At Thermopylae in the late summer of 480 the Spartan king Leonidas held out for three days with a mere 300 hoplites against thousands upon thousands of the best of the Great King's troops. It has also been the site of several battles in antiquity besides this most famous one. In 279 BC the Greeks faced Brennus and his Gauls there (Paus. 10. 20-23, Justin 24. 4-8); in 191 the Romans under M. Acilius Glabrio (and teamed with Philip V) defeated Antiochus of Syria and the Aeotolians (Livy 36. 17-19, Plut. M. Cato 13); and in 1941 the New Zealanders fought a rearguard action there against the Germans, in the course of the war which interrupted the excavation begun under the direction of Sp. Marinatos in 1939.

Clearly Thermopylae was a location of great strategic importance, because it commands the pass through which one goes after traveling south from Thessaly through Lokris and into Boeotia. Holding the pass could block an invader and even turn him back, though on all three of the famous occasions the defense of the pass failed. The Athenians took up a position there in 352 and discouraged Philip II from invading. In 323 during the Lamian War, the last-ditch effort by Athens to break free from Macedonian control, the general Leosthenes blocked the Macedonian Antipater by stationing troops at Thermopylae. However, the pass at Thermopylae was not the only way south from Thessaly into Central Greece; it was merely the best and easiest route.

In 480, in 279, and in 191 the invaders were able to get over the mountains and take the defenders in the rear. Examining the question of exactly what route was taken on each occasion, although admittedly a matter of primarily antiquarian interest, nonetheless illustrates some important trends in modern historical research. It also helps to answer the question of why Thermopylae should even be thought of as a pass. Herodotus' description of the location suggests that there are cliffs on one side and the sea on the other:

The pass through Trachis into Hellas is 50 feet wide at its narrowest point. It is not here, however, but elsewhere that the way is narrowest, namely, in front of Thermopylae and behind it; at Alpeni, which lies behind, it is only the breadth of a cart-way, and it is the same at the Phoenix stream, near the town of Anthele. To the west of Thermopylae rises a high mountain, inaccessible and precipitous, a spur of Oeta; to the east of the road there is nothing but marshes and sea. (Hdt. 7. 176)

But the modern visitor to the site sees two not very imposing looking hills; they lie to the south, not to the west. This discrepancy has led some scholars to assert that Herodotus never even saw the site, and that if he could make so basic an error all of his topographical information about the site, which is copious and detailed, must not be trusted; others tried to save his credibility by positing that he saw the site around noon, so that the sun was directly overhead and it was impossible to orient himself. W. Kendrick Pritchett, who is generally credited with injecting new life into the study of ancient topography, has mounted a vigorous defense of Herodotus's reliability on this and other sites. Pritchett points out that Herodotus seems to have done a very careful study of the site despite the error over the directions; he gives many distances in stades and plethra, and his account also includes an unusually high number of obscure toponyms.

More puzzling for the tourist who arrives at the site with his Herodotus in his hand is what lies to the south of the hills, beyond the modern roadside monument: a broad expanse of scrubby ground stretching out for about four miles to the sea. It looks today like no pass at all. The reason for this is no mystery. Due to what geologists call "alluvial fans", a process by which rivers deposit silt (travertine and other sediments), the coastline of the Gulf of Malea has advanced from 3-5 miles over the last 2500 years (Kraft et al., Journal of Field Archaeology 14 (1987) 181-197). Kraft and his team calculated the sea level in 480 using a mathematical formula known as the "eustatic curve". Together with the results of radiocarbon dating on the deposits and stratographic interpretation of the layers of the new land, they were able to account for the fact that travelers of only a few centuries ago reported the pass to be much narrower than we would expect if the process of buildup were proceeding at a steady rate. Rather, according to Kraft :

Fluctuations in the width of the pass at Thermopylae [have been] common, as expected in an unstable structural configuration along the flank of a major graben (i.e. a rapidly subsiding block of the earth's crust). (187)

Kraft concluded that the pass was not more than 20-30 meters wide in 480. That was too wide for Pritchett, who attacked the findings in volume VI of his Studies in Greek Topography (Herodotus says that the pass at Thermopylae was narrower than that at Alpeni, which he puts at half a plethron or roughly 15 meters wide).

The Battle

The confrontation at Thermopylae took place in the late summer of 480. Some modern accounts seem to know exactly on what dates the battle fell, because Herodotus says (7. 206) the the festival of Apollo Carneia was on at Sparta and that the Olympic games were also in progress. This confidence about the precise dating has lately been called into question (e.g. by Sacks in CQ 1976), but it is still possible to describe the battle in terms of relative chronology and that in many ways turns out to be more revealing. For example, we know that when Xerxes and the Persian imperial army arrived at Anthela, just west of the pass, they encamped and waited for five days before attacking. The reason for this is fairly straightforward. First, although the Persians could be confident that they would outnumber the enemy, they had as yet no idea how many hoplites were waiting on the other side of the pass, hidden by a hastily reconstructed wall. Second, Xerxes was waiting for his battered fleet to catch up; it had been damaged and delayed by bad weather yet again, the hand of the gods on the side of the Greeks (7. 188, the storm off the coast of Magnesia). A quick victory over the Greek fleet would allow him to simply land troops in the rear of the enemy, obviating the advantage offered to the Greeks by the terrain at the pass.

Xerxes used the time waiting for the fleet to arrive to good advantage. First he sent a spy to see what the Greeks were doing; the astonished horseman returned to report that he had seen the Spartans stripping for exercise and fixing each other's hair. It seems unlikely that this scene aroused the contempt in the Persian commanders Herodotus said it did, at least to judge from the next move, which was to send a herald to propose that the defenders of the pass should surrender and become allies of the Great King. In return they would be allowed to depart unharmed, and they could expect to get some of the land of those who refused to surrender. This tid-bit is reported by Diodorus (11.5, derived from Ephorus) but it is credible, since Xerxes had made similar pronouncements to the other Greek states before; Herodotus rather reports it as a conference held among the Greek contingents before Xerxes had arrived (7. 207). The offer will not have been expected to sway the Spartans; indeed, Xerxes had shown a disinclination to make further overtures to the Athenians and Spartans after the heralds of Darius had been executed both at Sparta and at Athens (Hdt. 7. 133). But if we can believe Ephorus the offer did expose the differing preoccupations of the various Greek contingents. The Peloponnesians, presumably including the Tegeans, Arcadians, Corinthians, and Phlians as well as some contingent of the Spartans, were for abandoning northern Greece and falling back on the Isthmus; only the insistence of Leonidas restrained them, and naturally the Phocians and the Locrians will have opposed this idea, since the non-combatants of Phokis and Lokris were for the most part still not evacuated. This debate among the Greek states typifies the distinctive feature of their foreign relations in the period, namely that each state tended to support its own regional interests, and it is worth reflecting on how this is usually portrayed in modern historical writing. The sense one gets is often that this was the curse of the Greeks; had they only been able to cooperate better, as they did for just long enough at Salamis, they could have ruled the world, or they would never have become the subjects of the Macedonians or (later) the Romans. Perhaps our postmodern penchant for "diversity" makes it easier for us to see how such sentiments are misguided: the cultural homogeneity which greater unity and cooperation among the Greeks would have inexorably brought about, would have brought with it, as it did in the much reviled Hellenistic Age, the sapping of their creative spirit which drew its energy from that very contentiousness which marked their interrelations.

In any event Leonidas was able to hold the Greek force together. He had only 7,100 troops; Herodotus says that Xerxes had 2.5 million troops and as many again of camp followers, but the figure is widely acknowledged to be fantastic. A more realistic estimate is had by lopping off a zero: perhaps 200,000, not all of whom had arrived at Thermopylae by the time Xerxes decided he had waited long enough.

At first, the battle went entirely according to the plan of the Greeks. The narrowness of the pass at the middle gate negated the advantage of numbers for the imperial troops. Moreover, the Greek hoplite was better equipped, with his long thrusting spear, heavy hoplite shield, and body armour; the Persian had a shorter javelin-type spear, a wicker shield which did indeed provide superior mobility in the open field but was much less useful than bronze at close quarters, and thick-woven linen corselets. For two days the Spartans held off lesser elements of the imperial army: Medes and Cissians were succeeded by the crack troops, the Immortals, to little avail.

Then the tide turned when a local man, a Malian named Ephialtes, offered to show the Persians a way around the back of the defending force, a way to get past the "Middle Gate" and turn the Greek position. Xerxes agreed, sending what was left of his 10,000 "Immortals" off at dusk. The precise route taken by the Persian troops that night is disputed. The standard view used to be that the path corresponded to the gorge of the Asopos river (so e.g. Leake, Grundy, Hignett), but this has two problems. First, the Asopos river gorge is too rocky to be negotiated at night without numerous broken ankles; second, Herodotus says that the path began from the Asopos river "which flows through the gorge" and not, as the standard view insists, "where it flows through the gorge" (7. 216). Two other main candidates have been put forward: the Vardates route (favored by Myres, Burn, and Wallace); and the Chalkomata spring route, favored by Pritchett. Whichever of these two it was may never be known for certain, but both would bring the Persians to the peak of Sastano (Kallidromos) near ancient Drakospilia by dawn. From there the paths converge.

Now, according to Herodotus Leonidas had been aware from the beginning of the existence of the Anopaia path. He stationed 1000 Phokians there to stop any encircling movement. The Phokians, according to Herodotus, were taken by surprise and put up little resistance. But word got through to Leonidas that the position had been outflanked, and there seems to have been time to abandon the position and withdraw to the south before the Immortals under Hydarnes arrived. Why did Leonidas refuse? There have been various answers to this question. Herodotus represents it as an act of deliberate self-sacrifice carried out in accordance with an oracle, which had said that the death of a Spartan king would save Sparta from destruction. One may observe that the pronouncements of the oracle in the late 480's have a distinctly pro-Persian cast; it seems likely that the priests, whose job after all was to predict the future, simply believed that the victory of the Persian army, whose immense size was known well in advance of its arrival, was inevitable. It may be that this oracle, if genuine, actually meant that the recommended course of action was for the Spartans to depose one of the sitting kings and take back Demaratus as the vassal of the Persians. Alternatively it is possible that the oracle is a post-eventum falsehood, put out by the oracle and its partisans to make it appear that Apollo had successfully predicted the outcome. There is also available the so-called "military" solution to the question, as formulated by Dascalakis. He argues that Leonidas remained in order to give the allied contingents, whom he dismissed (with the exception of the Thebans and the Thespians), time to get away.

There is an interesting sidelight here which sheds light on the interstate politics of the Persian Wars. Thebes had officially surrendered to Xerxes, and in the years after the was the Thebans had a very hard time living this down. Herodotus says that the Theban contingents who remained with the Spartans did so under compulsion, but moderns have seen that this makes little sense. At so crucial a time, Leonidas would be insane to choose to have hostiles in his midst. It is more likely that the Theban contingent consisted (as Diodorus says, 11.4.7) of exiles who had opposed the surrender to Xerxes, and that Herodotus was taken in by the anti-Theban propaganda which was flying thick and fast at Athens in the years before the outbreak of the Peloponnesian War.

There is a final dispute to be noticed concerning the identity of the hill to which Herodotus says the defenders retreated before finally being overwhelmed (7. 225). Until the excavations by Marinatos, it was generally assumed that this was the westernmost of the hills, Hill 1 by the remains of the Phokian Wall. However, the excavations proved that Kolonos Hill must be identified with Hill 2, due to the discovery of a large number of arrowheads similar in type to those found at Marathon, in a well at the Agora, and on the north slope of the Acropolis. The stone lion, the memorial to the heroism of the defenders, has never been found (though there is a modern restoration in the wrong place for the tourists) nor have the bones of the dead.


Béquignon, Y. La vallée du Spercheios Paris, 1937.

Burn, A.R. "Thermopylae and Callidromos" [Studies presented to D.M. Robinson, vol. I, 480-489] St. Louis, 1951.

- - - - - - - Persia and the Greeks London, 1962.

- - - - - - - "Thermopylae revisited, and some topographical notes on Marathon and Plataiai" [Greece and the Eastern Mediterranean in Ancient History and Prehistory = Studies to F. Schachermeyer, New York, 1977] 89-105.

Bury, J.B. "The Campaign of Artemisium and Thermopylae" ABSA 2 (1895) 83-104.

Dascalakis, A. Problèmes historiques autour de la bataille des Thermopyles [École Française d'Athènes - Travaux et Mémoires, fasc. XII] Paris, 1962.

Evans, J.A.S. "The final problem at Thermopylae" GRBS 5 (1964) 231-237.

- - - - - - - "Notes on Thermopylae and Artemisium" Historia 18 (1969) 389-406.

Grant, J.R. "Leonidas' last stand" Phoenix 15 (1961) 14-27.

Grundy, G. B. The Great Persian War and its Preliminaries London, 1901.

Hammond, N.G.L. "Thermopylae and Artemision" in CAH IV (2nd ed., Cambridge 1978) 546-561.

Hignett, C. Xerxes' Invasion of Greece Oxford, 1963.

Kase, E.W. & G.J. Szemler "Xerxes' march through Phocis" Klio 64 (1982) 353-366.

Kontorlis, K. The Battle of Thermopylae Athens, 1972.

Kromayer, J. & Veith, G. Schlachtenatlas - griechische Abteilung Leipzig, 1922.

Leake, W.M. Travels in Northern Greece II London, 1835.

Mackay, P.A. "Procopius' de Aedeficiis and the Topography of Thermopylae" AJA 67 (1963) 241-255.

Marinatos, Sp. N. & M. Robinson "Archaeology in Greece 1938-39" JHS 59 (1939) 199-200.

- - - - - - - & E.P. Blegen "News Items from Athens" AJA 43 (1939) 699-700.

- - - - - - - Thermopylae - An Historical and Archaeological Guide Athens, 1951.

Müller, D. Topographischer Bildkommentar zu den Historien Herodots Wasmuth, 1987.

Myres, J. L. Herodotus - Father of History Oxford, 1953.

Pritchett, W. K. "New Light on Thermopylai" AJA 62 (1958) 203-213.

- - - - - - - Studies in Ancient Greek Topography - Part IV (Passes) [U.C. Pub. Class. Stud. 28] Berkeley, 1982.

- - - - - - - Studies in Ancient Greek Topography - Part V (Roads) [U.C. Pub. Class. Stud. 31] Berkeley, 1985.

Sacks, K.S. "Herodotus and the dating of the Battle of Thermopylae" CQ n.s. 26 (1976) 232-248.

Schoder, R.V. Ancient Greece from the Air London, 1974.

Simpson, R. Hope "Leonidas' decision" Phoenix 26 (1972) 1-11.

Wallace, P.W. "The Anopaia Path at Thermopylai" AJA 84 (1980) 15-23.

Greek History Main Page