Kynosura is the promontory that closes the bay of Marathon and Keos is an island at the very tip of Attika off Cape Sunion. In the eighteenth century the distinguished Hellenist Pierre-Henri Larcher understood that the Keos and Kynosura mentioned by Herodotos were the places generally known by these names; but in 1829 William Martin Leake asserted, without any textual evidence, that the name of Kynosura applied to a promontory of the island of Salamis and that Keos was some island thereabout. (82)
Later Stein added the further gratuitous contention that Keos was the common name of Kynosura. When J. W. Blakesley objected to Leake that the text of Herodotos itself indicates that Keos and Kynosura must be the places known by these names, (83) he was treated with ridicule. Only Grote was considerate enough to admit that no other meaning can be given to Herodotos' account, but subjoined that Herodotos must have been totally mistaken. (84)
The theory of Leake continued to be accepted as dogma until 1935, when Henri Gregoire (who was a famous specialist of Byzantine culture and as such could read Greek, but was not used to begin with the assumption that the authors he read were fools, impostors, or madmen) having noticed how Herodotos' text is intepreted by critical historians, expressed his surprise and indignation in an article entitled "La Legende de Salamine, ou comment les philologues ecrivent l'histoire." (85)
He observed that the text indicates that only a part of the Persian fleet fought at Salamis, while the rest formed a line extending from Munichia to the coast of the Peloponnese. In 1952, followers of Gregoire, G. Smets and A. Dorsinfangs-Smets, reviewed carefully not only the evidence provided by Herodotos, but also that provided by other sources, and conluded: "the very evidence itself indicates that the entire Persian fleet was not at Salamis and that only its western wing was engaged in the battle." (86)
When confronted with the arguments of Gregoire, some scholars decided to ignore them and some decided to rehash some old discarded contentions. The first choice is that of the commentary on the The Persians by E. D. Broadhead. (87)
The other course was chosen by Legrand, who as a specialist of Herodotean studies tried to reply to Gregoire by repeating the argument of Grote: what Herodotos relates cannot be taken seriously because it is full of "incoherences"; "Herodotos has gathered together anecdotes that are more or less tendentious and which he picked here and there, from the right and from the left." (88)
But Legrand grants by implication that if Herodotos' statements about the second section of the Persian fleet are not taken at face value the entire narrative becomes preposterous and the entire Persian strategy becomes erratic. His position is that it is better to classify the events of Salamis as an "enigma" than to assume the impossible, namely that Herodotos provided a reasonable account. That a ferry operation had been started is indicated by the statement of Herodotos that the second division of the Persian fleet moved "to hold the entire ferry line up to Munichia" (VIII 76). This passage is usually disregarded, but Macan who, though doggedly partisan in his interpretation of the evidence, did not ignore it, observed: "It is curious that the roadstead up to Munichia should be described as a porjmos, a term properly used of a ferry, a strait, or narrow waterway." Of course it is not Herodotos who gives a "curious" meaning to Greek terminology. In the tragedy Agamemnon by Aischylos, the body of water between the peninsula of Attika and the Peloponnese, across which a signal is sent by fires, is called porjmos (line 306). The old commentary of J.C.F. Boehr does less violence to Greek usage when it tries to explain the word porjmos of Herodotos by assuming that in peacetime there used to be a ferry service between Munichia and the island of Salamis.
In 1953, Myers replied to Gregoire by granting that there was a second section of the Persian fleet which was stationed at Keos and Kynosura, but tried to discount the importance of this second section; it would have been composed of triremes that had arrived late or had remained behind because of the need for repairs. (89)
A similar argument had been used by Munro in 1926 when he claimed that the second section was composed of the 100 triremes of the Pontic Greeks which according to Herodotos were absent at the time of the muster of Doriskos. This contingent would not have caught up with the rest of the fleet in about four months. (90)
Myers tries to explain Herodotos' statements about the position of the second section by some sort of optical illusion that he describes in these cryptic words: "With the Phenicians on the western wing now on converging courses the gulf (porthmos) outside the straits [of Salamis] seemed indeed 'filled with ships.'" (91)
In truth, the porjmos (outside the sound of Salamis where the first section of the Persian fleet was stationed), that went from Munichia along the island of Aigina to the Peloponnese, was filled with ships. Aristeides (VIII 81) who arrived at Salamis on a small boat to report on the Persian movements, related that he had great difficulty in crossing over from Aigina because he had to slip through the blockading enemy fleet. The enemy of which he speaks was somewhere between Aigina and Salamis. Aigina is close to the coast of the Peloponnese and relatively distant from Salamis and the coast of Attika. If the entire Persian fleet had been lined up between the harbors of Athens and Salamis, the line of communication between Aigina and the southern shore of the island of Salamis would have been unimpeded. Aischylos (line 368) speaks clearly when he mentions "the other (Persian) ships that were all around the island of Aiax," that is, Salamis.
According to Aischylos (The Persians) and to Herodotos, the Persian force that attacked at Salamis suffered disaster, but in spite of it the Greeks were expecting a new attack by the Persian fleet (Herodotos VIII 97) and were surprised when this fleet was withdrawn (VIII 107, 108). After they learned that the Persian fleet had withdrawn, the Greeks considered pursuing it and attacking the bridges at the Dardanelles. This means that the Persian fleet had still so many ships that it could be expected again to form the two bridges.
The fact that the Persian fleet remained all-powerful even after Salamis indicates that only a part of this fleet was engaged in this battle; but critical historians, by denying this, are forced to discount the Greek accounts of a great victory and reduce the outcome of the battle of Salamis to something close to a draw.
My interpretation of the Persian strategy at the time of the battle of Salamis is the only one which is in agreement with the summation of the events presented by Thukydides (I 73):
It was the battle of Salamis that prevented the Persians from attacking the Peloponnese by sea in order to destroy the cities one by one; given the number of the Persian ships, the cities could not have found a way to organize a common resistance. The best proof of it is provided by the conduct of the Persians themselves: once they suffered a naval defeat, they realized that their forces were no longer adequate to their plan and, hence, they promptly withdrew the greater part of the army.
The authoritative commentary of W. W. Gomme, in following the current interpretation of the events, totally distorts the clear meaning of the words of Thukydides, asserting that he spoke of "the danger that if the Greek fleet retreated from Salamis it would disperse each to its own home."
From Aischylos (The Persians, 366) it may be inferred that at the battle of Salamis the Persian fleet consisted of three squadrons. Herodotos (VIII 85) mentions the Phoenicians as forming the left wing and the Ionian Greeks as being on the right wing. According to Herodotos, the Ionians with the Dorians of Asia Minor and the Karians contributed a squadron of 200 triremes to the original strength of the fleet, whereas the Phoenicians had contributed 300 triremes. Possibly the Persian formation at Salamis consisted of a squadron of 200 Phoenician triremes on the left and of an Ionian-Dorian-Karian squadron of 200 triremes on the right, with a mixed squadron of 100 Phoenician triremes at the center, plus 100 more contributed by sundry Greek allies. It may be concluded that the Persian force consisted of 600 triremes. Munro recognized that the Persian fleet consisted of 3 squadrons, but irresponsibly placed the Phoenicians at the center, the Ionians at the right, and the Egyptians at the left. It is certain that the Egyptian squadron of 200 triremes did not participate in the battle.
In conclusion, it is difficult to doubt that the Persian fleet as a whole had a force of 1200 triremes. In 1956 Hammond, although he tried to refute by some obscure argument the contention that on the eve of the battle of Salamis part of the Persian fleet was stationed at Keos and Kynosura, granted that it is no longer possible to ascribe a smaller figure to the Persian fleet. But a fleet of 1200 triremes required about 240,000 men as crews, without counting the embarked marines, so that it is reasonable to presume that the Persian naval units, including transports, required the service of about half a million men, as indicated by Herodotos. But if Persia, which was not a naval power, mobilized such a naval force, the land forces must have been much larger. Hammond tries to avoid the issue by mentioning only a comprehensive figure: "the total of combatants and non-combatants in the army and navy was probably in the range of 500,000 men." (92)
Thereby he continues the practice of treating quantitative data in a flippant manner.
After the defeat at Salamis the ferry operation was no longer possible, so that the grand plan to clean up Greece from top to bottom had to be abandoned. Because the season of equinoctial storms, dreaded by all ancient navigators of Greek waters, was at hand, the fleet was withdrawn in a great hurry to the waters of Asia Minor. In the month of October the Persian army withdrew from Attika to Thessalia. There a decision was reached to split the army into two parts. The King withdrew to Asia Minor, leaving his general Mardonios in Greece. The King took with him half of the army plus a contingent under the command of Artabazos drawn from the other half.
Since the winter was approaching the retreat had to be accomplished in forty-five days from Thessalia to the Dardanelles. At this speed the army could not be followed by supply trains, so that the retreat turned into a disaster because of famine, plague, and dysentery. The development of the plague and dysentery must be explained by the use of polluted sources of water by the undernourished troops. The army was so large that it could not rely on the ordinary sources of water, whereas during the advance provisions had been made for an orderly supply of water. Aischylos stresses the lack of food and water. According to Herodotos, many soldiers died upon arriving at the Dardanelles, where there were abundant supplies of food and water, because of overeating, "combined with the change of water" (VIII 117). Perhaps from the Persian point of view the fact that the King was able to return quickly to Asia Minor with a part of his army was a positive achievement, since it squelched the danger of revolts within the Empire.
Among the recent writers Richmond Lattimore takes a rather moderate position by stating that "This terrible retreat has been exaggerated by Aischylos and Herodotos alike, though want of supplies may have created serious difficulties and distress." This much can be granted, but I would not accept Lattimore's contention that whereas Aischylos may be believed as an eyewitness to the battle of Salamis, what he said about the retreat of Xerxes may be false. One can argue that both Herodotos and Aischylos because of national pride ascribed to the enemy army a size that had no relation to reality, but they would hardly have invented a version of the events by which the Persian army fell under its own weight. The positive result of the King's retreat with the army was that he was forced to realize that he could not keep more than 300,000 landfighters in Greece, a force that the Greeks could hope to match once they were able to gather together 100,000 of their own soldiers.
"Die Perserkriege," in Vortraege ueber alte Geschichte, Vol. I, p. 407f.
On the Demi of Attica (London, 1829), pp. 144-146.
Herodotus (London, 1854), vol. II, pp. 400-419.
George Grote, A History of Greece (London, 1862), Vol. III, pp. 470-471, n. 2.
Les Etudes Classiques, IV (1935), pp. 519-531.
G. Smets and A. Dorsingfang-Smets, "La Bataille de Salamine: Les Sources," Annuaire de l'Institut de Philologie det d'Histoire Orientale et Slave, (Brussels) 12 (1952), p. 426.
Aeschylus, Persae, ed. by E. D. Broadhead (Cambridge, 1960).
Ph. E. Legrand, "A Propos de l'enigme de Salamine," Revue des Etudes Anciennes, XXXVIII, (1936), pp. 55-60.
John L. Myres, Herodotus, Father of History (Oxford, 1953), p. 274. See also Paul W. Wallace, "Psyttaleia and the Trophies of the Battle of Salamis," American Journal of Archaeology, 73 (1969), p. 300.
J. A. R. Munro, "The Deliverance of Greece" in The Cambridge Ancient History, Vol. IV (1926), p. 305. Munro accepted Keos and Kynosura as the localities known by these names today; he postulated that the second fleet was waiting at a distance and was summoned by Xerxes when the battle began.
Myres, op. cit., p. 275.
H. G. L. Hammond, A History of Greece (Oxford, 1959).