Knidos: topography for a battle
p. 435-451
Remerciements
I would like to extend warm thanks for their help to the editors of this volume, Riet van Bremen and Jan-Mathieu Carbon, as well as to Simon Hornblower and Jaime Curbera.
Texte intégral
1In early August 394 BC, in the sea off Knidos, was fought one of the major sea battles of antiquity. This battle had long-term consequences for the history of Greece and Asia Minor. The fleet of Sparta and her allies under Peisandros was destroyed by the ‘Persian’ fleet under Konon and Pharnabazos. Xenophon briefly mentions that Agesilaos, the king of Sparta, then in central Greece, received the news of the battle at the time of a solar eclipse which is that of 14 of August 3941. Thus the battle itself took place during the first days of the same month. If the chronology is certain, the development of the battle has not yet been clarified.
2Xenophon’s report is brief and does not tell much about the conditions in which the battle was fought2. A fragment of Philochorus gives information that can also be found in Diodorus (see below)3. Other fourth-century authors provide mere allusions to the battle4. Diodorus offers more details5. However, on a crucial issue his presentation remains confusing. Modern scholarship – W. Judeich in his Kleinasiatische Studien (1892) and G. Barbieri in his Conone (1955) – has underlined the difficulties involved in his narrative6. Besides this, Polyaenus makes mention of a trick devised by Konon to concentrate his enemy’s attack on his right wing, but this adds little to the general comprehension of the battle7. The summaries by Cornelius Nepos, Justin and Orosius are full of rhetoric and of little or no historical value8.
3This paper will not propose a global reconstruction of the strategic developments that led to Peisandros’ defeat or of the consequences of the Spartan defeat. It will focus on topographical matters and will try to make sense out of the admittedly confusing report of Diodorus, by exploiting other ancient sources that give us a valuable description of the Hellenistic and Roman town of Knidos. However, before coming to the main point, it will first be necessary to summarize the debate relating to the topography of the whole city of Knidos, viz. the question of the relocation of Knidos, and to evoke the conditions under which the two rival fleets fought the battle.
4The problem of the topography of the city of Knidos has been debated among scholars in the last half century. Until G. E. Bean and J. M. Cook’s 1952 article in the Annual of the British School at Athens everyone believed that there was only one possible site for the town of Knidos: that of Cape Tekir, at the tip of the Knidian peninsula. It was there that the Hellenistic and Roman city of Knidos was located. The site is well known from its impressive remains. In the modern period it has been excavated in four separate phases: between 1833 and 1837 by a French team under C. Texier, from 1857 onwards by a British one under C. T. Newton, between 1967 and 1976 by the American team of I. Love, and lastly since 1997 by the joint British and Turkish team led by I. Jenkins and R. Özgan9. A strong fortification, splendid constructions and Hippodamian town-planning are two of its main characteristics.

The Knidia and its neighbouring regions.
5In 1952, however, Bean and Cook questioned the view that the town of Knidos had always been at Tekir10. Their hypothesis was based on an analysis both of the surface remains in the whole Knidian peninsula and of the ancient literary and epigraphic sources. Their conclusion was that the town at Cape Tekir was founded only in the late Classical or early Hellenistic period. Previously, the capital city was located at Datça, in the centre of the Knidian peninsula. This was the origin of the debate11. Ever since, scholars have been divided between those who have accepted their theory and those who have rejected it. At first, Bean and Cook’s view found acceptance (among others by as distinguished a scholar as L. Robert), while later their theory tended to be rejected12. In more recent years, however, the hypothesis of relocation has again gained ground13. It is thus necessary to review briefly the main arguments here.
6Bean and Cook based their theory of relocation mainly on archaeological arguments. For them, the site of Cape Tekir offered no pre-Classical remains and the town looked as if it had been founded only in Hellenistic times. There was, on the other hand, plenty of Archaic or Classical material on the site near the modern Turkish village of Datça, in the centre of the Knidian peninsula. This is also the site of a rich plain, the best of the whole Knidian territory. For Bean and Cook, there was no doubt that Datça was the site of Old Knidos. The detail of their argument needs, however, to be revisited, especially the lack of pre-Classical remains at Cape Tekir. Even so, the difference between the development of the two sites remains striking.
7Bean and Cook also took into account literary sources. Plutarch mentioned that Kimon’s fleet for the campaign of the Eurymedon c. 469-466 BC ‘set sail from Knidos and Triopion’14. Above all, Thucydides drew a clear distinction between the town of Knidos and Triopion15. In 412/411 BC, the Lacedaemonian Hippokrates divided his fleet into two squadrons: ‘When they were informed of the arrival (of the twelve ships), those in command in Miletos ordered him to leave half of his ships to guard Knidos, while the others would cruise round Triopion and seize the merchantmen coming from Egypt. Triopion is a promontory that projects from the Knidian territory and sacred to Apollo’. Afterwards, the Athenians coming from Samos, that is from the north, first took by surprise the six ships left at Triopion, then attacked Knidos. For Bean and Cook, a detail of this account was worth considering: it took one night for the crews from Triopion who escaped with their lives to reach Knidos and reinforce the garrison, which would not make sense if at that time Triopion and the town of Knidos had been one and the same location16. Finally, Pseudo-Scylax, admittedly a text of the first part of the fourth century, makes a clear distinction between the ‘sacred promontory of Triopion’ and the ‘Greek town of Knidos’17. The existence of two sites, Triopion and Knidos, is beyond doubt. There is no need to argue the point anew at greater length, since Bean and Cook’s arguments remain valid.
8Bean and Cook also had to face the difficulty of the location of Triopion, the common sanctuary of the Dorians. If for them there was nothing of note at Cape Tekir, there was nevertheless another site worth considering. They suggested that Triopion might have been at Kumyer, about ten kilometres to the east of Tekir, with suggestive Archaic remains and one Archaic inscription. However, this site does not correspond to the description given by Herodotus and Thucydides, who both clearly locate Triopion at a site at the tip of the Knidian peninsula18. Another location was proposed by D. Berges. While strongly supporting Bean and Cook’s ‘Old Knidos’ theory, he preferred to identify Triopion with the sanctuary of Apollo at Emecik, east of Datça, which dates back to the Archaic period19. The excavations he conducted on this site have produced an impressive amount of material20. Besides, two Hellenistic inscriptions from Emecik mention the god Apollo Karneios21. However, this sanctuary of Apollo Karneios (not Triopios) does not fit the testimony of our ancient literary sources, as he himself recently acknowledged in his final publication of the site22. This led him to conclude that the precise location of the Triopion should remain open until some new discovery sheds new light on the issue23.
9The present author has already contributed to the debate24. On the issue of relocation, there is every reason to follow Bean and Cook: both the archaeological remains of Datça and the analysis of literary sources support the hypothesis of an Old Knidos vs. a new Knidos. There definitely was a transfer of the capital from Datça to Tekir. The recent excavations at Datça-Burgaz have proved the existence of a Classical town on this site. Stretching along the sea and about one kilometre long, it had an acropolis overlooking the sea, strong Classical fortification walls and several harbours. Although the origin of the site dates back at least to the seventh or sixth century, there is no evidence of occupation later than the fourth century25. In addition, since 1952 and Bean and Cook’s initial publication, several new inscriptions have come to light, which also vindicate the relocation theory. Three decrees and a treaty from Datça (two of them securely dated) are from the first part of the fourth century26. Other Knidian public documents have been published by W. Blümel, this time from Tekir. While we still lack a full photo coverage of these inscriptions, the decree that W. Blümel considers to be one of the earliest of the Tekir group is to be dated after 350, which favours the theory of the transfer of sites probably at some date around 330-320 BC27.
10The only argument that could be adduced against this theory would be the possibility that the Datça stones were in fact transported from Tekir to Datça in modern times. The site of Tekir was abandoned in the medieval and modern periods. Indeed, like those of many other ancient sites along the coast of Asia Minor, the ruins of Knidos-Tekir were obviously used for many centuries as a stone-quarry. Stones from Tekir were transported to other sites, and some of them may have found their way to Datça, which was the place where life concentrated in modern times28. However, it would still be necessary to prove that Datça was the main destination of stones transported from Tekir. It would also need to be explained why the stones bearing Archaic or Classical public inscriptions coming from Datça would have been transported from Tekir, while the Hellenistic ones remained on the spot. The fact that the decrees dated to the first half of the century come from Datça, while those seemingly dated to the latter part of the same century come from Tekir, where besides so many later inscriptions have been found, is a strong argument for rejecting the transportation hypothesis. Thus the epigraphic dossier vindicates Bean and Cook’s relocation theory.
11As for Triopion, following what seem the obvious testimonies both of Herodotus and Thucydides the present writer has proposed to locate it at the tip of the peninsula. Triopion was not only the site of a sanctuary but must have been the site of a little, homonymous town, Triopion, which is mentioned by Stephanus29. Arrian alludes to the capture of ‘Kos and Triopion’ by Orontobates in 333, which would not make sense if Triopion had not been at that time a town, or at least a valuable fortified stronghold30. Admittedly (thus Bean and Cook), the site at cape Tekir has often been considered to lack any remains from the Archaic and Classical times. However, the American and Turkish excavations on the mainland have brought to light some (poor) Archaic material31. The problem is to determine what kind of remains we may hope to find and where to search for them. On the site of Tekir, two types of architectural structure are clearly visible, the first on the former island, the other on the mainland, which used Hippodamian planning. The Hellenistic and Roman type of masonry is dominant, often of isodomic or pseudo-isodomic type. An ‘older style’ is clearly visible on the island, with its circuit wall that parallels that of Classical Nisyros32. On satellite photos a Hippodamian grid can also clearly be traced on the southern part of the island situated at the tip of the peninsula. But most of the terrace walls on the island, in rough polygonal masonry, look older than what can be observed on the mainland part of the town. In fact, they do not differ much from what can be observed regularly on pre-Hellenistic Karian sites in the hinterland of Halikarnassos or Mylasa33. Some (admittedly rare) black varnish potsherds are visible on the surface in this part of the town. For what they are worth (they clearly are no substitute for a regular survey or for excavations, which would be the only way to achieve a precise chronology), these notes make the existence of a small Archaic and Classical town on the island at Tekir an even more attractive hypothesis. As for the Triopion sanctuary, there is good chance that it was located on the island. As for the competitions, they must have taken place on the mainland, where there was enough room to accommodate them, whereas this is obviously not true of the tiny island. So far, no Archaic or early Classical inscription has ever been found at Tekir. This does not mean that this will always be so: if indeed Cape Tekir was the site of the Triopion, some early dedication to Dorian Apollo might come to light one day.
12To conclude, in the present state of our knowledge every argument, whether archaeological, literary or epigraphic, supports the theory of a relocation of the capital from Datça to Tekir at the end of the Classical period or in the very first years of the Hellenistic age. Besides, Cape Tekir must have been the site of the Triopion, the common sanctuary of the Dorians. Before the Hellenistic period, it must also have been that of a small town which took on more importance during the fourth century, until it seemed logical to transfer the capital of the city and build there a splendid new town.
13This setting is necessary to understand the history of the battle fought in 394 BC, known as ‘the battle of Knidos’. Conversely, the topographical aspect of the interpretation of this battle is another argument in favour of the relocation or two-town theory. After the Peloponnesian war, the Spartans wanted to maintain their dominion over the Aegean cities and had to challenge the Persian empire for control over the Greek cities of western Asia Minor. In that war, the naval operations proved once again to be the key to success. On the instigation of Pharnabazos, the Great King decided to make a special war effort. He summoned the Athenian Konon, who by then had taken refuge with king Evagoras of Cypriot Salamis. The kernel of the new fleet was composed of Cypriot ships. The aim was to man a fleet of 100 ships. Before that goal was reached, Konon started for Cilicia with 40 ships. According to Diodorus, Artaxerxes’decision and Konon’s departure both date to 399/834. In fact, everything invites us to consider that he departed only in 397/6.
14From Cilicia, Konon moved to Kaunos35. The importance of that port was clear to everybody. In 440, at the time of the siege of Samos, Perikles himself had come to Kaunos to stop the possible arrival of a Phoenician fleet36. In 412, 27 Peloponnesian triremes, fearing the Athenian fleet, came to Kaunos waiting for further developments in the Aegean37. Kaunos played an important role in the negotiations between the Peloponnesians and Tissaphernes in winter 412/138. Going to Kaunos and Phaselis, Alkibiades could boast that he had blockaded the arrival of the Phoenician fleet and won the favour of Tissaphernes39.
15At first, in 397, Konon was blockaded in the port of Kaunos by the Lacedaemonian navarch Pharax40. His 40-ship fleet could not match the 120 vessels of his opponent. However, Konon resisted the Lacedaemonian attack. Artaphernes and Pharnabazos sent him a significant reinforcement and Pharax had to abandon the siege and come back to Rhodes. But Kaunos was only a starting point for Konon. From there, he could launch secret negotiations with the Rhodians41. He had now a fleet of 80 ships and he headed towards the Chersonese42. This movement possibly aimed at attracting the Spartan fleet out of the port of Rhodes. But the result of the previous secret negotiations with the Rhodians was that now the Rhodian ports remained closed to the Spartans and that they were open to Konon’s fleet43. The event took place in 396. This was a major setback for the Spartans, given both the loss of such a strategic position as Rhodes and that of its neighbouring places. Loryma, at the tip of the Chersonese, was thus also in Konon’s hands and this fits well with the fact that that port was his main offensive base in 394. With the loss of Rhodes, the Spartan fleet also lost the contribution of the c. 10-20 Rhodian ships, while we may suppose that Konon’s was reinforced by the same number. Thus the Rhodians and Konon were able to seize an important grain convoy sent by the Egyptian pharaoh to the Lacedaemonians44. The Rhodian leaders’attitude was unexpected, insofar as Rhodes had an oligarchic government since 411. Yet it was the leaders of the oligarchic faction who opened the doors of the city to Konon (before being themselves overthrown by a revolution in 395)45.
16On all these aspects, our main source is Diodorus, whose narrative has been the subject of much debate. For instance, his chronology of the Spartan navarchs does not match what we know from other sources. As often, Diodorus has compressed events that really covered several years into one single year. In comparison, his (incorrect) dating of the battle to the Athenian archonship of Diophantos (395/4), instead of that of Euboulides (394/3) as correctly stated by Lysias, is a minor mistake46. For this period the Oxyrhynchus historian offers highly valuable additional information. It reveals how Konon managed to favour a democratic upheaval in Rhodes in 395, although he pretended to take no part in it because he was in Kaunos47. It also shows Konon’s financial weakness: he had difficulties paying his troops and sailors and had to face an upheaval both in Rhodes and Kaunos48. Unfortunately, for the events in Asia the text does not go any further. For the next developments, which led to the battle of Knidos, we are left with Xenophon and Diodorus.
17Thus, there was a kind of stalemate of more than three years, which corresponds to the ‘war around Rhodes’ mentioned by Isocrates49. From a strategic viewpoint, between 397 and 394 the Lacedaemonians did effectively prevent any access to the Aegean from Konon’s fleet, and this was no mean success. But, because of Konon’s skill, the Spartans missed their chance to crush his fleet when they still had the upper hand. Then the balance of power altered: it was now Konon who possessed superior means and who was in a better strategic position. While originally he was locked in Kaunos, through their loss of Rhodes and Loryma – and probably the rest of the Chersonese – his opponents were now pushed in the corner and kept only Knidos, the last position from which they might block access to the Aegean.
18Indeed, the balance of forces was now clearly in favour of Konon. Xenophon does not mention the numbers of ships, but says only that Konon’s fleet outnumbered that of Peisandros50. Diodorus gives figures: 85 ships for Peisandros, more than 90 for Konon51. There has been a scholarly disagreement on the numbers of triremes of Konon’s and Pharnabazos’ fleet. It may be suspected that the figure was significantly higher than 90, but no precise figure can be reached. If Konon’s and Pharnabazos’ fleet really counted well over 90 ships, then this large fleet could probably not be concentrated in one single port but must have been divided into two main bases, Loryma and Rhodes, with Kaunos as a possible supply base (in 396 and 395, Konon’s forces were not all concentrated at Rhodes: at that date, one section of his fleet and of his forces was still in Kaunos). This was justified by strategic, but also by technical reasons (for instance, having enough food, water and lodging for the rowers and soldiers). To face their enemy, Konon and Pharnabazos would have needed to concentrate their own forces beforehand.
19As previously mentioned, the main sources for the battle are Xenophon and Diodorus. While Xenophon probably obtained reports from men having participated in the battle, he remains deliberately vague on this Spartan defeat. His version obviously aims mainly at emphasizing the courage of Peisandros, the Spartan navarch. Fighting against the odds with a fleet outnumbered by the enemy, and with his allies at his left wing fleeing from the very beginning of the battle, Peisandros could only fight to the end and die with his honour intact, which is what he did. It is quite possible that Xenophon wanted to respond to criticism made against Agesilaos’choice of Peisandros as commander of the fleet52, but this is certainly not the image of Agesilaos that Xenophon wanted to popularize.
20While Diodorus’ account is more detailed, it cannot be said that his narrative produces a clear report of the battle. His source must have been a fourth-century historian. Theopompus ended his Hellenica with the battle of Knidos, but we have no idea how he reported the battle53. Thus Ephorus is considered to be the main or sole source of Diodorus for these events54. But Ephorus himself (born around 405) must have based his narrative on one of his predecessors and the Hellenica Oxyrhynchia is the most likely source. The fragments of the Hellenica Oxyrhynchia we possess suffice to prove that, whoever was its author (Cratippus?), he was first rate55. There is no reason to doubt that his description of the battle of Knidos was of the same level. Thus we are left with a sequence Hellenica Oxyrhynchia – Ephorus – Diodorus, where Diodorus gives us only a third-hand version of the battle. But, if indeed Ephorus was his source, Diodorus’ narrative contradicts Polybius’ view, which compared Ephorus’precision in the description of sea-battles with his obscurity in rendering land battles56. But there is more: the examples Polybius gives to illustrate Ephorus’ clarity in the description of sea-battles are precisely the battles of Cyprus and Knidos. Admittedly, none of these qualities can be traced in Diodorus’ reference to the battle of Knidos, which at best gives us a brief and poor digest of the original version. A sound conclusion is thus that the weakest link in our tradition is Diodorus, not Ephorus; but one may also wonder whether Ephorus’ clarity in the description of sea-battles, and especially of the battle of Knidos, was not simply due to the quality of his model, the Hellenica Oxyrhynchia.
21Who took the initiative to fight? As pointed out by Barbieri, and despite Diodorus’lack of clarity, there should be no doubt that the strategic initiative came from Konon, for ‘when he was informed that the Spartan fleet was at Knidos’ he moved his own fleet towards the north57. If Diodorus is to be followed, it is precisely at that time that Peisandros decided to leave Knidos and to move his fleet to another location. Starting from Knidos, Peisandros would have sailed to ‘Physkos of the Chersonese’ and from there he would have attacked Konon’s fleet. Diodorus’ wording parallels the way he mentions ‘Loryma of the Chersonese’ or ‘Bybastos (sic, for Bybassos) of the Chersonese’58. There seems to be no doubt that Diodorus thought of Physkos, the port in the bay to the north-east of the Chersonese, the present day Marmaris.
22Diodorus’ account poses many questions. P. Foucart thought that Peisandros wanted to reach Physkos, but that this movement was prevented by the approach of the enemy. However, this clearly contradicts the Greek. Peisandros not only ‘aimed at reaching Physkos’, he actually reached that port, whatever it was59. L Pareti considered that it was on Peisandros’ way back from Physkos, in the sea off the island of Syme, that the two fleets would have met60. Though clearly absurd, Peisandros’ movement would have been without consequence as in the end the battle would have been fought on the west side of the Chersonese, off Syme and the Knidian coast. J. Buckler has also recently taken Diodorus at face value, asserting that Peisandros moved from Knidos to Physkos-Marmaris61. This would be a proof of his lack of experience; and indeed this is underscored by Plutarch, for whom Agesilaos’choice was based on favouritism, as Peisandros was his wife’s brother, rather than on the sense of the interest of his country, as he passed over older and more experienced men62. Konon and Pharnabazos would have grasped this unexpected opportunity. Fighting in the open sea, Peisandros would have been defeated. According to Buckler, the battle took place off the east coast of the Chersonese63. It should be observed that, as the majority of the sailors of Peisandros’ fleet who managed to swim to the coast were not taken prisoners, we should admit that not only Physkos, but the Chersonese, or at least its northern portion, was still in Spartan hands. It would also remain to be explained how the ships that escaped managed to go another time around Loryma and the southern tip of the Chersonese – this time from the east to the west – and to come back safely to Knidos.
23These reconstructions raise so many difficulties that they are clearly untenable, all the more as our sources, starting from Xenophon, and continuing with Isocrates, Deinarchus, Polybius, Plutarch, Pausanias (with the precision that the battle took place at the foot of Mount Dorion), and Cornelius Nepos repeatedly make mention of a battle that took place ‘near Knidos’64. Unless we consider that the Chersonese had by then become Knidian territory – a bold hypothesis, but one that would not solve all the difficulties, see below – all these authors would have been mistaken and the battle should have been called the ‘battle of Physkos’ or ‘of the Chersonese’. But is this really what we should believe?
24In fact, as Judeich and Barbieri clearly saw, it is highly unlikely that Diodorus’Physkos was Physkos-Marmaris, and this for several reasons65. First, Physkos-Marmaris was not in the Chersonese, the ‘peninsula’. Starting with Judeich, all recent research has also proved that the ancient ‘koinon of the Chersonesians’ did not encompass Physkos (to the east) or Kedreai (to the west)66. Diodorus’expression ‘of the Chersonese’ is suitable for Loryma and Bybassos, but cannot be a correct definition of the location of Physkos-Marmaris, unless we admit that Diodorus did not quote a fourth-century source but spoke very loosely. But strategic reasons also make this solution unlikely. Reaching Physkos-Marmaris after starting from Knidos was simply out of the question for Peisandros. Sailing around Rhodes to the south with a huge fleet along enemy coasts and in the open sea was impossible. For Peisandros would have had to pass the channel between two of the main bases of his enemies, Loryma and Rhodes. Physkos-Marmaris being also neighbour to Kaunos, he would have reached a port right within the triangle formed by the three main bases of his enemy. This would have been perfect nonsense since at any moment, starting from its bases, Konon’s fleet could have attacked the long file of Peisandros’ ships and cut it into pieces. Besides, it should be mentioned that, even though the bay of Physkos provides an excellent harbour, the straits giving access to it are very narrow. Indeed, the pass was easy to defend, but losing it, even to a small force, would have been sufficient to provoke the loss of the whole fleet, not to mention that Physkos could have easily been attacked by land. Leaving the excellent base of Knidos (out of reach for an attack by land) for Physkos-Marmaris would have been nonsense. Peisandros may have lacked experience, but this does not mean that he was a fool. And had he been one, his enemies would have undoubtedly seized the opportunity to utterly destroy his fleet during his movement around the Chersonese.
25Is there a solution to this difficulty? Should we consider that Peisandros sailed to another, unknown Physkos on the west coast of the Chersonese? This was Judeich’s suggestion (followed by Barbieri), who considered that Physkos must have been some small port ‘to the south-east of Knidos’, from where Peisandros could have a view to the south, from where the danger came67. Judeich pointed out that Strabo made mention of another Physkos, near Mylasa, and thus he considered that the name must have been more or less common68. However, if not definitely proved, it is very likely that the whole Chersonese fell to Konon in 396. This assumption would be fully justified by the mention of Konon’s fleet moving ‘towards the Chersonese’, which preceded the seizure of Rhodes, but also by the fact that in 394 a large body of Konon’s fleet was anchored at Loryma, at the tip of the peninsula69. This port was a key strategic point and the bay of Loryma was later the site of an important Rhodian naval base70. But it would not have made sense for Konon and Pharnabazos to have a significant naval base at Loryma when the rest of the Chersonese was in enemy hands. Thus, there should be no doubt that at that date (and probably since 396) the Chersonese was in Konon’s control. Finally, it should be added that no source, literary or epigraphic, has come to light to confirm the hypothesis of a Physkos in the Chersonese region.
26If such is the case, could this Physkos have been the name of the long and narrow bay that gave access to the isthmus of Bencik, to the east of the Knidia? As is developed below, physkos can mean ‘bowel’ or ‘intestine’, which might have been a correct description of the topography of this bay. But no ancient source gives a hint in favour of this hypothesis. Moreover, for Peisandros to go to Bencik would have been absurd. This would have taken him to a zone which had no strategic importance, and would have left the gate to the Aegean wide open71. Besides, concentrating his whole fleet in a zone that was in immediate contact with the enemy (who certainly held the Chersonese, as mentioned before) and where the fleet could be easily trapped, would not have made any more sense.
27We are apparently left with no easy explanation and it seems that no reasonable solution can emerge. Should we consider with C. D. Hamilton that ‘the details of the battle are too obscure to permit any accurate reconstruction’72? This is what this paper would like to challenge. Leaving aside the debate about Diodorus’ report, the starting-point of any reconstruction should be that all ancient authors admit that the battle was fought in the immediate neighbourhood of Knidos. Moreover, by differing from Judeich, Pareti and Barbieri (who published his book only three years after Bean and Cook presented their hypothesis of the relocation of the town of Knidos and did not know of their hypothesis), we have now a different view of the Knidian territory and of the evolution of its occupation. This opens the way to a new hypothesis. We should consider that Peisandros left Knidos – Datça, not Tekir – for another point of the Knidia. Peisandros’main target was to prevent Konon reaching the Aegean: thus keeping watch over Triopion (Tekir) and its ports was absolutely necessary. It so happens that Pausanias gives a valuable piece of information that may help to reconstruct the events, one which was of no value for scholars before Bean and Cook, as they assumed that the Knidian capital was then at Tekir73. Among the offerings he describes at Olympia, Pausanias believed that one had been dedicated by one part of the town of Knidos-Tekir: the inhabitants of the island facing the continent at Cape Tekir. This island is also described by Strabo, who gives valuable information on it: it rose high, had a theatre-like shape, protected the two harbours and was the place where a good part of the local population lived74. The only difference between Strabo’s and Pausanias’description relies on the fact that for the former the island and the mainland were linked by moles (chômata), while for the latter it was by a bridge (gephyra). Pausanias missed the point of the dedication at Olympia: the Chersonesians who made it must have been the ‘koinon of the Chersonesians’, still an independent city in the fifth century, but later absorbed by Rhodes75.
28Yet, in passing and quite unwittingly, Pausanias gives us a valuable piece of information. The island which was part of the town of Knidos of his time was called Chersonesos. To the present author’s best knowledge, one single modern author has commented on this designation. While not proposing a meaningful solution for the dedication of the ‘Chersonesians’ (he considered that Pausanias meant that this dedication had been made ‘by the people of the Knidian continental territory’), A. von Gerkan rightly noticed that Pausanias’version of the dedication of the Chersonesians had no authority76. However, he went so far as to consider that Pausanias’designation of the island by Knidos-Tekir was doubtful, for if in Greek the word ‘island’ (nèsos) could commonly apply to a peninsula, it would have been bizarre to call a ‘peninsula’ a true island. But this argument is not valid. The people of Triopion, then of Knidos, can perfectly well have styled the island Chersonesos, insofar as it was linked to the mainland by a jetty (even originally only a wooden one) or mole or any other similar construction. Pausanias gives this information on Knidos so to speak en passant: finding the bizarre mention of a dedication made by these ἐν Κνίδῳ Chersonesians, he tried to explain it using his personal knowledge of the topography of Knidos-Tekir. Thus it cannot be supposed that he invented the name for the sake of his own explanation. Besides, Pausanias makes two other valuable mentions of Knidos: the first when he alludes to the battle of Knidos and to the fact that it was fought at the foot of Mount Dorion (he is the only author who gives this information)77; the second when he makes the valuable comparison between the topography of Megalopolis on the one hand and Mytilene and Knidos on the other hand, towns that are all divided by a river (the former) or by straits (the latter)78. Pausanias himself originated from Asia Minor and it seems likely that he had a personal knowledge of the site of Knidos, which might explain his topographical precision79. There is thus no reason to throw doubt on his naming of the island.
29Thus, we now have a Chersonesos, and from this single piece of information the movement of Peisandros’ fleet becomes meaningful. Provisionally, the problem of ‘Physkos’ must be left aside. If one considers that Diodorus’ source mentioned that Peisandros made for the ports by the island of Chersonesos, then his fleet left Knidos-Datça in a westerly direction, to reach Triopion (at Cape Tekir). One can imagine he must have left there a few ships to keep watch on this fundamental strategic point, just as his colleague Hippokrates had done a few years before, in 412/1 BC. The island offers an excellent screen from the winds coming from the high sea80. If (according to Strabo) the northern port was large enough to receive twenty triremes, together with the larger southern port, there is no doubt that Peisandros’ fleet of eighty-five ships could find an anchorage there. (As noted above c. 469/6 BC, for the campaign of the Eurymedon, Kimon’s fleet had ‘set sail from Knidos and Triopion’ with a fleet of two hundred ships.) Peisandros’ movement was perfectly reasonable as his mission was precisely to prevent the huge fleet of Konon and Pharnabazos reaching the Aegean and irremediably damaging the Spartan positions in the region.
30But what of Physkos? According to the dictionaries, ἡ φύσκη designated the ‘large intestine’, especially when stuffed with pudding, and was thus equivalent to ‘sausage’, ‘blackpudding’. The diminutive was φύσκιον or the parallel forms φυσκία or φύσκος (also ‘botellus’). The nickname φύσκων or φυσκών, ‘pot-belly’, was given to Pittakos by Alcaeus and by the people of Alexandria to Ptolemy VIII Euergetes II for his exceptional obesity81. Physkos was, however, also a traditional Greek personal name, one of those which conveyed a clear meaning, but not necessarily a disparaging one, although it might seem so at first glance. It is historically well attested82. It had been borne by the Lokrian hero Physkos, the founder of the homonymous town of Physkos in Lokris83. In the case of the Lokrian town, the origin of the name is thus perfectly clear. But this does not exclude the possibility that other toponyms could be based on the banal meaning of physkos.
31The detailed topography of the town of Triopion – later Knidos – must now be examined. Describing the site of the city of Megalopolis, in the Peloponnese, Pausanias mentions that the river (potamos) Helisson divided it just as the towns of Knidos and Mytilene were divided by their straits (euripoi)84. Indeed, the topography of Mytilene and that of Knidos were very similar. In both of them, a small island a few yards from the mainland determined the existence of two ports. Interestingly, Strabo’s description of the town of Mytilene echoes that which he gives for Knidos85. A similar situation, with a small island off the coast determining the presence of two ports, was to be found at Syracuse, with the island of Ortygia, and at Halikarnassos, where the Hekatomnid fortress and palace had been situated on the island, where is now located the Hospitaller castle of St. Peter. In his description of Halikarnassos, Pseudo-Scylax mentions that the town possessed two ports, a closed one (which presumably must be the main port corresponding to the present port of Bodrum), and another ‘by the island and ‘river’ ’86. This ‘river’ would seemingly designate the channel between the island and the mainland. As for the word potamos used by Scylax – if indeed it was, which is not certain – it normally designated a river or any other natural waterway, not a canal or other man-made construction87. However, instead of ποταμός, K. Jeppesen has proposed to read ὁ τάφος in Scylax’ text and to see there an allusion to the famous tomb of Maussolos, ὁ τοῦ Μαυσώλου τάφος, an emendation which is far from certain88. In any case, the existence of a waterway between the fortress-island and the mainland should not be doubted. The existence of two ports at Halikarnassos is also mentioned by Vitruvius89.
32From Xenophon in his Anabasis and his companion Sophainetos of Stymphalos in his Anabasis of Cyrus, we know that a river (potamos) of eastern Babylonia, near Opis – with a width of one plethrum (c. thirty metres) and ‘which had a bridge over it’ – bore the name of Physkos90. It has thus been suggested that the name of this river was derived from the Aramaic root psḫ – (‘passage’, ‘ford’)91. This root can also be found in the name of the town of Thapsakos of the Greeks, the Tiphsah of the Bible, a famous crossing-point on the Euphrates. This is where Kyros’army forded the Euphrates in 401 BC92. But if indeed the derivation from an Aramaic root is correct, the Greek Physkos conveys no idea of passage. The construction is purely phonetic and it remains interesting that the word physkos, ‘bowel’ or ‘sausage’, was chosen: from the vague phonetic assonance with a local word could be shaped a name that could make sense in Greek.
33As far as the the topography of Physkos-Marmaris is concerned, it seems likely that the rounded shape of the bay and its small entrance might have suggested the designation of the place as Physkos, as a nom parlant93. The fact that three different places bore the same name in the same wider region might also not be mere chance, but this is only a guess. At Triopion, at least, the island bore also the (frequent) nom parlant of Chersonesos. Thus it is legitimate to suggest that at Triopion, the site of Hellenistic and Roman Knidos, the channel that separated the island of Chersonesos, or better the site where the two ports of Triopion-Knidos-Tekir were installed, was locally styled Physkos. This is what the Sicilian Diodorus (who unlike Pausanias who was from Asia Minor, probably had not seen Knidos in person) might have found in his sources and completely misunderstood. From the description he read, he may have created a ‘Physkos of the Chersonese’ because he knew of the existence of a port of that name in the same region.
34The reality was that Peisandros moved from Knidos-Datça to the port at the site of Triopion, east of the island of Chersonesos, that locally bore the name of Physkos. He probably did so in order to get the reinforcement of ships that were based there, in the two excellent ports benefiting from the shelter of the island of Chersonesos. Judeich identified the Mount Dorion mentioned by Pausanias with the Boz Daǧ, a few miles east of Tekir94. If this identification is correct (and it is quite an attractive one as the ‘Mount of the Dorians’ must have been located just above the site of the common sanctuary of the Dorians), then we must understand that the main phase of the battle of Knidos was fought in the waters at the foot of Mount Dorion95. Peisandros’ships may have been pushed back to the coast either in the sector of Deveini or in that of Kumyer.
35Admittedly, this reconstruction will have to remain a hypothesis until new evidence confirms it. However, for the present, its advantage for the topography of the battle of Knidos is twofold. Firstly, it fits perfectly with what all ancient sources suggest was the scene of the battle (in the sea off Cape Tekir). Secondly, it can be supported by the hint offered by Pausanias’ mention of Chersonesos as the name of the island at Cape Tekir. We should thus admit that our late Hellenistic or Roman sources made several errors on the historical geography of Knidos. On the present hypothesis, one should consider that Diodorus could not correctly render Peisandros’ movement from Knidos-Datça to Triopion. But this is paralleled by Pausanias’ confusion of the koinon of the Chersonesians with the people inhabiting the island of Chersonesos at Cape Triopion. The common origin of this misunderstanding should be explained by the difference between the local political geography of the Classical period and the later political organization of the Hellenistic or Imperial period. The city of the Chersonesians disappeared as early as the fourth century BC. The transfer of the capital of Knidos from Datça to Tekir left no trace in general history, while the territory of the city of Knidos was left unchanged, for there was no creation ex nihilo of a new town, but only a rearrangement. This clearly contrasts with Rhodes, where in 408/7 BC out of the three poleis of the island a new, unitary city emerged that had no previous existence, with a homonymous capital founded on a new site. For those who had no knowledge of local Knidian developments, many details found in previous authors became almost incomprehensible. General historians of the late Hellenistic or Imperial period, like Diodorus, Pausanias or Plutarch, could thus make inadvertant mistakes in their own narratives. This was the case, for instance, when Pausanias erroneously attributed the destruction of Haliartos of Boeotia to the ‘Persians’, probably because he had read that the town had been destroyed in the war against the Macedonian king Perseus96. Plutarch had invented the story of an ‘Athenian’ at Sparta, when he should have understood from Xenophon that the man was a Spartan bearing the name Athenaios97. Diodorus notoriously mangled the chronology of the early 390s and everyone agrees that his report of the battle of Knidos is of poor quality. It is sometimes the task of modern historians to separate the wheat from the chaff and try to reconstruct a plausible sequence of events from the scraps of evidence ancient authors nevertheless provide us with. This is the case with the problem of the topography of the battle of Knidos of 394 BC. One can only hope that some new discovery will give further information on the topography of ancient Knidos, helping to confirm the hypothesis presented here.
Annexe
Appendix: Physkos as port of Mylasa?
In his notice about Mylasa, Strabo mentions in passing that ‘the city comes nearest to the sea at Physkos, which is their harbour (epineion)’98. However, a notice of Stephanus of Byzantium also alludes to Passala as the harbour (epineion) of the Mylasans99. Besides, two inscriptions found at Mylasa bearing imperial letters of the pre-Byzantine period (c. 427/9 AD) conferred tax privileges to the Mylasans for the harbour tax levied in Passala, defined as a ‘kome of the Mylasans’100. Passala must probably be identified with the ancient site located at Sakız, not far from the mouth of the present Sarı Çay, where inscriptions and a Hellenistic tower have been found101. Thus if Strabo actually thought of Physkos-Marmaris, he would have simply confused one name for another102. It has also been proposed that Physkos was simply another name of Passala103.
Should we really believe that Strabo mistook Physkos for Passala? Indeed, this is quite possible. But the chronology of the documents should also be taken into account in the reflection. The only formally-dated documents alluding to Passala are the two inscriptions of the early fifth century AD. At that date, the topography of the gulf must have been different from what it is today, but also from what it had been in the Archaic, Classical and Hellenistic periods. Indeed, a considerable process of silting has occurred in this area, as in other portions of the Turkish littorals (the most famous case being the modification of the coast in the lower Meander where, among others, the present Lake Bafa was, in fact, a maritime gulf in Antiquity). West of Mylasa, over the centuries, the silting of the Sarı Çay has almost completely filled up a former maritime gulf (see figs. 1 and 2 and map p. 47 in G. Reger’s paper, this vol.). It is obviously in that gulf that by the time of Alexander the Iasians had obtained from the king a fishing privilege in the mikra thalassa, the ‘little sea’ 104. Today, the surface left to the sea is residual and what is now a zone of plain and marshes is for a large part occupied by the recently-built Milas airport. But between Archaic and Classical times and the Imperial period (and even the late Imperial period) the process of silting in the present Sarı Çay plain must also have been significant. In other words, even if the site of Sakız-Passala was already occupied at an early date, nothing proves that it was from the start the epineion of the Mylasans. If the shape of the mikra thalassa was significantly different in the early history of Mylasa and penetrated far inland, another site, nearer to Mylasa and more convenient might well have been preferred. This might justify the mention of an epineion of Physkos by Strabo. This would, of course, have been a Physkos different from that of Physkos-Marmaris. After a process of continuous silting, the site would later have been abandoned. As for the origin of the name, however, it would have alluded to the bowel-shape of the bay, as at Physkos-Marmaris and at the port of that name at Tekir whose existence has been hypothesized above.
Which should one suppose: a mistake on the part of Strabo, or the existence of another Physkos? In this case also, further discovery may help to solve the riddle.
Notes de bas de page
1 Xen. Hell. 4.3.10. On this solar eclipse, von Oppolzer 1887, 80 no. 1959 (see Stephenson 1997, 366, and ibid., 30, on the conversion between Julian and Gregorian calendars)..
2 Xen. Hell. 4.3.11-2.
3 FGrHist 328 F 144-5.
4 Lys. 19.28; Isoc. 9.68; Din. 1.75. See also below for other, later, authors.
5 Diod. 14.83.4-7.
6 Judeich 1892, 74-5; Barbieri 1955, 144-60, esp. 145-7.
7 Polyain. Strat. 1.48.5.
8 Corn. Nepos 9.4.4; Justin 6.3.1-12; and Orosius 3.1.12-5.
9 For a presentation of the history of the excavations and of the present state of the discoveries, see Bruns-Özgan 2002, and Berges 2006, 19-34.
10 Bean&Cook 1952.
11 On which see Hansen & Nielsen 2004, 1123.
12 See the historiographical references to the debate in Bresson 1999b, 85.
13 Bresson 1999b for a detailed historiography of the debate that there is no need to repeat here, and Berges 2006, 30-4, with N. Tuna’s annual excavation reports in the AST.
14 Plut. Cim. 12.2.
15 Thc. 8.35.1-4.
16 See also the comment of Gomme et al. 1945-1981, V, 78, and see now Hornblower 1991-2008, III, 847-51.
17 Ps-Scyl. 99.
18 Hdt. 1.174.2; Thc. loc. cit. In a coded allusion, Lycophron (Alex. 1388-96, esp. 1391) also situates Triopion at the tip of the Chersonese. Schwyzer’s correction (1926, 448) to Ληκητρίαι (the name of a cape on the island of Kos) has been adopted by Fusillo et al. (1991, 55) and Hurst (2008, 306), but neither by Mascialino in the Teubner ed. of 1964 (p. 62, relegating it to the app. crit.) nor by Lambin 2005 (99, who ignores it).
19 Berges 1994. For other alternative solutions that have been proposed, see Bresson 1999b, 85.
20 Berges 2006.
21 J. Nollé in: Berges 2006, 60-2, nos. 6-7.
22 Berges 2006, 23.
23 Berges 2006, 22-3.
24 Bresson 1999b.
25 Berges 2000, 164-5 with Fig. 3 (cf. also Berges 2006, fig. 5).
26 Respectively Blümel, I. Knidos 603 (a decree for the tyrant Iphiades of Abydos, around 360 BC = SEG 44, 901), 604, and Blümel 1994 (a decree for Epaminondas, of 364 BC). Treaty with Chalke: ibid., 605.
27 See Bresson 1999b, 85-6 and 99-100, for a detailed argument.
28 Bankel 2004, 109, n. 33, who however missed the present author’s 1999 article and has not been aware of the issues linked to the epigraphic dossier; for the issue of Knidos αἰπεινή (Hom. HAp. 43), see ibid. 90-1. The argument that the inscriptions from Emecik (on which see now Nollé, in Berges 2006, 60-2) were also transported from Tekir is even less convincing.
29 Steph. Byz. s.v.
30 Arr. Anab. 2.5.7. See Bresson 1999b, 89 and 102.
31 Stampolidis 1984; Bankel 2004, 111.
32 Bresson 1999b, 98, with pls. p. 96-7.
33 See Bresson 2011.
34 Diod. 14.39.1-2 (cf. Justin 6.1.4-9).
35 Diod. 14.39.4 and 79.4-5.
36 Thc. 1.116.1-3.
37 Thc. 8.39-42.
38 Thc. 8.57.1.
39 Thc. 8.88 and 108.1.
40 Diod. 14.79.4-5.
41 Paus. 6.7.6.
42 Diod. 14.79.6.
43 Diod. 14.79.6.
44 Diod. 14.79.7-8.
45 Hell. Oxy. 18. On these events, Debord 1999, 243-4, with 251-2.
46 Diod. 14.82.1 (beginning of the archonship of Diophantos), 14.85.1 (beginning of the archonship of Euboulides). Cf. Lys. 19.28, and Stylianou 1998, 35, on the reasons for this mistake.
47 See above and n. 46.
48 Hell. Oxy. 22-3.
49 Isoc. 4.142. See also Isoc. 9.64, with Judeich 1892, 54 n. 1.
50 Xen. Hell. 4.3.11-2.
51 Diod. 14.83.4-5.
52 See below.
53 Pédech 1989, 40-64, espec. 57 (but no fragment of Theopompus in our possession alludes to the battle).
54 On Ephorus, see briefly Marincola 2007b, 172-4, with references.
55 For the Oxyrhynchus historian as source of Ephorus, Marincola 2007b, 173. As for his identity, the Cratippus option (rather than the Theopompus one or that of a still anonymous historian) is strongly advocated, with good arguments, by M. Chambers’introduction to the Teubner ed. of Hell. Oxy. (1993), xxii-xxv. See also briefly Tuplin 2007, 162-3.
56 Polyb. 12.25f. 2-3.
57 Barbieri 1955, 147.
58 Diod. 14.83.4 (Loryma), 5.62.4 (Bybassos).
59 Foucart 1909, 166, followed by Swoboda 1922, 1327, but this had already been criticized by Pareti 1911.
60 Pareti 1911.
61 Buckler 2003, 73, but (admittedly in a general book) with no mention of the scholarly debate and only two modern references, respectively to Beloch 1922, 76, and Hamilton 1979, 228-9 (but see below on these authors).
62 Plut. Ages. 10.6.
63 Buckler 2003, ibid.
64 Isoc. 9.68; Din. 1.75; Polyb. 12.25f. 2-3 (explicitly following Ephorus); Plut. Ages. 17.2; Paus. 6.3.16; Corn. Nepos 9.4.4.
65 Judeich 1892, 74-5 n. 1; Barbieri 1955, 144-8.
66 Judeich 1892, ibid.; Fraser & Bean 1954, 84-7.
67 In the absence of sources, this solution was qualified by Pareti 1911, 115 n. 1, as a ‘leap into the void’.
68 Str. 14.2.23. On the issue of the port of Mylasa, see the Appendix below.
69 Diod. 14.79.6. Diodorus does not say explicitly that Konon took over the Chersonese in 396, but this seems a reasonable inference, as this allowed him to have contact with the Rhodian leaders who soon abandonned the Spartan side and opened their ports to his fleet.
70 On Loryma, see Held 1997, 2000, 2001, 2002, 2003b, 2005, 2006a and b, 2009 and D. Blackman in this volume. Was the Chersonese Rhodian already before 396? And, if not, at what date did it come into Rhodian hands? Held 2009 favours an incorporation c. 300 BC. The issue should be rediscussed in detail.
71 The region had a special strategic importance and it is no chance if another (admittedly less important) battle was fought a few years before (411 BC) and in the same region, the so-called battle of Syme (Thc. 8.42.1-5; see Chêne 2002).
72 Hamilton 1979, 228. Same view: Bleckmann 2006, 42-4; cf. also 36-54 on the battle.
73 Paus. 5.24.7.
74 Str. 14.2.15.
75 Bresson 1999b, 104-14.
76 Von Gerkan 1924, 117 n. 1.
77 Paus. 6.3.16.
78 Paus. 8.30.2.
79 For the origin of Pausanias (probably Magnesia-by-Sipylos), Habicht 1985, 13-15.
80 For a more detailed treatment of navigation issues around cape Triopion, see Bresson 2011.
81 See respectively D. L. 1.81 and Str. 17.1.11 (with Nadig 2007, 66-72).
82 See occurences in LGPN, II (Athens), 1; III. A (Western Greece), 4; III. A (Central Greece), 3.
83 On the hero Physkos and Physkos in Lokris, see Oldfather 1941.
84 Paus. 8.30.2.
85 Cf. Str. 13.2.2 (Mytilene) and 14.2.15 (Knidos).
86 Ps-Scyl. 99.
87 Chatelain 2001, 88-9, and app. 2, p. 106.
88 Jeppesen 1986, 87-91, with the doubts of Hornblower 1988.
89 Vitruv. 2.8.14.
90 Xen. Anab. 2.4.25; Sophainetus, quoted by Steph. Byz., s.v. Physkos.
91 Barnett 1963, esp. 3 (n. 8), 18 and 25.
92 Xen. Anab. 1.4.11-18.
93 For the modern toponym Physkos on Rhodes see this vol., Held, p. 358, n. 11.
94 Paus. 6.3.16.
95 Pareti 1911, 113, suggested correcting ὄρος τὸ Δώριον into ὄρος Τριόπιον, but this is hardly justified, as no source makes mention of Τριόπιον as an ὄρος.
96 Paus. 9.32.5 and 10.35.2. See Habicht 1985, 99, with other references in Bresson 1999b, 106 n. 90.
97 Plut. Ages. 13.3; Xen. Hell. 4.1.39. As they accepted Plutarch’s authority, modern authors have also missed the meaning of Xenophon’s text, see Bresson 2002.
98 Str. 14.2.23.
99 Steph. Byz. s.v. Passala.
100 I. Mylasa 611-2.
101 The whole ancient information on Passala is discussed in detail by Blümel, I. Mylasa I p. 117-19 and need not be reproduced here.
102 This is for instance the view of Fraser & Bean (1954, 57) and Blümel, I. Mylasa I p. 17 and 117.
103 Smith 1857, 2.626.
104 Inscription of Iasos mentioning the mikra thalassa: I. Iasos 30; identification of the mikra thalassa with the bay occupying the present Sarı Çay plain: Blümel, I. Iasos I p. 46; see also S. Mitchell 1998-1999, 157, and the papers by K. Konuk and G. Reger in this volume.
Auteur
University of Chicago
Le texte seul est utilisable sous licence Licence OpenEdition Books. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
La transmission de l’idéologie impériale dans l’Occident romain
Milagros Navarro Caballero et Jean-Michel Roddaz (dir.)
2006
La circulation de l’information dans les états antiques
Laurent Capdetrey et Jocelyne Nelis-Clément (dir.)
2006
Villes et territoires dans le bassin du Douro à l’Époque romaine
Milagros Navarro Caballero, Juan José Palao Vicente et Maria Ángeles Magallón Botaya (dir.)
2007
Le monde de l’itinérance
En Méditerranée de l’Antiquité à l’époque moderne
Claudia Moatti, Wolfgang Kaiser et Christophe Pébarthe (dir.)
2009
Les seigneuries dans l’espace Plantagenêt
(c. 1150-c. 1250)
Martin Aurell et Frédéric Boutoulle (dir.)
2009
Studies in Greek epigraphy and history in honor of Stefen V. Tracy
Gary Reger, Francis X. Ryan et Timothy Francis Winters (dir.)
2010
Ombres de Thucydide
La réception de l’historien depuis l’Antiquité jusqu’au début du xxe siècle
Valérie Fromentin, Sophie Gotteland et Pascal Payen (dir.)
2010