphora and wine production, the discovery
of workshops where the most common and circulated type of Byzantine
amphoras (those that we are familiar from Serqe Limani wreck, type Iof the
author's typology were manufactured), provides us with a whole new
perspective on the region, namely that Ganos was a very important wine
production and exportation center.
Still continuing practices in Gazikdy today strongly suggest the

possibility of an uninterrupted cultivation and production of wine from
Byzantine times to our era. In the Byzantine period the vast amount of wine
production of the region was most probably exported to Constantinople,
from where it would have been distributed to the Black Sea countries, the
Balkans and the Mediterranean.
By the rule of thumb method the present author has followed the route
of the ships that departed from Ganos to arrive at Constantinople. The
Marmara Islands, which stand in the Sea of Marmara along the Ganos-
Constantinople route, have been the wrecking points of some of the ships
which departed from Ganos loaded with wine amphoras. The huge number
of amphoras found in the shipwrecks near the Marmara Islands indicates
that this cargo was headed for Constantinople. And the amphoras carrying
the Ganos wine, which have been found all around the Black Sea, lead us to
think that the city was also a stopover port for trade ships.
During the 93/94 survey seasons 10 shipwrecks of the Byzantine period

have been identified around the Marmara Islands, 7 among them loaded
with Ganos amphoraslwine and with a destination tolthrough
Constantinople. One of them is a roof-tile wreck and the other is a water-pipe
wreck. Both of them are dated to the Thcentury by the help of the anchors
and an amphora sample belonging to the wrecks. Another wreck found is
loaded with the last amphoras of the maritime commerce (type IV of the
author's typology).
The aim of this paper is to acquaint you with the discovery of the
shipwrecks around the Marmara Islands, and focus your attention to the
GanosIConstantinople route via Prokonnessos.
Nergis Gunsenin
Bogazici University
Tourism Administration Program
Hisar Campus
80815 Bebek
Istanbul, Turkey
BIBLIOGRAPHY
<<Analyses Chirniques Comparatives des Arnphores de Ganos, de I'ile de Marmara et de
I'Epave de Serqe Lirnani (Glass Wreck)., Anatolia Antiqua V (L'lnstitut Franqais
d'Etudes Anatoliennes George Durnezil - Istanbul), Paris, (1997), p. 249-260 (with
Helen Hatcher).
<<Recentes decouvertes sur I'ile de Marmara (Procennese) a I'epoque Byzantine : Bpaves et
lieux de chargementn, VllmISBSA, Archaenautica, Paris, (1999), in-press.
<<LeVin de Ganos : les amphores et la mer>>, Eupsycia, Melanges offerts a Helene Ahrweiler,
Byzantina Sorbonensia, Paris, (1998), p. 281-291.
"From Ganos to Serqe Lirnani: an essay on social and economic activities in the middle ages,
through the evidence of recent archaeological and historical discoveries", The INA
Quarterly, Texas, Spring 1999, in-press.
THE KYNOS SEA-FIGHTERS: EXCEPTION OR RULE?
Thucydides (1, 13) tells us the first sea battle of Corinthians against
Corcyraeans, was fought around 664 BC. This fits poorly with vase paintings
of ship-to-ship combat on some Late Mycenaean kraters of c. 1200 BC from
Kynos Livanaton on the mainland shore of the Northern Euripos, which by
their excavator Fanouria Dakoronia have been presented in "Tropis 11, Tropis
IV, Tropis V' (Fig. 1-2)'. The bias might render useful a lapidary review of
what is known about the beginnings of combat at sea. I shall leave aside its
general setting (in wars or piracy between communities such as "states" or
tribes, or in piratical actions by private "entrepreneurs") but for cases where
the sources allow for specification. A clear-cut separation of what to us is
piracy from war is most likely to be anachronistic. 1 agree with Elizabeth
Schofield's definition of piracy as "informal war".
The Kynos paintings date from the last sub-period of the Late Bronze
Age in Greece (LH Ill C1). This time seems to be the setting of an episode in
the Odyssey (16, 355 sq., 471sq.): Penelope's suitors set about to intercept
young Telemachos' ship returning from Pylos, in the strait off Ithaka. Only
Athena's advice saves the young prince from encountering the suitors' ship
which is called "bristling with arms" - i.e. fit for attacking the other ship at
sea. This is a first hint that ship-to-ship combat may have been an everyday
event in Late Mycenaean Greece. The palace archive of Pylos seems to
furnish more, namely recruiting, and sailing orders for rowers, which in the
general atmosphere of military measures in an emergency pervading most
of the archive entries, seem to refer to the manning of warships2. These
sources, however, leave open if fighting at sea was special of Mycenaean
Greece or common all over the Eastern Mediterranean.
In an attempt to find an answer, let us start by reviewing Bronze Age

sources from the Aegean. The earliest representation of warriors in a ship
was found in MH Ill strata of the Kolonna site on Aegina, dating from c. 1600
BC (Fig. 3)? a painted sherd shows warriors with long spears in a longship
(i.e. a low-sided sleek vessel propelled by a considerable number of
oarsmen, leaving little space for cargo), while two such vessels, the right one
apparently carrying a midship cabin (see below), with summarily rendered
unarmed crews decorate a large pithos from Kolonna (Fig. 5)4. More
longships (in the former case two bows only are preserved) are shown on
MH Ill sherds from lolkos in Thessaly (Fig. 4)5, and in graffiti from
HyriaIDramesi on the Euripos, of MH Ill or LH I date where all-purpose squat
OLAF HOCKMANN TROPIS VI
roundships are to be seen as well: Bronze Age Greece knew both types,
which suggests that here ships had been used for some kind of warfare
earlier than elsewheres. Among the longships there seem to be two sub-
types defined by either more or less symmetric outlines of their bows and
sterns (Kolonna: sickle-shaped hulls, Hyria: vertical stem- and sternposts) or
differing ones (lolkos, where the keels at the bows protrude in a shape which
has been likened to the rams of much later Greek warships; the stern bent
upward in a marked curve is conjecture. But even if it should have been
angular there hardly would have been a "pseudo-ram"; see below).
We should not neglect that sleek longships already are known from Early
Bronze Age drawings and models from the Cyclades, a millennium prior to
the Kynos finds7.The fact that Cycladic settlements at that time used to be
fortified, suggests these little 3rdmillennium towns needed protection from
attacks by what may have been raiders in such longships as the Cycladic
sources show.
There, however, is no need to think that employing ships for some kind
of warfare was unique among the antagonists of the Aegean. In 3'*
millennium Egypt's Old Kingdom there existed terms for warlike matters as
"arraying boats for battle"; and groups of reed boats fighting on the Nile are
shown by some reliefsg.This goes beyond what is shown at Kynos where not
flotillas but only individual ships enlocked in long-range combat prior to
boarding are preserved, but rather suggests formal styles of naval warfare.
There might be an Early Bronze Age group of longships, comparable with
the Kolonna and lolkos vase paintings, engraved into a sword blade from
Dorak in NW Turkey'', but it should be treated with due reserve as long as
the authenticity of the "find" cannot be substanciated.
About 1500 BC, a flotilla of Aegean longships was painted onto a wall of
the West House at Akrotiri on Thera, one group of fragments showing what
seems to be a landing operation facing opposition from defenders on shore
(Fig. 6)". There are preserved the ends of three longship hulls, gently turning
up like the bows and sterns of the vessels forming the famous Fleet
processionI2,one of them showing a man holding a long lance standing in a
kind of bow castle, while naked dead bodies and shields are floating in the
sea between the ship and the shore. I take them to be those of defenders
killed by long-range weapons of the sea-borne attackers, toppled into the
sea. This is not the only source for landing battlest3:they seem to have been
common. Even the Iliad (11, 701 sq.) mentions the Achaean Protesilaos being
killed when attempting to enforce a landing at the outset of the Trojan War,
which is likely to fit into a Late Bronze Age settingI4.
The Thera frescoes, of Cretan style, call for attention for more than one
reason. On the one hand they might be taken to substantiate Thucydides'
report (1, 4) that the mythical Cretan king Minos established a thalassocracy
in the Aegean by suppressing Carian pirates who then inhabited the
Cyclades and scavenged on Cretan shipping, which seems to imply sea
combat to have been practised in the Aegean by the mid-2"d millennium.
Minos is said to have forced the former pirates to serve in his fleet (Thuc. I,
8). "Minoan thalassocracy" has been the subject of a symposium at the
Swedish Institute at Athens in 198215,when the majority of the participants
thought a Minoan thalassocracy to be no more than a myth. This, however,
seems to be a matter of definition since there cannot be any doubt that by
c. 1500 BC in the Aegean conditions existed when Cretan "influence", unless
domination or even the presence of Cretan settlers at a number of sites both
on the islands and the west coast of Asia Minor, is ascertained by
archaeological findsq6. Maybe Thucydides himself spoiled the case by
applying the term Thalassocracy, up to date and well-defined at his time, to
a remote past. We are ignorant of his sources.
For meeting our second argument, all Thera boats display a feature
which earlier had characterised the Kolonna vase paintings already, namely
stems and sterns curving up more or less symmetrically in gentle curves.
This feature separates them from Early Bronze Age representations in the
Aegean which had been characterised by markedly different outlines of bow
and stern, one at least being angular1'.
Symmetric sickle-shaped outlines, going back to a particular way of hull

construction, had been common in Egypt from earliest times on1'. I think
their being shown at Kolonna, Thera, and in other Late Bronze Age pictorial
sources from Greecet9denotes some knowledge of Egyptian ships, superior
to earlier local types. At the time of the Thera frescoes this not only pertains
to the general layout of the hulls but also to such details as the richly
ornamented "captain's seats" (ikria)", cabins2', and masts laid down". I
share Morgan's view (1988, 131 sq.) that the evidence of influence from
Egypt is compelling. One even might wonder if the high rounded stern of the
lolkos boat in Theocharis' reconstruction shown in fig. 4 (if correct: it is not
preserved on the extant vase fragments), might not go back to the example
of Egyptian wooden ships imitating papyrus boats23.
There is no reason for considering Egyptian influence unlikely. The Nile

country had been in contact with Crete, and maybe the Aegean, from the 3d
millennium onz4,and by the mid-Pd millennium Crete (Keftiu, Kaphtor)
formed a member of the civilised world of the Eastern Mediterranean known
as far east as at Mari on the Euphrates. At Ugarit, in Lebanon, the artist god
"Kothar-and-Khasis" was believed to be a Kaphtoritez, and tomb paintings
at Thebes, from Thutmoses' Ill reign (1490-1436 BC), show envoys from
Keftiu as delivering tribute to their overlord the Pharaohz6.
This very situation in the earlier part of the Aegean Late Bronze Age
implies general political and naval conditions in Crete, and perhaps all over
the Aegean to have been similar to those in Egypt's sphere of political
interest in the Levant. Here the great pharaoh Thutmoses Ill had extensively
employed Egypt's "navy" for supporting his many campaigns in Palestine,
Lebanon, and Syria to an extent not known earlier2'. It is generally thought
that these "warships", as they are called, only served for logistic purposesz8.
It should not be neglected, however, that fighting on the Nile had

occurred when the founder of Egypt's New Kingdom, Ahmose, conquered
the capital of the preceding Hyksos dynasty of foreign invaders, Awaris, in
the eastern Nile Deltaz9,and in Nubia. By the way, I wonder if the many
"ships' soldiers" mentioned in New Kingdom texts30did not use to fight it out
at sea when encountering hostile vessels. It would be funny enough to
imagine the opponents making for the shore to stage the battle on land.
Such a model becomes even more ironic by the fact that shores fit for
landing are rare on the coasts of Palestine and Lebanon. Perhaps skirmishes
at sea were everyday events at that time already, not worth being recorded
in the Egyptian archives for just this very reason. It should be noticed there
are mentions of Thutmoses' "warships" taking foreign freighters at sea3', so
to speak acts of piracy which imply shipboard fighting.
Elisha Linder (1973; 1981) evaluated the host of written documents from
the El-Amarna period (the century BC) in Egypt. He demonstrated that
the naval activities of Thutmoses Ill on the Levant coast not only tightened
the pharaoh's control of the local vassal kingdoms, in the first line Ugarit and
Gebal (Byblos), but also fostered the growth of sea trade. The Uluburun
wreck, of the end of this period, gives a vivid impression of how precious
ships' cargoes could be during this Golden Age of intense international
contacts and trade within the civilised world of the Eastern Mediterranean3'.
A wreck of Cypriote or Levantine origin from Cape lria in the Argolid, of c.
1200 BC (Lolos et al. 1995) testifies to the range of Cypro-Levantine sea
trade even during the last years of the Golden Age.
Linder no doubt is right in thinking the rich sea trade in the Eastern
Mediterranean raised the appetites of less civilised tribes, "states" (?), or
individuals at the western margin of the Levant-Cyprus zone, who now
started to reap a profit from piratical inroads upon the shipping in the East
(Linder 1981, 38 sq.). We may also follow his thinking that these Westerners
employed warships of longship type, as had been common in the Aegean
for centuries (see above) but as yet was unknown in the east.
The Eastern states reacted by creating navies of their own for protecting
their shipping from the pirates (Linder 1981, 39). They perhaps followed
Western models for their warships. Pictorial sources dating to the last years
of the Golden Age, namely a seal from Ugarit (fig. 7) and a graffiio from
Enkomi in Cyprus (fig. 8)" show vessels with straight keels and angular
stems and sternposts, and sails without a boom like those of the Peoples of
the Sea (fig. 9; see below). At that time Ugarit disposed of a navy of 150
vessels (Linder, loc. cit.), which made it a first-rank sea power. Other
kingdoms of the Levant had smaller navies.
It is hard to believe they were exclusively employed for fighting pirates

from Lycia, and the Aegean. Here, dominance had passed from Crete to
Mainland Greece with its foremost royal seat of Mycenae. The Odyssey
leaves no doubt about piracy having been an "honest" profession; just think
of Odysseus' fancy tale of having lost his ships when raiding EgypP. At any
rate the Eastern Mediterranean navies became professional in fighting at
sea, be it against pirates or each other.
This is the background for the upheavals which, from the late 13mcentury
BC on, put an end to the Golden Age of peaceful trade among the civilised
Late Bronze Age states in the Eastern Mediterranean. Their outset is marked
by the decline of Egypt's navy, Ugarit becoming a vassal of the Hittite empire
instead of Egypt's and, as a consequence of that, the rise of the Hittite
empire to the rank of a sea power which by c. 1200 BC was able to conquer
CypruP.
The latter event immediately precedes, or is connected with, the first

forebodings of the intrusion of hostile foreign seafarers, in Egyptian sources
called Peoples of the Sea and vaguely derived from "the islands of the Great
Green [Sea]", which seems to mean the Aegean [36]. They attacked Cyprus,
and c. 1200 BC were thrice beaten by the Hittite fleet which in the first place
was formed by contingents from Ugarit (Linder 1981, 40). The Hittite king
Shuppiluliumash II boasts of "having burned the enemy fleet in the midst of
the sea"36.This cannot but mean that the Hittite navy then was expert in
fighting at sea, applying tactics not paralleled until late Hellenistic times.
Instead of Thucydides' event in 664 BC, Shuppiluliumash's victories in
Cypriote waters deserve being called the first sea battles recorded in history.
They failed to put an end to the incursions of the Sea Peoples, who even
intensified their activities. A group of migrant tribes was instrumental in the
collapse of the Hittite empire, then laid waste the Levantine kingdoms, and
finally approached the Nile Delta on land, while another sea-borne group
entered the Delta with a fleet. The pharaoh Ramses Ill c. 1186 BC crushed
both invading forces in land and sea battles, both shown in his temple at
Medinet Habu (Fig. 9).
The sea battle in particular (if the surprise attack by Egyptian warships
full of archers, supported by archers on land, on the invaders' fleet in one of
the Nile branches in the Delta may be called thus), has been the subject of
many studies3', among which one (Wachsmann 1995) deserves special
attention for pointing out close similarities between the Sea Peoples' ships
and those shown on the Late Mycenaean kraters from Kynos3'. An origin of
some of the Sea Peoples in Southeastern Europe had earlier been claimed
on antiquarian and archaeological grounds, in the first place of the
Philistines whose pottery goes back to Late Mycenaean prototype^^^. This
closely connects the Medinet Habu reliefs of Egyptian warships attacking the
Sea Peoples' fleet, with the Kynos vase paintings of ships engaged in
combat at sea.
For summing up, there can be no doubt either about the Late
Mycenaean Greeks having been familiar with shipboard combat, or of their
predilection for piracy which in the Aegean can be traced back to the Middle
if not to the Early Bronze Age. It is likely enough that longships specialised
for sea-borne war, formal or informal, first became known in the Egypt-
Levant-Cyprus zone of the Eastern Mediterranean where close ties had
emerged by peaceful cooperation both commercial and political, by inroads
from the West. There is direct testimony of this in a letter of Ugarit's last king
Hammurapi to the Hittite viceroy of Cyprus (?), in which Hammurapi
complains about being attacked at home while his army is in the Hittite
heartland in Asia Minor and his fleet in Lycia, in the far WesP. We may
suppose the Hittite king had ordered it there for intercepting the Westerners
in their home waters. The letter being among the last entries of Ugarit's royal
archive, immediately preceding Ugarit's final destruction by the Sea
Peoples, clearly enough demonstrates the Hittite operation in Lycia was

abortive.
What I wanted to point out is twofold. On the one hand only Aegean
tribes had known longships for warlike purposes in the Early and Middle
Bronze Ages. On the other, at the time of the Kynos paintings shipboard
combat had become common all over the Eastern Mediterranean. The
Kynos sea-fighters, far from being unique, were just up to the international
standard of their age.*
Olaf Hockmann
Taunusstr. 39
D-55118 Mainz, Germany
NOTES
1.My fig. 1-2, technically modified, go back to Dakoronia 1987 (1990), 122 fig. 1.3; 1989 (1995),
147 fig. 1-2. Cf. also: Dakoronia, forthcoming (more finds in recent years). Wachsmann 1995,
+
26 sq. fig.; 28 fig.
2.Ventris & Chadwick 1956,185sq. Linder 1973,321; 1981,41sq. Gray 1974, G 54. Wachsmann
1995,23 sq.
3.Hiller 1984, 28 fig. 1.
4.lbid. 29 fig. 2.
5.Theocharis 1958. Gray 1974, G 16 no. B 22; G 43 fig. 8 c. Laffineur 1984,138 fig. 8. Hdckmann
1985,42 fig. 17. lmmerwahr 1985,86 sq., fig. 1. Basch 1987,92 fig. 191.
6.Blegen 1949, pl. 7,6. Casson 1971, 40 fig. 25. 32. Gray 1974, G 17 no. C 6; pl. G I, b.
Hockrnann 198542 fig. 16. Basch 1987.92 sq., fig. 191; 144 sq.fig. 300-302. Morgan 1988,
pl. 165.
7.Renfrew 1967,5; 18 nos. 12-14; pl. 1,12; 3. Casson 1971, fig. 22. Gray 1974, G 35 fig. 3; pl. G
I,a. Hockmann 1985,38 sq. fig. 4.6-7. Basch 1987,78 sq. fig. 152 sq.- Morgan 1988,135 fig.
86. Wachsmann 1995, 12 sq. fig.+
8.Jones 1988, 261 no. 14. "To sink enemy's boats": loc. ck 223 no. 81.
9.Landstrcim 97 fig. 304.
10.Gray 1974, G 29 no. AA 5; 39 fig. 5. Hockmann 1985,38 fig. 5. Basch 1987,90 sq. fig. 189-
190.
ll.S. Marinatos 1974, pl. G XV.- Warren 1979, 124; colour pl. Aa.- Hockmann 1985, 41 fig. 15.
Morgan 1988, 159 sq.; pl. 2.3.42. 139. 144. 189. Wachsrnann 1995, 20 + fig. Marinatos and
Morgan think the scene show the aftermath of ship-to-ship combat. Marinatos even calls a
floating (!) object a grapnel, the earliest unambiguous representation of which is to be seen
at Medinet Habu (note 37), and bases his interpretation on a group of marching warriors, but
these Aegean warriors on land rather make Warren (1979, 125 sq.), Wachsmann, and me
think of a landing battle being shown.
12.S. Marinatos 1974. Casson 1975. Tilley & Johnstone 1976. Brown 1978. Warren 1979.
Wachsmann 1980. Laffineur 1984, 134 sq. (pointing out Mycenaean elements in ships'
decoration, mostly interpreted as a Minoan-basedAegean 'koine' unless rejected [L. Morgan]
in the discussion [p. 138 sq.]). Hijckmann 1985, 41 fig. 12. Toby 1986. Basch 1987, 119 sq.
fig. 232 sq. Morgan 1988, 121 sq., fig. 67; 143 sq.; pl. 8 sq.;103; 159 sq.; 168 sq. Gillmer 1989
(1995).- Raban 1989.- Wachsmann 1995, 15 sq. + fig.
13.Gray 1974, G 17 no. C 5; G 128 sq. Warren 1979, 121 sq.; 125 sq.; fig. 5. Hockmann 1980,
E 280 sq.; fig. 66. Morgan 1988, 150 sq.; 153; pl. 191-4. References to naval raids in the Iliad:
V,640 sq.; IX,328 sq.; XIV,250 sq.; XV,24 sq.; XV111,207 sq.; Odyssey: 9,39 sq.; 14,85 sq. 199
sq.; 16,427 sq.; 17,286 sq.; 21,15 sq.
14.Cf. the perfect correspondence of many of the Iliad's descriptions of weaponry and tactics
with what is shown in Mycenaean pictorial sources (e.g. Triimpy 1950, 51 sq. Gray 1974, G
125. Hockmann 1980, E 313 sq.; E 317 sq.).
15 Hagg & Marinatos (ed.) 1984.
16.Branigan 1981; 1984. Barber 1984. Benzi 1984. Coldstream & Huxley 1984. Davis 1984.
Hood 1984. Laviosa 1984. N. Marinatos 1984. Schiering 1984. Schofield 1984.
17.Renfrew, loc. cit (note 7). Gray 1974, G 14 sq. no. A 1-16; G 41 fig. 6; G 42 sq. fig. 7; 8 b.
Hockmann 1985,39 fig. 10.- Basch 1987,83 sq. fig. 170-171, 175; 98 sq. fig. B 1 sq., C 1 sq.,
D 1 sq.- Melas 1988. Morgan 1988, 135 fig. 86-7; 137 fig. 90. Wachsmann 1995, 18 fig.
18.Casson 1971 fig. 5. 7-9. 16-19. 57. Gray 1974, G 32 fig. 2; G 37 fig. 4; G 49 fig. 12 a-b.
Landstrom 1974, fig. 10. 14. 18. 22 etc.; 83 ff.; 95 ff. etc.- Basch 1987, fig. 70-71. 86. 94 sq.
110 sq. 117. 227-228. 250. 252. Morgan 1988, 125 sq. fig. 74-5; 139 sq. fig. 94-5; pl. 195.
Wachsmann 1995, 11 fig.; 23 fig.
19.Gray 1974, G 35 fig. 3 j; G 41 fig. 6 h.i.k.1.n-q; G 42 sq. fig. 7; 8 e; G 45 fig. 9 a-j. Basch 1987,
95 sq. fig. A 4 sq.; 102 fig. C 15; 104 sq. fig. F 7-8, F 12 sq.; 116 fig. 224; 148 fig. 311. Morgan
1988, 122 fig. 67-8; 125 fig. 72-3; 130 sq. fig. 79-81; 136 fig. 89; 142 fig. 88-9.
20.For Thera see S. Marinatos 1974, G 148; pl. G XIII; G XVI, a-b; G XVII. Shaw 1980 (painted
ikria at Mycenae ?). Tzamtzis 1985 (1989).- Basch 1987, 118 fig. 229. Morgan 1988, 137 sq.;
143; pl. 175 sq. Egypt: Landstrom 1974, fig. 317. 320. 322-4. 326-332. 336-7. 339. 351. 354.
375. 383-4. In Egypt such ornate stands were placed in both ends of the ships whereas the
Thera vessels have them in the stern only. Plain stands are shown in the Thera "landing
battle" (S. Marinatos 1974, G 150; our fig. 6), and in many Egyptian representations
(Landstrom 1974, 111; fig. 319. 334. 338. 343. 345. 348. 351. 372. 376-7. 408).
21.Thera: S. Marinatos 1974, G 148; pl. G XIV. Egypt: Landstrom 1974, fig. 166-9.205.207.213.
276.383-4. 287-8. 292-3. 357-8. 361-2. 365. 369-371.
22.Thera: S. Marinatos 1974, G 149. Egypt: Landstrom 1974, fig. 137. 143-4. 187. 190. 193. 196.
199. 249.252. Morgan 1988, 139 fig. 94.
23.For such sterns on real papyrus rafts, where they are formed by the ends of reed bundles
bent forward: Landstrom 1974, fig. 30-47. 78. 80. 298. 300. 303. 305. Similar, on wooden
ships: ibid. 23 sq.; 56 sq. + fig. 166 sq.; 90 sq. + fig. 274 sq.
24.Egyptian objects in Crete and on Asia Minor's Aegean coast: Branigan 1970, 181 sq.
Renfrew 1972, 446 sq. Coldstream & Huxley 1987, 137 sq. Hockmann 1987, 61. 75. 93 sq.
For Cretan trade with Egypt note Ipuwer's lament after the collapse of Egypt's Old Kingdom
that "no oil comes from Keftiu any more" (Pritchard 1969, 441. Hockmann 1987, 61).
25.Gordon 1966, 23; 49 n. 10; 58 n. 24.
26.Strom 1984, 192 sq., fig. 1-3 (suggesting LM I B Crete actually had asked Thutmoses Ill's
support for overcoming Mycenaean competition in the trade with the Levantine kingdoms,
then Egypt's vassals or allies).- Wachsmann 1987, passim.
27.Save-Soderbergh 1946.- Linder 1973, 317.
28.Save-Soderbergh 1946, 31 sq.; 36; 42. Landstrdm 1974, 109.- Linder 1973, 318.- Urk. 372
sq. no. 292 f; 294 A.
29.Urk. IV, 3 sq. Save-Soderbergh 1946, 3 sq.- Jones 1995, 63.
30.Save-Soderberg 1946,83 sq. Landstrom 1974, 109.- Jones 1988,58 no. 40; 125 no. 9; 1995,
64. "Warships, fighting ships": 1988, no. 103, 249, 251, p. 111 no. 4, 87 no. 165, 129 sq. no.
1, 5, 13, 15, 16, 36, 79, 234 sq. no. 19-20.
31. Save-Soderbergh 1946,34 sq.; 42. Urk. 202 no. 3 d. Ugaritic references to ships seized and
a naval campaign (?): loc. cit. 60 sq.
32.Bass et al. 1984. Wachsmann 1987, 129 sq. Pulak 1995.

33. Seal from Ugarit: Basch 1987, 70 fig. 131. Wachsmann 1995, 24 fig. Graffito from Enkomi,
Cyprus: Casson 1971, fig. 27.- Basch 148 fig. 312.
34. Od. 14,199 sq. (for other references to raids cf. note 13). Thucydides (1,5) was aware of the
rdle of piracy in Greek past.
35.0tten 1963,3 sq., 20 sq.; 1976. Giiterbock 1967.- Gray 1974, G 123; G 132. Cornelius 1973,
276; XVI. Johnstone 1980, 79 sq. tinder 1973, 319; 1981, 39.- Gray 1974, G 132. Lehmann
1985,26; 28 sq.
36.Wachsmann 1987,93 sq.; 98 sq.; 111 sq.
37.Casson 1971,36 sq.; 41; fig. 61. Gray 1974, G 83; pl. with p. G 88. Landstrom 1974, 111 sq.,
fig. 345-8. Hockmann 1985,45 fig. 21. Wachsrnann 1981; 1982; 1995,28 sq. Basch 1987,68
sq. fig. 123 sq. Morgan 1988, 133 fig. 84. Raban 1989; 1989 (1995). de Boer 1991. Jones
1995,59 sq. Wachsmann 1995,29 sq. fig. +
38.Years ago i had come to identical conclusions in a paper still in press (Hockmann,
forthcoming).
39.Harding 1976. Lehmann 1985, 45. Briquel 1986.
40.Gray 1974, G 123. Lehmann 1985,28 sq.
*My sincere thanks are due to H. Konen (Marburg) for supplying Egyptological literature not at
my disposal, and helpful suggestions.
REFERENCES
Barber, C. 1984: The Status of Phylakopi in Creto-Cycladic relations. In: Hagg & Marinatos (ed.)
1984, 179-182.
Basch, L. 1987: Le Musee imaginaire de la marine antique (Athenes).
Bass, G.F., Frey, D.A. & Pulak, C. 1984: A Late Bronze Age shipwreck at Kas, Turkey. IJNA 13,
271-279.
Benzi, M. 1984: Evidence of a Middle Minoan settlement on the acropolis at lalysos (Mt.
Philerimos). In: Hagg & Marinatos (ed.) 1984, 93-105.
Blegen, C.W. 1949: Hyria. Hesperia Suppl. 8, 39-42.
Branigan, K. 1970: Minoan Stone Vases (London).
1981: Minoan colonialism. BSA 76, 24-33.
1984: Minoan community colonies in the Aegean? In: Hagg & Marinatos (ed.) 1984,49-
52.
Briquel, D. 1986: Le fonti dei Popoli del mare. In: Cispadana e letteratura antica. Congr. lmola
1986, 1-25 (Bologna 1987).
Broodbank, C. 1989: The Longboat and society in the Cyclades in the Keros-Syros culture. AJA
93, 319-337.
Casson, L. 1975: Bronze Age ships. The evidence of the Thera wall paintings. IJNA 4, 3-10.
Coldstream, J.N. & Huxley, G.L. 1984: The Minoans of Kythera. In Hagg & Marinatos (ed.) 1984,
107-112.
1987: Die Minoer auf Kythera. In: H.G. Buchholz, ~gaischeBronzezeit, 137-148
(Darmstadt).
Dakoronia, F. 1987 (1990): War-ships on sherds of LH lllC kraters from Kynos.
Tropis 2 , 117-122.
1989 (1995): War-ships on sherds of LH Ill kraters from Kynos? Tropis 3, 147-148.
1993 (1999): Representations of sea battles on Mycenean sherds from Kynos,
Tropis 5,119 128
1991 (1996): Kynos ... Fleet. Tropis 4, 159-171.
Davis, J.L. 1984: Cultural innovation and the Minoan thalassocracy at Ayia Irini, Keos. In: Hagg
& Marinatos (ed.) 1984, 159-166.
De Boer, J. 1991: A double figure-headed boat type in the Eastern Mediterranean and Central
Europe during the Late Bronze Ages. In: Thracia Pontica 4, 43-50 (Sofia).
Gillmer, Th. 1989 (1995): Further identifications of functional parts on Thera fresco ships. Tropis
3, 177-192.
Gordon, C.H. 1966: Ugarit and Minoan Crete (New York).
Gray, D. 1974: Seewesen. Archaeologia Homerica, chapter G (Gottingen).
Guterbock, H.G. 1967: The Hittite conquest of Cyprus reconsidered. JNES 26, 73-81.
Hagg, R. & Marinatos, N. (ed.) 1984: The Minoan Thalassocracy. Myth and Reality. Proc. Third
Internat. Sympos. at the Swedish Inst. Athens, 1982 = Skrifter Svenska Inst. Athens, 4"
XXXll (Stockholm 1984).
Harding, A.F. 1976: Illyrians, Italians and Mycenaeans. Trans-Adriatic contacts during the Late
Bronze Age. Iliria 4, 157-162.
Hiller, S. 1984: Pax Minoica versus Minoan Thalassocracy: Military aspects of Minoan culture.
In: Hagg & Marinatos (ed.) 1984, 27-31.
Hockmann, 0. 1980: Lanze und Speer. In: H.G. Buchholz (ed.), Archaeologia Homerica:
Kriegswesen Teil2, Abschn. X, E 275-E 319 (Gottingen).
1985: Antike Seefahrt (Munchen).
1987: Fruhbronzezeitliche Kulturbeziehungen im Mittelmeergebiet unter besonderer
Berucksichtigung der Kykladen. In: H.G. Buchholz, ~gaischeBronzezeit, 53-120
(Darmstadt).
Forthcoming: Graffiti of ships from Razlog, Bulgaria. In: [Working title] Studies in
Memory of James Hewey Gaul (Sofia).
Hood, S. 1984: A Minoan Empire in the Aegean in the 1 6 and ~ ~1 5 Centuries
~ ~ BC? In: Hagg &
Marinatos (ed.) 1984, 33-37.
Immerwahr, S.A. 1985: Some pictorial fragments from lolkos in the Volos museum. Arch.Eph.
1985 (1987) 85-94.
Jones, D. 1988: A Glossary of Ancient Egyptian Nautical Titles and Terms (London - New York).
1995: Boats. Egyptian Bookshetf, British Museum (London).
Laffineur, R. 1984: Mycenaeans at Thera: further evidence? In: Hagg & Marinatos (ed.) 1984,
133-138.
Landstrom, B. 1974: Die Schiffe der Pharaonen (Munchen - Giitersloh - Wien).
Laviosa, C. 1984: The Minoan thalassocracy, lasos and the Carian coast. In: Hagg & Marinatos
(ed.) 1984, 183-185.
Lehmann, G.A. 1985: Die mykenisch-fruhgriechische Welt und der ostliche Mittelmeerraum.
Rhein.-Westf$il. Akad., Vortrage G 276 (Opladen).
Linder, E. 1973: Naval warfare in the El-AmarnaAge. In: D. Blackman (ed.), Marine Archaeology.
Colston Papers 23, 317-322 (London).
1981: Ugarit: a Canaanite thalassocracy. In: G.D. Young (ed.), Ugaritin Retrospect, 31-
42 (Winona Lake).
Lolos, G., Pennas, Ch. & Vichos, G. 1995: Der Schiffsfund von Kap lria (Golf von Argos). In: In
Poseidons Reich. Archaologie unter Wasser. Antike Welt, Sonderheft, 59-62.
Marinatos, N. 1984: Minoan threskeiocracy at Thera. In: Hagg & Marinatos (ed.) 1984, 167-178.
Marinatos, Sp. 1974: Das Schiffsfresko von Akrotiri, Thera. In: Gray 1974, G 141-G 151.
Melas, M. 1988: Minoans overseas. Alternative models of interpretation.Aegaeum 2, 47-70.
Morgan Brown, L. 1978: The Ship Procession in the Miniature Fresco. In: C. Doumas (ed.),
Thera and the Aegean World 1, 629-641 (London).
Morgan, L. 1988: The Miniature Wall Paintings of Thera. A Study in Aegean Culture and
Iconography (Cambridge - New York etc.).
Otten, H. 1963: Neue Quellen zum Ausklang des hethitischen Reiches. MDOG 94, 1-23.
1976: Zum Ende des Hethiterreichs aufgrund der Bogazkoy-Texte. Jahresber.1nst.Vor-

u.Fruhgeschichte Frankfurt 1976, 27-29.
Pritchard, J.B. 1969: Ancient Near Eastern Texts (3d ed., London).
Pulak, C. 1995: Das Schiffswrack von Uluburun. In: In Poseidons Reich. Archaologie unter
Wasser. Antike Welt, Sonderheft, 43-58.
Raban, A. 1989: The Medinet Habu ships. Another interpretation. IJNA 18, 163-171.
1989 (1995): The Sea People and Thera Ships. Tropis 5, 353-366.
Renfrew, C. 1967: Cycladic metallurgy and the Aegean Early Bronze Age. AJA 73, 1-20.
1972: The Emergence of Civilisation (London).
Save-Soderbergh, T. 1946: The Navy of the Eighteenth Egyptian Dynasty. Uppsala Univ.
Arsskrift 1946:6 (Uppsala - Leipzig).
Schiering, W. 1984: The Connections between the oldest settlement at Miletus and Crete. In:
Hagg & Marinatos (ed.) 1984, 187-189.
Schofield, E. 1984: Coming to terms with Minoan colonists. In: Hagg & Marinatos (ed.) 1984,
45-48.
Shaw, M.C. 1980: Painted "ikria" at Mycenae? AJA 84, 167-179.
Strom, 1. 1984: Aspects of Minoan foreign relations, LM I-LM II. In: Hagg & Marinatos (ed.) 1984,
191-194.
Theochares, D.R. 1958: lolkos, Whence Sailed the Argonauts. Archaeology 11, 13-18.
Tilley, A. & Johnstone, P. 1976: A Minoan naval triumph? IJNA 5, 285-292.
Toby, A.S. 1986: World's first warships: tubs or ocean greyhounds? IJNA 15, 339-346.
Tzamtzis, A.I. 1985 (1989): lkria on Minoan seals. Tropis 1, 275 sq.
Urk.: Urkunden der 18. Dynastic. ~bersefzungzu den Heften 5-16 (ed. E. Blumenthal et a/.)
(Berlin 1984).
Urk. IV: Urkunden des agyptischen Alterturns, Abt. IV,1 (ed. K. Sethe) (Leipzig 1914).
Ventris, M. & Chadwick, J. 1956: Documents in Mycenaean Greek (Cambridge).
Wachsmann, S. 1980: The Thera waterborne procession reconsidered. IJNA 9, 287-295.
1981: The Ships of the Sea Peoples. IJNA 10, 187-220.
1982: The Ships of the Sea Peoples: additional notes. IJNA 11, 297-304.
1987: Aegeans in the Theban Tombs. Orientalia Lovanensia Analecta 20 (Leuven).
1995: Paddled and oared ships before the Iron Age. In: The Age of the Galley, 10-35
(London).
Warren, P. 1979: The Miniature fresco from the West House at Akrotiri, Thera, and its Aegean
setting. JHS 99, 115-129.
1984: The Place of Crete in the Thalassocracy of Minos. In: Hagg & Marinatos (ed.)
1984, 39-44.
Aegean longships of the Middle and Late Bronze Ages
112 Kynos Livanaton (LH Ill Cl).

315 Kolonna, Aegina (MH Ill).
4 lolkos (MH IIIILH I).
6 Akrotiri, Thera: fresco (LM I B).
7 Ugarit: seal (c. 1200 BC).
8 Enkomi: graffito (12mcent. BC).
9 Medinet Habu: relief of Sea Peoples' ship (c. 1186 BC).
(7 after L. Basch; 819 after L. Casson.)
Aegean longships of the Middle and late Bronze Ages
1-2 Kynos Livanaton (LH Ill CI) 3.5 Kolonna, Aegina (MH Ill).-
4 lolkos (MH IIIILH I).-6Akrotiri, Thera: fresco LM I B).-
7 Ugarit: seal (c. 1200 BC).-8 Enkorni: graffito 12th cent. BC.-
9 Medinet Habu: relief of Sea Peoples' ship (c. 1186 BC). 7 after
L. Basch.- 8.9 after L Casson.
UNDERWATER FINDS OF SHIP AND BOAT MODELS'
Objects from the seabed can be classified according to their find context.
On the one hand there are finds recovered from shipwrecks and underwater
structures, while on the other there are objects without such contexts, which
are usually more difficult to interpret and to date. Of the latter, a considerable
number are chance finds, making it impossible to further investigate their
place of discovery. We can imagine them having been lost by accident or
having been consciously thrown into the sea from ships as an emergency
measure or as rubbish.
Curiously, a small number of such objects have the shape of ships or

boats of terracotta or bronze, and seem to be of cultic character. They
suggest that religious ceremonies may have been conducted on board ships
and boats, which involved the dedication of objects to marine deities. Ten
possible examples from the Mediterranean are known to me at present, and
are listed in the following short catalogue.
Nos. 112:
Two terracotta models, one from the sea outside Amathus in Cyprus, the
other probably from the same location2.The city is known for a number of
terracotta models found mostly in Archaic graves3,but the models discussed
here date from the Late Cypriote period, i. e. 1600-1050 BC.
No. 3:
A terracotta model from the area of Lake SirbonislSabhat al-Bardawil on
the northern coast of Sinai4. It seems to represent a logboat and has been
dated to the 13m112m centuries BC.
Nos. 415:
Two terracotta models found off the Lebanese coast near Tyre. They
have been identified as Phoenician and probably date from the beginning of
the first millennium BC5.
No. 6:
A terracotta model found at Gytheion, the naval harbour of Sparta6.The
dating of this well-known warship model is disputed; it is commonly believed
to belong to the period around the late first century BC or the early first cen-
tury AD, but an Archaic date has also been suggested7.
KARIN HORNIG TROPlS VI
No. 7:
A bronze model of a warship found during investigations of the harbour
at Knidos8,probably of Hellenistic date.
No. 8:
A fragment of a Roman terracotta lamp discovered in the harbour area of
Fos-sur-Mer".
No. 9:
Another terracotta lamp found in the sea near PuteolilPouuoli, of
Knidian production and dating from the second half of the first century AD''.
Its top is decorated with a representation of lsis and Sarapis, and its tabella
ansata carries the inscription EUPLOIA, the epithet of Knidian Aphrodite.
No. 10:
A bronze lamp shaped like a warship from the third-century AD
shipwreck of Aghia Ghalini off the southern coast of Crete; it is the only
example found in a wreck1'.
Although these models all belong to different periods and backgrounds,

it is possible to consider them, for the present purpose, as one group on
account of their shared significance. This approach appears justified on the
basis that one can frequently observe different religions sharing similar kinds
of rituals.
An example within the framework of the present topic is, for instance, the
custom of placing ship and boat models in graves, which is common to
Mesopotamia, Egypt, Cyprus, Crete and other regions. The background for
this practice surely differs, but the act itself is the same. This holds true also
for the postulated practice of dedicating ship and boat models at sea.
A small number of written sources can potentially provide some

indication as to the background of this practice.
The Roman author Apuleius (2ndcentury AD)'* can be cited as a

reference to ships sunk for cultic reasons. He mentions a festival of lsis at
Kechreai held at the beginning of the seafaring season, in the course of
which the priests fill a full-scale ship with offerings and set it adrift on the sea.
Sunken ship models occur in Apollodoros' Bibliotheka (la or 2""century

AD)13,which records the story of Kinyras, King of Cyprus, who pledged to
UNDERWATER FINDS OF SHIP AND BOAT MODELS
provide fifty ships to aid the conquest of Troy. Only one of these ships,
however, was a real, large-scale vessel, while the others were made of clay
with clay figures inside and were put out onto sea. Perhaps it is not too far-
fetched to see in this account traces of an actual magical act performed by
the king in order to ensure the success of the Greek expedition.
If we look again at the ship models listed above, we notice that they can
be tentatively divided into two groups, ignoring for the present any problems
with dating. One group of five objects from the southeastern Mediterranean
belongs to the second or the beginning of the first millennium BG (nos. 1 -
-
5), another group of five models (nos. 6 10, if we include those of Aghia
Ghalini and Gytheion) to the Hellenistic period and the time of the Roman
-
Empire. Three of these later models are lamps (nos. 8 10). The ship model
from the Aghia Ghalini wreck (no. lo), however, may not really fit in here, as
it could easily have been carried on the ship for some purpose other than
dedication at sea.
Now we turn to the deities possibly associated with these ship and boat
models.
The examples from Tyre (no. 415) and Sabhat al-Bardawil (no. 3) may be
connected with the worship of the Syrian god Baal SapunalZaphon14,known
in the Greek world as Zeus Kasios. A ship belonging to this god is mentioned
already in a papyrus of the lgmdynasty (12mcentury BC) from Memphis1'.
The historian Prokopios (ca. 500-560 AD) reports that, in his Greek guise, the
god received a ship made of stone as a dedication from a merchant on the
island of Kerkyral"; votive anchors with inscribed dedications to him were
found at the Hispanic coast near Cape Polos". His main cult center was
Mons CasiusIGebel al-Aqra in northern Syria, with a subsidiary cult at Mons
CasiusIRas Qasrun (Katib al-Gals) on the northern border of Lake
SirbonisISabhat el-Bardawil, where the logboat model is said to has been
found (no. 3).
One of the female protectresses of seafarers is Aphrodite Euploia, who is

known to have had a sanctuary at Knidos in the Hellenistic period. It is
noteworthy that of the second, later group of ship and boat models one
example was made at Knidos (no. 9) and a second, bearing the goddess'
epithet, was found in the harbour of this city (no. 7).
For the other models one may consider a link with lsis (Pelagia or
Pharia), who - as mentioned above - is known to have received full-scale
ships as dedications.
KARIN HORNIG TROPIS VI
Although we cannot be certain that the same explanation holds true for
all of the examples, it nevertheless seems likely that most of the models were
used in some sort of cultic activity on board a vessel. Concerning the
question of how such rituals were conducted, G. Kapitan has pointed out the
equipment that would have been required, such as louteria, offering cups
and altars, which have been mostly discovered in shipwrecksI8.Perhaps the
above-mentioned ship and boat models at least partly constitute the remains
of offerings given in the course of such ceremonies.
In addition to these models, there are also other kinds of objects that can
be interpreted in a similar way. One example is from a written source
recording an offering thrown into, and later retrieved from, the sea: a golden
tripod, said to have been dedicated by Helen on her return from Troy, and
found later by fishermen from Kos, Miletos or other regions. The story is
refered to in several versions by Diodorus Siculus (1" century BC), Valerius
Maximus (ISthalf of the 1" century AD), Plutarchos (ca. 46-ca. 120 AD),
Diogenes Laertios (3'dcentury AD), and otherstg.
As for archaeological finds, a bronze statuette of a Syrian god may

provide a parallel case, as it was discovered in the sea between Selinus and
Sciacca on the south coast of Sicilyz0.It probably dates from the beginning
of the first millennium BC.
The small number of ship and boat models I have discussed here is, of
course, a very limited base for interpretation, but appears perhaps more
significant if one takes into consideration the slim chances of ever finding
such objects.
Much work has been done in the past on the types of these models, on
the use of ship and boat models as lamps, and on marine deities. It would
also be of interest to look further into possible ethnological parallelsz1.There
are thus more aspects to the subject than it has been possible to cover this
article. My intention has been to look for a.general explanation of the
occurrence of ship and boat models in underwater contexts. Even if the
interpretation given here may not necessarily apply to all the objects
discussed, it should be kept in mind as a possible explanation for any future
finds.
Karin Hornig
Wasserstr. 4
79098 Freiburg
Germany
UNDERWATER FINDS OF SHIP AND BOAT MODELS
ABBREVIATIONS
Gottlicher, Modelle A. Gottlicher, Materialien fur ein Corpus der Schiffsmodelle im Altertum
(Mainz 1978)
MlMA L. Basch, Le musee imaginaire de la marine antique (Athens 1987)
Parker, Wrecks A.J. Parker, Ancient Shipwrecks in the Mediterranean & the Roman Provinces,
BAR International Series 580 (Oxford 1992)
SIMA Studies in Mediterranean Archaeology
Wachsmuth D. Wachsmuth, Pompimos ho daimon. Untersuchungen zu den antiken
Sakralhandlungen bei Seereisen (Berlin 1967)
NOTES
1.1 would like to thank Miss Alexandra Villing, Mr. Stuart Rae, Mr. Michael Wedde and Mr.
Gerhard Kapitan for helpful suggestions and support in the creation of the English version of
this contribution.
2.No. 1: Limassol, private Collection of Nicos Kirzis, L. 45cm;
K. Westerberg, Cypriote Ships from The Bronze Age to c. 500 BC, SIMA Pocketbook 12
(Gothenburg 1983) 14f. no. 8, 82 fig. 8;
No. 2: Limassol, private collection of Phr. Nicolaides, L. 26cm;
K. Westerberg, op. cit. 16 no. 11, 85 fig. 11; MlMA 257 fig. 554.
3.K. Westerberg, op. cit. nos. 20. 24. 31-40. 42f. 47f. 50f.
4.No. 3: Haifa, National Maritime Mus., inv. no. not given, L. 36cm;
MlMA 56 fig. 92.
5.No. 4: Location unknown, L. 28cm;
MlMA 305f. fig. 645;
No. 5: Location unknown, L. 28cm;
MlMA 305.307 fig. 646.
6.No. 6: Sparta, Mus. Arch., inv. no. 5712, L. 58cm;
MlMA 428. 432ff., fig. 936-943; A. Delivorrias (ed.), Greece and the Sea, Exhibition Cat.
Amsterdam (Athens 1987) 22%. no. 126. This model has already been interpreted as a votive
offering by 0 . Hockmann, Antike Seefahrt (Munich 1985) 158 and A. Delivorrias, op. cit. 229
7.0. Hockmann, op. cit. 99; id. Some thoughts on the Greek Pentekonter in: Tropis Ill(Athens
1995) 215 note 20.
In J.S. Morrison , J.F. Coates, Greek and Roman Oared Warships (Oxford 1996) 238 No. 37
the later date is given preference
8.No. 7: Location unknown, L. unknown;
P.F. Johnston, Ship and Boat Models in Ancient Greece (Annapolis 1985) 122 no. Hell. 37.
9.No. 8: Istres, Mus. du Vieil Istres, inv. no. not given, L not given;
C. Beurdeley, L 'Archeologie sous-marine. L'Odyssee des tresors (Paris 1991) 112 fig. 90.
Because of the circumstances of excavation it is not entirely clear, whether this object was
found in a place, which was in fact under water in ancient times.
10.No. 9: London, British Mus., inv. no. GR 1862.4-14.1 (lamps Q 2722). L. 63cm;
D.M. Bailey, A Catalogue of Lamps in the British Museum 3, Roman provincial lamps, London
(London 1988) 33%. fig. 20, 28, 138, 151 pl. 80; Varen, Vechten en Verdienen. Scheepvaartin
de Oudheid. Exhibition Cat. Amsterdam (Amsterdam 1996) 43 fig. 75, 48 Cat. NO.48.
11.No. 10: Rethymnon, Arch. Mus., inv. no. not given, L. ca. 28cm;
Gottlicher, Modelle 88 No. 534 pl. 42, 534; MlMA 452 fig. 996; for the wreck see Parker,
Wrecks 62 No. 68.
KARIN HORNIG TROPIS VI
12.Apuleius X1,5,5; X1,16,4-8.

13.Apollodoros' Bibliotheka, Epitome 3, 9; this version is recorded also by Eustathios 827, 38ff.
(M. van der Valk, Eustathii Commentarii 3, Leiden 1979, 139) and schol. T (H. Erbse, Scholia
Graeca in Homeri lliadem 4, Berlin 1974, 126f. No. 20b).
14.Wachsmuth 398. note 1819-1825; E. Lipinski (ed.), Dictionnaire de la Civilisation
phenicienne et punique (Turnhout 1992) 60f. S.V. Baal Saphon (Bonnet); I. Cornelius, The
Iconography of the Canaanite Gods Reshef and Ba'al, Orbis Biblicus et Orientalis 140
(Gtittingen 1994) 151f. no. BR 11 note 7.
15.Papyrus Sallier IV verso 1.6; W. Helck, Ein lndiz fruher Handelsfahrten syrischer Kaufleute,
Ugarit-Forschungen2, 1970, 35f.; for further references see the preceding note.
16.Prokopios, de be110 gothico IV, 22.
17.Wachsmuth, 396 note 1825.
18.G. Kapitan, Louteria from the sea, IJNA 8, 1979, 97-120; id., Archaeological evidence for
rituals and customs on ancient ships in: Tropis I(Athens 1989) 147-162; 1. Radic, Three more
louteria finds in the eastern Adriatic, IJNA 20, 1991, 155-160. New finds of louteria were made
in the Sicilian shipwreck "Ognina 4", see H.G. Martin, Ognina 4, Vorlaufiger Grabungsbericht,
DEGUWA-Rundbrief 10, 1995, 21f. Small clay altars and offering vessels from the Gela
shipwreck dated to the sixth century BC were mentioned by R. Panvini during a 1995
convention, the lomRassegna di Archeologia Subaquea held at GiardiniINaxos.
19.Plutarchos, Solon 4; Valerius Maximus, IV, 1, 7, for the further mentioned ancient authors see
B. Snell, Leben und Meinungen der sieben Weisen (Munich 1938) 108-113.
PO.Palermo, Mus. Arch. Regionale, inv. no. 3676, H. 36.2cm;
S. Moscati (ed.), The Phoenicians, Exhibition Cat. (Venice 1988) 48,422,424,655 no. 423; C.
Bonnet, Melqart. Cultes et mythes de I'HBracles Tyrien en MBditerranBe, Studia Phoenicia 8
(Leuven 1988) 266f. fig. 21. A connection of the statuette with the worship of Baal Sapuna is
likely, regardless of the actual find context and date.
21.Wachsmuth 130 note 214; 450 note 2142 for instance refers to a similar Arabian custom,
according to which a jug decorated like a ship, filled with bric-a-brac and adorned with
candles, is presented as an offering in the hope of freeing a becalmed ship.
THE BYZANTINE SHIPWRECK (TANTURA A) IN THE TANTURA
LAGOON, ISRAEL
HULL CONSTRUCTION REPORT
Introduction
Tantura Lagoon is situated on the Israeli coast 30km south of Haifa,

Israel. The general setting of the site and a wider view of this joint INA-CMS'
underwater excavation project, is presented in this volume by S.
Wachsmann. This article focuses on the hull construction of one of the
wrecks located and excavated at this site - Tantura A.
Following two seasons of excavation, a significant part of the wreck was

revealed for investigation. The preserved timbers, which measure 9.02m in
length, lie northwest to southeast. The wood includes remnants of keel, post,
strakes, frames, nail and bolt attachments, and caulking in seams. No
mortises or tenons were employed in the construction of the hull. The vessel
is believed to be a small coaster, with an estimated length of 12m. Byzantine
pottery and carbon-14 dating place this vessel at the end of the 5'h to
beginning of the Thcentury AD. Thus, there is evidence for what is at present
the earliest example of Mediterranean hull construction in which frames
preceded planks. The reasons for this conclusion are presented below.
The Keel (fig. 1)
A 5.2 meter length of keel was preserved (Figure 1). The keel terminates
in the northwest at its junction with the post, while its southeastern terminus
exhibits signs of a break that have been degraded by teredo. The keel is
curved downwards, which may be due to stresses subsequent to deposition.
Additionally, the keel is twisted toward the southwest and is further distorted
at the post junction. The keel has a rectangular cross-section; its average
dimensions are 1l c m sided and 18cm molded. There was no keel rabbet or
chamfered edges for the garboard, nor was there a false keel.
YACOOV KAHANOV TROPIS VI
Two 26cm hook scarfs are present in the keel section, both cut in the
same direction. The first is 3.5m from the southeastern terminus of the
keel. The second, more complex, is located 1.7m further northwest, and
served to attach the keel to the post.
The Post
A naturally curved Aleppo pine timber that was attached to the

northwestern end of the keel has been labeled as a post. It is a transition
piece that formed the change between the keel and the post. No convincing
evidence has yet been found to indicate whether this was the bow or the
stern.
From the keellpost scarf, the post rises in a curving angle of about 55" It
is 1.16m long and 52cm high. Its molded dimensions at the keellpost scarf
are 16.5cm, tapering to 14.7cm along the majority of its length and
narrowing to 6cm at its end. The average sided dimension is 9.5cm. The
back rabbet carved into the upper surface of the post is 5cm high and has a
width of 1.2cm.
In addition to the keellpost scarf, the post has a second hook scarf which
is located at the uppermost extremity of the timber. This hook scarf, aligned
in the same direction as the previous two hook scarfs, has a bolt hole 1.5cm
in diameter, which probably housed the bolt that secured the scarf.
Frames (fig. 2)
On the southwestern side of the keel, fragments of eight frames (F1-F8)

were preserved, situated in seven framing stations. Linear staining patterns,
imprinted on the inner surfaces of the planking and the upper surface of the
keel, indicate seventeen additional frame stations whose timbers did not
survive. Nail holes and remnants of pitch within these stained surfaces attest
to the method of attachment. In all, 24 frame stations were evident (frame
stations A-X).
The average dimensions are frame sided 9cm, molded 9.5cm, and
center-to-center spacing 32.4cm. The extant frames differed widely in
dimension and wood type. Each was either broken at the keel, or was
severely truncated in length. Frame 8 (fig. 3) is the longest remaining framing

timber. It is curved upwards at the turn of the bilge and its total horizontal
length is 1.31m. Each frame possesses a Icm deep mortise for seating the
frames atop the keel. Nail concretions and nail holes confirm that the frames
were fastened to the upper surface of the keel by iron nails. Limber holes
measuring 4cm square were found in the frames near the keel and at the
turn of the bilge.
The extant frames are truncated on their keel end, leaving no remnant on
the northeast side of the keel, and are degraded by teredo and erosion. We
have no direct evidence for the use of floors, full- or half-frames. Futtock
remnants were present at three different frame stations.
Frame stations J through P form a contiguous line of widely spaced

frames with some of the largest sided dimensions in the vessel, averaging
10.33cm. Considering there may have not been a specific midships frame,
rather midships may have fallen somewhere between two frames, the
estimation is that midships is somewhere between frame stations K and M.
Planking (fig 4)
Northeast of the keel, only a 1.475m section of garboard, near the post,
and a loose strake fragment have survived. Southwest of the keel, eight
strakes are preserved, including the garboard, which is 8.78m in length and
2.5cm thick. It is comprised of three Aleppo pine planks, joined by two butt
scarfs. The garboards were placed 5cm below the upper surface of the keel.
They continue into the post, where after 14cm they were especially carved
to sit in the post's rabbet. The garboards were not nailed to the keel; their
nail holes probably represent attachment to frames or inner timbers.
All other strakes are also made of Aleppo pine, and are 2.5cm thick.
Plank widths vary greatly, from 3.8cm at the post to 26cm amidships. At
midships, there are eight strakes, which decrease to five at the post through
the use of drop strakes. This narrowing of the hull at frame stations H and Q,
also through the use of drop strakes, indicates that midships may lie
between these two frame stations. We only have evidence for the use of butt
scarfs in the joining of planks. Each of these was located in conjunction with
a framing station, and each plank is secured with nails.
Iron nails, driven from the outside, were used to fasten the planks to the
frames. A typical nail hole was 6mm square, with a typical spacing on a
plank of 8cm at a frame station.
There were no mortises, tenons, or treenails observed anywhere.
Evidence of charring on the northwest extremities of the strakes and on

the southeast extremity of the garboard indicates bending through a heating
process. This was not present on the frames or keel.
Caulking was evident in the planking seams in several areas, pitch was
found on the frame-plank junctions, and there was a yellowish resin on the
inner surfaces of the strakes.
Conclusions
From the preserved hull remains, there is good evidence as to the overall
dimensions and shape of the vessel. This evidence leads to an estimate of
12m for the total length of this vessel. The angle of deadrise was very small,
and the hull was almost flat-bottomed until the turn of the bilge. The
estimated beam length of this vessel is 4m.
Both carbon-14 and pottery analysis place this vessel no later than the
5m,possibly the beginning of the 7"' century AD. All evidence indicates that
at least some frames preceded planks in the construction process. The
plank edges were not connected to one another and there is no evidence of
mortise-and-tenon joints used for either plank joining or alignment. Planks
were fastened to frames with iron nails and caulking material sealed seams.
As this vessel predates any previously known archaeological parallel, it is, at
present, the earliest evidence, in the Mediterranean, for the transition from a
shell based mortise-and-tenon hull, to some form of skeletal-frame based
construction. It is likely that smaller vessels, such as this local coaster, would
experience the transition before larger craft.
Acknowledgments
The research was carried out together with Mr. Royal from INA, while the
preliminary analysis was performed together with Mr. Breitstein from the
CMS. We were privileged to have the attention of R. Steffy who was present
at the site, and the opportunity to obtain advice from S. McGrail, who also
visited during the excavation. We are grateful to both of them. We also thank
L. Basch who supported our preliminary dating by drawing our attention to
his article Le Navire d'~nee.
Yaacov Kahanov
The Recanati Center for Maritime Studies
University of Haifa, Israel
NOTES
1.INA- The Institute of Nautical Archaeology, Texas A&M University.

CMS- The Recanati Center for Maritime Studies, University of Haifa.
SUGGESTED READING
Bass, G.F. and Van Doorninck, F.H. 1978. An 1lth-century Shipwreck at Cerce Liman, Turkey.
IJNA 7.2, pp. 119-132.
Jezegou, M.P. 1985. Elements de construction sur couples observes sur une 6pave du Haut
Moyen-Age dbcouverte a Fos-sur Mer (Bouches-du-RhBne). In VI Congreso
lnternacional de Arqueologia Submarina. Cartagena 1982. Madrid. Pp. 351- 356.
Kahanov, Y. and Breitstein, Stephen. 1995. A Preliminary Study of the Hull Remains. INA
Quarterly 22.2, pp. 9-13.
Kahanov, Y. and Royal, J. 1996. The 1995 INAICMS Tantura A Byzantine Shipwreck Excavation
- Hull Construction Report. CMS News 23, pp. 21-23.
Steffy, J.R. 1982. Reconstructingthe Hull. In: Bass G.F. &Van Doorninck F.H. (Eds.) Yassi Ada
volume I.A Seventh-Century Byzantine Shipwreck. College Station. Pp. 65-86.
Steffy J.R. 1982. The reconstruction of the llm-Century Serce Liman Vessel. A Preliminary
Report. IJNA 11.1, pp. 13-34.
Steffy, J.R. 1994. Wooden Shipbuilding and the Interpretation of Shipwrecks. College Station.
Van Doorninck, F.H. 1976. The 4th-~enturyWreck at Yassi Ada. An Interim Report on the Hull.
IJNA 5.2, pp. 115-131.
Van Doorninck, F.H. 1982. The Hull Remains. In Bass G.F. & Van Doorninck F.H. (Eds.) Yassi
Ada volume I.A Seventh-Century Byzantine Shipwreck. College Station. pp. 32-64.
Basch, L. 1985. Le Navire d'~n6e.Neptunia 158, pp. 23-27.
ILLUSTRATIONS
Fig. 1. Tantura A, Keel, Post and Keel Scarfs.

Fig 2. Tantura A, Frame and Frame Station Locations.
Fig 3. Tantura A, Frame 8.
Fig 4. Tantura A, Planking Pattern.
PYRAMIDAL AND OTHER PIERCED STONE -
WHAT PASSED THROUGH THE TRANSVERSE HOLE? *
So-called pyramidal stone anchors have been used in Greece and in

Magna Graecia during the classical period. Their employment and fitting
remained as yet rather enigmatic.
In a paper on these anchors presented to the 1"'Symposium in 1985,

Honor Frost proposed a rigging consisting of a wood bar and a rope lashed
around the bar which is shown passing through the transverse hole (Frost,
1989:fig. 4)'. In explanation she states:
"The bar's projecting ends probably served as handles for lifting the
anchor and casting it overboard." (p. 99).
However, a wooden bar as long as that shown in her sketch would act
on the sea-bottom as anchor stock and turn the stone onto one side without
the hole.
Since this study ten more pyramidal stone anchors have been found.
One was lifted in 1988 near Crotone, Calabria (Fig. I), from a mooring place
in a creek at the north side of a landspit on which the ruins of an Aragonese
castle named "Le Castella" stand and where in Greek times a fortified
settlement had existed.' Nine examples were discovered in 1993 by the
Hellenic Institute of Marine Archaeology at the island Antidragonera, on the
east side of Kythera, in context with a late 4th-century BC shipwreck
(Kourkoumelis, 1 992)3
Examining the shapes of the new finds and re-examining those

published previously (Frost, 1989:figs. 3,5-8),1 noticed that the transverse
hole is always pierced through the wider sides of the stones. Moreover,
underwater photographs of pyramidal stone anchors which now for the first
time are available show that these anchors lay always in positions on the
sea-bottom in which one broad side with the transverse hole is visible from
above (Figs. 2 and 3). The conclusion is that no long wooden bar could have
been lodged in it.
Nevertheless I do think that Frost is right when she argues for a wooden
bar in the transverse hole and reasons that
"the latter is larger than most 'normal rope-holes"'.
GERHARD KAPITAN TROPIS VI
In this context she points to the apical piercing, a hole in the anchor's
topside which runs down to the transverse hole, and writes:
" ... the wooden bar ... explains the connection between the apical and
the horizontal (i.e. the transverse) piercing..." and adds:

"In some such anchors the lead that filled the apical piercings, contains
traces of corroded iron from bars, nails or pins".
Later she adds "iron rods" and mentions:

" I recall a suggestion that the iron represents the remains of apical
rings..." (Frost, 1989, 102).
The iron ring on the apex would have served for fastening the anchor
line, and this would explain why the iron rods were cast into lead with which
the apical holes were filled up. In a similar way, certain types of ancient
sounding leads have on top a metallic ring, the ends of which would have
been held into the smelted lead during founding. In pyramidal anchors a still
more solid junction may have existed, if the iron rod embraced the wooden
bar. With some skill this could be achieved with a pre-bent rod inserted into
the apical hole, before the bar was passed through the transverse hole.
Future examinations of remains of lead and iron in apical holes may reveal
the evidence.
Another question can now be answered, that of the shape of the bar in
the transverse hole. It cannot have protruded much from the openings. A
short timber with pointed ends would have been a useful fitting. The
protruding points could have been used for lifting the anchor, while on the
sea-bottom one pointed end would have performed the function of a
gripping anchor arm. With the arrangement here proposed (Fig. 4)
pyramidal stone anchors would not have been weight-anchors, but much
more efficient composite anchors, although of a type with only one arm
timber.
Pyramidal stone anchors found in harbours and on roadsteads are likely

to have served as mooring anchors. As heavy composite anchors they
would have been still more suited for this employment. The argument,
however, will not be discussed here. More insights may result from future
discoveries, and perhaps one day a pyramidal anchor will be found deep
under mud or sand, well preserved with all its fittings.
Despite their quite particular design, there is no doubt that pyramidal

stone anchors are related to other pierced stone anchors. They were not
WHAT PASSED THROUGH THE TRANSVERSE HOLE?
developed from the shape of pyramids with which they have a remote
similarity when standing upright, namely their sloping sides. These are,
however, not triangular, but trapezoidal, and their base or underside is
rectangular, not square. Direct forerunners of the pyramidal stone anchors
were probably similar pierced stone anchors without apical hole. Of these a
few were found in the harbour of Zea Liman, together with the improved
type, and I think it significant that these, only a bit simpler anchors, are of
smaller sizes (Frost, 1989: figs. 3 (2/73), 6.11 and 12)'. Other big pierced
stone anchors may also have influenced the thickness of pyramidal anchors,
e.g. Frost, 1963: fig. 7.' Moreover, the rather triangular and sometimes quite
big Bronze Age stone anchors from Byblos are in some way similar to them
(Frost, 1963: fig. 4; Galili, 1985: fig. 3).
Finally there is a forerunner for an apical piercing in a triangular Bronze

Age stone anchor in Malta (Frost, 1963: fig. 15). Its fitting may likewise have
consisted of a short arm timber in the transverse hole and of cords in the
apical piercing lashed around the timber (Fig. 5), but this arrangement would
have been removable6.
Arm timbers fastened to the upper part of primitive anchors are known
from ethnographic examples. A flat triangular stone from the Gilbert Islands,
East-Micronesia, has a pointed arm timber lashed onto the apex together
with a rope loop (Sarasin, 1938: 16 f., text fig.3). In Sri Lanka I found an
anchor made from an almost rectangular stone and two parallel arm timbers
which are transversally tied over the upper side, together with a short, round
wooden cross-bar over the arms. To the latter the rope is lashed in turnable
fashion (Fig.6)'. Similar in arrangement is a primitive anchor device made
from a tree branch with hook-like ramifications and a ballast stone bound to
its lower end (Van Nouhuys, 1951: 9, fig. 1-rights). In the same article (p. 24,
fig. 4 - 3" and 4"' row), among anchors of fishing craft on the Spanish coast
and the Canary Islands, two anchor stones are depicted, each with an arm
timber in a single transverse hole.
This leads us to the question, whether certain prehistoric pierced stone

anchors also may have had an arm timber in the transverse hole. Indeed,
Bronze Age stone anchors present several features, which suggest this.
The one-holed anchors from Byblos are distinguished by grooves

chiselled between transverse hole and apex and over the apex. The
transverse hole has been considered the "rope hole" and accordingly the
grooves were called "rope grooves", but no anchor line was lashed there or
laid into the grooves. The grooves were made for keeping in place the cord
lashings of a rope loop tightly fastened to the apex. The almost triangular
Byblian-type anchors have a very pointed apex. If they were perhaps without
grooves, cords and loop could slip down from there. The same is true for
Egyptian stone anchors having a rounded apex. Or, in other words, without
grooves the pull from the anchor line would be exerted on the transverse
hole. In this case the diagram of forces would be less favourable and the
resistance of the anchor minor.
One advantage of a well-fastened rope loop is an optimal holding

capacity of the anchor; another is that it saves anchor line. If the anchor line
were lashed to the so-called "rope loop" - as it was initially on pierced
anchor stones - it would continuously be subject to wear. For this reason
all improved anchors have a loop or a rope ring. Even the simplest
ethnographic anchors consisting of tied-up stones are provided with a loop.
There is no doubt that during the Bronze Age and even much earlier a rope
loop was lashed to stone anchors. A triangular Neolithic anchor stone from
Khirokitia, Cyrpus, dated from context to the 6'h millennium BC, is provided
with grooves running from its piercing to the apex and over it (Frost, 1984),
exactly in the same way as later on the stone anchors from Byblos.
It would seem that loop cords kept in grooves are largely protected from
wear, and doubtless this is the case. However, protection was not the
purpose of grooves. Otherwise they would have been chiselled also on other
Bronze Age stone anchors, e.g. on those having trapezoidal or almost
rectangular sides. Here, the rope loop fastened to the apex could not glide
sideways, and this is why stone-masons saved themselves the time to cut
grooves. An example of grooveless anchors are the 24 specimens
discovered in the late 14'h-centuryBC Uluburun shipwreck in Turkey (Pulak,
199557, and 1996). For the most part they may have belonged to the cargo
(Frost, 1995).'
For tying loop cords to one-holed stone anchors, there are two
possibilities: either they are lashed through the transverse hole, or they are
fastened, on both sides of the stone, to an arm timber lodged in the hole
(Fig. 7). The first way of lashing is used on one-holed weight anchors. The
second system would be characteristic of composite anchors with one arm
timber.
Lashing the loop cords to an arm timber presents an essential

advantage; they do not come immediately in touch with the sea-bottom,
since the anchor's upper part lies at first somewhat raised, resting upon the
gripping arm. On the other hand, the loop cords of grooveless weight
anchors touch the seafloor directly and would wear out soon. This must have
encouraged the change to one-armed composite anchors.
On Bronze Age stone anchors still other features indicate a fitting with
arm timber. In many cases their transverse holes are much larger than is
needed for fastening to them big anchor lines, and far larger than is needed
for lashing the smaller loop cords. Furthermore, the transverse holes are in
various cases not round but square or almost square. The Uluburun anchors
are again an example. Square or angular holes are best suited for lodging
arm timbers.'
To prepare a list of one-holed Bronze Age stone anchors which were or

could have been composite anchors with one arm timber requires
reconsidering studies with detailed examination of all known anchors. Here
I mention only that three-holed stone anchors of the Middle and Late Bronze
Ages are also likely to have had a third arm timber in their usually very large
and often square upper transverse holes (see Shaw, 1995; Frost, 1989:
fig.1). With three arm timbers they would have been a type of super gripping
anchor, and this may be the reason, why many have now turned up in
temples and other sacred or privileged places, where they had been put as
ex-votos.
POSTSCRIPT
From tests of the terracotta models (Fig. 8)'' in a water basin it resulted
that both pyramidal stone anchors with and without apical piercing settle
onto any side, no matter if a short arm timber is lodged in the transverse hole
or not. The reason is that their sides are not sufficiently different in width.
If pyramidal anchors always settled onto one broader side with

transverse hole, as is evident from underwater photographs, they must have
had in addition a stock timber fastened onto the apex, crosswise to the
transverse hole or the arm timber. The stock would have been attached
separately, by means of cords lashed either through the transverse hole of
an anchor without arm timber, or to the arm timber lodged in this hole. The
stock could be quickly fixed, before the anchor was cast from aboard; it had
to be sufficiently strong and at least about twice as long as the arm timber.
GERHARD KAPITAN TROPlS VI
*The terms "horizontal piercing" (ore hole) and "rope hole" used elsewhere
are here replaced by the word "transverse hole", and that for the following
reasons: First, in the anchoring position of pierced stone anchors this hole
runs vertically. Only when the anchor is set up is the piercing found in a
horizontal position. Secondly, on simple pierced anchor stones this hole was
initially used for fastening through it an anchor line, but later the anchor line
was always tied to a rope loop or ring on the apex.
Gerhard Kapitan
Viale Tica 53 (v. Regia Corte 4)
1-96100 Siracusa
Italy
NOTES
1. In a former reconstruction of the rigging of pyramidal anchors Frost had

proposed a rope lashed through both piercings (Frost, 1963: fig. 18).
2. A defensive wall of the Greek settlement built from blocks in sparing
technique and from stones filled into the spaces was included in the
construction of the Aragonese castle.
I am obliged to Dr. Alice Freschi for having made available to me her
unpublished research report "Ricerche subacquee nel Crotomese, 1988".
She also generously allowed me, in arrangement with the Soprintendenza
Archeologica di Calabria, to use her scale drawing of the pyramidal stone
anchor and an underwater photograph of it in my illustrations Fig. 1 and 2.
3. 1 am indebted to Nikos Tsouchlos, President of H.I.M.A., for having sent
me a copy of ENALIA with the report here-cited, and to Dimitri
Kourkoumelis for his letter of 27-11-95 in which he kindly answered
questions. The underwater photograph by N. Tsouchlos which illustrates
my Fig. 3 was kindly made available as proof of the evidence.
4. A roof-shaped pierced stone, which is perhaps related to pyramidal
anchors, was found in 1967 in the ruins of Epidaurus Limera, near
Monemvasia, Laconia. For information I am indebted to Frank Frost, Santa
Barbara, California, who also sent a colour slide. The stone is about 62-
64cm long and has two steeply sloping rectangular sides with a large
round piercing just beneath the ridge-shaped apex, and two vertical
triangular sides. Whether it had been used as anchor or as a weight, e.g.
suspended from a beam, is unknown.
5. A rather big, but roughly cut, almost "pyramidal" stone anchor with a small
saddle-shaped apex, in the Dover Museum, but perhaps originating from
the Mediterranean (Frost, 1963:4-fig. 20, 6, 12, 20), may likewise have to
be considered in this context.

6. An alternative lashing of the rope loop to this Maltese stone anchor, if it
was without arm timber, would be similar to the rigging proposed at first by
Frost for the pyramidal anchor; see note 1.
7. This killick was used on a small sailing oruwe, a single-outrigger logboat
canoe at Talahena, south of Negombo (Kapitan, 1987: 141 ff.) It was cast
offshore during line fishing in deep water, when the canoe had to ride at
anchor. This year I managed to relocate the killick, which was now
damaged and out of use. The fisherman owning it expressed the opinion
that it is insignificant to which end of the killick the arm timbers are tied. At
Talahena I also photographed a double-armed killick of reverse
arrangement.
8. The anchors may have been stowed in the hull in unrigged state, that is to
say, without arm timbers in their transverse holes, as this would have
saved space (see the following statements). To insert an arm timber in the
transverse hole, to fasten cords to the timber and to tie from the cords a
loop on the anchor's apex can quickly be done by experienced mariners,
when the stone anchors have to be used.
9. It was thought at all times that the smaller lower holes of three-holed stone
anchors were for arm timbers. This is also demonstrated by ethnographic
examples (Van Nouhuys, 1951: fig. 4 - last row). Occasionally these holes
are angular (Frost, 1963: figs. 21-22, 24-25), but such square or angular
lower holes already appear on three-holed Bronze Age stone anchors
(Shaw, 1995: figs. 3(c), 4,8,9). Arm timbers of corresponding sections can
be solidly affixed in such holes and sit more securely than round timbers
in round holes.
10. The terracotta models shown in Fig. 8 were produced by the author in
order to check the described fittings. The author is indebted to Prof.
Franco Giudice, lnstituto d'Arte, Syracuse, for having provided clay and
taken care of the baking of the models.
REFERENCES
Frost, H . , 1963, From rope to chain. The Mariner's Mirror 49.1,1963: 1-20.
Frost, H., 1984, Khirokitia; une pierre d'ancrage. Fouilles recentes a Khirokitia (Chypre) 1977-
1981 (2 tomes). Ed. Recherches sur les Civilisations. Paris: 125-126, 146, PI. XXX.
Frost, H., 1989, "Pyramidal" stone anchors; an inquiry. TROPlS 1: 97-114.
Frost, H., 1995, Where did Bronze Age ships keep their stone anchors? TROPIS 111: 167-175.
Galili, E., 1985, A group of stone anchors from Newe-Yam. The International Journal of Nautical
Archaeology 14.2: 143-153.
Kapitan, G., 1987, Records of native craft in Sri Lanka - I : The single outrigger fishing canoe
oruwa - Part 1 . Sailing oru. The lnternational Journal of Nautical Archaeology 16.2: 135-
147.
Kapitan, G. & Naglschmid, F., 1982, A 4m-centurydispersed amphorae cargo on the Secca di
Capo Ognina, Siracuse, Sicily (site Ognina 4). Proceedings of the Diving Science
Symposium (6mInt. Scientific Symposium of CMAS, Edinburgh 1980) 1982: 229-239.
Kourkoumelis, D., 1992, Avayvoplcrrl~!junoppux~adpeuva o-rq BaA6uma nep~o)(liAukp8va
Kuefipov nepto6ou 1993. Enalia (Athens) IV. 112: 6-1 1.
Pulak, C., 1995, Das Schiffswrack von Uluburun. In Poseidons Reich. Archaologie unter Wasser
(ed. by DEGUWA e.V.). Zaberns Bildbande zur Archaologie, Band 23. Mainz am Rhein:
43-58).
Pulak, C., 1996, Dendrochronological dating of the Uluburun ship. The INA Quaterly 23.1: 12-
13.
Sarasin, F., 1938, Uder die Geschichte des Ankers. Verhandlungen der Naturforschenden
Gesellschaft in Base1 (Basel) XLIX, 1937-38: 9-53, Taf. I-VIII.
Shaw, J.W., 1995, Two three-holed stone anchors from Kommos, Crete: their context, type and
origin. The InternationalJournal of Nautical Archaeology 24.4: 279-291.
Van Nouhuys, J.W., 1951. The Anchor. The Mariner's Mirror 37.1: 17-47.
CAPTIONS TO THE ILLUSTRATIONS
1. Scale drawing of the pyramidal stone anchor from "Le Castella", Calabria from the excavation
report by Alice Freschi (see note 2).
2. The pyramidal stone anchor from "Le Castella", Calabria photographed in situ by Alice
Freschi (see note 2).
3. The pyramidal stone anchors A5 and A6 photographed by Nikos Tsouchlos on a 4m-century
BC shipwreck site at Antidragonera, Kythera. (Photo by courtesy of H.I.M.A., Athens; see note
3) -
4. Pyramidal stone anchor from the 4m-centuryBC shipwreck Ognine 4, Syracuse; scale drawing
(after Kapitan & Naglschmid, 1982: fig. 6), completed with the author's sketch of its proposed
rigging, but see the postscript for the additional stock timber.
5. Scale drawing of the Bronze Age stone anchor in Malta having an apical piercing (after Frost,
1963: fig. 15), completed by the author's sketch of its suggested rigging which consists of an
arm timber in the transverse hole and of cord lashings for a rope loop on its apex. The cords
are passed through the apical hole and embrace the arm timber; on the apex they are knotted
over a wooden counter-part in the piercing.
6. Killick-type anchor with two parallel arm sticks between the anchor stone and an upper
transverse rod to which a rope loop is lashed in turnable fashion. The killick was in use on a
sea-going outrigger fishing canoe at Talahena near Negombo, Sri Lanka (author's photo
1986).
7.One-holed Bronze Age stone anchors of grooveless type such as were found in the Uluburun
shipwreck, Turkey:
a. unrigged;
b. with a rope loop on the anchor's apex the cords of which are lashed through the transverse
hole;
c. with a rope loop the cords of which are tied to an arm timber in the transverse hole.(sketch
by author).
8. Terracotta models of ancient pierced stone anchors discussed or mentioned in this treatise,
provided with the suggested fittings of a rope loop or ring and - in some cases - of an arm
timber in the transverse hole.''
-
Fig. 1
Fig. 6
KAPITANGERHARQ TROPlS VI
o20-49
crn
_ Fig. 4
Fig. 7
I
Fig. 8
3 14
ABSTRACT
NOTES ABOUT THE ROMAN NAVIGATION

IN THE MIDDLE DANUBE AREA
Underwater archaeological research on the territory of Serbia has not

been carried out, so this paper will present the summary of the data related
to the navigation on the Danube and its tributaries which have been available
on the basis of archaeological material from this area. We shall deal with the
findings of the fibulae, which are supposed to have been made for the needs
of the members of the fleet on this section of the Danube limes. We shall also
refer to the diffusion of the bricks with the stamps "classis" and, in relation to
this, to the question of the river ports, as well as to the cults of the gods
connected with the water and the sailors. This paper will be concluded by
the illustrative review of bronze oil lamps made in the shape of ship models.
Gordana Karovic
Republicki zavod za zastitu spomenika kulture
(Institute for the Protection of Cultural Monuments of Serbia)
Bozidara Adzije 11,
11000 Beograd, Yugoslavia
THE SHIP AS REALITY AND SYMBOL:
HOW IT WAS PERCEIVED IN HELLENISTIC AND ROMAN PALESTINE
The practical and symbolic attributes of the ship made it a recurrent

theme in the literature and art of ancient civilizations. The perception of the
sea in the Eastern Mediterranean was greatly influenced by the fundamental
role of ships and navigation in Greek civilization. Homer and his works still
remain the mirror and model of the Greek attitude to the sea, at least in its
literary form. Surprisingly, however, there are few modern works that deal
specifically with the significance of maritime aspects of Graeco-Roman
literature1.
This paper attempts to explore the extent to which the concept of the
ship permeated the life and thought of Hellenistic and Roman Palestine,
according to ship motifs in their geographic and historical context. Secondly,
we wish to distinguish the ship as an artistic convention from its symbolic
representation as related to a broader cultural background, especially in
Jewish sources. This perspective may help to understand the relationship
between realistic and metaphorical elements in the examples under view.
I. Biblical perceptions of the ship
Maritime passages in the Bible are particularly suitable in this context,

because they convey the literary thematic traditions of Palestine, as of the
first millennium BCE. A characteristic of these texts, typical of biblical style,
is the mixture of realistic and metaphorical elements in the depiction of the
ship.
The prophecy at Tyre in Ezekiel 27, describing the rise and fall of a
Phoenician thalassocracy, is probably the most richly detailed maritime
chapter in the Bible in which the ship is a principal theme2.The text manifests
a basic comprehension of the technological aspects of navigation, and
enumerates the materials and parts used in the construction of vessels -
wood, shell, mast oars and keel; linen and sails, ropes; crew and weapons.
Then comes a detailed description of international commerce, products,
places and routes. The prophecy concludes by foretelling the dramatic
collapse of Tyre, i.e. the sinking of the shiplempire, and the moral-historical
lesson.
NADAVKASHTAN TROPIS VI
While such 'complete' descriptions are rare in biblical texts, shorter

references are nonetheless significant. The realistic, or practical, side of
shipping is emphasized by the famous declaration of the maritime vocation
of the tribe of Zebulun - "Zebulun shall settle at the shore of the sea; he
shall be an haven for ships, and his border shall be at Sidonn3.The double
terminology for coast and haven (hof yamimlhof oniot in Hebrew) appears to
be the only explicit biblical reference to a harbour or harbour installations4.
In a broader sense, the passage conveys the idea that geography, economy,
and culture are inevitably interdependent.
For future generations of inhabitants in Palestine, Zebulun would

exemplify the expert in practical knowledge of shipping, as we shall see from
a later source.
However, the Bible uses the ship more often as a metaphor or allegory
for a broader moral lesson. Two short illustrations in Proverbs are: The praise
of the woman, and comparing her to a merchant ship: "She is like the ships
of the merchant, she brings her food from far awayn5.The author of the text
believes the diligence, intelligence, and initiative of the 'ideal woman' are
characteristic of the ship. Beyond this progressive approach to woman, the
comparison between woman and ship shows how the maritime theme was
chosen. Here, the ship represents strength, wealth, and practical ability.
Another clear statement of the virtue of the ship is the well-known

passage in Proverbs - "Three things are too wonderful for me; four I do not
understand; the way of the eagle in the sky, the way of the snake on the rock,
the way of a ship on the high seas, and the way of a man with a maid"6.The
universe is thus divided into its basic elements, and the ship embodies the
sea. The "wonderful" aspect here is attributed, in some commentaries, to the
disappearing trace of the passage of the ship through the water, as it is
explicitly expressed in an apocryphal text'. Moreover, analogy with the eagle
and the snake shows that a human technological invention was highly
enough considered to be used to represent the entire maritime world.
II. Graeco-Roman perceptions of ships: reality and symbol
The Hellenistic and Roman Age in Palestine was a time of relatively

intense maritime activity, to judge by the role of ports and fleets in the events
of the period. Naval engagements and increased maritime commerce under
Ptolemaic, Seleucid, and Roman rule gave the coast and its cities a central
THE SHIP AS REALIlY AND SYMBOL:
place. From a Jewish perspective, the Maccabaean leaders and King Herod
added important chapters to the maritime history of the country, of which
several proofs are to be found in the sources.
The Hellenization of Palestine, given impetus by the arrival of Alexander

the Great (332 BCE), had a significant impact on most spheres of life. It also
created new forms of expression which were endemic to the confrontations
of civilizations in the dramatic events of that era.
Texts of the period combined biblical and post-biblical forms with

Graeco-Roman views and concepts in a Hellenized environment. In the
background, the Iliad and Odyssey, Hesiod, and other sources, continued to
inspire writers to use maritime themes, especially ships. The works of Flavius
Josephus, the Books of the Maccabees, the New Testament, and texts of the
Pseudepigraphia and Apocrypha, from which several examples are chosen
here, all contain passages presenting the ship, both as reality and metaphor.
Together, they developed the literary traditions of the age.
Realistic themes
The pratical skill of Zebulun in maritime affairs reappears explicitly in the

Testaments of the Twelve Patriarchs, a text probably composed in the Late
Hellenistic period (Pd-1" centuries BCE). Zebulun boasts of the way he
realized his old maritime vocation: "Iwas the first to make a boat to sail on
the sea, for the Lord gave me the necessary knowledge and skill. And I fitted
a rudder behind it, and stretched a sail on an upright piece of wood in the
middle of it"'.
The simple recounting of constructing a vessel reflects the increased

activity of fishermen and others operating along the Mediterranean coast of
Palestine. Shipping and navigation developed considerably during the
struggles between the Ptolemies and Seleucids, and consequently in the
time of the Maccabees. The construction of larger vessels for commerce and
warfare, also reflected by iconographic representations, would become a
rewarding and highly esteemed occupation in the country.
In the New Testament, particularly in the Acts, the journeys of Paul across
the Mediterranean are a good illustration for texts of the Roman period, with
rich maritime information using realistic descriptions. The episode of arrival
in Malta is a famous example: " ... but they noticed a bay on which they
NADAVKASHTAN TROPlS VI
planned to run the ship ashore, if they could. So they cast off the anchors
and left them in the sea. At the same time they loosened the ropes that tied
the steering oars; then, hoisting the foresail to the wind, they made for the
beach. But striking a reef, they ran the ship aground; the bow stuck and
remained immovable, but the stern was being broken up by the force of the
wavesng.
This account of an attempt to reach anchorage ending with a shipwreck

is given credibility by the order of events and the enumeration of parts of the
boat - anchors, ropes, steering oars, foresail, bow, and stern. The entire
episode reminds one of the story of Jonah, and follows the dramatic pattern
of navigation-storm-shipwreck-rescue seen originally in Homer1' and
repeated in Hellenistic sources. If Paul's 'schematic story' was not written by
a witness or participant, it was certainly transmitted by someone
experienced in the arts and techniques of navigation.
From reality to metaphor
The metaphors and allegories derived from the ship contain realistic
representations, themes, and literary forms inspired by classical traditions
and sources. The maritime information they include, when examined in its
historical context, points to the relative importance accorded to the sea and
navigation by writers of the Graeco-Roman period.
Flavius Josephus is of special significance here, because his career and

works express the fusion of cultures in Hellenistic and Early Roman
Palestine. As a Jew writing in Greek and holding pro-Roman views, he was
naturally aware of maritime affairs. Personal and political contacts with Rome
also necessitated journeys by sea, which are mentioned in his books. Most
of the descriptions concern King Herod and his successors, and their travels
along the central route connecting Palestine, Asia Minor, and Rome1'.
Josephus' accounts, in spite of inaccuracies and exaggerations contribute to
our knowledge of the active maritime routes in the Eastern Mediterranean''.
In The Life (Vita), there is only one reference to a journey Josephus

himself made to Rome in 64 CE. It ended, he says, with shipwreck in the
Adriatic and the miraculous rescue of eighty out of the six hundred
passengers on board, by a Cyrenaic ship. Modern commentaries tend to
agree that the tale lacks credibility in its details - the size of the ship and
number of passengers1=;at all events, it confirms that the shipwreck motif
was adopted as a dramatic theme by Josephus, as it was by other writers of

the age.
Josephus' personal history and experiences perhaps reinforce the

meaning of ship allegories in his works. In one case, he recalls the difficult
times in Palestine under Antiochus the Great and Antiochus Epiphanes
towards the end of the 3'dcentury BCE. To explain the effect of instability, war
and change on the country, he says: " ... so that they were no different from
a storm-tossed ship which is beset on either side by the heavy seas, finding
themselves crushed between the successes of Antiochus and the adverse
turn of his fortunes"14.The passage alludes to the wars of succession in
Palestine, in particular the Battle of Raphia (217 BCE), which was one of the
decisive confrontations between the Ptolemies and Seleucids on the
southern part of the coastal plain.
It is relevant that Marcus Agrippa, mentioned by Josephus a few

passages earlier for his defense of Jewish privileges, had himself travelled
across the Mediterranean on his journeys to Alexandria and Rome15.The
analogy between nation and ship, or between political hardship and storms,
makes a convincing argument in oratory. The phrase "turns of fortune"
associated with the idea of a ship, reminds us of Tyche in her maritime form
as the goddess-protector of ports and navigators. Evidence of the worship
of Tyche, presented with the attribute of a steering-oar, has been found in
several maritime cities of Pa1estinel6.
The analogy between the nation of Israel and the ship is elaborated in the
Testament of Naphtali, an apocryphal text which opens with: "Whilst we
stood with Jacob our father on the shore of the Great Sea, behold, a ship
came sailing in the middle of the sea without sailors and without people"17.
This is part of a detailed account of the shipwreck of a merchantman, lost
through the inability of its crew to unite and overcome the stormy sea.
Evidently, the story is an allegory in biblical or midrashic style, of the conflicts
between the tribes, and their impact on the people of Israel. In the story, the
vessel, found "without a sailorn- in a Hebrew version1' - is finally saved
and repaired by Jacob, the father. The mention of two masts, rudders,
steering oars, sails and cargo, shows at least some experience in navigation.
Apart from such maritime terminology, an analogy between the fate of a
nation and that of a ship is rarely found in writings of the period. The nearest
equivalents, both in length and importance, are the accounts of Paul's
journeys in the Acts mentioned above, which contain credible maritime
references.
NADAV KASHTAN TROPIS VI
The ship is a metaphor for the reverses of fate in yet another dramatic
situation, in which Josephus himself is involved when he sees that the battle
of lotapta is lost. The defenders try to influence him not to flee, saying:
"Moreover, it would be unworthy of him to fly from his foes, to desert his
friends, to leap in the storm from the vessel on which he had embarked in a
calm"lg. Commanding an army, fighting a war, or manifesting loyalty to the
nation are compared to the mission and duties of a captain at sea. The
metaphor, though a commonplace, gains importance in a crucial moment of
the Jewish War and of Josephus' career.
Ships were perceived as symbols of rule and power in Hellenistic

Palestine, according to a singular passage in Maccabees I.It is the
description of the mausoleum in Modiin which Simon built to commemorate
his family. The monument, it is said, was erected with seven pyramids,
representing the members of the Maccabaean dynasty. "For the pyramids
he devised an elaborate setting, erecting about them great columns, and on
the columns he put suits of armour for a permanent memorial, and beside
the suits of armour he carved ships, so that they could be seen by all those
who sail the seapQ0.
The choice of the ships to adorn the mausoleum of the Maccabaean

dynasty is particularly significant here, and attests to the interest the
Maccabaean leaders took in the sea and in navigation. Their involvement in
maritime affairs probably resulted from a conviction that the coastal districts
should be included in their territory. For they knew that without dominating
the coast, and without prosperous ports, the security and independence of
the state would be at risk. Not all leaders of the dynasty shared this view, or
gave priority to maritime affairs.
Of all of them, Simon seems to have been the most conscious of the
economic and cultural benefits of ties with neighbours in the Mediterranean.
After several campaigns along the coast, he fortified the city of Jaffa2',
assuring its position as the main port of the country. His decision to erect a
monument carved with ships for his ancestors is less surprising in this
context. Simon could have easily chosen other symbols to commemorate
battles on land or victories of the Maccabees; but his choice, if we accept the
facts, affirms in yet another way the importance of navigation at sea as a sign
of power and independence.
Ill. The symbolic connotations of ship depictions
Allegories and metaphors involving ships can be better understood in

their broader context, not merely as literary devices. The texts discussed
above are witnesses to periods of increased maritime activity in Palestine
and depend on political circumstances and the personalities involved. They
can be corroborated by the archaeological evidence of the period, though
ship depictions are more problematic, both as reality and metaphor. The
concluding remarks do not refer to technological data, which can be derived
from ship representations, but only to their geographical context and
symbolic value.
One would expect to find marine motifs in the coastal areas of Palestine,
but most of the iconographic representations of ships, dated to the
Hellenistic and Roman periods, have been discovered in Jerusalem, the
Judaean Desert, or other inland regions. The natural correlation between
demographic centres and archaeological evidence does not seem adequate
to explain this.
The geographical locations of depictions of ships, mostly inland, must be

related to their symbolic and metaphoric context. Marine motifs are primarily
artistic expressions, in different media, and for various purposes. Ships,
parts of ships, fish, and other marine motifs, obviously, appear more often
on coins2' and oil lamps; they should be examined separately, taking into
account the political and economic consequences of their dissemination in
large quantities in Graeco-Roman times. There are, on the other hand,
drawings and graffiti of ships found in Israel, which reinforce the symbolic
aspects discussed.
The examples chosen belong mostly to depictions discovered at well-

known sites in Israel, and published during the last 40 years. These include
a recently discovered mosaic floor (see illustrations); their meaning has still
not been clearly understood.
The graffiti on 'Jason's Tomb' in Jerusalem (excavated in 1956) are the

only ones which show a Hellenistic warship equipped with a battering ram,
and possibly pursuing two other shipsz3.It can be assumed that a vessel of
this type would have participated in the naval engagements of the second or
first centuries BCE. A similar warship could have decorated the Maccabaean
monument discussed earlier. The drawing at Jason's Tomb contributes to
our knowledge of form, equipment, construction and, more generally, naval
NADAVKASHTAN TROPIS VI
warfare. Nevertheless, the choice of a ship to decorate the tomb of a

Hellenized Jew in Jerusalem remains an enigma. We shall probably never
know whether if it indicates the involvement of the owner in maritime affairs,
or if it was merely a posthumous symbol. The latter theory is more
acceptable, judging from the common motif of "the last journey" often
expressed by the ship in Graeco-Roman burial sites.
Another drawing, of a merchant ship, was discovered on a stone

beneath the Church of the Holy Sepulchre in Jerusalem, in 1971. The
"Jerusalem Ship Drawing" has been recently examined and dated to the first
or second century CEZ4.It is suggested that the drawing was executed by a
Christian pilgrim during the construction of the basilica, or by a "merchant
sailor and his companions who had journeyed inland from Caesarea, selling
or buying goods in Aelia Capitolina for shipment to the Westnz5.
Similar problems of interpretation exist for other graffiti and mosaics of

merchant ships from the Roman period in Palestine. Ship engravings were
found in Beth-Shearim, the Jewish necropolis east of Haifa in 1936/3726. It is
another example of depictions of ships in burials of the later Roman period
(third century CE). More graffiti have been found in isolated sites of Judaea,
including two in Herod's northern palace at Masada". This can be attributed
to Herod's interest and involvement in the sea, particularly to his maritime
journeys, or perhaps to navigation on the nearby Dead Sea.
A recent important discovery is the mosaic floor at Lod, east of Tel Aviv,
dated to the third or fourth century CE, and found in spring 199628. One
section presents a maritime scene with fish and two merchantmen, one of
which is perfectly conserved. The only other mosaic of Palestine with a ship
was found in 1977,at Migdal (Magdala) near the Sea of Galilee. The mosaic
represents a conventional outline of a fishing or cargo boat29 which, because
of its location, may refer to navigation and battles on the lake during the
Jewish War against Rome as described by Josephus30.
Conclusion
The ship depictions we have presented here can be interpreted in two

ways:
a) as realistic commemorations of persons directly involved in a maritime
profession or occupation, such as shipowners or shipbuilders;
b) as symbols and metaphors for wealth, political or military power, or
religious influence. Both these alternatives are applicable to iconographic

representations, and both express the realistic and metaphorical aspects
of the ship. Further conclusions may be derived from the preceding
examples:
1) Literary representations of ships often combine realistic data and

allegorical themes. They are conveyed in a simple, conventionalized
manner, or in detailed and sophisticated form.
2) Texts, and certainly iconographic depictions of ships, remain enigmatic

pieces of information, but their historical, geographic, and cultural
background is crucial for an understanding of maritime activities in
Hellenistic and Roman times. The Hellenization of Palestine encouraged
the absorption of Greek forms and ideas, including the ship and ship-
symbols. The Maccabaean monument is a good example of this.
3) Maritime literary traditions were inspired by Homeric, classical and

biblical references, such as episodes of the Odyssey, Ezekiel's prophecy
concerning Tyre, and later, the Testament of Naphtali and Paul's travels.
These were also popular literary models for Jewish and Christian writings
of the Second Temple Period.
4) Detailed descriptions of ships and their parts generally appear in texts

concerned with journeys and battles, where technical data are of primary
importance. Conversely, storms and shipwrecks are popular themes for
allegories, employed both as dramatic material and as vehicles for moral
or theological lessons. The adventures of Jonah and Paul are examples
for this.
5) Ultimately, the use of allegorical elements reflects a degree of

experience and involvement in such maritime occupations as shipping,
fishing, commerce. What it basically implies is the assimilation and
adoption of a maritime culture in its artistic and spiritual forms.
Dr. Nadav Kashtan

University of Haifa, Israel
NADAVKASHTAN TROPlS VI
NOTES
1. Symbolic aspects are discussed in: E.R. Goodenough, Jewish Symbols in the Graeco-Roman
Period, N.Y. 1958, (vol. 8, pp. 157-165). For iconographic references: L Basch, Le Musee
imaginaire de la marine antique, Athens, 1987.
2. Ezekiel, 27, 1-36.
3. Genesis, 49, 13.
4. Two allusions to ports: Psalms, 107, 30: "... and he brought them to their desired haven".
(Hebrew: mehoz heftsam); Ezekiel, 27, 3: "..Tyre, who sits at the entrance to the sea".
(Hebrew: mevo'ot yam).
5. Proverbs, 31, 14.
6. Proverbs, 30, 18-20.
7. " As a ship passing through the billowy water, Whereof when it is gone by, there is no trace
to be found". The Wisdom of Solomon, 5, 10-13.
8. The Testament of Zebulon, VI, 1-2 in: The Testaments of the Twelve Patriarchs (Apocrypha).
9. Acts, 27, 39-41.
10. Jonah, 1 , 4-12. Cf. the Homeric shipwreck at Scheria, Odyssey, 5, 367-463.
1 1 . Josephus Flavius, Jewish Antiquities (hereafter:JA), 14,374-380; JA, 16, 16-26; The Jewish
War (hereafter: BJ), 1, 290 (Herod's travels); Herod Archelaos: BJ, 2, 14-18; Pontius Pilate:
JA, 18,89; BJ, 1 , 634.
12. J. Rouge, Recherches sur I'organisation du commerce maritime en Mediterranee sous
/'Empire romain, Paris, 1966, pp. 85-93.
13. Josephus, Vita, 14-16. Cf. commentaries in: L. Casson, Travel in the Ancient World, Toronto,
1974, pp. 158-160; T. Rajak, Josephus and Modern Scholarship, Philadelphia, 1983, pp. 43-
44; G. Hata, "Imagining Some Dark Periods in Josephus' Life", in: Josephus and the History
of the Graeco-Roman Period, Leiden, 1994, pp. 312-313.
14. Josephus, JA, 12, 130. Translation by R. Marcus in the Loeb edition, vol. VII (1961), p.65.
15. JA, 18, 155 ff.
16. For Dor, see: R. Sofer-Ovadia, " A Bronze Statuette of Tyche", Sefunim, 1 (1966), pp. 21-24
(HebrewIEnglish); for the Tyche of Caesarea Maritima, cf. A. Gottlischer, Nautische Attribute
Romische Gottheiten, Bremen, 1981; R. Gersht, "The Tyche of Caesarea Maritima", Palestine
Exploration Quarterly, 116 (1984), pp. 110-114.
17. The Apocrypha: The Testaments of the Twelve Patriarchs - The Testament of Naphtali
("About Natural Goodness"),ch. VI. Cf. Odyssey, VII, 557-563 for the self-navigating ships of
the Phaecaeans.
18. See the text in: D. Sperber, Nautica Talmudica, Leiden, 1986, pp.86-91, with notes and
references to Greek and Hebrew versions.
19. BJ, 3, 193-196. Cf. BJ, 2, 556, with a similar metaphor for Jerusalem.
20. 1 Macc., 13, 27-30.
21.1 Macc., 12, 33-34; 1 Macc.,14, 5, 34.
22. A. Ben-Eli (ed.), Ships and Parts of Ships on Ancient Coins (Pub. by the National Maritime
Museum, Haifa), Haifa, 1975. Ships do not appear as symbols on Jewish coins, not even on
those issued by the Maccabaeans; the anchor was evidently the most popular maritime
theme. Cf. E.W. Klimowsky, On Ancient Palestinian and Other Coins, their Symbolism and
Metrology, Numismatic Studies and Researches (Pub. by the Israel Numismatic Society), Tel
Aviv, 1974, pp. 21-50.
23. L.H. Rahmani, "Jason's Tomb", lsrael Exploration Journal, 17 (1967), pp. 69-75; figs. 5a and
5b. The graffiti have today almost disappeared from the walls.
24. Gibson, Shimon & Taylor, Joan E. Beneath the Church of the Holy Sepulchre, Palestine
Exploration Fund Monograph Series Maior, London, 1994, pp. 25-48.
25. Ibid. p. 48.
26. B. Mazar, Beth Shearim, Vol. I,lsrael Exploration Society, Jerusalem, 1957.
27. The drawings were found on the wall of a storeroom for liquids. See: E. Netzer, Masada 111.
The Yigael Yadin Excavations 1963-1965 Final Reportsh el( Buildings), lsrael Exploration
Society, Jerusalem, 1991, pp. 119-120; ills. 193-194.
28. M. Avissar, in: Hadashot Arkheologiyot (Archeological News), Published by the lsrael
Authority of Antiquities, Jerusalem, 1996, vo1.105, pp. 157-160 (in Hebrew); id. "The
Representation of Two Merchant Ships on a Late Roman Mosaic Floor in Lod (Lydda), Israel".
(Paper read at the 6mSymposium on Ship Construction in Antiquity, Lamia, 1996.)
29. A. Raban, "The Boat from Migdal Nunia and the Anchorages of the Sea of Galilee from the
Time of Jesus", IJNA (International Journal of Nautical Archaeology), 17 (1988), pp. 311-329,
suggests that the hull and cutwater resemble the Ginossar (Gennesareth) fishing boat (Is'c.
CE) discovered nearby.
30. Josephus, BJ, 2, 635; BJ, 3, 523-531, mentioning the small boats which participated in the
battle of Tarichaeae.
All illustrations are published by courtesy of the lsrael Antiquities Authority

(IAA) and authors mentioned in corresponding notes.
1.Bronze coin of Tiberias, 120 CE, rev. showing galley with ram and 4 oars.
See: A. Ben Eli, fig. 46 (note 22).
2.Graffito of Hasmonean Warship from Jason's Tomb, Jerusalem, I * c. BCE
(note 23).
3.Drawing of merchantman from the Holy Sepulchre, Jerusalem, 1"-2"d c.
CE.
Source: S. Gibson, 1994 (note 24).
4.Graffito of boats from Herod's Northern Palace at Masada.
See E. Netzer, 1991 (note 27).
5.Marine scene with two merchantmen of the mosaic floor in Lod (Lydda), 3rd
to 4'h C. CE (note 28).
See also the recent publication by R. Talgam, "Mosaics in lsrael in the Light
of Recent Discoveries", in: Qadmoniot, XXX1/2 (116), 1998, pp. 74-89 (in
Hebrew).
6.Mosaic with boat from Migdal (Magdala), Sea of Galilee, I * c. CE.
(cf. A. Raban, note 29).
NADAV KASMTAN TROPlS Vt
Fig. 2
Fig. 3
328
3NNS31Vd NVWOtl CINV 311.SIN3713H NI 03A133U3d SVM 1
1 MOH
: ~ O B W A aNv
S unwtl sv ~ I H S~ H L
PRELIMINARY REPORT ON THE REPRESENTATION
OF SOME POST-BYZANTINE SHIPS FOUND IN A CISTERN SOUTH
OF THE TRlGONlAN TOWER IN THESSALONIKI
In the summer of 1971 the Municipality of Thessaloniki wanted to

landscape the site in front of the first square tower south of the Trigonion
Tower. The 9th Ephorate of Byzantine Antiquities carried out an exploratory
excavation within and around the tower. The ground floor of the tower was
found to be occupied by a chapel with an arch spanning the width of the
eastern wall. The apse was decorated with a fresco, probably depicting
Mother Mary praying. Within and outside the floor of the chapel there were
graves. We investigated twelve of them. The four western ones contained
only skulls (around 50), which bore wounds mainly from swords. The chapel
and graves date from the mid Byzantine period.
A new landscaping project, in 1994, that was considered necessary for

the site, where the gthEphorate for Byzantine Antiquities were carrying out
restoration works at a large scale, offered the opportunity for further
investigation south of the Trigonion Tower, below the more modern anti-
sliding wall. As one can see from old photographs of the beginning of the
century, a mount of earth used to be there, that had never been investigated.
The mount had been created by the accumulation of earth carried over
between the wall, the Tower, and the anti-sliding wall. As soon as the earth
started being shifted, simple plastered walls appeared. This is how the
greatest part of the small chapel 6.00 x 2.00m. was revealed. Its apse, 1.20
m. across and 0.50m. deep, spans the width of the wall. At the center of the
apse there was a portion of a smooth pillar embedded in the floor, which at
this point was 0.30m. higher than over the rest of the chapel. The walls were
constructed through slow drying masonry work, using earth as connecting
material. At a distance of 1.50m. from the eastern wall of the chapel, there
was a shallow groove all along where the presbytery barrier had been. This
is supported by the figures, unfortunately preserved only from the waist
downwards, which decorated the chapel . In what we consider to be the Holy
Sanctuary there were two archbishops and the rest of the chapel was
decorated by military saints. The apse was also decorated with a fresco, but
the depiction is today so faint as to be unintelligible. The chapel, according
to the style of the preserved depictions, is dated in the mid 15thcentury.
The earth fill of the chapel revealed fragments of wall frescoes, depicting
clothing items, and parts of nude bodies or faces. One such fragment
preserves a figure from the eyes downwards, which probably represents
Christ. Among the fresco fragments, there were many extracts from
inscriptions, which were either parts of particular pictures, or elitaria.
The typology characterizing the illustrations of most of the revealed

fragments, especially those with Christ's figure, leads to dating them from
the last quarter of the 13thcentury. It's impossible to know what happened
to this small chapel. It is not possible for the wall fresco fragments to have
belonged to this building. The fragments were part of a large scale depiction.
Where was the church the frescoes belonged to? Could there possibly be a
larger building, below the anti-sliding wall below the Trigonian Tower?
Maybe the larger church was destroyed in the beginning of the 15thcentury,
and the vestry was turned into a chapel. Another explanation could be, that
debris from monumental ruins was carried over and placed here, which is
rather improbable.
To the south of this chapel three cisterns were revealed. The western one
2.00 x 1.60 and +0.50m. deep was built on a strong stone support, its walls
constructed using slow drying masonry work. The interior wall surfaces were
covered with strong hydraulic mortar, which had been worked well, into a
smooth even surface. The second cistern in the center, was similar in
construction, but smaller.
To the east of these small cisterns, at a distance of +0.60m, a third

cistern was revealed 2.80 x 2.50m. on the interior and +1.20m. deep. Its
walls were 0.60m. wide, built using slow drying masonry work, with strong
connecting mortar. Internally the walls are plastered with hydraulic mortar
just like with the other two cisterns. However, along the four side edges, in
order to offer maximum impermeability to the cistern, extra mortar was
added to the full edge length, projecting interiorly. At the center of the cistern
there is a square hollow 0.50 x 0.50m., +0.25m. deep.
This hollow was made so that any solid particles contained in the water,
could sink there, and be more easily removed. At the top of the side walls,
where they would be connected to each other, the remnants of the quadro-
spheres, that would be supporting the low cupola covering the cistern, can
be discerned. Pipes started from these cisterns - especially the large one -
OF SOME POST-BYZANTINE SHIPS FOUND IN A CISTERN SOUTH OF THE
TRlGONlAN TOWER IN THESSALONIKI
and traveled in a south-western direction; they were possibly connected with

other cisterns situated at short distances from each other, all over the Upper
City. This system of cisterns and pipes supplied the city with water. The
cisterns date from the early Christian period but they seem to have been in
use for a long time. At a certain point they were abandoned, possibly
because their ceilings collapsed. That is probably when the ships were
drawn, by one or more individuals, who were privileged to be in a position to
have a view of Thermaikos Bay, where the ships moored. It seems that
shortly after the ships were drawn, the cisterns were filled in to create the
enbankment, to support the chapel construction. This landfill helped to
preserve the drawings, which were made using a dye, which was very
susceptible to weather conditions.
In the north part of this cistern on a surface 2.40m wide and 1.40m high,
ships No 1 and 2 are drawn, and an unclear sketch of what is probably ship
No 3 on the north-west part of the side surface.
Ship 1
It's a rowing vessel of the galley family (if it dates from the 14thcentury).
However, the number of oars (15 in total) indicates that this vessel is smaller
than a galley; maybe it is a <delucca>> or a <<londra>> of the eastern
Mediterranean, as at least one mast with a <<Latini>~ antenna, is clearly
discernible, and there is possibly another one on the bow. The deck
projection on the stern, that all these vessels used to have, is also clearly
discernible. Another interesting feature is that of the three lines on the stern,
which seem to indicate that the vessel had two side - rudders instead of one
central stern rudder. This is a piece of evidence indicating a rather earlier
period since such kind of steering was abandoned in the Mediterranean
from the mid 14mcentury onwards; by the end of the 15" century the use of
one central stern rudder was fully established.
Ship 2
This is the depiction of a vessel with indications that it used both oars
and sails (possibly two-masted). The originality of the perspective of this
design is quite an unusual feature on either depicted ships or graffiti.
Perspective drawings are not common.
Ship 3
Unfortunately, the elements preserved from the sketch are not sufficient
to define the type of vessel that is drawn on the eastern part of the north side.
The depiction of ship No 4 (0.80 m wide and 0.54 m high) dominates the east
cistern side, 2.65 m x 1.40 m. To the east of this vessel there are two smaller
ships (0.20 m and 0.30 m) ; both of them have superstructures similar to
those of ships No 5 and 6.
Ship 4
In contrast to ship No 1, this vessel has features that classify it in the
family of boats with a wide hull and several superstructures. The part of the
bow preserved shows a convex post (0.20 m) with a characteristic curve
(0.10 m) at the end of the rail gunwale. There is an extensive superstructure
that starts from the middle mast, and goes all the way to beyond the stern,
where it comes to a sharp end, while on another small region of the stern,
there seems to be another higher superstructure. There are also discernible
decorations on the superstructures (complex line compositions) as well as a
strake that probably indicates there was an open lower deck on the vessel.
Another interesting feature is the pronounced drawing of the stern
hollow, which probably indicates the position of the rudder, within the vessel
where it would join onto the steering system. Hardly any of the sailing
elements have been preserved; there seem to be two masts, one baw
spright on the stern and some riggings.
The fact that there is no extensive forecastle, while the stern
superstructures are quite long (0.42m) but rather shorter than in other ship
depictions, as well as the concave indentation on the stern for receiving the
rudder and the curved stern post leads us to comparing this vessel with the
early Mediterranean galleons. It is also known that the first attempts for the
construction of galleons took place in Venice during the first half of the 16th
century, while later on the construction of these vessels was continued by
the Spaniards, the Portuguese, as well as other Northern European peoples
(the English, the Dutch, the French). Finally, on the west side, the
dimensions of which are identical to those of its symmetrical counterpart on
the east (2.40m wide and 1.40m high), three more ships are preserved.
Ship 7
This vessel has lines on the stern that seem to create a superstructure
which, however, is disproportionate to the rest of its dimensions. It also
seems that the stern was designed to bear sails, which nevertheless would
not provide enough power for such a vessel.
Ship 8
This is an unfinished draft of a small vessel with a superstructure on the
stern.
Ship 9
Another rowing vessel, a round ((nave tonda,,. Ships of this type carried
more voluminous cargo. There are two masts and the two lines projecting
from the bow are obviously either riggings or observatory towers. Finally, it
is unclear whether the lines on the stern are an attempt to depict the plating
of the folds or a folded down sail .
*We would like to thank Mr. Kostas Damianides for all his counsels about
naval architecture that he kindly offered us.
Vassilis Koniordos & Elli Pelekanidou

Charalambous Moushou 26a
54 634
Thessaloniki
BIBAIOrPA@IA
r.r o ~ v a p qecTa
, T&imq q @&~~ahOviKqCp
K. M&p?$Oq, <<Mvqp&ia M Q K E ~ O VlCITopiClq,
L K ~ ~ ~ @&CY~ahoviKIl 1947
M. XaT<IlloawO~,<<Amuypacpia @&~aahOViKqq~ @ E U U ~ ~ O V 1880
~K~
0. Tafrali, Topographie de Thessalonique, Paris 1913
I. Bau6pap&Mq, 'Imoputa A p x ~ i aMaK&60viaq>> Topoq A'.
Ah. A h a , 'lmopia q q 8eaaahovi~qq~~ O ~ a o a h o v i ~1961
q
Av. Ophavi5op, <c Ta ukK6 60pfiq ~ o Apxaiov
v EMfivov>>Topoq B' ABfiva 1960
K.A. Ba~ahonouhou,'H M ~ K E ~ara o v1715,,
~ ~ M Q K E ~ O V1L1,1971
K~
B. Aqpq~pta6qq,~TonoypacpiaTIlq e&~CJahOViKqq KaTa q V E n O ~ f TIIS
i T o u ~ K o K ~ ~ (1430-
T~~C,
1912)rn @ E U U ~ ~ O V 1983
~K~
M. E. Cousinery, <<Voyage das la Macedoinem Paris 1831
z ~ f i p q Kiaaa, ~ O ~ u p o p a&pya ~ ~ TOu

~ a U O U ~ T ~ V O UBayla<i~Drq @&~CrahOviKIl:~ O T O ~ L K ~ ~
n p o a & y y lT~ L~ ~~ ~ T Topoq
L K O ~ a e q yN.
. Mou~aonouhou
J. M. Spieser, <<Thessalonique et ses monuments du IV au VI sieclen Athenes 1984
r. Z ~ O y d y h o<<H
~ , &v eE~aah0ViKqnaTplapx1Kfi pO\rilTov BhaTa6ovrn @&aaahOvi~q 1971
M. Vickers, Further Observations on the chronology of the walls of Thessaloniki, Makedonika 12
(1972) p,230
Ch. Bakirtzis D' une porte inconnue des ramparts occidentaux de Thessalonique, Balkan
Studies 14 (1973) p.306
Ship no. 1
Ship no. 2
no. 7
Ship no. 8
KONlOROOSV. & ELEK4NIDOUE TROPIS VI
The Trigonian Tower with the excavation area

and the cisterns protected by roofing
7
The cistern with the majority of the depictions
TRSGONIAN f OWER IN THESSALONlKF
I --.=--- - '*
Plan of the Acropolis of ~hksaloniki
Plan of the Excavation site

with the two cisterns
SUMMARY
THE SHIP OF THERA

A DIFFERENT INTERPRETATION, ON THE CONSTRUCTION AND
UTILITY OF THE STERN APPENDAGE
1. This presentation refers exclusively to the type of "paddle-propulsed"

ships of Thera, which have the Stern Appendage as their most significant
common feature.
2. Contrary to the wooded mainland Greek environment, (with its

numerous tranquil lakes and protected bays, which had been particularly
propitious for the use/evolution of dugouts), the semi-dry environment of the
Aegean islands, with slim vegetation and rough seas, had been the most
appropriate for a different steadier floating-craft, constructed of locally
available logs tied together, i.e. the raft.
3. The bad hydrodynamics, the inadequate wave-protection and the

limited displacement of the flat raft gradually led to the crosswise and
lengthwise curve of its structure until it shaped a vessel that solved these
specific problems.
4. A vessel constructed in this way (rope-lashedlsemi-rigid),would have

the physical tendency to inside-collapse, under the outerlhydrostatic
pressure. That requests a flexible support with enough elasticity for the shell
of the vessel. In this case, the stern appendage seems to constitute a point
of tightening of the lateral arched beams which pre-stretches exteriorly and
upwards the structure of the vessel.
5. The rather inexistent water-tightness of such a vessel could not have

been confronted with additionallsolid watertight materials. The solution was
given by an external flexible cover made by leather of water-proof cloth
(resin-coated).
6. Using as a base the anthropometrical equivalent, which results from

the proportions of the "most important passengers of the ship" (officials and
officers but not rowers), we are in position to re-estimate the dimensions and
the size of the ship, being smaller than it was believed until today.
APOSTOLOS KOURTIS TROPlS VI
7. The above-described method of construction, evolutionarily allows the

adoption of a lighter structure with obvious profit, the smaller displacement
in relation to its dimensions. Thus, the supposition that the "ship of Thera"
had the ability to carry out inter-island voyages across the Aegean Sea by
paddles becomes more realistic.
Captain Apostolos Kourtis - HCG

Canaris 2
Holargos 15 562
Athens
LONGBOATS AND TUNA FISHING IN EARLY CYCLADIC PERIOD:
A SUGGESTION*
Every scholar dealing with the problems concerning the Early Cycladic
Period, and especially the Grotta-Pelos and Keros-Syros cultures (ECP), is
faced with a serious disadvantage. Traces of this civilization have been
preserved almost exclusively through the cemeteries of the period. The
evidence about the settlements of this era is insufficient due to a large
number of factors: scarcity of sites; perishable materials of construction;
later occupation in the same sites blocking the expansion of in-depth
research; overgrowth of building activities during the last three decades,
resulting in rescue excavations in restricted properties; poor documentation
and unpublished material.
If someone intends to study the fishing activities of this period the

evidence is extremely limited. Later periods with pictorial art offer preserved
piscatorial depictions on pottery, frescoes or seal-stones. In the case of ECP,
the only direct evidence that could be used comes from a reference to two
fishing hooks, published by Tsountas in the late lgthcentury1.The collection
of fish remains from archaeological context is a relatively recent concern, as
they were not appreciated as a source of information until the late 60s'. In
order to estimate the nature and the extent of fishing activities we are obliged
to study the following periods of the Bronze Age and, namely, the directly
preceding period, hoping that we might in this way bridge the gap of absent
or lost evidence. In the last case the only existing and most valuable source
of information is a Neolithic settlement excavated by Evans and Renfrew in
Saliagos, a small islet between Paros and Antiparos, in the heart of the
Cyclades3.
Studying the bone material of the above-mentioned publication, we

come across a unique phenomenon: huge quantities of fish bones have
been discovered on the islet and their thorough examination has lead to an
astonishing knowledge of the every day nutritional habits in the site4(table I).
A first observation is that nearly all the bones found belonged to very large
fish. Secondly, 97% of all the identified fish bones belong to tunas, often of
very large size, weighing over 135 kilograms5.The counting that has been
made for statistical reasons in Pit A, Cliff 17 and in Square N3 allows us a
comparison between animal bones and fishbones and a 'translation' of
these sizes in weight of consumable meats.
GEORGE KOUTSOUFLAKIS TROPlS VI
In Pit A, tuna represents 88% of the total meat quantity while in Square
N3 28%. These numbers appear amazingly large. Some further estimations
make it possible to understand what they represent in absolute sizes. In Pit
A, 1930 tuna vertebrae were collected. Calculating 39 vertebrae per fish
gives us a minimum of 48 tunas. If we multiply this number by 135 - number
that has been estimated as the average tuna weight in the site - we get a
total of 6.500 kilos of meat. In the same pit the sheep and goats are
represented with 490 kilos, the bovines with 210 and the pigs with 180. In
Square number 3, the situation seems to be different although the fact that
the tunas represent the second greatest source of meat should not be
underestimated. To understand better these quantities, I report that the total
catch of tunas in Greece in the year 1936 was not more than 400 metric tons,
only seventy times the lowest quantity by estimation discovered in Pit A
alone, at Saliagos7.
The first undoubted conclusion concerning these statistical data, is the

primary importance of tuna fishing in the Saliagos diet plan and economf.
It is remarkable, therefore, that nothing in the cultural assemblage can be
with certainty associated with fishing: not a single fishing hook has been
found on the isletg.This fact leads to the question of how the large tunas
were captured.
The percentage of tunas in the total fish presence indicates fishing on a

large scale. Typical tuna behavior is migration in large shoals. Tuna is an
open water fish. Today it approaches the coastlines of the Aegean area in
the period between February and June, appearing also in smaller numbers
in late summer and autumn, but it may well be that the migration pattern of
tunas was very different six thousand years ago from the present one. The
traditional practice of tuna fishing until the first half of the 20thcentury was
fishing by means of enclosure (nets fixed or movable). Tuna fishing from the
shore, using line and hook is not customarylO.Furthermore it seems
improbable for the period under consideration, taking into account the
technology available. It is difficult to imagine that a bone-made hook and a
line of that era would be strong enough to resist the fight of a 135-kilo tuna.
In addition, the adoption of such a method would not explain the large-scale
capture1' reported in Saliagos.
The most characteristic ways of tuna fishing in the Mediterranean12are

twofold: the first and most common is called 'Tonnara' by the Italian
fishermen. Tonnara is a permanent fish-catching net installation. In a
developed form, it can reach a length of more than two miles and it is set
A SUGGESTION
vertically to the coastline. One end of the net is attached to the shore and the
other end is moored on the sea floor. In the middle there is a corridor
connected with another square net, the so-called 'camera de la morte' (room
of death). It is the place where the tuna flocks are enclosed and then killed.
Tonnara has a strong tradition in the Mediterranean. Oppian, a later Greek
writer of the Pd century AD, describes in one of his books these fishing
installations13:
'First of all the fishers mark a place in the sea which is neither too
straitened under beetling banks nor too open sky and shady coverts. There
first a skilful Tunny-watcher ascends a steep high hill, who remarks the
various shoals, their kind and size, and informs his comrades. Then
straightway all the nets are set forth in the waves like a city, and the net has
its gate-warders and gates withal and inner courts. And swiftly the Tunnies
speed on the line, like ranks of men marching tribe by tribe - these younger,
those older, those in the mid season of their age. Without end they pour within
the nets, so long as they desire and as the net can receive the throng of them;
and rich and excellent is the spoil"4.
The second practice followed in tuna fishing is a kind of movable net.

When the alarm by the tuna-watcher1' is given, people run to the shore and
pull the boats to the sea. The boats are divided into two or four squadrons
and they extend their nets in the sea, trying to surround the tuna flock. If they
succeed, they pull slowly the net towards the coast. In the shallow waters,
the fishermen jump in the net circle and kill the tunas with spears and clubs1".
This way of fishing - with the involvement of boats - has also a very old
tradition in the Mediterranean" and is attested to by the ancient writers1".
In this second case of movable nets, the role of the boats is very
important since tunas can easily escape to the open sea. This is why the
boats that are involved are lengthy and narrow, with a good number of
oarsmen, in order to reach high speed, with an almost flat keel, so that they
float into shallow waters, and a low profile, so that they have resistance
against the wind. Generally they are not heavy constructions, so that they
can be easily dragged in or out of the sea, only by their own crew. In other
words they are boats of high operational readiness1'.
But also in the case of the previously-described 'Tonnara', boats are

often used to hunt tunas and drive them to the net. These are, in general, the
two possible ways of large-scale tuna fishing. In the case of Neolithic
Saliagos, it would be an exaggeration to infer permanent fishnet installations
in the form of 'Tonnara'. It would be more plausible to imagine that the tunas
may have been driven into shallow water by a series of boats and nets, and
then caught by clubbing or shooting with arrows, tipped with obsidian
points. The unusually shallow and narrow configuration of the bays at
Saliagos, both north and south of the isthmus, may have been particularly
favorable to this kind of fishing. A large number of obsidian points have been
found in a Saliagos settlement and their use for hunting non domesticated
animals is not justified at least under the light of the bone material found2'.
The complete lack of hooks - which are known to have existed in the
Aegean as early as the Early Neolithic period - in the islet, seems to mean
that in the fourth millennium BC in the Cyclades, as far as the capture of big
fishes is concerned, the hook was already technologically surpassed by
more developed means of fishing".
This is the evidence concerning the fishing activities in Neolithic

Saliagos. The annual catch should have been - to an extent - an event of
crucial importance. Something more than a thousand years separate the
Saliagos culture from the Early Cycladic II Period, in which longboats with a
fish depiction on one end are for the first time portrayed. As it has been
already noted, no direct archaeological evidence for tuna fishing exists in
ECP". And this is where a question of archaeological deontology arises: are
we allowed to take tuna fishing in the Early Cycladic Period for granted, since
no traces of this activity appear in the archaeological record?
In general, the transition from the Neolithic to the Early Cycladic Age,
seems to be nonviolent and the culture level is not seriously interrupted. The
knowledge of a vital survival activity for the community can not be easily lost.
There is no reason why the inhabitants of the Cyclades would overlook or
ignore the obvious advantages of such an activity. Scholars like Bintliff paid
great attention to the study of environmental factors of the prehistoric
settlements in the Cyclades and it is remarkable that, of the vast number of
small inlets around the coast, in islands like Melos or Mykonos, just those
with prehistoric finds mainly of ECP, coincide with the migratory fish run
locations. Many sites on promontories of the coastline have been interpreted
as temporary camps from which the approach of the tunas in adjacent bays
could be observed. These bays are key tuna grounds and the sites are little
more than scatters of obsidian and flint 'fish points', like those of the
Saliagos culturez3. No matter how plausible this theoretical approach
appears to be, the tunas compelled an economic interest for the food they
represent - producing quantities and qualities throughout ancient historyz4.
Many later authors expound at length on tunas' multitude, migrations, habits
A SUGGESTION
and size. The economic value of tuna as a food source, then and now, finds
ample recognition by writers such as Aristotle and Apostolides, separated by
over two thousand years, and in regard to its consideration as the 'Manna of
the MediterraneadZ5.Tuna migrations are predictable, repeated every year
and following the same passages. This means plenty of food supply and a
diet enriched with proteins. Furthermore it is a food supply that is not
affected by weather conditions, which can destroy an agricultural
production, or diseases, which can kill the livestock of a community.
Longboats in the Early Cycladic Period
The corpus of Early Cycladic Period longboat representations is not

large. Thirteen two-dimensional depictions are incised on the frying pan
vessels26(fig. 1) and other four lead models are known from the island of
Naxos2' (fig. 2). TO describe briefly the first group, they are long and
undoubtedly narrow vessels with many oars or paddles. They have a high
and a low end and their shape is defined by abrupt changes of inclination
and angles2'. In the extension of the high end there is a fish depiction on an
antenna. In all cases, one boat is depicted on every frying pan, except on n.
27, where two boats are depicted. These two boats, as well as boat number
54, have no oars. The lead models group presents general similarities with
the frying pan group. The differentiation consists of more gentle angles in
prow and stern and the absence - perhaps for technical reasons - of the
oars, the fish and the spur. This group provides us with significant
information concerning the proportions of these vessels. In the best
preserved example the beam-length ratio is 1:12.1. If we accept a minimum
beam of 1.5 meters, this should suggest a length of about 19 meters. If we
estimate the length of the frying pan vessel through the numbers of paddles
depicted, we have almost the same size. Taking into account the smallest
longboat, we estimate a number of 24 + 1 rowers, i.e. 25 persons for the
propulsion of just one boat - if we interpret correctly the depictions, and this
number is just the minimum2*.
One thing is certain: in any case a good number of rowers was involved
for the propulsion of these vessels, a fact which has provoked much
discussion in relation to the use and the final destination of these boats. In a
recent article C. Broodbank makes some very interesting remarks relating to
the demographic situation of Early Cycladic II settlements and the number of
men needed to crew the longboats30.
The whole analysis is based on the hypothesis that the number of

inhumations in the early cycladic cemeteries are reliable indicators of the
approximate size of the relevant settlements. Taking as a fact the ratio of one
cemetery for every settlement and one burial for every grave, Broodbank
tries to calculate the population of the settlements by counting the graves in
the known excavated cemeteries of the Early Cycladic II Period. This number
corresponds to an exact period of time, going from 100 years, which is the
least a settlement can survive, to 400 years, which is the entire span of the
Keros-Syros culture. He counts 5 persons for every family and 25 years for
every generation. In table 2 we can see the results of this demographic
approach, which of course does not have an absolute character, but gives
us a general idea of how large these settlements might have been.
Thus, for example, in the Krassades cemetery we have 50 burials, which

means uninterrupted habitation of 13 people if the settlement's duration was
100 years, and of only 3 persons if the duration was 400 years. In
Chalandriani of Syros, which is the largest cemetery with about 600 graves,
we have a number of 150 inhabitants for 100 years, 75 inhabitants for 200
years, 50 for 300 years and only 38 for 400 years. Assuming that paddling
was exclusively a male activity, we have to abstract in these numbers the half
which represents the female population and another quarter for the old or
sub-adult persons. By looking at this table it is clear that even in the most
optimistic case of population distribution, the majority of the communities
were not able to crew their own longboat. The only settlement on this table
that is able to support a longboat is Chaladriani, if we accept a settlement
duration of not more than 200 years. Even in this case, the community would
have to lose a very important part of its labor potential. Somehow we have
to admit that if this was happening, it was happening very rarely, for a very
short period of time and for specific reasons.
As Broodbank points out: '... the idea that the longboat was in any sense
a common and normal phenomenon in the Cyclades should give way to an
acknowledgement that it must have been an unusual and highly specific
development'. If so, we have to clear up its function and its use as a boat.
We are accustomed from Tsountas' time to consider the longboats as the
key for the understanding of inter-regional exchange in the Early Cycladic II
Period. In part this seems to derive from the very fact that the evidence of
longboats and the rise of trade appear simultaneously in the archaeological
record.
To come closer to the question of the longboat's function we have to

A SUGGESTION
focus again on its design. It is surely capable to transport low bulk goods
such as marble, metals or obsidian, although its use for such purposes
seems to be excessive. This kind of material could be transported easily with
smaller boats and less crew. On the other side, longboats are hardly
appropriate for the transportation of large cargoes because the space
needed for paddlers would decrease in direct proportion to the quantity of
cargo loaded. Transportation and trade activities could be served much
better with boats propelled by sails. The fact that no sails are reported in ECP
does not necessarily mean that they didn't existed. This is something that
has to be estimated as 'no evidence' rather than 'negative evidence'. Sails
could hardly be present anyway in the three-dimensional lead or clay ship
models and as for their absence in the 'frying pan' incisions this could be
interpreted in two different ways: either sails didn't exist indeed or the type
of ship pictured had to be for some reason propelled by oars.
If the commercial use of the longboat is rejected, the two other functions
a boat can have are mainly for raiding activities and fishing. The possibility
that longboats might have been used for raiding has been expressed in the
past and Renfrew estimates this function as very possible3'. Indeed
longboats combine high speed with transportation of a good number of
paddlers-warriors. We have many examples of similar use of the many
paddled boats from primitive societies in the South Pacific. In the case of
Maori and Mortlock tribes, although they have many different boats for
activities as trade, personnel transportation and fishing, the many-paddled
canoe is used only for offensive tasks. This theory is very attractive although
it is not very proper to relate civilizations that lie chronologically and
geographically so much apart. Nevertheless the concept we have for the
culture of the Keros-Syros period doesn't justify this function. A function like
that could be well related with the period (EC IIIA) directly following when, by
all indications, the trouble in the Aegean islands started. In regard to the EC
II Period, the increased commercial exchanges, the transportation of raw
materials as well as the location of many settlements in the coastline give us
mostly an impression of free transference for both people and goods.
However, we can not exclude the possibility that trading and raiding could
form parts of the same system.
The increased need for a longboat in crew can only be justified in the
case of an activity that requires the almost total participation of the labor
potential for a small period of time. The existence of a boat requiring a large
number of paddlers - in the social context that has already been described
- creates the impression that the purpose of this kind of boat would be to
GEORGE KOUTSOUFLAKIS TROPIS VI
obtain maximum speed for a very short period of time. Tuna fishing has
these requirements: once the tunas appeared in the area, the islanders knew
that they would not stay forever. Additionally, the islanders knew the time of
the year and perhaps also the exact place. The use of sails in such a venture
would be highly inappropriate, and this might be a good reason for their
omission in the 'frying pan' vessels.
A first indication for a piscatorial use of the longboat could be considered

the depiction of the fish which stands on the highest end of the vessel,
somehow attached on an antenna or spear, repeated in all thirteen examples
with religious devotion (fig 3). The oars or the spur are omitted sometimes,
but the fish never. It should have been an important symbol of the boat -
beyond any constructional significance, a symbol understood by the society
- and the potters - on its own terms. Its meaning as a totem or symbol can
most easily be explained in piscatorial terms3'. Perhaps it was standing there
in order to bring good luck to the fishing expeditions as an image of the
desired prey, or it could be a demonstration of the fishermen's belief that the
figure, which possessed a prominent position in the boat, could somehow
exercise control over the flocks. If this argument is pushed a bit further it
might also be said that in shape most of the fishes incised have the typical
tuna features (strong outlines, compact bodies and extended dorsal or
coastal fin)33.Even the rows in the back of fishes 25 and 13 remind us of the
lines of the tuna species Lineatus, Alleteratus and Affinis.
The hanging 'banner' together with the supporting antenna could

represent the means of capture and final death, the net and the spear,
although there is no further evidence to support a view like that. Fish and
hanging 'banner' might be depicted once more together, some centuries
later, on the Phaistos disk, before they both disappear as ship emblems34
(fig. 4).
Special mention should be made of n. 27, the only example presenting

two boats in a single 'frying pan' vessel. In this depiction there are no lines
representing paddlers, instead there is a row of stamped triangles on each
deck and a zigzag design both on and above the crafts. Broodbank favored
the idea that the doubling of motifs may be read not as a numerical
duplication but as a doubling of the power and attribute of the image3';
however, the portrayal of two boats in a somehow narrative scene should not
be rejected a priori. A naval battle in the EBA Aegean is completely out of
question. We might see two boats in converging course, cooperating in a
nautical activity. It seems rather impossible that the stamped triangles
A SUGGESTION
represented paddles, as they are not present in either side of the hull -
probably paddles were never intended in this depiction. Stamped triangles
could however stand for men on board. If so, we could summarize the
situation depicted as follows: two boats not far from one another with the
paddles raised and all men on board. Let us think that in most cases of
ancient and modern references to tuna fishing, many boats seem to
cooperate in the venture. An extra explanation should be given for the zigzag
design on the hulls: could it be nets that are raised out of the water by
standing men on deck, as it is seen on fig. 5 (tuna fishermen from Sicily
1947)? This would explain the absence of paddles and the presence of men
'on deck'.
In nautical archaeology we are used to separate the ships of specific

functions into two groups, war ships and trading vessels, often ignoring that
also a third group, fishing boats, existed. When I first started relating
longboats to piscatorial activities, I was convinced that they were boats of
specific function and not boats for all purposes. For the collection of the few
indications, I had to take huge chronological steps from the Neolithic period
to modern times. I realize that these scattered elements of information can
not fill up our ignorance for Early Cycladic Period settlements and their
activities. Under these circumstances the piscatorial use of the longboat can
neither be taken as obvious nor self-evident. But it should certainly be
considered as an open matter to discussion.
G. B. Koutsouflakis
Ag. Paraskevis 60
Athens
NOTES
* I have received valuable advice during various stages of work on this paper. My thanks go to
H. Spondylis, from the Department of Underwater Antiquities, for the encouragement and the
information concerning the traditional ways of tuna fishing in Turkey, to Th. Webb, from the
Hellenic Institute of Underwater Archaeological Research, for his many constructive comments,
and also to colleagues D. Mytilineou and P. Micha, for their advice and recommendations on
the final form of this manuscript. Also to H.E. Tzalas for the opportunity to present this paper
although my participation came after the deadlines set by the organizing committee. Drawings
and tables presented are based on publications referred to in the list of illustrations.
Abbreviations of the archaeological bibliography are the established ones.
1. Chr. Tsountas, 'Kykladika', Arch. Ephemeris 1899, 104, PI. 10, nos. 38-39. From the island of
Syros (National Archaeological Museum, no. 5209).
2. However some fish remains are reported in some of the old publications: in Troy II (R. Vichow
in Ilios, p. 360, Troja, p. 165) and Troy V (large vertebrae possibly of tuna and shark, Troy II,
part I, p. 158, 268). Hadjidakis reported 'fish bones' from the Middle Minoan levels of Tylissos
GEORGE KOUTSOUFLAKIS
(Arch. Ephem. 1912, 232-233), Marinatos from the grave at Krasi (Deltion 12, 1929, 133) and
Keramopoulos burned vertebrae from the Late Helladic Ill palace of Kadmos in Thebes ( 3,
1917, 179). Furthermore Tsountas and Wace have found fishbones in the houses and the
citadel of Mycenae (Mycenae, p. 106, n. 4).
3. J.D. Evans-C. Renfrew, Excavations at Saliagos, BSA Suppl. Vol. 5 (1968), from now on
'Evans-Renfrew 1968'.
4. J.M. Renfrew-P.H. Greenwood-P.J. Whitehead, 'The Fish-Bones', in Evans-Renfrew 1968,
Appendix VIII, 118-121.
5. This weight seems reasonable although in some cases the weight of the largest fishes has
been overestimated by the experts. The equation of a five feet tuna with the weight of 800 Ib.,
presented on page 119, is unrealistic. Calculations of the food represented by the Saliagos
finds have been however reassessed: see M. Rose, 'Neolithic Fishing in the Aegean: New
Evidence from Franchthi Cave', paper given at the 8 7 general meeting of the Archaeological
Institute of America, abstract in AJA 90 (1986), 177.
6. Evans-Renfrew 1968, 'Enviroment and Life', 77-81, table 23.
7. Of course the Saliagos quantity represents an accumulation over a wide period of time -
however the number remains impressive taking into consideration that only a limited part of
the real tuna-catches survived through buried fishbones.
8. However it should be noted that the case of Saliagos can stand only for itself: some miles to
the north, in the island of Keos, a late Neolithic sea-side settlement and cemetery was
excavated by modern methods (J.E. Coleman, Keos I, Kephala, Princeton 1977), resulting in
largely differentiated material (J. Coy, Animal Remains, Appendix 4, 129-133). Fish bones
were almost completely absent- not only due to the bad state of preservation. It is difficult to
decide which from the two sides forms the 'exceptional' case.
9. Despite the two bronze EC II hooks reported from Syros, large numbers of bone-made hooks
have been discovered in Neolithic stratums of a cave, in the island of Gioura in the North
Aegean (I am very grateful to Dr. A. Sampson for showing me this material and discussing it
prior to its publication).
10. Although it is reported in ancient times, see Aelian, On the nature of animals 13.3: <<.. .
BKOUU62 KEAToL?~ ~ a MauuaA[Li)Taq
i ...BYK impolq TO& eljvouq Brjp(iv~It is interesting
that Platon in commenting the piscatorial activities distinguishes 'sea hunting' (nepi 86ha~-
Tav Bqpa) from hook and line fishing ' a y ~ ~ m p e i(Lores a' VII 823 d-e).
11. Another case of massive tuna capture is reported by Pausanias (X, 9, 3) during his visit to
the sanctuary at Delphi, where he refers to the bronze bull statue dedicated by the people of
Corfu in response to the gods' advice for the capture of a large number of tunas (for the votive
monument see CI. Vatin, 'Monuments votifs de Delphes', BCH 105 (1981), 440-449, P.
Amandry, 'Notes de topographie et d'architecture delphiques: Ill, Le Taureau de Corcyre',
BCH 74 (1954), 20-21). Another bronze bull for the same purpose was dedicated in the
sanctuary of Olympia (Pausanias V, 27, 9), both votives being financed by the d & ~ a r(1qI ~
10) of what seems to be an extremely successful catch. Pausanias' text however doesn't give
any further information about the way of capture: << ...oi 6& (2A~ivyap TOUC Buvvouq npoal-
WUC
P O U ~ E V O I~ r j v T a ~ ~ l I 7 q IEi a
~ ~O V~ E)O P O U CCI~OTBMOUULV 6~ A~Acpou~. Kai OUTU
f l 0 0 ~ [ 6 0 v iTE ~ K E ~ V O eVd 0 ~ U lTOV Taop0V Kai CldTiKa ~ E T UT ~ ~V U U ~ C(lip0oUl ~V TOUC
kBaq, ~ aucpiui i TO (ivaeqpa Cv 'OAupniq TE ~ aCV i AcAcpoiq Bmiv r j ~ E K U T~ ~i j qiiypaqn .
Pausanias uses the word iiypa to describe the capture, while the monument's votive
inscription refers to the Brjpa of tunas.
12. For an exhaustive monograph on the subject see Paulus Rhode, Thynnorum Captura,
Lipsiae 1890. Especially for tuna fishing in the Greek archipelagus see M. Athanassopoulos,
'Sur les Thonnides en Grece', Comptes Rendus de I'Academie des Scienes, vol. 177 (1923),
501-2 and by the same author 'Note Complementaire sur les 'Thonnides' en Greee', Bull. lnst.
Oceangr. Monaco, no 440 (1924).
A SUGGESTION
13. Oppian, Halieutica 111, 633-648: .... Xdpov pCv n a p n p o ~ o v&nc@pauuav~o Bdauqq
OCTEAiq v m c ~ v ~ n o &m7pc@dcu~v
v u n ' &ealq OCTEAiq v a v d p o ~ u ~dni6popov,
v aMa ~ a i
ai'epn ~ aoKcnavoiq
i ~ ~ u e p d udvaiu~pa
lv pdrpa @dpovra. "Eve' I)'TOL n p d ~ o vpdv in'
opelov uwr ~oAwvovi i r q dnap6aiv~1Buvvou~onoq,d m c ~ r o u u a qnavroiaq aydAaq
T & K ~ L ~ @ & T ~TEI , Kai OOGC~~, lll@abU~&l 6' &Tap0101. TU 6' U U T ~ K6iKTLJa~ I ~ ~ V GUT&
T U
noA~qnpo6&6q~&vdv oi'6paulv. &v 66 nuAopoi ~LKTI&,&v 66 ncAal, puxa~orT' a M d v ~ q
E"a01V.Oi 66 BOW U & ~ O V T&lli ~ ~T ~ U C &UTE, @MayycqdV6pdV ~ ~ x o ~ ~ Kv ~WT VU @ U A ~ ~ O V .
Oip6v Eaur v o n A o ~ c p oroii ~ , 6' ciui y ~ p a i r ~ p ooil ,6' dvi p i u q a n , anc1pdu~0166 Aivuv
EVTOUOE~ ~ O U O E~UOKE
E"<oxoq ibTam1 aypq ... .
I V , v ipc@uul ~ aa iypopdvouq ~iVdAli'T~1 ~ ~ K T U OaV@ v ~ l f 66
14. Oppian-Colluthus-Tryphiodorus,translation by A.W. Mair, Loeb Classical Library, 1928, 398-

i ~ a i
401.
15. Aelian, On the Nature ofAnimals 15.5: N O u ~ o r r d qi6dv. . .Ady&r p&v roiq Orlparaiq dnoO&v
6 @ 1 ~ ~ 0 L i v. ~. a. l&p&i y& pfiv noMaK1q ~ a TOV i d v ~ ciprepov>>
a According to Plutarch
(Moralia 980 A), the tuna-watcher, Buwou~onoq,was helped in his estimations by the cubical
formation of the shoal: -... o y o ~ Buvvou~dnoq,
v Bv a ~ p r 6 d Aa60
q TOV aplepdv rflq dm@a-
v c i a ~cueuq dno@aiv&~al nouov ~ a anav i TO nAflBoq Curiv, ~ i 6 6 qor1 ~ a TO i 6a8oq
( I ~ T ~ dVVi'0p T & T ~ Y ~ & V O(VT T O ~ X n& p~ ~d ~TE TO ~ U T O C& m i KCli TO pqK0p. The lookout,
8uvvoo~on&iov, was sometimes a high mast (Philostr. /mag. i.13 ~ o ~ O n l W p &y@ i ~ an~q a@'
li@qAoU<uAom), sometimes a more elaborate platform (Ael. 15.5 N O ~ f i a v ~ o n r d v@UAUT-
TWV p M a o<u 6~60rjuaq Ady~r61Ci)~~rv i~cieel~ a TOU i n~Aayouq&~&TTEIV clj0Li. Oi 66
&<ap~l@lVT& &AaTqq
~ T ~ TOVV UKOITOV U V & X O U O ~ V T?7qd~dpaqm[one of the two ITpdjlVa
&Aarqq u@qAawhich support the platform of the Buvvoa~oneiov])orjust an elevated place
on a rocky shore (Aristoph. Eq. 312, hence metaphorically <<omlq(KAdwv) dpdv ~ a q
' A & jvuc & K K & K ~ @ W K Q60dv,/
~ KUTIO T d V I E T ~ BvWeE~ v v TO* @ O P O U ~~ U V V O U K O .~ ~ ~ V ~ ~
Some coastal elevated sites in the Aegean islands, that provided obsidian blades but no
pottery, have been interpreted by their surveyors as lookouts for tuna-watchers (J.F. Cherry-
R. Torence, 'The Earliest Prehistory of Melos', in An Island Polity: the Archaeology of
Exploitation in Melos, eds. C. Renfrew-M. Wagstaff, Cambridge 1982, 23, 26).
16. Aischylus (Persians 424) refers to the lamentable position of the Persians during the sea
battle of Salamis, saying 'they were striked with oars and broken ship-parts as if they were
tunas in a net' (Toi 6' d u r ~Buvvouq I)' TIV' ixeLjuv 6oAov / ayaiul ~ ~ n Bpaupauiv d v
T ' & p c l n i ~ v /Enaro\3. A similar lively image is given by Homer (Od. X 124). some centuries
earlier, for the companions of Odysseus, whose boats had been smashed by the thrown
rocks, and are being speared in the water like fish by the Laestrygones: ekBUq 6' k q nci-
pOVT&q dT&pITda 6aiTa @&~ovTo.~>
17. For a more recent account see C. Apostolides, La P6che en Grece, 1883, p. 31: 'Au mois
de mai plus de 20 bateaux de Spetzia, quelques-uns de Skiathos se livrent... a la p6che des
thons. Quand I'arrivee des thons dans les parages de ces iles est annoncee, les p6cheurs
font leurs preparatifs de campagne. Tous les bateaux... se placent a I'entree du golfe
d'Argolide, que les poissons traversent toujours pour penetrer dans I'interieur de ce golfe; le
p6cheurs approchent de la cBte, y jettent I'une des extremites du filet, et, en avanGant vers le
large, ils y jettent le reste. Cela fait, ils enfoncent dans I'eau une poutre et y laissent un
gardien [Buwoa~onoq]. Le bateau revient a terre en decrivant une courbe et trainant apres
lui une corde, avec laquelle, en tirant I'extremite placee du cBte de la mer, ils font decrire au
filet une ligne circulaire. AussitBt que le gardien annonce, par des signaux, a ses camarades
qu'un nombre assez considerable de thons se trouve a leur portee, ceux-ci tirent de la terre
le filet oh ils englobent les poissons'.
18. Aelian, On the nature ofAnimals 15.5: K . . . c i ~ aGrraMrjAorq ~ a i qvauuiv & p & ~ o u u uK ~~ T U
moI;)(ov i x o v ~ a TE i a M r j A ~ vdnci
, ror ~ a TO i ~ ~ K T U O&V @ ' & ~ a m q6rTjpv~al.
l ~ aI)' i yc
ITphTI) Tfiv Caurrjq &K~C~/\OUUCY p0@V TOO ~ ~ K T L ? O U ~Va)(wp&i, E ~ T U fi d&UT&pa 6po
GEORGE KOUTSOUFLAKIS TROPIS VI
TOOTO, ~ ar j rpirq,
i ~ a 6&i
i ~ a 0 ~ i vnaj vl r s ~ c j p ~ qov i, 66 rrjv nipnrqv ~ ~ ~ T T O VETIT E ~
piMouul, TO& 66 C n i T ~ U ouxprj T ~ ~a&ivairro.E ~ T Ui p i ~ ~ o u~u o~ lv i i M ~g aciyouu~ i
TOO 61~rLiou njvpoipav, rjouxa<ouul~.Philostratus Imagines i. 13: w. . .K&Vip6&o-
v~aq rouq kBOq i6g, 6oGq TE hq wyio-qq 6 ~abT@ i npoq rouq Cv roiq a ~ a r i o l q
~ a TOV
i
apl0pov A ~ Y E I ~ araq i pupla6aq aLir&iw. The earliest written attestation of a method like
that, might be suggested by Homer (Od. XXll 383-387) although nothing is being said about
the kind of fish and the involvement of boats: 'But he' (Odysseus after the slaughter of the
suitors) 'found all the sort of them fallen in their blood in the dust, like fishes that the fishermen
have drawn forth in the meshes of the net into a hollow of the beach from out the gray sea'
(TOUC66 i6cv p M a navraq Cv ai;uar~~ aK Oi V ~ ~ U/In ~ m ~ & i r anoMoLiq, q (;jC T' Mciaq, ouq
0' MI~EC / K O ~ O Vi q aiy~aAovnoAri]q EKTOU~E BaAauqq / ~IKTL~QJiS&puuav noAuun@).
19. Multi-oared boats of the plank-canoefamily are the typical vessels involved in such activities,
throughout the Mediterranean to the Atlantic shores. The tradition of boats like that is still
living in Portugal with the 'saveiro' boat (see O.L. Filgueiras, 'The Xavega boat. A case of
study on the integration of archaeological and ethnological data', in Sources and Techniques
in Boat Archaeology (ed. S. McGrail), BAR Suppl. Series 29 (1977), fig. 9.6.), in the Black Sea
and Bosporus (the 'dalian' boat), and in the Turkish coasts ('piyante' boat).
20. Their use in fishing activities is suggested to an extent: Evans-Renfrew 1968, 58, 79, 119.
21. Waisted weights found in the Saliagos site might be net sinkers (Evans-Renfrew 1968, 71,
79, fig. 87).
22. Two vertebrae of fish however have been found on the floor of an Early Helladic House in
the coastal settlement of Aghios Kosmas in Attica - a settlement with very strong Cycladic
influences (G. Mylonas, Aghios Kosmas, Princeton 1959, 11, 35, 48). The excavation was
performed in the 1930s. Nothing is reported about the size of these vertebrae and their bad
stage of presewation did not permit a definite identification.
23. J.L. Bintliff, Natural Environmentand Human Settlement in Prehistoric Greece, Oxford 1977,
117-125, 538-542, 594-595. Against this approach stands the direct archaeological evidence
of at least one key site, excavated by modern methods. Bone material from Phylakopi
indicates that the marine portion catch was not contributing greatly to the diet of the town.
The few fish remains recovered in the excavation were of small fish with no bones equivalent
to the large migrating tunas. The possible reasons for this obvious neglect of a substantial
resource - which are questioning hard Bintliff's ideas about man's adaptation to the mobility
of resources - are discussed by C. Gamble, 'Surplus and Self-sufficiency in the Cycladic
Subsistence Economy', in Papers in Cycladic Prehistory, eds. J.L. Davis-J.F. Cherry, 1979,
126-7.
24. H. Michell, The Economics of Ancient Greece, 1957, 288.
25. W. Radcliffe, Fishing from the Earliest Times, London 1921, 100.
26. For the most recent discussion on the 'frying pan' vessels and the catalogue of incised
depictions see J.E. Coleman, 'Frying Pans' of the Early Bronze Age Aegean', AJA 89 (1985).
191-219, fig. 5. The very badly damaged example in Berlin (W. Zschietzschmann,
'Kykladenpflannen', AA 50 (1935). 657, Abb. 3) is omitted.
27. C. Renfrew, 'Cycladic Metallurgy and the Aegean Early Bronze Age', AJA 71 (1967). 5.
28. For an estimation of the nautical qualities of the shape of these boats see r. Bfixoq, 'Ta
nhoia m a 'qyavouxqpa' O K E U ~t-qq ZU~OU', EV&Q 1.8 (1989), 14 - 15. An important
issue for many years has been the identification of the prow with the high or the low end of
the craft. Scientists like C. Renfrew, L. Casson, D. Fimmen, G. Glotz, N. Platon and D. Levi
favored the first view while others like S. Marinatos, Ch. Doumas, J. Morrison, A. Koster, P.
Johnstone, Y. Vichos and L. Basch supported the second one. I believe that the first one is
more convenient on the basis of the fish emblem, which, as one would expect, should be
facing forward.
29. Such estimations are very popular between the nautical archaeologists especially in the
A SUGGESTION
cases where the data is limited. Conclusions, however, based on such estimations can be
very fragile: we have no idea if these boats are paddled or rowed, and if the latter is the case,
the number of crew depicted could be reduced immediately to the half now estimated. And
most importantly, we have always to keep in mind that the persons who made the incisions
and lead models were potters and metal craftsmen and not naval architects. They did not
share the same interest to the subject we have, and probably they wanted just to express the
idea of a large (narrow?) vessel, with what they estimated 'a good number of oarsmen'. If this
is the case, then the enumeration of oars, beam-length ratios and interscalmia can not have
any real value.
30. C. Broodbank, 'The Longboat and Society in the.Cyclades in the Keros-Syros Culture', AJA
93 (1989), 319-337.
31. C. Renfrew, The Emergence of Civilization, London 1972, 398.
32. A. Macnpanaq, 'To nhoco m ~ KuKAa&$
q ma q v npcliiprl X a k o ~ p m i a 'Apxa~oAoyia
, 35
(6 - 1990), 0.82.
33. Similar fishes are depicted in much larger scale on a 'frying pan' of the Athens National
Museum (n. 6140) originating from Naxos, but not related, however, to a longboat (Coleman,
AJA 1985, pl. 36, fig. 23). The identification of the fish emblem as a tuna was proposed also
by Bintliff, 'Natural Enviroment', 121.
34. L. Basch, Le Musee imaginaire de la marine antique, Athens 1987, 137-8.
35. C. Broodbank, AJA 93, 328, 336.
Table I: Weight of meat represented in Cliff 17, Pit A, and in Square N3 by species in Saliagos.
After Evans-Renfrew 1968, p. 79, tab. 23.
Table II: Nine large Cycladic cemeteries, with implied settlement populations and numbers of
nuclear families estimated for durations of cemetery usage from one to four centuries
(after Broodbank, AJA 93 [1989], p. 325, tab. 5.
Fig. 1. Longboat depictions on Early Cycladic 'frying pan' vessels. Numbered examples
redrawn after Coleman AJA 89 [1985], 199, fig. 5.
Fig. 2. Example of Early Cycladic lead model (after C. Renfrew AJA 71 [I9671 pl. 1)
Fig. 3. Fish and 'banner' ornaments on Cycladic ships inscribed on 'frying pan' vessels.
Numbered examples redrawn and magnified after Coleman, AJA 89 (1985). 199, fig. 5.
Fig. 4. A ship from the disk of Phaistos. Photo: J. P. Olivier, after L. Basch, MlMA (1987), p.
137, fig. 285F.
TABLE 23. W E I G H T O F M E A T R E P R E S E N T E D I N CLIFF 17, P I T A,

A N D I N SQUARE N3 BY S P E C I E S
c u m 17, PTT A QUARE ~3

No. of Min. tlum6sr Meat wt. Total nuat Min. manber Tolal meat
Spcc'es fmgmmls of animals pcr animal wt. (kg.) of a n i d wt. (kg.)
SheepIgoat 2595 17 29 kg. (65 Ib.1 490 22 640

Bovine 32 I 210 kg. (450 Ib.) 210 6 I ,260
Pig 96 4 45kg.(roolb.) 180 10 450
Tunny 1930. 48 135kg. (300 lb.) 6,500 7 940
Other Fish 358 .. .. .. .. ..
Patella 1900 I900 3-5 gm. 6.5 1970 6.90
Momdonlo 340 340 1.0 gm. 0'35 650 0'65
Mwex 140 140 5.0 Bm- 0'70 216 I .&
Orartodmna 90 45 1.0 gm. 0'05 53 0.05
Table 1
CEMETERY BURIALS 100 YEARS 200 YEARS 300 YEARS 400 YEARS
(Approx.) Families Pop. Families Pop. Families Pop. Families Pop.
Table 2
A SUGGESTION
Fig. 1
Fig. 2
SUDDEN SEALEVEL CHANGES: CAUSES AND CONSEQUENCES.
FIRST HISTORICAL AND ARCHAEOLOGICAL EVIDENCE FOR ICE-
SHEET DECOUPLING EVENTS?
INTRODUCTION
Have you ever wondered what would happen if the level of the sea rose
suddenly by 2 or 3 meters? It is true, of course, that traditional thinking
connected with sea level rises considers such occurrences as gradual
events, taking place during periods of climatic amelioration'. The
accompanying melting of glaciers results in the slow rise of the sea.
Relatively recent evidence, however, shows that within this gradually rising
sea level there has been a number of abrupt and quite unexplained changes
in the level of the sea. Such events are by no means confined to the
Holocene or the last ten thousand years, but are also observed towards the
end of the last interglacial ca 120,000 year ago, and quite possibly during
other periods of the past.
Geoscientists studied the event of the last interglacial in the limestones

of the stable platform of the Bahamas in the Caribbean. There 120,000 years
ago and in a period of a century the sea level rose 20 feet above and then
plunged 30 feet below modern levels, prompting scientists to call a
"madhouse" these pulses of catastrophic changes capable of dramatically
reshaping the landscape2.Today, with half of humanity living in coastal
areas, the results of such changes may be easily imagined.
Similar if not so marked sea level changes appear to have taken place
during the last ten thousand years, or what geologists call the Holocene
Epoch, when real havoc could have befallen port installations or maritime
settlements, by sinking access roads under the sea and drowning all
sources of sweet water. It is the purpose of this paper to attempt to explain
these phenomena. To do so, the Middle and Late Holocene occupation and
desertion of the Israeli coast is used to focus on the geological processes
and the archaeological and historical evidence, as a means of
understanding such odd, potentially risky event^.^
CONNIE LAMBROU-PHILLIPSON & JOHN PHILLIPSON TROPlS VI
FLUCTUATING OCCUPATION AND DESERTION
In his description of the ancient harbors of Israel, Avner Raban says that
periods of notable activity associated with flourishing coastal settlements,
were followed by periods of negligence and outright desertion, when the
Jews and the Arabs lived with their backs to the sea4. This intermittent
occupation of the coast is attested since Upper Palaeolithic times5.It seems
particularly marked with prosperous sites during the 5thand dthmillennia BC
or from Neolithic B to Early Bronze I, followed by desertion during the 3'*
millennium, and again by occupation at the beginning of the second
millennium or in Middle Bronze 116. One wonders what made this coast
attractive in the 5thand 4thmillennia BC, repellent in the rd, and attractive
again in the Pd?
Various processes have been advanced to explain this situation in the

Syro-Palestinian coast, ranging from a eustatic sea level rise, to crustal
downwarping, to climatic deterioration. All are unsatisfactory for a variety of
reasons, rapidly reviewed below.
DESERTION CAUSED BY EUSTATIC RISE
A eustatic sea level rise cannot explain settlement on this coast during
the Eith and 4'h millennia when all indications show a relatively rapid sea level
rise, and desertion in the 3rdduring a much slower rise. The reverse would
appear much more likely. Besides, it is improbable that all settlements were
founded at exactly the same elevation above MSL, at the same distance from
the coast, so that a rise of the sea affected all of them in the same manner,
to the same extent, and at the same time. This is a little too improbable to be
seriously entertained as a viable proposition. Even a difference of one or two
meters in the elevation of the various sites would have resulted in a gradual
abandonment, not the wholesale desertion attested on the Israeli coast
during the specific interval of the 3rdmillennium BC.
Of course, any rise of the sea would tend to silt up existing harbor
facilities as Raban suggests7,by diminishing the stream velocity at or near all
river mouths and causing a drop of the suspended load. But in the East
Mediterranean there is an additional reason for the silting of harbors, which
is of marine not continental origin that complicates this simple picture.
This reason is the combination of counterclockwise near shore currents

SUDDEN SEALEVEL CHANGES: CAUSES AND CONSEQUENCES
FIRST HISTORICAL AND ARCHAEOLOGICAL EVIDENCE FOR ICE-SHEET
DECOUPLING EVENTS?
along the Syro-Palestinian coast, and the Coriolis deflection in the northern
hemisphere, which is to the right of the direction of flows. But this would tend
to move sand and silt into all existing harbors, not only those at river mouths
on the Israeli coast. Something which may help explain the existence of
Bronze Age desilting facilities in the form of water reservoirs and sluices
associated with some ancient island harbors, first explored by Pere
PoidebardQ.
DESERTION PRODUCED BY CRUSTAL DOWNWARPING
Crustal downwarping and its opposite upthrust on the other hand cannot
be assumed for successive millennia. Neither the apparent geomorphology
of this coast, nor its calculated rate of sinking, nor its seismicity, nor the
absence of a large river to overload this coastal belt with sediment, argue for
crustal downwarping.
First, there is no obvious evidence of downwarping in the

geomorphology of the area. The Israeli shoreline is as straight as few other
shorelines, the coastal sandstone ridge more or less parallel to the shore,
save for the break at Tel Mevorach. The lack of indentations and
embayments is rather notable. This is decidedly no evidence of
downwarping.
Second, the rate of sinking calculated from a lucky underwater find is no

different from other known Holocene transgression rates of between 0.18
and 0.15cm per annumlO.The find consists of an oak trunk encircled by a
man-made stone fence found in situ under 5.4m of water, dated by Carbon
14 to the 4thmillennium BC (corrected C-14 date, 5700 BP 2 140), indicating
a sea level 9 to 10m below the present one1'.While a submerged Pre-pottery
Neolithic water well at Atlit-Yam seems to confirm this Middle and Late
Holocene variation in sea level12.
Third, there is no large river delta on the Israeli or more generally on the
Syro-Palestinian coast, the accumulated sediment of which might have been
responsible for the readjustment of the underlying crust. That is, excessive
loading causing the gradual isostatic sinking of the continental shelf, such as
one may reasonably postulate for the Nile Delta for examplet3.
Fourth, the seismicity of the Syro-Palestinian coast is very low by any

standards. It is extremely low by comparison to the seismicity of the Aegean
CONNIE LAMBROU-PHILLIPSON& JOHN PHILLIPSON TROPIS VI
area, where virtually half the seismic energy of Europe is released in an area
which is a small fraction of the total area of that continent. This does not
argue for crustal downwarping either.
Finally, if crustal movements were the causes of the successive

habitations and desertion of the Israeli coast, that would imply downwarping
during the 3rdmillennium BC and upthrust in the Pd,which are highly
improbable from what we know today. For a recent review of the problem
based on different evidence but which arrives at similar conclusions, see
Galili et a/. (1993).
DESERTION ORIGINATING FROM CLIMATIC DETERIORATION
Another reason that may be advanced is climatic deterioration causing

the desertion of the Israeli coast in the 3rdmillennium, just as it was probably
climatic amelioration which brought perennial occupation in the Sth
millennium. The latter could have been caused by a northward progression
of the Intertropical Convergence Zone, and a corresponding hydrological
optimum in an area generally poor in surface water resources.
That would have caused increased planctonic productivity of the littoral

due to excessive water volume and nutrients channelled through the Nile,
improved fish-catch, etc. The coincidence of the beginning of Mediterranean
river delta formation ca 8,000 to 7,000 BP14,and the beginning of perennial
occupation of this coast in the Sm millennium BC is too striking to pass
unnoticed. Therefore, a corresponding climatic deterioration could have
been caused by the reverse process, that is a southward progression of the
Intertropical Convergence Zone, which would be equivalent to a
hydrological minimum.
The suggestion has undoubtedly the merits of logic, neatness and

balance to recommend it. Unfortunately, none of these otherwise desirable
qualities are safe guides to meteorological or geological processes.
Evidence is at hand, which contradicts the assumption of climatic
deterioration, at least to the extent of causing wholesale desertion. This is
that Early Bronze II sherds are found in at least a few early settlements, but
"always on sites which are located high up on the lee side of the coastal
ridget5."Surely no climatic deterioration could have made life unbearable for
the coastal settlements, but spared the sites located on the sheltered side of
the coastal ridge that is barely a few meters above sea level. No climatic
DECOUPUNG EVENTS?
deterioration could be so confined as the difference in height and distance

between the coastal sites and these situated on the lee side of the coastal
sandstone ridge appear to imply.
ICE SHEET DECOUPLING AND EPISODIC SEALEVEL CHANGES
In the absence of other viable explanations, a mechanism that is worth

considering is episodic sea level changes caused by ice sheet decoupling
eventsI6. Recent data regarding the Holocene glacial history indicates that
deterioration of marine ice sheets did not occur at a steady rate". Unlike
terrestrial ice sheets, marine ice sheets or ice sheets bordering the sea are
far more sensitive to sea level fluctuations than they are to climatic variations;
and under certain conditions are liable to rapid mass wasting. Antarctica
seems like a prime suspect for harboring such delinquent ice sheets,
because marine geological studies of its continental shelf indicate that ice
sheets were grounded over large portions of the shelf during the last glacial
maximum of about 18-20,000 BC, but their removal was very rapid1'.
This is not hard to understand. Marine ice sheets melt mainly at the
interface of the ice with the sea. A gradual rise of the sea would increase that
surface. But since ice is lighter than water, this would also tremendously
increase the hydrostatic pressure on the grounded ice sheet, in some cases
resulting in dislodging it from the floor. What is more, this hydrostatic
pressure would be far greater in Antarctica than on any other continent. If all
water in glaciers was released, flood waters would cover 7% of Africa, 15.5%
of Australia, 17% of North and South America, 20% of Eurasia, but 34% of
AntarcticaIs.
In other words, the hydrostatic pressure would work far more effectively,
precisely where glaciologists and geologists maintain that ice sheet
decoupling events might have taken place, i.e. in Antarctica and as indicated
in Figure 2.
Decoupling would be followed by a "draw-down", as it is called, or a

slide into the sea. How fast such a slide could take place depends on a
number of variables such as floor smoothness and slope, or the gradient of
the continental margin; the shape of the ice sheet, or its length versus its
broadness; its thickness which measures the downward force component;
and the increasing hydrostatic pressure of the rising sea level, all illustrated
in Figure 3.
CONNIE LAMBROU-PHILLIPSON & JOHN PHILLIPSON TROPIS VI
What should be the size of an ice sheet, the complete draw down of
which could result in a significant rise of sea level, say 2m? Calculating from
present circumstances and assuming all other factors being equal, a rise of
2m over all oceans with an area of 362 million km2would require a marine
ice sheet 362,000 km2 in area and an average thickness of 2km. That is an
ice block 630km long, 630km broad, and 2km thick, including a correction
for the fact that water occupied 9110th~the space of ice. The Ross Ice Shelf
in Antarctica is over 50% larger in size; and ice sheets of 2-3km thickness are
known in present day Antarctica. But today ice covers only about 10% of the
continents. During the last glacial maximum, the cover was well over 30%.
So there is nothing a priori improbable about the figures presented here.
Additional recent evidence from the N Texas Gulf shows sea level has
risen in an episodic manner during at least the last 12,000 years, in episodes
of similar duration as the earlier described Bahamas casez0.The evolution of
river valleys, fluvial deltas, coasts and estuaries, and the facies architecture
of these depositional systems have been profoundly affected by the episodic
nature of sea level rise2'.
Unfortunately, verifying such episodes is usually extremely difficult.

Relatively rapid rises in sea level caused by decoupled marine ice sheets
would fall in the realm of "noise" in oxygen isotope studies of deep sea cores
for example. Here often 100,000 years may be represented by a few meters
of core length".
EVIDENCE FOR ICE SHEET DECOUPLING FROM TEXTUAL SOURCES?
It is for this reason that other evidence is necessary, and perhaps why
one piece of ancient testimony deserves our careful scrutiny. Writes the
ancient Greek geographer Strabo:
"And when I was residing in Alexandria, in Egypt,

the sea about Pelusium and Mt Casius rose and
flooded the country and made an island of the
mountain, so that the road by Mt Casius into
Phoenicia became navigable" (1.3.17).
Before discussing the implications of this passage, it is essential to clear

some uncertainties (Fig. 4). Mt Casius is not the similarly named mountain
DECOUPLING EVENTS?
near the Orontes River in Syria. This Mt Casius is not around ancient
Pelusium and has no place on the road from Egypt to Phoenicia. Strabo
rather refers to a striking range of sand dunes 60m high, which is an
impressive height on the Egyptian coast. These are located to the immediate
east of Pelusium and now known as Katib el-Qals (31 13" N, 33 05" E), 45
miles E of Port Said. There the coastal strip is barely 2km wide and very low
except for the dunes of Katib el-Qals, which are identified as the ancient
Mount Cassius by the Mediterranean Pilop. A 2m sea level rise or even less
could have made an island of the high sand dunes and the road to Phoenicia
"navigable", as Strabo says using the explicit words, ploten genesthai
(1.3.17).
Now we happen to know that Strabo was in Alexandria using the facilities
of the famous library there between 20 and 25 BCZ4.So we have some
accurate dates of when this sea level rise took place and can be on the
lookout for it in coastal excavations. But the primary question here is, what
could have caused the road to Phoenicia to become navigable?
Obviously Strabo does not refer to a tsunami. He has done so elsewhere

when describing a battle between Tryphon, the regent of Antiochus VI, and
Sarpedon, a general of Demetrius II Nicator of Syria, "on the shore between
Tyre and Ptolemais" ca 141 or 139 BC (16.2.26). There is no similarity in the
two passages, naturally enough, since tsunamis would be described for their
destructive power, not for making coastal roads navigable.
Strabo is not referring to a Nile flash flood either, when the river could
have jumped its banks and flooded the countryside. It is extremely unlikely
that the Nile would have jumped its banks, and Alexandria then joined to the
Nile by several canals2kould have remained intact. Besides, the last great
Nile floods recorded in Pharaonic inscriptions fall in the first half of the Pd
millennium BC26,and nowhere near the time of Strabo.
The same conclusion would be reached by assuming that this rise in sea
level could have been caused by the weight of Nile sediment resulting in a
gradual sinking of the delta. If in five years the delta sank about 2m, then
Alexandria should be over 900m under the level of the sea by now.
It is the absence of any other alternatives that has forced the authors to
accept ice sheet decoupling as a probable explanation; not only for Strabo's
description of the road to Phoenicia, but equally for the sequence of
occupation-desertion-occupationof the Israeli coast from the 5thto the 2"*
millennium BC. But such an acceptance implies also the obligation to accept
that other sites in the East Mediterranean might have had similar histories.
An Antarctic ice sheet decoupling event can hardly be construed as a local
phenomenon; its effects though varied in different places must be
widespread.
EVIDENCE FOR ICE SHEET DECOUPLING FROM ARCHAEOLOGY?
As it happens, Strabo's testimony is powerfully supported by the

excavations carried out at Ras ibn Hani, located 9km NNW of Lattaquie and
5km WSW of Ras Shamra. Ras ibn Hani is a low-lying cape about 500m
wide, projecting 2.5km westwards forming a tombolo or tie-bar made up of
marine sediments. According to the excavators, during the Hellenistic
occupation during the 3'dand Pd centuries BC, the sea was more or less at
its actual level, but deposits of fine sediments indicate an elevation of the
sea.
The site seems to have been "submerged during the largest part of the
Roman period and then re-occupied in the 4'h century AD", before being
destroyed by an earthquake probably around 526 AD2'. The Roman period
of the Syro-Palestinian coast dates from the second half of the 1" century
BC. In fact, the language of the excavators place the submersion of Ras ibn
Hani as near Strabo's residence in Alexandria as it is possible to come within
this context.
What are we make out of this? Shall we say with Lord Byron,
A "strange coincidence", to use a phrase

by which such things are settled nowadays?
Lord Byron Don Juan VI. 78
We don't think so. We think that when in excavations of coastal sites,

excavators come across indications of rapid rises in sea level, they should
not assume that they are wrong simply because we have all been taught that
such things are gradual events. We know better now. And such events do
not seem confined to any specific period. Thus for example, Strabo quotes
the earlier Posidonius that,
"He infers the migration of the Cimbrians from their native

country occurred as the result of an inundation of the sea
DECOUPLING EVENTS?
that occurred all of a sudden" (2.3.6)
The German tribe of the Cimbrians was known to inhabit the Jutland
peninsula of present day Denmark, precisely one of the places that would be
inundated from a sudden sea level risez8.
Of course, as Raban said during the discussion of this paper, it is not

very likely that excavators would fail to notice such a sudden rise in sea level.
But that would be only if they are excavating a site directly affected by such
rises. If the upward movement of the sea drowned the sweet water resources
of the site, this may or may not be noticed, depending on other factors. And
a smaller sea level rise may not be noticed if one is not on the lookout for it".
There are far too many examples from archaeology, when the excavators or
their successors returned for a second look as a result of later knowledge or
testimony.
DESERTION BROUGHT ABOUT BY UNDERWATER LAVA FLOWS
In the previous discussion of the causes of desertion of the Israeli coast,

one cause that was not reviewed is the distinct possibility of underwater lava
flows. This possibility was discussed by one of the authors (JP), a consulting
geologist, with Alan E Johnson, another consulting geologist in Nairobi
during July 1996, and the authors hereby express their gratitude.
It is well known that lava flows are continuously extruded in mid-ocean

ridges and from active volcanoes. It is therefore conceivable that a sea level
rise could be caused by an underwater lava flow, the lava or other extrusive
products replacing an equal volume of water.
The problem with this suggestion is one of scale. The volume of extruded
products from the largest known volcanic explosions ranging from Krakatoa
to Santorini have been estimated to between one and three scores of cubic
kilometers30.For such a figure to reach the 362,000 cubic kilometers required
for a 2m sea level rise necessitates an increase of no less than four orders
of magnitude. Underwater lava flows may be considerably larger than these
on land but four orders of magnitude during a short period such as Strabo
implies in his narrative have to be considered as extremely unlikely.
CONCLUSION
It is certainly striking that the available written and archaeological

evidence implicates parts of the coasts of Egypt, Israel, Syria, and the Baltic
as areas of the ancient world, where a sudden rise of sea level would have
been the cause of widespread flooding. Geologically speaking, there are still
areas liable to marine transgressions after only a small sea level rise. So one
must assume that such sudden inundations did take place in the past, and
it is highly desirable to find archaeological proof of such floodings wherever
this is possible.
Aside from casting some light on many unexplained desertions of

coastal sites, there is one more reason to archaeologically trace and
describe such events. Archaeology borrows methodologies from the earth
sciences, and it is no secret that most archaeometric methods have a
geological provenance. The archaeology of coastal sites is one area where
archaeology can repay the debt. Earth scientists are assiduously looking for
confirmation of their findings. And the combination of historical and
archaeological testimony can be a powerful confirmation of such events.
Events that may have drastically altered the life of coastal peoples in the
past, but which may also do so in the future, with far more dire
consequences than at any time in the past.
Therefore the precepts of old models regarding gradual sea level

changes should not be allowed to guide the interpretation of excavations or
other archaeological data. Abrupt, erratic and bizarre changes of sea level
have been established beyond any reasonable doubt by firm geological
evidence. What is the explanation for such events is an open question. One
thing is certain: it need not be a dramatic climatic change as it has been
assumed all along. A gradually rising sealevel could be a far more effective
mechanism for the decoupling of ice sheets, which could result in sudden
and devastating inundations of low lands, a change in local atmospheric
conditions, etc.
There is obviously more at stake here than the reliability of

methodologies and models. And earth scientists investigating such events
are anxious to have confirmation from other disciplines. Especially the
definitiveness that may result from the material record of archaeological
excavations.
DECOUPLING EVENTS?
ACKNOWLEDGEMENTS
The authors would like to express their thanks to Leib Wolofsky,

consulting geologist, who read the initial version of this paper and made
many valuable comments.
Connie Lambrou-Phillipson & John Phillipson

P.O. Box 3771, 10210 Athens
Greece.
NOTES
1. Sea level curves are notoriously unreliable in general (Fig. 1); see for example Naldretl 1990:
51, Figs 1 and 2. This is mainly because they are usually recorded in a specific place and
show the resultant of all movements the specific location might have been subjected to such
as these caused by glacioeustatic factors, isostasy, faulting, tilting, downwarping, glacial
rebound, etc. Since aside for Holocene rising sea level, various places have different geologic
histories, sea level curves do not, and could not, show any but accidental agreements, given
the role of the previous parameters on a dynamic earth.
2. Some of these are briefly reviewed below.
3. See a brief review of the subject in Stock 1995.
4. Raban 1985: 11.
5. As for example at Ras Beirut, Adlun, Tabun, Kabarah, Atlit, etc; see Garrod 1970: 75-84.
6. See Raban 1985 for relevant references.
7. Raban 1985: 11.
8. McDonald 1952.
9. Poidebard 1939.
10. KrafI et a/. 1989; Pirazzoli & Suter 1986.
11. Raban 1983: 22.
12. Galili & Nir 1993.
13. Stanley 1989.
14. /bid.
15. Raban 1985: 14.
16. Anderson & Thomas 1989.
17. Denton & Hughes 1981 quoted by Anderson &Thomas 1989.
19. Ryabchikov 1975: 30 fn. Another factor that would increase the hydrostatic pressure is the
sinking of the continental shelf under the weight of the ice. It is known that the center of
Finnoscandia had sunk about 850m during the last glaciation, and that during deglaciation
and glacial rebound great geological disturbances occurred, including seismic activity and
even the remelting of some rocks. Similar occurrences could have taken place in Antarctica
and hastened the waisting of marine ice sheets; see Morner 1989.
20. Kidson 1982; Carter et a/. 1986.
22. See for example, Rossignol-Stricketa/. 1982.
23. Mediterranean Pilot. Vol V, 1976, p. 83, sect 3.141. Raban said in his intervention that other
CONNIE LAMBROU-PHILLIPSON& JOHN PHILLIPSON TROPlS VI
high dunes exist on this coast, but in either case there are on the way from Alexandria to
Phoenicia, and would have become "navigable" with a 2m rise of sea level as assumed here.
24. Strabo, Introduction&mi.
25. Strabo 3.1.9.
26. Dunham & Janssen 1960: 135, PI. 95A (RIS 16); Bell 1970; Verner 1972.
27. Bounni nd: 106 & n 3, and 1978; see Bounni et al. 1976.
28. Elsewhere Strabo shows that he does not believe Posidonius's explanation (7.2.1). But his
comments clarify that here he is thinking of the tidal process ("natural and eternal... increase
and diminution... regulated and periodical... twice every day..."), not the process he
described for Alexandria. This difference in outlook here and in several other instances shows
to the authors of this paper that Strabo did not write his Geography around AD 18 and 19 as
B Niesse thought (1878), or around 7 BC as Ettore Pais contended (1908), but rather during
a much longer period than these dates indicate, perhaps starting as early as 25 BC in
Alexandria. But this inquiry cannot be pursued further here.
29. The 2m rise used here is simply an assumption. There is no evidence in Strabo to show the
height reached by the sea when he resided in Alexandria. This might have been less than
what is conveniently assumed there.
30. See for example Wilson 1978; Thorarinsson 1978; Sigurdsson etal. 1990; Pyle 1990.
REFERENCES
ANDERSON JB & THOMAS MA 1989, Marine ice-sheet decoupling as a mechanism for rapid,
episodic sea-level events. The record of such events and their impact on
sedimentation. Abstracts, 28" lnternational Geological Congress, Washington, 1: 41.
BELL B 1970, The oldest records of the Nile floods. GeographicalJournal 136: 569-73.
BOUNNI A, LAGARCE J & SALIBY N 1976, Rapport preliminaire sur la premiere campagne de
fouilles (1975) a Ibn Hani (Syrie). Syria LIII: 233-79.
BOUNNI A 1978, Rapport preliminaire sur la deuxieme campagne de fouilles (1976) a Ibn Hani
(Syrie), Syria LV: 233-301.
BOUNNI A nd, La Syrie, Chypre et I'Egee d'apres les fouilles de Ras ibn Hani. In V Karageorphis
ed, The Civilizations of the Aegean and their Diffusion in Cyrpus and the Eastern
Mediterranean, 2000-600 BC: 18-24 September 1989, Larnaca, 105-110.
CARTER RM, CARTER L & JOHNSON DP 1986, Submerged shorelines in the SW Pacific:
Evidence for an episodic post-glacial transgression. Sedimentology, 33: 629-49.
DUNHAM M & JANSSEN JMA 1960, Second Cataract Forts. 1 Sernna-Kumma, Boston.
FAIRBRIDGE RW 1961, Eustatic changes in sea level. Physics and Chemistry of the Earth, New
York, Vol. 4: 95-185.
GALlLl E & NIR Y 1993, The submerged Pre-pottery Neolithic water well of Atlit-Yam, northern
Israel and its palaeoenvironmental implications. The Holocene 3: 265-70.
GARROD D 1970, Primitive man in Egypt, Western Asia and Europe in Palaeolithic Times. The
Cambridge Ancient History, TdEd, Cambridge, UK, Vol. I, 1: 70-89.
KlDSON C 1982, Sea level changes in the Holocene, Quaternary Science Reviews, 1: 121-51.
KRAFT C, KAYAN I & ASCHENBRENNER E 1985, Geologic studies of coastal change. In G
Rapp & JA Gifford eds, Archaeological Geology, New Haven & London, 57-84.
KRAFT C, CHRZASTOWSKI MJ, STEDMAN SM, HI-IL Y 1989, Sedimentation rates in coastal
marshes as indicators of relative sea level rise. Abstracts, 28mlnternational Geological
Congress, Washington 1989, 2: 220-1.
McDONALD JE 1952, The Coriolis Effect, Scientific American, May 1952: 72-6.
MORNER N-A 1989, The Swedish failure in defining acceptable bedrock depository for nuclear
DECOUPLING EVENTS?
waste deposition. Abstracts, 28" lnternational Geological Congress, Washington DC, 2-

468.
NALDRET AJ 1990, lnternational Geological Correlation Programme: An example of
collaborative geoscience. General Proceedings, 28" lnternationalGeological Congress,
Washington, DC, USA, July 9-19, 1989, Reston, VA, 50-6.
NlESE B 1878,Beitraege zur Biographie Strabos. Hermes 33.
PAlS E 1908,The Time and place in which Strabo composed his Geography. Ancient Italy
(English transl.), London.
PlRAZZOLl PA & JR SUTER eds 1986,Late quaternary sea-level changes and coastal evolution.
Journal of Coastal Research, Special Issue 1.
POIDEBARD A 1939,Un grand port disparu: Recherches aeriennes et sousmarines 1934-36.
Paris.
PYLE DM 1990,New estimates for the volume of the Minoan Eruption. Thera and the Aegean
World Ill,11: I 13-21.
RABAN A 1983, Submerged prehistoric sites off the Mediterranean coast of Israel. In NC
Flemming & PM Masters-lnman eds, Quaternav Coastlines and Marine Archaeology,
London.
RABAN A 1985, The Ancient harbours of Israel in Biblical times. Harbour Archaeology,
Proceedings of the First lnternational Workshop of Mediterranean Harbours. BAR
lnternational Series 257,Oxford.
ROSSIGNOL-STRICK M, NESTEROFF W, OLOVE P & VERGNAUD-GRAZZINI C 1982,After the
Deluge: Mediterranean stagnation and sapropel formation. Nature 295:105-10.
RYABCHIKOV A 1975,The Face of the Earth, Moscow.
SIGURDSSON H, CAREY S & DEVlNE JD 1990, Assessment of mass, dynamics and
environmental effects of the Minoan Eruption of Santorini Volcano. Thera and the
Aegean World 111, 11: 100-12.
STANLEY DJ 1989,Evolution of Nile and other Mediterranean deltas during Holocene and role
of eustatic sea-level oscillations. Abstracts, 28" lnternational Geological Congress,
Washington, 3:170.
STOCK C 1995,High Tidings. Ancient, erratic changes in sea level suggest a coming swell.
Scientific American, August 1995:15-16.
STRABO 1960, The Geography of Strabo, Transl. by HL Jones, Loeb Edition in 8 volumes,
London UK and Cambridge MA.
THORARINSSON S 1978,Some comments on the Minoan Eruption of Santorini. Thera and the
Aegean World 1: 263:75.
VERNER M 1972,Periodical water-volume fluctuations of the Nile. Archiv Orientalni 40: 105-23.
WILSON L 1978,Energetics of the Minoan Eruption. Thera and the Aegean World, 1: 221-8.
bGE . 13' YE-S BEFaK PRESENT
-3- V1-Lc
.
..I A
Fig. l a
Fig. 1b
Low relative sealevel: Hydrostatic pressure small
If all glaciers melt. water will tend to cover

7% of Africa
1 5 . 5 % of Australia
17% of N 6 S America
20% of E w a s l a , but
34% of Antarctica
-
~ l g hrelative sealevel: Hydrostat~cpressure large Fig. 2
DECOUPLING EVENTS?
REMARKS ON THE HYPOZOMA
Greek inscriptions at Piraeus from the Sthand 4mcentury BC tell us that

men of war had hypozomata, made of rope (Morrison, Williams, 1968, pp.
305-307). Polybios wrote that these were laid on at the eve of a campaign
00(Vll, 3, 3). Nobody tells us how and where they were fitted. Apollonios
Rhodios comes nearest (1, 367-370). It was done on both sides (ekaterthen),
with a well-turned rope (eustrephei oploi). Between adjective and noun we
find the word endothen (from the inside). This has been interpreted as
meaning that the rope was 'turned on itself'(Bude, 1974.1, p. 67), but it might
mean that the Argonauts were 'teinamenoi' (tightening) from the inside. The
purpose of the 'undergirding' was 'to make the strakes fit well around the
gomphoi and to make them resist the impact of the waves'.
The part about the gomphoi may be compared with a statement made
by professor Steffy when writing about the function of tenons (Steffy, 1995,
p. 422). Longer, larger hulls required additional support, if only to reduce the
possibility of tenon shear; the stringers on the Madrague de Giens ship are
a good example (Pomey, 1978. PI. XXVII, XXXVI). Long, narrow warships
perhaps required even more stringers and hypozomata as well. Hogging
apparently was the principal hazard for these ships (Coates and Macgrail,
198(4?), pp. 71-73). So it is logical to think of the Egyptian 'hogging-truss'
that acted like a bowstring, the hull being the bow. It actually encouraged
sagging, to counter which, and hogging at the same time, perhaps the
beams sewed, that stuck out through the hull as shown in the reliefs of
Queen Hatshepsut's ships at Deir-al-Bahari (Steffy, 1994, fig. 3-6) and
probably in the Dashur boats (Steffy, 1994. pp. 33-37). These beams may
have had grooves in them to receive the edges of 1 or 2 indented strakes,
like in the Bremen cog (Lahn, 1992. pp. 66-72). The very word girding
suggests that the Greek hypozoma was laid around the outside of the ship.
The written evidence for this is slender and not nautical. Most important
is perhaps Kallixeinos' description of Ptolemaios Philopator's
tessarakonteres, recorded by Athenaios of Naukratis (V. 203c-204d), that
tells us that its hypozomata had twice the length of a hull and 117 of that
length to spare. Plato's fantasy of the light being like the hypozoma of a
trieres (Politeia, 616 b-c) is equivocal. It suggests a hogging-truss following
a centerline as well as something going round the kosmos. There is
Vitruvius' contention (X, 15, 6), that round the battering-ram in Hegetor's
L. TH. LEHMANN TROPlS VI
'tortoise' go 3 (Loeb, 1956), or 4 (Bude, 1986) ropes tied from the head to
the heel of the beam in the way that ships are kept together (continentur) 'a
prora ad puppim'. Then, again, at given distances the whole was tied by
transverse ropes going round the beam.
There is a textual problem here, whether the mss. give navis or naves.
We translated as if the latter is the case but Fleury (Bude, 1986) thinks that
'navis funes' (the ropes of the ship) 'continentur'. As these are already
'religati' (tied), this sounds as if the longitudinal ropes were held in placed by
the transverse ones, which would 'brake' the ship. There may be an echo of
Vitruvius' words in the 'Origines' by lsidorus of Sevilla, where (XIX, 44) a
'tormenturn' is a rope supposed to keep a ship together 'a prora ad puppim',
whereas a 'mitra' is one tied round it midships. The latter may be a rope used
in undergirding as an emergency measure, mentioned by Paul in Acts, 27,
17. We have no other indications whether hypozomata still existed in the Th
century AD. Perhaps Vitruvius was hasty in his nautical association; there is
a description of Hegetor's tortoise extant in Greek, by Athenaios
Mechanikos, who is thought to have lived earlier than Vitruvius in the 1"
century BC (Bude, 1986. Livre X, pp. 299-300). It mentions 3 ropes and the
verb 'hypozonnumi' but not girding from bow to poop (Wescher, 1867. p.
24). Instead of transverse ropes we here find chains.
There is some visible evidence for what may be hypozomata running

round the outside of the hull. In the Akropolis museum in Athens there is a
bronze lamp in the shape of a warship (fig. 3). Midships, for the greater part
of its length, a rope is modelled, running under the gunwale (Basch, 1987,
p. 228, figs. 477-478). Here Lucien Basch warns us, that a rope round a ship
may mean: 1. a fender, 2. decoration, of which he gives a striking illustration
(fig.479). To the latter remark we should like to add, that decoration on a ship
may be a rudiment of something that once was functional, like the ornate
beakheads, reminiscent of those on 17-century sailing-ships, found on
some 1gm-centuryiron steamers.
There are also some reliefs, that show a rope running horizontally from
the epotides round the stempost (Basch, 1987, p. 435). These are the least
convincing ones, some of them seem to show ropes hanging slack, which
cannot be safely explained away as attempts at perspective, such as are
sometimes found in Roman representations of ships (Basch, 1987, figs. 923-
924, 926).
The most interesting are however some reliefs, and coins, that show 2
ropes, or one twice, wound around the hull of a ship, where it tapers and
curves upward, near the steering oars. So these ropes may actually pass
underneath the keel, but in a place that in normal circumstances does not
touch the water. Possibly this is also the place where the tightening was
effected 'endothen', whether the visible rope is a 'collar' to which the rear
end of the hypozoma is fastened, as Morrison suggests (GOS, 1968, p. 297),
or the 'spare' length of hypozoma after going round. We know 3 of those
reliefs: One from Kos (Basch, 1987, figs. 789-790), the one from Lindos
(Basch, 1987, figs. 782-785) and the one from Pergamon, on the great altar,
that in the 1880s was in its entirety transported to Berlin (fig. 1). This relief is
part of the so-called 'Telephos' frieze (Schmidt, 1961, pl. 66, center).
It is a tempting and not altogether unlikely conclusion, that these 3

glimpses of rope, laid around the ship, represent 3 parts of the hypozoma.
Where did it run? Perhaps the middle part was hidden by the parexeiresia,
as has been suggested by Kennedy (1976, fig. 1).
Then what of the bowstring-function? This is connected with the

question: Did Greek men-of-war have sheer? The iconography is not
altogether decisive, which may be the result of the best representations
being found on the convex exteriors of vases and if, as many of those
suggest, they had sheer, there could be features like oarports and
parexeiresiai, that did not follow the sheer. Parallels for that are the rowing-
frames called apostis on the galleys, that had less sheer than the gunwale,
and the gundecks of lTh-centuryships, that had no sheer at all. Now if the
trieres or other oared ship of the time had sheer, one might suppose that the
hypozoma might be made to follow leads to act as a bowstring. One can
imagine that fitting the hypozoma was a finicky job, which perhaps is implied
in an incomplete inscription from Peiraeus, that was to tell us the minimum
number of men employed at it, probably 50 (Morrison & Williams, 1968, p.
294, p. 305 n. 31). Anyway the hypozoma did pull at the rear part of the keel,
or the sternpost, forward of the aforementioned 'collar'. But, we may with
reason object; the Egyptian hogging-truss acted on both ends of the hull.
That the Classical Greek hypozoma did not is probably due to the presence
of the ram at water level, that needed a solid 'backing', incompatible with a
keel that curved up above that level. Greek ships were not 'amphielissai'
(Homeros, passim) anymore. The prow was reinforced in several ways, as
we can learn from Steffy's observations of the vestiges of wood within the
Athlit ram (Steffy, 1983, pp. 235-240), and probably from Pollux' (really
Polydeukes') 'Onomastikon'. In this Greek lexicon from the Pdcentury AD we
read (1, 89) that ships had, inside their bows, 'tropoi', that were laid over their
stemposts and stuck out on both sides. Triereis are specially mentioned as
having up to 7 of those 'tropoi', that sound very much like breasthooks. So
perhaps the whole ship was hung from the stempost.
It is highly unusual for wooden structures to hang together by means of

ropes. Where did the idea come from? Possibly the answer is in the way the
ship ends aft of the encircling ropes and the steering oars. There the keel (or
sternpost?), wales and strakes curve strongly upward to give shape to the
so-called aphlaston, that, especially with a rope round it at its base, looks like
nothing so much as like a bundle, and bundles are tied. From Egypt it is
known, that, a millennium or more before the invention of the trieres, boats
were made of bundles of papyrus and that many Egyptian wooden vessels
owed the shape of their prow and stern decoration to those prototypes
(Johnstone, 1980, pp. 69-72). Torr (1894, p. 68) and Tzalas (1995, p. 447)
remarked on the kinship of these decorations with the Greek aphlaston. The
latter moreover makes it very probable that the shape of the aphlaston
derived from the ends of Greek reed boats (fig. 2) (Tzalas, 1995, fig. 5c). It is
remarkable that the aphlaston keeps its 'bundle' construction in Hellenistic
and Roman times, whereas the Cheops ship, more than 2 millennia earlier,
had its prow and poop decorations 'copied' in solid wood, consisting of 2
halves and a cap piece and, when put together, hollow and detachable
(Jenkins, 1980, pp. 70, 98, 102, figs. 41, 47, 48, 79, 134). Alas we do not
know what, in warships of the Classical and Hellenistic period, happened
under the water line. If the double 'strap' round the poop in these reliefs is
really part or accessory of the hypozoma, it would be satisfying if the keel
and all the strakes were encircled by it. All wrecks retrieved, if complete
enough, show a stempost making an angle with the keel, including the lsola
Lunga wreck. The angle is blunt, but the garboard strake ends there (Frost,
1981, fig. 9). Above water again we find one mercantile aphlaston on an Attic
black-figured vase from the 6'" century BC (Casson, 1971, figs. 81 & 82), but
later builders for traders apparently have settled for the cheniskos (little
goose), an aquatic bird's head and neck looking backward, as a stern-
ornament. Its history seems to go back to the 8thcentury BC and also had a
forward-looking variation (Wachsmann, 1993. 11, pl. 114). Another Attic black-
figured vase shows two men-of-war with a cheniskos in an aphlaston
(Casson, 1971, fig. 90).
Was the hypozoma a rudiment of an earlier way of shipbuilding, like the

beakheads on the steamships? If so, it was a very reduced one, for the reed
boats could not be built without transverse ropes tying the component
bundles and thus passing underneath their bottoms, while hogging-trusses
would likewise have been useful. A cylinder-seal from the Djemdet-Nasr

period seems to show truss that holds a reed aphlaston in place
(Johnstone, 1980, fig. 2.5). In the same book (fig. 2.8) we see a reed-boat
from New Zealand with a long whip-like elongation of the prow that looks
very much like the ones on the ships of the Akrotiri fresco. Far be it from us
to suggest a science-fiction like connection between the Aegean and the
Pacific! This only shows the possibility of the material.
The author thanks Dr. Ellen Schraudolph-Gautier of Berlin for her readily
given help in locating the relevant slab of the Telephos relief. Likewise for
procuring for him a photograph of it (by Johannes Laurentius). Also Prof. Dr.
C.J. Ruygh for his estimation of the language of Apollonios Rhodios.
L. Th. Lehmann
Koestraat 15 B
10 12 BW Amsterdam
REFERENCES
Apollonios Rhodios. Argonautika. Athenaios Mechanikos. Peri Mechanematon. In: Wescher, C.

1867, Poliorcetique des Grecs, pp. 3-40. Paris.
Athenaios of Naukratis, Deipnosophistai.
Basch, Lucien. 1987, Le Musee imaginaire de la marine antique. Athens.
Casson, Lionel. 1971, Ships and Seamanship in the World. Princeton.
Coates, John F. & Sean MacGrail. 1984, The Greek Trireme of the 5m Century B.C. Greenwich.
Frost, Honor e autori vari. 1981, Lilybaeum. In: Notizie degli Scavi di Antichite. Suppl. al Vol.
XXX.
lsidorus Hispalensis. Origines.
Jenkins, Nancy. 1980, The Boat beneath the Pyramid. London.
Johnstone, Paul. 1980, The Seacraft of Prehistory. London.
Kennedy, Don H. 1976, Cable reinforcement of the Athenian Trireme. In: Mariner's Mirror 62.2,
pp. 158-168.
Lahn, Werner. 1992, Bauteile und Bauablauf. Die Kogge von Bremen. Band I. In: Schriften des
Deutschen Schiffahrtsmuseums. Band 30. Brernerhaven.
Morrison, John S. & R.T.Williarns. 1968, Greek Oared Ships. Cambridge.
Plato. Politeia.
Polybios. Historiai.
Polydeukes. Onomastikon.
Pomey, Patrice. 1978, La Coque. In: Tchernia, A,, P. Pomey & A. Hesnard. L'Epave romaine de
la Madrague de Giens (Var). XXXIV" supplement a Gallia, pp. 75-100. Paris.
Schmidt, Eva Maria. 1961, Der Grosse Altar von Pergamon. Leipzig.
Steffy, J. Richard. 1983, The Athlit Ram. In: Mariner's Mirror 69.3, pp. 229-247.
Steffy, J. Richard. 1994, Wooden Shipbuilding and the Interpretation of Wrecks. College Station,
Texas.
Steffy J. Richard. 1995, Ancient Scantlings: The Projection and Control of Mediterranean Hull
Shapes. In: Tropis I l l , pp.417-428. Athens.
Torr, Cecil. 1894, Ancient Ships. Cambridge.
Tzalas, Harry. 1995, On the Obsidian Trail: With a Papyrus CrafI in the Cyclades. In: Tropis I l l ,
pp. 441-469. Athens.

Vitruvius Pollio. De Architectura Libri X. Loeb edition vol. 11, 1956. Transl.& comm. Frank
Granger. London & Cambridge, Mass., Collection Bude, 1986, Livre X. Transl. &
Comm. Louis Callebat & Philippe Fleury. Paris.
Wachsmann, Shelley. 1993, Seagoing Ships in the Late Bronze Age Levant 1-11. Jerusalem.
ILLUSTRATIONS
Fig. 1 Part of the <<Telephos,,releif from Pergamon (with the kind permission of: Staatliche
Museen zu Berlin).
Fig. 2 .;'A@Aamov- of a modern reedboat under construction. (H. Tzalas, Tropis 111)
Fig. 3 Side and top view of a bronze lamp in the Acropolis Museum, Athens.
Fig. 4 Reconstruction drawing of ships with horizontal straps around one end: from Trajan's
Column, Rome and the Triumphal Arch, Orange, France. (after A. Evans, Palace of
Minos).
REMARKS ON THE HYPOZQMA
Fig. 3
ABSTRACT
MOBILITY OF CRAFTSMEN AMONG GREEK & PHOENICIAN

SHIPWRIGHTS
A WORKING HYPOTHESIS
Literary evidence from ancient Near Eastern texts attests to the mobility
of craftsmen and artisans among the city-states of that region, especially
from the Amarna Age onwards. Demand for these skilled professionals
originated in the centralized palace economies of the Bronze Age, but after
the disruptions brought about by the Trojan War and the movement of the
Sea Peoples, a new order emerged, which saw the rise of Greek and
Phoenician maritime powers. The migrations of the Greeks and Phoenicians,
whether to colonize or to trade, then saw these two groups interact both in
the eastern and the western Mediterranean. The single best example of the
transmission of ideas - opposed to products - by skilled personnel is the
diffusion of the Phoenician alphabet to the Greeks. In the realm of
shipbuilding and seamanship, Greek historical sources refer to the impact of
Phoenician navigational skills, while the evidence recovered by marine
archaeologists suggests such close affinities in the details of shipbuilding
that it presupposes independent free craftsmen were at work in the
construction of merchantmen in diverse localities. Such socio-economic
aspects of ancient seafaring complement our growing database
accumulated by the ongoing progress of marine archaeological research.
Dr. Elisha Linder

University of Haifa
Ma'agan Michael Ancient Ship Museum
Haifa 31905, Israel
ABSTRACT
KAPHEREUS AND KYME:

LATE BRONZE AGE SHIPWRECKS OFF EUBOEA ISLAND
This paper examines the literary and archaeological evidence for the
existence of Late Bronze Age shipwrecks off the coast of Euboea.
In the first part of the paper, we consider the information, contained in

Pausanias' Description of Greece (11, 23, 1, also IV, 36,6) and other sources,
about Early Greek (Mycenaean) shipwrecks off Cape Kaphereus (the
modern Cavo d'Oro) in the southeastern part of Euboea, a crucial point of
reference for navigators of all periods in the Aegean Sea. These "epic"
shipwrecks, like Odysseus', may be assigned to a Late Mycenaean horizon
(ca 1200 BC), as they were caused during the unfortunate return of the
Greeks from Troy.
The main aim of our paper, however, is to highlight the importance of a

Late Bronze Age shipwreck documented from a wrecked cargo of nineteen
(19) copper oxhide ingots of "pillow type", raised from the sea near Euboean
Kyme by Captain L. Kalamakis and now deposited in the Numismatic
Museum of Athens. This group of ingots, a chance find known since the
beginning of this century from reports by L. Pigorini (1904), J.N. Svoronos
(1906) and Arthur J. Evans (1906), has been dealt with in specialized studies
by H.-G. Buchholz (1959), G.F. Bass (1967) and Parise (1968). It now
emerges as the third largest underwater "hoard" of copper oxhide ingots in
the eastern Mediterranean, after Akroterion (Ulu Burun) and Chelidonion
(Cape Gelidonya), and certainly the earliest, dating from the earlier Late
Bronze Age.
Major questions concerning the exact find-place of the Kyme cargo as

TIANNOZ AOAOZ TROPlS VI
well as the date and circumstances of its discovery are discussed in our
paper. The precise date and typology of the Kyme oxhide ingots are further
considered, using new comparative material from shipwrecks and sites
excavated on land, and in the light of recent surveys of the metal trade in the
Mediterranean during the Late Bronze Age.
In line with the approach followed by A. Sampson (1980, 1981) and E.

Sakellaraki (1984) in their summary treatment of the Kyme find, we shall
proceed towards a fuller assessment of the cargo within the context of
Minoan seafaring and trade activity in the North Aegean in the era of the
Thalassocracy (i.e. in the Middle Minoan Ill - Late Minoan I period). Special
attention will be drawn to the existing evidence, both literary and
archaeological, for Minoan presence in Euboea and on the smaller islands
of the North Aegean (e.g. on Skopelos and Samothrace). Late Bronze Age
coastal sites in the wider area of modern Kyme that may possibly be linked
to the wreck will also be taken into account.
Finally, brief reference will be made to another wrecked cargo, that at

Sheytan Deresi (Devil Creek) off the Carian coast of Turkey (see G.F. Bass
1974, 1975, 1976, and now R. Margariti), largely consisting of plain pottery
of arguably Minoan character, and, like the Kyme cargo, also attributable to
the broader horizon of the Thalassocracy. This ceramic cargo should now be
considered in relation to abundant Minoan finds in the Dodecanese (e.g. on
Rhodes, Karpathos, Kasos, Telos, Kos) and especially against the
background of the spectacular Minoan discoveries made, over the past two
years, by W.-D. Niemeier at the Cretan colony of Miletus in lonia.
Prof. Yannos G. Lolos

Dept. of History and Archaeology
University of loannina
Greece
MORE EVIDENCE ABOUT NEOLITHIC INLAND CRAFT
(DISPILIO, LAKE KASTORIA)
Introduction
A paper presented at the 5'hShip Construction Symposium described a

feature, which was discovered in 1992 on the prehistoric site near the village
of Dispilio, Lake Kastoria, and which was interpreted as a dugout gunwale in
outline (Marangou 1993). The find is still there, but further work has not yet
been resumed on the precise spot.
Nevertheless, with the progress of the excavations, more evidence about

neolithic craft at Dispilio is now available: boat models and/or boat-shaped
vases, as well as some more features probably showing boat gunwales in
outline. Of course the entirety of the Dispilio material is still under study and
the excavations continue, therefore the comments presented here are
provisional.
I will just recall the location of the site, in North-Western Greek

Macedonia, on the southern bankside of the lake of Kastoria, the short
crossing distance to the opposite shore, and the present marshy
environment (shallow waters and reeds) in the area of Nissi, where the
excavations take place, under the direction of Professor George
Hourmouziadis (University of Thessaloniki, Hourmouziadis 1996).
Evidence comes from different parts of the site. Boat models and boat-
shaped vessels are concentrated mostly in one area (centre-north),
occasionally elsewhere, boat outlines in a quite distinct area (south-east) of
the trenches under excavation'.
Besides, evidence is not contemporaneous: boat models come for the

moment from earlier strata than boat outlines. According to provisional
estimations from the ceramic evidence (cf. Anagnostou et a/. 1997),
stratigraphy shows an early Late Neolithic date, preceding the Dimini phase,
and even the end of Middle Neolithic2for the earlier models: pottery sherds
with typical painted decorative patterns and Danilo "cult vessels" fragments
were found in corresponding early strata. Boat outlines were found in the
upper strata, dating most probably from the end of the Late Neolithic period;
black-topped and black-, brown- and red-burnished pottery was found in the
strata of both boat outlines.
CHRISTINA MARANGOU TROPIS VI
Description of boat models and boat oulines
A first approach of the finds reveals a diversity of sizes of models and

boat-shaped vases3and varied types of both boat outlines and models.
According to the evidence available in summer 1996 (a total of about ten

objects, mostly fragments), models had an estimated initial overall hull
length of 6 or 7 to 22cm; a maximum breadth of 2.5 to I l c m ; and a
maximum depth of 2 to 7.2cm. Boat-shaped vases (oval-shaped, often
shallow, elongated containers) reached an initial overall length of 12.86 to
29cm, a maximum breadth of 7.44 to 13cm and a maximum height of 3.75
to 6.66cm. Length to breadth ratio isin average about 2:1 for both models
and vases. Dimensions and lengthlbreadth ratio are comparable to those of
some Neolithic or Chalcolithic boat models (cf. Marangou 1991 and idem
1996), such as the one from Osikovo (Kodzadermen Chalcolithic culture,
length 22cm, breadth I l c m ) (Frey 1991: 197, fig. 2) or Telis-Redutite
(contemporary to the Krivodol-Salcuta culture, length 19.2cm) (Frey 1991,
fig. 1, 2).
Types of complete models are difficult to establish, since most of them

are fragmentary, preserving only one end. The latter can be rounded or
roughly sub-rectangular, of trapezoidal transversal section with a flat bottom,
which could help the models lay on a horizontal surface. They presumably
belonged originally to asymmetric boat types, possibly dugouts. Some
fragments show a rounded hull and extremity, probably indicating variety of
types of the originals.
A boat model (end of Middle Neolithic) presents a rather symmetric

breadth shape (fig. l ) , a flat bottom, and an irregular gunwale, while the
bottom of the hull becomes thicker near one pointed, slightly raised end.
Near the same end, a lateral external protrusion is puzzling, but it could also
be considered accidental, because of the rough modelling. Furthermore, a
cavity can be seen in the interior of the same pointed extremity. Heights of
stern and bow would be almost equal. Both sides of the boat are higher
amidships.
Boat outlines, now, have an overall length of 3.30m and 3m, and a
maximum breadth of 0.80m and 0.73m respectively, with a 1ength:beam ratio
of 4.12 and 4.10. The shapes of these small boat outlines (fig. 2), either
asymmetric with one pointed and one sub-rectangular end, or symmetric
with pointed ends, correspond to those of boat models, either from Dispilio
(DISPILIO. LAKE KASTORIA)
or from other sites. The ends of asymmetric outline A (the northern one)
could be compared to various fragments from Dispilio or, among others, to
the above-mentioned Osikovo model, outline B to the Selevac model
(Tringham et a/. 1990, pl. 10.5) or to the previously described Dispilio model
(fig. 1).
Boat types
The dugout seems to be the most obvious watercraft in a Neolithic lake

environment, according to technological possibilities, availability of tree
trunks and evidence from excavated dugouts (cf. for example Arnold 1995,
1996) and most Neolithic boat models (Marangou 1991; idem 1996;
Hockmann 1996). However some characteristics of the Dispilio models
seem exaggerated, even if we take into account the fact that they are simply
clay imitations, not necessarily exact scale copies.
The exceedingly large breadth of some Dispilio models with a

1ength:breadth ratio of 2:1 is particularly astonishing. It has been suggested,
about Balkan examples, that two or three trunks or longitudinal elements had
been joined together for the construction of large extended dugouts (Frey
1991: 196, 197, fig. 2 and 199, fig. 4). A similar explanation has been
proposed by Hockmann (1996, fig. 4, 2) about the Tsangli Thessalian model
(Marangou 1991, pls. IV, Vllb-IX; idem 1996, fig. 2-3). However, the also
Neolithic but later in date boat outlines of Dispilio have a more credible ratio
of 4.1:1.
In addition to large breadth, the pointed extremities, the irregular

gunwale and the uneven thickness of the sides of the Dispilio model (fig. 1)
are striking, although according to its section it could be a dugout. If the
Dispilio model represents a dugout, and if its morphological details are
deliberate, the thinner hull bottom towards one end could suggest a boat
with a fish-tank; the thinner walls would help buoyancy, if water was entering
inside (Arnold 1983: 272). Pointed ends, on the other hand, characterise
expanded dugouts, yet the latter were technologically possible from the
Bronze Age onwards (McGrail 1987: 87). Besides, the transversal hull
section of the Dispilio model is trapezoidal, not rounded, as would be the
case on expanded dugouts (see for example Arnold 1995: 150-151);
furthermore, their sides would be low - the Dispilio model being, on the
contrary, comparatively deep - unless, of course, strakes had been added.
It is interesting to note the similarity of the pointed and slightly raised ends
CHRISTINA MARANGOU TROPlS VI
of a Neolithic clay model from the Lake of Bracciano (illustrated in

L'archeologo Subacqueo 1, 3, 1995: 3)4.
Besides the dugout, alternative possibilities5should therefore remain

open; other materials available at the site, such as reed and hide, attested
by important quantities of big animal bones, could also have been used for
boat construction6.In fact, even the papyrella sections are not very different
from the ones of the Dispilio model (cf. Tzalas 1989: 11; idem 1995: 463, fig.
1). Besides, some Neolithic watercraft representations have occasionally
been interpreted as hide boats, such as, for example, the graffito from the
island of Hvar in Dalmatia (Bonino 1983: 66 and fig. 78) or a number of clay
models from Hungary and Germany (Hockmann 1996: 37, fig. 9, cf. Basch
1987, Cypriot clay models).
As it has already been said, work has not been resumed in the trench of
boat outline A since its discovery. Boat B, where digging continued, is not
preserved. During excavation, the dark, almost black contour of gunwale B
on the soil was clearly differentiated from the remaining grey area. Some
other dark patches tended to disappear after a while. There were no more
traces of the remainder, which should have disintegrated. As a matter of fact,
wood has not, generally, been preserved in the upper strata. Nevertheless,
the northern part at least of the outline seemed to "move" towards the west
for a depth of some centimeters during excavation; it is not impossible that
this was due, not to preservation reasons (as if the bottom of the boat had
disappeared, see further), but rather to its being the wooden frame of a hide
boat7,the only part of it to have left some traces in the soil.
Context and surroundings of boat models and boats
The context of boat models and boat-shaped vases includes probable

house floors and garbage areas. One of the oldest (end of Middle Neolithic)
fragments has been found, among large quantities of animal bones and
sherds, in an uncovered space, possibly a garbage pit. This area was
separated by means of a stone construction from a probably covered,
internal space, which contained a lot of big (restored) vases and small
objects (figurines, rings and other ornaments, including a stone fish-
pendant).
An intact boat-shaped vase comes from a stratum consisting of

disintegrated and burnt organic material, postholes, wooden and clay
elements, presumably belonging to a floor and/or floor-structures. The

stratum contained several large vases, mostly coarse.
Most of the other more or less fragmentary boat-shapedvases and boat

models of different sizes and types were concentrated in an area of
approximately 8m2, during two or three successive strata of early Late
Neolithic date (Tsangli and Arapi phases-before Dimini). These also
contained a lot of fine and coarse pottery, restored vases of different sizes
and types, for cooking and storage (small jars, fruitstands, bowls etc.), some
human figurines, micrographic vases, burnt and disintegrated wood remains
from posts, structures and floors, a lot of animal bones and some fish
remains, as well as a number of postholes, possibly from light structures.
This discovery of boat models in a Neolithic house is not unique. In

Middle Neolithic Selevac a concentration of pottery, tools, figurines and
miniature objects, including a boat model, has been observed in association
with the floor of a posthole house (House 4), related to an oven (Tringham
etal. 1990, fig. 4.28). It is possible that a similar situation, a scene of several
figurines and a boat(?) was represented in miniature (Marangou 1996: 280-
281) in an early Late Neolithic house model founds near the hearth of a real
house in Thessaly. The role of boat models as images was apparently as
important as that of the images of human beings (figurines). However, it is
impossible to ascertain, for instance if this role was protective or magic.
Besides, this does not necessarily exclude a parallel practical function of the
artefacts.
Both boat outlines were situated outside a settlement area, delimited, on

the east and the north, by a narrow "ditch" or "channel" (maximum width
15cm, depth 5cm). A portion of this feature is visible at a short distance to
the west of outline B (fig. 2). Although no postholes have been observed in
close association with this ditch, some clay elements (bricks?) were found in
proximity. If these features are the remains of an enclosure, the boats should
be located in the immediate eastern periphery of the settlement or, at least,
of an area marked off by boundaries or a barrier, but of course this has still
to be confirmed, since the area to the east of the outlines has not yet been
investigated.
The outlines, of approximately the same size, are separated by a

distance of about 2m and present some persistent features. Attempts at
interpretation of the relationship of the latter with the boats are mere
hypotheses for the moment. All the same, whatever their meaning, their
occurrence on both boat outlines seems significant:

1) the outlines have the same South-EastlNorth-Westorientation.
2) a pair of small (diameter 5-10cm) postholes on the external limit of the
eastern side (the one further away from the "enclosure"), which is the
starboard side, at least for the first boat (A), separated by a constant span
of 65-70cm. On the second boat (B) these postholes are arranged at
approximately 1.05m and at 40cm from the southern end. On outline A,
two sets of postholes repeat the same pattern, at a distance of ten
centimeters, at, respectively, 1m and 30cm from the stern for the first set,
90cm and 20cm for the second set.
3) Near the middle of the western side (which is the port side for boat A),
inside the gunwale outline, an oval, dark brown feature, containing two
later(?) postholes, on boat B, and, almost on the same spot, on boat A, we
note a hard, white circular feature. The distance of the centre of these
features from the southern end of the boats is respectively about 1.40m (A)
and 1.90m (B).
4) Remains of clay elements are found in proximity to both outlines; those
situated close to boat B could be related to the above-mentioned
"ditch"/enclosure.
There are some alternative interpretations for this ensemble and its
location, which should remain hypothetical, since the exact position of the
lake at the time is not known yet; it is situated to the north of the excavation
now. Boats could simply have moored on the shore, close to the settlement.
Some of the dugouts at Federsee, including a Late Bronze Age one, were
found on a platform or landing place, to which access was only possible by
boat. The bow of the dugout must have been moored on a post structure,
obliquely to the direction of the river (Paret 1930: 80). Typical beaching
places for prehistoric boats were situated on the shore, where they were
drawn up at a short distance from the water. The sole signs of such landing
places are provided by the posts, to which boats were fastened by means of
a rope. Such a vertical post was fixed directly near the bow of a dugout at
Wustrow (Ellmers 1973: 60). The extremity of the northern end (bow?) of
boat B at Dispilio is touching the southern part of a large posthole ( 30 x
22cm). The corresponding post could in fact have been used for mooring, in
the same way as for modern local monoxyla (flat planked boats of roughly
dugout shape). The same could have happened with boat A, to which some
circular features are related (cf. Marangou 1993).
Furthermore, the repeated occurrence of pairs of postholes on one side

of the gunwale, at a constant distance from one end of the boat, seems
deliberate. If these small posts were not used to secure the boat, they could
have been used to fix some kind of nets (cf. Torke 1993). They could belong
to the prongs of a bag net, which rest on or are fixed in the bottom in shallow
water during fishing (Stewart 1977). The boat is anchored or tied above the
fishing area. Rows of stakes can otherwise be the remains of fish traps or fish
fence weirs, situated in a fishing ground, just off the settlement, also
functioning as a dock for the beaching of boats ("fishing location", Andersen
1985: 55). Two mesolithic dugouts were found in such a location in the reed
swamp outside Tybrind Vig (Denmark). The area also contained settlement
debris (Andersen 1987: 89, fig. 3). As a matter of fact, under the stratum of
boat outline 6 , an intermediate stratum contained the remains of a light
construction of clay and reed - possibly a fence-fishing trap. The following
stratum repeats again, in the same area, a third boat outline (C)' with a
slightly different orientation.
Alternatively, the boats might have sunk or been abandoned on the spot.
Changes of water level and periodic inundations, frequent in lake
environments, could be involved. It is not excluded either that boat 6 had
been inverted, in this case accidentally; its bottom would not have been
noticed during excavation before arriving to the better preserved remains of
the gunwale, or it would have disappeared (cf. above).
On the other hand, half-finished dugouts were frequently kept under

water in a convenient place off the settlement, for a period of time, even for
some years, awaiting for the final stage of the construction process (Paret
1930: 114; Ellmers 1973: 59). This submersion was said to soften the wood
or make the boat more durable (Rasmussen 1953: 43). Sometimes the
dugout, when not used, was also kept under water, in order to prevent wood
from drying and splitting (Andersen 1987: 94). Wooden posts of a small
diameter (8cm) often marked the place over water level, so as to be able to
find the boat again. When two posts were fixed with a distance between
them, close to the side of the boat, then its length direction was also
indicated. In one case (dugout of Wustrow), remains of seven posts were
found; it is not certain whether all of them were used simultaneously, or
whether some of them were replacing broken posts (Ellmers 1973: 59-60).
The process could be repeated on the same spot for each new boat. The
second pair of postholes close to boat A at Dispilio is then understandable.
Some other postholes have been discovered in the trench of boat 6 ,

including very near the outline, and it is possible that some of them were also
related to boat 6. A few others might also be related to other boats, which
CHRISTINA MARANGOU TROPlS VI
were not preserved. One of the interpretations which have been proposed
about a stone age Danish dugout surrounded by vertical poles was that it
had constituted part of a burial ceremony: a human skeleton found nearby
would have lain originally in the boat (Christensen 1990, fig. 9). Graves in the
shape of a boat or using a real boat or a part of a boat, occasionally inverted,
are well known in Northern Europe. However, the available data at Dispilio
do not permit to advance, till now, funerary associations for the outlines.
Submersion of dugouts under water could be achieved by means of

stones. Their weight kept the dugout in place. There might be several heaps
all the long of the boat (Ellmers 1973: 59). Accumulations of stones near
lakeshores can thus be found, originally used for this reason (Paret 1930:
114, 115). Some stones were found in a row near boat B at Dispilio, but their
direction is divergent (East-West), although it is not excluded that they had
been used in such a way. Stones can otherwise be used on a dugout as
ballast, stabiliser or counter-weight (Arnold 1983: 272; Andersen 1987: 94).
As with overturned dugouts, if they had not accidentally capsized (see

above), they may have been turned upside down during the construction
process, in order to work on the underside of the trunk. After the felling of the
tree and hollowing the trunk in winter, the hollowed trunk was transported to
the edge of the lake and left on the bank till spring, the hollow part towards
the soil; it stayed upside down in order to work on the exterior of the bottom.
It was at the end of the process that the hollow part was turned again
upwards to dry (Arnold 1983: 272).
Therefore the different interpretative possibilities of the outlines'

surroundings include a mooringllanding space, a fishing location, a boat-
building area, and a garbage or waste ground, where the boats out of use
had been abandoned. The solution depends naturally on our understanding
the character of this area in connection to water: was it dry land, a lake
bottom, or an amphibian zone? The progress of the specialists' study should
give an answer to this question and help us interpret the environment of the
boats and hence their purpose.
Functions, uses and users(?) of boats and boat models
Functions of Neolithic inland boats could include food acquisition,

communication and conveyance. Bulky material, such as large posts, or
reeds, frequently used in constructions at Dispilio, needed to be carried.
Fishing and waterfowl hunting from prehistoric dugouts (Ellmers 1973: 62)
are often attested. Fish vertebrae and fishing gear, even a fish-shaped stone
pendant have been discovered at Dispilio; bird bones and probable fishing
weights (Hourmouziadis 1996: 44, fig. 13) were found, among other places,
in the trenches with the boat outlines.
Granted the size and weight of the adapted trunks, it seems certain that
dugout-building was a collective task. Yet hide-boat- and possibly reed-
bundle-boat-building would also need more than one individual. The
relatively small size of the Dispilio outlines implies that two or at the most
three individuals could be carried aboard at the same time. This is the
maximum number of paddlerslpassengers (usually fishermen) in present
monoxyla of the region, which have comparable dimensions. The total
number of boats used in the settlement simultaneously is unknown, but one
could suggest a collective use - if not ownership - of the boats by the
inhabitants, or at least by several social groups, since the collection of
dugouts found in each one of the Neolithic or Mesolithic sites till now is
restricted and does not permit to maintain that every household possessed
its own watercraft.
The function of boat-shaped (oval) vases, as well as their identification,

is open to any suggestion. Boat-shaped vases are not necessarily related to
fish and fishing either as food (fish) containers, or as representing the means
for the activity. As a matter of fact, some Neolithic containers have an animal
and even human form, and such figures are occasionally applied as
appendices on ceramic vessels, including at Dispilio; this would not
necessarily mean that their contents were related to their form. The shape of
a boat could nevertheless be adapted, as a convenient - and possibly
meaningful - form of container.
The action of fire is sometimes attested on boat-shaped-vases: the

exterior of one of them is carbonized; circular marks of burning can be seen
on the base, inside and outside a second boat-shaped vase. A convincing
use hypothesis is, though, still missing, in the absence of any analyses.
As boat-shaped vases, boat models also come from house floors and
garbage; consequently they seem to have been connected to everyday
activities, although we can not guess their precise function during the latter.
Besides, apparently they were not present in all buildings. On the contrary,
they seem to be related only to a few households, and there is only one
important concentration in a particular space. However, it is impossible for
the moment to attribute this particularity to specific economic or social

activities of some households or social groups. Ownership of models, as
well as of real boats, is elusive.
It is impossible till now to attribute precise uses to boat models of

different sizes and types, in their quality as implements, or to assess exactly
their importance and symbolic meaning, as miniatures. This difficulty also
occurs concerning other representations of human, animal or inanimate
prototypes, modelled in various scales. Moreover, to their morphological
diversity as artefacts and as images of life-size originals and to their
presumed multi-functionality as symbols is added the large spectrum of
economic, social and possibly ritual roles of their prototypes.
Conclusions
Concluding, we may stress that boats, real as well as symbolic, were

inter-linked with various facets of life at Dispilio. They attest a variety of
watercraft types as well as of their possible uses in a Neolithic lake
environment. Their micrographic representations apparently fulfilled some
symbolic needs.
It is obvious that boats conveyed some considerable significance, both

as images, and as real structures, yet their exact social and ritual meanings
remain unknown. The continuation of the excavations is expected to give
answers to many open questions. Hopefully, some day, real dugouts will be
excavated, not only their reduced copies and fleeting outlines ...
Aknowledgements
The author is grateful to the following for most helpful comments and
discussions on primitive craft during or after the Lamia Ship Construction in
Antiquity Symposium: Ssren Andersen, Marco Bonino, Charlie Christensen,
0le Crumlin-Pedersen, Avner Raban, Richard Steffy. A lot is due to a study
visit at the Centre of Maritime Archaeology in Roskilde in April 1997 and to the
most kind help and advice of 0le Crumlin-Pedersen. Finally, warmest thanks
are addressed to George Hourmouziadis for entrusting me with the study of
the Dispilio material, and for stimulating discussions; and to colleagues of the
Dispilio excavations for providing information about the trenches excavated
and/or supervised by them: A. Almatzi, I. Anagnostou, Th. Kougoulos, G.
Mortzou, F. Seroglou, K. Touloumis. T. Yagoulis.

Christina Marangou
Rue du Bailli 95
1050 Brussels
Belgium
NOTES
1. An area of 29 x 24m (east-west x north-south) was under excavation or prepared for

excavation till summer 1996.
2. This was confirmed later by C14 dates for the lower strata. Unfortunately, no radiocarbon
dates are available for the upper strata as yet.
3. It is sometimes difficult to distinguish between models and vases. Generally, boat models
show some features of real boats, not just an elongated oval shape. Nevertheless, it is
possible that some models were used as containers.
4. 1 am grateful to Marco Bonino for suggesting this reference.
5. This aspect has been examined in Marangou, Chr., Neolithic craft: evidence about boat types
and uses, Proceedings of the 8 Symposium of the Greek Archaeometric Society, Athens,
November 1996 (in press) and idem, On the omnipresence and elusiveness of the Neolithic
Aegean watercraft, Foutth Annual Meeting of the European Association of Archaeologists,
Thematic Block Ill, Interpreting the Archaeological Record, Session 2, Maritime Archaeology,
Goteborg, September 1998, Abstracts: 85-86.
6. Further discoveries on the site seem to corroborate this hypothesis, at least for the Middle
Neolithic.
7. 1 wish to thank Avner Raban for suggesting this possibility.
8. Gallis, K., 1985, A Late Neolithic foundation offering from Thessaly, Antiquity LIX, 20-24, pls.
XIV-XVI.
9. The third boat outline, discovered under outline B, was studied later, but it has not been
included in this paper. It has been presented in Marangou 1998 (see note 5).
REFERENCES
Anagnostou, I., S. Thomaidou, G. Stratouli, M. Sofronidou, K. Touloumis, 1997, Excavations of

Dispilio, Kastoria: the chronological problem, Proceedings of the Archaeological Work in
Macedonia and Thrace Conference 1993 (7). Thessaloniki 1997, 13-17.
Andersen, S., 1985: Tybrind Vig. A preliminary report on a submerged Erteborlle settlement on
the West coast of Fyn, Journal of Danish Archaeology 4, 52-69.
Andersen, S., 1987: Mesolithic dug-outs and paddles from Tybrind Vig, Denmark, in Acta
Atchaeologica 57, 1986 (1987), 87-106.
Arnold, B., 1983: Les Dernieres pirogues monoxyles d'Europe Centrale, Helvetia Archaeologica
14, no 55/56, 271-286.
Arnold, B., 1995, 1996: Pirogues monoxyles &Europe Centrale. Constructicn, typologie,
evolution, tomes 1, 2. Archeologie Neuch&teloise 20-21, Musee Cantonal d'Archeologie,
Neuchatel.
Basch, L., 1987: Le Musee imaginaire de la marine antique, Hellenic Institutefor the Preservation
of Nautical Tradition, Athens.
Bonino, M., 1983: Le imbarcazioni monossili in Italia, in Bolletino del Museo civico di Padova
( W I I ) , 51-77.
Christensen, Ch., 1990: Stone Age Dug-out Boats in Denmark: Occurrence, Age, Form and
Reconstruction, in Experimentation and Reconstruction in EnvironmentalArchaeology, D.E.
Robinson (ed.), Oxford, Oxbow Books (Monograph 5), 119-141.
Ellmers, D., 1973: Kultbarken, Fahren, Fischerboote, Vorgeschichtliche Einbaume in
Niedersachsen, Die Kunde N.F. 24 (1973), 23-62.
Frey, 0.-H., 1991: Varna-ein Umschlagplatz fur den Seehandel in der Kupferzeit?, in J.
Lichardus (ed.), Die Kupferzeit als historische Epoche, Symposium Saarbrucken und
Otzenhausen 1988, Teil I, Saarbrucker Beitrage zur Altertumskunde Band 55, R. Habelt,
Bonn (1991), 195-201.
Hockmann, O., 1996: Schiffahrt in der Steinzeit, in Omaggio a Dinu Adamasteanu, a cura di
Marius Porumb, Clusium (Cluj-Napoca), 25-60.
Hourmouziadis, G.H., 1996: Dispilio (Kastoria). The prehistoric lake-side settlement. Codex,
Thessaloniki.
Marangou, Chr., 1991: Maquettes d'embarcations: les debuts. Aegaeum 7, Thalassa, Actes de
la 3" Rencontre Internationalede I'Universite de Liege, Calvi 1990, R. Laffineur and L. Basch
(eds.), Liege 1991, 21-42, pls. Il-IX.
Marangou, Chr., 1993 (in press): Evidence about a Neolithic dugout (Dispilio, Kastoria). Tropis
V, Proceedings of the 5m lnternational Symposium on Ship Construction in Antiquity,
Nauplia 1993, Hellenic lnstitute for the Preservation of Nautical Tradition, H. Tzalas (ed.).
Marangou, Chr., 1996: From Middle Neolithic to Early Bronze Age: consideration of early boat
models. Tropis IV, Proceedings of the 4mlnternational Symposium on Ship Construction in
Antiquity, Athens 1991, Helienic lnstitute for the Preservation of Nautical Tradition, H. Tzalas
(ed.), Athens 1996, 277-293.
McGrail, S., 1987: Ancient boats in North Western Europe. The archaeology of water transport to
AD 1500. Longman, London and New York.
Paret, O., 1930: Die Einbaume im Federseeried und im ijbrigen Europa. Praehistorische
Zeitschriff MI, 76-116.
Rasmussen, H., 1953: Hassela-Egen. Et bidrag ti1 de danske stammebades historie. Kuml,
Arbog for Jysk Arkaeologisk Selskab (Arhus), 3, 15-46.
Stewart, H., 1977: Indian fishing: early methods on the northwest coast. University of
Washington Press, Seattle.
Tringham, R., and Krstic, D. 1990: Selevac. A neolithic village in Yougoslavia. Monumenta
Archaeologica 15, UCLA lnstitute of Archaeology. Los Angeles, California.
Tzalas, Harry, 1989: 0 0 6popoq TOU o~tdtavou, ApxaloAoyia 32, September 1989, 11-20.
Tzalas, Harry, 1995: On the obsidian trail. With a papyrus craft in the Cyclades. Tropis Ill,
Proceedings of the Zdlnternational Symposium on Ship Construction in Antiquity, Athens
1989, Hellenic lnstitute for the Preservationof Nautical Tradition, Athens 1995, ed. H. Tzalas,
441-469.
FIGURES
1. Clay boat model from Dispilio (end of Middle Neolithic) (photograph by the author). Length:
20.5cm.
2. Boat outlines A (to the North) and B at Dispilio (end of Late Neolithic?) (rough sketch based
on documentation from the Dispilio excavations).
MORE EVIDENCE ABOUT NEOLITHIC INIAND CRAFT
(DISPILIO.
-- LAKE
-
KASTORIA)
Fig. 1
6--
I-; a-7
m-
2 ,
,
,
..
y
I;( r-
Fig. 2
A SHIPBUILDING SCENE ON AN UNPUBLISHED RELIEF
At the recent exhibition Seafaring in Classical Antiquity in the Allard

Pierson Museum in Amsterdam (October 1995-March 1996), part of a
sarcophagus depicting a shipbuilding scene was on display, thanks to the
kindness of a private collector. (fig 1) The relief sheds light on various
aspects of methods of shipbuilding in the Roman empire, which will be the
subject of this paper.
Characteristics
The relief is 97.3cm long, 23.8cm. high, and 6.9cm. thick. It is made of
Carrara marble, which was often used in imperial times. It dates from the
third century AD. On the left side of the relief we see two men standing inside
a boat hammering and chiselling, on the right side two men are sawing a
plank with a frame-saw.
According to the information provided by the owner, the relief was found
built into a wall in Rome. From the third century AD onwards it was the
practice in Rome and other Italian cities to use ancient marble reliefs in
public buildings, monuments, triumphal arches, and Christian basilicas. This
practice is designated by the Latin term spolia.
Faber navalis
Since our relief shows a construction scene, we may assume that the
sarcophagus of which this relief was part, constituted a funerary monument
for a faber navalis, or shipwright. Several other funerary monuments for
shipwrights have been found in the Roman harbour towns of Ostia, Ravenna,
Aquileia and Pisa. Most of these depict the tools of a carpenter, or some
other image related to his occupation, most commonly a ship1.
Only the well-known relief of P. Longidienus, which I shall discuss below,

shows human figures building a ship. The other reliefs rarely show more than
one person. This is not surprising, as most carpenters were working in small
FIK MEIJER TROPlS VI
workshops, and were generally of low social status. They did not earn
enough money to finance the costs of a conspicuous burial monument. For
this reason the iconographical information on craftsmen in general, and on
carpenters in particular, is scarce. The relief we see here differs from others
in that it shows four persons engaged in the building of a ship. This brings
me to the suggestion that the shipwright who was honoured with this
monument must have been a man of some means. If it can be accepted that
the relief shows scenes from the life of the deceased, it is reasonable to
suggest that he was the owner of a shipyard employing several craftsmen2.
The tools
On the left side of the relief, inside the ship, there are two men working
with hammers and chisels. The Romans used different types of hammers
(malleus or malleolus) with metal or with wooden top edges. The fact that we
see in this fragment hammers being used together with chisels, recalls the
practice described by Livy (27,49,1): Fabrile scalprum cum malleo habebant
('They used a chisel with a hammer'). Only Columella (3.6.3) informs us on
the form of hammers. An original hammer with a metal edge is known from
Herculaneum3, and an original wooden hammer was found in Main?.
Although the practice of using hammers and chisels was widely known in the
ancient world, it was depicted only rarely5.
This makes our scene more interesting. Woodworkers used chisels

(scalprum) to cut mortises for the mortise and tenon construction, and to
mould decorations and bowed surfaces. There were different types of chisel,
but the ancient authors do not give information about their shapes; the
number of chisels found in ancient cities is limited6.
On the right side of the relief two men are working with a frame-saw
which had already been in use in the Roman empire from the first century
AD, and probably earlier7. The frame-saw on the relief has a narrow blade,
which is stretched between the lower ends of wooden side-pieces.
Interpretation
The main point of interest of this relief is that it enables us to venture a

few suggestions as to the methods of shipbuilding in the third century AD.
Most wrecks of ancient freighters found in the Mediterranean suggest that
ancient ship builders started by constructing the hull. This suggestion is

confirmed by the Longidienus relief, the only iconographical evidence we
had until now of ship construction in classical antiquity. (fig. 2) L. Casson,
who studied this relief, set the scene in the context of the results of
excavations of ancient wrecks. The carpenter is working on a frame near the
completed hull and is ready to put this frame inside the ship8. This
interpretation is not only confirmed by archaeological data, but also by the
step-by-step description of how Odysseus built a ship on the island of
Calypso (Odyss. 5,244-257). This text shows that the mortise-and-tenonand
shell-first method of Graeco-Roman shipbuilding was already in existence in
the eighth century BC.
Over the last decades some scholars have criticised the assumption that
the shell first-method was the only method of shipbuilding in the ancient
Mediterranean world. Some wrecks dating from the first to the third centuries
AD suggest a different working method. Some frames appear to have been
pre-erected before the completion of the upper parts of the planking. Already
in 1972 L. Basch suggested that there may have been mixed building
processesg, and more recently P. Pomey has discussed the problems
connected with the shell-first conception and skeleton process in ancient
Mediterranean shipbuildinglo.
Cuomo and Gassend, the excavators of the wreck of La Bourse de

Marseille which dates from the end of the second or the beginning of the
third century AD, have argued that in this case a construction of frames was
placed inside the ship before the planking strakes were set1'. Other wrecks
show traces of being constructed according to the shell-first system, with an
admixture of 'skeleton-first' elementst2.
Literary authors do not often display a great interest in ships and

seafaring, but we can find some literary evidence for other methods of
construction. Ovid's description of the ship that took Paris and his beloved
Helena away from Laconia suggests that its method of construction had not
been shell-first (Heroides, XVI, 112). Ovid writes Texitur et costis panda
carina suis. Although this is a difficult verse to translate, in the context of the
verses 110-116 it must mean that after setting in some principal frames the
hull was constructed with a mixed construction of alternating frames and
strakes ('construction alternee')13.The third-century AD author Athenaeus in
his description of the Syracusia, the gigantic ship that Hiero II built in the third
century BC, gives a further indication of a method of construction which
deviates from the traditional shell-first method (DeipnosophistaiV. 207a-b).
FIK MEIJER TROPlS VI
In this case the shape of the hull would seem to have been determined
entirely by the frames, and not by the strakes14.
Does our relief depict a traditional shell-first construction, or a mixed

construction? Looking at the ship we can see two projections above the
highest plank of the hull near the stem and the stern. These could be some
sort of bitts to which mooring lines were attached. In that case the scene
would not inform us about the construction. The plank, which is being sawed
on the right side of the relief, could be part of the ceiling, the internal
structural planking of the huIll5.If we, however, take these projections as the
top edges of frames, another interpretation is possible. In that case our relief
depicts the construction of a boat by a method which we might describe as
a 'mixed construction'.
Fik Meijer
Archaeologisch-Historisch lnstituut
Oude Turfmarkt 129
1012 GC Amsterdam
NOTES
1. See G. Zimmer, Romische Berufsdarstellungen, Berlin 1982, 33-34.

2. Cf. L. Casson, 'Documentary evidence for Graeco-Roman shipbuilding (P.
Flor. 1 69)', Bulletin of the American Society of Papyrologists 27 (1990), 15-
19, which is a papyrus record of payments to shipwrights and sawyers
working on a boat.
3. See G. Maggi, Ercolano, fine di una citta, Naples 1985, tav. 51.
4. W. Gaitzsch, Eiserne Romische Werkzeuge. Studien zur Romischen
Werkzeugkunde in ltalien und den nordlichen Provinzen des lmperium
Romanum, Oxford 1980, 90-9 1.
5. The best known picture of a carpenter with a hammer and chisel is a wall
painting in the Casa dei Vettii in Pompeii, triclinium P, north wall, in the
scene with Daedalus and Pasiphae.
6. See S.T.A.M. Mols, Houten meubels in Herculaneum. Vorm, techniek en
functie, Nijmegen 1994, 92 (an English translation is forthcoming).
7. A painted example of a frame-saw comes from the face of a workshop in
Pompeii, now in the Museo Nazionale in Naples, inv. 8991; cf. Mols, o.c.,
92-93.
8. L. Casson, Ships and Seamanship in the Ancient World, Princeton 1972,
203-204.
9. L. Basch, 'Ancient wrecks and the archaeology of ships', International
Journal of NauticalArchaeology 1 (1972), 39-50.
10. P. Pomey, 'Shell conception and skeleton process in ancient
Mediterranean shipbuilding', in: Chr. Westerdahl (ed.), Crossroads in
ancient shipbuilding, Oxford 1994, 125-130.
11. J.M. Gassend and J.P. Cuomo, 'La construction alternee des navires
antiques et I'epave de la Bourse de Marseille', Revue Archeologique de
Narbonnaise 15 (1982), 263-273; contra P. Pomey, 'Principes et methodes
de construction en architecture navale antique', in: Cahiers d'histoire 33
(1980), 297-312.
12. Cf. Pomey, o.c., 125-130.
13. Cf. F. Meijer, 'Ovide Heroides XVI 112 et la construction navale romaine',
Mnemosyne 93 (1990), 450-452.
14. F. Meijer and A. Wegener Sleeswyk, 'On the construction of the
'Syracusia' (Athenaeus V. 207 a-b)', (forthcoming in Classical Quarterly
996, vol. 11).
15. This suggestion was raised in the public discussion after my lecture in
Lamia.
FIK MEIJER TROPIS VI
ABSTRACT
In the miniature frieze of the wall paintings in the West House at Thera
which belong to the mid-16'h or early 15'h century BC a number of large
paddled ships, a smaller oared ship and a number of boats are depicted. All
have been of unusual interest to nautical archaeologists but their precise
function has not been identified beyond a general agreement that they
appear to be celebrating some kind of cult. JSM's paper will attempt to
sharpen this interest by identifying the cult and thus the function which each
type of vessel is shown as performing.
A number of Greek myths concerning Dionysos, Theseus and the Cretan

Kouretes which have been recognised to reflect a primitive rite de passage
will first be described, and then other relevant ritual practices indicating the
acquisition of knowledge and the beginning of a new life.
Finally, with the help of slides the characteristics of a rite de passage

which will have been identified will be recognised in the various stages of the
cult illustrated in the miniature frieze; from the rivers of the wilderness at the
outset to the divers at the conclusion of the sequence of pictures. The
precise function of the various vessels should then become clear.
t John Morrison
Trireme Trust
Granhams
Great Shelford
Cambridge CB2 5JX, U.K.
ABSTRACT
THE USE OF CATAPULTS IN HELLENISTIC NAVAL WARFARE
It is generally believed that during the fourth and third centuries BC

Greek warships grew larger and larger to accommodate ever increasing
contingents of marines and batteries of catapults (cf. Casson, SSAW, 107).
The standard view (which goes back to W.W. Tarn) holds that sea battles
between these larger and larger units evolved essentially into contests
between the marines on the decks as grappling and boarding became the
tactics most favored by the Macedonians and Romans (Tarn, Hell. Milita~y&
Naval Developments, 144-45). Catapults played an important role, so the
argument goes, by "softening up" an enemy vessel before the ramming,
grappling and boarding actions were initiated to finish off-one's enemy (Tarn
HMND, 152; Casson, SSAW, 121-22). In 1981, V. Foley and W. Soedel
(Scientific American [April 19811 160-62) suggested that catapults were also
used against enemy oar-crews by firing projectiles that penetrated the decks
of their enemies. This view has appeared recently in John Coates'
contribution to the volume on ancient galleys edited by J.S. Morrison (The
Age of the Galley, p. 135).
In this paper, I will demonstrate that Foley and Soedel's deck-penetrating

catapults are not supported by ancient evidence that explains how naval
catapults were used, and what their limitations were (a handout will
accompany this section of the paper). I will then briefly set forth a new
theory, which explains the naval arms race of the late fourth and third
centuries as an evolution of Alexander's use of the navy to project power. His
particular brand of naval power was adopted by Antigonus, exploited
vigorously by Demetrius and then adopted in turn by Lysimachus, Gonatas,
Ptolemy I and II. The important role of siege craft combined with the
relentless drive of Antigonus and Demetrius to reconstitute Alexander's
empire led directly to the astounding development of larger and larger ship
classes. The phenomenon is purely Hellenistic, firmly rooted in the political
events of the late 41hand early 3'dcenturies BC and heavily influenced by the
reliance on naval siege craft demonstrated by the rulers who succeeded
Alexander the Great.
Prof. William M. Murray

6613 Baybrooks
Temple Terrace, FL 33617
USA
LES EPAVES GRECQUES ARCHAIQUES DU VIe SIECLE AV. J.-C. DE
MARSEILLE :
épaves Jules-Verne 7 et 9 et César 1.
La fouille de la place Jules-Verne à Marseille, menée d'août 1992 à la fin

octobre 1993 à proximité immédiate de l'actuel Vieux Port, a mis en évidence
une partie du rivage antique avec ses aménagements portuaires'. Elle a
aussi livré deux épaves grecques archaiques et plusieurs épaves romaines2.
Les deux épaves grecques, Jules-Verne 7 et 9, ont été retrouvées gisant

l'une contre l'autre dans les sédiments qui ont envasé le port à la fin du VIe
siècle av. J.-C. (Fig. 1 et 2). Manifestement abandonnés à la même époque
au cours des vingt-cinq dernières années du siècle, les deux bateaux ont
donc vraisemblablement été construits et ont navigué dans la seconde
moitié du VIe S., soit moins d'un siècle après la fondation de la cité
phocéenne.
Enfin, tout dernièrement, en 1997, une troisième épave grecque

archaïque a été découverte sur la fouille du chantier du projet de << musée
César >>. Située place Villeneuve-Bargemon mitoyenne de la place Jules-
Verne, cette fouille a mis en évidence la suite des aménagements du port
antique de Marseille3.L'épave, dite César 1 , semble avoir été abandonnée à
l'extrême fin du VIesiècle av. J.-C.
Epave Jules-Verne 9
Cette épave est celle d'une grande barque de pêche qui est conservée
sur 5 m de longueur et 1,40 m de largeur. Les vestiges, homogènes,
correspondent à une extrémité et à une partie du centre de I'embarcation
(Fig. 3). Des petits fragments de corail rouge, retrouvés emprisonnés dans
la résine d'étanchéité interne de la coque, indiquent que I'embarcation a
servi, notamment, à la pêche au corail. Les formes de la carène, à section
arrondie au maître couple et aux extrémités élancées, et ses dimensions,
que l'on peut restituer à environ 9 m de longueur sur 1,60 m de largeur, sont
celles d'une embarcation côtière, légère et rapide, très vraisemblablement
propulsée à rames ce qui n'exclue pas la présence d'une petite voile
auxiliaire.
La structure de I'embarcation est légère. La quille, de section presque

carrée (7 x 6,8 cm), est dépourvue de râblure. Elle se prolonge, peu avant
PATRICE POMEY TROPIS VI
l'extrémité conservée, par une allonge qui lui est assemblée par un écart en
trait de Jupiter >? à clef verticale et qui amorce la remontée des fonds avant
de faire place à l'étrave ou à l'étambot. Un fragment provenant de la pièce
d'extrémité montre la présence de râblures qui n'existent, de ce fait, qu'au
niveau de l'étrave et de l'étambot. Les planches du bordé (ép. 2,7 à 3 cm ;
larg. 15 à 20 cm) sont assemblées à franc bord et des joints en biseau
unissent les bordages d'une même virure. La membrure, dont seuls
subsistent les empreintes, une varangue en place et un couple de revers
isolé, était composée de varangues de fond largement espacées (maille de
0,90 m) vraisemblablement alternées avec des couples de revers situés
uniquement dans la partie haute de la muraille. Des allonges étaient
assemblées aux extrémités des varangues par un écart à croc chevillé.
Enfin, le départ d'une petite épontille au centre de l'unique varangue en
place indique la présence de baux transversaux servant sans doute de banc
de nage dans la partie centrale et de support de petits ponts de couverture
aux extrémités.
Mais l'intérêt essentiel de cette épave réside dans son mode de

construction faisant appel à la technique archaïque d'assemblage par
ligatures. Nous sommes en effet en présence d'un bateau entièrement
<< cousu dont tous les éléments essentiels de la structure sont liés entre eux
>)
au moyen de ligatures en lin (Fig. 4 et 5). Ainsi, les planches du bordé étaient
assemblées entre elles et à la quille au moyen de liens passant à travers des
canaux obliques (diam. 0,6 cm) prenant naissance à partir d'évidements
tétraédriques (1,5 à 1,7 cm de côté) régulièrement ménagés (tous les 2,5
cm) le long du bord d'assemblage de chaque élément. De petites chevilles
enfoncées dans les canaux obliques à partir des évidements tétraédriques
bloquaient les ligatures en place. Des chevilles horizontales (diam. 1 cm,
écart 20,5 cm), disposées au préalable dans les plans de contact, avaient
pour objet de maintenir les planches en place, tout d'abord lors de leur
assemblage et, par la suite, pour éviter le cisaillement des ligatures.
L'étanchéité était assurée par une bande de tissus disposée au-dessus de
chaque joint avant le ligaturage et par une épaisse couche de résine
appliquée sur toute la face interne de la carène. C'est grâce à cette résine
que de nombreuses ligatures ont pu être exceptionnellement conservées en
place. Les membrures étaient de même ligaturées à la coque selon le même
principe mais au moyen de canaux obliques situés transversalement au
centre de chaque virure. Les membrures, à dos arrondi et au pied étroit
régulièrement entaillé d'évidements, ont une morphologie très particulière
qui se justifie par la technique d'assemblage utilisée: le dos arrondi et le pied
étroit permettant un meilleur serrage; les évidements évitant l'écrasement
LES EPAVES GRECQUES ARCHAIQUES
DU Vle SIECLE AV. J.-C. DE MARSEILLE
des liens du bordé. L'ensemble témoigne d'une grande régularité et d'une

extrême minutie obtenues grâce à de nombreux tracés préliminaires
effectués à la pointe sèche par les charpentiers. De même, des marques
incisées sur le dos de la quille à l'emplacement de chaque varangue devait
guider les constructeurs lors de la construction du bateau réalisée selon les
principes et les méthodes de la construction bordé premier >>'.
CC
Au total, cette épave illustre d'une façon remarquable la technique

archaïque d'assemblage par ligatures qui est signalée par les textes5,
notamment grecs, et qui n'était mise en évidence jusqu'à présent que par
les vestiges fragmentaires des épaves du VIesiècle av. J.-C. de Giglio et de
Bon-Porté dont elle vient confirmer et préciser le système d'assemblage6.Le
caractère local de l'embarcation permet d'être assuré de son origine
massaliète et l'on peut donc rattacher son système de construction à une
tradition grecque, et plus précisément phocéenne, directement héritée des
fondateurs de la cité. A ce titre, l'épave Jules-Verne 9 constitue un
témoignage exemplaire de cette tradition grecque de bateaux cousus de
l'époque archaïque à laquelle appartiennent aussi les épaves de Giglio et de
Bon-Porté, elles-mêmes d'origine grecque voire massaliète7.
Epave Jules-Verne 7
Cette épave est en revanche celle d'un navire de plus grandes dimensions.
Elle est conservée sur 14 m de longueur et près de 4 m de largeur (Fig. 6).
Cependant, malgré de nombreuses ruptures, la coque est assez complète
dans la mesure ou les divers éléments peuvent être remis en place. Ainsi, la
quille est entièrement conservée, et les formes des extrémités, en dépit de
la disparition de l'étrave et de l'étambot, sont en partie connues par les têtes
des virures. L'une des murailles offre même trois niveaux de préceintes dont
la dernière se situe vraisemblablement à proximité de la ligne de plat bord.
La plupart des membrures peuvent être remises en place et certaines
d'entre elles sont entièrement conservées jusqu'à leur extrémité supérieure.
Enfin, il est possible de restituer le dispositif d'emplanture du mât par la trace
de certains éléments de fixation et quelques pièces isolées. L'épave
correspond à un petit, navire de commerce à voile d'environ 15 m de
longueur sur 3 m de largeur à la carène de section transversale arrondie et
aux extrémités pincées et fortement élancées.
La structure de la coque est très proche de celle de l'épave Jules-Verne 9.

La quille, conservée sur 10,70 m de longueur, est de même composée de
deux éléments assemblés par un trait de Jupiter ,, à clef verticale. De

section presque carrée (11 x 10 cm), elle est dépourvue de râblure à
I'exception toutefois des extrémités où les râblures apparaissent à environ
un mètre en avant des écarts de liaison, toujours du même type, avec
l'étrave et I'étambot. Les bordés, assemblés à franc bord, sont constitués de
virures d'une largeur et d'une découpe très variables (ép. 2,5 à 3 cm ; larg.
14 à 28 cm) qui introduit une grande asymétrie dans l'ensemble du plan de
bordé. Des joints en biseau, dans les fonds, et en ligne brisée ou curviligne,
au-delà, relient les bordages d'une même virure. Trois préceintes
polygonales ou semi-circulaire (9, 11: 13" virures), étroites (7 à 11 cm) et
épaisses (8 à 12 cm) en renforcent la structure. La membrure reprend le
même dispositif que celui de l'épave Jules-Verne 9. Aux extrémités des
allonges des varangues des encoches latérales avec chevillage indiquent la
présence de baux transversaux. Des mortaises au centre des varangues
correspondent à la présence d'épontilles dont un exemplaire complet a été
retrouvé.
Là encore, la technique de construction est remarquable à bien des égards.

Ainsi, I'assemblage de la quille et des bordés est réalisé d'une façon tout à
fait originale au moyen de deux techniques différentes employées
concurremment : par tenons et mortaises » et par ligatures végétales ,,.
La technique par tenons et mortaises est utilisée très majoritairement pour la
plus grande partie du montage du bordé. Les tenons longs et étroits (14 x 3
cm), chevillés de l'intérieur (diam. int. 1,1 à 1,4 ; diam. ext. 0,9 à 1,2), et les
mortaises largement espacées (20 cm) dénotent un système précoce dont
le réseau n'atteint pas encore la densité qu'il connaîtra par la suite (Fig. 7).
Les assemblages par ligatures, identiques à ceux de l'épave Jules-Verne 9,
sont en revanche utilisés d'une façon limitée : dès l'origine, aux deux
extrémités de la quille et pour la fermeture des virures de bordé sur l'étrave
et l'étambot ; par la suite, en divers endroits de la coque pour des
réparations intervenues après l'assemblage initial par tenons et mortaises.
La membrure présente, de même, un double aspect. Par son dispositif

général alternant membrures de fond à maille large (0,90 m) et couples de
revers dans les hauts, la morphologie particulière des varangues et de leurs
allonges (dos arrondie, pied étroit, base régulièrement entaillée
d'évidements) et leur liaison à croc chevillé, elle est caractéristique de la
membrure des bateaux cousus. Pourtant, elle n'est plus ici assemblée par
ligatures. Les varangues et leurs allonges sont en effet directement clouées
sur le bordé avec des clous rabattus et repénétrant dans le dos des pièces,
alors que les couples de revers sont chevillés à I'exception de leur pied, et
de leur pied seulement, qui sont encore ligaturés.
Parmi les autres caractéristiques remarquables de cette épave, il convient de

noter la présence de deux séries de marques de charpentier en relation avec
les phases de construction du navire. La première, incisée avec le fer d'un
outil tranchant du type ciseau à bois sur le dos de la quille, marque
l'emplacement du pied de chaque varangue comme sur l'épave Jules-Verne
9. La seconde, en forme de pointe de flèche tracée à la pointe sèche sur la
huitième virure, indique l'emplacement de l'extrémité des allonges de
varangues. Cette dernière série de marques indique que les membrures des
fonds n'ont été posées qu'après l'assemblage des huit premières virures et
avant la poursuite du montage du bordé. La huitième virure, plus large que
les autres, jouait alors un rôle particulier de virure de réglage ,,. C'est à son
niveau qu'étaient effectués les contrôles, et au besoin les rectifications de
forme et de symétrie de la carène8.
Les nombreuses similitudes de structure et de détail de construction,

notamment au niveau des marques de charpentier, indiquent que ce navire,
dont le rayon d'action n'interdit pas a priori l'hypothèse d'une origine
étrangère, fut construit lui aussi dans les chantiers navals massaliètes. Aussi,
par ses caractéristiques et son double système d'assemblage, nous
sommes manifestement ici en présence d'un navire de transition qui reflète
d'une façon exceptionnelle l'évolution des techniques de construction et le
passage de la technique d'assemblage par ligatures à la technique par
tenons et mortaises dans le contexte précis de la Marseille grecque de la
seconde moitié du VIeS. av. J.-C.
Epave César 1
Réduite aux vestiges d'un fond de carène conservé sur 6,10 m de

longueur totale et 0,90m de largeur, mais séparé en deux parties par une
petite tranchée, cette épave comporte la quille et ses deux extrémités
s'achevant par des écarts en cc trait de Jupiter ,, à clef verticale, un fragment
de la pièce d'étrave ou d'étambot en place dans l'un des écarts
d'assemblage, des éléments de cinq virures, dont un galbord offrant une
extrémité complète, et deux maigres fragments de membrures (Fig. 8). Mais
elle présente le grand intérêt d'être semblable à l'épave Jules-Verne 9 par sa
structure, ses formes et ses dimensions et d'être identique à l'épave Jules-
Verne 7 par sa technique de construction. En effet, si le bordé est assemblé
essentiellement par des tenons chevillés dans des mortaises selon un
PATRICE POMEY TROPlS VI
échantillonnage correspondant à celui de Jules-Verne 7, des ligatures sont

toujours utilisées aux deux extrémités de la quille pour la fermeture des têtes
de virures sur l'étrave et l'étambot et pour des réparations. De même, la
membrure des fonds est clouée au bordé.
Ainsi, vers la fin du VIe S. av. J.-C., cette épave témoigne à nouveau de
l'évolution des techniques de construction navale à Marseille mais cette fois
avec l'introduction de la technique d'assemblage par tenons et mortaises
sur une embarcation plus modeste que le caboteur de I'épave Jules-Verne 7
et de même type que la barque côtière de I'épave Jules-Verne 9.
Au-delà de leurs caractéristiques individuelles, l'intérêt de ces trois

épaves est de présenter, à la même époque, la deuxième moitié du VIeS. av.
J.-C., et dans le même contexte, Marseille grecque, deux stades de
l'évolution des techniques de la construction navale grecque. Le premier,
représenté par I'épave Jules-Verne 9, marque l'aboutissement de la
technique archaïque d'assemblage par ligatures. Le second, illustré par les
épaves Jules-Verne 7 et César 1, témoigne de l'adoption de la technique de
construction par tenons et mortaises qui s'imposera par la suite comme la
technique dominante de la construction navale gréco-romaine. Mais cette
phase d'adoption apparaît, du moins à Marseille au cours de la seconde
moitié du VIe S., à ses débuts. Le système d'assemblage par tenons et
mortaises ne semble pas encore totalement maîtrisé ainsi qu'en témoigne
l'usage des ligatures pour la fermeture des extrémités et les réparations et il
n'a pas encore la densité qu'il aura par la suite. De même, la structure
d'ensemble et surtout la morphologie des pièces découlent directement de
la tradition d'assemblage par ligatures alors que dans le cas des membrures
leur morphologie particulière n'a plus de raison d'être. En fait, la
contemporanéité des trois épaves et leurs similitudes montrent que l'on se
limite, dans un premier temps, à appliquer une nouvelle technique
d'assemblage au sein d'un mode de construction traditionnel et encore
vivace sans que cela ait de conséquence sur l'évolution de la structure et de
la morphologie des navires.
Par rapport aux ligatures, la technique par tenons et mortaises présente

un double avantage : d'une part, elle assure une plus grande longévité aux
assemblages qui n'ont plus besoin d'être refaits régulièrement ; d'autre part,
elle confère à la coque une plus grande solidité qui ouvre des possibilités
plus importantes de développement des formes, des dimensions et du
tonnage. Si I'adoption de ce nouveau système d'assemblage sur le navire
de I'épave Jules-Verne 7, le plus grand, paraît répondre aux avantages qu'on
lui suppose, son utilisation sur le bateau de I'épave César 1, très proche de
l'embarcation de pêche de I'épave Jules-Verne 9, montre que la diffusion de
la nouvelle technique fut rapidement étendue à toutes les unités et ne se
limitait pas seulement aux plus importantesg.Cependant, il faudra plusieurs
siècles pour que la nouvelle technique soit totalement assimilée et que ses
conséquences soient sensibles sur l'évolution des formes et des structures.
Au tout début du Ve siècle av. J.-C., 9 navire grec de Géla paraît encore très
proche de I'épave Jules-Verne fi. A la fin du V", soit un siècle après
l'adoption de la nouvelle technique, I'épave Ma'agan Mikhael (Israël), qui
relève à l'évidence de la même tradition grecque, présente déjà un caractère
beaucoup plus évolué qui se traduit, notamment, par un usage encore plus
réduit des ligatures, un réseau de tenons et mortaises plus dense, une
système par ligatures est déjà atténué &

carène à retour de galbord et une morph sgie des pièces où l'influence du
Encore un siècle, et à la fin du
IV" S. av. J.-C. le navire de Kyrénia représew le stade finale de l'évolution
de cette longue tradition dont il est issul~L'assemblagepar tenons et
mortaises apparaît totalement maîtrisé y compris dans les réparations, la
carène à retour de galbord comporte désormais une quille entièrement
râblurée, les couples de revers sont devenus des demi-couples alternés
avec les varangues et toute réminiscence des assemblages par ligatures sur
la morphologie des pièces a disparue. Dès lors, l'évolution est telle que le
navire de Kyrénia peut être considéré autant comme le stade ultime de la
longue évolution de la tradition grecque des bateaux cousus que comme le
point de départ de la nouvelle tradition de la construction navale gréco-
romaine.
Patrice POMEY
Directeur de recherche au CNRS
Centre Camille Jullian, CNRS-Université de Provence
France
NOTES
1. Les fouilles, conduites par le Service Régional de l'Archéologie avec le concours de la Ville
de Marseille, ont été dirigées par Mme A. Hesnard, directeur de recherche au CNRS (Centre
Camille Jullian, Aix-en-Provence). Cf. Hesnard 1994.
2. Pomey 1995. En outre, les épaves romaines ont fait l'objet d'une présentation au 5*
International Symposium on Ship Construction in Antiquiiy qui s'est tenu à Nauplie en 1993
(Porney 1999) et les épaves grecques au 7" International Symposium on Boat and Ship
Archaeology de l'île Tatihou en 1994 (Pomey 1998b).
3.Hesnard 1998
4. Pomey 1988, 1998a.

5. Pomey 1985, Bonino 1985.
6. Pomey 1981, Bound 1991.
7. Pomey 1997, p. 199.
8. Pomey 1998a.
9. 11 serait évidemment très important de pouvoir établir avec précision la chronologie relative
de la date de construction de ces trois bateaux. Dans l'état actuel de l'étude archéologique
des sites, il est évident que les épaves Jules-Verne 7 et 9 , situées dans le même niveau
stratigraphique, ont été abandonnées au même moment. L'épave César 1 semble, en
revanche, avoir été abandonnée peu après. Mais les dates d'abandon ne permettent pas de
préciser les dates de construction ni de préjuger, pour les épaves Jules-Verne 7 et César 1,
de leur chronologie relative.
10. Freschi 1991.
11. Linder, Rosloff 1995 ; Kahanov 1996, 1998. Sur l'appartenance de le cette épave à la
tradition grecque d'assemblage par ligatures et sur sa place dans son évolution, cf. : Pomey
1997.
12. Steffy 1985, 1994. Sur la place de cette épave dans l'évolution de la tradition grecque
d'assemblage par ligatures, cf. : Pomey 1997.
BIBLIOGRAPHIE
BONINOM., 1985, Sewn boats in Italy: sutiles naves and barche cucite, dans S. McGrail, E.
Kentley (ed.), Sewn Plank Boats (B. A. R. Int. Series 276), Oxford, p. 87-104.
BOUNDM., 1991, The Giglio wreck (Enalia, Sup. 1), Athènes.
FRESCHIA., 1991, Note techniche su1 relitto greco arcaico di Gela, Atti IV Rassegna di
Archeologia Subacquea, Giardini Naxos 13-15 ottobre 1989, Giardini Naxos, p. 201-
210.
HESNARD A.,1994, Une nouvelle fouille du port de Marseille, place Jules-Verne, Comptes Rendus
Académie des Inscriptions et Belles-Lettres, janvier-mars, p. 195-217.
HESNARD A., 1998, Marseille, Place Villeneuve-Bargemon (<<Musée César 2 m), Bilan Scientifique
de la Région Provence-Alpes-Côted'Azur 1997, Service Régional de l'Archéologie, Aix-
en-Provence, p. 78-84.
KAHANOVJ., 1996, Conflicting evidence for defining the origin of the Ma'agan Mikhael
shipwreck, Tropis IV, 4mlnternational Symposium on Ship Construction in Antiquity,
Athens 1991, Athènes, p. 245-248.
KAHANOVY., 1998, The Ma'agan Mikhael ship (Israël). A comparative study of its hull
construction, dans P. Pomey, E. Rieth (dir.), Constructicn navale maritime et fluviale.
Approches archéologique, historique et ethnologique, 7' Colloque international
d'archéologie navale - 7" 1S B S A., île latihou 1994, (Archaeonautica 14, 1998), p.
155-160.
LINDER E., ROSLOFFJ., 1995, The Ma'agan Mikhael shipwreck, Tropis 111, 3"' lnternational
Symposium on Ship Construction in Antiquity, Athens 1989, Athènes, p. 275-281.
POMEY P., 1981, L'épave de Bon-Porté et les bateaux cousus de Méditerranée, Mariner's Mirror,
67, 3,p. 225-243.
POMEY P., 1985, Mediterranean sewn boats in Antiquity, dans S. McGrail, E. Kentley (ed.), Sewn
Plank Boats (B. A. R. Int. Series 276), Oxford, p. 35-47.
POMEY P., 1988, Principes et méthodes de construction en architecture navale antique, Navires
et commerces de la Méditerrannée antique. Hommages à Jean Rougé, (Cahiers
DU Vle SlECLE AV. J.-C. DE MARSEILLE
d'histoire, XXXIII, no3-4, p. 397-412.

POMEYP., 1995, Les Epaves grecques et romaines de la place Jules-Verne à Marseille, Comptes
Rendus Académie des Inscriptions et Belles-Lettres, avril-juin, p. 459-484.
POMEYP., 1997, Un exemple d'évolution des techniques de construction navale antique : de
l'assemblage par'ligatures à l'assemblage par tenons et mortaises, dans ~echni~ues
et économie antigues et médiévales :le temps de l'innovation, Collogue international,
Aix-en-Provence, mai 1996, Paris, p. 195-203.
POMEYP., 1998a, Conception et réalisation des navires dans l'Antiquité méditerranéenne, dans
E. Rieth (dir.), Concevoir et construire les navires. De la trière au picoteux,
(Technologies, Idéologies, Pratiques, Revue d'anthropologie des connaissances, XIII,
l), p. 49-72.
POMEYP., 1998b, Les Epaves grecques du VI" siècle av. J.-C. de la place Jules-Verne à
Marseille, dans P. Pomey, E. Rieth (dir.), Construction navale maritime et fluviale.
Approches arch6ologique, historique et ethnologique, 7" Colloque international
d'archbologie navale - 7h 1 S B S A., île Tatihou 1994 (Archaeonautica 14, 1998), p.
147-154.
POMEYP., 1999, Les Epaves romaines de la place Jules-Verne à Marseille : des bateaux
dragues?, Tropis V, 5mInternational Symposium on Ship Construction in Antiquiiy,
Nauplia 1993, Amènes, p. 321-328.
STEFFY
J. R., 1985, The Kyrenia ship. An interim report on its hull construction, American Journal
of Archaeology, 89, 1, p. 71-101.
STEFFY
J. R., 1994, Wooden ship buildoing and the interpretation of shipwrecks, Texas A & M
University Press, College Station.
ILLUSTRATIONS
Fig. 1 Vue d'ensemble des deux épaves grecques de la place Jules-Verne en cours de fouille.
(Cliché CNRS-Centre Camille Jullian).
Fig. 2 Plan d'ensemble des deux épaves grecques de la place Jules-Verne. (Relevé et dessin
M. Rival, CNRS-Centre Camille Jullian).
Fig. 3 Vue de I'épave Jules-Verne 9 en cours de fouille. (Cliché CNRS-Centre Camille Jullian).
Fig. 4Schéma du système d'assemblage par ligatures de I'épave Jules-Verne 9. (Dessin M.
Rival, CNRS-Centre Camille Jullian).
Fig. 5Vue axonométrique partielle du système d'assemblage par ligatures de I'épave Jules-
Verne 9. (Dessin M. Rival, CNRS-Centre Camille Jullian).
Fig. 6Vue d'ensemble, au premier plan, de I'épave Jules-Verne 7. (Cliché CNRS-Centre Camille
Jullian).
Fig. 7 Schéma axonométrique du système d'assemblage par tenons et mortaises de I'épave
Jules-Verne 7 (Dessin M. Rival, CNRS-Centre Camille Jullian).
Fig. 8 Vue de I'épave César 1 en cours de fouille. (Cliché J. Castay).
PATRICE POMEY
-- - --- - - -- -- - - -- - - --- -
'TRQPIS VI
-- -
Fig. 3
-. -*.L
c-
- y , -y-
- .. - -
*
n
:%-&--
d -*
* -*
%
-y.
..
.-
1
'
I, - --
-P,
---
-i-\
Fig. 2
Uece J.Verne - Epmm w9

Détail & I'aessmblage dae viru-
et dw ie membrure
mm
Fig. 4
435
PATRICE POMEY
- - TROPIS VI
Fig. 5
'Ti-
LES EPAVES GRECQUES ARCHAIOUES
-- -- --- - --
DU Vle
- --
SIECLE
-
AV. J.-C. DE MARSEILLE
- . . - -- - -
Fig. 8
ABSTRACT
THE ULUBURUN SHIPWRECK-AN UPDATE
Since the first hull remains of the Uluburun shipwreck were exposed in the
summer of 1984, we had known that the ship's planking was assembled with
mortise-and-tenonjoinery similar to that found on later Greek and Roman ships,
making the use of this construction technique in the Uluburun hull the earliest
known in the history of seagoing ship construction.
After the completion of the Uluburun shipwreck excavation in 1994 all hull
wood, including the sections of poorly preserved fragmentary planking found
during the last campaign, were taken to the Bodrum Museum of Underwater
~rchaeologyfor storage and conservation. Recent preliminary examination and
study of some hull pieces revealed several unexpected explanations for what
had been previously observed. We had realized from the beginning that the
Uluburun ship's joinery was more robust and more widely spaced than that
found in Greek and Roman ships of similar size, i.e., 15-18 meters in length.
Unlike most Graeco-Roman mortise-and-tenon joints, however, those in the
Uluburun hull were found to be extraordinarily deep and extending to within a
few centimeters of the opposite plank edge. Moreover, all examined planking
pieces revealed that each joint cut in one plank edge is positioned immediately
next to the nearest joint cut from the opposite plank edge. Consequently,
mortises often intrude on one another. Such a practice, which required removal
of a large volume of wood over nearly the entire width of the plank, would seem
to have compromised the structural integrity of the planks and thus the hull. Yet,
it was observed steadfastly that such pairs of tenons extended up the sides of
the hull planking every 24-26 centimeters, center to center. While this may
simply have been a convenient way of maintaining consistent joint spacing,
more likely it represents a specific, conscientiously executed structural practice.
The latter view gains additional support in light of our failure, after repeated
examinations of the extant hull sections, to reveal any evidence for the
employment of frames in the building of the Uluburun hull. These tenon <<belts,,>
therefore, may have functioned as internal ((stiffeners,,, or (drarnes.,,
Cemal Pulak
Institute of Nautical Archaeology
at Texas A&M University
College Station
Texas
THE ENIGMA OF THE ANGULAR SAILING SHIPS
IN THE RED SEA SINCE THE 4"' MILLENNIUM BCE
The beginnings of navigation and sea-borne connections, between the

various parts of the ancient Near East and farther away, may have taken
place as early as the initial phase of human appearance on the islands of the
Mediterranean, or even earlier. Yet, the earliest surviving documents as for
the types of vessels used for these early water voyages are only the
iconographic depictions - clay models, rock-drawingsand painted vases of
Late Neolithic and Early Chalcolithic eras in Egypt and Mesopotamia. Some
of these depictions are either without datable context or are too crude and
simplified to be of any real value when trying to study the technological
aspects of the actual nautical vessels these ancient artists attempted to
illustrate. Others, more detailed and better executed, still might be
controversial as to the proportions, size, means of propulsion, the raw
materials used and the technology of their construction. For that reason, the
scholars studying ancient shipping are cautious, tentative and
argumentative in their use of these iconographic documents as evidence for
the origins of the earliest sails (Bowen, 1960; Casson, 1991: 4) and the
identification of the materials of which the illustrated types might have been
made (Casson, 1971: 5-40; Hornell, 1946: 181-193; Vinson, 1994: 11-15).
Yet, every student of shipbuilding technology would accept the logical
correlation between the availability of certain suitable raw materials for
construction of floating navigable boats and the technology used for shaping
the final product.
The vast repertory of boats depicted in clay models, rock drawings and
paintings from Pre-dynastic Egypt may be grouped in three main categories,
based on the shape of the hull:
1.Reed, or papyrus-made hull, characterized by up-curving ends, narrow
to a point. This type is presented by clay models dated as early as the 5*
millennium BCE from the Badarian culture (Vinson, 1994, Fig. 2) and
continue all through the Amratian and the Gerzean periods (Kantor,
1944, Fig. 5).
2.A long, crescentic hull, depicted as having sides of even breadth to their
entire length with angular cut ends. A type of hull which might be
dictated by either a dug-out, monoxyle trunk (Basch, 1987: 55-56) or
long timber planks (Vinson, 1994: 12). This type is the most
characteristic one in the Upper Egypt culture of the Naqada II period
AVNER RABAN TROPlS VI
(Petrie, 1921; 1933; Kantor, 1944: 115; Landstrom, 1970: 12; Bass, 1972:
12-13; Basch, 1987: 57-60; Vinson, 1994: 12-15), though few datable
depictions of the type are even earlier, of the Amratian era, of the first half
of the 4Ih millennium BCE (Bass, 1972: 13, Fig. 2; Casson, 1971, Fig. 3).
3.A rather similar type, as for its raw material (wood), but with either only
the prow or both ends terminated with solid vertical post of significant
size (Kantor, 1944, Fig. 4; Engelmayer, 1965, PI. XII, 4; Williams, 1980:
16). This "square", or angular type was still rather common among
depictions of boats dated to the eve of the first dynasty period, mostly in
the eastern desert of Upper Egypt and on rock drawings from Nubia
(Arkell, 1950, Fig. 1; Emery, 1961, Figs. 4, 10, 12). This type was
designated as "foreign" and "non-Egyptian" by most scholars (Kantor,
1944: 129; 1965: 10; Frankfurt, 1951: 110-11; Bass, 1972: 13; Vinson,
1994: 16-20).
The argument that this foreign type represents a Mesopotamian vessel

of the Protoliterate period was suggested already by Frankfurt (1951),
supported by Kantor (1952), strongly opposed by Helck (1962: 6-9) and
questioned ever since. Recent finds in the Delta have somewhat "made
passe" the earlier scholars' claim for 4th-millenniumdirect sea routes and
sea-borne connections between the Gulf and Egypt, through the Indian
Ocean and the Red Sea (Kantor, 1965: 10-14), in favour of the more
"conventional" land route along the Fertile Crescent (Moorey, 1990).
Though it is most probable that both overland and sea routes along the
Levantine coast had been used prior to the unification of dynastic Egypt, at
least since the mid-4Ih millennium BCE, or even earlier (Prag, 1986;
Andelkovic, 1995), this may not necessarily contradict the well-established
data from Upper Egypt, Wadi Hammamat and Nubia, indicating
Mesopotamian importation of artifacts and its direct technical and cultural
influences on a society which was, in that era, much more complex and
advanced than that of Lower Egypt (Baumgartel, 1960: 139; Kantor, 1965:
12; Bard, 1994: 1 11-118). Recent socio-anthropological studies would
suggest that the shift of developmental focus from Naqada and
Hierakonpolis in Upper Egypt to Buto in the north had occurred only during
the later Pre-dynastic period, partly because of the growing importance of
that Syro-Palestinian trade route (Wenke, 1989: 142; Andelkovic, 1995: 72).
The fact is that all iconographic depictions of the so-called "foreign

boats", which are dated to the Pre-dynastic era, were found exclusively in the
south, in the Eastern Desert and along the ancient route from El-Quseir on
IN THE RED SEA SINCE THE 4'hMILLENNIUM BCE
the Red Sea to the Nile Valley, near Naqada, through Wadi Hammamat.
What is the significance of these boats?
Among the illustrations published so far, there are three varieties (see
Figs. 1-3):
There is the sub-type depicted on the famous ivory knife handle from
Gebel el Arak (Emery, 1961: 38, Fig. 1) which is similar in shape and
decorations to the ceremonial, divine boat, depicted on a cylinder seal from
Uruk in Mesopotamia. Gebel el Arak is situated at the eastern edge of the
Nile Valley, at the western end of Wadi Hammamat. A somewhat similar boat
type is inscribed on the side of a Predynastic clay vessel found at a nearby
site (Kantor, 1944: Fig. 4, E). Whether this variant represents a real
ceremonial boat made of reed bundles, which was used in southern
Mesopotamia during its Protoliterate period (Frankfurt, 1924: 138-142; Arkell,
1959), or a conceptual symbol of foreign and rival culturai unit, it is hard to
say for sure (Basch, 1987: 60-62).
The second sub-type is actually a hybrid version of the high, vertical

ended boats and the crescentic vessel (Raban, 1996). Such is the rather
long boat from the wall painting at tomb 100 in Hierokonpolis (Quibell &
Green, 1902: PI. UO(V). In its context, this boat, painted black, is shown amid
five white ones, of the typical Upper Egypt Gerzean crescent, or
"boomerang" type. The black vessel carries the same attributes as the other
five, such as the tree branch at the fore end and the down-dropped bundle
under its prow. It also carries the standard double shrine-like cabins, though
the black one differs in shape, being arched instead of having a flat roof. The
same high prow type of boat is to be found depicted on rock drawings at
Wadi Hammamat, on the way to the Red Sea (Kantor, 1944: 138, Fig. 3, J.K)
and at the Sayala region of the eastern desert of Sudan (Engelmayer, 1965:
PI. X11.4). This last drawing depicts a boat without cabins, furnished with
eight oars on each side and helmsman, operating the steering oar, sitting at
the low-lying stern (Fig. 4).
Because it is difficult to make any reasoning of the prominent high and

rather heavy prow of that subtype, its resemblance to the hull shape of the
Early Cycladic boats of the Aegean Sea, which are dated to the following
millennium, is most intriguing (see below).
The third sub-type is relatively close in its general hull shape to the first
one. Yet it is depicted always without oars, frequently furnished with a square
sail and the triangular shape of its vertical sternpost, with the even width of
AVNERRABAN TROPIS VI
its hull and prow, indicating wooden construction, rather than reed bundles.
Though some iconographic documents of that sub-type are of uncertain
date, others are considered to be either of the late Gerzean, or early First
Dynasty period. The most famous one is the sailing boat painted on the late
Gerzean vase now in the British Museum (No. 35324, A) and two others -
from the eastern desert in North Sudan (Basch, 1987: 50, Figs. 79, 80, 81).
Another, recently published, was carved on a stone-made censer found at
Qustul, in southern-most Egypt (Williams, 1980: 16). That boat has a cabin
with a forward sloping roof, similar to that which is depicted on the vase in
the British Museum. On it a human figure is illustrated sitting with his hands
pulled back behind his back (Fig. 5). Another man is standing behind him,
at the stern, as if holding him in captivity, much like the petroglyph scene
from Sudan, dated to the early First Dynasty time of King Djer (Emery, 1961:
60, Fig. 22). There are theories among prominent scholars, that these
angular vessels belonged to the invading "Dynastic Race" that came by sea,
probably from Mesopotamia, either through Syria and the Nile Delta (Emery,
1961: 38-40), around the Arabian peninsula to El-Quseir (Derry, 1956), or
both to Mesopotamia and Egypt, from some unknown common provenance
in the Indian Ocean (Rice, 1990: 35-44). This last presumption, which
attributes common cultural and ethnic origin to the Pharaonic Race and the
Sumerians cannot be attested by any linguistic resemblance. The alleged
interpretation of the scenes depicted in the painted tomb at Hierakonpolis,
the carved tusk handle of the flint knife form Gebel el Arak and the Nubian
petroglyphs, as historical illustration of such invasion (Emery, 1961: 38), is
too farfetched. It is quite clear that in both scenes, from Gebel el Arak and
from Hierakonpolis, the winning side is the local, Gerzean one. The
"Menacing black ships" (Rice, 1990: 74, PI. 24) are more likely non-local
ships of an alien naval (?) power of which the people of the upper Nile Valley
had to be aware. These 4'h-millennium marines might have crossed the
eastern desert on their way from the coast of the Red Sea to the Nile Valley,
either through Wadi Hammamat, or farther south, but not necessarily as
aggressive invaders (Rice, 1990: 45-47). It is more likely that their aim was
trade. Probably seeking gold and bringing in their own goods, of which
some were the fine products and technical innovations of southern
Mesopotamia (Kantor, 1965: 10-16).
As stated above, the three variants of angular boats are fundamentally

different in their function and construction. The "ceremonial" type is less
angular and its ends are turned up and backward, narrowing to their floral
decorated points as if they had been made of papyrus, or bundles of reeds.
This variant is the only one that matches the Protoliterate period boats
IN THE RED SEA SINCE THE 4mMILLENNIUM BCE
carved on cylinder seals from Mesopotamia and Elam (cf. Rice, 1990: 71, PI.
12-13; Collon, 1987: 158, Nos. 712-714). The other two variants, which were
most probably made of wood and carried a functional square sail, might be
considered as the only sailing iconographic document of non Egyptian
marine sailing crafts of the 4'h millennium BCE (Vinson, 1994: 16).
As we have seen, the third angular variant differs radically from the so-
called Mesopotamian "Divine Boats", had no prototype in earlier depictions
from the Nile Valley and is rather rare among boat types of dynastic Egypt
up to the time of the New Kingdom. The few that are characterized by vertical
stems and stern posts were heavy cargo carriers on the Nile, such as the
long, plank-built, heavy duty boats depicted at the Valley temple of Unas, the
last Pharaoh of the 5thdynasty, carrying granite columns from the quarries of
Elephantine; or the sarcophagus carrier illustrated at the tomb of Chief
Justice Senezerrib, which is shown with stitched gunwale - a boat that
according to the accompaning text belonged also to King Unas (Landstrom,
1970: 62, Figs. 185, 186). A single wooden model of that type of boat
belongs to the early days of the 6'h dynasty and is exceptional among 15
other models found at the same context (Poujade, 1948: 40). The best
known depiction of vertical posts hull is of the seagoing ships, manned by
Syrian merchants and crew, which decorate the mortuary temple of Sahure,
the Pharaoh of the 5'hdynasty (Borchardt, 1913: 127-134). Much has been
written on these boats, their technical qualities (Landstrom, 1970: 63-69;
Casson, 1971: 20) and historical context (Viston, 1994: 23). Yet it is
interesting that the surviving text which is next to the scene of the "Syrian"
fleet, or an Egyptian one, returning from the Levantine coast of the
Mediterranean, tells us of ships that were sent to Punt, the Ophir of the
Pharaohs, in East Africa, at the 13"' regal year of king Sahure, bringing back
vast quantities of myrrh, electrum and ebony wood (BAR, 1.161). Strangely
enough these ships were called "Byblos" (KBNT) ships (Faulkner, 1940).
These ships are of clear-cut Egyptian technical heritage, with their keel-less
flat bottom, the "hugging truss", or "overhead" queen note which replaced
that missing keel, the high bipode mast and the stitched gunwales. Yet the
crew is not Egyptian. The leading merchants are "Canaanites", the type
called "Byblian" and the ship sailed also to Punt. Less Egyptian and less
ambiguous are the iconographic documents for ships with vertical posts,
which date to the New Kingdom era. The most famous one is the scene of a
"Canaanite" fleet of merchantmen reaching the quay at Thebes and
unloading their imported cargo, from the decorated wall of the tomb of
Kenamun, the superintendent of the granaries of Amun's temple during the
reign of Amenhotep 111 1407-1372 BCE. Another rather similar type of vessel
is depicted on the wall of the tomb of the chief physician of Amenhotep II

1450-1425 (Save-Soderbergh 1957: PI. XXIII). And the third, from the tomb
of Iniwia, probably of the 13'h century BCE, depicts "Canaanites" unloading
wine(?) jars from moored ships, of which only the forepart of three ships
have been found (unpublished, No. EM 11935 in Cairo Museum, and see
e.g. Landstrom, 1970: 138, Fig. 403). For some reason, Landstrom restored
that type as if it were of keel-less Egyptian type, though no "hugging truss"
is depicted on either one of the three documents (1970: 139, Fig. 407).
Others would consider these ships to be either true "Canaanite" (Basch,
1987: 62-66; Vinson, 1994: 40-44), or Canaanite type of merchantmen, which
were built at the Royal Egyptian shipyards at Pro-Nefer, by Canaanite
craftsmen (Save-Soderbergh, 1946: 39-60). Basch was the first to suggest
that the Egyptian name for this Canaanite type was MNS (menesh), a term
to be found in Egyptian texts since the time of Amenhotep Ill (Basch, 1978).
Later, these vertical posts, square type is to be found as representing the
"Sea Peoples" fleet at the famous depiction of Ramesses Ill defeating them
at sea, on the south wall of his temple at Medinet Habu (Nelson, 1943;
Raban, 1989: 165-167). This type continued to be characteristic for small
coastal and riverine log carriers of the Phoenicians, both in the Levant,
Cyprus and on the Euphrates, serving their Assyrian lords.
Such are the boat models found at Akhziv, Israel (Basch, 1987: Figs.
642-643), which are dated to the 9th-8'hcenturies BCE; the repertory of clay
models from Amathus and other Phoenician sites in Cyprus (Basch, 1987:
253-258, Figs. 543-557); and the Hippoi depicted on Assyrian reliefs (Basch,
1987: 305-20, Figs. 648-674). Just this type contrasted with the local New
Kingdom vessels in Egypt, so they differed from the crescent-shaped cargo
vessels of the Aegean and "Etheo-Cypriot" hulls of the first half of the last
millennium BCE, not to mention the war galleys and the longboats of the Iron
Age and the Archaic Period in the Mediterranean. How far west this type was
known and at least artistically depicted is hard to guess. So far, the statistical
analysis made by Basch (1987: 94-137) counted only two Early Minoan
seals, a painted pithos and the famous disk from Phaistos (Basch, 1987:
Figs. E l , E2, 273, 285) out of over 250 iconographic items. A similar
conclusion derives from Wedde's Ph.D. research (summarized: Wedde,
1995). In mainland Greece there is so far only one picture of that type, or
rather its derivation, painted on a LH Ill (12"-c. BCE) crater from Kynos
(Dakoronia, 1995: Fig. 2).
Summing up the iconographic data from the ancient Near East in a

combined spatial and chronological order, one would find that this special
IN THE RED SEA SINCE THE 4"' MILLENNIUM BCE
type of angular ship, with vertical prow and stern posts, is to be found in the
following order.
1.In the upper Nile Valley and the wadis of the Eastern Desert, on the way
to the Red Sea, since the mid-4'h millennium BCE, the Gerzean, or
Naqada II period, continuing into the Proto-dynastic and the Archaic
periods.
2. On various artifacts, in ceremonial and religious context, in both Egypt
of the First Dynasty (cf. Landstrom, 1970: 23-25) and Mesopotamia of the
Zd millennium BCE (Rice, 1990: 45-46).
3. Around the mid-3rdmillennium BCE, mainly in a sea-going voyage
context, both in the Mediterranean and the Indian Ocean, from the 5'h-6'h
dynasties of Egypt and from Early Minoan Crete. (From that period we
do not have a single iconographic document of boats or ships from the
Levant).
4. Syrian sea-going merchantmen, depicted in the tombs of high officials
of the Royal administration in Egypt of the New Kingdom (18'h-19'h
dynasties, 15'h-13'hcentury BCE).
5."Sea Peoples" coasters of the 12'hcentury BCE.
6. Phoenician and Phoenico-Cypriot boat types during the first half of the
last millennium BCE.
All scholars agree that this type was alien to Pre-dynastic Egypt, most
probably predates the Sumerians, and is unlikely to be of Mesopotamian
origin. Having been depicted first in the geographical sphere between the
Upper Nile and the Red Sea it would be best to search for its provenance in
the Mediterranean.
With all that in mind, the remaining potential origin of this type of sea-
going vessel should be searched for in the north-western corner of the
Indian Ocean; and more precisely along the south or the eastern coasts of
the Arabian peninsula. This understudied area has been opened for a full-
scale modern archaeological research only in recent years. Such recent
studies seem to verify some notions that were popular during the 1930s
(Oppenheim, 1954). There is some recently discovered data concerning the
societies of Bahrain, Qatar and Oman, which suggest that agricultural
communities involved in trade and seafaring had thrived there as early as the
5'h-4'hmillennia BCE (Potts, 1984; Zarins, 1992; Rice, 1994). The sea-borne
contacts of these people with the African continent and maybe even with the
Nile Valley might be attested among other facts by the introduction of the
sorghum crop plant to the Gulf (Qatar). In the same context, dated to the late
4"' millennium BCE, in which typical Mesopotamian pottery of Jamdat Nasr
AVNER RABAN TROPlS VI
style have been exposed (Potts 1994: 238-239).
It is tempting to attribute to these peoples of the southern and eastern

coasts of Arabia the role of seafarers who carried goods, cultural ideas and
technological innovations from the head of the Gulf (the "Sea Land" of
ancient Mesopotamia) to the Egyptian ports on the Red Sea; and probably
across the Eastern Desert to the Nile Valley, as Kantor suggested in 1956. It
is also quite probable that such alleged maritime endeavours were carried
out on board sailing ships of the angular type discussed above.
The later spatial distribution of that type is correlated quite intimately with
the maritime sphere of the West Semitic people of the Levantine coast of the
Mediterranean, known from the Bible as "Canaanites", and later, since the
Iron Age, by the name the Greeks gave them: "Phoenicians". It is not within
the scope of this paper to deal with the issue of Canaanite involvement in
Egyptian sea-going shipping and the connections of both with Early Minoan
Crete. All we are trying to present is an independent case, based solely on
the type of marine vessel which is characterized by a unique hull shape and
predominant vertical posts, which might indicate, when followed through
time and space, that the combined evidence of ancient texts, whether
Biblical, Ugaritic, Greek or Latin, concerning the origins of the Canaanites
from the Red Sea (for a full length up-to-date discussion, see Rollig, 1983;
Salles, 1993), might not be dismissed so easily.
In this context there is room here to refer the reader to two additional
texts, aside from those of Homer, Herodotos and Strabo. The first is chapter
10, verse 6 in the Biblical book of Genesis, in which Canaan is designated
as the son of Ham and a brother to Cush (Nubia), Mitzraim (Egypt) and Put.
Among the offspring of Mitzraim are the Caphtories (the ancient people of
Crete), from whom the Philistines were descended (Gen. 10:14). The
second is the Ugaritic epos of King Kreth who had sought a bride as far
south as Udum by the Red Sea (Gordon, 1949), as it was the custom in
those days to marry within the nation; going back to its place of origin (as
Isaac went back to Aram-Naharaim for Rebecca, Gen. 24:lO). The last items
are the petroglyphs from Nahal Gishron near Eilath, on the ancient road from
the Red Sea to the Mediterranean (the Kounthilas Road), in which two
angular ships with upright sterns and stem posts are depicted (Rothenberg,
1967: 158-59, Fig. 231), here (Fig. 6a).
The second sub-type of the angular boats from Pre-dynastic Egypt is

best depicted as the "Black Ship" from the painted wall of tomb 100 in
Hierakonpolis (Quibell & Green, 1902, PI. W ;Landstrom, 1970: 14, Fig.
17). Though considered "foreign", its alleged Mesopotamian origins have
been refuted by scholars (Frankfort, 1924: 93-95). Other pictures of that
rather strange style of hull are to be found among the Pre- and Proto-
dynastic petroglyphs from the Eastern Desert and Nubia (Engelmayer, 1965,
Pls. XII, 4, 14). A variant of this hull has its stern rising at an angle of about
50", which first appears during the first dynasty era (e.g., Williams, 1980: 16;
Landstrom, 1970: 25, Figs. 73-75). In all these pictures the almost vertical
post is clearly at the fore end side, as indicated by the fixed bench behind it
and the dangling bundle from its top (Landstrom, 1970: Figs. 17,42,79). The
clearly depicted helmsman at the lower side of the rock-engraved boat from
the Nubian desert (Engelmayer, 1965, Pls. XII, 4) just verifies this conclusion.
It is difficult to explain the function of such a high and heavy prow post, and
its effect on the hydrodynamic navigability of that type of vessel, even when
assuming its relative size and prominence as artistic bias. In the case of sea-
going vessels, which might have sailed on high seas for long distances, such
high prows could be used as a navigational aid, during night sailing, for the
helmsman to "shoot" stars on the vertical line of the prow post, the mast and
his eye. Yet, what could have been its function for riverine craft, or in a boat
propelled by paddlers? Whatever function this high and heavy prow may
have served, its uniqueness may be used as a cultural benchmark; and as
such, its resemblance to the Early Bronze Age boats from the Cycladic
Islands of the Aegean (Basch, 1987: 77-84) is rather intriguing.
Again, it is not the aim of this paper to repeat all the known arguments
concerning this strange type of marine vessel and the tantalizing issue of
defining its stern from its prow (see, for example, Casson, 1971: 30-31;
Basch, 1987: 83-85: Vinson, 1994: 15; Wedde, 1995: 489-491). The relevant
issue here is the actual similarity between the late 4th-millenniumexceptional
variant of hull from Egypt and the earliest depicted type of sea-going vessel
from the Aegean and Crete (the famous three-dimensional clay model from
Palaiokastro, dated to the Early Minoan Period; and see for example,
Marinatos, 1933: 173, Fig. 19). To the "technical" similarity of the unique
profile of the hull (including the raised angular aft), one might add the
"dangling bundle", which characterizes both the Egyptian depiction of
ceremonial context and all the items from the frying pans, or "Pollens" from
Syros (Basch, 1987: Fig. 159-168). The only change is the omitted palm
bench and the additional fish above the tip of the bow post at the later
groups. Some scholars define these Aegean boats as an autochtonic type of
dug-out canoe, which would be ideal for a geographic area abundant with
long, straight conifer trees (Renfrew, 1972: 348; Casson, 1971: 30-31,41-42;
Wedde, 1995: 491, n. 12); others would reconstruct their hulls as having
been composed of planks, sawn or fixed by mortises and tenons (Basch,
1987: 85-88; Vinson, 1994:15). None realized that it would have been almost
impossible to sail these boats in open seas without an outrigger, in order to
avoid eventual capsizing. One should also wonder how a long, narrow
canoe, with a heavy and prominent prow, which is hardly suitable for a
riverine voyage, became the earlier iconographic representative type of
marine vessel in the Early Bronze Age Aegean Sea.
Having only a handful of rather sketchy iconographic evidence, one

cannot produce a well-based explanation for these alleged discrepancies.
Yet, referring to the heated arguments as concerning the origins of the
Aegean and Early Minoan civilizations, the resemblance of this rather
"strange" type of Cycladian boat and a particular variant of an earlier, Proto-
dynastic vessel from Egypt might add something of substance on the side
of the Diffusionists who would follow the old Biblical claim that Ham (=
Africa) begat Mitzraim (Egypt) who begat Caphtor (Crete) (Gen. 10:14).
These Aegean canoes, which were far from being primitive and would
represent continuous technical development in nautical engineering over
many centuries (Basch, 1987: 81), may be used as an additional argument
for other aspects of the alleged "Libyan Diffusion" into Crete and even
mainland Greece, including actual artifacts dated to the late Neolithic and
Early Bronze Age (Bernal, 1991: 95-99).
Prof. Avner Raban

The University of Haifa
BIBLIOGRAPHY
Andelkovic, B., The Relations between Early Bronze Age I Canaanites and Upper Egyptians
(Belgrade: The University of Belgrade, Center for Archaeological Research, Vol. 14,
1995).
Arkell, A., "Varia Sudanica", Journal of Egyptian Archaeology, 36 (1950), pp. 24-40.
Arkell, A., "Early Shipping", Antiquity, 33 (1959), pp. 52-53.
BAR = Breasted, J.H., Ancient Records of Egypt, Vol. 1 (Chicago: Chicago University Press,
1906).
Bard, K.A., From Farmers to Pharoahs - Mortuary Evidence for the Rise of Complex Society in
Egypt (Sheffield Academic Press, 1994).
Basch, L.,-''L~~ a v i mnS
r et autres notes de voyage en Egypte", Mariner's Mirror, 64 (1978),pp.
99-123.
Basch, L., Le Musee imaginaire de la marine antique (Athens: lnstitut Hellenique pour la
Preservation de la Tradition Nautique, 1987).

Bass, G.F. (ed.),A History of Seafaring Based on Underwater Archaeology (London: Thames &
Hudson, 1972).
Baumgartel, E.J., The Cultures of Prehistoric Egypt, vol. 2 (Oxford: Oxford University Press,
1960).
Bernal, M., Black Athena, vol. 2: The Archaeological and Documentary Evidence (Brunswick, NJ:
Rutgers University Press, 1991).
Borchardt, L., Des Grabdenkmal des Konigs Sahu-Re, vol. I1 (Leipzig, 1913).
Bowen, R. La Baron, Jr., "Egypt's Earliest Sailing Ships", Antiquity (1960), pp. 117-131.
Casson, L., Ships and Seamanship in the Ancient World (Princeton University Press, 1971).
Casson, L., The Ancient Mariners, 3'dedition (Princeton University Press, 1991).
Clowes, G.S.L., Sailing Ships (London, 1932).
Collon, D., First Impressions - Cylinder Seals in the Ancient Near East (The University of
Chicago Press, 1987).
Dakoronia, F., "War-ships on Sherds of LH I l l Kraters from Kynes?" in H. Tzalas (ed.) Tropis 111,
Proceedings of the Third International Symposium on Ship Construction in Antiquity
1989 (Athens: Hellenic Institute for the Preservation of Nautical Tradition, 1995), pp.
147-148.
Derry, D.E., "The Dynastic Race in Egypt", Journal of Egyptian Archaeology, 42 (1956), pp. 80-
85.
Diop, C.A., The African Origin of Civilization, Myth or Realrty? (Translated from the French by M.
Cook) (New York: Lawrence Hill & Co., 1974).
Emery, W.B., Archaic Egypt (Harmondsworth: Penguin Books, 1961).
Engelmayer, R.. Die Felsgrarirungen im Distrikt Sayala-Nubien, 1 - Die Schiffdarstellungen
(Vienna: Oesterreichische Academie der Wissenschaften, Philosophisch-Historiche
Klasse, Denkschriften 90, 1965).
Faulkner, R.O., "Egyptian Seagoing Ships", Journal of Egyptian Archaeology 26 (1940), pp. 3-9.
Frankfort, H., Studies in Early Pottery of the Near East (London, 1924).
Frankfort, H., The Birth of Civilization in the Near East (Bloomington: University of lndianna
Press, 1951).
Gordon, C.H., Ugaritic Literature (Rome: Pontifical Biblical Institute, 1949).
Helck, W., Die Berichungen Aegyptens zu Vorderasien im 3 und 2 jahrtausend v. chr.
(Wiesbaden, 1962).
Hornell, J., Water Transport (Cambridge University Press, 1946).
Kantor, H.J., "The Final Phase of Predynastic Culture: Gerzean or Semainean?",Journal of Near
Eastern Studies 111 (1944), pp. 1 10-136.
Kantor, H.J., "Further Evidence for Early Mesopotamian Relations with Egypt", Journal of Near
Eastern Studies XI (1952). pp. 239-250.
Kantor, H.J., "The Relative Chronology of Egypt and Its Foreign Correlations Before the Late
Bronze Age" in R.W. Ehrich (ed.), Chronologies in Old WorldArchaeology (University of
Chicago Press, 1965). .
Landstrom, B., Ships of the Pharaohs, 4000 Years of Egyptian Shipbuilding (Garden City, NY:
Doubleday & CO., 1970).
Marinatos, S., "La Marine creto-mycenienne." Bulletin de Correspondence Hellenique 57
(1933),pp. 170-235.
Moorey, P.R.S., "From Gulf to Delta in the Fourth Millennium: The Syrian Connection", Ere&
Israel 21: Ruth Amiran Volume. Jerusalem (Israel Exploration Society, 1990), pp. 62*-
69*.
Nelson, H.H., "The Naval Battle Pictured at Medinet Habu", Journal of Near Eastern Studies I1
(1943),pp. 40-45.
Oppenheim, A.L., "The Seafaring Merchants of Ur", Journal of the American Oriental Society, 74
AVNER RABAN TROPIS Vl
(1954), pp. 6-17.

Petrie, W.M.F., Corpus of Prehistoric Pottery of Egypt (London, 1921).
Petrie, W.M.F., "Egyptian Shipping", Egyptian Archaeology, 18 (1933), pp. 1-14.
Potts, D., "The Chronology of the Archaeological Assemblages from the Head of the Arabian
Gulf to the Arabian Sea (8000-1750 BC)" in R.W. Ehrich (ed.), Chronologies in Old
World Archaeology, 3" edition (University of Chicago Press, 1984).
Potts, D., "Contributions to the Agrarian History of Eastern Arabia, II: The Coltivars: Arabian
Archaeology and Epigraphy 5 (1994), pp. 236-275.
Poujade, J., Trois Flottilles de la VI" Dynastie des Pharaons, Facs. 1 (Paris: Documents
d'archeologie navale, 1948).
Prag, K., "Byblos and Egypt in the Fourth Millennium BC", Levant, 18 (1986), pp. 59-74.
Quibell, J.E. & F.W. Green, Hierakonpolis, Vol. 11 (London, 1902).
Raban, A,, "The Medinet Habu Ships: Another Interpretation", The lnternational Journal of
Nautical Archaeology and Underwater Exploration, 18.2 (1989), pp. 163-171.
Raban, A., "The Enigma of the Long Planks Predynastic Boats on the Upper Nile" in H. Tzalas
(ed.), TROPIS IV: Proceedings of the Fourth lnternational Symposium on Ship
Construction in Antiquity, Athens 1991 (Athens: Hellenic Institute for the Preservation of
Nautical Tradition, 1996).
Renfrew, C., The Emergence of Civilization: The Cyclades and the Aegean in the Third
Millennium BC (London: Methuen Press, 1972).
Rice, M., Egypt's Making, the Origins of Ancient Egypt 5000-2000 BC (London-New York:
Routledge, 1990).
Rice, M., The Archaeology of the Arabian Gulf, c. 5000-322 BC (London-New York: Routledge,
1994).
Rollig, W., "On the Origins of the Phoenicians", Bewus, 31 (1983), pp. 79-83.
Rothenberg, B., Negev: Archaeology in the Negev and the Aravah (Tel Aviv: Massada Press,
Hebrew, 1967).
Save-Soderbergh,T., The Navy of the Eighteenth Egyptian Dynasty (Uppsala & Leipzig: Uppsala
Universitets Arsskrift, 6, 1946).
Save-Soderbergh, T. Four Eighteenth Dynasty Tombs (Private Tombs at Thebes I ) (Oxford
University Press, 1957.
Salles, J.F., "Les Pheniciens de la Mer ~rythree", Arabian Archaeology and Epigraphy 4 (1993),
pp. 170-209.
Vinson, S., Egyptian Boats and Ships (Buckinghamshire: Shire Egyptology, Shire Pub., 1994).
Wainright, G.A., "Pharaonic Survivals Between Lake Chad and the West Coast", Journal of
Egyptian Archaeology, 35 (1949), pp. 170-175.
Wedde, M., "Bow and Stern in Early Aegean Bronze Age Ship Imagery - Areanalysis" in H.
Tzalas (ed.) Tropis, 111, Proceedings of the Third lnternational Symposium on Ship
Construction in Antiquity, Athens 1989 (Athens: Hellenic Institute for the Preservation of
Nautical Tradition, 1995), pp. 485-506.
Wenke, R.J., "Egypt: Origins of Complex Societies", Annual Review of Anthropology, 18 (1989),
pp. 129-155.
Williams, B., "The Last Pharaohs of Nubia", Archaeology, 33.5 (1980). pp. 12-18.
Zarins, J., "The Early Settlements of Southern Mesopotamia: A Review of Recent Historial,
Geological and Archaeological Researches",Journal of the American Oriental Society,
1 12 (1992), pp. 55-77.
CAPTIONS FOR THE ILLUSTRATIONS
1. One side of the ivory knife handle from Gebel el 'Arak (after Emery 1961, Fig. 1).
2. The boat "Procession" from Hierakonpolis tomb 100 (after Quibell & Green, 1902, PI. UO(V).
Note the "Black Ship" in the center.
3. The sailing boat depicted on a Gerzean vase BM. 35324A (after Casson, 1971, Fig. 6).
4. Petroglyph of high-bowed boat from Nubia (after Engelmayer 1965, PI. X11.4).
5. Angular sailing boat of the late 4'" millennium BCE, carved on a stone censer from Nubia
(drawn by H. Dinkel, after Williams, 1980: 16).
6. Two petroglyphs from Nahal Gishron, near Eilath, by the Red Sea:
a. Probably 4'" millennium BCE one (drawn by the author after Rothenberg, 1967, Fig. 231).
b. Probably 12mcentury BCE one (photo by the author).
AVNER RABAU -- -- -- -- .-- - .
TROPIS VI
- -
1 I
Fig. 1
Fig. 2
THE ENIGMA OF THE ANGUMR SAILING SHIPS
- -- -- IN THE RED SEA SINCE
- THE 4'" MILLENNIUM
-- - BCE- --
Fig. 5
Fig. 6b
Fig. 6a
THE COASTAL LANDSCAPE BETWEEN THERMOPYLAI AND
DEMETRIAS
FROM A MARITIME POINT OF VIEW
Introduction
Archaeology has contributed much to our knowledge of ship constructi-

on. Ships and boats are complex structures, so the emphasis in maritime
archaeology on ship construction is understandable. Maritime archaeology,
however, is not restricted to the technical aspects of boats and ships. In this
paper I would like to pay attention to the coastal landscape between the
springs of Thermopylai and the town of Demetrias, the area where the city of
Lamia, the host town of the Sixth International Symposium on Ship
Construction in Antiquity, is situated.
The aim of this paper is to determine the importance of the cities along
the coast from a maritime point of view and to find an answer to the question
why there were so many ports in this area in Hellenistic times. In the first
place I would like to pay attention to the available written sources that
provide information on the coastline between Thermopylai and Demetrias,
the sailing routes and the landmarks along the coast. Information on the
shifting of the coastline in ancient times has been obtained in geological and
paleogeographical investigations. This paper is not restricted to 'antiquity',
because maritime activity in this area continued until the Ottoman conquest
of Thessaly in AD 1393.
Overseas contacts of Demetrias and New Halos
During the past two decades archaeological excavations have been

conducted in two Hellenistic towns along the Pagasitikos Gulf: Demetrias
and New Halos. The research focused on the location and the layout of the
cities, the public buildings and residential houses, and the relation between
the city and its territory. Demetrias was founded by Demetrios Poliorketes
around 290 BC. In the 3'" century BC it was, after Pella, the second capital of
the Macedonian kingdom, and served as a naval harbour for the Antigonids
(Marzolff, 1980 & 1994). New Halos was in all probability also founded by
Demetrios Poliorketes (Bakker & Reinders, 1996), in 302 BC, before he
departed from Thessaly for Asia Minor to assist his father Antigonos in his
struggle against Lysimachos and Seleukos. New Halos lay at a strategic
point between a spur of Mount Othrys and a back-swamp bordering the
Pagasitikos Gulf.
The strategic position of New Halos was evident from the very start
(Reinders, 1989), not only from the town's situation, near a narrow passage
along the coastal route between northern Thessaly and central Greece, but
also from its impressive 4.7-km-long enceinte, reinforced with at least 117
towers. The town may indeed have been founded specifically for strategic
reasons. In our first publications on the investigation of New Halos we paid
little attention to maritime aspects of the city. However, those aspects
certainly deserve attention because maritime strategy was Demetrios's
strong point: in 'poliorcetics' he was often less successful. The excavation of
the remains of six houses in New Halos yielded information on the period of
habitation, but also evidence for a different interpretation. In addition to
agricultural implements, storage jars, amphoras and animal bones, some
150 coins were found, which showed that the city of New Halos was
abandoned around 265 BC (Reinders, in prep.). The provenance of these
coins indicates with what cities New Halos was in contact. The coins point to
contacts with neighbouring cities in Achaia Phthiotis, like Peuma, Thebai,
Ekkara and in particular Larisa Kremaste, but they also suggest contacts
with cities on the island of Euboia, such as Histiaia and Chalkis, and with
Lokris. As Furlwangler (1990) has already demonstrated, most of New Halos'
contacts were maritime, oriented towards the south; coins from other cities
in Thessaly are rare, as are coins struck by the Macedonian kings (fig. 1).
Unlike New Halos, which struck its own coins, the nearby Hellenistic city
of Demetrias relied on the emissions of the city of Larisa. The coins found
during the excavations in Demetrias reflect contacts with the north (fig. 2),
with the cities in the northern part of Thessaly and Macedonia. Virtually no
coins from cities in Achaia Phthiotis or the island of Euboia were found.
Furlwangler (1990,235-40) assumes that in the first quarter of the 3rdcentury
BC Dernetrias' contacts were restricted to Magnesia, the eastern part of the
Pagasitic Gulf and, possibly, the city of Histiaia on the northern shore of the
island of Euboia.
Trade and transportation are reflected in other artefacts, too. Stamps on

amphora handles provide information on the importation of wine. The site of
Demetrias yielded a large number of amphoras from Thasos and a relatively
small number from Rhodos. When we compare these data for the period
THE COASTAL LANDSCAPE BETWEEN THERMOPYLAI AND DEMETRIAS
around 200 BC with the numbers of amphora stamps found in Athens we

observe a preference for wine from Rhodos; the number of stamps from
Thasos is low (Furtwangler, 1990). Unfortunately no amphora stamps were
found in Halos, so we know nothing about this city's long-distance contacts
in the early rdcentury BC. Although the exact nature of the contacts of New
Halos is not clear at the moment, the coins suggest overseas contacts with
the ports of Euboia and Lokris. Does this mean that New Halos served as a
port of call for part of Achaia Phthiotis? It was this question that led us to
reconsider the situation of New Halos and to study the coast between
Thermopylai and Demetrias from a maritime point of view.
Skylax, Strabon and Artemidoros
The oldest description of the coast between Thermopylai and Demetrias

is a periplous ascribed to Skylax of Karyanda, who lived around 500 BC
(Miiller, 1965 = GGM 1). The preserved version of thisperiplous (GGM 1, 15
ff.) is however a compilation of later date, presumably written in the 4m
century BC, possibly between 338 and 335 BC (Gisinger RE, Skylax;
Kretschmer 1909, 153). Along the coast,Cni 8 a A a ~ ~(GGM
g 1, 49-51), this
periplous mentions, amongst other locations, Thermopylai and the river
Spercheios and the cities of Lamia and Echinos in Malis; Echinos became
part of Malis in 351 BC. Skylax continues with the towns of Achaia Phthiotis:
Andron, Larisa, Melitaia, Demetrion and Thebai. The absence of Halos in this
enumeration is not surprising, because the city of Old Halos had been
destroyed after a siege by a Macedonian army under Parmenion in 346 BC
(Reinders, 1988 and 1993). To the north of these towns, Amphanaion and
Pagasai are mentioned along the shores of the Pagasitikos Gulf. In addition
to these towns and rivers along the coast, Skylax also mentions the distance
of a voyage circumnavigating Attica, Boiotia and Lokris; the distance
between Cape Sounion and the border of Malis near Thermopylai is said to
be 1300 stadia. Skylax gives only a few names of Thessalian cities in the
hinterland, Cv (a&)p&ooy&q.
Another important source of information on the coast between Piraeus

and Thessalonike in Hellenistic times is Strabon's Geography. Strabon was
born in Amasia in Asia Minor in 64 or 63 BC. He travelled to Rome many
times and went on voyages to many other countries, but he never visited
Thessaly. In his Geography he quotes passages from many authors, so his
information comes from different sources and different periods.
One of the sources Strabon used for his description of Greece is
Apollodoros' Commentary on the Homeric Catalogue of Ships. Strabon
hence interwove the geography of his own time with that of Homeric times.
The information about Homeric times comes from the Catalogue of Ships in
book 6 of the Iliad. This Ship Catalogue enumerates the towns that provided
troops for the Trojan War. It is generally accepted that the list to a certain
extent refers to Late Bronze Age geography, from before 1200 BC. Some of
the towns mentioned, such as Pylos, are known to have flourished until that
period. The site of Pylos has yielded no archaeological evidence for
occupation after the Mycenaean period (Hope Simpson & Lazenby 1970,
82). The names of the towns in some regions, like the Peloponnesos, can be
associated with archaeological sites. Many of the towns in our area,
however, cannot be identified on the basis of archaeological evidence.
The information that Achilles came from the area northeast of Lamia has
engendered and indeed continues to engender much speculation about his
home. A small number of 'Homeric' place names can be identified with more
or less certainty. Excavations in present-day lolkos, a short distance to the
west of the city of Volos, have yielded evidence for occupation in the Late
Bronze Age, suggesting that the site may be identified as lolkos (Hope
Simpson & Lazenby 1970, 136), although it has recently been proposed that
the Mycenaean site Dhimini may have been the Homeric city of lolkos
(Intzesiloglou, 1994). The site of Pyrasos was on a large magula at a short
distance to the east of the village of Nea Anchialos (Hope Simpson &
Lazenby 1970, 132). The southern part of the Almiros Plain, the Krokion Plain
of classical times, is known to have contained two large towns: Old Halos
and New Halos. Hope Simpson (1970, 126) places Homeric Halos near the
city of Lamia, although another site that was also occupied in the Late
Bronze Age was found during an archaeological survey in Voulakaliva, a little
to the north of Hellenistic Halos (Reinders, 1993). Smaller archaeological
sites have been identified as Antron and Pteleon, but no indisputable
evidence for Late Bronze Age settlements has yet been found.
If we omit the Homeric information Strabon borrowed from Apollodoros,

we are left with a less complicated account. The description of the
geography of the coastal area of Thessaly is almost certainly based on
Artemidoros's. Around 100 BC Artemidoros of Ephesos wrote a work
comprising eleven books. Strabon's description contains sections from
Artemidoros' work or from similar periplous-like descriptions. The
description of the coast from Thermopylai to Demetrias forms part of a
longer account describing the route from Piraeus to Thessalonike. Strabon,
following Artemidoros (?), divided this route into six sections, each
characterised by an important landmark, namely Cape Sounion, Euripos
Strait , the springs of Thermopylai, the city of Demetrias, Peneios River and
the city of Thessalonike (fig. 3). The length of each section is given in stadia
(table 1). Strabon also mentions other landmarks that were important for
sailors, such as Cape Poseidion, the projecting headland of Keraion, the
small island of Myonnesos, a submarine reef, the Krokion Plain and Cape
Pyrrha.
The springs of Thermopylai lie at the beginning of the section covering

the districts of Malis, Achaia Phthiotis and part of Magnesia. Various names
of cities are given, with the distances between them expressed in stadia. The
majority of these cities flourished in the third century BC. Along the Maliakos
Gulf, Strabon mentions the city of Lamia and its port Phalara. Next come the
towns of Echinos, Larisa, Antron and Pteleon, which formed part of Achaia
Phthiotis. Along the western shore of the Pagasitikos Gulf the cities of Halos
and Thebai lie at either end of the Krokion Plain; Demetrias lies on the
southern shore of Volos Bay (fig. 4).
A few sites in the hinterland are mentioned, like Homeric lton and
Phylake. It is doubtful whether the sites lton and Phylake were still known in
Hellenistic times. Distances in stadia to Halos and Thebai are given, but no
Hellenistic remains have been found at the implied sites. The information on
the geography of the inland Thessalian Plains is very poor in comparison
with that on the coastal region. The geography of the coastal area was
important for sailors. The town of Larisa Kremaste, die schwebende genannt
wegen ihrer vom Meere aus gesehen himmelhohen Lage (Stahlin 1924, 182),
was even named from a sailor's viewpoint. In his description of the section
of the coast between Thermopylai and Demetrias, Strabon mentions
Artemidoros (9.2.5). Assuming that Strabon based his description on
Artemidoros, we have attempted to 'reconstruct' this part of his periplous
(appendix 1).
The shifting of the coastline
The greater part of the coast between Thermopylai and Demetrias is

steep and rocky, here and there intersected by small bays offering sheltered
anchorage. This part of the coast is covered in the British Admiralty, chart
no. 1556 printed in 1890. From this map we may infer that the coastline has
not changed that much since the lgthcentury. Palaeogeographical research
has shown that further back in time, however, certain parts of the coast
underwent dramatic changes, in particular after the Classical period.
The coastline a short distance to the north of Thermopylai has shifted

under the influence of the river Spercheios (fig. 5). Strabon's description of
this area, the distances between various landmarks and cities, the outlet of
the Spercheios and the courses of the rivers Dyras, Melas and Asopos has
been closely studied (Radt 1994, 33-4). Strabon mentions that the city of
Lamia lay 25 stadia from the sea, whereas the present shoreline lies more
than 10km from Lamia.' lnvestigations in the area of the mouth of the
Spercheios (Kraft et a/. , 1987) revealed that in 480 BC the shoreline lay close
to Thermopylai. The site where in 480 BC a Greek army resisted a Persian
army led by Xerxes nowadays lies buried beneath a 20-m-thick sediment.
After 480 BC in particular the river transported enormous amounts of
sediments into the Maliakos Gulf. Deforestation of the mountains caused
denudation of the soil and large amounts of sediments were washed
downhill during heavy rainfall. This led to the formation of a wide mud plain,
which also covered the original site of Thermopylai on a narrow stretch of
land between the mountains and the sea. lnvestigations by Kraft (eta/. 1987)
have shown exactly how the coastline shifted between 2500 BC and AD 1972
and demonstrated the aforementioned rapid sedimentation between 480 BC
and 1972.
The coastline in the area of Volos Bay, the northern part of the
Pagasitikos Gulf, has also shifted over the ages (fig. 6). In 3500 BC Volos
Bay extended up to the Late Neolithic site of Dhimini. In the period 3000-350
BC considerable sedimentation occurred here, which stagnated after that
time. We know that the Late Bronze Age site of lolkos and the city of
Demetrias lay close to open water. According to Zangger (1991), the
sedimentation in Volos Bay was also caused by deforestation, but on a much
smaller scale than that at the mouth of the Spercheios.
Smaller bays along the coast were closed off from the open sea by
beach ridges, which created lagoons that later evolved into back-swamps.
This development has been studied in a back-swamp along the shore of
Sourpi Bay, close to the area where the cities of Old and New Halos lay
(Reinders, 1989). The investigation showed that an inlet of Sourpi Bay was
gradually filled with marine sediments and that, around 1500 BC, a beach
ridge was formed with an open lagoon behind it (Bottema, 1988; Van
Straaten, 1988). It was on this beach ridge that the port Old Halos was
situated (fig. 7). The lagoon was gradually filled with sediments transported
by the Amphrysos, the small river originating at Kefalosi spring. Nowadays a

beach ridge constitutes an almost uninterrupted shoreline from the outlet of
the Platanorrema to the outlet of the Salamvrias near Korakonisi Peninsula.
It was not only perennial rivers like the Spercheios that discharged
sediments. In 1994 we observed that tremendous amounts of sediments are
occasionally washed down from the mountains by rivers whose beds are dry
in the summer. In October 1994, after a short period of heavy rainfall, the dry
beds of the Xerias and the Platanorema in the Almiros Plain changed into
wild brown torrents. With a deafening noise, the 60-m-wide and 3-4-m-deep
Xerias River transported boulders, sand, pebbles, cobbles and the garbage
of illegal dumps to the Pagasitikos Gulf. In the upper part of the plain the
riverbeds are quite wide, but along the rivers' lower courses further towards
the coast the beds are narrow and shallow. Near the mouth of the
Platanorema large areas of arable land were flooded and covered with
cobbles, pebbles and other sediments. The discharge of rainwater and
sediments from the mountains also led to the creation of small underwater
deltas beyond the mouths of the four rivers in the Almiros Plain.
Geological and seismic research have shown that the Pagasitikos Gulf is
gradually silting up. In the central and eastern parts of the gulf, which are
from 80 to over 100m deep, there are Holocene marine sediments with
thicknesses of up to 16m (Mitropoulos & Michalidis, 1988). The greater part
of the sea bed is covered with silt. Coarser sediments, sand, sandy silt and
silty sand are to be found in the northern part of the gulf and at Volos Sill, the
opening of the gulf to the south between Trikeri and the mainland
(Perissoratis et a/., 1988). The western part of the gulf, along the Almiros
Plain, is relatively shallow, with depths from 20 to 40m. Coarser sediments,
washed down from the mountains by the four rivers, are to be found along
the west coast.
Generally speaking, the coastline has shifted only there where rivers like
the Spercheios and Amphrysos empty into the Maliakos and Pagasitikos
Gulfs. Unfortunately, however, it is precisely in these areas that some
important ports, such as Demetrias, Halos and Phalara, were situated in
ancient times. The beach ridge on which Old Halos was situated is now
covered with 1-2 metres of sediment. Because of this, and the complicating
factors of relative changes in sea level and relative land motion due to
tectonic activity, we know fairly little about the ancient ports in this area.
REINDER REINDERS TROPIS VI
Ports in Classical and Hellenistic times
Between Lamia and Demetrias there was a large number of towns which,
as we know from written sources and archaeological research carried out
over the past 25 years, flourished in Hellenistic times in particular. The
number of towns is indeed surprisingly high in comparison with what we
know about other regions along the coast between Piraeus and
Thessalonike. Demetrias and New Halos were both founded in the early 3'
century BC (Marzolff, 1980 and 1994; Reinders, 1988). Some of the other
towns were founded at earlier dates, but were still occupied in Hellenistic
times, in many cases with an enlarged enceinte. Only the towns of Antron
and Pteleon are relatively unknown. The remains of the towns along the
Maliakos Gulf - Lamia, Phalara and Echinos - now lie buried beneath later
buildings, but rescue excavations have shown that these towns, too,
flourished in Hellenistic times, and also in the Roman and Byzantine periods
(Pantos, 1994).
Strabon describes the coast and the ports largely as they were in Late
Hellenistic times, but he occasionally also refers to classical times, for
example with respect to Pagasai, which information he may have borrowed
from Apollodoros (figs. 8 & 9). Some of the towns varied in importance in
different periods. In classical times Pagasai was the main port of Thessaly.
In 377 BC Jason, the tyrant of Pherai, sent grain from Pagasai to the
Boeotian city of Thebai (Garnsey et al., 1984; Xen. Hell. 5.4.56). Grain was
transported from Thessaly to Kos (Sherwin-White 1978, 110) and Rome in
the 3rdand Tdcenturies BC, respectively. In Hellenistic times Demetrias
came to be the chief port in this part of the Pagasitikos Gulf. Pagasai became
a fairly insignificant town.
Pagasai was in classical times the port of the inland city of Pherai
(Strabon 9.5.15). The town of Pagasai was probably situated on the Soros,
a conical rock, so a nearby stretch of beach will have served for beaching
the vessels (Marzolff 1994, 256). lntzesiloglou (1994, 49-50), however, is of
the opinion that the town on the Soros was Amphanai and that Pagasai was
situated on the southern shore of Volos Bay, near Pefkakia. The topography
of the area around the Gulf of Volos is a much-discussedtopic. In Hellenistic
times a newly founded city, Demetrias, was the naval station and residence
of the Macedonian kings. The town must have had a good harbour near
Pefkakia, where there is now a shipyard (Marzolff, 1980). Perhaps the
southern shore of the small Volos Bay was less silted up in Hellenistic times
than it is today. Demetrias was undoubtedly an important port in Hellenistic
times, although Strabon (9.5.15) mentions that the city's power had already
dwindled somewhat by his time.
The next town along the coast of the Almiros Plain was Pyrasos, the port
of Phthiotic Thebai, situated near present-day Nea Anchialos. Strabon
(9.5.14) mentions the name of the port that succeeded Pyrasos in Hellenistic
and Roman times: Demetrion. In the Byzantine period this town took over the
position and the name of the inland city of Thebai and became the principal
centre in this area (Asimakopoulou, 1982).
On the opposite side of the Almiros Plain the important harbour of Old
Halos was favourably situated on the beach ridge south of the Platanorema.
Herodotos (7.173) reports that the Greek fleet disembarked an army of
10,000 men at Halos, as Xemes' army was approaching Tempe. In its
sheltered position in Sourpi Bay, the beach near Old Halos was evidently
excellently suitable for beaching the vessels and disembarking such a large
army. From Halos, the Greek army set off overland in the direction of Tempe,
but returned to their ships when Xemes avoided the route through the
Tempe defile.
On another occasion, in 346 BC, an Athenian embassy en route to Pella

disembarked at Halos when it was besieged by a Macedonian army
commanded by Parmenion (Demosthenes 19.163). The city of Halos
managed to resist the siege of the Macedonian army thanks to the help of
the Athenians, who were able to reach the Thessalian city overseas. The
Athenians made peace, but excluded Halos from the peace treaty and left
the Halians to their fate. The city was subsequently taken and its territory was
given to the Pharsalians. The 11-ha-large Magoula Plataniotiki has been
identified as the site of Old Halos (Reinders, 1993). At the end of the 4m
century BC New Halos was founded. Hellenistic Halos was situated at a
strategic location, 2km from the shore. We may assume that the beaches
south of the Platanorema will have been excellently suitable for beaching in
Hellenistic times, too.
Harbours, or rather landing places, along the Maliakos Gulf are also
mentioned in written sources. In 302 BC Demetrios Poliorketes disembarked
32,000 men near Larisa Kremaste (Diodoros 20.1 10.3), which lay on a
promontory, 2.5km from the sea. The city of Echinos is not referred to as a
port in written sources, but the town was in that respect ideally situated, at a
distance of 1.5km from a beach in a sheltered bay. Larisa Kremaste and
Echinos both formed part of Achaia Phthiotis. At Echinos remains from the
classical period up to Roman times have been excavated (Fotini-
Papakonstantinou 1994,231-232).The next town along the coast of the
Maliakos Gulf was Phalara, the port of Lamia, situated in the district of Malis.
At the beginning of this century, Stahlin was of the opinion that Phalara was
situated near the mouth of the river Achelos. Although there is no sound
evidence to support it, the present-day port of Stilis has been identified as
the site of ancient Phalara, situated close to the sea (Pantos 1994,221).
Phalara was destroyed by an earthquake in 476 BC, but excavations at Stilis
have shown that the site was re-occupied from the second part of the 4m
century BC onwards.
Figures 8 and 9 present surveys of. the main ports in classical and
Hellenistic times. Beaches near the aforementioned towns served as
'harbours', landing places or intermediate stations for voyages and troop
movements; in those days there were presumably no harbours in the proper
sense, with moles and quays. Ships could easily be hauled onto the
beaches, or goods and people could be brought aboard and unloaded from
vessels at an anchorage or roadstead with the help of smaller vessels, or
simply by carrying the cargo through shallow water from the beach to a
vessel (Casson 1971,fig. 191).Only occasionally does Strabon use the word
'port' for one of the towns mentioned along the coast from Thermopylai to
Demetrias. The harbour of Demetrias he refers to as a vaOo-raepoq, which
is translated as a 'naval base' in the Loeb translation. He uses different
words for Pagasai and Pyrasos, namely Crtive~ov and eljhipevoq,
respectively, which are translated as 'sea-port' and 'with a good harbour'.
The words vaOo-raepoq and Cnivs~ovare sometimes however translated as
'anchorage' or 'roadstead'. Considering the nature of the area, the latter
translations are preferable in the case of sites like Demetrias, Pagasai and
Halos.
Maritime activity in the Roman and Byzantine periods
The coins recovered in the excavations of the remains of New Halos

show that this Hellenistic city had overseas contacts, but the city was
abandoned in 265 BC already, presumably after an earthquake (Reinders,
1989). Phthiotic Thebai, in the northern part of ttie Krokion plain, then
became the most influential city of this plain. Pyrasos, or Demetrion, the port
of Phthiotic Thebai, is one of the three cities with harbours mentioned in an
inscription of c. 151-150 BC. The towns of Thessaly undertook to deliver
430,000kophinoi of grain to Rome. 'That each of the cities arrange the
transport of its allocated grain down to the harbdur, whether that of the
Demetreion or at Phalara or at Demetrias' (Garnsey et a/. 1984, 37). Stahlin
(1924, 173) assumed that the present-day harbour of Nea Anchialos, in his
days 'ein kleiner Weiher am Meere', is a remnant of the old harbour. The
position of this harbour and its long-distance trade are referred to in the
inscriptions of many gravestones. The 2"d-centuryBC inscription mentioning
the Thessalian grain transports shows that in Late Hellenistic times
Demetrias, Demetrion and Phalara were the main ports of Thessaly (fig. 10).
The importance of Phthiotic Thebai gradually dwindled as the Byzantine

city of Thebai rose to power. This city was situated at the site of the port of
Demetrion, present-day Nea Anchialos (fig. 11). Remains of this city,
including the ruins of three basilicas, have been excavated in Nea Anchialos.
Surveys in the southern part of the Almiros Plain led to the discovery of a
large number of small rural sites dating from Late Roman and Byzantine
times. Near Aidinion archaeologists found the remains of a small rural site
with an olive press (Nikonanos 1971, 312-313). Byzantine Thebai flourished
in the early Byzantine period, when it was the main centre in the Almiros
Plain. No other towns are mentioned in written sources and most of the
archaeological remains from this period that were found during the survey
represent sites of only modest dimensions (Avramea, 1974; Malakasioti et a/.
1995).
Byzantine Thebai retained its powerful position until the end of the Tm
century AD. Traces of a fire were found during excavations in Thebai. There
is evidence for occupation in the 8'" and 9'" centuries, but by that time Thebai
was no longer an important port (Koder & Hild 1976,271). The other port in
the northern part of the Pagasitikos Gulf, Demetrias (Dimitriada), is
described as a port in written sources until the Ottoman conquest of
Thessaly in 1393 (Koder & Hild 1976, 271).
In mid-Byzantine times, before the Ottoman conquest, Almiros was an

important port. In the I T h and 13'" centuries Almiros and Saloniki were
among the principal ports of the Byzantine empire. Communities of Jews
and of people from Venice, Pisa and Genoa lived in Almiros (Koder & Hild
1976, 170-171; Sawidis, 1993; Rizos 1996). Within an enceinte, close to the
outlet of the river Xerios, east of the present-day town of Almiros, glazed
sherds dating from the 12'" and 13'" centuries have been found. In these
centuries Almiros was the main port along the coast of the Pagasitikos Gulf.
Benjamin of Tudela, who visited Almiros, mentions that the town had some
400 Jewish inhabitants (Adler 1907, 11). Archaeological surveys in the
Almiros Plain yielded evidence for the existence of a fort, monasteries and
rural sites in this area in the same period. These sites indicate that medieval
Almiros was not merely a commercial city, but presumably also had a
hinterland in the Almiros Plain (Dijkstra et a/. , 1997).
Periploi were used in Byzantine and Medieval times, too. Little is known
about navigation in Byzantine times, but the I ~ a 6 l a a p o qii rtspinhouq ~ i i q
MsyaAqq 8aAaocqq provides specific data on the distances along the
coasts of the eastern Mediterranean and also information on the nature of
harbours, prevailing winds, the presence of anchorages and water, etc.
(Delatte 1947, XIX; Miiller 1965, CXXVI). This periplous is thought to be a
Byzantine adaptation of a stadiasmb from the 3"1 century AD. Unfortunately
it contains no information on the coasts of the Greek mainland.
The medieval portolans and portolan charts were to a certain extent the
successors of the periploi and stadiasmoi. The importance of the coastal
stretch between Thermopylai and Demetrias is reflected in these portolans.
After the fall of Constantinople in 1204, the territory of the Byzantine empire
was divided amongst the crusaders. Although Thessaly was assigned to the
Peregrini, the non-Venetian crusaders, the harbour of Fteleo remained under
Venetian control. A series of important ports and various other towns and
landmarks are shown on a number of portolan charts drawn in the 14m-16th
centuries (Nordenskiold, 1897). The contours of the coast roughly indicate
protruding capes and sinuosities; the coastline resembles a serrated holly
leaf.
A number of places between Talandi (Atalandi) and Monester (Mount

Pilion) are indicated on almost all the maps from this period (fig. 12 and table
2). Many of the capes and ports can be identified without problems. Only the
names in the region of the Spercheios river are not clear: place-names like
Longiton, or Lo Giton, and Lambena, or Ladena, occur on some charts,
sometimes in reversed order. The city of Lamia - also called Zeitoun or
Zitouni in the Middle Ages - is not mentioned by either of those names on
the portolan charts. The name Zitouni was used to indicate the Kastro of
Lamia: Castrum de Situm super Ravenica, which was rebuilt after 1204
(Fotini-Papakonstantinou 1994, 10). Ravenica was a medieval city that was
probably situated south of Lamia (Koder & Hild 1976, 251). According to
Koder & Hild (1976,283-284), the names Gripton and Gitone were also used
to indicate the city of Lamia between 1218 and 1259, the period that the city
was again in Greek possession. The place names Longito, Lomgito, lo
Gittom, and also Lambena and Ladena occur on the Portolan charts in the
neighbourhood of the outlet of the Spercheios; they presumably refer to the

city of Lamia (Kretschmer 1909, 673).
An important stronghold in medieval times was the castle of Bondenitsa,

present-day Mendenitsa, south of Thermopylai, defending a pass on the
route from northern to central Greece. There was evidently also a port with
the same name (Koder & Hild 1976, 222) - habemus portum et castrum
Bondonicie - which was important for the grain trade (Rizos, 1996). Among
the other towns occurring on the portolan charts besides Almiros and
Dimitriada are Gardikion (Gardikia hetera, Pelasyia) and Fteleo (Pteleo).
Gardikion is said to have a harbour on a late portolan. This must refer to
Gardikios Bay, because the city itself lay inland, at the site of ancient Larisa
Kremaste (Koder & Hild 1976, 161). Fteleo was an important Venetian
stronghold until the fall of Negroponte in 1470, when the town was
surrendered to the Ottomans and its inhabitants were deported to
Constantinople (Koder & Hild 1976, 241). Table 2 presents a survey of the
towns and landmarks on the portolan charts.
Protruding capes give the coastlines on the portolan charts their peculiar
shapes. Two capes were clearly important landmarks along the coast. Cape
San Nicolo is consistently mentioned after Fteleo, so it cannot be identified
as the Cape Poseidion of ancient times or the Cape Stavros of present-day
maps. Along the coast north of Fteleo, Mitzela Bay in the shelter of Ayos
Nikolaos Island provided good anchorage. Byzantine remains have been
found on this island (lakovidis, pers. comm.), suggesting that it may well be
the Cape San Nicolo of the portolan charts, as Kretschrner (1909, 637)
already suggested. Cape Sepias, on the south-east coast of Magnesia, was
called Cape Ayos Yeoryos Zagora in medieval times, Zagora being the
Slavic name for Mount Pilion. Almost all the portolans indicate Monester,
another medieval name for Mount Pilion. Mount Pilion was called KcMla,
(monks') cells, in Greek, because of the great number of monasteries that
were to be found on it (Koder & Hild 1976, 186). A written portolan also
mentions a Cape Monester situated along the east coast of Magnesia: Da
cauo di moster a cauo di verliqui 35 miglia tra maestro e tramontana (Kretsch-
mer 1909,323). If cauo di verliqui is Kavo Dhermatas (Cape Kissavos), cauo
di moster may be Cape Damochari. Apart from the capes and Mount Pilion,
the outlet of the Spercheios is indicated on many portolan charts.
The written Italian portolans prove that navigation was not restricted to
the route between Euboia and the mainland, but also extended to the
Aegean Islands and routes further afield, if weather conditions permitted.
The results of archaeological surveys in the Almiros Plain suggest that most
mid- and Late Byzantine sites lie in the coastal zone (Dijkstra et a/., 1997),
but detailed information obtained in surveys covering the entire region is still
scarce. A large number of mid-Byzantine forts is moreover known along the
east and west coasts of Magnesia, the east coast of Lake Karla and to the
west of New Halos. These strongholds and observation posts were intended
to protect the Byzantine trade routes against piracy and incursions, rather
like Frankish strongholds such as Bondenitsa and Platamona.
The Ottoman conquest of Thessaly
The Venetian documents of the 1lth-13th centuries mention only Almiros

as a port in Thessaly. Important Thessalian trade products were grain from
the Thessalian plains and foothills, oil from the olive groves around the
Pagasitikos Gulf and wine from Fteleo (Rizos, 1996; Avramea 1974, 65-66).
In the 13mcentury Venetian influence in Negroponte (Chalkis) and Fteleo
increased and the latter took over Almiros' position as the most important
port in the region. At the beginning of the ISth century the island of via
(Negroponte), the Maliakos Gulf and the Gulf of Fteleo were still under
Venetian control, whereas Thessaly and Macedonia had already been
incorporated in the Ottoman Empire. After 1393 the town of Volos (Golos),
which was first mentioned in 1333, took over Demetrias' position. A Turkish
garrison was stationed at Golos, which was situated at the site of present-
day lolkos (Koder & Hild 1976, 165-166). During the Ottoman domination,
from 1393 until 1881, it was the only important city and port in the area of the
Pagasitikos Gulf. In this period, too, grain was transported from Thessaly to
Venice. The sultan's permission was needed for this transport and this led to
grain smuggling (Rizos, 1996; Van Keulen, 1728). The portolan charts do not
pay much attention to the political changes, although the Turkish port of
Volos (Golos) is indicated on portolan charts from the end of the 15thcentury.
Besides portolan charts there are also written portolans. Delatte (1947)
published a number of 16th-centuryGreek manuscripts containing coastal
descriptions dating from the Late Byzantine period, or perhaps from after the
Ottoman conquest of Thessaly, but while Negroponte and the island of via
were still under Venetian control. The choice of words reflects influences of
both the Frankish and Venetian languages. In comparison with the periploi
and the late-medieval portolan charts, these written portolans provide
strikingly little or no information on the ports along the coasts of the Maliakos
and Pagasitikos Gulfs; the ports of Almiros, Dimitriada and Fteleo along the
Pagasitikos Gulf are not mentioned at all. Instead, detailed information is

given on the island of via (Eypinou) and the harbours of Negroponte
(Chalkis) and Karistos (Delatte 1947, 224-226). Unlike the portolan charts,
the written portolans do mention the city of Lamia (Zeitun). The distances
between Lamia, Platamona and Saloniki are only roughly indicated, but
much attention is paid to the passage between Skiathos and the mainland
(Delatte 1947, 296-7).
The lack of information on the Pagasitikos Gulf is all the more surprising
in view of the abundant, detailed information that is provided on the Aegean
islands, for instance on the islands of Skiathos, Skopelos and Alonnisos.
This must mean that in Late Byzantine times, or perhaps in the period that
Thessaly was under Ottoman rule while Euboea was still under Venetian
control, maritime activity shifted to the Aegean islands, which then started to
play a prominent role in shipping. The written portolans mention the
harbours and ports of every single island in the Aegean, even the tiniest.
A corresponding amount of detail with respect to the islands and lack of

information on the Maliakos and Pagasitikos Gulfs is observable on the
maps of 163 ports in the Mediterranean published by Roux and Allezard in
1800 (fig. 13). Among these maps there is however one of the Pagasitikos
Gulf, indicating depths in fathoms, shallows, and the rock between Skiathos
and Ayos Yeoryios on the mainland: Roches a fleur d'eau, which was called
Myrmex in classical times. Moreover, many anchorages in the Gulf are
indicated, for example near the Turkish towns of 'Armiro' and 'Vollo' and in
Port Olive behind Ayos Nikolaos Island, and also various as landmarks like
the Isles des Rats, which in the late medieval portolans are indicated as
Pondiki, present-day Pondikonisi. The map by Roux and Allezard is one of
the two printed maps of the Gulf of Volos mentioned in Zacharakis'
catalogue (1992) - a small number compared with the hundreds of maps
available for the Aegean Islands.
We also find evidence for maritime activity and an orientation towards the
Aegean islands in the Greek communities on Mount Pilion. The Greek
villages on Mount Pilion engaged in shipping from the 16" century onwards;
Zagoriana was the general term used for the vessels of these mountain
villages (Makris 1982, 182). One of these maritime villages without a harbour
was Mitzela on the east coast of Mount Pilion. When the Turks set fire to the
village in 1828, the villagers fled to the Aegean Islands; before that year
many of them had fought in the struggle for independence as a sailor or a
soldier. Many families went to the nearby island of Skopelos; 70 families from
REINDER REINDERS TROPIS VI
Mitzela are known to have been living on that island in 1829: a total of 295
persons, 147 male and 148 female. The heads of 37 families were sailors,
others were soldiers or labourers; the heads of 10 families were widows
(Kalianos, 1984). In 1834 the Greek authorities granted the fugitives from
Mitzela permission to found a new settlement, which they called Nea Mitzela,
also known as Amaliopolis.
During Ottoman rule, the Aegean islands were relatively independent. It

was on the Greek islands that a Greek merchant navy developed
(Sphyroeras 1985, 9-21). Well-known Greek merchant vessels, like perama,
trechandiri and bratsea, can be traced as far back as the 1Thcentury. These
vessels originally weighed 3 tons. Later the maximum capacity ranged
between 30 and 40 tons (Damianidis & Zivas, 1986), but larger vessels were
also built. A network of maritime activity linked the individual islands, but the
islanders also sailed to Smyrna and Constantinople and traded with
European cities (Sphyroeras eta/., 1985). The Ottoman control of Thessaly
and the Ottomans' dense network of land routes were not the only reasons
why the sea route between Euboia and the mainland fell into disuse; the
islanders also had larger sea-going vessels, which were not restricted to
sheltered passages and coastal navigation.
The importance of the Pagasitikos and Maliakos Gulfs in maritime history

is reflected in the number of ports that were to be found along their coasts
through the ages. Phalara, Halos, Pyrasos, Pagasai and Demetrias in
classical and Hellenistic periods; Phalara, Demetrion, Thebai and Demetrias
in Roman and Byzantine times; Bondenitsa, Fteleo, Almiros and Dimitriada
in the 12'h-14'hcenturies AD. After the Ottoman conquest of Thessaly in 1393,
all the ports along the Maliakos and Pagasitikos Gulfs fell into decay and lost
their power to the Turkish port of Volos (Golos), the only port in Thessaly
after the conquest of Negroponte and the deportation of the inhabitants of
Fteleo to Constantinople in 1470.
Land and sea routes
Maritime sites along the Pagasitikos Gulf played an important role in

maritime history as ports of call for Thessaly, but what do we actually know
about land and sea routes? Nowadays there are many land routes covered
by buses and - less common - trains in Greece. Airplanes connect the
capital Athens with remote parts of the country and a dense network of ferry
boats links the major ports on the mainland with the islands in the Aegean
and lonian seas. Most of these connections are related to the tourist
industry, but wooden merchant vessels still operate between the harbour of
Volos and the Northern Sporadhes: Skiathos, Skopelos and Alonnisos.
Until 1881, Thessaly and northern Greece formed part of the Ottoman
Empire. The coastal section between Thermopylai and Demetrias had been
split into two parts, divided between Greece and the Ottoman Empire.
Between 1834 and 1881 the border ran from the Gulf of Arta in the west to
the Pagasitikos Gulf in the east. Volos and Almiros were in Turkish hands
while the northern shore of the Maliakos Gulf up to the peninsula of
Amaliopolis formed part of the Greek kingdom. In the early 20thcentury, after
the liberation of Thessaly, it was quite unusual to travel from Athens to
Thessaly across land. From the schedules published by Baedeker (1910) we
know that boats departed from Piraeus to Volos every day. The steamers of
the early 20mcentury followed the route between the mainland and the island
of via and halted at Lavrion, Aliveri, Chalkis, Limni, Atalanti, Edipsos, Stilis,
Orei and other places (table 3 and figure 14). Baedeker (1910, 213-217)
describes this route as excursion no. 16, D'Athenes a Volo par mer, and
provides detailed information on coasts, harbours and antiquities,
continuing as it were in the tradition of the periploi of classical and Hellenistic
times.
Compared with travelling by train, a voyage by boat was relatively cheap

and comfortable in those days. When the archaeologists Wace and
Thompson excavated at Magoula Zerelia in the Almiros Plain they went by
boat from Piraeus to Volos. We left the Piraeus by the steamer 'Ares' at 8 P.M.
on Wed. June Pd 1908 and reached Volo on June 3&at 2 PM. The voyage was
without incident- the steward provided an excellent lunch, and Droop threw
over board Ruskin's 'Art of Drawing'. After they had visited the ephoros in the
Museum of Volos, they travelled overland to Almiros. The steamer
occasionally stopped at Nea Mitzela and Tsingeli, on the shore of Sourpi
Bay, close to the site of Old Halos (Baedeker 1910, 214; Reinders 1988,
161). From Piraeus the steamer followed the route between Boeotia and
Euboea through the narrow passage at Chalkis and from there to the
Maliakos Gulf and the Strait of Trikeri into the Pagasitikos Gulf. The voyage
from Piraeus to Volos took about 18 hours.
From the periplous of 'Skylax' and Strabon's description of the coast

between Piraeus and Demetrias we may already infer that ships in classical
and Hellenistic times followed the same sea route as those sailing in the
early 20th century. This is confirmed by three other written sources.
Herodotos (7.173) refers to the disembarkation of 10,000 men near Halos in
480 BC, while Demosthenes (19.163) mentions that in 346 BC the Athenian
embassy sailed to Halos via the Euripos. In 302 BC Demetrios Poliorketes
gathered his fleet and his troops at Chalkis, sailed along the coast and
landed near Larisa Kremaste (Diodoros 20.1 10.2).
At first glance the sea route between Piraeus and Demetrias offered
sailors many landmarks for orientation and involved only few problems. The
19th-century sea maps of the British Admiralty show an abundance of
protruding capes, towns and conspicuous trees for orientation. Near the
shore were shallows and submarine rocks, but long stretches of the coast
presented no problems, with only one exception. On their route to the north,
the sailors had to pass an obstacle: the Euripos, a narrow channel between
Chalkis and the mainland. In the Euripos a strong current changes direction
at irregular intervals, a phenomenon already described by Strabon (9.2.8;
quotations from Strabon's Geography have been taken from the edition of
the Loeb Classical Library, translated by H.L. Jones): Concerning the Euripus
it is enough to say only thus much, that they are said to change seven times
each day and night.
Under normal conditions the route from Chalkis northwards involved no

problems, but the channel between Lokris, Malis and Euboea is bordered by
two active tectonic faults. Earthquakes occur in this area. We know that
earthquakes occurred here in classical and Hellenistic times, too. In 426 BC
the town of Phalara was destroyed by a 'tsounami', a sea wave: and as for
Echinus and Phalara and Heracleia in Trachis, not only was a considerable
portion of them thrown down, but the settlement of Phalara was overturned,
ground and all. In Atalanta near Euboea a trireme was liffed out of the docks
and cast over the wall, This information Strabon (1.3.20) derived from
Demetrios of Kallatis, who wrote a treatise on earthquakes.
These circumstances were of course exceptional. In 480 BC the Greek

army, after landing at Old Halos, headed for Tempe overland, and the
Athenian embassy with Demosthenes, on mission to Pella, sailed without
problems to the Pagasitikos Gulf, from where it travelled overland to its
destination. The embassy had reason to travel partly overland instead of
going all the way to Tempe and Macedonia by ship: the sea route was safe
for only part of the year. The whole voyage along the coast of Pelion is rough,
a distance of about eighty stadia; and that along the coast of Ossa is equally
long and rough (Strabon 9.5.22). With eastern winds, the route along the
east coast of the Pelion and Ossa was particularly dangerous. The voyage
along the east coast of Pelion and Ossa was however considerably longer
than Strabon suggested. In 480 BC Xerxes' Persian fleet was lying in wait (on
a small beach between Kasthania and Cape Sepias) when, a violent east
wind bursting forth, some of the ships were immediately driven high and dry
on the beach and broken to pieces on the spot (Strabon 9.5.22). Another
obstacle on the route to the north along the east coast of Magnesia is a rock
between the mainland and the island of Skiathos. The Persians left a stone
column on the rock, called Myrmex, as a sign for the other vessels
(Herodotos 7.183).
On account of storms and poor visibility in the winter, the sailing season
in classical times was generally restricted to the summer; the season lasted
from 10 March to 10 November at best (Casson 1971, 270; Vegetius, mil.
4.39). The sailors had to rely on landmarks and especially with poor visibility
the route between the island of Euboea and the mainland was too
dangerous. In wintertime the seas were 'closed' (Vegetius, mil. 4.39). Even
nowadays sailing along the east coasts of Euboea and Magnisia in
wintertime is not without danger. In 1996, on the 2gth of December, the
Dystos, carrying a cargo of cement, was hit by two large waves on the
broadside and capsized off Kimi on the east coast of Euboea.
In early and mid-Byzantine times the route to Thessaly between Euboea

and the mainland was still used for the transportation of Thessalian products
to Venice, Pisa, Genoa and other maritime cities (Koder & Hild 1976, 101-
102). In the 12Ihcentury a city like Almiros also had trade connections with
Saloniki and Constantinople. The written portolans, although dating from a
later period, show that the route to Saloniki ran from Zeitoun to Gardika, Aios
Yeorgios Zagora, along the east coast of Magnesia, via Seta (Cape Pourion),
Verliqui and Platamonas to Thessaloniki (Delatte 1947, 226). Koder & Hild
believe that there were also connections between the ports of the
Pagasitikos Gulf and the northern Sporadhes, as suggested by the wreck of
a Byzantine vessel that was found near the island of Alonnisos in 1970. The
cargo of this vessel, consisting of over 1500 plates and amphoras, yielded a
mid-12th-centurydate for the wreck (Asimakopoulou-Atzaka 1982, 168-169;
Kritzas, 1971). The glazed plates were beautifully decorated with birds,
animals and geometrical and floral patterns. The same kind of pottery was
found during surveys in the Almiros Plain (Dijkstra et a/. 1997). According to
Throckmorton (1971), the vessel 'resembles a modern caique rather than a
tenon-fastened Roman ship'.
Strabon's Geography, the periploi and the portolans give us a sailor's

view of the coastal area between Thermopylai and Demetrias. From the 4m
century BC until AD 1393 Thessaly, especially the districts of Malis and
Achaia Phthiotis, had a maritime orientation. Thessaly was reached overseas
from the south via a great number of ports of call along the coasts of the
Maliakos and Pagasitikos Gulfs. After the Ottoman conquest the old ports
disappeared completely; their role was taken over by a dense network of
land routes, khans and annual markets (Rizos, 1996).
The network of land roads, cobbled kalderimis and bridges covered the
whole of northern Greece and the Balkans. European travellers in the 19Ih
century made use of these land roads and mentioned long caravans along
their routes. At regular intervals were khans with overnight facilities for
human and animal travellers. Small sites called 'to chani' are still to be found
here and there today, like the khan at the Fourka pass north of Lamia and
that at the pass near Bralos. There was also a khan 'just west of Kefalosi
spring', as we know from Wace's diary, but it lay in ruins by the time Wace
and Thompson carried out their excavations in the area east of Platanos.
In Turkish times goods were transported by ox-drawn carts, by pack

animals and also by camels, to be traded at annual fairs (Ritzos, 1996).
These annual fairs more or less took over the function of the ports of call in
previous periods. Although no documentation is currently available on land
roads, we may assume that there were land roads before the Ottoman
conquest between the various districts of Thessaly and a nearby harbour, or
between an inland town and its harbour, as for instance in the case of Pherai-
Pagasai, Thebai-Demetrion and Lamia-Phalara. These land roads will
however not have constituted an interrelated dense network. The Ottomans
provided the infrastructure for this network and Christian communities
enhanced it by building bridges (Rizos, 1996). The network was not
restricted to Thessaly, but also connected Thessaly with central Greece,
Macedonia, Constantinople, Anatolia and central Europe.
We know almost nothing about the role of the harbour of Volos and the
sea route between Euboea and the mainland during Ottoman rule. Two
Turkish portolan charts in the Khalili collection (Soucek, 1996) show
accurate, detailed geographical representations of the coast of Volos (Koloz)
and the island of Euboea (Igriboz). Both charts, which are in the tradition of
Piri Reis, indicate that the Gulf of Volos and the passage along Negroponte
were important for the Ottomans, at least for the Ottoman navy, for which Piri
Reis' Book of the sea was intended. A portolan chart of the Aegean by
Mehmed Reis, dated 1590, shows the same attention to detail for the whole
Aegean (Biadene 1990, 94-95). The dates of the first two portolan charts are
not known with any certainty, but the Euboea chart closely resembles the
chart of Euboea from the isolario of Bartolomeo dalli Sonetti (1485) and the
maps of Benedetto Bordone (1528); the Turkish portolan charts are
obviously based on the Italian printed atlases of the Aegean Islands
(Sphyroeras et al. 1985, 28).
West European sailors never used the old sea route to Thessaly through
the Euripos. Dutch trade in the Mediterranean began to flourish after 1590.
A passage in Van Keulen's Zeefakkel shows that it was simply impossible for
large vessels to pass the Euripos; it was only just wide enough for a galley
without oars: 'maer van 't Kasteel tot de stad is een windbrug om op te halen
als 'er kleyne scheepen en Galeyen door halen, want t'is er niet wy'er, als dat
er een galey zonder riemen deur kan' (Van Keulen, 1716). West European
sailors occasionally called at the port of Volos. Van Keulen describes the sea
route from Makronisos, the island east of Cape Sounion, via the islands of
Skiros, Skopelos and Skiathos to the entrance of the Pagasitikos Gulf (table
4 and fig. 15). He describes the anchorages in the Gulf of Volos, among
which were anchorages near the village of Seigne (Fteleo) in Pteleos Bay, in
a bay near Moordenaars Eyland and between the island of Trikeri and the
mainland. Although Van Keulen's map is not accurate, Moordenaars Eyland
is without doubt Ayos Nikolaos Island, which is indicated on the portolan
charts as Cape San Nicolo.
What had West European sailors to look out for in this part of the
Aegean? Van Keulen provides one warning: in describing an anchorage off
the west coast of Skopelos, in Panormos Bay, he mentions that although this
anchorage is good, he nevertheless advises sailors to anchor behind Dhasia
Islands, two small islands north of Panormos Bay. He explains that it is
difficult to leave Panormos Bay with western winds and warns sailors against
the risk of the sudden appearance of Turkish galleys in search of vessels
loading grain without permission: om niet beset te worden van de Turcksche
Galeyen die de scheepen hier kooren ladende somtijds komen betrappen.
From this we may infer that the Turkish tsiflikia in the Almiros Plain produced
a surplus of grain and that it was worthwhile to attempt to export the grain
without obtaining the necessary Turkish permit.
Summary
The section of the coast between Thermopylai and Demetrias formed

part of the sea route from Piraeus to Thessalonike. In classical, Hellenistic
and Byzantine times there were many ports along the shores of the Maliakos
and Pagasitikos Gulfs. Written sources inform us about the sea route from
Piraeus via the Euripos channel to these ports, which served as ports of call
for the large Thessalian plains. They were also important as intermediate
stations for voyages and troop movements to Tempe and Macedonia,
because the sea route from Demetrias to Thessalonike was dangerous. The
coast between Cape Sepias and the river Peneios was rough; only few
beaches provided shelter in the case of gales.
During a long period, spanning seventeen centuries, the predominance

of maritime activity shifted from one port to another, but the ports of call
along the Maliakos and Pagasitikos Gulfs retained their important positions
until Thessaly, Lokris and Euboea came under Ottoman control in the 15m
century. With the exception of Volos, almost all the ports lost their former
importance. After 1393, when Thessaly came under Ottoman rule, a dense
network of land roads connected Thessaly with central Greece and, via
northern Greece, with the Balkans and Constantinople.
Acknowledgements
I am indebted to Prof. Dr. S.L. Radt and Prof. Dr. H.T. Waterbolk for
reading the manuscript and providing useful suggestions. I would also like
to thank Susan Mellor for correcting the English text and Hans Zwier for
preparing the drawings.
Reinder Reinders
Groningen Institute of Archaeology
BAl Poststraat 6
9712 ER Groningen
The Netherlands
REFERENCES
Adler, M.N., 1907. The Itinerary of Benjamin of Tudela. London.

Asimakopoulou-Atzaka, P., 1982. Palaiochristianiki kai Bizandini Magnisia.
In: G. Chourmouziadis (ed.) , Magnisia, to chroniko enos politismou,
105-175.
Avramea, A.P., 1974.1Bizandini Thessalia mechri tou 1204. Athina.

Bakker, A. & R. Reinders, 1996. Tetrobolen en tetradrachmen. Een
muntvondst in Nieuw Halos. Paleo-aktueel7, 66-69.
Biadene, S. (ed.), 1990. Carte da navigar. Portolani e carte del Museo Correr
1318-1732. Marsilio Editori, Venezia.
Bottema, S., 1988. A Reconstruction of the Halos Environment on the Basis
of Palynological Information. In: New Halos, a Hellenistic Town in
Thessalia, Greece, by H.R. Reinders, 216-226.
Casson, L., 1971. Ships and Seamanship in the Ancient World. Princeton
University Press, Princeton & New Jersey.
Damianidis, K., & A. Zivas, 1986. To trechandiri stin Elliniki navpiyiki techni.
Eommex, Athina.
Delatte, A., 1947. Les Portulans Grecs (= Bibliotheque de la Faculte de
Philosophie et Lettres de I'Universite de Liege, Fascicule 107). Liege.
Dijkstra, Y., H.R. Reinders, V. Rondiri & Z. Malakasioti, 1997. Van
Duivelsberg tot Rode Rots. Paleo-aMueel8 (in press).
Fotini-Papakonstantinou, M., 1994. To notio kai to ditiko tmima tis Achaias
Phthiotidas apo tous klasikous mechri tous romaikous chronous. In
R. Misdrahi-Kapon (ed.), La Thessalie, Quinze annees de recherches
archeologiques, 1975-1990, Bilans et perspectives 2, 229-238.
Athenes.
Fotini-Papakonstantinou, M., 1994. The Kastro of Lamia. Archaeological
Receipts Fund, Athens.
Furtwangler, A.E., 1990. Demetrias, eine Makedonische Grundung im Netz
Hellenistischer Handels und Geldpolitik. Habilitationsschrift,
Saarbrucken.
Hope Simpson, R. & J.F. Lazenby, 1970. The Catalogue of the Ships in
Homer's Iliad. Clarendon Press, Oxford.
Intzesiloglou, B.F., 1994. lstoriki topografia tis periochis tou Kolpou tou
Volou. In: R. Misdrahi-Kapon (ed.), La Thessalie, Quinze annees de
recherches archeologiques, 1975-1990, Bilans et perspectives 2, 3 1-
56. Athenes.
Keulen, J.G. van, 1728. De nieuwe groote ligtende Zee-fakkel, deel 3.
Amsterdam.
Koder, J. & F. Hild, 1976. Hellas und Thessalia (~sterreichischeAkademie
der Wissenschaften, Philosophisch-HistorischeKlasse Denkschriften,
Band 125). Wien.
Kallianos, K., 1984. 0 katalogos ton Thessalomagniton paroikon tis
Skopelou sta 1829. Thessaliko lmeroloyio 7, 3 1-38.
Kraft, J.C., G. Rapp, G.J. Szemler, C. Tziavos & E.W. Kase, 1987. The Pass
at Thermopylae, Greece. Journal of Field Archaeology 14, 181- 198.
Kretschmer, K., 1909. Die italienischen Portolane des Mittelalters. Ein Beitrag
zur Geschichte der Kartographie und Nautik. (= Veroffentlichungen
des lnstituts fur Meereskunde und des Geographischen lnstituts an
der Universitat Berlin, Heft 13). Mittler und Sohn, Berlin.
Kritzas, Ch., 1971. To vizandinon navayion Pelagonnisou-Alonnisou. AAA
4.2, 176-182.
Makris, K.A., 1982. Metavizandini kai Neoteri Magnisia. In: Magnisia, to
Chroniko enos Politismou. Athina.
Malakasioti, Z. V. Rondiri & R. Reinders, 1995. Groninger bijdrage aan
Griekse monumentenzorg. Paleoaktueel6, 71-74.
Marzolff, P., 1980. Demetrias und seine halbinsel (= Demetrias Ill). Bonn.
Marzolff, P., 1994. Antike Stadtebau und Architektur in Thessalien. In: R.
Misdrahi-Kapon (ed.), La Thessalie, Quinze annees de recherches
archeologiques, 1975-1990, Bilans et perspectives 2, 255-276.
Athenes.
Mitropoulos, D., & S. Michalidis, 1988. Seismic stratigraphy and structure of
Pagasitikos and Maliakos Gulf and the surrounding areas, Aegean
Sea, Greece. Rapp. Comm. int. Mer Medit. 31.2.
Muller, K., 1965. Geographi Graeci Minores (Volume 1; reprint of the edition
of 1855). Georg Olms, Hildesheim.
Nikonanos, N., 1971. Aidinion. AD (Chronika), 312-313.
Nordenskiold, A.E., 1897. Periplus, an essay on the early history of charts and
sailing directions. Stockholm.
Pantos, P., 1994. La Vallee du Spercheios - Lamia exceptee - aux epoques
hellenistique et romaine. Quinze annees de recherches, 1975-1990.
In: R. Misdrahi-Kapon (ed.), La Thessalie, Quinze annees de
recherches archeologiques, 1975-1990, Bilans et perspectives 2,22 1-
228. Athenes.
Perissoratis, C., P. Zacharaki & A. Andrinopoulos, 1988. Texture and
composition of the bottom sediments of Pagasitikos Gulf and Trikeri
Strait, Thessaly (Greece). Rapp. Comm. int. Mer Medit. 31.2.
Radt, S.L., 1994. Aus der Arbeit an der neuen Strabonausgabe. Pharos 2,31-
35.
Reinders, H.R., 1989. New Halos, a Hellenistic town in Thessalia, Greece.
Utrecht.
Reinders, H.R., 1993. 1 topothesia tis Alou. In: V. Kondonatsios (ed.)
Achaiophthiotika 1, 49-59. Almiros.
Reinders, R. (in prep). Earthquakes in the Almiros Plain and the
abandonment of New Halos (Stuttgart Congress paper, May 1996).
Rizos, A., 1996. Wirtschaft, Siedlung und Gesellschaft in Thessalie im
ubergang von der Byzantinisch-Frankischen zur Osmanischen
Epoche. Hakkert, Amsterdam.

Sawidis, G.K., 1993. 1 emporiki akmi tou Almirou kata to 120 aiona m. Chr.
In: B. Kontonatsios (ed.), Achaiophthiotika 1, 203-211. Almiros.
Sherwin-White, S.M., 1978. Ancient Cos. An historical study from the Dorian
settlement to the Imperial period (= Hypomnemata, Untersuchungen
zur Antike und zu ihrem Nachleben, Heft 51). Vandenhoeck &
Ruprecht, Gottingen.
Soucek, S., 1996. Piri Reis & Turkish mapmaking after Columbus. The Khalili
Portolan Atlas. Nour Foundation, Azimuth Editions, Oxford University
Press, London.
Sphyroeras, V., A. Avramea & S. Asdrahas, 1985. Maps and map-makers of
the Aegean. Olkos, Athens.
Stahlin, F., 1924. Das Hellenische Thessalien. Stuttgart.
Straaten, L.M.J.U. van, 1988. Mollusc Shell Assemblages in Core Samples
from Ancient Halos. In: New Halos, a Hellenistic Town in Thessalia,
Greece, by H.R. Reinders, 227-235.
Throckmorton, P., 1971. Exploration of a Byzantine wreck at Pelagos Island
near Alonnessos. AAA 4.2, 183-185.
Zacharakis, Ch. G., 1992. A catalogue of printed maps of Greece 1477-1800.
Samourkas Foundation, Athens.
Zangger E., 1991. Prehistoric Coastal Environments in Greece: The
Vanished Landscapes of Dimini Bay and Lake Lerna. Journal of Field
Archaeology 18, 1- 15.
CAPTIONS TO FIGURES
Figure 1. Provenance of the coins found at the site of New Halos.

Figure 2. Provenance of the coins found at New Halos and Demetrias (data
for Demetrias after Furtwangler 1990).
Figure 3. The coast between Peiraieus and Thessalonike, with distances
indicated in stadia.
Figure 4. Towns and landmarks along the coast between Thermopylai and
Demetrias in Classical and Hellenistic times.
Figure 5. Maliakos Gulf. The shifting of the coastline near the outlet of the
Spercheios (after Kraft et a/. , 1987).
Figure 6. Volos Bay. The shifting of the coastline north of Demetrias (after
Zangger, 1991).
Figure 7. Sourpi Bay. Salt marsh between Old and New Halos.
Figure 8. Thessalian ports in Classical times.
Figure 9. Early Hellenistic ports.
Figure 10. Ports in Late Hellenistic/Roman times.
Figure 11. Early Byzantine ports.
Figure 12. Ports, towns and landmarks between Talandi (Atalandi) and
Monester (Mount Pilion) on medieval portolan charts.
Figure 13. Gulf of Volos (Roux & Allezard 1800; reprint, Maritime Museum of
Greece, 1981).
Figure 14. Sea route between Piraeus and Volos (Baedeker, 1910).
Figure 15. Sea routes from Macronisi (Makronisos) to Negroponte (Chalkis)
and Volo (Volos), according to Van Keulen (1728).
Appendix 1
Strabon's description of the coast between Thermopylai and Demetrias

(quotations have been taken from the edition of the Loeb Classical Library,
translated by H.L. Jones).
Thermopylae, then, is separated from Cenaeum by a strait seventy

stadia wide; but, to one sailing along the coast beyond Pylae, it is about ten
stadia from the Spercheius; and thence to Phalara twenty stadia; and above
Phalara, fifty stadia from the sea, is situated the city of the Lamians ...
( 9.5.13). The Spercheius is about thirty stadia from Lamia, which is situated
above a certain plain that extends down to the Maliac Gulf ( 9.5.9).
... and then next, after sailing a hundred stadia [from Phalara] along the
coast, one comes to Echinus, which is situated above the sea; and in the
interior from the next stretch of coast, twenty stadia distant from it, is Larisa
CremastQ(it is also called Larisa Pelasgia) (9.5.13).
Then one comes to Myonnesus, a small island; and then to Antron ...
Near Antron, in the Euboean strait, is a submarine reef called "Ass of Antron"
( 9.5.14).
... and then one comes to Pteleum and Halus (Strabon 9.5.14). And
Artemidorus places Halus on the seaboard as situated outside the Maliac
Gulf, indeed, but as belonging to Phthiotis; for proceeding thence in the
direction of the Peneius, he places Pteleum after Antron, and then Halus at
a distance of one hundred and ten stadia from Pteleum (9.5.8).
The Phthiotic Halus is situated below the end of Othrys, a mountain

situated to the north of Phthiotis, bordering on Mount Typhrestus and the
country of the Dolopians and extending from there to the region of the Maliac
Gulf. ... Halus is called both Phthiotic and Achaean Halus, and it borders on
the country of the Malians, as do also the spurs of Othrys Mountain ... Halus
(either feminine or masculine, for the name is used in both genders) is about
sixty stadia distant from Itonus... It is situated above the Crocian Plain; and
the Amphrysus River flows close to its walls (9.5.8).
... and then to the temple of Demeter; and to Pyrasus, which has been
rased to the ground.... Pyrasus was a city with a good harbour; at a distance
of two stadia it had a sacred precinct and a holy temple, and was twenty
stadia distant from Thebes (9.5.14).
Below the Crocian plain lies Phthiotic Thebes (Strabon 9.5.8). Thebes is
situated above Pyrasus, but the Crocian Plain is situated in the interior back
of Thebes near the end of Othrys; and it is through this plain that the
Amphrysus flows ... Now Phylace is near Phthiotic Thebes ... it is about one
hundred stadia distant from Thebes, and it is midway between Pharsalus
and the Phthiotae (9.5.8).
... and then to Cape Pyrrha, and to two isles near it, one of which is called
Pyrrha an the other Deucalion. And it is somewhere here that Phthiotis ends
(9.5.14).
The sea-port of Pherae is Pagasae, which is ninety stadia distant from

Pherae and twenty from lolcus... Now Pherae is at the end of the Pelasgian
Plains on the side towards Magnesia; and these plains extend as far as
Pelion, one hundred and sixty stadia (9.5.15).
Demetrias, which is on the sea between Nelia and Pagasae, was

founded by Demetrios Poliorcetes, who named it after himself ...
Furthermore, for a long time this was both a naval station and a royal
residence for the kings of the Macedonians (9.5.15). lolcus is situated above
the sea seven stadia from Demetrias.
Table 1
Distances along the coast between Piraeus and Thessalonike, in stadia

(1 stadium is about 185m, depending on the foot used).
from to distance reference
Piraeus Sounion 330 Strabon 9.1.2

Sounion Euripos 670 Strabon 9.2.8
Euripos Peneios 2,350 Strabon 9.5.22
Euripos Thermopylai 530 Strabon 9.4.17
Spercheios Demetrias 800 Strabon 9.5.22
Demetrias Peneios 1,000 Strabon 9.5.22
Peneios Thessalonike 660 Strabon 8.8.5
Table 3
Weekly Piraeus-Volos schedule of the John MacDowall & Barbour

company: 'Tson', Athens (Baedeker, 1910).
day time harbour day time
Monday 19.00 Piraeus Friday 17.00

Monday 22.30 Lavrion Friday 13.00
Tuesday 06.00 Chalkis Friday 02.00
Tuesday 09.30 Limni Thursday 22.30
Tuesday 12.00 Stilis Thursday 19.00
Tuesday 20.00 Volos Thursday 12.00
Prices (without food; first and second class): Pireas-Lavrion, 6 dr., 4 dr.;
Piraeus-Chalkis, 8 dr., 5 dr; Piraeus-Volos, 15 dr., 10 dr.
A
03 Table 4
0,
Sea route from Macronisi (Makronisos) to Negroponte (Chalkis) and Volo (Volos), according to Van Keulen (1728).
from to course distance
Macronisi (Makronisos) Asturi Isl. (Petali Isl.) northeast 4-5 miles

Asturi Isl. Suhi de Basa (Stira, KavalianJ northwest 3 miles
C. Negro (C. AlivBri) Volie northwest 3-4 miles
Volie Strait S. Marco (Euripos) west 1 mile
Strait S. Marco Negroponte (Chalkis)
C. Martelo (C. Mandili) Schiro (Skiros) north I 1 miles

P. S. Georgio de Schira Schoppelo (Skopelos) northwest 6 miles
Schoppelo (north point) Sciatta (Skiathos) west 2 miles
Sciatta (southcoast) Gulf of Volos west-southwest 3-4 miles
THRAKE
MAKEDONIA
THESSALIA
LARISA
DEMETRIAS
ACHAIA PHTHlOTlS
HALOS
MALIS - LOKRIS
EUBOIA
PHOKIS-PELOPONNESOS
AIGYPTOS
Fig. 2
487
REINDER REINDERS TROPIS V1
Fig. 15
OLBIA-SARDINIA WRECK OF THE SICILIAN0 (fig.1)
In the autumn of 1995 the Soprintendenza Archeologica per le provincie

di Sassari e Nuoro, directed by Dr. Fulvia Lo Schiavo, set in motion a season
of excavations on the remains of a boat which may be dated between the
end of the Pdand the beginning of the Tdcentury AD. It lies in just over two
metres of water in a wide sand pool lying between the rocks of the shore and
a small reef to seaward. It was probably the impact against the reef caused
by the northerly winds that was the cause of the sinking. The wreck was
reported in 1990 by the Centro Sub Tavolara and a preliminary survey
suggested it was the remains of an old boat left to rot in a place where it
could not foul access to the nearby landing place of Cala Finanza. On the
other hand the results of the 1995 investigation have given rise to the
supposition that there was a violent episode caused by the northerly winds
which drove the boat onto the reef. Amongst the rocks of the reef and all
around there were many materials which presumably belonged tothe wreck
and these were mixed with materials of different dates which suggests other
disasters. The work was carried out by the staff of the Centro Sub Tavolara
and led by Dr. Rubens D'Oriano of the Soprintendenza and by the present
writer. Students and tourists were present during operations as observers.
Of the wreck there remain two main pieces and a few scattered fragments
(fig.2). The first of these consists of the lower part of the bottom of the aft
section (photo 1-2). The second is part of the side, presumably the right side,
and under this second part there are some unidentifiable pieces of wood,
which probably belong to the upper works (photo 3). The remains derive
from a little less than half the craft, and the remainder must have been
destroyed and carried away by the movements of the sea since a detailed
search amongst the stones where the prow's remains should have been
have yielded no trace of the break-up. The results of the enquiry are under
study for final publication and this paper is confined to a series of
observations on the construction details. At the moment we may suppose
that the length of the boat was from 15 to 18 metres, that it had been very
carefully built, but that at the moment of the wrecking it must have been very
old and badly kept. Considerable repair-work is evident and its overall
condition is very poor.
EDOARDO RlCCARDl TROPlS VI
SIDE
There is one fragment of this, presumably from the right side of the boat,
and it lies a little more than one metre from the remains of the bottom. It is
concave to a remarkable degree. Few fragments of the inside planking
remain, and they are planks of from 12 to 19cm in width and 3cm thick,
apparently joined only by iron nails. The remaining frames are incomplete
with two of them obtained from trunks that had been only roughly hewn,
while eight were made with care. The planking bears the marks of three
frames, now gone. The lockpins between the frames and the planking have
a diameter of 18mm and some are pierced by copper nails. The frames are
numbered OR1 to OR10. The planking consists of the remains of 17 planks
of from 10 to 13cm in width and about 3cm in thickness. They are numbered
t l to t17. The mortises are about 7cm apart, rectangular and 7.5cm wide with
a thickness of 7mm. The tenons are rectangular and fit the mortises
precisely. The lockpins are 8mm in diameter. On a level with frames OR7 to
OR10, plank t3 displays a simple dovetail joint into which has been inserted
a piece of planking of triangular shape, 1lOcm long. Two pieces of lead
sheeting are nailed into the top of the triangle: this could be a normal joint
made during construction which had to be made watertight afterwards but it
could be that part of the plank has been replaced (photo 4). The outer part
of the craft has evident remains of a covering of watertight resins while
inside, the pitch is preserved only underneath the frames: it has been worn
away and abraded in the space and room. This is another symptom of old
age and poor maintenance of the boat.
BOTTOM
There remains the keel, 11 fragments of frames numbered from 0 1 to

0 11, four incomplete planks from the right side numbered TS1 to TS4 and
eight incomplete planks from the left side numbered TD1 to TD8.
The keel is preserved for a length of 5.65 metres and is made of two
beams of the same thickness of 21-22cm where they join (the lower surface
is in a very poor state), and it has varying widths: the aft section is 17cm at
the top and 14 at the bottom. The central part is a constant 13.5cm. At the
extreme aft end the keel is 16cm wide at the top and 13cm at the bottom. At
the rotted end of the centre of the boat it is 15cm at the top and 13cm at the
bottom. The aft part of the beam exhibits a deep rabbet for the insertion of
the barboard and it is curved upwards at the point of attachment of the stern
post. The central beam is straight and without rabbets. The two woods are
OLBIA-SARDINIA WRECK OF THE SICILIAN0
joined together by a complex toothed scarf joint with a key and nail
reinforcement (fig. 3). At this point their thickness exceeds 4cm afhvards and
l c m in the centre. To obtain a good view of the joint it was necessary to
remove two pieces of the planking which were put back after the
investigation without having suffered any sort of trauma (photo 5-6-7). The
most similar example of a scarf functioning this way is in the wreck of Jassi
Ada from the Thcentury AD (fig. 4-5-6-7-8).
Frames: they are in a very poor state of preservation, incomplete and

worn in particular on their upper surface. Apparently only floor timber 0 4 is
joined to the keel by an iron pin, which crosses the joint. Inside the wreck
there are a number of thin stray pieces of woods that are the same width as
the frames and which exhibit holes of the same diameter as the lockpins
(18mm): these are thicknesses originally applied between the frames and
the planking.
Planking: on the right side, the barboard, which is 9cm thick on the keel
side and 7cm on the side of the second strakes, is fitted in a rabbet in the
keel which runs aftward from the joint, whilst the rabbet is not present
towards the prow. The maximum width is 12cm. The distance between the
centres of the tenon lockpins is a constant 16cm. The barboard is in a very
bad state of preservation between frames 0 4 and 0 9 and it shows signs of
repair, and because of this a further investigation was considered necessary.
The second strake is 5cm thick and measures 13.5cm at its point of
maximum width. The distance between the centres of the lockpins varies
from 14 to 16cm. The mortises are rectangular and from 6.5 to 7.5cm wide,
and at least one of them was commenced with a drill hole. The other two
planks have a thickness of about 3.5cm. The distance between the centre of
the lockpins varies from 13 to 18cm. Some mortises were begun with drill
holes and the lockpins are double in at last two cases. The tenon lockpins
were all fitted from the inside and have a variable diameter of from 7 to 8mm.
From the extreme aft end up to beyond the keel joint there is an external
reinforcing plank which covers the surfaces of the keel and, in part, the
barboards. This plank is situated to the rear of the point of insertion of the
sternpost and is probably a repair. It covers a thick layer of resin. It measures
2cm thick to aft and 2.5cm at the joint. The width varies from 14 to 22cm and
it is fixed only to the keel by iron nails with a distance of about 60cm between
them. There is no sign of reinforcement like this at the centre of the ship.
Inasmuch as comparisons are possible the left side is a mirror of the right.
Of particular interest is a repair, which had the purpose of making the

EDOARDO RlCCARDl TROPIS VI
barboard watertight at the point where it joins with the keel on the part under
the frames numbered 0 4 to 0 9 (photo 8). Along the whole of this length
there is visible a strip of caulker made of vegetable fibres (note 1) which was
inserted from the inside of the craft between the two woods of the barboard
and the keel (fig 61 photo 9). Some fragments of barboard at a point
corresponding to frame 0 6 were removed to allow an improved view of the
details. One tenon was completely exposed to view so as to ascertain
whether it had been inserted from the outside by cutting a mortise which
passed through the side of the planks or not. The tenon is of normal length
and the mortise into which it is inserted is blind. A careful check was carried
out on the outer surface of the barboard and traces of iron nailing applied
from the outside were found. A careful check was carried out on the joints
between the barboard and the second strake and they also turned out to be
normal (photo 10). So it is clear that the barboard was not replaced, but that
attempts were made to make it watertight both with the caulking and with the
nails. The result is a beginning of skeleton first construction even though the
caulking was put into the upper corner of the woods and not into the lower,
and even though we do not know if the frame, which is no longer present,
had been nailed in.
Anatomical investigations of the woods are in note 2.
A suggestion has been made that the wreck should be salvaged and
preserved as an "exploded" exhibit, that is, with its various parts side by side
but not assembled so as to allow ease of observation of the details of the
construction and repairs, which possess features of very great interest.
Dott. Edoardo Riccardi

Archeologia Subacqua Navale Marittima
Via A. Faggi, 13
17042 Bergeggi
Savona, ltalia
Translation: Michael Chamberlain

Drawings: Giovanni Sedda and Antonello Piccinnu
Photo: Egidio Trainito
OLBIA-SARDINIA WRECK OF THE SICILIAN0
NOTES
1. Botanical analysis of the caulking was done by prof. Daniele Arobba. It can be broom fibre
(Spartium junceum) or, but less probably, tow of hemp. Are excluded cotton, hemp, flax and
esparto.
The analysis was done by O.Pignatelli, lab DENDRODATA and G.Giachi, lab RESTAURO
Soprintendenza Archeologica Toscana. The results are:
Keel Ulmus minor
Tenons Quercus ilex and Quercus suber
01 Fraxinus excelsior
02 Olea europea
03 Picea abies ?
07 Acer campestre
010 Ulmus minor
Fraxinus excelsior
Ulmus minor
Fraxinus excelsior
Picea abies ?
Acer campestre
Fraxinus excelsior
Fraxinus excelsior
Ulmus minor
Cupressus sempewirens
Ulmus minor
Frame pins Fraxinus excelsior and Olea europea

Plank pins Fraxinus excelsior and Olea europea
L1 - L2 - L3 Quercus sp.
As far as the planks are concerned we have:

Tdl - Td2 - Td5 - Td6 - Td7 - Tsl - Ts2 Pinus sylvestris
but for the planks Td3 - Td4 - Td8 - Ts3 - Ts4 the two labs gave different results. The two
possibilities are Picea abies and Larix decidua so we need further investigations, which
are forthcoming.
Also the samples of the side planks (tl to t17) are still under investigation.
CAPTIONS
Drawings:
1: Map of the Mediterranean, showing the site of Olbia

2: General view of the wreck
3: Reconstructed axonometry of the keel joint
4: Section of the CH area
5: Section of the OR4 area
8: Points where sections were measured
Photographs:
1: Photocomposition of bottom fragment

2: View of stern
3: General view of the two fragments of the wreck
4: Lead repair-work on side fragment
5: Dismantling planks to examine the keel joint
6: Upper view of the keel joint
7: Side view of the keel
8: Side view of the repair area (left: an iron nail)
9: The caulking in the repair area
10: The repair area after removing a fragment of the garboard
-OLBIA-SARDINIA
- -- - .-- - WRECK OF THE SICILIAN0
_--- --
Fig. 2
1 Fig. 1 -'4---
Fig.
Fig. 5
Fig. 6
- OLBtA-SARDINIA
-- -- - .---WAECK
- -
OF THE SICILIAN0
Fig. 7
Fig. 8
EDOARDO RlCCARDl
-- .- - --- -
Ph. I
-mm-q
I I
I
91
Ph. 2
v-- --- -
'
---I!
>
Ph. 3
502
- - --
. - -- -- OLBIA-SARDINIA
--
, - - --
- -
WRECK-
OF- THE
- - -
SICILIANO
- --
Ph. 4
Ph. 6
II
Ph. 7
503
EDOARDO RlCCARDl TROPE VI
Ph. 9
Ph. 10
584
THE LINEAGE OF THE TRIACONTOR
A verifiable hypothesis
The prototype of a new type of ship is usually of modest size.

Understandably so: investing in building a ship of a new type is a costly
gamble. Success leads not only to a demand for more ships of this type, but
also, and nearly invariable so, to a demand for larger ships of the same type.
It seems that in all ages and all places shipwrights who were confronted with
the demand for larger units of a new type of ship, conserved the shape and
the proportions between the principal dimensions, I, b and h, of the hulls of
the smaller ships, that is, they kept the ratios I:b :h constant when building
larger units. As a consequence, series of geometrically similar hulls of
varying size resulted. It explains why in early treatises on shipbuilding, e.g.
those of Witsen (1691) or Chapman (1762) proportionality rules for
dimensioning a given type of ship could be given. In what follows, ships with
such similar hulls are said to belong to the same lineage. Their hulls may
then be compared with only one size parameter as a variable, for example,
their length I.
Other dimensions, of course, could not be enlarged proportionally,

for example those which had to do with the body size of crew members.
Oars could be enlarged in proportions before reaching a limit which made it
necessary to use either more oars per interscalmium, the repeat distance
between oars, or to place more than one man on each oar. The
interscalmium itself depended on the manner of rowing and not directly on
the size of the ship. In sailing ships, rigs could be enlarged somewhat before
it became necessary to use more masts. Details of the development towards
larger sizes of ships of the same lineage could thus obscure its origin,
although the latter could still be traced by comparing the proportions of the
principal dimensions of the hull. In principle, it is easy to do that, but in
practice the data which is needed is not available.
In oared ships of the same lineage, the number of oarsmen that

could be put on board, rt, should vary as a function of the size of the hull, and
so should the number of oarsmen that was placed in an interscalmium, n,
although the function would be a different one. The idea that a number of
Greek oared fighting ships could have belonged to the same lineage is
explored here by comparing data from the literature on such oared ships
with the predicted mathematical relationship between the number of men
A.W. SLEESWYK TROPlS VI
per interscalmium, ri, and the total number of oarsmen on board, rt.
The reason for deriving such a relationship is simply that it lends

itself to verification, because in a few instances both rj and rt are given for
ancient Greek oared ships. If the derivation, or one of the assumptions on
which it is based, is flawed or wrong, or if it does not apply to the data, it
shows up immediately because the results fail to agree with those predicted
by the relationship between r, and r, If, on the other hand, one finds
reasonable agreement between data gleaned from the literature and the
theoretical predictions over a large range, there is good reason to regard
both the reasoning and the assumptions on which it is based as correct.
The mathematical relationship between ri and r,
The most ancient types of Greek oared fighting ships to be considered

here were called triakonteres or pentekonteres (triacontor and pentecontor in
the following), which, translated literally, means "thirty-fitted" and "fifty-
fitted". It has generally been assumed that the number referred to the total
number of oarsmen on board. After the Thcentury BC, new types of oared
ships were obviously distinguished differently, although they were named
similarly, e.g. trieres [trireme] - "three-fitted" - or tetreres, "four-fitted",
penteres "five-fitted", etc. up to tesserakonteres, "forty-fitted", which was a
ship of monstrous size built for Ptolemy IV Philopator [r. 240-204 BC].
Occasionally, ancient texts give the total number of oars or oarsmen on

board such a ship as well. For example, in the ancient naval inventories
found in Piraeus which date from about the middle of the 4thcentury BC, the
number of oars in use on board the Athenian trireme is given as 170 [Casson
1973:84]. Thucydides refers [2.93.2] to members of trireme crews each
carrying "his oar"; it may safely be deduced that there were 170 oarsmen on
board the trireme in his day, with one man on each oar.
The later convention for naming these types of ships, then, did not refer
to the total number of oarsmen on board, but to something else. In principle,
it could either be the number of oars, or the number of men in an
interscalmium, which for the lower denominations would have been the
same, but not so for the higher ones. If there were many men on a single
large oar in each interscalmium, the appellation might also have referred to
the number of oarsmen on an oar. But if there was more than one
interscalmium it is most unlikely that the number of oarsmen on an oar could

have been a distinguishing characteristic. More probable by far is that "five-
fitted" etc. referred to the number of oarsmen fitted in an interscalmium,
which would have been a measure of the technically important power per
unit length of shipboard. Accordingly, in what follows rj is thus regarded.
The change in the system of naming the different types of ships may be
explained if one- and two-level oared ships of different types and different
lineages were fairly common, and if the larger and later ships belonged to a
few lineages only. The total number of oarsmen, r,, would have been a more
useful distinction for the early ships and the number of men on an
interscalmium, ri, for distinguishing the larger and later ships.
The derivation of the relation between ri and r, with varying size starts with
the assumption that the fraction of the cross-sectional area of the hull, which
is available for seating oarsmen in, is a constant in oared ships of the same
lineage. The assumption is illustrated in Figure 1; the cross-sectional area
available for seating oarsmen stands to the cross-sectional area of the hull
in the same proportion for large (Fig. la) and small (Fig. Ib) ships of the
same lineage. It does not appear at all improbable that ancient shipwrights
used a rule of thumb, which determined that proportionality.
The assumption is supplemented by one in which it is assumed that

each oarsman requires the same constant cross-sectional surface area. In
other words, the number of oarsmen per interscalmium ri varies with size,
grosso modo, as the cross-sectional area, i.e. as b x h, or, equivalently, as
the second power of the linear size, say, as I;:
The fraction of the length of the ship that could have been occupied by
oarsmen was some 70 to 75 percent; it follows from the similarity between
c.
ships of the same lineage that that fraction was in principle a constant, The
number of oarsmen in one file, i.e. the number of interscalmia nj that could
have been placed over the length fi x Iwould then been equal to:
nj =Jx/,
li
in which li is the interscalmium, defined as the heart-to-heart repeat
distance between oars. Consequently, the total number of oarsmen may be
given as:
ri = ni x ri.
A.W. SLEESWYK TROPIS VI
As remarked in the preceding section, the value of the length 1, of the

interscalmium length depends on the mode of rowing, i.e. on whether the
ships are rowed a zenzile (I, = I
,= 1m) or a scaloccio (I, = I ,= 1.25) or with
men on both sides of the oars (I,, = I , = 1.75m) (Sleeswyk 1995). The
number ri being proportional to P , dr, is proportional to I.The implication is
that the number of interscalmia ni wou!d be proportional to drill,, and the
number of oarsmen rt (= r, x n,) to ridrill, (= r?"/l,). The total number of
oarsmen may then conveniently be given as:
rt = R x ri3" x (Iz//,), (1)
where R is, within a certain error margin, equal to the number of oarsmen
in a monoreme of the given lineage, which, of course, is rowed a zenzile.
Verification
From the various assumptions on which Formula (1) is based it is clear

that it cannot possibly be very accurate. The values of both f, the length
fraction of the hull occupied by oarsmen, and the various values of the
, as given in the literature show considerable
interscalmium I,, I, I, and I
scatter. Consequently, applications of Equation (1) in which these values are
regarded as constants produce results which are first approximations at
best.
Nevertheless. there is agreement between the values of r, predicted by

Equation (1) for given ri-values and the values from the literature as
presented by Casson (1973) and Morrison (1995) within relatively narrow
error margins of a 8%, as may be observed in Table I. The adjustable
parameter R must be equal to the number of oarsmen in a monoreme of the
given lineage, which, of course, is pulled a zenzile. For R the value of 32 has
been chosen, which is within the 8% error margin equal to the number of
oarsmen in a triacontor (30); it gives reasonable agreement over the entire
range.
TABLE I
SIZE OF OARCREWS (first approximation)
'i 1 1- 2 2- 3 4 5 6 16 40
rt (literature) 30 - 120 170 280 3-400 360 1600 4000
rt(theory) 32 58 90 126 166 256 286 376 1638 4626
li = Iz Iz Iz Iz Iz Iz Is Is Is Id
The paucity of reliable data renders it impossible to improve much on

this rough-and-ready comparison. Nevertheless, examination of the
available material in some more detail results in a number of corrections.
Refinements
The naval inventories found in Piraeus throw some light on the effect of
the use of an outrigger on the value of the length fraction fi occupied by
oarsmen. The three categories of oarsmen on board the trireme, thalamians,
zygites and thranites, number 54,54 and 62,respectively. Assuming that the
total length of the hull is 40 metres and the value of the interscalmium 1
metre, the thranites, whose oars were supported by the outrigger, occupied
a fraction fi = 0.78of the length of the hull, and the thalamians and zygites
= 0.68.The average number of oarsmen in these three categories is 56z3,
occupying a fraction fj = 0.71 of the length.
It seems probable that outriggers on Greek ships were only in use on

ships larger than the bireme or dieres. About the seating arrangements of
oarsmen on board ships larger than triremes nothing is known with certainty.
Accordingly, one cannot do more than assume that the average length of the
hull occupied by the oarsmen was the same as that of the trireme, and that
in smaller ships without outriggers the length fraction of the hull was the
same as that of the thalamians and zygites in the trireme. The different length
fractions of the hull occupied by oarsmen in the smaller and larger ships
stood then to each other in the ratio 27 : 28'" (= 1 : 1.05),which may be
accounted for by assuming values of R in the same proportion. Thus, if one
assumes R, = 30.5 for ships smaller than the trireme, the following values of
r, result for the 'l', the '1 I"', the '2'and the '2'12'30,
: 56,86and 120.Assuming
for the trireme and larger ships R,, = 1.05X 30.5 = 32.0,the same values as
in Table I result.
A.W. SLEESWYK TROPIS VI
Regarding the distinction between the interscalmia I ,and I, it must be

remarked that it existed only if the men pulling a scaloccio at the end of the
loom stood up and performed the pas du galere (Burlet and Zysberg 1990),
which required a larger interscalmium then when the men remained seated.
It must immediately be added that there is no positive evidence that rowing
in this manner was performed in Antiquity, but if it may be assumed that
there was an unbroken tradition of building and using galleys from Antiquity
to the Renaissance in Italy via Byzantium, it seems probable enough.
The larger interscalmium was probably necessary only when five or more
men were pulling on one oar. As there could be three oars in an
interscalmium, it means that the oarsmen in a "twelve" with three oars may
have been spaced at the smaller interscalmium value of I, and those in a
"five" with one oar per interscalmium at the larger value I,.There could be
considerable overlap, which necessitates considering theoretical r, values for
both interscalmia in that region. Table I may then be completed as follows
,: rt = 358 for the "5" and:
by the values of rtfor the interscalmium value of I
rt = 470 for the "6".
We now examine somewhat more closely the assumption that the two-
dimensional profile of the oarsmen determined their number seated in an
interscalmium. It implies that in the hemiolia ("1'12")and trihemiolia ("2'12")a
single file of oarsmen was seated in the midline of the ship. It must
immediately be remarked that the available evidence for the hemiolia does
not support a hypothetical arrangement of this type. Casson (1973: Fig. 117)
interpreted the well-known black-figured vase picture of a ship from the 6thc.
BC in the British Musuem having an additional layer of oars over the gunwale
forward of the mast only, as that of a hemiola. The literary evidence strongly
suggests that originally, i.e. before the c. BC, the hemiola of this type was
exclusively used as a pirate ship, but that later it was incorporated in several
navies, too.
In this instance obviously no men were seated along the midline, and
Equation (1) for ri = 1'" does not apply here. It may be remarked that of the
theoretical values for the number of oarsmen rt on board of ships of the
lineage of the triacontor that for the "1'"" comes closest to that of the
pentecontor. Nevertheless, the value, which even after correction is still as
large as 56, differs considerably from 50. Moreover, there is no evidence
whatever supporting the notion that the pentecontor and the hemiolia were
identical. More probably, the pentecontor belonged to a different lineage.
The trihemiolia appears to have been a respectable naval vessel from the
beginning; it is first mentioned for 304 BC as a type of ship in the Rhodian
navy. Whether it was related to the hemiolia is a matter of conjecture. So far,
no representation of an oared ship has been identified as that of a
trihemiolia. But if it is admitted that oarsmen, seated in a single file along the
midline, would each have pulled an oar either on starboard or on port, the
well-known picture of the stern end of a Roman ship with three layers of oars
from the column of Trajan (r. 98-117 AD) (Figure 2) could be interpreted as
that of a Roman trihemiolia. Alternate thalamians, seated in a single file,
would have pulled an oar either at starboard or at port. Seen from the side,
it would have looked as if every second thalamian oar in the ship was
missing.
Morrison (1995:75) derives the number of oarsmen on board the

trihemiolia from an inscription on a monument (265-260 BC) at the approach
of the acropolis of Lindos in Rhodes which gives the names of 288 men who
served in two trihemiolia. Each of these ships could have had "60 oarsmen
a side and 120 all told. A hyperesia of 24 non-rowers added (compared with
30 of the three and 45 of the four) brings the total up to the 144 of the
inscription". That number of oarsmen agrees exactly with the corrected
theoretical value of r, derived above, suggesting that the trihemiolia was
indeed of the type discussed here, and therefore belonged to the lineage of
the triacontor.
On the other hand, the mention of the fifty ships of the Boeotian fleet in
the so-called "ship's catalogue" in the llliad (11, 509-510) with complements
of 120 men should not be interpreted as an indication that these ships were
trihemioliai; the time interval of many centuries between the first mention of
this type of ship and the historical context of the Iliad is much too large for
that. One can only say that it is highly improbable that these ships belonged
to the lineage of the triancontor.
The assumption that the surface area of the profile of the arrangement of
oarsmen in an interscalmium is directly proportional to their number
obviously ceases to be valid if the oars are double-manned, i.e. if opposite
the men pulling an oar are placed men pushing the same oar, because the
profiles of these two categories of men would overlap. In that case it is
clearly the outer profile to the two overlapping arrangements of oarsmen that
matters.
In a recent study, the author concluded that the model of the galeass La
Royale in the Musee de la Marine in Paris originally had been intended to be
equipped with double-manned oars (Sleeswyk 1995). Apparently, the model
illustrated a proposal for a ship which was never built, and as far as we know,
double-manned oarage was not practiced in the modern era. But was it in
Antiquity? It seems so: in a paper which is to appear shortly, Sleeswyk and
Meijer (1997) argue that only an arrangement of 150 double-manned oars on
each side of the ship in 50 interscalmia, which each contained 40 oarsmen,
would fit the dimensions and all other particulars of the "forty" as given in
Athenaeus' text (V.203e-204b).
Double-manned oarage could also explain the mystery of the

Leontophoros, a ship which according to the Greek writer Memmon was an
okteres, "remarkable for its size and beauty. In each file, 100 men rowed so
that there were 800 in each part, 1600 in both". This ship had been built by
Lysimachos c. 330 BC to meet the challenge of his enemy Demetrios
Poliorcetes who had built a "fifteen" and a "sixteen". Casson (1973:113) has
pointed out the difficulties of interpretation of Memnon's text: how is it
possible that an "eight" could have been built to be more than a match for a
"sixteen"?
A solution is that the oars in the Leontophoros were double-manned by

8 men on the two sides of the loom of each oar; the file of hundred men
which Memnon mentions pulled 50 oars. With some justification in each
case, the ship could have been called either an "eight" of a "sixteen". From
a comparison of the lengths which a file of 100 men would have occupied it
follows that the interpretation as double-manned oarage is the probable one.
If they pulled 100 single-manned oars, one oar per interscalmium,that length
would have been some 125m, but if they pulled and pushed 50 oars, the
length vrould have been reduced to c. 87.5m. Assuming that that length was
at most three-quarters of the total length of the ship, the latter would have
been 167m for single-manned oarage, and 117m for double-manned. The
first option must be rejected out of hand, the second is somewhat shorter
than the 124m reported for the "forty". If Demetrios' "sixteen" had two oars
in each interscalmium and eight men on each oar (thus having sixteen files
of oarsmen on each side), he could have had the c. 1,600 oarsmen on board
which Equation (1) predicts (see Table I), and his ship would have been c.
84 m long.
The number of 1,600 men in Lysimachos' Leontophoros was to match

that in the "sixteen" of Demetrios Poliorcetes. The cross-sectional profile of
the men on the double-manned oar in Leontophoros was equivalent to that

of an "eight", requiring a width which was 1/42 of that of the "sixteen".
However, the length would have been greater, because an interscalmium
value I, = 1.75m would have been necessary to house the same number of
rowers, and the hull would have been correspondingly longer. It so happens
that the ratio between I,, and I, is close to 42, which means that the wetted
surface of the hull of the Leontophoros and the "sixteen" might have been
virtually the same, giving rise to the same resistance force at low speeds. At
higher speeds the wave resistance would have become an important
component of the total resistance force, and then the longer and slenderer
Leontophoros would have been at an advantage. But there can be no doubt
that she was a dangerously long ship as regards strength of the hull.
Although the Leontophoros would have been derived from Demetrios
"sixteen", which was assued to have been a ship of the triancontor lineage,
she herself did not belong to it, because her hull was proportionally much
longer.
The various corrected values of r, as a function of ri derived from the

literature which have been discussed in the foregoing, are plotted as circles
in the double-logarithmic diagram presented here as Figure 3. The
theoretical values of r, are given here as straight lines, which have a slope of
3 to 2, in accordance with Equation (1). It may be observed that there is
satisfactory agreement - somewhat better than in Table I - between the
values derived from the literature and the values predicted by the theory
presented here.
The weights of the two known bronze rams, those of Bremerhaven, 54kg,
and of Athlit, 465kg, are plotted in the diagram too. These weights should be
proportional to Ix b x h, i.e. to r;l2, if the ships which carried them were of the
same lineage. Informed guesses what types of ships these two were have
been made in the past (Murray 1986). Most probably both ships were
Roman; the ram of Bremerhaven can only have belonged to a monoreme,
the type of ship to which the Athlit ram belonged was probably a quadrireme
or "4". As: 465/4312kg= 58 kg comes close to 54kg, the two ships may
indeed have been of the same lineage, although it was not necessarily the
lineage of the Greek triancontor. Had the larger ship been a trireme or a
quinquereme of the same lineage as the monoreme, the corresponding
weights of the ram of the smaller ship would have 89 or 42 kilograms.
In conclusion, it may be said that the similarity between hulls implied by

the concept of lineage correctly gives the relationship between the number
of men ri in an interscalmium and the total number of men r, on board of a
number of important Greek oared fighting ships, probably reflecting a design
rule followed by Greek shipwrights. The weights of the two known bronze
rams of Roman origin may be related by the same similarity rule, indicating
that the Roman shipbuilders may have followed the same rule in designing
their ships.
Andre Wegener Sleeswyk
Rijksuniversiteit Groningen
Nijenborgh 4
9447 AG Groningen
The Netherlands
REFERENCES
R. Burlet and A. Zysberg, 1990, Mais comment pouvait-on vivre et voguer sur les galeres du Roi-
Soleil? Quand voguaient les galeres, 152-167. Paris.
L. Casson, 1973, Ships and Seamanship in the Ancient World. Princeton.
F.H. af Chapman, 1775 (1979), Architectura Navalis Mercatoria. London.
DKP (Der Kleine Pauly), 1979, Lexikon der Antike, Stuttgart.
J.M. Morrison, 1995, The Trireme. The Age of the Galley, 49-65. London.
W.M. Murray and P.M. Petsas, 1986, Octavian's Campsite Memorial for the Actian War,
Philadelphia.
A.W. Sleeswyk, 1995, The Oarage of the galeass La Poyale. The international Journal of Nautical
Archaeology, 24.3: 21 1-218.
A.W. Sleeswyk and F.J.A.M. Meijer, 1997, Quantitative analysis of Philopator's "forty".
Mnemosyne.
N. Witsen, 1671 (1994), Aeloude en hedendaegsche scheeps-bouw en bestier. Amsterdam
(Franeker).
ILLUSTRATIONS
Figure 1 Assumed similarity: in oared ships of the same lineage the cross-sectional area
available for seating oarsmen is the same fraction of the cross-sectional area of the
hull for large ships (a) and small (b).
Figure 2 Ships depicted on Trajan's column (c. 115 BC); the ship in the middle may be
interpreted as a trihemiolia, the others were biremes.
Figure 3 Diagram giving the number of oarsmen on board of ship, and the weight of the ram
in kg, as a function of the number of oarsmen per interscalmium. 'Z', 'S' and 'D'
indicate ships rowed a zenzile, a scaloccio and double-manned; 'R,' and 'R,,' refer to
ships without and with outriggers. 'A' gives the weight of the ram of Athlit, 'B' that of
the ram of Bremerhaven.
- -- -- -- - . --
- ---THE LINEAGE_
- -- -
OF_THE
.
TRIACONTOR
- _- -- -
nc 07 o a r s r n P n , w l o l r a m .n k~
n
Fig. 1
A MEDITERRANEAN SHIP CONSTRUCTION DATABASE;
DATING AND CLASSIFYING SHIPWRECKS BY THEIR HULL REMAINS
The word "database" is usually associated with computers. However, my

first database for ship construction was compiled in 1974, and it had nothing
to do with computers. It was presented in a series of lectures to field school
students and interested professionals who were excavating a fourth-century
wreck at Yassi Ada, near Bodrum, Turkey. I titled it "basic data to be
recorded for ancient ship construction" (in those days, data processed by
computers was usually called a "databank") and it amounted to little more
than a list of details that had to be recorded or investigated during our
excavation. By today's standards, it was small and reflected our limited
knowledge of ancient ship construction. But that database still exists,
although now it has expanded enormously and much of it is stored in a
FoxPro 2.6 database system housed in a computer with several gigabytes of
memory. Its meaning and purpose remain the same, though - a base of hull
construction data, arranged for efficient comparative evaluation, that is
intended to provide as much information about a vessel or group of vessels
as possible. Although the work is usually easier to store and manipulate on
a computer, that device is by no means a necessary component. In fact,
laboratory models and graphics (on computer and/or drafting board) are
sometimes the best or the only methods by which to compare or interpret
certain details. The key words in this process continue to be "data" and
"evaluation." The processing vehicle is a matter of choice.
That base of ship and boat construction data had already become quite
sizeable by 1987 when, at the second symposium on ancient ship
construction at Delphi, I presented a paper entitled "Problems and Progress
in Dating Ancient Vessels by their Construction Features1."While that paper
predicted a bright future for structural interpretation of Mediterranean vessels
and cited progress in both quantity and quality of hull recording over the
previous decade, it also revealed many shortcomings. Coins and pottery
continued to be the most reliable sources for dating shipwrecks. The
greatest discouragement was centered in the recording and publishing of
hull remains. Far too many wrecks still went unpublished; many others were
insufficiently documented. The paper ended with a plea for more accurate,
more extensive recording, and included a list of structural features that
should always be documented in the hope that some sort of standardization
might evolve.
J. RICHARD STEFFY TROPlS VI
When the invitation to this symposium arrived, I recalled the second

symposium and wondered how far we have progressed in the nine years
since the meeting at Delphi. Certainly, underwater equipment, procedures,
and safety have improved considerably, and some of those improvements
have simplified recording. Laptop and network computers have greatly
enhanced recording flexibility and the interchange of ideas. FAX and E-mail
have eliminated most of the problems of excavating far from headquarters.
Wood identification, metal analysis, computer graphics, biochemistry,
conservation, and various forms of electronic analysis have all been refined
and thus have enhanced the interpretation of hull remains. But how about
the dating and comparative studies of hulls that were found so sadly lacking
at Delphi? For an answer, I turned to my existing database on Mediterranean
vessels, which soon became the subject of my paper for this symposium.
Any good collection of publications on hull remains can become the

foundation of a good database; it is the manner in which the data is
compiled and interpreted that is important. While there are many ways to
accumulate a bank of data on hull structures, whatever method is employed
must be flexible enough to accommodate all forms and sizes of watercraft.
Furthermore, it must be adaptable enough to allow the researcher to
recognize functions and properties that are revealed for the first time. Ships,
you see, are not really datable or identifiable in the context of amphoras and
coins as mentioned earlier. They are far too complex for that. Even where
hull survival is extremely limited, dozens or even hundreds of entries will go
into such a database. Our ancient Mediterranean file alone lists seventeen
types of fastenings that are known to have been applied in thirty-one
different ways, and certainly we are still far short of having discovered all that
were used. One fragment of portside planking on the Kyrenia wreck2had six
types of fastenings within a 50cm span - regular planking nails, wooden
sheathing nails, lead sheathing tacks, regular mortise-and-tenonjoints, parts
of patch tenons, and treenails. Although joinery is still being processed,
there are already several dozen forms of joins; probably we will wind up with
a hundred or more ways to scarf, lap, hook, dovetail, butt, or rabbet two
pieces of wood together by the time our recording is completed. Shape and
fabric, those two all-important features for identifying amphoras, are equally
important in hull identification. In shipbuilding, however, the fabrics - wood
types - often varied within single containers - hulls - and required
hundreds of methods and devices to mold them into those big, complex
shapes. Ships and boats can be dated solely by their construction features,
but neither you nor I will ever classify ships with the same finesse by which
amphora experts ply their specialty. With hundreds or thousands of parts
comprising its structure, one ship could not have been exactly like any other,
even if it was made in the same yard to the same design. Such duplication
would require identical trees to be converted under identical conditions.
Even sister ships varied somewhat throughout the wooden ship era, and that
is what makes this database interesting. It is the variety of methods and
materials that shipwrights employed to achieve a certain hull shape that is
most important in evaluating the history of shipbuilding technology.
Consequently, it is as much a study of shipwrights and economics as it is of
hull construction.
My own system is directed toward analyzing construction techniques

and utilization of materials in such a way that it will assist in dating hulls,
recognizing their nationality or geographic origin, and revealing such details
as economic influences, the use of practical mathematics in design or
construction, the conversion of timber, hull strengths and weaknesses, and
a host of other characteristics; there is even a category to accommodate
possible factors that are as yet unknown to us. To make the entire process
more manageable, there are six major divisions containing vessels built
before 500 BC, from 500 BC to 100 BC, from 100 BC to AD 100, from AD 100
to AD 500, a medieval category extending to 1500, and a post medieval file.
All data from one category can be compared or interchanged with that of any
other category.
There is a profile of each wreck, which is essentially an expansion and

modernization of the profile list presented nine years ago at Delphi. A well-
preserved vessel might provide two or three hundred entries in such a
profile, while more sparsely presefved remains will contribute much less.
Sources for these profiles come mainly from published reports or
symposium papers. Where details have not been published, I have
sometimes been able to contact those who recorded the site (or, sadly, did
not record it but may remember something). Anthony Parker's list of
Mediterranean shipwrecks has contributed valuable information3. The
bibliographies published by Parker and John lllsley have also been a
tremendous help in locating original sources4.
In addition to the profiles, there are sixteen specialized files that establish
a base for interpreting the evolution of certain hull features, technological
progress, and similar details. These files, which may be altered or expanded
as research dictates, contain from six to twenty-six entries. They include
project information, general hull design data, wood characteristics for
timbers and fastenings, and data concerning keels, keelsons, posts,
J. RICHARD STEFFY TROPIS VI
planking, wales, mortise-and-tenon joints, frames, fastenings, and various

types of hull sheathing. All entries or files do not necessarily apply to all
vessels. For instance, mortise-and-tenon joint information would not be
applied to a Venetian galley file, nor would lead sheathing data apply to the
Kinneret boat. The part of the Mediterranean base discussed here contains
data from seventy-eight vessels dated between 400 BC and AD 1025.
Although we eventually hope to make contact with people knowledgeable in
a couple hundred more sites with hull remains, we have so far listed fifteen
ships that are known or estimated to have been over 30 meters in length. In
addition, there are twelve ships between 20 and 30 meters long, twenty-four
vessels between 12 and 20 meters long, six boats, and twenty-one wrecks
whose lengths cannot be determined. Many of these vessels are poorly
documented. Many more include questionable comments and details.
Sadly, the great majority of Mediterranean wrecks listed by Parker have
provided little or no usable hull information because their publications are
too sparse, too erratic, or completely non-existent.
The following is an overview of some of the more interesting results of

these Mediterranean hull studies to date. Additionally, because so many of
the people attending this symposium direct shipwreck projects to record or
research hull remains, I will pass along suggestions for improving the
recording of timbers as revealed by various database comparisons. More
detailed results of this research have been, and will continue to be,
published as separate journal articles or project reports. In fact, it was during
the preparation of the final report of the Serqe Limani merchant vessel5that
the database system described above was found to be too limited and
unable to deal properly with design features such as proportions and
mathematics. At the ISBSA conference in France two years ago, I related that
we had at least limited proof that the Serqe Limani merchantman, which
must have been built late in the tenth century or very early in the eleventh,
had been constructed with the Byzantine foot as its unit of measurement6.
Furthermore, I noted the predominance of proportions and multiples of that
dimension throughout the hull. Since that time, both of these features have
been proved beyond doubt and another very important aspect has surfaced.
For at least two years Fred Hocker, who did a lot of work on Serqe Limani
research models and built the sectional replica on display in the Glass Wreck
museum in Bodrum, tried to convince me that the hull had standing
tailframes in addition to the standing frames we established in the central
part of the hull. In other words, he was convinced that complete frames were
in place at the after end of the bow and the forward end of the stern before
any planking was installed. I remained skeptical because I did not believe
these shipwrights were able to project tailframe shapes; it was not

compatible with the existing frame projections. And even if there were pre-
erected tailframes, I saw no way we could prove it with such limited hull
survival. The matter came up again last spring when the final report was
being illustrated with a model that was intended for display in the Glass
Wreck museum in Bodrum. The planking did not work well with our
hypothesized bracing system; it did not seem to go on properly without
some rigid, predetermined form to dictate and hold its hard curvature into
the stem and stern posts. And so, to update our previous cataloging of
fragments and to make the final report as accurate and complete as
possible, I returned to Bodrum for additional research and restudy of the
timbers. The results were far more gratifying than anticipated. With the help
of Sheila Matthews, who reassembled most of the S e r ~ eLimani hull remains,
we found plenty of evidence for units of measurement. We now had a couple
hundred locations that confirmed original planking thicknesses and many
more that confirmed frame dimensions and spacings. In fact, the basic unit
of measurement used to build the S e r ~ eLimani hull has been definitely
established and falls within a few millimeters of the value established for the
construction of Hagia Sophia in Istanbul7.
There were indeed tailframes, too, one at each end of the hold where the
bow and stern planks began their hard inward bends toward the posts (Fig.
1). In section, these tailframes resembled the ones Marco Bonino reported
for the first Contarina wreck, although they were made completely
differentlf. Originally, each tailframe had about one and one-half times the
cross-sectional area of the other standing frames and each consisted of two
half-frames whose lower ends overlapped, crossed the keel, and most likely
were attached to each other. One half-frame in each pair was fastened to the
keel. These appear to be the only true half-frames in the hull. Additional
details of this interesting feature can be found in the final report on the hull,
which is about to go to press. What is important here is that our database,
as described above, had no provisions to detect such things and is presently
being revised. Furthermore, it made me wonder how much we really know
about the design and fabrication of ancient hulls, especially the Greek,
Roman and early Byzantine vessels already documented. If there are ten
standing frames and a pair of tailframes, then it seems likely that the builder
of this 15-m-longfreighter could pre-determine rising and narrowing lines, a
feature that has previously been attributed to Genoese, Venetian and later
shipwrights. Even if these lines were not yet recognized by the Serqe Limani
shipwright, at least the potential and required technology were there. In fact,
such lines were produced automatically when the planking was installed.
J. RICHARD STEFFY
And that, in turn, raises a couple of other interesting questions. If this builder
could pre-erect frames and produce a hull with such practical construction
and efficient proportions, what might shipwrights of this period have been
able to do with larger, more sophisticated vessels? Furthermore, if Byzantine
shipwrights were already capable of such techniques by the tenth century,
were there already clues for similar geometric projections in the ships of the
earlier Byzantine and Greco-Roman periods? Had we neglected to
appreciate such proportions, angles, and other clues on earlier vessels?
Undoubtedly we had. A few suggestions resulting from our database studies
may help all of us recognize these obscure hull properties in the future.
To first answer the questions prompted by the 1987 Delphi symposium

paper, we are steadily improving in recording hull remains, both in breadth
of coverage and interpretation of details. Although several of the earliest
underwater excavations included excellent hull recording, the profiles for this
database reveal a general improvement with the passing of time. Not only
have there been new shipwreck reports since 1987, but older ones have
been restudied and there has been a lot of updated or new infcrmation on
those earlier excavations. Sadly, however, there are still far too many hull
remains that have been excavated or intruded upon, but never published. In
too many instances, they are remains which might add appreciably to our
present bank of knowledge.
As for dating and establishing the nationality of shipwrecks solely by their

construction, our database has illustrated the naivete of my remarks in 1987.
Ships and boats are so complex and subject to so many variations or
revisions of structure that they can never be singly datable in the way that
amphoras are recognized. A few of the more important reasons for this
statement follow.
DESIGNS
Our database essentially confirmed what we already knew about general

hull designs. Early in our period of investigation (400 BC to AD 1025),hulls
with wineglass-shaped cross-sections predominated, although only the
Ma'agan Michaelg, Kyrenia, and Marsala Punic'' wrecks were sufficiently
preserved and recorded to confirm this. Hull shapes became increasingly
fuller, stronger, and more efficient as time progressed. While many hulls
maintained a V-shaped entry of the garboards into the keel, others were
even fuller and flatter with garboards abutting keels horizontally or nearly so.
There seems to be an orderly evolution in the use and arrangement of basic

hull timbers and edge fastenings, the former becoming stronger and more
efficient internally and the latter becoming smaller and more sparse until it
disappeared altogether.
Within this orderly evolution of design and construction, however, there

was an independence of thought and technique that, at least to the present,
cannot be assigned to any specific date or geographic area. Timber
applications, fastening methods, and construction processes vary so much
from hull to hull, even within similar hull designs, that dating or assigning
nationality to a certain combination of hull assemblages is risky at best.
Some applications appear to die out and then reappear centuries later, so
that with our present database of information we are still not the most reliable
dating source on a wreck with a variety of common artifacts and cargo. But
we can, and must, improve our approach to this problem through better and
more extensive recording. Some examples of design contradictions and
shortcomings in recording are listed below.
KEELS
There does not appear to have been an orderly evolution of keel

development. With the exception that frames were not attached to keels in
the first part of our study period, there was not even a sequence of types of
fabrication. For instance, where wineglass-shaped hulls were used, keels
were sometimes keystone-shaped; garboards entered their upper corners
and their bottom surfaces were dimensioned smaller than their sided
dimensions at the rabbets (Fig. 2a). Most, probably all, keystone-shaped
keels were used in cases where the garboards entered at angles exceeding
thirty degrees from the horizontal. In such cases, keellgarboard tenons were
pegged from the side of the keel; where the angle was extreme, the
garboard side of the tenon was also pegged from the outside. This was
especially true for the second garboard to be installed, since there was no
room to swing a mallet with the first garboard in place. In double-planked
hulls with similar bottom shapes, there were two rabbets but all other factors
remained the same.
By far the larger percentage of keels were shaped rectangularly,

regardless of the angle of garboard entry. Some had rabbets, others did not,
and there seems to be no pattern of development here. For instance, the
garboards in the Ma'agan Michael wreck enter the sides of the keel at about
the same angles as those of the Kyrenia and Marsala Punic wrecks, yet the
Ma'agan Michael keel has no rabbets at all; the inner garboard edges were
simply set flush with the side of the keel (Fig. 2b).
Flat-bottomed hulls followed similar illogical patterns, even for similar

hulls of similar periods. The Kinneret boat had no keel rabbet or false keel
(Fig. 2c)''. The Herculaneum boat, of similar age and size, had beveled
upper keel edges to receive similarly beveled garboard edges (Fig. 2d)I2.As
with all hulls whose garboards entered the sides of the keel at angles of less
than thirty degrees, the keellgarboard tenons of both vessels were pegged
from the tops of their keels.
Incidentally, the Kinneret and Herculaneum boats are prime examples of

the value of comparative studies. Both were approximately the same size
and probably constructed within a few decades of each other. No artifacts
could be associated with either hull, although the destruction of the
Herculaneum boat by the eruption of Mt. Vesuvius assures a precise dating
for that hull. The Kinneret boat was in danger of destruction and underwent
an emergency excavation, which permitted only a preliminary study before
conservation could begin. That conservation process is now completed and
more thorough studies of the hull are underway. Publication in the near
future will assure additional information on this hull. The Herculaneum boat
was completely carbonized and was found upside down on the ancient
beach. A preliminary study of the exterior of the hull was made and
published, although these very delicate remains had to be inverted and
cleared of pyroclastic material internally before the interior and part of the
exterior of the hull could be recorded. The hull has since been turned upright
and conservation is said to have been completed. Hopefully it, too, will be
further documented and published because the Herculaneum boat is one of
the finest crafted vessels I have ever examined. In spite of its inverted and
charred condition, one could still recognize this exceptionally graceful hull
design, a masterful planking plan that featured strakes running the full length
of the hull without the use of a single scarf, and the sculpted wales and rub-
rails. In other words, its form and methodology are completely counter to
that of the Kinneret boat. Complete reports on the construction of these two
hulls will not only add a lot of new information to our base of structural data;
it will provide an interesting comparison between craft and craftsmen in
Rome and its outposts.
After keellgarboard tenons were no longer pegged, as on the seventh-

century Yassi Ada hull, keel rabbets could be placed further down the sides
of the keel (Fig. 2e). This allowed the frames to overlap the top of the keel for
extra security, a practice that was frequently employed until the end of the
wooden ship era. Curiously, the first and last vessels in our period of
investigation had rectangular keels of nearly identical dimensions (Figs. 2b
and 2f). Neither had a rabbet. However, the Ma'agan Michael keel, which
survives in its entirety as a single piece, was more sophisticated in that it was
more carefully crafted and had a false keel attached to it with mortise-and-
tenon joints. The Serqe Limani keel was roughly sawn from at least three
pieces and had no false keel. In summation, the march of time did not reveal
an overall improvement of keel fabrications, either in design, strength, or
quality. Nevertheless, each was probably excellently attuned to its own hull
structure.
PLANKING
Most projects seem to have recorded planking the most extensively.

Planking also occurs most frequently, even on sparsely preserved wrecks.
There is a predominance of pine throughout the period. So far, twelve of
those vessels are known to have been planked and framed completely in
pine. Twenty-six more were completely planked in pine, with many more
partially planked with that kind of wood. And why not? Pines with grain
structures like Allepo and Brutia were very sea-kindly, moderately strong,
and readily accessible for harvesting in most areas. Most importantly, they
were easily worked with the tools of antiquity, an important factor
considering all the shaping, scarfing, and mortising required of ancient
planks. Perhaps other species of wood were more desirable for long, narrow
galleys, but for full-bodied merchant vessels, pine has so many advantages
that I find myself asking why when I profile a hull planked in anything other
than pine. Apparently a lot of ancient shipwrights concurred, regardless of
the size or period in which their hulls were built. From the little Kyrenia
merchantman to the big Caesarea ship, estimated to exceed 40m in length,
hulls were often framed and planked exclusively in pine13.And that, too, was
the case from Ma'agan Michael to Serqe Limani.
Wood species, scarfs, planking shapes, and most other factors

concerning planking seem to follow a general use pattern. In other words,
this database does not indicate many datable characteristics about hull
planks before the decline of mortise-and-tenonjoinery, but it does cry out for
more extensive recording in some cases. For instance, the hooding (forward
and after) ends of planks on the Serqe Limani hull and all later hulls I
examined were reduced in thickness at or before they reach the posts. That
has not been the case on the ancient hulls I studied. Probably edge-joined
plank ends needed more thickness because of their joints, but we need to
know a lot more about these ancient shell terminations. Thicknesses should
be measured every few centimeters, all fastenings should be carefully
recorded, and the extreme ends and edges carefully documented. This
becomes doubly important where the posts have not survived.
Planking thicknesses should be measured in the centers of all planks at

frequent intervals, because sometimes this is the only place where the
original thickness of the plank has not been dubbed away with an adze.
Thicknesses should also be measured where original saw marks survive on
opposite sides of a plank - this is usually the best indicator of original
thickness. From the sparse records we now have, it would appear that
classical vessels were shaped with an adze more extensively than were
medieval hulls, but the recording is so rare that this must remain an
assumption. The Serqe Limani planks and timbers had many original saw
marks on opposing surfaces, while adzes were used less frequently than on
any other hull I have studied. Again, there is a need for more extensive
recording of tool marks.
Mortise-and-tenon joints might also be more carefully recorded in some

cases. So many of our profiles list dimensions for either mortises or tenons,
but not for both. Tenons did not necessarily fit their mortises precisely, so
both should be recorded. Also, joint spacing may vary at different locations
in the hull, so spacing must be recorded at several appropriate side and
bottom seams and in the ends of hulls as well as centrally.
FRAMES
For all but the last few centuries of the period we are investigating,
frames were shaped to fit shells of planking. Hence there were fewer flats cut
into their planking surfaces at the turn of the bilge, fewer assembled frames,
and less symmetry to framing plans. Most notably, the centerlines of frame
stations were seldom straight, and individual frame timbers wandered all
over their hull surfaces.
Floor timbers for wineglass-shaped hulls are a case in point. On most

hulls, they appear to be crotch timbers; i.e., timbers formed from the
junctions of tree trunks and/or two or more of their branches (Fig. 3). 1
noticed several such crotch timbers while visiting the Ma'agan Michael
project, and most of the frame drawings of the Marsala Punic wreck are
similarly illustrated. Frame 19 of the Punic hull has an extension piece
attached to its lower surface to fill the gap between garboards, but it still
appears to have been made from a crotch timber. Even the floor timbers of
the big merchantmen, such as the Madrague de Giens vessel, were made
from crotch timbers and pierced with large holes above the keel14. The
selection of so many crotch timbers to span floors of ancient hulls must have
required a lot of careful timber selection and shaping. It makes some sense,
from the standpoint of strength, to use crotch timbers for those hulls whose
frames were fastened to the keel, but I fail to see the reasoning behind their
use on vessels where none of the frames were attached to the keel. The
Kyrenia builder may have had a more practical, if not more economical,
solution. In that hull, floor timbers were cut from simple curved stock and the
triangular sections over the keel were cut separately and aligned with their
floor timbers by means of a pair of unpegged tenons (Fig. 4). There was one
exception, however. Frame 52, which replaced one that probably rotted and
was installed long after the ship was built, was made of a crotch timber. It
was quite obviously fashioned by a different ship carpenter, probably one
who belonged to a later generation and worked to a different philosophy.
And that is the complexity, and the danger, of attempting to date ships by
their construction. Even in a single hull, within a decade or two of its launch,
different procedures were followed. In this case, the later procedure
matched those of a century before and a century later.
Within this realm of alternating floor timbers and half-frames, there are
many more variations in fabrication and methodology. I can only suggest
that framework be recorded as carefully as possible. Perhaps the greatest
difficulty of listing frame data lies in the measurement of frame spacing,
known in later years as "room and space." Room and space doesn't quite
satisfy ancient and medieval frame spacing because of their erratic
placement and curvature. Average frame spacing is a better measurement,
because ancient and early medieval vessels were not built to the strict frame
spacings of later craft. But where does one take the average spacing
measurements? Along the keel centerline was the desired location in latter-
day construction, but frames of ancient vessels had very irregular spacing
and most half-frames never reached the keel. Irregular spacing was
unavoidable because trees seldom grew branches that were correctly
curved in one plane and perfectly straight in another. The shipwright, in
selecting the proper curvatures to match his shell of planks, often had to
settle for timbers that curved fore or aft slightly. Sometimes, fore and aft
curvature was quite radical. One example is a frame found in Tantura lagoon
in northern IsraelI5.Actually, it was an olive branch and it was positively the
least worked frame timber I have ever seen. This one could hardly be
designated with sided and molded dimensions. It was actually round, about
12cm in diameter, with the part that touched the planking flattened just
enough to seat itself and accept nail shafts. At some places it did not touch
the planking at all; at others, only a centimeter or two of surface made
contact. None of the bark had been removed. Most importantly, it made a
perfect S-curve laterally, winding far off any proposed centerline in either
direction. Quite obviously, this frame was added after the planking was in
place, and the shipwright placed it so that it best covered and supported the
inner planking surface (see the wreck plan in Y. Kahanov's paper above). In
such cases, room and space is not nearly so much a factor as is good
coverage of the overall support of the framework.
While this frame may have been exceptionally crude and curvy, it was by
no means a rarity. A glance at almost any ancient wreck plan reveals how
much the frames angled or curved away from any straight athwartships line,
the futtocks sometimes taking a radically different angle from the centerlines
of their floor timbers. Where planks preceded frames, futtocks did not have
to be attached to floor timbers, nor did either timber have to follow a
designated centerline. It was merely necessary for them to be distributed so
that they provided the greatest support to the hull. Consequently, average
frame spacing should be the determining dimension of distribution for such
hulls.
Is average frame spacing all that important? And how does one
determine average frame spacing? First of all, it is very important. Remember
those proportions on the S e r ~ eLimani hull and our questions about the
disciplines that lead up to them and the standing frames? If we are ever to
determine the origins of documented forms of early naval architecture, we
must find ways to seek them out. Certainly one path would be to determine
proportions, strength factors, and the like in earlier hulls. One proportion, or
set of values, that keeps cropping up throughout the Greco-Roman period is
mortise-and-tenon joint spacing, average frame spacing, and the
relationship between the two. Average joint spacing for all vessels recorded
so far between the early 4" century BC and the 3rdcentury AD is about 12.5
centimeters. In fact, hulls of all sizes have an overwhelming majority of
recorded joint spacings between 11.5 and 13.5cm. Even double-planked
hulls have effective spacings in that range. Frames, on the other hand, are
frequently spaced twice that distance; both little Kyrenia and big Madrague
de Giens have alternating floors and half-frames spaced on approximately

25cm averages. But frame spacing is usually so poorly documented or
illogically measured that it is so far largely unreliable for accurate database
use.
Whether increments in the neighborhood of 12.5 or 25cm relate to

ancient units of measurement cannot be substantiated as yet. But one thing
is certain. If we are ever to find the ancient proportions or increments that led
to the meualunas and the projections of frames and rising and narrowing
lines used by Venetian and Genoese shipwrights, then we must measure
things like frame spacing more carefully. Frame spacing should be
measured between the centers of frame timbers at a number of places along
each frame station, depending upon the distribution and distortion of the
timbers. Certainly, one set of measurements should be taken along the keel,
another at the turn of the bilge, with one or two more along intermediate
planking seam locations. A meticulous, well-drafted wreck plan can
sometimes fulfill all of these demands in a single illustration, and they can be
confirmed or expanded by others. Alternatively, tables listing several rows of
spacings are helpful.
Proportions in timber sizing may also provide clues, so they should be

recorded carefully. For example, the Kyrenia ship has wales that are one and
one-half and twice as thick as the planking, while the average sided
dimensions of the frames were also double the plank thickness. Frame
spacing is twice the mortise-and-tenon joint spacing, and on and on.
Whether such proportions are simply a part of that ship carpenter's work
ethic, or whether they are part of a larger discipline, remains to be seen.
Indeed, whether such proportions eventually will provide clues to the origins
of formal shipwrightery when compared with values from other vessels
cannot be confirmed. But certainly we must investigate such matters, and
therefore such details should be recorded.
In summation, our database has already made two things abundantly

clear. The first is that additional, more precise information must be fed to the
existing bank. And so my system is being redesigned and much of the
original recording will be done over again with broader, more precise data
input. The second was that determining a vessel's date or nationality, while
an important step in itself, is but one of many very important revelations a
good database will provide. It is important that those who record or analyze
hull remains realize fully the value of those remains and their potential of
unlocking the secrets of the past. Nothing a sunken ship might have been
carrying was as complex as the carrier itself. No artifact required as much

thought and time to produce, no artifact touched the lives of as many people
as did the hull that carried it, nor did any artifact have as profound an effect
on society, either technologically, economically, or socially. And for most of
us, nothing in that hold could have been nearly as mysterious or as beautiful
as this ship whose rotted remains we now investigate.
J. Richard Steffy
Yamini Professor Emeritus
at Texas A&M University
College Station, Texas
NOTES
1. J. R. Steffy, "Problems and Progress in Dating Ancient Vessels by their Construction

Features" in H. Tzalas (ed.), Tropis 11. Proceedings of the 2" lnternationalSymposium on Ship
Construction in Antiquity (Athens, 1990), pp. 315-320.
2. J. R. Steffy, "The Kyrenia Ship: An Interim Report on its Hull Construction", American Journal
of Archaeology 89.1 (1985): 71-101.
3. A. J. Parker, Ancient Shipwrecks of the Mediterranean & the Roman Provinces, BAR-S580,
Oxford, 1992.
4. J. S. Illsley, An Indexed Bibliography of Underwater Archaeology and Related Topics,
lnternational Maritime Archaeology Series, Volume Ill, Oxford, 1996.
5. J. R. Steffy, "The S e r ~ eLimani Vessel", in Wooden Ship Building and the Interpretation of
Shipwrecks (College Station, 1994): pp. 85-91.
6. J. R. Steffy, "Seldom Discussed Features of Ancient and Medieval Ship Construction",
Proceedings of the Seventh lnternational Symposium on Ship and Boat Archaeology,
Tatihou, July 1994. In press.
7. Descriptions of Byzantine units of measurement relating to Hagia Sophia can be found in E.
Shilbach, Byzantinische Metrologie, Munich (1970), 13-16. See also R. J. Mainstone, Hagia
Sophia: Architecture, Structure, and Liturgy of Justinian's Great Church, London (1988). 117.
8. M. Bonino, "Lateen-rigged Medieval Ships: New Evidence from Wrecks in the Po Delta (Italy)
and Notes on Pictorial and Other Documents". lnternationalJournal of Nautical Archaeology
7.1 (1978): 9-28.
9. E. Linder and J. Rosloff, "The Ma'agan Michael Shipwreck" in H. Tzalas (ed.), Tropis 111.
Proceedings of the 3" lnternational Symposium on Ship Construction in Antiquity (Athens,
1995), pp. 275-81.
10. H. Frost, "First Season of Excavation on the Punic Wreck in Sicily", lnternationalJournal of
Nautical Archaeology 2.1, (1973): 33-49.
11. J. R. Steffy, "The Boat: A Preliminary Study of its Construction" in S. Wachsmann et al., The
Excavations of an Ancient Boat in the Sea of Galilee (Lake Kinneret), 'Atiqot (English Series)
XIX, Jerusalem (1990), 30, 31.
12. J. R. Steffy, "The Herculaneum Boat: Preliminary Notes on Hull Details", American Journal
of Archaeology 89.3 (1985): 519-21.
13. M. A. Fitzgerald, "A Roman Wreck at Caesarea Maritima, Israel: A Comparative Study of its
Hull and Equipment", Doctoral Dissertation, Texas A&M University, 1995, 18-33.
14. M. Rival, La Charpenterie Navale Romaine, Paris (1991), 208.

15. See Y. Kahanov, "The hull construction of the Byzantine wreck at Tantura lagoon, Israel"
presented later in this symposium.
CAPTIONS FOR ILLUSTRATIONS
Fig. 1. A sternward view of the reconstructed forward tailframe of the eleventh-century Serqe
Limani vessel.
Fig. 2. A selection of typical keel cross-sections: (a) A keystone-shapedkeel typical of the fourth
and third centuries BC; (b) a cross-section of the Ma'agan Michael vessel's keel of about
400 BC; (c) a cross-section of the first-century Kinneret boat's keel; (d) a cross-section
of the first-century Herculaneum boat's keel; (e) a cross-section of the keel of the
seventh-century Yassi Ada ship; (9 a cross-section of the eleventh-century Serqe Limani
vessel's keel.
Fig. 3. One area of a tree from which a crotch timber could have been cut.
Fig. 4. A floor timber of the Kyrenia ship, showing the separate chock over the keel centerline.
(All drawings by the author, based on illustrations in the project reports cited.)
Fig. 3
Fig. 2
Fin A
THE NUMBERS IN THE NAMES OF ANCIENT WARSHIPS:
SOME PROPOSED COMPROMISES
Introduction
Many types of ancient warship were described by a name which

contained a number: the dieres ('bireme' in English) where the di means
'two', the trieres ('trireme' in English) where the tri means 'three', and so on.
Some years ago, as part of Tropis Ill, it was suggested that the number in
the name indicated the number of oarsmen in cross-section (Tilley, 1995).
That was completely at odds with the widely accepted view that the number
in the name indicated only half the number of oarsmen in cross-section. It
required biremes with only two (not four) oarsmen in cross-section and
triremes with only three (not six), a drastic and revolutionary change in ideas
about ancient warships, which implied that whole libraries of books and
learned articles are fundamentally in error.
The suggestion has been opposed by many people who are regarded as
authorities on ancient ships, who were all committed to six-banked triremes
of one sort or another before the idea of triple-banked triremes was mooted.
It appears that the new theory is simply too radical to be accepted by the
present generation of ancient ship authorities. They have not, however, been
able to show any discrepancies between the ancient evidence and the new
theory. Opposition to it has concentrated on asserting that triple-banked
rowing is impracticable, on ignoring the evidence in its favour and on
concocting pseudo-evidence for three-level triremes. This paper proposes a
number of compromises by which a good deal of the relevant evidence
could be accepted and pseudo-evidence discarded, without the need to
relinquish faith in three-level triremes. It also appeals for trials to test the
assertion that triple-banked rowing is impracticable.
ALEC F. TILLEY TROPIS VI
English terminology
The first proposed compromise concerns the English language:

upholders of the traditional view on ancient warships should use the terms
'bank' and 'room' correctly or not at all. This (Fig. 1) is a double-banked
boat. In most English-language literature on ancient ships it is called single
banked. But a single-banked boat is like this (Fig. 2) with only a single line
of oarsmen. All modern European languages use the same system, and
misuse occurrs only in English. No French author would call the
arrangement in Fig. 1 'arme a point' in the way that it is miscalled 'single
banked' in English. No Greek author uses the term 'diplokopos'to describe
this (Fig. 3), nor a German 'doppel ruderig', nor an Italian 'doppio ordine di
reme', in the way that it is so often miscalled 'double banked' in English. The
ambiguity is unscientific and ought to be abolished. Think how it would
complicate discussions in zoology if some zoologists insisted on calling a
cow a biped because it has two legs on either side.
The same confusion surrounds the word 'room' introduced into English
from Scandinavia to describe a single unit in an oared vessel. Landstrom
(1961, 64) rightly explained that in Scandinavia the size of a ship was
I...
measured by the number of rooms as they called them... Each of these

rooms on a fighting ship meant a pair of oars, and for each pair of oars there
was a thwart'. Morrison described a trireme as 'a ship in which there were
three oarsmen to each unitary division or "room", called in Latin
"interscalmium"' (Morrison and Williams, 1968, 339). Those words describe
my idea of the original trireme. But Morrison's diagrams make it clear that six
men in each room was what he had in mind.
Faith in three-level triremes requires the acceptance of such

improbabilities as the ship having been too difficult for ancient artists to draw
or sculpt or paint (Morrison and Williams, 1968, 169). Such a strong faith
would surely not be shaken by calling its object 'six-banked' or by saying
that it has six oarsmen in every room.
Triple-banked rowing
The second compromise which upholders of the traditional view ought to

accept is that ancient navies used triple-banked vessels. There is evidence
of ancient ships rowed like this (Fig. 4) with three men on each bench, each
man using his own oar (Fig. 5), an arrangement seen on the Siren Vase (Fig.
6) and in the Victory of Samothrace ship (Fig. 7). But the evidence is rarely
acknowledged in print. It is, for example, completely omitted from The Age
of the Galley, a recently published book which purports to take 'full account
of the latest research', and it was ignored in the discussions that preceded
the building of the Olympias. It is quite often ridiculed, though. Professor
Morrison conflated two ways of rowing triple-banked, the Siren Vase method
(Fig. 4) and a two-level method (Fig. 8) and ascribed the resulting confection
to me:
'The name trieres was first adopted to describe a system of 'benches'
one behind each other throughout the rowing compartment of the ship, on
which sat sets of three oarsmen. In each set the port and starboard oarsmen
row normally while the midships oarsman sculls at a lower level (how then is
he on the same 'bench'?) pulling a pair of longer oars' (Morrison, 1978, 204).
Lucien Basch wrote of the arrangement in Fig. 5:
'Les rames no2 et 3 a compter de I'avant appartiendraient, selon Tilley, a
cette troisieme file' (Basch, 1987, 271). In English:
'The second and third oars counting from forward belong, according to
Tilley, to this third file'.
Coates (1995, 160) transformed the arrangement in Fig. 8 into 'two
superimposed levels' and then asserted that a ship using it would capsize.
But such lampoons are not essential to faith in six-banked triremes. If a

trireme had six oarsmen in cross-section, a vessel with only three might have
been named after the number 'one-and-a-half', and there is indeed an
ancient Greek ship-type name which implies that number: hemiolia. Before
he confused the system in Fig. 4 with that in Fig. 8, Morrison commended
the idea that the Siren Vase showed a triple-banked vessel, writing:
'His [Tilley's] main argument is that a red-figure vase... shows a ship with
three banks in the modern English nautical sense. That is, I think, a valuable
suggestion' and going on to say that it might well provide a solution for the
hemiolia puzzle (Morrison 1969).
But Morrison never published those views, and later put forward different
ideas for the hemiolia, without mentioning the Siren Vase or the oarage
shown on it (Morrison, 1980, 121-6).
Casson has kindly allowed me to quote his opinion:

'Your solution is the only one offered so far that makes sense of the
features shown on the Siren Vase and is at the same time demonstrably
workable'.
For Casson, the ship on the vase is named after the number 'one', a
moneres, having all the oarsmen on one level.
Upholders of the traditional view of ancient warships ought to be able to

accept ancient triple-banked ships, naming them either after the number
'one' or 'one-and-a-half', whichever they prefer. At present, ridiculing or
ignoring the idea of rowing with three banks of oars inhibits experimentation.
A simple experiment, long overdue, one for which I would dearly like
support, is to make a mock-up of the Victory of Samothrace ship, to test its
suitability for rowing Siren Vase fashion.
Ancient terminology
The third compromise is to accept that at least some ancient authors,

some of the time, used the same nomenclature as European languages do
today. That is not a new idea. The English classicist, Anderson, suggested
that an 'eight' and a 'sixteen' might have had the same oarage, differently
described (Anderson, 19 41, 323), and the American Admiral Rodgers
suggested (1937, 256) that the gigantic 'forty' had forty, not eighty, oarsmen
in cross-section.
Consider single-banked rowing. If the ancient trireme, named after the

number three, had six oarsmen in a room, then single-banked boats should
have been named for the number 'a half' (zero decimal five). But there is no
such rowing word in ancient Greek or Latin. The lowest number used to
describe a rowing system is 'one', which suggests that to describe a single-
banked boat the ancients used the same nomenclature as we do.
Now double-banked rowing. A Latin poet, Manilius, wrote about a man

swimming:
'nunc alterna ferens in lentos bracchia tractus... nunc aequore mersael
Diducet palmas furtiva biremis in ipso', which Goold translates:
'Now lifting one arm after the other to make slow sweeps... now like a'
hidden bireme he will draw apart his arms beneath the water'.
The second swimming action is the breast stroke. A man's arms
swimming breaststroke are like the oars of a double-banked boat. They give
no suggestion of a vessel with oars at two levels.
Now triple-banked rowing, as in Fig. 4 - 6. There was a small Greek

merchant vessel called a phaselos, which could in an emergency be
converted into a phaselos trieretikos. No one supposes that a phaselos
trieretikos had six banks of oars. It could well have had three, the conversion
consisting of adding a third bank of oars and oarsmen down the middle of
an originally double-banked phaselos.
Now consider four banks. There was an oared ship named tetreres after
the number four. It is accepted that the tetreres was the first warship to use
more than one man to each oar. The new system is likely to have been
introduced in its simplest possible form, with two two-man oars to a room.
Morrison assumes that the tetreres had eight men pulling four two-man oars
in each room, but that is improbably complicated for a new system. The
question is obscured by the fact that Morrison wrote:
'The tetreres could then have had four men to each "room" ... rowing two
men to each of two oars' (Morrison and Williams, 1968,291). He meant eight
men rowing, with two men to each of four oars. Errors of that sort make the
subject unnecessarily difficult.
Now five banks. Alexander the Great was criticised for ostentation
because his barge (keletes) was rowed in a manner designated by the
number five (Ephippos, apud Athenaeum, viii. 38, cited in Torr, 1894, 109).
Nobody believes that it had ten banks of oars. It could well have had five.
If it is assumed that in ancient Greek and Latin there was the same
ambiguity that we have experienced in English at this conference, then much
of the linguistic evidence can be accepted without renouncing faith in six-
banked triremes.
Oars at two levels

The fourth proposed compromise concerns representations of ships with
oars at two levels, like this (Fig. 9) Phoenician warship of about 700 BC.
The origin of the trireme has been summed up as follows:

'Thucydides says that the Corinthians were the first Greeks to build
triremes, c. 700 [BC]. We may perhaps guess that the tradition cited by
Clement of Alexandria that the Sidonians invented them is true' (Harden,
1962, 125).
The Sidonians were Phoenicians. One should therefore expect the first
trireme in the iconography to be a Phoenician of around 700 BC, such as this
(Fig. 9). But to sustain the three-level trireme theory, it is better to call a ship
like that in Fig. 9 a bireme (even though there is no mention by ancient
authors of a dieres until the first century AD) and to imagine that the first
ALEC F. TILLEY TROPlS VI
triremes came later than 700 BC and were Greek. So in Greek oared ships
(p. 158), Morrison's translation of Thucydides omits altogether the qualifying
words tes Hellados, and makes it appear that the Corinthians were the first
absolutely, a mistake one would not expect to find outside the trireme
controversy. Later, Morrison supported the idea that Thucydides was wrong
in his chronology and that the invention of the trireme could be put at about
650 BC, thus excluding the possibility that a ship of 700 BC could be a
trireme (Morrison and Coates, 1986, 39-40).
But Morrison's latest position is that the ship in Fig. 9 may indeed be a
trireme. He has in mind a ship built for oars at three levels, but only using two
levels (Gardiner (ed.), 1995, 54-5).
Morrison as consultant editor congratulates Morrison as author on the

originality of his contribution, although Morrison as referee has been
resisting for more than twenty-five years the idea that the ship in Fig. 9 was
a trireme. The important thing, though, is that his faith in six-banked triremes
no longer requires the mistranslation of Thucydides or the elaboration of
special pleading intended to show that Thucydides' chronology was wrong.
Substantially more of the evidence is accepted and the argumentation
becomes appreciably less fanciful.
The idea could be extended to cover other two-level ships which the
evidence indicates are triremes. In the Olympias the lowest level of rowers is
'almost wholly ineffective' (Shaw (ed.), 1993, 62) and 'not worth its place in
the ship' (Coates, 1988,77). If we assume that ancient seamen had the same
experience, they would soon have discontinued using the lowest-level oars,
and might well have boarded up the oarports, as English captains boarded
up the oar-ports of their frigates in the 18Ihcentury. Thus all the two-level
ships, which the evidence requires to be triremes, could be accepted as
such by everyone.
The acceptance of that idea would open the way to the next and fifth
compromise: to renounce pseudo-evidence. Several ship representations,
which do not show oars at three levels, have been 'improved' with imaginary
additions or distortions and presented as pseudo-evidence in favour of
three-level triremes.
This (Fig. 10) is the sherd known as the Vienna fragment. It is not very
impressive in itself, but it illustrates the use of pseudo-evidence most clearly.
The rowing arrangement that the artist has actually shown consists of three
semi-circular oar-ports at two levels. Morrison and Coates see it as evidence
of a warship with three levels of oars, like the Olympias. The Olympias
uppermost row of oars is suggested to them by a line of thole pins which
through 'rough drawing' the artist has entirely neglected to depict. They
interpret the lower-level oarport as a large, circular, Olympias-type oarport,
transmuted by the same 'rough drawing' into the equal-sized, semi-circular
oarport that we can actually see (Morrison and Coates, 1986, 150). Thus
their interpretation is amazingly like the Olympias. And thus the study of
ancient warships is reduced from science to crystal gazing.
This drawing (Fig. 11) was made in the seventeenth century AD. It shows
a ship with oars at two levels. It is used by Morrison & Coates as evidence
that triremes had oars at three levels, by first imagining that it is a copy made
by an ignorant artist of a three-level original and then assuming that the
imaginary three-level original was a trireme (Morrison and Coates, 1986,
142). It would be no less scientific to divine the nature of ancient triremes by
examining the entrails of chickens.
This (Fig. 12) is the well-known Lenormant relief. What the sculptor has
actually shown is a single row of oars and oarsmen. The other oblique and
horizontal lines resemble the side of the sixteenth-century galeasse
illustrated in Fig. 13.
To transform what we actually see into something resembling the

Olympias, Morrison assumes ancient paint (of which there is now no trace)
to run the oblique features across the horizontal ones so that they can be
interpreted as a second and third level of oars (Morrison and Williams, 1968,
171). The undisputed oars do cross the horizontal wales. The other features
do not.
Basch rightly points out that even with the addition of ancient paint, the
features would not form straight lines. He therefore proposes (Basch, 1988,
178) -, an ancient original of which Fig. 12 is an inaccurate copy by an
ignorant artist - just what Morrison and Coates proposed for the 17"'-
century drawing (Fig. 11).
It has been remarked that the oars could not have been parallel to each
other at the point in the stroke shown in Fig. 12. One is asked to presume
ALEC F. TILLEY TROPE VI
that the Lenormant sculptor really saw oars like this (Fig. 14) but carved what
he actually did carve 'for the sake of art' (Shaw (ed.), 1993, 1). Coates
observes that there is not enough space between the wales on the actual
sculpture for oarports as large as the lowest ones on the Olympias (Morrison
and Coates, 1986, 234). We are to assume, and mentally correct, an error on
the part of the sculptor, not in the design of the Olympias.
These heroic efforts to make the evidence fit the theory transform the
actual relief into something exactly like the Olympias; but in science, theory
must be adjusted to suit evidence. If we allow ourselves to argue from
supposed ancient paint of which there is now no trace, or from supposed
originals of which the actual monuments are supposed to be erroneous
copies, then we will be able to find ample evidence for the proposition that
ancient pigs had wings.
If one were to extend Morrison's new idea (that the ships with oars at two
levels on the wall relief from the palace of Sennacherib were really ships
designed for three levels but with one level out of use) to the many
representations of two-level ships that appear in the iconography just at the
time when triremes were coming into use, then faith in three-level triremes
would not need pseudo-evidence to support it.
Three classes of oarsmen
The sixth and last compromise proposed in this paper concerns the
seating of the three classes of oarsmen in triremes. Before the trials of the
Olympias, the evidence that the three classes of oarsmen in a trireme sat
forward, amidships and aft was generally accepted, the thranites furthest aft
and the thalamians furthest forward (Morrison, 1941, 20). The original
arrangement in the Olympias could, with a measure of goodwill, be said to
conform. The three classes were assigned to the three levels, the
thalamanians at the bottom. On either side, three rowers, one from each
level, were regarded as a 'triad', the uppermost furthest aft and the lowest
furthest forward. But it was found impossible to keep time with that
arrangement, and the triads were reformed with the thalamians furthest aft,
where they could be seen by the others of the triad. This discrepancy with
the ancient evidence is never remarked upon in pro-Olympias literature, and
the fore-and-aft evidence is no longer mentioned.
In the Olympias, when the thalamanians row alone, they change places
with rowers of a higher level. That is understandable. The lowest level is

hotter and more airless than the others, and the rowers cannot see their oars
In ancient triremes, when only one of the three classes rowed, the oarsmen
stayed put. That is evidence that in ancient triremes the three classes of
.oarsmen were not at three different levels.
To insist, against the evidence, on different levels for the three different
classes, is certainly not essential to faith in three-level triremes. Seating the
three classes of oarsmen forward, amidships and aft is suitable for any oared
warship, including a three-level one.
Conclusion
The unscientific methods of argument used in support of three-level

triremes have a harmful effect that goes far beyond the trireme problem: they
infect the whole subject of ancient ships, so that outsiders are apt to regard
it as something (like extra-sensory perception or the Bermuda triangle) to be
avoided by scholars jealous of their intellectual reputations. Ignoring or
lampooning the idea of rowing with three banks of oars makes it difficult to
gain support for trials. Trials of the Olympias have shown that it is impossible
to row efficiently with six oarsmen in cross-section with oars of the two
lengths used in Athenian triremes. But, irrespective of conclusions, there is
considerable merit in compromises, which reduce the need for unscientific
argumentation.
Alec F. Tilley
Fieldfare
Hambledon
Hampshire PO7 4RX
England
REFERENCES
Anderson, R.C. 1941. Triremes and other ancient galleys. The Mariner's Mirror 27: 314-23.
Basch, L. 1987. Le Musee imaginaire de la marine antique. Athens.
Basch, L. 1988. The Eleusis museum trireme. The Mariner's Mirror 74: 163-97.
Coates, J.F. 1995. Tilley's and Morrison's triremes. Antiquity 69: 159-62.
Coates, J.F. etal. 1990. The Trireme Trials 1988. Oxford.
Gardiner, R. (ed.) 1995. The Age of the Galley. London.
Harden, D. 1962. The Phoenicians. London.
Landstrom, B. 1961. The Ship. London.

Morrison, J.S. 1941. The Greek trireme. The Mariner's Mirror 27: 14-44.
Morrison, J.S. 1969. Referee's report to the editor of Antiquity.
Morrison, J.S. 1978. Rowing the trireme. The Mariner's Mirror 64: 203-8.
Morrison, J.S. 1980. Hemiolia, trihemiolia. International Journal of Nautical Archaeology 9: 121-6.
Morrison, J.S. and J.F. Coates. 1986. The Athenian Trireme. Cambridge.
Morrison, J.S. and R.T. Williams. 1968. Greek Oared Ships 900-322 B.C. Cambridge.
Rodgers, W.L. 1937. Greek and Roman Naval Warfare. Annapolis.
Shaw, T . (ed.) 1993. The Trireme Project. Oxford.
Torr, C. 1894. Ancient Ships. Cambridge.
Tilley, A.F. 1995. Warships of the ancient Mediterranean. Tropis 111. Athens.
THE NUMBERS 1N THE NAMES OF ANCIENT WARSHIPS:
_ SOME PROPOSED COMPROMISES
--
- . -- -
*.Ii:;r,:.4.;.. b=.>.-.
-,
1
,
'Y-,.?
?>'.
i,;), ; -.
-:-
--.--
'I,
+-\
--.
Fig. 1 Fig. 2 Fig. 3
i- ~ q 4.
;- -- K,,, !:.;>*:.,-,. . . bT!.l+;,/
A,
d</;-
--
-.._
-.-. . ,
. . , ' A*!,,
: : 8.- : . %
I ;&.. :<my---
--..
, - .
<.
-:
-- ' . "
- - . _ -, . . ,. ,,
- L
.:-!?-->::
>
k '
h-,,! . L
,
,> .=- -2---
.-i
-.
Fig. 7
.A.
,, , , . Fig. 5
. , - Fig. 8
I Fig. 9
Fig. 13 Fig. 14
A TRIREME ON A FUNERARY LEKYTHOS
Introduction
The earliest known literary evidence on triremes1appears in a sentence

by Hipponax2,a sixth-century BC poet from Ephesos. Later evidence comes
from Herodotus and Thucydides, the historians of the second half of the 5th
century BC. From these sources, it is evident that triremes appeared in
eastern Mediterranean between 540 and 525 BC3. However, the earliest
known representation of a trireme is dated to the middle of the !jth
century
BC4.
The funerary lekythos in the Athens National Museum (fig. 1)
The monument presented here is an unpublished lekythos of pentelic

marble in the Athens National Museum No. 9167, which has a max. diameter
of 38cm and a total height of 73.5cm. In a temple-like frame, 26cm wide and
30cm high, the starboard bow of a trireme, with a hoplite standing on its
deck, is depicted. The hoplite is out of proportion in comparison to the size
of the ship. The ship, which is moving to the right, is shown from the lower
wale upwards. The surface of the sea is not shown.
The main features of the ship are given with parallel bands of varying
thickness. The lower band corresponds to the heavier lower wale. At its fore
end there is a ram with two blades. The second intruding band, just above
the lower wale, corresponds to the ship's hull. The extruding third band,
corresponding to the upper wale, is extended beyond the cutwater, forming
the fore-ram. The fourth, thin, intruding band corresponds to the upper hull
of the ship. The distance between the two wales is bigger than the one
between the upper wale and the lower timber of the outrigger.
The outrigger is depicted with two parallel timbers, connected with short
uprights. To the right of the outrigger, the rectangular epotis or ear timber is
shown projecting laterally. The upper timber of the outrigger is extended
toward the stem, which projects upwards, forming an S.
The apotropaic eye is rendered in high relief. Its representation suggests

EVANGELOS E. TZAHOS TROPIS VI
that this eye was not painted but rather made of marble and mounted on the
ship's hull. We know that there were two eyes on each side of the ship. The
second eye was located between the two wales, and was usually painted on
the hull of the ship. The presence of the second eye is clearly depicted on a
Hellenistic rhyton (D 201) in the British Museum and on a coin of Demetrius
Poliorketes.
There are three oars visible on the left side of the prow. They are parallel
to each other, directed from the upper right to the lower left. The lowest oar
emerges below the upper wale, whereas the other two emerge between the
timbers of the outrigger. Because of the restricted space in the bow, the first
oar, corresponding to the zygites, is not shown. This oar arrangement fits the
oar distribution thranites: zygites: thalamites = 31: 27: 27, attested to by the
naval inventories found in Piraeus, and by the reconstructed trireme
"Olympias".
The deck of the ship is supported on curved stanchions, which descend

down to the upper timber of the outrigger. Their continuation beyond this
timber is not shown. The area below the foredeck is boxed in, with a solid
parapet at its sides and its aft end is curved.
On the deck, there is a standing hoplite extending his left leg forwards.
With his left hand, he carries a shield, while with his right hand, he holds his
weapon, a sword or a lance. The weapon was probably painted, like all the
other details of the depiction. The warrior wears a corselet and an Attic
helmet.
The depiction of a hero who was killed in a sea battle is very rare for the
period under consideration. The only example I am aware of belonging to
the same period is the funerary stele of Demetrius (Glyptothek Muenchen
No. GI 522).
The funerary stele of Demokleides in the National Museum of Athens No.

752 depicts the hero sitting on the deck of the ship. There are also other
examples from the Hellenistic period, like the relief in the museum of Delos
No. A7245, representing Timokrates with his sword and shield on the deck
of a ship, and the relief in the museum of Paros No. 1 which represents
Difilos on the bow of a ship.
The lekythoi, which have an oval rippled body and a representation

within a temple-like frame have been dated to the second half of the fourth
century BC by Avgi Proukaki5. However, with reference to the historical

events of this period, I will try to date this monument more precisely.
After the victory of Philip the II over the Athenians on the battlefield of
Chaeronia, there was a peace treaty signed in 338 BC. The sea battles
between the Athenians and the Macedonians resumed only after the death
of Alexander the Great. The sea battles at Abydos and Amorgos took place
in 322 BC6.Accordingly, our lekythos was probably produced either in the
first decade of the second half of the 4'" century or near 322 BC. However,
considering also its oval shape, it seems that the later date is more probable.
Known trireme depictions
I shall now give a brief description of known trireme representations and

compare them to our funerary lekythos. The other known trireme
representations are:
1. The Acropolis relief No. 1339 and parts of the same monument7
la. The votive relief in the Acropolis Museum, No. 1339, known as the
Lenormant relief, which portrays a midship section of the starboard side
of a trireme, with oarsmen pulling their oars. It is dated to the last decade
of the 5'h century BC (410 BC). (fig. 2)
1b. A relief fragment in the Acropolis Museum No. 2544, which depicts the
upper part of a youth, considered to be part of the Lenormant relief.
Ic. Another relief fragment, originally in the Athens National Museum No.
5240 and now in the Acropolis Museum No. 16479, which depicts an
oarsman, is also considered to be part of the Lenormant relief.
On the deck of the Acropolis relief there are human figures sitting or
reclined. Supporting the deck, which is narrower than the overall beam,
there are successive aft curving stanchions. The outrigger consists of two
heavy horizontal timbers, connected with short uprights (stiles and
tholepins) and seems to be in higher relief than the rest. The outrigger is
supported by brackets, which rest on the lower wale. Nine oarsmen,
identified as thranites, are shown pulling their oars, through the timbers of
the outrigger. Immediately under the structure of the outrigger emerge the
oars of the zygites. The thalamian oars emerge above the lower wale,
EVANGELOS E. TZAHOS
probably through portholes, originally painted and fitted with leather sleeves,
the askomata, to prevent the entrance of water.
Note that the lower wale of the Lenormant relief is slightly heavier than
the upper one. The relief illustrates the relative position of the oarsmen.
According to the naval inventories, the thranites were 31 and the zygites and
the thalamites 27, that is 85 on each side, making a total of 170 oarsmen
altogether. The distance between two tholepins, according to Vitruvius, was
equal to two cubits.
2. The Eleusis relief9(fig. 3)
A votive relief in the Eleusis Museum No. 5255 depicts the midship
section of the port side of a trireme with oarsmen pulling their oars. It is dated
around 350 BC.
On the deck of the ship, there are human figures either sitting or reclined.
The stanchions supporting the deck are almost vertical. The outrigger
consists of two heavy timbers, of higher relief than the rest, connected with
short uprights. There are 11 oarsmen, the thranites, pulling their oars
through the timbers of the outrigger. Between these oars, pairs of oars
emerge below the lower timber of the outrigger, on a lower level. These oars
apparently belong to the zygites and the thalamites. There are no wales to
be seen.
3. The Demokleides stelei0
The funerary stele of Demokleides, son of Demetrius, in the Nat.

Archaeological Museum of Athens No. 752 depicts in outline form the port
side of a bow with a stem curved forward and a warrior seating on the deck
with shield and helmet. It is dated to the beginning of the 4ih century BC.
4. The Demetrius stele" (fig. 4)
The funerary stele of Demetrius from Panormos, now in the Glyptothek

of Munich Nr GL. 522, depicts the starboard bow of a trireme with an armed
warrior on its deck. The stele is dated to the middle or the third quarter of the
4ih cent. BC.
The stem of the trireme on the Demetrius stele, considering also the
trace of the missing part, forms a right angle. The deck, which is narrower
A TRIREME ON A FUNERARY LEKMHOS
than the overall beam, is supported on stanchions curved from lower right to
upper left. The space below the foredeck is boxed in with a solid parapet,
and its aft end is curved.
The outrigger consists of two thin timbers, of which the upper one
extends up to the stem. The deck and foredeck are narrower than the
outrigger. At the right end of the outrigger a rectangular block protruding
laterally forms the epotis. The lower wale is heavier than the upper one and
it probably ends in a two-bladed ram. The surface of the sea, the oars and
oarsmen are not depicted.
5. The Vienna fragmentl'(fig. 5)
A fragment of a red-figure vase, in the University of Vienna No. 503.48,

depicts part of the side of a trireme without oars or oarsmen. This fragment
is dated to the middle of the 51h century BC. The stanchions of the Vienna
fragment, which support the railed deck, are curved from lower left to upper
right. Again, the width of the deck is less than the overall beam. The
horizontal timbers just below the curved stanchions without uprights
correspond to the outrigger. There are no tholes, stiles and supporting
brackets depicted.
Between the two pairs of horizontal timbers, which constitute the upper
and lower wales, there are 2 oar ports for the zygites and one for the
thalamite. However there are no oars or oarsmen.
6. The Red-figure Krater13(fig. 6)
The red-figure krater in the Jatta collection No. 1501 in Ruvo represents
the port side of the stern of a trireme with three human figures. It is dated
around 400 BC.
The upper part of the ship on this krater is depicted with outstanding
craftmanship. The deck is supported by stanchions with an exaggerated
double curvature, the lower ends of which rest on the upper wale.
Considering the position of the right arm of the man ascending the ladder,
one can conclude that the deck is narrower than the overall beam. The
outrigger is supported by a bracket resting on a lower wale. The two timbers
of the outrigger are connected with vertical uprights. There are no oarsmen
or oars. On this krater, the man on the ladder conceals part of the stern.
Therefore, it is impossible to tell whether the thranite tholes or the zygian
EVANGELOS E. TZAHOS TROPlS VI
oarports continue aft. If two thranite tholes and two zygian ports are
concealed, then the arrangement proposed by J. Morrison and J. Coates
would fit the distribution between the classes of oarsmen attested by the
naval inventories.
7. The Dal Pozzo drawingI4
The drawing made by Cavalier0 dal Pozzo at the beginning of the 1 7 ~ ~

century and now in the British Museum shows the starboard bow of a trireme
with oarsmen pulling oars. Again, the deck is narrower than the overall
beam. The foredeck continues forward to the stem. The space below the
foredeck is boxed in with solid parapets. The foredeck does not extend
laterally over to the thranite oarsmen. The thalamian and the thranite oars are
shown correctly. The ram and the ear timber are not shown.
8. The Acropolis fragment No. 13533''
In addition to the above well known depictions I will mention now the
fragment of a votive relief in the Acropolis Museum No. 13533, which depicts
the starboard side of a bow. It is dated between 350 and 300 BC. This
fragment, although presented during the temporary exhibitions in Athens
and Lisbon in 1987, was exhibited unidentified. On this relief, I recognise the
following features particular to a trireme: The lower timber of the outrigger
continues foward below and beyond the epotis. The upper timber of the
outrigger continues forward to the stem and is curved upward. There are two
wales shown: the lower being heavier than the upper. The space below the
foredeck, is boxed in with a parapet and at its aft end there is a curved
stanchion. The epotis is rectangular but its fore face is inclined. Again, there
are no oars or oarsmen. Also shown is the wavy surface of the sea.
Comparison of the funerary lekythos with the above representations
1. the bow and the ram
The bow of a trireme is depicted on four of the above-cited funerary

monuments (3,4, 7, 8). On the Demokleides stele (3), the wales, the hull, the
outrigger and the oars are not depicted. They were probably painted on the
white marble surface. However, the outline of the bow, the ram, the fore-ram,
and the deck are given in the same way as on our lekythos and the
Demetrius relief. On the relief of Demetrius the bow is similar to our lekythos
and the ram and fore-ram are depicted the same way. However the fore end
of the lower wale is curved downwards.
The form of the prow of the Dal P o u o drawing (7) and the relief of the
Acropolis Museum No. 13533 (8) are similar to the bow of our lekythos.
The stem of the Demokleides stele (3) is curved fore, while the stem of
the Demetrius stele (4), considering the trace of the missing part, forms a
right angle. The form of the stem of the Dal Pozzo drawing is similar to the
one of our lekythos.
The ram in three of the four cases has two blades only. In our lekythos
and the Demokleides stele the two-bladed ram is more evident. In the case
of the Dal Pozzo drawing there is no ram shown. It seems that the structurally
advanced three-bladed ram of AthlitI6, of the rhyton in the British Museum'',
of the coin of Demetrius Poliorketes, of the stele of Diphilos in the museum
of Paros is a later innovation.
The apotropaic eye is depicted only on our lekythos. The eye shown in
high relief is located fore of the epotis. The way of representation suggests
that this eye was not painted but was made of marble and mounted on the
ships' hull. We know that there were two eyes on each side of the ship. The
second was placed lower between the wales; it was smaller and was usually
painted on the hull. This eye, although not shown on our lekythos, was
probably painted on the marble surface. The presence of the second eye is
clearly depicted on the Hellenistic Rhyton of the British Museum and on a
coin of Demetrius Poliorketes.
2. the deck with its supports and the foredeck
The deck is depicted on all the cited monuments, and is supported on

curved stanchions. The stanchions of our lekythos and those of the Vienna
fragment (5) are curved from lower left to upper right, whereas the
stanchions of the Acropolis fragments (la, lc), of the Dal P o u o drawing (7)
and those of the Demetrius (4) stele are curved from right to left. The
stanchions of the Eleusis relief (2) are almost vertical.
On the deck of the monuments (1, 2, 3, 4, 6 , 7) there are human figures
standing, sitting or reclined.
The Demetrius stele (4), the Dal P o u o drawing (7), the Acropolis
fragment No. 13533 (8) and our lekythos depict a parapet, which boxes in
the space below the foredeck. The end of this space is curved aft on all
monuments.
In the case of the red-figure krater (6) the deck is supported by

stanchions with an exaggerated double curvature, the lower ends of which
rest on the upper wale.
In all cases the deck is narrower than the overall beam.
3. the outrigger-parexeiresia and the thranites
The outrigger is clearly shown in most of the examples cited. On the

Acropolis relief (la) nine oarsmen, the thranites, are shown pulling their oars,
which pass through two heavy horizontal timbers, connected with short
uprights (stiles and tholepins) and seem to be in higher relief than the rest.
In the Dal Pozzo drawing (7) the outrigger consists of two horizontal timbers,
the lower of which is heavier. Five oarsmen identified as thranites pull their
oars in the same position as those of the Lenormant relief.
The Eleusis relief (2) shows 11 oarsmen, the thranites, who pull their oars
through the timbers, connected by short uprights. On the Vienna fragment
(5) the horizontal timbers just below the curved stanchions without uprights
correspond to the outrigger. However there are no oars or oarsmen. On the
Demetrius stele (4) the outrigger consists of two thin timbers, of which the
upper one extends up to the stem. At the right end of the outrigger a
rectangular block protruding laterally forms the epotis. On the Ruvo krater
(6) the outrigger is supported by a bracket resting on a lower wale. The two
timbers are connected with vertical uprights. No oarsmen or oars are st~own.
On the Acropolis relief No. 13533 (8) the lower timber of the outrigger
extends beyond the epotis and probably rests on the upper wale with short
uprights. The upper timber of the outrigger extends beyond epotis towards
the stem. The epotis of the same relief is rectangular but its fore face is
inclined. Again, there are no oars or oarsmen.
On our lekythos, the outrigger consists of two horizontal timbers

connected with short uprights. At the right end of the outrigger there is a
rectangular epotis. Two oars, identified as those of the thranites, pass
through the timbers of the outrigger, extending from upper right to lower left.
They are parallel to each other and their extension above the outrigger is not
shown. The supports of the outrigger are also not shown.
4. the wales and the lower oar files
The oar files of the triremes corresponding to the zygites and thalamites
usually pass through oarports. In the case of the Acropolis relief (1)the oars
of the zygites and the oars of the thalamites are clearly shown. The zygian
oars emerge under the structure of the outrigger, probably through ports,
which are not shown. The thalamian oars emerge above the lower wale,
again through ports, which are visible. The Dal Pozzo drawing (7) shows
correctly the thalamian but incorrectly the zygian oars. The latter should be
continuing upwards across the lower and upper wales, up to the lower
timber of the outrigger. In both cases the lower wale is slightly heavier than
the upper wale. On the Eleusis relief (2),between the oars of the thranites,
there are pairs of oars, emerging under the lower timber of the outrigger.
These oars belong apparently to the zygites and the thalamites. On the stele
of Demetrius (4) the lower wale is heavier than the upper one and there are
no oars or oar ports.
On the Vienna fragment (5), between two pairs of horizontal timbers,

which constitute the upper and the lower wales, there are oar ports for the
zygites and the thalamites. However, there are no oars shown. In the case of
the Acropolis fragment No. 13533 (8), the lower wale is heavier than the
upper one; however there are no oars or oar ports.
On the red-figure krater (6),the man on the ladder conceals part of the
stern. Therefore, it is impossible to tell whether the thranite tholes or the
zygian oar ports continue aft.
If two thranite tholes and two zygian ports are concealed, then the
arrangement proposed by J. Morrison and J. Coates would fit the
distribution between the classes of oarsmen attested by the naval
inventories.
In our lekythos the lower ale is extremely heavier than the upper one.
Because of the restricted space in the bow area, the first oar corresponding
to the file of the zygites is not shown. The first oar of the file of thalamites
appears to be in accord with the arrangement of the Ruvo krater.
Conclusions
The funerary lekythos of the National Museum is unique and significant

because, for the first time, we have a representation of a bow in which the
oars are also shown. Its features offer support to a number of previous
hypotheses, while bringing into question certain others. In addition some
features suggest entirely new hypotheses.
1. Conclusions concerning the form of the stem and the ram
The stem of our lekythos forms an S, while the stem of the Damokleides
stele is curved and that of the Demetrius stele forms a right angle.
Contrary to the general assumption the ram of our lekythos has two
blades only. I also recognise that the rams of the Demokleides and
Demetrius steles have two blades each. It seems that the three blades of the
Athlit ram, of the rhyton in the British Museum, and of the coin of Demetrius,
is a later innovation. Apparently the number of the blades depended on the
size of the ram, its weight and the strength of the lower wale.
2. Conclusions concerning the width of the deck, the curving of the

stanchions and the existence of a protected cabin below the foredeck
All the known fragments have a deck which is narrower than the overall
beam. This is more obvious on our lekythos. The space below the foredeck
of the Demetrius stele, the Acropolis fragment No. 13533, the Dal P o u o
drawing and our lekythos seem to have a solid parapet on each side. This
could mean that there was a protected structure for the officers, which was
narrower than the overall beam.
The curved stanchions used to support the deck continue aft forming the
interface between the deck and the foredeck.
The inward curving of the stanchions would facilitate the use of curtains,
which could be rolled down easily to protect the oarsmen from projectiles.
3. Conclusions concerning the outrigger, the epotis and the apotropaic

eye
The outrigger of the lekythos corresponds to the outrigger of the other

known monuments but the uprights beyond the lower timber are not shown.
The epotis has a rectangular shape and the apotropaic eye is unique in
its style, suggesting a mounted relief. There were two eyes on each side of
the ship.
4. Conclusions concerning the oar arrangement
The two oars of the thranites and the single thalamite oar correspond to
the arrangement proposed by J. Morrison and J. Coates for the Ruvo krater,
the reconstructed trireme Olympias and by the naval inventories. Therefore,
the absence of the first zygite oar on our lekythos is justified.
5. Conclusions concerning the lower wale
The lower wale in the case of the Demetrius stele, the Ruvo krater, the
Acropolis fragment No. 13533,the Dal P o u o drawing, and the lekythos in
the Athens National Museum is heavier than the upper wale. Although this
development was attributed to the ships of the Hellenistic period such as the
ship of the Victory of Samothrace and the ship of the lsola Tiberina
monument in Rome, it is obvious that triremes had broader and heavier
lower wales, at least in the Late Classical period.
The heavier lower wale became necessary to increase the defensive

capability of the triremes against the ramming of bigger ships like fours and
fives, which were sailing in the post-Classical period.
Dr. Eng. Evangelos E. Tzahos

26 Athinas Street
Voula 16673
Greece
ABBREVIATIONS
AT: J. S. Morrison, J. F. Coates, The Athenian Trireme, Cambridge 1986

GROW: J. S. Morrison, J. F. Coates, Greek and Roman Oared Warships 399-30 BC, Oxbow
Monograph, Oxford 1996
AG: R. Gardiner (Ed.), The Age of the Galley, London 1995
GOS: J. S. Morrison, R. T. Williams, Greek Oared Ships, Cambridge 1968
NOTES
1. AG, p. 49-65.
2. l m h v a ~lapl lo^ F 28 W: 103.
3. Walllinga H. T., Ships and Sea-Power before the Great Persian War, The Ancestry of the
Ancient Trireme, E. J.Brill, Leiden 1993, p.104.
EVANGELOS E. TZAHOS
4. AT, p. 47.
5. Avgi Maria Proukakis, The Evolution of Attic marble Lekythoi and their relation to the
problem of identifying the dead among the figures shown on the funerary reliefs
(Dissertation London 1971), p. 45 and 53.
6. Diodorus Siculus, 18.1.5.9.
7. AT, p. 15-16, 111.13.
8. AT, p. 17, 111.14.
9. T<&ou-AA&Cavt5pfi 0.- Zna8apq E.: TaCdcu3vraq p& TO d o i o q q Kupfiv~taq,ABjvat
1987, p. 84, No. 6C.
10. IG 112 11114; AHMOKAEIAHZ AHMHTPIOY.
11. Vierneisel-Schloerb, Glyptothek Muenchen, Katalog der Skulpturen, Band Ill, Klassische
Grabdenkmaeler und Votivreliefs, Munich 1988, p. 59-64, 111.24.
12. GOS, p. 169.
13. GROW, p. 187-88, 111.10.
14. AT, p. 13, 111.11.
15. T<axou-A&tav6pfi 0.- Znaeapll E: T a ~ d ~ u o v r ap&
q TO nhoio q q Kupfivctaq, ABjvat
1987, p. 85, 111.61.
16. AT, p. 130.
17. AT, p. 144.
ILLUSTRATIONS
1. Funerary lekythos of the National Archaeological Museum, Athens

2. The Acropolis Relief, No. 1339
3. The Demokleides Stele, No. 752, National Archaeological Museum, Athens
4. The Demetrius Stele, GI. 522, Glyptotek of Munich
5. The Vienna fragment
6. The red-figure krater
. . -- -- A TRIREME ON
-..-
A FUNERARY
.. - .- - - .
LEKYFHOS
- -- --
Fig. 1
587
EVANGELOS
--- - -E. -.
TZAHOS
- --- ---- - -- - - --
T R O P E VI-
Fig. 2
Fig. 5 Fig. 6
588
TWO NEW REPRESENTATIONS OF ANCIENT SHIPS FROM ATTICA
The lekythos of Koropi
In 1994 a labourer from Koropi, Attica, presented the Museum of Brauron

with a lekythos of white marble which according to his declaration had been
for generations exposed in his field'.
This lekythos of medium size is broken at the neck and the foot as such
monuments are most often found. Both extremities are missing and the
breakage is obviously ancient. The provenance is probably from a
secondary workshop of Attica.
The total height of the preserved body is 63cm and the maximum width
at the upper part (the shoulders) is 36cm. (Fig. 1)
A ship, represented on the belly of this lekythos, is the subject of the

present research.
Let us briefly say that lekythoi as well as loutrophoroi are white marble
carved vessels developed from earlier versions made of clay. They were
used mainly in Athens and Attica in the 5th c. and 4th c. BC as funerary
decorative objects placed in pairs or individually on graves, together with
marble stelae and ornamental statuary, called the &n~urrjpa~a.
The rich adornment of tombs belonging to wealthy and predominant

citizens was a funerary custom particular to Athens and the region of Attica,
but could also be found in other Greek cities. The fashion started in the
second half of the 5Ih c. B.C. but was a la mode mainly in the following
century. The edict of Demetrius Phalereus in 317 BC put an end to this
expensive custom (that was discriminating between rich and poor citizens
even after death), allowing only the placing, as grave marks, of modest
marble ~ i o v i o ~ othe
i , columellae, bearing usually the name and the demotic
of the deceased. So the lekythoi as part of the monumental tomb-adorning
practice were short-lived and in use for a little more than a century. There
are, however, known exceptions as in some remote areas, in private rural
properties, burials secretly continued to be adorned and decorated with
monumental sculptures even after the prohibition.
- The representation of a ship on a funerary lekythos is an unusual theme
- HARRY E. TZALAS TROPIS VI
and it must be stressed that ships and naval representations in general are
scarce on monuments of Classical and Early Hellenistic times. The majority
of depictions on such grave monuments show the scene of the 6&Siwuq,the
farewell scene, where the departed is often represented in low relief seated,
saluting close family members, friends or beloved pets. There are instances
where the representation characterises a particular activity of the deceased:
a military scene for an officer, an athletic event for an athlete, etc. There are
also representations that are mythological, related to the belief of the Greeks
in an after-life.
Typical is the scene of young Myrsine, on an Attic marble lekythos, work

of the BC, showing Hermes Psychopompos escorting the young girl, while
on another similar monument she is introduced to Hades.
Before trying to explain the ship's technical characteristics as

represented on the Koropi lekythos, let us attempt to interpret the presence
of this ship on a funerary monument. This can mean that the tomb belonged
to a mariner, a marine officer - perhaps a trierarch or a proreus - or to a
shipowner; there are examples of such themes. But if the ship was a stylized
representation of the boat of Charon, then the activity of the departed could
be totally alien to the ship and the sea.
By studying the ship itself we can try to get an answer to our above
question.
The represented ship, including the bow protrusion, has a total length of
46.5cm while its height is 17.5cm. We note that the hull is disproportionally
deep and the body unusually bulky, thus it cannot realistically be attributed
to a warship of that time.
In fact we have only the contour of the vessel carved in low relief from the
end of the keel to the upper part of the bulwark, excluding any
superstructure, mast and rigging, steering oars and oars. Other
indispensable structural parts are also missing, as the gunwale, the oarports,
etc.
As we know that it was usual to paint details on such marble lekythoi

representations, it is well possible to imagine that structural details,
equipment as well as decorative features, i.e. the ophthalmoi, were painted
with vivid colours that have faded and disappeared in the course of the
centuries.
TWO NEW REPRESENTATIONS OF ANCIENT SHIPS FROM ATTICA
Can we interpret the exaggeration in the body of the hull as an attempt

of the artist to harmonise the ship's profile with that of the lekythos? Was the
marble cutter sensitive to the horror vacui and was he trying to fill as much
as possible the half section surface of the monument that was usually
visible? This eventuality cannot be dismissed and may explain the distortion
of the ship's curves. In fact the stone carver made use of the maximum width
of the body of the lekythos at its wider part. The 46.5cm of the total length of
the ship from stern to the end of the bow's protrusion represents more than
half of the circumference of the lekythos at its wider part. It is obvious that
the artist's attention has been concentrated on the hull itself: that is what he
wanted to represent.
It is known that some of the lekythoi were cut and carved beforehand
and that the family of the departed could select a ready-made piece adding
only the name of the deceased and the features of his face when such a
representation was depicted. Is this the case of this ship or was this ship
depiction ordered and executed for that particular burial?
Looking into the details of the hull contour we see a nearly straight keel
at its lower end, then we have a curvature forward and aft, smoothly
connecting its ends to the stem and sternposts. The "aphlaston" has four
extensions (Fig. 2a), suggesting that they are the stern ends of the keel and
of the three wales, top (of the hull), middle and lower (at or just above the
water line). At the bow the lower three protrusions are certainly a ram that
has been rendered with great care (Fig 2b). In fact, the extremities of the ram
show a peculiar three-fold rendering: the lower is leaning downwards, the
middle and the upper blades are horizontal, but the middle is thicker than the
upper.
The upper part of the ship is delimited by what seems to be a straight

bulwark that extends from the base of the "aphlaston" to nearly the foreward
extremity of the bow. A few centimeters before the extremity, the bulwalk
ends abruptly and the contour marks what is the deck level, which in an
absolutely rectilinear manner extends forward, above the ram, to form a
strange protrusion.
What type of ship is represented on this lekythos and what was its
relation to the deceased? Is this a real ship naturalistically rendered or a
ritual vessel2?
There are instances where the boat of Charon is painted on funerary clay
HARRY E. TZALAS TROPlS VI
pottery or carved on funerary marble monuments. In the Iconographic

Section of the Lexicon lconographicum Mythologiae Classicae3there are no
less than 29 depictions of Charon and his boat on clay lekythoi, one on a
mosaic and 9 on stone reliefs. Charon's boat is represented usually as a
rowing or paddled craft where the ferryman of the underworld is paddling,
rowing or polling and collecting his toll.
One of the best known and most important representations of the boat
of Charon4is the funerary relief of Kerameikos. On this large stele Charon is
shown on his boat with the typical mariner's hat, while ashore three figures
are waiting their turn to be ferried on the Acheron. Our ship does not seem
to be such a boat as it bears no resemblance to the numerous
representations depicted on painted lekythoi or carved on stone. As far as is
known, with only two exceptions, Charon's boat is a small flat bottom
~ K ~ T L aOcraft
V , limited to the navigation on the Acheron, the river of the
afterworld, the reign of Hades. The exceptions are a painted lekythos of the
third quarter of the 5" c. BC in the Munich Museum, where Charon's ferry has
a warlike appearance with an "aphlaston" and a prow protuberance that
could be a ram; the other, in the Vatican Museum, is also a ship attributed to
Charon, sculpted on a cylindrical Roman marble altar dated to the 1"' c. BC5.
So the Charon boat should be excluded as none of its characteristics are

found on the ship of the Koropi lekythos.
It should be noted at this point that neither the "aphlaston", nor the well-
rendered ram, nor the round shape of the hull can be considered unusual
features. The bow protrusion is however unusual if not unique and we should
pay some attention to this feature.
This massive protrusion starts from the level of the fore deck and extends
from its end for 10cm. If we assume that the depicted ship was of the size of
a trireme, i.e. 37m, then this protrusion represents no less than a length of
10 meters. What can be this structural detail or piece of equipment that the
stone carver specifically wanted to show? If the lekythos was not dated to the
4'". BC but was representing a Late Roman vessel - which is not the case
- this could have been interpreted as a bowsprit or a beak6. However the
bowsprit and beak only appear centuries later on Roman ships.
This protrusion has no parallel with the exception of a little-known ship

painted on a krater in the Bologna Museum.(fig. 3) This is the ship of Phaon,
a masterpiece of classical painting attributed to the Polygnotos painter7(ARV
TWO NEW REPRESENTATIONS OF ANCIENT SHIPS FROM AlTICA
1056; No. 86 of the Museo Civico di Bologna).
Phaone is the handsome ferryman who, according to legend, showed

indifferenceto Sappho's passionate love for him and was responsible for her
suicide when she leaped from the promontory of Leucas and drowned in the
sea. The Bologna Krater shows precisely the moment when Sappho is about
to fall in the sea, while Phaon bearing the typical mariner's hat is paddling
his boat in an attempt to rescue her. Eros is gently landing on the foreward
deck from where an unusual protrusion is extending and intends to crown
Phaon9.Aphrodite is also about to step on the ship. It is obvious that the
protrusion in this instance is carefully shown as being lashed, which recalls
the sewing building technique, sutilus navis, similar to the Etruscan and local
Villanovan boats. This bow projection must be something peculiar to ferry
boats''.
In conclusion, we must admit that the ship represented on the Koropi

lekythos, although certainly not the work of a great artist, presents a specific
interest for the naval iconography of Classical and Hellenistic times.
It shows a vessel that combines the characteristics of a round hull - a

large merchantman - the attributes of a warship - the "aphlaston" and the
ram (being the parts that constituted the trophy should a warship be
captured) - and an unusual protrusion, perhaps a peculiarity of the ferries.
Such a ship, in our opinion, never existed, but perhaps the stone carver
was requested to decorate the tomb of a wealthy shipowner from Messogea
- perhaps the owner of a passenger-ferry or of a number of ferries - and
gave priority in his relief to the shape of a bulky commercial vessel, adding
that characteristic protrusion to denote a particular type of vessel owned by
the deceased.
Bearing in mind however that at that particular period of naval warfare it

is the warship that was considered outstanding and had a predominant
place in the iconography, the "aphlaston" and the ram were additions made
to enhance the depiction.
Of course, the above is only a hypothesis and can be refuted by later

iconographical finds.
The author would like to thank Clairy Efstratiou of the Museum of

Brauron for permitting the publication of this ship.
HARRY E. TZALAS TROPlS VI
The rock carving of Lavrium
The second iconographic document is a modest rock carving of a ship

on the mountainside of Souriza, in the Lavrium, about 3km at bird's eye
distance from the sea. The area called Souriza is scattered with important
remains of ancient metallurgic activities'.
This is not a graffito but a rock carving made on a flat surface of hard
rock by a stone-carver who used a chisel and a hammer with dexterity (Fig.
5). Not far away from this representation is a site with metallurgic activities
datable to Classical times. It is known however that by-products of early
metallurgic activities were processed until Late Roman times. A deep hull
ship is rendered very schematically. Only the features considered essential
are shown. (Fig. 6)
We see the contour of a hull that has an overall length of 32cm and a
height of 23cm. The deck is not rendered by a horizontal line but is concave,
with the foreward and aft parts rising. The prolongation of the pointed bow
is particularly stressed. A portion of a mast is shown placed towards the
stern and goes through the deck, nearly touching the lower extremity of the
ship in an "X-ray" depiction, certainly not unusual in ancient representations.
The absence of any sternpost extension and decoration as well as of any

bow protrusion that could be interpreted as a ram indicates that the
represented craft is a merchant ship. Its profile characteristics find a parallel
with several ship depictions of Roman times and can be dated to the 3rdc.
AD or later.
The rock-carver that made this ship was perhaps a slave working in a
metallurgic installation that was processing by-products of earlier activities.
He was, however, familiar with ships and had seen some out of the water,
and had a basic idea of their lines. He knew that the mast goes through the
deck and steps above the keel.
Not far from the site, about 4km off on the coast, half way between the
present port of Lavrium and the promontory of Cape Sounion, is a deep and
extremely well protected bay that has retained its medieval name of
Passalimani2.Nearby are the remains of the Agora of the Salaminians. There
was probably in the region an ancient shipyard that specialised in the lead
sheathing of aging hulls. A small piece of a lead sheath with two marks of
nails was found during excavations in the mid-80s at that precise spot. As
TWO NEW REPRESENTATIONS OF ANCIENT SHIPS FROM AlTICA
the Lavrium mines were famous for their lead production it should not be
considered as improbable that a yard, or more than one yard, in the area
specialised in the sheathing of small merchant vessels right on the site
where lead was produced and was readily available to be used.
Had the carver of the Souriza ship seen the sleek lines of the vessels
slipped in the aforementioned bay? The question will never be answered but
a methodical survey of the Passalimani bay followed by an excavation in the
shallows of its inner part will certainly provide precious information for what
may have been one of the important portuary installations in Attica.
Harry E. Tzalas
Hellenic Institute for the Preservation
of Nautical Tradition
Skra street 94
Kallithea, Athens
NOTES
The lekythos of Koropi
1. It is the same lekythos described by Alexander Conze, in Die Attischen Grabreliefs, Berlin-
Leipzig (1890-1922), vol. Ill. Referred to as lekythos no. 1324 it is said that in 1892 this white
marble artefact with a representation of a ship could be seen in the Municipality of Koropi;
earlier it was lying in the field of a certain Sotiris Andreas.
2. The profile of this ship recalls in some ways later Roman low relief carvings. An example is
the stele no. 1465 of Kosmetes Aurelius as well as stelae 1466 & 1468 in the National
Archaeological Museum, Athens, all showing a game of "naumachia".
3. For representations of Charon see LIMC, 111.1, p. 149, 111.2, pl. 168-174.
4. Ibid.
5. Ibid.
6. For the appearance of the beak instead of the ram, see Pryor and Jeffrey.
7. ARV 1056; No. 86 of the Museo Civico di Bologna.
8. For Phaon representations see LIMC, VII.l & V11.2, text 364-367, Fig. pl. 317-319.
9. 1 am grateful to Dott.ssa Govi Morigi, director of the Museo Civico di Bologna for allowing the
photographing of the Phaon lekythos.
10.1am indebted to Dr. Marco Bonino for his three-dimensional interpretation of the Phaon ship
(Fig. 4).
The rock carving of Lavrium
1.1would like to thank Dr. Evangelos Kakavoyiannis, the excavator of several

of the Lavrium sites with metallurgic activities, for indicating this
representation and permitting its publication.
HARRY E. TZALAS TROPIS VI
2. The author reported in a communication made in 1994 at the Sib Scientific

Encounter of Southeastern Attica that there are in the inner part of the bay
of Passalimani - now called Possidonia - important submerged portuary
constructions that need to be surveyed. See Harry E. Tzalas <<5* scientific
encounter southeastern Attica, proceedings Paiania (1994), pp 281-
~ lEnio-qpovl~fiq
2 8 8 ~ , ( n p a ~5"$ ~6 ~u<fiTrptlq
N. A. A n i ~ f i qIlalavia
, 1994,
USA. 281-288).
ILLUSTRATIONS
1.Drawing of the Koropi ship made in 1892 by Alexander Conze, in Die Attischen Grabreliefs
(1890-1922), Berlin-Leipzig, vol. Ill.
2. The marble lekythos of Koropi (photos by the author).
a) The prow of the ship of Koropi b) and the bow.
3. The ship of Phaon. Courtesy of the Museo Civico Archeologico, Bologna.
4. Structural interpretation of the Phaon ship by Dr. Marco Bonino.
5. Ship Graffiti, Lavreotic-Peninsula, Attica (drawing by the author).
6. Photograph of the Ship Graffiti of Lavreotic-Peninsula (photo by the author).
OGRAPHY
Kokula, G. Marmorlutrophoren (Diss.), AM Beiheft 1-184.

Kurtz, Donna C. & John Boardman. Greek Burial Customs, Thames & Hudson, London 1971.
Picard, Charles. Manuel d'Archeologie Grecque. La Sculpture, Paris 1963.
Prukaki-Christodoulopoulou A. Einige Marmorlekythen, AM 85, 1970, 54-99.
Schmartz, B. Griechische Gravereliefs (Ertraege der Forschung Bd., 192) Wissenschaft,
Buchgeeneinraft, Darmstadt 1993.
TWO NEW REPRESENTATIONS_ OF ANCIENT SHIPS FROM AlTICA
Fig. 1
Fig. 2
Fig. 2a Fig. 2b
597
HARRY E. TZAtAS
- -- --- - - -
TROPlS
- .VI
Fig. 4
-- - - - TWO .NEW REPRESENTATIONS
-. . -. . -- - SHIPS FROM ATTICA
OF ANCIENT -
Measurements in cms. Fig. 5
Fig. 6
599
THE INAICMS JOINT EXPEDITION TO TANTURA LAGOON, ISRAEL:
Report on the 1994-1995 Seasons of Excavation
Tantura Lagoon is one of the few natural harbors along Israel's

Mediterranean coast. It has served as a port facility for one of the country's
largest ancient mounds, Tel Dor, and its immediate environs, for at least four
millennia. The cove is also a natural mechanism for preserving the remains
of shipwrecks and associated finds, as it tends to bury vessels under sand
soon after a sinking event. These two considerations make Tantura Lagoon
an ideal location for the study of historically and archaeologically significant
shipwrecks.
Since 1994, the Institute of Nautical Archaeology (INA) at Texas A&M

University and Haifa University's Recanati Centre for Maritime Studies (CMS)
have joined forces to carry out a study of shipwrecks preserved inside the
cove. The 1994 and 1995 seasons focused primarily on the excavation and
the recording of the Tantura A hull. The portion examined constitutes
approximately twenty-five percent of the bottom of a small local coaster,
which was preserved up to about the turn of the bilge.
Of three radiocarbon tests carried out on a single splinter of wood

removed from Tantura A's keel, two supplied dates fell in the early fifth to
early sixth centuries AD (AD 425-530 and AD 440-540 respectively). The third
sample fell within two possible earlier calendric date ranges with unequal
weights (AD 268-280 [11%] and AD 330-415 [89%]). Byzantine-period
sherds found glued solidly to hull planking by resin also confirm a date for
the hull in the waning years of Byzantine rule in Palestine.
The date of this hull is of particular interest, as it is constructed in frame-

based technique, without the use of unpegged mortise-and-tenon edge
joinery, which until its excavation had been considered standard for the
seventh century AD1. This makes Tantura A the oldest recorded hull in the
Mediterranean to have been built in the innovative construction methods that
were to evolve more fully and to standardize during medieval times.
Tantura A lacked a cargo. Apparently, it had been washed off in the

process of the sinking event. We found remains of a Byzantine-period cargo,
perhaps from this hull, spread out beneath the sand south of the shipwreck
in Trenches 1, Ill and IV.
SHELLEY WACHSMANN TROPIS VI
A hydraulic probe survey carried out in the immediate vicinity of Tantura

A in 1995 located significant sections of two other coherent hulls, as well as
timbers of several additional vessels. Trench VII revealed an assembly of
loose timbers along with organic remains, consisting of lengths of rope
(including a knot and an eye-splice), basketry and colored cloth. We found
another large hull, lacking cargo, in Trench Vlll. This ship is apparently of late
- perhaps medieval - date. Next to it, lay a single plank bearing two
pegged mortise-and-tenon joints and associated with a collection of Late
Roman ceramics, raising the distinct possibility that a second, earlier, hull
may be located nearby. In Trench IX we found a large, strongly built hull
bearing Byzantine-period ceramics. The hydraulic probe survey also
revealed two lead-filled wooden anchor stocks of wood anchors (Trenches
IA and V), at least one of which dates to the Persian period. Additional
ceramic horizons discovered during the survey - ranging from the Middle
Bronze Age IIA to the Persian period - hint at the strong possibility of
additional hulls buried in the lagoon's sands.
Postscript
Immediately following the Sixth Symposium, in fall 1996, we carried out a

third season of exploration in Tantura Lagoon. Work focused on excavation
and study of the large shipwreck in Trench VIII, now termed Tantura B, which
dates to the early ninth century AD, and appears to be a galley. This hull lay
atop another shipwreck, dating to the Late Roman-period. Nearby the
expedition found timbers of a third hull, built with mortise-and-tenonjoinery.
Trench X revealedyet another shipwreck, of early medieval date. Thus, by the
end of three seasons of work, we had uncovered hull remains of seven ships
in an area about the size of a regulation basketball court.
The final excavation report for the 1994-1996 seasons of exploration is

currently in preparation.
Shelley Wachsmann
P.O. Drawer HG
College Station, TX 77841-5137
USA
NOTE
1.See Y. Kahanov's contribution in this volume.

THE INNCMS-
JOINT EXPEDITION TO TANTURA LAGOON, ISRAEL-
SELECT BIBLIOGRAPHY
Bryant, V.M., 1995. Preliminary Pollen Analysis of Sediments Collected from Tantura Lagoon.
INA Quarterly 2212: 18-19.
Carmi, Y. and D. Segal, 1995. How Old is the Shipwreck from Tantura Lagoon? The
Radiocarbon Evidence. INA Quarterly 2212: 12.
Charlton, W.H., 1995. The Rope. INA Quarterly 2212: 17.
ITF = In the Field. National Geographic Magazine 191(January 1997): 103-109 (see pp. 104-
105).
Kahanov, Y., 2000. A Byzantine Shipwreck (Tantura A) in Tantura Lagoon, Israel: Hull
Construction Report. Tropis VI: Proceedings of the 6th Symposium on Ship
Construction in Antiquity, Lamia, 1996.
Kahanov, Y. and S. Breitstein, 1995A. A Preliminary Study of the Hull Remains. INA Quarterly
2212: 9-13.
Kahanov, Y. and S. Breitstein, 19958. Tantura Excavation 1994: A Preliminary Report on the
Wood. C.M.S. News 22 (August).
Kahanov, Y. and J.G. Royal, 1996. The 1995 INNCMS Tantura A Byzantine Shipwreck
Excavation - Hull Construction Report. C.M.S. News 23 (December): 21-23.
Royal, J.G. and Y. Kahanov, in press. A Byzantine-Period Merchant Vessel at Tantura Lagoon,
Israel. lnternationalJournal of NauticalArchaeology.
Sibella, P., 1995A. The Ceramics. INA Quarterly 2212: 13-16.
Sibella, P., 19958. Notes on the Architectural Marble. INA Quarterly 2212: 19-20.
Sibella, P., 1998. Light from the Past: The 1996 Tantura Roman Lamp. INA Quarterly 2414: 16-
18.
Wachsmann, S., 1995A. The 1994 INAICMS Joint Expedition to Tantura Lagoon. INA Quarterly
2212: 3-8.
Wachsmann, S., 19958. Return to Tantura Lagoon. C.M.S. News 22 (August).
Wachsmann, S., 1996A. A Cove of Many Shipwrecks: The 1995 INAICMS Joint Expedition to
Tantura Lagoon. C.M.S. News 23 (December): cover, 17-21
Wachsmann, S., 19968. Technology Before its Time: A Byzantine Shipwreck from Tantura
Lagoon. The Explorers Journal 7411: 19-23.
Wachsmann, S., and Y. Kahanov, 1997. Shipwreck Fall: The 1995 INA/CMS Joint Expedition to
Tantura Lagoon, Israel. INA Quarterly 2411: cover, 3-18.
Wachsmann, S., Y. Kahanov and J. Hall, 1998. The Tantura B Shipwreck: The 1996 INNCMS
Joint Expedition to Tantura Lagoon, Israel. INA Quarterly 2414: cover, 3-15.
Wachsmann, S. and K. Raveh, 1984. A Concise Nautical History of Dornantura. lnternational
Journal of Nautical Archaeology 13: 223-241.
ABSTRACT
WATERCRAFT FOR HEAW TRANSPORT IN ANCIENT EGYPT
The construction of massive vessels used for transporting colossal

statues, obelisks, and building components was a vital element of ancient
Egyptian nautical technology. A recent experiment in loading a converted
modern vessel with a 1.5-ton granite obelisk allows the investigation to have
a practical basis that can be combined with previous knowledge about
Egyptian cargo carriers. The utility of the experiment also will be evaluated.
Dr. Cheryl Ward

Texas A & M University at Galveston
Pelican Island
P.O. Box 1675
Galveston, Texas 77663
USA
ON THE ROLE OF MULTI-FUNCTIONAL HYBRID HULLS
IN THE CONSTRUCTION OF A NARRATIVE OF EARLY GREEK SHIP
ARCHITECTURE*
Introduction
Crafting a narrative purporting to reproduce the historical evolution of

early Greek ship architecture is an undertaking fraught with difficulties and
pitfalls**. The three categories of available evidence, to wit, shipwrecks,
texts and representations, lack individually the necessary
comprehensiveness to contribute decisively on their own, while the simple
(and simplistic) addition of the accumulated information results in an
unsatisfactory account due to the specificity of individual wrecks, the
vagueness of the texts, and the generic nature of images. A measured
blending of input, the careful collating of complementary statements under
exclusion of obvious incompatibilities does not alone, however, result in an
acceptable narrative. Wrecks, texts and images, as they have come down to
the modern beholder, do not represent the sum total of the ancient maritime
experience. Nor does the attainable degree of detail offered by the data
provide a full spectrum of information. To this must be added factors more
difficult to evaluate: the impact of the functional environment, the role of
regional building traditions, and the influence of political conditions, as well
as aspects derivative of the archaeological record's partial visibility. These
latter include absent hull types or regional traits, misunderstandings ancient
and modern, and the hobby-horses and personal agendas of scholars.
The evidence raises a number of issues and contradictions requiring

analysis and resolution. An implicit - and not always sufficiently remarked
upon - leitmotif in the scholarly literature is that the database is composed
by images of oared galleys and wrecks of merchantmen, while the texts
generally concentrate on activities requiring warships to the quasi-complete
MICHAEL WEDDE TROPlS VI
exclusion of all other maritime undertakings. In addition, while the bulk of

ship images in Greek art is largely pre-Classical in date, sufficiently well-
documented wrecks permitting at least a partial reconstruction, an estimate
of capacity, and an ethnic attribution are all relatively late in date. The
ensuing narrative, based on the wreck/text/image triad, structures the
evolutionary sequence on "warships" alone, with the merchantman treated
as a given regardless of whether attested or not, while allowing for some
movement of goods on merchant galleys. This reconstruction, in a sense, is
not incorrect, but would, it is suggested, bear revision. Emending the
cumulative image projected by the textbooks, justifiably oft quoted and
constituting the references employed by non-specialists, amounts to a major
undertaking, when not an act of hubris. Nonetheless, recent developments
in the relevant bibliography render such a tentative desirable. In practical
terms, the standard narrative is constituted by three fundamental
assumptions, without whose acceptance the account would collapse. The
first involves a distinction between merchantmen and warships from the
earliest times onwards, the second an equation of oared galleys with
warships, and the third a cavalier encounter with visibility in the
archaeological record.
Establishing a terminology
A critique of the textbook reconstruction of ancient Greek ship

architecture does well to commence with terminological issues. Although
glossaries are appended to facilitate the reader's navigation through the
unavoidable technical terms, the accounts do not offer precise clarification
regarding the exact sense in which key words are employed1.The notable
exception is offered by the Classical Greek type designations such as
triakontoros, pentekontoros, trieres. These, however, are adopted as givens
from the texts, despite indications that usage did not remain constant: the
term pentekontoros may, or may not, have been employed to designate both
single- and double-level variants2, while the trieres is known to have gone
through at least two major stages in its design histow. This definitional
insouciance finds its roots in the ancient authors, who frequently employ
generic terms for specific ship types4,and has adhered to scholarship since
the time when the study of Greek ship building was in the hands of
philology5.
A rigorous terminology constitutes a prime desideratum, despite the

inconveniences involved. Forging a vocabulary largely predates, by
ON THE ROLE OF MULTI-FUNCTIONAL HYBRID HULLS IN THE CONSTRUCTION
OF A NARRATIVE OF EARLY GREEK SHIP ARCHITECTURE
necessity, the complete analysis of the database since through assigning

labels to the material under study the differences and similarities become
apparent. Undesirable since apt to steer the enquiry through the external
imposition of concepts, this is unavoidable due to the absence of specificity
inherited from the ancient texts. In addition, a strict terminology imposes
adherence upon the responsible scholar, even in cases when a hazier
vocabulary would facilitate encounters with recalcitrant evidence. Yet it is
exactly such unfocused linguistic usage which permits lax argumentation.
The objections raised here against the textbook accounts stem in part from
insufficient attention to terminology. An unpreventable blemish ensues when
precise definitions may not correspond to ancient thought and usage. Yet
modern thought about ancient topics reqirire specified usage so as to create
a basis for discussion.
For the present purpose, five definitions will be proposed to cover a

range of craft at varying extents testified to by the database. (Clearly generic
terms such as "ship", "boat", "craft", "embarkation", "vessel" must remain in
their innate loosely defined state to account linguistically for all-inclusive or
generalizing statements6.)
A merchantman will be defined as a hull conceived to maximize cargo
capacity, implying a minimal crew, and a reliance on the cheapest form of
locomotion: wind power. Oars constituted a viable alternative only in
limited contexts such as maneuvering.
A galley, while a capable sailer, is here understood as designed to seat a
large crew of rowers so as to attain high speeds regardless of wind
conditions. Its cargo capacity is reduced, although not non-existent. The
crew may also double up as warriors, but a galley is not, by definition, a
warship7. In its many permutations, the galley may, however, eventually
become functionally specialized, evolving into the backbone of ancient
pre-trieres navies.
A warship is a special purpose craft designed for participating in sea
battles or in activities related to warfare exclusively. Secondary uses are
quasi-excluded due to the design, the size of the crew, and the primary
purpose (unless major structural modification is undertaken). In the Greek
context, a warship is conceived almost exclusively for ramming: the trieres.
In this definition, a craft employed for troop transport is not a priori a
warship - even if adhering to the galley architecture.
The more recent literature has argued for the existence of galleys with
increased capacity to carry trade goods, yet little has been done to identify
such craft in the representational record. Almost by definition - I'etat des
choses oblige - physical remains of such hulls are non-existent. Two
different terms may be employed.

A cargo galley is a galley with improved cargo-carrying ability. The accent
is placed on galley characteristics as opposed to those of a hull conceived
to carry merchandise, resulting in a wider-bellied galley, conceived to
move primarily under oars.
A merchant galley places the stress on the cargo capacity, yet the hull is
designed along galley lines: sleeker, faster, with a galley bow, and
designed to be sailed extensively as well as rowed.
At first glance the distinction may seem a case of hair-splitting, but given
the gaps in the evidence, it may be of some significance: to eliminate the one
or the other, or to employ them as synonyms, would imply, as the terms are
defined, a specific procedure through which the shipwrights approached
this hybrid form. It would appear too early to operate such a deterministic
intervention on the language employed to categorize the datas.
Obviously, it may be argued that many designs would fall outside these
five categories. One may safely assume the existence of a vast, largely
invisible, population of small craft capable of moving at speed under oars,
sail well, carry cargo, passengers, military equipment or warriorsg. It should
also be kept in mind that any attempt to create type listings uniting all
instances thought to represent the available population for any one category
would butt against the problem of identifying models as statements of
specific design, and of determining minimum dimensions for admission to
the above categorieslO.
Hybrid designs between merchantman and oared galley
A major failing of the established narrative is to operate an unreflecting

distinction between merchantmen and warships in early Aegean ship
building without engaging in the prerequisite holistic analysis of all forms of
water transport for the period under consideration. Whereas Egyptian hull
forms exhibit two distinct tendencies (abstraction made of papyrus craft and
embarkations employed in religious contexts) caused by two largely
incompatible operational environments, riverine vs. maritime, such a
differentiation is not possible in the Aegean1'. Although there are faint
indications of several, in a sense, conflicting traditions in the Early Bronze
ship imagery, each with the potential to have remained in production despite
their later invisibility, Minoan ship building appears to have developed out of
the Early Cycladic II craft depicted on "frying pans" from Syros, two plaques
from Naxos, a sherd from Orkhomenos, and a model from Palaikastro". The
subsequent development down to the ships on the Miniature Wall Painting

from Akrotiri can be shown to ensue in an evolutionary manneri3.
A contrasting architecture, sufficiently different to warrant speaking of a

break, emerges towards the end of the Late Bronze Age, in the form of the
first Mycenaean vessels. Admittedly, this hull form appears, on the testimony
of the available imagery, rather suddenly, and it is not to be excluded that
this is connected to changes in ceramic decoration, the addition of figurative
designs to the existing abstract and vegetal repertoire, rather than to
exclusively ship constructional factors. Nonetheless, even if allowing for a
partial invisibility of the earliest statements of the Mycenaean hull type, or an
extremely selective representational strategy on the part of Minoan
craftsmen leading to exclusion of concurrent forms from the pictorial record,
the Minoan ship type appears to the modern beholder as the sole hull in use
during the acme of Minoan society. There being no contrast against which
to hold up this ship type, the distinction between merchantmen and
warships becomes inoperable on the level of hull architecture, and thus has
no place in the vocabulary of Aegean Bronze Age ship studies up until
around 1400 BC14.
This train of thought introduces the concept of multi-functionality: what

was, as far as the data allow to distinguish, a single ship type was employed
- on the testimony of the Akrotiri Miniature Wall Painting - as required,
either to carry goods on trading missions, ferry people in religious
processions, or transport warriors to theaters of war''. Functional diversity as
translated into differing architectural forms can only be argued when the
necessary contrasting types, and sufficient evidence for specialized (and
incompatible) use, are marshaled in support. This mono-functionality cannot
be approached solely from the ship architectural evidence, be it physical or
representational, but requires close readings of the historical testimonia, and
the factoring in of the functional and political environment.
Navigation in the Aegean, and the Eastern Mediterranean if the coastal

route is employed, is characterized by constant visual contact with land,
causing cabotage to be the main mode of movement. Landfalls for eating
and sleeping led to limited autonomy being required of a ship and crew,
potentially allowing a more specialized use of the hull. If speed was the
premium characteristic, the Greek ship architect employed almost the entire
hull as the motor section by emphasizing locomotion by oars. If cargo
capacity was advantaged, the hull became a wind-driven hold. In either
case, on-board livability was of minor import. Political conditions, on the
other hand, argued against mono-functionality. Centralized control of the

Aegean and the attendant political stability was a rare phenomenon,
essentially imposed only by the Minoans, the Corinthians, the Athenians, the
Rhodians, and, intermittently, the Romans16. When decentralized into
regionally restricted maritime fiefdoms, or entirely outside any form of
judiciary, the Aegean became the home of pirates. The invention of the
galley by the Mycenaeans, and its continuity throughout ship building history
in the region, indicate that the times of peace were few and short-lived. Only
with safe navigation can an economy rely on sail-propulsed merchantmen.
Unless a political entity could escort its merchantmen with a fleet of galleys,
something few were able to do, cargo had to be shipped in hulls capable of
moving at speed independent of wind conditions, of deterring attack from
marauders (whether privateers or state-sponsored), and of defending itself
in man-to-man combat at sea.
Whereas it is possible to interpret the textual evidence as supportive of

reading a limited capacity for cargo into oared galleys, it is clear that the
standard galley design, be it of single-, double-, or triple-level design, did not
offer adequate stowage to sustain a complex economy dependent, as any
such entity would be in Greece. on imports to supplement production and
natural resources within its own territory". Yet if, as argued above1',
merchantmen required a protected environment, and oared galleys, when
reduced to seating capacity alone, could not compensate for reduced bulk,
a third way became necessary. That a hybrid hull type midway between a
galley and a merchantman, combining speed under oars with increased
cargo capacity, existed in the Archaic period is evidenced by the statement
from Ploutarkhos that Polykrates ordered the construction of the so-called
samaina, a craft which was "low and flat in the prow, so as to look snub-
nosed, but wide and large and well-spread in the hold, by which it carries a
large cargo and sails well'g.'"
A number of images depict what the author would characterize as a

cargo galley, a vessel designed along traditional galley lines, and retaining
oars as the primary mode of propulsion, but endowed with a roomier hold
for an increased cargo capacity.
1. White-painted IV oinochoe of unknown provenance, Lefkosia 194711-

1611, Cypro-Archaic 1 (700-600 BC): flat hull with vertical stempost, small
projection at junction of keel and stem, non-descript stempost terminal

possibly related to the Mycenaean bird-headed post, stern with post
crowned by inward-turned duck-head, loose-footed brailed sail, two
steering-oars (Fig.
2. White-painted IV oinochoe from Ormidia (Cyprus), Metropolitan Museum

74.51.511, Cypro-Archaic I: flat hull with vertical stempost and non-
descript terminal, forecastle, inward-turned bird-headed sternpost, and
aftercastle, loose-footed brailed sail, two steering-oars (Fig. 2)".
3. Bichrome IV oinochoe from Karpas (Cyprus), British Museum 1926.6-

28.9, Cypro-Archaic I: flat hull with vertical stempost, bird-headed
terminal displaced by crew member hoisting anchor, forecastle, stern
with inward-turned, highly stylized bird-headed terminal, loose-footed
brailed sail, two steering-oars. Large amphora on either side of mast
(Fig. 3)23.
4. Stone relief from the palace at Kujundjik (Nineve), British Museum, reign
of Sennacherib (705-681 BC): flat hull with keel rising to vertical stempost
crowned by duck-headed terminal, sternpost curving slightly in over hull,
lower level of rowers rowing through ports, upper over gunwale, deck
with protective sidings, from which are hung shields, raised on
stanchions, two steering-oarsz4.
5. West Greek krater of unknown provenance by the Aristonothos Painter,

Palazzo dei Conservatori, 700-650 BC: curving hull with rising triangular
bow terminated by projection at base of oblique post, three
"proembolia"(?) extending beyond bow, cleft sternpost curving in over
hull, deck raised on stanchions, mast and stays, two steering-oars (Fig.
4y5.
6. Hebrew seal of Oniyahu, son of Merab, in a private collection, 8thor T hc.

BC: flat hull with vertical stempost crowned by bird-headed terminal,
vertical stern, shields along gunwale, sail, single steering-oar (Fig. 5)".
7. Seal from Roman tomb on Siphnos, probably T hc. BC: flat hull with
triangular bow and short projection, sternpost curving in over stern,
wale- and gunwale-lines running along hull, fore- and aftercastle, one
steering-oar (Fig. 6)27.
8. Ivory situla from Chiusi, end T hc. BC: curving hull with triangular bow
and large projection, stempost turned inward, stern curving in over hull,
sail, one steering-oar (Fig. 7)28.
9. Clay model from Amathous (Cyprus), Metropolitan Museum 74.51.1752,

Cypro-Archaic (c. 600 BC): wide-bellied, deep hull with vertical stempost
and short projection at waterline, curving sternpost rising into
anthropomorphic terminal with attached aftercastle, two wales along hull
and railing on gunwalea.
10. Painting in a tomb near Kef-el-Blida (Tunisia), 6thor 5thc. BC: flat hull with
triangular bow, stern curving into vertical post, sail, two (?) steering-
oars3'.
11. Amphora from Vulci, British Museum H230, beginning 6thc. BC: flat hull
with triangular bow with projection and possible animal-headed
stempost terminal, stern curving in over hull, sail, one steering-oar (Fig.
8)31.
12. Fragment of painted clay plaque from Corinth, 6th c. BC: curving
bifurcated stern with insignia, mast with lowered yard and sail, rigging,
row of pitchers along upper edge of plaque (Fig. 9)32.
13.Painting on Black-figure kalpis, Rijksmuseum Meermanno-

Westreenianum 6191836, c. 510 BC: flat hull with almost vertical
stempost and non-descript terminal, sternpost curving up, rowers rowing
over gunwale, loose-footed brailed sail, two steering-oars33.
14. Painting on Red-figure stamnos, the "Siren Vase", British Museum E440,
beginning !jthc. BC: flat hull with projectionlram and concave stempost,
stern rising into outward-turned post-terminal, rowers rowing through
ports, loose-footed brailed sail, two steering-oars?.
Comments
The three Cypriote vases combine features associated with

contemporary galleys and aspects of hull morphology sufficiently unusual to
merit attention. None have the characteristic triangular bow profile
introduced in the late Middle Geometric period in Greece, yet all display the
vertical stempost typical of earlier galleys. The first ship (1) has the small
spur known from Mycenaean and Protogeometric to early Middle Geometric
galleys, and the loose-footed brailed sail. The second (2) lacks the spur, but
has fore- and aftercastles. The third (3) again lacks the spur, but has the bow
figure so typical of Mycenaean galleys, yet carries two large recipients.
These ships are nor galleys in the traditional sense, nor merchantmen.
The inclusion of the Kujundjik "roundships" (4) employed by King Luli to

evacuate Tyr in the face of the onslaught of Sennacherib in 701 BC may
surprise35.The best-preserved ship exhibits traits which indicate that the
Assyrian artisan commissioned to recreate the scene for the palace at
Kuyundjik committed some "artist's errors". The right extremity has a vertical
post terminated by an avian figurehead, whereas the left curves up gently
from the keel-line. Yet two steering-oars are shown to the right, and the
rowers face in this direction. The steering-oars are incorrectly placed at the
bow and the oarcrew is inverted accordingly. Although less well preserved,
the ship to the right confirms this reading: the steering-oars are shown at the
left extremity, and the crew faces the stern. The ship is more roughly carved,
and damage obscures the differences in the lines at bow and stern observed
on the first ship, but it confirms the post with figurehead as the bow.
The right ship on the Aristonothos krater (5) offers a striking contrast to
its opponent. Rather than stress its otherness in terms of ethnicity, it would
appear more profitable to note morphological features in harmony with the
reading suggested here. The stern is that of a galley, whereas the bow, while
not equipped with a prominent projection, has a triangular profile, and the
continuations of the wales beyond the post known from galleys. The hull is
decked and deep, suggesting an increased cargo capacity"6. Although no
oars are shown, all galleys and galley derivatives were designed to be
rowed. The triangular bow profile encountered on the Aristonothos ship is
characteristic of the Kef-el-Blida (10) and Meermanno-Westreenianum (13)
vessels; if quoted here, it is only to underscore the suggestion that there
existed, throughout the Mediterranean, ships that were galleys, but not
designed quite like the run-of-the-mill pentekontoros, or dieres. The
existence of a hybrid multi-functional galley design is thus suggested by a
scatter of images from the 8thdown into the 5mcentury37.
Dating the inception of diversity in ship architecture
From the material catalogued above, if correctly interpreted as cargo

galleys, an initial date for the inception of the type in the late 8mto early 7
century BC may be suggested. Since the earliest images suggest a
MICHAEL WEDDE TROPIS VI
developed hull type, it is probable that the dating needs to be revised

upwards - despite the lack of representations. An argument in support
thereof can be constructed by reference to developments in the late
Mycenaean period, and in particular to the Pyrgos Livanaton ships3'. The
invention of the oared galley by the Mycenaeans some time in the 14"
century BC3', a de facto rejection of the Minoan ship type as unsuitable to
their needs, led to a greatly diminished ability to carry merchandise. A
double vocation as traders and warriors rendered a second design
imperative: deeper, somewhat slower, decked, spurless, combining speed
with cargo capacity, traits offered by the Minoan ships only under sail4".From
the ensuing dichotomy rose not only the lineage of Greek decked galleys
leading to the dieres, and, ultimately, to the trieres, but also the cargo
gallev'. The development down to the late 81hcentury BC cannot be charted
with certainty in the absence of sufficient data. Nonetheless, decked galleys
appear to have remained in production, as attested by the Middle Geometric
II Metropolitan Museum krater and Lefkandi pyxis, the latter which points
directly at the two-leveled Dipylon ships42.TOwhat extent the decked single-
level type, or the double-level design, can be considered cargo galleys
remains unknown.
The most significant contribution to a hypothetical answer to the

question would be to liberate the discourse from the needless burden of
automatically equating oared galleys with warships as a specific type. In the
definition suggested above, the warship is mono-functional in its conception,
although a partial multi-functionality can be assumed in the context of naval
operations. The warship, understood as a weapon in itself, was primarily
designed for battle at sea, but also suited for the showing of strength as a
deterrent or threat. This could include blockading ports, or escorting
convoys of defenseless ships. The sole ship design to fit the definition is the
trieres (and its subsequent developments), suggesting that the warship -
as defined here - appears in Aegean waters when the Greeks invent or
adopt this type as the main component of their navies. It is thus argued that
all prior developments of hulls primarily designed to be rowed should be
considered oared galleys, able to fill a number of functions. In turn, this
argument raises two crucial issues, to wit, the date when mono-functional
warships became viable, and when their necessary obverse, the dedicated
cargo carrier, the merchantman, begins to appear in appreciable numbers in
the Aegean, developments which do not exclude the continued use of oared
galleys or hybrid designs of greater versatility.
It is customary to consider the merchantman a constant of ship design,

given the high volume of seaborne trade which can be reconstructed from
import distribution patterns. As merchantmen are very rarely represented,
this requires postulating non-depicted merchantmen for the pre-Classical
period, and arguing from to date unexcavated wrecks. Whereas arguments
ex silentio are a frequently employed, and necessary, tool of archaeology,
and a recurrent feature of the research presented here, they should be
employed only when other channels of thought have been exhausted. The
merchantman is a case in point. It is obvious that hulls, large and small,
primarily conceived for transporting goods over short or long distances,
were constructed at all times. A subsistence economy in an island or coastal
context requires suitable embarkations, with small communities rarely
capable of manning a large, or even medium-sized, oared galley. But to term
such craft merchantmen requires a step not necessarily supported by the
earlier data.
Statistically, the merchantman is a Classical phenomenon. If it is argued

that the shipwrecks catalogued to date are cargo-carrying hulls, the sine qua
non of their detection, excavation, and documentation as archaeological
sites since the unballasted galley does not sink, a database of 1149
Mediterranean wrecks is available for the time period 2200 BC to AD 15m
centuw. It is clear that not all these sites constitute wrecks, but for an initial
appreciation the problem of calibrating the proportion of incorrect
identifications may be considered a constant over time. From the earliest
wreck down to 500 BC, some 1700 years, the catalogue contains 38 entries
(the 6thcentury accounts for 23 items). The 5thcentury alone numbers 38
wrecks. There follows a gradual increase until triple-digit figures are attained
for the time span Pdcentury BC-AD 2ndcentury. Even if there are numerous
factors involved in constituting a database which so favors the Late
HellenisticILate Roman Republic to Early Roman Empire periods (there is a
sudden drop and gradual decrease from the AD 3rdand later centuries), the
numbers cannot be purely random. The volume of trade carried by
dedicated cargo carriers likely, on every trip, to be in a state conducive to
preservation as an archaeological site if lost at sea is infinitely smaller in the
period prior to 600 BC".
Notwithstanding the realization that the pictorial evidence cannot be

expected to chart the development of ship architecture in its every detail, it
cannot but be noted that depictions of recognizable merchantmen are
extremely rare prior to the Roman period. For the time span under study
here, down to the 5Ihcentury BC, only five instances can be catalogued, of
which three are dated immediately before and after the turn of the century"5.
Similarly, and for well-known reasons, merchantmen appear rarely in the

literature down to the court, cases involving shipwrecks in the corpus of
speeches by the Athenian orators in the Classical period. A topic in itself, it
may briefly be noted that although the merchantman existed in the time of
Homeros, as indicated by similes employing some aspect of the
merchantman as a comparandum, the type never appears in corpore in the
Homeric epics. It is erroneous to argue that the twenty-oared ship was a
merchantman: the texts indicate that it is a swift oared galley of multiple
purpose, including cargo transpop.
The reasons for the absence of the pre-Classical merchantman in the
wreck/text/image data triad are surely manifold, including inclement
conditions for conservation, and irrelevance to most literary and
representational themes. A further element towards an explanation may be
sought in the political conditions. Through its inability to escape or defend
itself against a pursuer, a merchantman requires the rule of law over the seas
it is to travel. If, in addition to the natural dangers of the sea, hostile behavior
from foreign galleys would render crossings unnecessarily dangerous, a
partial or full curtailment of non-military traffic if restricted to merchantmen
would result. The mono-functional merchantman became viable only when
the Aegean was pacified, that is, dominated by one or more states capable
of exerting control with a substantial fleet. Despite the occasional presence
of what appears to be a standing navy in the Archaic period (Polykrates), this
does not happen until the inception of trieres-based fleets and the rise of
Athens. This capital advance in ship architecture created the first mono-
functional warship (as defined above), incapable of doing extra-duty as a
cargo carrier. Although other forms of oared galleys, including cargo galleys,
remained in use, the appearance of the trieres is linked to the increased use
of merchantmen to transport goods.
The date for the introduction of the trieres constitutes a problem well
beyond solution within these pages4'. The parameters involved are manifold:
interpreting Thoukydides 1.13 and solving the conflict with Thoukydides 1.14;
testing the compatibility of further literary testimonia such as Herodotos
11.159, Clement 1.16.36, Plinius V11.56.207, Diodoros XIV.42.3, and Nikolaos
of Damaskos (FGrH 90 F.58); accepting Thoukydides' date of 704 BC, or
favoring the re-dating on grounds of a faulty generation count to c.650 BC,
or arguing for a late 6th c. date; gauging the consequences of these
conflicting dates, including their impact on the origins of the trieres, and on
the rate of innovation in early ship architecture; factoring in such aspects as
the role of ramming in galley design, and the speed of trieres incorporation
in fleets - and the reasons therefore. The analysis is rendered difficult by the
OF A NARRATIVE OF E.4RLY GREEK SHIP ARCHITECTURE
very uneven evidence at the scholar's disposition, being essentially a

handful of Greek texts which either ignore the Levantine situation, or
misinterpret or misrepresent it4'.
Contrary to some scholars, the present author does not consider the
evidence adequate to establish beyond doubt the trieres as a late 8th/earlyFh
century invention. The three key statements in Thoukydides 1.13 concerning
the Corinthians as (1) the first modern shipbuilders" and (2) the first builders
of trieres, and (3) Ameinokles as the builder of four ships for the Samiansso,
are preceded by formulations indicating doubf'. In addition, the statement
regarding the first naval battle is qualified as the first of which knowledge is
still available. Furthermore, the position of 1.13 before the Methodologia
should be warning enough not to read this passage as having the same
factual imprimatur Thoukydides attempts to place on his account of the
Peloponnesian Waf2.Whereas it is attractive to argue that Diodoros XIV.42.3
confirms ThoukydideP, the statement regarding Dionysos I of Syracuse (a
Corinthian colony) being spurred by the presumed Corinthian origin of the
trieres to initiate a major naval building programme may not do more than
attest to Thoukydides' sources being known in Syracuse. A later source (a
1' century BC text going back to a 4thcentury BC informant, in this case
probably Philistos of Syracuse) cannot confirm since an independent
transmission is not assured". It is significant that Thoukydides 1.14 bundles
the Corinthians, lonians, Samians, and Phokaians together as having had
navies almost without trieres, and dates the serious appearance of the type
to just prior to the Persian Warsss.
Herodotos tends to confirm the impression that regardless of when the

trieres was invented, it did not play any attested role whatsoever before the
middle of the 6thcentury. Although it may be argued that the plundering raids
of Polykrates requiredpentekontoroi rather than trieres, 111.39 and 111.44 allow
a fairly precise date to be advanced for the Samian tyrant's (re-)constructing
his navy around the trieres, and hint at an inability by his victims to face his
pentekontoroi with superior weaponP. If Thoukydides 1.14 (also before the
Methodologia) is correct in stating that the first to employ trieres were the
Sicilian tyrants and the Corcyraeans towards the end of the 6thcentury, then
a general lack of Greek trieres may be postulated in the Eastern Aegean until
they appear in Polykrates' fleef7. Herodotos V1.6-13 certainly indicates that
the proper use of the trieres was a largely unknown subjects for the lonians
just before the battle of Lade in 494 BC. The fact that the Phokaians
employed pentekontoroi at the battle of Alalia c. 535 BC is not proof in itself
that trieres were not in use at that time since they had, as told by Herodotos
111.163-166, left their homeland in pentekontoro~~. Their behavior in the

waters around Corsica prior to being confronted by the Etrusco-Carthaginian
fleet is akin to that of Polykrates around Samos, and requires more versatile
vessels than trieres. Fighting the enemy to a Pyrrhic victory with a fleet half
as large as that of their opponents suggests that the Etruscans and the
Carthaginians did not employ triremes5'. This would, in turn, be surprising as
far as the Carthaginians are concerned, hinting as it does that the
Phoenicians had not yet adopted the three-level galley.
An early adoption by the Phoenician in the Levant cannot be argued from

the available data, all which point to a late Gth/early5mcentury introduction,
possibly with Kambyses' establishing of a Persian fleet as the impulse60.The
sole other attested user of triremes is Necho, according to Herodotos 11.159,
yet it has been plausibly argued that the Egyptian kbn(w)t, translated by
Herodotos' informant as trieres, is a generic term in use since the Old
Kingdom, and, thus, with changing semantic context, trieres, then, merely
being an approximation for an oared galley employed for essentially
bellicose purposes, with no information on specificities of design and
oarage6'. Given the uncertainties involved in readings of Thoukydides 1.13,
the manifest anti-Greek slant of Clement 1.15-16, the testimonia of
Thoukydides 1.14 and Herodotos passim, and the chronological position of
Nikolaos of Damaskos6' and Diodoros' sources much later (both 1" century
BC) than Thoukydides casting doubt on their independence, it would appear
that Herodotos 11.159 should best be divorced from the trieres questioP.
The final item to be noted is that the trieres employed by the Greeks against
the Persians in 480-478 BC were not, according to Thoukydides 1.14,
decked, implying a more primitive stage in the development, perhaps
congruent with a reading underlining the newness of the design (as is the
information on the battle of Lade) at the beginning of the 5thcentury BC4.TO
retain the early to middle Thcentury as the date of invention of the trieres, a
date supported only by a generous and partially preconceived interpretation
of Thoukydides 1.13, would require accepting that a design slated to change
the nature of naval warfare in the Mediterranean - not to mention the
political impact on democracy in Athens - remained for three generations
without any discernible traces in history. This constitutes an unlikely
construct.
Thus a number of different enquiries converge on a shared historical

nexus: the concerted use of the merchantman becomes possible when
economic and political changes have rendered the appearance of the trieres
as the predominant ship of Greek fleets possible, if not necessary, a
development which coincides with the need for large-scale importation of

grain, and the battle for survival against the Persians, thus leading to two
mono-functional evolutions in ship design, each dependent on the other, the
merchantman to feed the economy that sustained the trieres which, in turn,
protected the merchantman. Although the trieres remains invisible in the
shipwreck data, there is an abrupt increase in the number of wrecks in the
second half of the 6thcentury BC, wrecks probably to be understood as
merchantmen. Prior to about 550-530 BC, movement over water was
dominated by multi-functional galleys. If the trieres did exist, it had no
historical impact.
Visibility in the archaeological record
The above account certainly requires further work. It may even border on
being that of a heretic. But it attempts to integrate all factors impacting on
ship design and use, and its study, into a holistic approach, into what may
be termed a "grand narrative". A tale spun across a millennium. It
acknowledges that interpretation in Archaeology entails constructing
plausible scenaria from disparate data, where the order in which evidence is
marshaled, and the unequal stress placed both by availability and the
scholar on given factors, will influence the results. Thus crafting a narrative
founded on the remains of a complex system becomes a study of visibility
requiring a substantial allowance for the invisible. Examining the evidence of
wrecks, texts, and representations cannot merely result in an account of
whatever happens to be available, but must be a study of method itself: how
to de- and reconstruct the database in constant reference to the
imponderables. Such an approach allows amplification of the available
evidence: the earliest representation or mention does not constitute a
beginning, a dearth of images should not be taken on face value alone,
textual under-representation requires enhancement. The bottom line is that
of plausibility, the placing of all elements within a coherent framework,
thereby avoiding overstressing a single wreck, text or image.
A consideration of the inadequacies of the database leads to inferences

(rather than speculation) regarding the invisible elements. The trieres itself,
if dated to the f hcentury, constitutes the major example, a design leaving no
other traces than the vague formulations of Thoukydides 1.13. It would be
necessary to postulate a sufficient production to sustain the technology at
one or more centers over three to five 30-year generations - given the
explosion of trieres towards the end of the 6m century, leading to the
appearance of comparatively large trieres fleets, several yards would be

required to maintain trieres-building know-how among their shipwrights
despite an almost non-existent demandffi, small cellules de veille, just in
case. Arguing in favor of a Greek origin for the trieres, and thereby probably
imposing an early date, would require a suitable two-level design, for which
there is no extant Greek representational evidence. The Dipylon ships
constitute a first attempt, the ships of Luli the evolution thereofs6. Either Greek
two-level ships with oarports existed, yet remain invisible to the modern
beholder, or a Th century Greek trieres developed a Phoenician advance,
itself derivative of a Greek design67.
The available textual evidence, too, hides more than it reveals

behind the use of the generic doru, naus, ploion and ploia makra. In fact,
apart from specific statements such as those made by Herodotos regarding
the Phokaians and Polykrates, or the use of type-specific designations in
Homeros (abstraction made of how modern scholars understand them), the
texts offer a very limited aperqu of early ship building and ship use. The
seabattle between the Corinthians and the Corcyraeans, introduced as the
first known such event by Thoukydides 1.13, obviously hides an enormous
range of sea-born activity of a more or less bellicose nature: the rise of
Corinth before and under the Kypselids, encounters during the colonization,
the Lelantine war, conflicts pitting sea-faring states against local
homologues or landlubbers, to mention a few potential contexts*. Clearly,
interpreting the available database requires a delicate balancing act, and a
willingness to go beyond the restrictions imposed by the visible6'.
Conclusions
It would, then, be foolhardy to argue that merchantmen were a rare sight

in the Aegean maritime economy before the Classical age solely on the
absence of representations, or on the low number of wrecks discovered and
dated to period before 600 BC. Similarly, rejecting Thoukydides 1.13 does
not automatically redate the appearance of the trieres to c. 550 BC. It is the
accumulation of indications culled from the catalogue of wrecks, the texts,
the representations, the functional environment, and the political conditions
reconstructed independently of reference to ship architecture, which leads
to the proposals made herewithin. The abrupt increase in the number of
wrecks around 550 BC70coincide with the testimony of Thoukydides 1.14 to
suggest that a trieres-induced reduction of piracy and other unsettled
conditions at sea could have led to an accrued viability of merchantman-
orientated transport in the later 6th century7'. A comparative absence of

protected sea lanes leading to a reduced use of merchantmen, in turn,
requires alternative means of cargo distribution. Homeros in the Odysseia,
Herodotos 111.39 with 111.44, and Ploutarkhos' Life of Perikles on Polykrates,
and Herodotos 111.163-166 on the Phokaians, coupled with a series of
images, suggest that the early Greeks, rather than escorting merchantmen
with galleys, may have resorted to the multi-functional and hybrid cargo
galley.
As the narrative stands it is clearly a product of the critique of the

literature which colored the epistemological framework at the outset. In its
present form, it is also under-documented, requiring a fuller treatment of the
texts, particularly in reference to the dating of the trieres. Yet in attempting to
point to areas requiring further work, such as the incorporation of data
lacking the visibility required for a positivist reading, it does place a more
than just discreet question mark next to some of the fundamental ideas upon
which the textbooks construct their accounts of the early Greek ship
architecture.
Michael Wedde
Loutropyrgos
GR-19006 Nea Peramos
Greece
NOTES
* The author is grateful to the organizers for extending an invitation to

present his views at the symposion. As usual, the text has been read and
criticized by Mrs. Ethel Wedde. Periodicals are abbreviated as laid out in
American Journal of Archaeology 90, 1986, 384-394, and 92, 1988, 629-630.
Note also:
FGrH Fragmente der griechischen Historiker (F. Jacoby, Leiden
1923-1958)
MarM Mariner's Mirror
**The present paper should be understood as experimental, owing its existence to discontent
with textbook accounts. It does not claim to solve the many problems involved in decoding the
data relative to early Greek ship architecture, but to offer a viable approach. The author's views
on the Bronze Age, Iron Age, and early Archaic developments are laid out in Wedde 1996.
1. Glossaries generally cover technical terms of hull construction, rigging, and ship handling.
Casson 1971:389-402 is exceptional in linking terms in the glossary to discussions in the
text, thus creating a book which can function as an extended glossary. Contrast
Morrison/Williams 1968:338-340 and GardinerIMorrison 1995:248-251. All three avoid the
contentious and generally loosely defined generic terms. The problems involved in providing
clear-cut definitions for ancient terms (witness eikosoros and pentekontoros) unfortunately
promote less than rigorous usage.
2. The term pentekontoros constitutes one of the major problems facing the scholar. From
having initially been understood as designating a single-level 50-oared open hull, as it clearly
is in Homeros, some scholars have come to believe it may also cover two-level vessels, due
to the absence of a commonly employed term for this design; cf. Morrison/Coates 1986:33-
35, Casson 1971:58-59 with 59n82, 61-63, GardinerIMorrison 1995:250 s.v. pentecontor.
Wallinga 1993 would argue otherwise, making the pentekontoros into a, by definition, two-
level craft seating 13 and 12, or 14 and 11 rowers (thus ignoring the Bronze Age and
Homeric evidence to the contrary), and functionally a merchant galley (although not
necessarily as defined here). His reconstruction of the oarage on the Dipylon ships, lower
level through ports, upper over an outrigger, has no factual basis (as noted by Morrison
1994). Whereas Herodotos, The Histories 1.164 on the Phokaians, offers proof that a
pentekontoros could carry cargo, the approach of Hockmann 1989 is to be preferred.
3. The Themistoklean trieres was, according to Thoukydides, The Peloponnesian War 1.14.3,
undecked (or only partially decked), functioning exclusively as a weapon in itself. Kimon
(Ploutarkhos, Life of Kirnon 12.2) added a wide deck, thereby preparing a potential move
away from ramming in favor of pitched battle between hoplites involving boarding, the
approach employed by the Corinthians at the battle of Sybota in 433 BC (Thouk. 1.45-54; cf.
MorrisonlCoates 1986:62-68). The issue of undeckedldecked hulls remains unsolved, but cf.
Wedde 1993 for an introduction to the problem.
4. Thouk. 1.13 is the locus classicus, causing disagreement on the crucial date of when the
trieres was invented (cf. below).
5. Not to be misconstrued as a blanket condemnation of philology as a research tool. On the
contrary, Torr 1894:105-124 provides ample evidence of the contribution of the texts to
understanding the range of ship types employed in Antiquity. cf. also Casson 1971:157-168.
6. To be complete, the present definitions ought to include the terms "navy", "fleet", "convoy",
etc. In the present text a generic ship term transcribed from the Greek and italicized refers to
the Greek type (e.g. trieres). The same obtains for Roman craft (e.g. liburna). The latin form
in roman type is employed for non-Greek, non-Roman variants (e.g. the Phoenician trireme).
7. As illustrated by a use as pirate ship: piracy is not an act of war unless perpetrated by one
state upon the subjects of another (making the dealings of Polykrates a borderline case). Yet
ship types originally introduced by pirates, such as the herniolia and the liburna, may find use
in the navies of a state. On the herniolia, cf. Casson 1958, Morrison 1980; on the liburna,
Anderson 1962:31-36, Casson 1971:340.
8. Casson 1971:157-168 discusses a number of terms by which hull types clearly to be
understood as merchant galley were designated. The multi-functionality, and the possibility
that any given term was applied to a range of related designs, argue against a too rigorous
application of such terms as pentekontoros.
9. Cf. Casson 1971:329-343 on small craft.
10. Based on experience with Aegean Bronze Age models, the present author is less sanguine
than others in evaluating the contribution offered by this category of finds. Frequently models
are too schematic, devoid of size indicators, and lack morphological uniformity to constitute
a confident basis for typological analysis (clusters!).
11. The state of the database imposes an unfortunate concentration on Aegean and Greek ship
building at the exclusion of other traditions. It is difficult to argue for foreign impulses when
the necessary evidence is lacking, yet it would be incorrect to argue against for exactly the
same reason.
12. Conveniently united at Basch 1987:79-82 figs 158-168, 78 fig.152,83 figs 169, 172, and 170-
171. For a discussion, cf. Wedde 1996:127-128, 137-140.
13. For extensive discussion, cf. Wedde 1996.
14. This distinction is latent in Morrison 1994 when he states that round sailing ships are
common in pre-Geometric imagery. To consider the Syros "frying-pan" craft and the Naxos
lead models as warships (Morrison/Coates 1986:25) is to misunderstand the evolution of
early Aegean ship architecture.
15. Cf. Marinatos 1974:color pls 7,9; Morgan 1988:figs 9-12, 189; Doumas 1992:figs 26,29,35-
38; Televantou 1994:foldout pls 1-2, 4, foldout figs I , 3.
16. Traditional or historical evidence support these candidates' entry on the shortlist. Other
seapowers such as the lonians and the Samians may or may not have sought to root out
pirates from their nests. The ability to protect shipping either through escorting cargo carriers
with galleys, or countering predation by employing cargo galleys capable of armed
resistance, does not imply an active anti-piracy policy. State-organized piracy may be
argued for as one aspect of the maritime involvement of the Samians under Polykrates. On
piracy, cf. Ormerod 1924.
17. According to Garnsey 1988 the inception of large-scale grain imports to Greece, chiefly
Athens, has been systematically up-dated in the scholarly literature. He prefers the late
Gm/early5mcentury as the turning point.
18. Admittedly in insufficient detail: the present purpose is merely to raise the issue, since an
adequate treatment would require a work of ancient history.
19. In John Dryden's translation of the Life of Perikles 26.3. Such a deep-hulled galley has been
postulated by Hockmann 1989, who suggests the ship on the "Siren Vase" (Cat. nr 14) as
an adequate approximation in pictorial terms. On the samaina, cf. Casson 1971:63 with
n.104. Wallinga 1993:93-99 is, as frequently on other issues as well, thought-provoking and
partly fanciful. For the literary testimonia, cf. Dunst 1972:159-161 (although his explanation
for the ram being described as either boar-headed or fish-headed is to be rejected
emphatically).
20. No more than a checklist is intended. It includes all instances considered by the author to
represent cargo galleys at the time of writing. No attempt is made at this early stage to
distinguish between cargo galleys and merchant galleys in the pictorial record. References
are restricted to one (Basch 1987 if available) or more illustrations and/or a more substantial
publication. The author does not claim to sail previously uncharted waters. The merchant
galley does appear in the textbooks (cf. Casson 1971:157-158, 1995, Wallinga 1993 etc.),
but little has been done to identify early examples and incorporate them into an overall
explanatory framework.
21. Basch 1987:260 fig.563; Westerberg1983:44-45 cat. nr 54 and 117 fig.54.
22. Basch 1987:261 fig.567; Westerberg1983:45 cat. nr 55 and 118 fig.55.
23. Basch 1987:261 fig.564; Westerberg1983:43-44cat. nr 53 and 116 fig.53.
24. Basch 1987:314 figs 660-661.
25. Basch 1987:233 fig.482; cf. Schweitzer 1955.
26. Avigad 1982:59 fig.1, Stieglitz 1984139. Discussed by Lionel Casson at the Third
Symposion as a merchantman, but without subsequent publication. The shields (or oar-
ports?) render a reading as a merchantman problematic.
27. Basch 1987:248 fig.522. The image may also represent an oared galley.
28. Basch 1987:409 fig.871. The shape reproduces that of the six ships on an Etruscan
oinochoe in the Maritime Museum of Haifa, dating to 725-625 BC, which may also be cargo
galleys. An almost identical vase with five such vessels is in the University of Missouri
Museum of Art and Archaeology, cf. Biers1 Humphreys 1977.
29. Basch 1987:252 fig.536; Westerberg 1983:41-42 cat. nr 50 and 113 fig.50. Contrast
Landstrom 1969:28-29 fig.61 (Basch 1987: 253 fig. 540) who reconstructs a merchantman
with the projection well above the waterline.

30. Basch 1987:397 fig.826.
31. Basch 1987:408 fig.868
32. Basch 1987:237 fig.494. The stern morphology suggests a galley (cf. Snodgrass 1983:17),
rather than a merchantman (cf. Basch 1987:235).
33. Basch 1987:228 fig.474; Galestin 1977:58-59 cat. nr 11.
34. Basch 1987:270 fig.574; Morrison~Williams1968:114 Arch.94. Included on the basis of
Hockmann 1989. On this ship, cf. also Tilley 1989:430-431for an interpretation as triple-
banked early trieres type. The present author employs the term "banked" according to
Tilley's definition as being a file of rowers down the length of the hull, as opposed to "level",
being all rowers of banks at the same height in the hull.
35. It is imperative that one avoids consulting the drawings by A.H.Layard since these
misrepresent the "roundships", as documented by Basch 1987:314 fig.661 for example. All
statements are based on id.: 314 fig.660. Cf. also DeGraeve 1981:pl.XL1.87a. Casson
1971:65 correctly terms the ships devoid of bow-projection "merchant galleys"; cf. DeGraeve
1981:67. Basch 1969:150 sees a fundamental relatedness between Luli's "longships" and
"roundships", but considers the former to be "simply a military version of the 'round' ships."
The reverse appears to be the case. The issue revolves around whether the Lulian
"roundships" are cargo galleys or merchant galleys, as defined above.
36. For a reading as decked, cf. Wedde 1993. The suggestion by Morrison1Coates 1986:28 that
the beak-like bow was "designed to engage and hold the upper works of an enemy ship and,
like the later Roman corvus, prevent disengagement while the boarding party did its work"
ignores the fundamental difference in the corvus being a manipulable piece of deck
equipment, and not part of the hull. To have the "beak" run up on an enemy hull would surely
result in the two becoming locked together, placing the attacker in danger of becoming, in
turn, a target, and would require either substantial redistribution of ballast and/or time-
consuming wood-chopping to free the "beak" from the upper works (an operation which
even a slight sea would render even more difficult). Rankov 1996:51 notes that the rowers of
the trieres accounts for 15 tonnes (or 36%) of the total displacement of 42 tonnes, and
function as ballast. The crew must remain seated so as not to destabilize the ship. Cf.
Coates/Platisl Shaw 1988:63-64. Sleeswyk 1991 suggests a limited use of the crew to trim
the craft for the attack, raising or lowering the ram, depending on the type of target.
37. It is debatable whether the graffito on an Etruscan vase from Veii, dating to 700-650 BC
(Basch 1987:408 fig.865) should be included. Further instances of the oblique stempost
without bow projection appear in later periods but need not be of concern here. In addition,
there are a number of models which defy confident type-designation.
38. In Wedde 2000. the author distinguishes between the Tragana type, an oared galley with a
projection at the bow, and the Skyros type, a related design without the projection, possibly
decked on the testimony of the Enkomi Grave 3 ships, to which type the Pyrgos Livanaton
ships were assigned when they were made available to scholars by Dakoronia 1987 (op. cit.
122 figs 1-3). These images, rather than the Skyros (Basch 1987:142 fig.295) or Enkomi
(id.:148 fig.311) ships, best illustrate the second Mycenaean hull type.
39. The date is suggested by the available imagery, the Late Helladic B and C periods, which
cover, according to Manning 1988:56, the timespan 136011325-1065 BC, according to
WarrenIHankey 1989:169,1340/1330-1065 BC. Since the appearance of the two Mycenaean
types is rather abrupt, one may calculate with a certain undocumented development time.
Whether this should include the Late Helladic lllA period (Manning: ?/1490/1450-1430-1400
BC; WarrenIHankey: c. 1390-134011330 BC) cannot be ascertained. Interestingly, the
Marmor Parium 15 ascribes the first pentekontorosto Danaos c. 1510 BC (Jacoby 1904:5; cf.
Davison 1947:19n1).
40. The large ships on the Akrotiri Miniature Wall Painting are generally considered first-rate
sailing vessels (cf. Gillmer 1975, 1978, 1985A, 1985B). The substantial overhangs and the
reduction of the motor-section when under oars to 50% of the length overall argues for a
disadvantage against the galley design when rowed, even though the maximum crew would
have attained 80% of the effectifs on a pentekontoros.
41. Cf. Wedde 1993, 1996.
42. The Metropolitan krater: Basch 1987:178 fig.374 (rejecting Basch's LG II date, and reading);
the Lefkandi pyxis: Kalligas 1987:83 fig.1, Popham 1987:357 fig.4.
43. Based on Parker 1992, excluding 70 entries outside the Mediterranean, and 42 undated
wrecks, adding the Cypro-Mycenaean wreck of Point lria (Pennas/Vichos/Lolos 1996), and
the 5m-centurywreck of Alonissos (Hadjidaki 1996).
44. Of the 38 pre-500 BC Mediterranean wrecks only 12 can be associated to varying degrees
with Greece, either by site or cargo (quoted with the catalogue numbers of Parker 1992):
362 Dhokos (near Hydra); 2200 BC; no hull remains
1079 Sheytan Deresi (Turkey); 1600 BC; no hull remains, but Minoan influence on pottery
544 Kimi (Euboia); 15Ihc. BC; no hull remains
[*I Point lria (Peloponnese); 13mC. BC; no hull remains to date
451 Giglio Campese (Italy); c. 600 BC; minor hull remains, Greek and Etruscan cargo
599 Lindos B (Rhodos); 6mC. BC; no hull remains
835 Plemmirio C (Italy); c. 550 BC; no hull remains, Greek pottery
106 Bon Porte A (France); 550-525 BC; hull remains, Greek and Etruscan cargo
317 Circeo (Italy); 550-500 BC; no hull remains, Greek pottery
113 Breganpon (France); 6-5mC. BC; no hull remains, Greek pottery
441 Gela (Italy); late 6*-early 5mC. BC; hull remains, Greek pottery
1243 Zakynthos B (Zakynthos); 550-450 BC; no hull remains.
There are obviously drawbacks in attempting to work with understudied and under-
published material, but the listing's sole purpose is to bring out the extent of the problem.
Giglio, Bon Porte, and Gela have permitted M. Bound to document the "GBG technique",
shell-first, edge-to-edge laced strakes (cf. Bound 1991:31), but, obviously, no other ship
constructional data can be won from the entries.
45. A list suffices here: the 7"' century Amathous model British Museum A202 (Basch 1987:259
fig.559); the 6mcentury model Metropolitan Museum 74.51.1750 (id.:258 fig.558); the Black-
figure bowl in the collection of the Archaeological Institute of Heidelberg University, dated to
530-520 BC (Casson 1996:263 fig.1); the Black-figure kylix British Museum 8436, dated to c.
510 BC (Basch 1987:221-222 figs 461, 462, 464), and the fresco from the "Tomba della
Nave" in Tarquinia, dated to 490480 BC (id.:411 fig.880). On the merchantman, cf. Ericsson
1984.
46. Wallinga 1993:27-28, 41-45 errs in believing Odysseia 9.322-323 adequate for interpreting
the eikosoros as a beamy freighter since a 20-oared ship, clearly a galley, appears
elsewhere, cf. llias 1.308-311, Odysseia 2.212-213, 4.669, 778-779 (employed by Odysseus
to bring Khryseis back to her father, by Telemakhos to go to Pylos, and by the suitors to lie
in ambush). To do so he is obliged to distinguish between the freighter and the galley without
sufficient evidence. Cf. also Wallinga 1995:38-39. Despite the pages Wallinga consecrates to
the eikosoros, and the speculations of Morrison/Williams 1968:46, the type remains largely
undefined. The author requests leave to return to the question of ships in literary mentions
in greater detail elsewhere (while noting the admonitions of Kirk 1949:139, and E. Linder
[peer review of Wedde 1996; the author is grateful to Prof. Linder for the kind and
encouraging remarks at the Lamia symposion]). Although the matter has been admirably
treated by J.S. Morrison in Morrison/Williams 1968 and by Casson 1971 a reconsideration
within the framework proposed herewithin, and in other writings of the present author, will
result in variant interpretations.
47. The author hopes to return at greater length elsewhere.
48. If it is assumed that a Phoenician advance in the Levant would rapidly diffuse to the
Carthaginians, the Central Mediterranean (or Etrusco-Carthaginian, since an alliance is
known at least for the mid-6mcentury, and contact likely in Sicilian waters) situation should
be included.
49. Wallinga 1993 and 1995 argues that the modernity involved a new organisation of the fleet,
not a new design. It is nonetheless attractive to think, although impossible to prove, that the
terminological change from triakontoroslpentekontoros to trieresltesseres etc., that is from
describing the full oarage to merely that of a unit per side, would have caused an impact on
the oral traditions behind the writings of the early historians.
50. Carpenter 1948:7 claims Ameinokles built pentekontoroi, echoed by Wallinga 1993:23
(katapharkoi pentekontoro~]and 1995:41 (pentekontoroi according to Corinthian state
specifications), but both scholars provide ample evidence for a flawed understanding of
early Greek ship architecture. Williams 1958:126 suggests an amendation from "~Campaq"
to " ~ ~ L K ~ o T o u ~ " .
51. Cf. Westlake 1977, 1989:8-9 on A6y&ra1-constructions in Thoukydides. Wallinga 1993:13n3
notes Thoukydides' caution.
52. The chapters before the Methodologia give the impression of being a mixture of hearsay and
tradition, a mere cavalcade through a millennium or more as preface to the main purpose of
Thoukydides. Too great a stress should not be placed on the information they contain.
53. Morrison 1979:58, 1994228.
54. All information on the Diodoros passage from Morrison 1994.
55. Morrison/Williams 1968:160 note that Thoukydides minimizes earlier conflicts so as to
aggrandize the subject of his oeuvre.
56. Morrison/Williams 1968:129 deem the pentekontoros "the armament proper to an aspiring
pirate chief" (cf. the assessment of Polykrates by Haas 1985:37-38, 46 as a "glorified pirate
chief" - despite the recognition that he strove for thalassocracy). Basch 1977:7, Morrison
1979:60, and Lloyd 1980: 196 concurr.
57. Morrison/Williams 1968:130 are troubled by the inference from a late Ern-centurydate for the
trieres that Polykrates could acquire enough hulls to sacrifice the 40 sent with dissidents to
aid Kambyses so soon after the design was introduced. The same authors (ibid.) and
Davison 1947:20-21 suggest that Polykrates would have had a fleet of 100 triereis, retaining
slightly more than half, a force inadequate to defeat the returning dissidents.
58. Davison 1947:20 fundamentally misunderstands the trieres vs. the pentekontoros design
when claiming that the Phokaians would have employed the former had the type been
available. Wallinga 1990:137 with 137n8, 1993:68 understands the Phokaian pentekontoroi
as merchant-galleys.
59. Wedde 2000.:Section 5.5 argues that the damage caused to the Phokaian ships derived
from hulls not designed to function as weapons in themselves being used to ram the enemy
craft out of desperation when faced by the 2:l odds in the enemy's favor. In this argument,
the trieres becomes the first ship type purposefully designed to ram, and the various
preceding forms did not fill the requirements of structural strength and momentum to the
extent of allowing designing battle strategy on ramming alone. The battle of Sybota suggests
that even as late as 433 BC, after the Persian Wars had proven the worth of ramming,
commanders could still resort to the pitched hoplite battle at sea - as Kimon reasoned in
the 460's by decking the trieres. Wallinga 1995:48 claims the Phokaians employed
"'ramming' or diekplous tactics" at Alalia (yet contrast id. 1993:34 claiming the Phokaian
ships to be functionally merchantmen), developped through repeated clashes with the
Carthaginians and their allies, and that they were forced to withdraw when the enemy
introduced the trikrotos naus, a two-level 50-oared ship with a third level added in the hold
(thalamians), accomodating 20-22 rowers for a total of 70-72. These, pace Wallinga, resulted
from installing the trikrotos system on galleys of the Egyptian kerkouroitype. Much the same
arguments appears in Wallinga 1993:lll-113, although the trikrotos system is explained as

adding thalamians to Lulian biremes.
60. On the Phoeniciantrireme, cf. Basch 1969, 1977, 1980, 1987:319-334. Morrison 1979:54-56,
62 rejects Basch's case, placing the difference in the supplementary manning, and argues
against a ~re-5~-century date. Id. 1995:56-57 accepts a 5m-centuryPhoenician trireme
without outrigger.
61. Wallinga 1993:104-105, 1995:46. Davison 1947:21n3, Lloyd 1975, and Morrison 1979:53
accept Necho's triremes, with the latter two considering them of Greek inspiration, Basch
1969:232 as Phoenician. Lloyd 1972:272-275 translates knbt-ships as "Greek-styled war-
galleys".
62. Accepted as support of a Corinthian origin of the trieres by Uoyd 1975:52-53.
63. As argued by Wallinga 1993:129.
64. Scholars disagree on reading the structural contrast between Themistokles' and Kimon's
triereis. Morrison 1941:41 argues for a change of tactics, from sea warfare centered on
ramming to pitched hoplite battle at sea, with Morrison/Williams 1968:162-163 seeing this as
a result of the conservatism of Kimon, with a possible influence of the Khians sailing with him.
Morrison 1979:56 and Basch 1987:295, 301 more or less concur on seeing an eastern
influence in that the Eastern Greeks and Phoenicians favored the second approach. Wallinga
1993:177n15 considers the Kimonian triereis to be troop transports since hoplites, not
epibatai, are mentioned, and therefore rejects a return to older mores.
65. Perhaps restricted to ships-of-state such as the Salaminia and the ParaloslParalia -
although not implying that these Athenian ships were trieres at an early date.
66. Morrison 1995:54-55 argues that Luli's biremes are dikrotos triremes. This allows him to
push the invention of the trieres back before the end of the 8* century, thereby validating
Thoukydides 1.13, and ressurecting Clement 1.16.42.3.
67. The earliest evidence for oarports is the Til Barsip fresco from the reign of Tiglat-Pileser II
(745-727 BC; DeGraeve 1981:pl.XXXIX.83), followed by the Lulian ships (c. 700 BC). The
earliest Greek instance is the left ship on the Aristonothos krater (c. 700-650 BC), a single-
level, decked galley (cf. Wedde 1993).
68. There is a fragmentary report to Tiglath-Pileser Ill, written soon after 738 BC by Qurdi-
Asshur-Lamur, recounting how seaborne lonians attack three cities and are pursued by ship
(cf. Braun 1982:15). In 715, Sargon II boasts in two different inscriptions that he "caught the
lonians out of the midst of the sea, like a fish", and that he "caught, like fishes, the lonians
who live amid the Sea of the Setting Sun" (id. 15-16). The AD 2'"'-century historian Abydenos
claims that in 696-5 "Sennacherib [Sinecherim] (...) on the seacoast of the Cilician land
defeated the warships of the lonians and drove them to flight" (FGrH 685 F.5 56; id. 18, cf.
Momigliano 1934), the Abydenos version being preferred to that of Berossos of Babylon (Sd
century BC), who speaks of a landbattle (FGrH 680 F.7 p.386).
69. With the reward likely to be derision when other scholars disagree. Wallinga 1993 and 1995
are a case in point: substantial extrapolation on not always sufficiently sturdy ground leading
to questionable readings - witness the discussions on the eikosoros, the pentekontoros,
and the samaina.
70. Employing the median date approach of Parker 1992 indicates that 19 of the 23 century
BC wrecks date to 550 or later. There is also a noticeable increase in ship representations c.
560-500 BC, coinciding with "-e augmentation in the number of known wrxks, and the
attested use of the trieres in Greek fleets.
71. Humphreys 1978:170-171 also suggests a rapid increase in the number and size of mono-
functional merchantmen coupbadto the surge in trieres-building around the time of the
Persian Wars.
MICHAEL WEDDE TROPlS V1
BIBLIOGRAPHY
ANDERSON, R.C. 1962 Oared Fighting Ships from classical times to the coming of steam, Kings
Langley.
AVIGAD, Nahman 1982 A Hebrew seal depicting a sailing ship, BASOR 246, 59-62.
BASCH, Lucien 1969 Phoenician oared ships MarM 55, 139-162, 227-245.
1977 Trieres grecques, pheniciennes et egyptiennes, JHS 97, 1-10.
1980 M. le Professuer Lloyd et les trieres: quelques remarques, JHS 100, 198-199.
1987 Le Musee imaginaire de la marine antique, Athens.
BIERS, Jane C., HUMPHREYS, Sally 1977 Eleven ships from Etruria, IJNA 6 , 153-156.
BOUND, Mensun 1991 The Giglio Wreck. A wreck of the Archaic period (c.600 BC) off the
Tuscan island of Giglio. An account of its discovery and excavation: a review of the main
finds, ENALIA Suppl. 1 , Athens.
BRAUN, T.F.R.G. 1982 The Greeks in the Near East, in Cambridge Ancient History 111.3,
Cambridge (2nd ed.), 1-31.
BROODBANK, Cyprian 1989 The longboat and society in the Cyclades in the Keros-Syros
culture, AJA 93, 319-337.
CARPENTER, Rhys 1948 The Greek penetration of the Black Sea, AJA 52, 1-10.
CASSON, Lionel 1958 Hemiolia and triemiolia, JHS 78, 14-18.
1971 Ships and Seamanship in the Ancient World, Princeton.
1991 The Ancient Mariners. Seafarers and Sea Fighters of the Mediterranean in Ancient
Times, Princeton, (2nd edit., 1959').
1994 Ships and Seafaring in Ancient Times, Austin.
1995 Merchant galleys, in GARDINER/MORRISON 1995:117-126.
1996 New evidence for Greek merchantmen, IJNA 25, 262-264.
COATES, J.F., PLATIS, S.K., SHAW, J.T. (eds) 1990 The Trireme Trials 1988. Report on the
Anglo-Hellenic Sea Trials of Olympias, Oxford.
DAKORONIA, Phanouria 1987 Warships on sherds of LH I l l C kraters from Kynos, in TZALAS
1987:117-122.
DAVISON, J.A. 1947 The first Greek triremes, CQ 41, 18-24.
DeGRAEVE, Marie-Christine 1981 The Ships of the Ancient Near East (c.2000-500 BC), Leuven.
DOUMAS, Khristos 1992 01 ~oixoypacpi~q rrlq Orjpaq, Athens
DUNST, Gunter 1972 Archaische lnschriften und Dokurnente der Pentekontaetie aus Samos,
AM 87, 99-163.
ERICSSON, Christopher H. 1984 Navis Oneraria. The Cargo Carrier of Late Antiquity, Studies in
Ancient Ship Carpentry, abo.
GALESTIN, M.C. 1977 De griekse oudheden, 's-Gravenhage.
GARDINER, Robert, MORRISON, John (eds) 1995 The Age of the Galley. Mediterranean Oared
Vessels since pre-classical Times, Conway's History of the Ship vol. 2, London.
GARNSEY, Peter 1988 Famine and Food Supply in the Graeco-Roman World. Responses to risk
and crisis, Cambridge.
GILLMER, Thomas C. 1975 The Thera ships, MarM 61,321-329.
1978 The Thera ships-a re-analysis, MarM 64, 125-133.
1985A The Thera ships as sailing vessels, MarM 71, 401-416.
1985B Theories on ship configuration in the Bronze Age Aegean, in TZALAS
1985:129-138.
GRAY, Dorothea 1974 Seewesen, Archaeologia Hornerica, Band I , Kapitel G, Gottingen.
HAAS, Christian 1985 Athenian naval power before Thernistocles, Historia 34, 29-46.
HADJIDAKI, Elpida 1996 Excavation of a Classical Shipwreck at Alonnesos (5th c. BC), Enalia
Annual 1992 Vo1.4, 19961, 37-45.

HOCKMANN,Olaf 1985 Antike Seefahrt, Miinchen.
1989 Some thoughts on the Greek pentekonter, in TZALAS 1989:207-220.
JACOBY, Felix 1904 Das Marmor Parium, Berlin.
KALLIGAS, Petros 1987 Early Euboean ship building. in TZALAS 1987:77-83.
KIRK, Geoffrey S. 1949 Ships on Geometric vases, BSA 44, 93-153.
LANDSTROM, Bjorn 1969 Seglande skepp, Stockholm.
LLOYD, A.B. 1972 Triremes and the Saite Navy, JEA 58, 268-279.
1975 Were Necho's triremes Phoenician?, JHS 95, 45-61.
1980 M.Basch on triremes: some observations, JHS 100, 195-198.
MANNING, Sturt W. 1988 The Bronze Age eruption of Thera: absolute dating, Aegean
chronology and Mediterranean cultural interrelations, JMA 1, 17-82.
MARINATOS, Spyridon 1974 Excavations at Thera VI, Athens.
MOMIGLIANO, Arnaldo 1934 Su una battaglia tra assine e greci, Athenaeum 12, 412-416.
MORGAN, Lydia 1988 The Miniature Wall Paintings of Thera. A Study in Aegean Culture and
Iconography, Cambridge.
MORRISON, J.S. 1941 The Greek trireme, MarM 27, 14-44.
1979 The first triremes, MarM 65, 53-63.1979
1980 Hemiolia, trihemiolia, IJNA 9, 121-126.
1994 Review of Wallinga 1993, MarM 80, 227-228.
MORRISON, J.S, COATES, J.F. 1986 The Athenian Trireme. The history and reconstruction of
an ancient Greek warship, Cambridge.
MORRISON, J.S., WILLIAMS, R.T. 1968 Greek Oared Ships 900-322 BC, Cambridge.
ORMEROD, H.A. 1924 Piracy in the Ancient World. An essay on Mediterranean History, Liverpool
(reprint 1978).
PARKER, Anthony J. 1992 Ancient Shipwrecks of the Mediterranean & the Roman Provinces,
BAR lnternational Series 580, Oxford.
PENNAS, Haralarnbos, VICHOS, Yannis, LOLOS, Yannos 1996 Point lria wreck 1992, 1993,
Enalia Annual 1992 po1.4, 19961, 4-5, 6-31.
POPHAM, Mervyn 1987 An early Euboean ship, OJA 6,353-359.
RANKOV, Boris 1996 The Second Punic War at sea, in Cornell, Tim, Rankov, Boris, Sabin, Philip
(eds),The Second Punic War. A Reappraisal, BlCS Suppl. 67, London, 49-57.
SCHWEITZER, Bernhard 1955 Zum Krater des Aristonothos, RM 62, 78-106.
SLEESWYK, Andre Wegener 1991 Mechanisms and tactics of ramming ships, in TZALAS
1991:429-449.
SNODGRASS, 1983 Heavy freight in Archaic Greece, in Garnsey, Peter, Hopkins, Peter,
Whittaker, C.R. (eds), Trade in the Ancient Economy, London, 16-26.
SPATHARI, Elsi 1995 App&vl<wvraq m o ~ p o v o .To d o l o q v EMqvi~rjT & ; x M ~Athens.
,
STIEGLITZ, Robert R. 1984 Long-distance seafaring in the ancient Near East, Biblical
Archaeologist 47.3, 134-142.
TELEVANTOU, Christina A. 1994 A~pwnjpiOrjpaq. 01 roixoypacpi&q q 6vsi~rjo i ~ i aAthens. ,
TILLEY, Alec F. 1989 Warships of the ancient Mediterranean, in TZALAS 1989:429-440.
TORR, Cecil 1894 Ancient Ships, Cambridge (in the Argonaut edit., Chicago, 1964).
TZAHOU-ALEXANDRI, Olga 1987 Contribution to the knowledge of 8th century B.C. ship
representations, in TZALAS 1987:333-361.
TZALAS, Harry (ed.) 1985 Tropis I. 1st lnternational Symposium on Ship Construction in Antiquity,
Piraeus 1985 [printed 19891.
1987 Tropis 11. 2nd lnternational Symposium on Ship Construction in Antiquity, Delphi
1987 [printed 19911.
1989 Tropis 111. 3rd lnternational Symposium on Ship Construction in Antiquity, Athens
1989 [printed 19951.
1991 Tropis IV. 4th International Symposium on Ship Construction in Antiquity, Athens
1991 [printed 19961.
1993 Tropis V. 5th International Symposium on Ship Construction in Antiquity, Nauplion
1993 [printed 19991.
WACHSMANN, Shelley 1980 The Thera waterborne procession reconsidered, IJNA 9, 287-295.
1981 The ships of the Sea Peoples, IJNA 10, 187-220.
1991 Bird-head devices on Mediterranean ships, in TZALAS 1991:539-572.
1995 Paddled and oared ships before the lron Age, in GARDINERIMORRISON 1995:
10-35.
1998 Seagoing Ships and Seamanship in the Bronze Age Levant, Ph.D diss.
WALLINGA, H.T. 1990 The trireme and history, Mnemosyne 43, 132-149.
1993 Ships and Sea-Power before the Great Persian War. The Ancestry of the Ancient
Trireme, Leiden.
1995 The ancestry of the trireme 1200-525 BC, in GARDINER/MORRISON 199536-48.
WARREN, Peter, HANKEY, Vronwy 1989 Aegean Bronze Age Chronology, Bristol.
WEDDE, Michael 1991 Rethinking Greek Geometric art: consequences for the ship
representations, in TZALAS 1991573-596.
1993 Decked vessels in early Greek ship imagery, in TZALAS 1993:505-526.
1996 From classification to narrative: the contribution of iconography towards writing a
history of early Aegean ship building, Mediterranean Historical Review 11.2, 11 7-164.
1999A Bronzezeitliche Schiffsdarstellungen. Vorgeschichte, Entwicklung und
eisenzeitliches Weiterleben der friihen Schiffsbaukunst Griechenlands, in Chrysos,
Evangelos, Letsios, Dimitrios, Richter, Heinz A., Stupperich, Reinhard (eds),
Griechenland und das Meer. Beitr3ge eines Symposions in Frankfurt im Dezember
1996, Peleus. Studien zur Archaologie und Geschichte Griechenlands und Zyperns,
Band 4, Mannheim und Mohnesee, 45-64.
19998 War at sea: the Mycenaean and Early lron Age oared galley, in Laffineur, Robert,
(ed.), Polemos. Le contexte guerrier en Egee a IPge du Bronze. Actes de la 7e
Rencontre Egeenne internationale, Liege, 465-474.
2000Towards a Hermeneutics of Aegean Bronze Age Ship Imagery, Peleus. Studien zur
Archaologie und Geschichte Griechenlands und Zyperns, Band 6, Mannheim und
Mohnesee.
WESTERBERG, Karin 1983 Cypriote ships from the Bronze Age to c.500 BC, SlMA Pocket 22,
Goteborg.
WESTLAKE, H.D. 1977 LEGETAI in Thucydides, Mnemosyne 30,345-362.
1989 Studies in Thucydides and Greek History, Bristol.
WILLIAMS, R.T. 1958 Early Greek ships of two levels, JHS 78, 121-130.
LIST OF FIGURES
Fig. 1 Lefkosia 194711-1611,from Wachsmann 1991565 fig. 20A.

Fig. 2 Metropolitan Museum 74.51.51 1 , from id.:fig. 208.
Fig. 3 British Museum 1926.6-28.9, from id.:fig. 20C.
Fig. 4 Aristonothos krater, from Basch 1987:233fig. 482 right.
Fig. 5 Seal of Oniyahu, drawing by author from Avigad 1982:59 fig. 1.
Fig. 6 Seal from Siphnos, from Basch 1987:248fig. 522.
Fig. 7 Situla from Chiusi, from id.:409 fig. 871.
Fig. 8 British Museum H230, from Hockmann 1985:49 fig. 33.
Fig. 9 Corinthian clay plaque, from Casson 1971:fig. 98.
- - -- - -
OF A NARRATIVE
--. - -- .
OF EARLY GREEK SHIP
. --- - ARGHITECTUHZ
- -- .- --
* Fig. 1 P
Fig. 6
,---,
Fig. 2
Fig. 7
Fig. 3
I Fig. 8
Fig. 4
Fig. 5 Fig. 9
ABSTRACT IIEPIAHUJH
H MOYZIKH Tf2N OAAAIZOllOPf2N W N THE APXAIOTHTAI
E q v n p a ~ u ~~f w
i 8aAaaawv
v Kal n o ~ a p ~ wps~acpop6v
v ~ a q ~v apxato-
a
WTa, E ~ E W P E ~ T)@fiCJl~q
O T) ( P o V T ~ T L K ~ ~Kal Q 0pyavlKfi poual~fi.BOQ~OUUC
T O U ~~wnqAa~sq va Ppio~ouvTOV puepo, napsixs, o ~ a v~ p s ~ a < o ~Eva av,
~ 6 6 1 snt~olvwviaq
~a Kat svicqus TO ~ 8 1 ~Ent 6 . n M o v - ~ a~l a n w qavs<dpq-
Ta an6 aunj q <<poual~fl &pyaaiaq>,-Q86Aaaaa, ps T L cs~wp~o-rCq ~ aKou-
o-rt~Cqq q lliloTq~&qKaL pu0oAoy~~Cq 6~aouv6Cas~q, a n o - r ~ k a snhouolo
nsliio avarrrucqq q q ~ M q v i ~ f i qpwpai~qq , Kat alyurrr~a~fiq pouot~fiq
cpavraoiaq
SHIP'S MUSIC IN THE ANCIENT WORLD
In the practice of sea and river transport in the ancient world, vocal and
instrumental music was regarded as useful. It helped rowers to find and keep
their rhythm, it provided a signalling code where necessary, and it acted as
a psychological tonic. More or less independently of this "work music",
moreover, the sea, with its distinctive acoustical properties and mythological
associations, provided a rich field for the exercise of the Greek, Roman, and
Egyptian musical imagination.
Dr. Richard Witt

The Open University
25 Kifisias ave.
Athens 115 23
Greece

On Ship Construction in Antiquity

Uploaded by

Copyright:

Available Formats

On Ship Construction in Antiquity

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

On Ship Construction in Antiquity

Uploaded by

Copyright:

Available Formats

I& VI HELLENIC INSTITUTE

FOR THE PRESERVATION

edited by Harry Tzalas

Published by the Hellenic Institute for the Preservation of Nautical Tradition

Cover: General plan of the two Greek shipwrecks of

THE EXECUTIVE COMMITTEE

Mrs. Danielle Auffray's communication at the 6m Conference: "Un site maritime

Dr. Dimitris Kourkoumelis' abstract on the "Pyramidal stone-anchors found at

An update of the work on the Uluburun Shipwreck was presented at the 7m

The publication of this volume was nearly complete, when we were

An authority in the field of marine archaeology and oared ships in particular,

I&paoptch~a-re,K& Ileptcpsp~tapxa, Your Reverence, Mr. District

'Orav anocpaotoa nptv Cva wove va When a year ago I decided to

I must not forget to mention the

Le Phare d'Alexandrie, ce bâtiment gigantesque, provoqua dès sa

Le Phare est souvent cité dans les textes anciens et contemporains. On

Un siècle après sa destruction, le sultan Qaitbay construisait sa citadelle

De nombreux auteurs ont écrit sur le Phare. On trouve des descriptions

Avec l'invention du scaphandre autonome, le Phare d'Alexandrie

En 1968, l'Unesco a envoyé la spécialiste d'archéologie sous-marine

Kamel Abou el Saadat, H. Frost a établi le premier plan du site et a localisé

En 1993, le site se trouvait menacé par la construction d'un mur de béton

Ce site immergé, de 6 à 8 m de profondeur, couvre 2,25 ha et comporte

La sculpture n'est pas absente du site : on trouve des statues colossales

Une seule inscription grecque a été découverte. Elle se trouvait sur un

Les résultats très importants révélés par la prospection ont encouragé,

Dans les années 1995 et 1996, il y a eu quatre saisons de fouilles :

Après la prospection du site, il a été décidé de cartographier en détail

Comme on l'a vu précédemment, le site était constitué de plus de 2000

D'où viennent ces ruines ? C'est la principale question qui se pose.

L'abondance des monuments égyptiens sur le site s'explique par le désir

Au fur et à mesure de la fouille l'idée que certains vestiges du site

Après toutes ces constatations, une question se pose en Egypte : est-il

justement l'obélisque 2001, 2026 a et b : daté sous I'eau de l'époque de

Mohamed Mustapha Abd-El Maguid

BLOCS RELEVES EN 1994

BLOCS RENFLOUES EN 1995 ET 1996

CARTE DES BLOCS PAR POIDS

The Ramesside information, a long term counterintelligence, has misled

There might be some jest in the presentation of these men. Their

decided, with poetic license, to use it in his composition. We can not

Altman, A. 1977 The Fate of Abdi Ashirta. UF 9:l-10.

Fig. 1 AkkoIKition .Fan,, type boat from Carmel Ridge.

Fig. 1: AkkoJKition Fan type boat from Carmel Ridge

Fig. 2:AkkdKition Far; type boats on alte from Akko

A salvage excavation initiated in April 1996 on Hehaluz Street in Lod led

The mosaic floor consists of two colorful rectangular carpets - the

The square middle panel, which is surrounded by a broad braided band,

A damaged panel with a decorative frame adjoining the north carpet on

Fig. 1 General view of the mosaic floor, looking south.

On doit à R. Burlet une heureuse définition de la voile latine

Les opinions relatives à l'origine de cette voile, si importante pour

En 1932, G.S. Laird Clowes n'hésite pas à affirmer : <<Thetriangular

J. Poujade, en 1946, se montre l'avocat ardent de l'origine arabe de la

solution du problème, et cela tant du point de vue arabe qu'européen ;

P.H. Dolley, en 1949, estime qu'il est impossible d'attribuer l'introduction

R. Leb. Bowen, après avoir écrit : <<theevidence seems to indicate that

Dans un livre important publié en 1963, G.F. Hourani raisonne de la

Dans ce débat, l'iconographie a joué un rôle considérable : on peut voir

Ces illustrations ont perdu une partie de leur importance depuis la