94 Herodotus probably arrived at these numbers from information picked up from various towns all round the Ægean, furnished by persons who had witnessed the passage of the army, or had heard of it from others who had been eye-witnesses. The enormous exaggerations of estimate made by eye-witnesses in estimating the numbers of a crowd of quite moderate dimensions, may easily account for the impossible numbers stated to and by Herodotus.

95 Those who state a much smaller number than this do not take into account the Oriental reliance on numbers, an Eastern characteristic which the Persian, unless he is much traduced, fully shared.

I hesitate to express any conjecture as to the possible maximum of the land force on this occasion. No evidence on this point can be said to exist.

96 The description of the route adopted renders it improbable that this body of troops included much more than the local levies of Western Asia. Sardes was the real place of rendezvous, though it is possible that some of the contingents did not join the main force until it arrived at Abydos.

97 Calculations show that the only eclipses about this time visible at Sardes occurred on October 2, 480 (mentioned H. ix. 10), and February 16, 478.

98 Herodotus says on the left. It is obvious that he has confused his point of view, and is speaking from that of a traveller voyaging in a ship up the coast.

99 The physical difficulties of the interior of the Balkan peninsula, together with the practical question of commissariat, render it probable that Herodotus is mistaken as to the nature of this inland march, and of the troops which undertook it. The main body of the army must almost certainly have pursued the coast road, which was easy, supplied with depôts of stores, and in touch with the fleet. The divisions which went inland were probably large detachments sent to inspire awe in the breasts of the inland tribes, so as to discourage them from any attempt to cut the line of communications.

100 Vide also p. 207.

101 The expedition to Thessaly must have been in the spring of 480. The Thessalian deputies came to the Isthmus “as soon as they were informed that the Persian was about to cross over into Europe” (H. vii. 172). Xerxes spent the winter of 481–480 at Sardes, and started thence “at the beginning of the spring” (H. vii. 37).

102 The Expedition to Tempe in Diodorus.—The account of the Tempe expedition as given by Diodorus differs in certain most important particulars from the brief narrative in Herodotus. It is possible that the manifest absurdity of the chronology in the historian’s work has created an undue prejudice against him as a source of evidence, a tendency increased by the second-hand nature of his information. It seems to me, however, that this latter characteristic constitutes a very strong reason for treating his information with respect. There is no reason why a bad historian should be a bad copyist; and whatever the man himself may have been, the writers from whose work he plagiarized so freely may in some cases have been most reliable witnesses. The passage on Tempe is in Bk. XI. ch. ii. 13, and it may be well to note the points in which he agrees with, and in which he differs from, the account in Herodotus.

It is quite plain that the account of Diodorus is not borrowed from Herodotus, nor does it show traces of having had, either in whole or even in part, a common source. In certain respects it is the more probable tale of the two. For example, the comparative smallness of the numbers sent is explained by the fact that the Lacedæmonians and Athenians hoped that the other Greeks would send contingents. Again, the size of the Persian force is given as a reason for going to Tempe, while Herodotus gives it as a reason for leaving the same, though he is inclined to reject the tale. It is inconceivable that the Greeks should at this time have been totally ignorant of the magnitude of the Persian force. Diodorus’ tale admits of the common-sense explanation that, having made up their minds to defend the North, they thought it best to choose a place where the Persian superiority in numbers would be of as little advantage as possible, though it was mainly for political purposes that so advanced a position was taken up.

In speaking of the medization of the States or clans, he evidently uses Θέτταλοι in Chapter II. in a general sense of the population of the region of Thessaly, whereas in Chapter III. he uses it of the Thessalians, properly so-called. There is nothing in Herodotus which contradicts Diodorus’ account of the medization. Both accounts may be true, though that of Diodorus is the fuller one. There is one point in it which is peculiarly supported by what Herodotus says. The Aleuadæ medized; but the mass of their subjects disapproved of this policy, and called in the Southern Greeks. The great Thessalian lords were lords of the plain rather than of the mountain. It was consequently the population of the plain which was opposed to their policy. It is therefore a remarkable fact that the tribes of the Thessalian region, which Diodorus mentions as having medized in the first instance, are those of the bordering mountain region; whereas those of the great plain remained true to the Greek cause until the withdrawal of the army from Tempe left them exposed to the overwhelming flood of the invasion.

103 Diodorus gives his name as Synetos.

104 The Thessalian appeal to the Congress at the Isthmus to send a force to Tempe is made to the προβόυλοι τῆς Ἑλλάδος, but the expression used by Herodotus in vii. 145 and 172, οἱ περὶ τὴν Ἑλλάδα Ἕλλήνες τὰ ἀμείνω φρονέοντες, has almost an official ring about it.

105 Cf. language used by the Greek embassy to Gelo at Syracuse (H. vii. 157).

106 It has been asserted that this passage from Diodorus is drawn from the passage of Ephorus, of which the fragment is a survival. It would seem to me that the essential difference between the two points to a difference of origin, and that they are two pieces of evidence on the question, and not one.

Note on the Sequence of Events in Relation to the Negotiations with Gelo.

It would, perhaps, be a mistake to lay too much stress on the indications of date in Herodotus with reference to the somewhat crowded incidents of this time. The actual dates cannot be settled, though it is possible to arrive at some idea of Herodotus’ views as to the sequence of events.

(a.) The Greek spies were sent to Asia at the time when Xerxes’ army was collected at Sardes (vii. 146).

(b.) The embassies to Gelo, Argos, etc., were sent after the despatch of the spies (vii. 146, ad init., 148, ad init.).

(c.) Gelo, after the departure of the joint embassy (vii. 163), and when he heard that Xerxes had crossed the Hellespont (ibid.), sent the treasure-ships to Delphi.

(d.) Before he sent these vessels he knew that he had to expect a Carthaginian attack (vii. 165, ad fin.).

The evidence is inconclusive. We lack the means of deciding the sequence of the departure of the embassy and of the acquisition of the information with regard to the coming of the Carthaginian expedition.

107 Bergk, “Pœtæ Lyrici Græci,” Ed. 4, v. iii. p. 485: Πολλὴν δὲ παρασχεῖν σύμμαχον Ἕλλησιν χεῖρ᾿ ἐς ἐλευθερίην.

108 Vide Note at end of chapter.

109 Livy, xxxvi. 15—“Hoc jugum (Œta) ab Leucate et man ad occidentem verso per Ætoliam ad alterum mare orienti objectum tendens ea aspreta rupesque interjectas habet, ut non modo exercitus sed ne expediti quidem facile ullas ad transitum calles inveniant;” and again (Livy, xxxvi. 17), Acilius Glabrio, speaking of Thermopylæ, “Quippe portæ sunt hæ, et unus inter duo maria clausis omnibus velut naturalis transitus est.”

110 There is considerable mule traffic through it at the present day.

111 It would, I reckon, be possible for a traveller to go from one plain to the other by this route without attaining a height much over a thousand feet.

112 It may be well to adduce one or two striking instances of this, apart from the one at present under consideration:⁠—

1. Circ. 350 B.C., Thermopylæ was the great obstacle to Philip’s advance south. His energies were centred in getting hold of the pass. He never attempted an assault upon it, but finally got hold of it by bribing Phalæcus, the Phocian condottiere. He then left a garrison at Nicæa near Thermopylæ (Dem. ad Ep. Phil. 4). There must have been some supreme objection from a military point of view to the Asopos pass, since Philip, who can hardly be suspected of military incapacity, never tried to turn Thermopylæ by using it. He was not pressed for time. He patiently allowed years to elapse before he got hold of Thermopylæ. Thermopylæ was all-valuable to him, and, what is more striking, absolutely necessary, in his opinion, for an advance southwards. (Vide Hogarth, “Philip and Alexander”).

2. In 279 B.C. (Pausanias, x. 20) Brennus, with more than 150,000 Gauls, invaded Greece. The object of the expedition was plunder and nothing else. If he could have got past Thermopylæ that object would have been attained. With such numbers he could have done what he liked, especially if, after getting through the Asopos defile, he had, before going south, turned Thermopylæ by way of Hyampolis and forced the Greeks to evacuate that pass. Pausanias, who seems to have ample information as to the details of this Celtic raid, gives the following list of the defending force:⁠—

   Bœotians, 10,000 infantry, 500 cavalry; Phocians, 3000 infantry, 500 cavalry; Megareans, 4000 infantry; Ætolians, 7000 infantry, with numerous light armed, and cavalry; Athenians, 1000 infantry, 500 cavalry, with numerous triremes; Mercenaries from Macedonia and Asia, 1000 infantry.

   The total cannot have been far short of 25,000 men.

   The force was so large that the commanders were enabled not merely to provide for the defence of the pass itself, but were also able to send out cavalry and light armed to dispute the passage of the Spercheios, a move which Brennus, who was, as Pausanias remarks, “not altogether wanting in understanding, nor, for a barbarian, without a certain amount of experience in devising stratagems,” frustrated by sending a number of his men across the bar at the river mouth. On this the Greek advanced guard retreated to the pass. Brennus then had bridges thrown across the river, and attacked Heraklea. The Heraklea of that date was probably situated on the mountain immediately west of the mouth of the Asopos ravine, on a site now known as Sideroporto. He did not take the place; and Pausanias adds in reference to this, that Heraklea was to him “a matter of lesser moment: he considered the main point to be to drive out of the pass those who were in occupation of it, and to make good his passage into Greece south of Thermopylæ.”

   It is of course manifest that such a passage as the Asopos ravine would be easily defensible, and the Herakleots may have blocked it. There is no question that, had it been passable, Brennus might have used it. The striking fact is that he did not use it, but spent his strength on a terrific failure at Thermopylæ.

3. In 224 (vide Polyb. xi. 52) Antigonus, wishing to get to the Isthmus, marched with his army by way of Eubœa. “He took this route,” says Polybius, “because the Ætolians, after trying other expedients for preventing Antigonus bringing this aid, now forbade his marching south of Thermopylæ with an army, threatening that, if he did, they would offer armed opposition to his passage.”

   It is to be noted that, as at the time of Brennus’ assault, the defenders of the pass were also in possession of Heraklea.

4. In B.C. 208 (vide Polyb. x. 41) the Ætolians, seeking to prevent the passage of Philip of Macedon southwards to aid his allies, “secure the pass of Thermopylæ with trenches and stockades and a formidable garrison, satisfied that they would then shut out Philip, and entirely prevent him from coming to the assistance of his allies south of the pass.”

   In this case also (vide chap. 42) the Ætolians were in possession of Heraklea.

5. In a passage already quoted, Livy (xxxvi. 15) is most emphatic in his statement that the only practicable military route by Œta is that through Thermopylæ. He is describing the attack of the Romans under Acilius Glabrio upon the troops of Antiochus who were defending the pass, and it is again reported that the allies of the defenders were in possession of Heraklea.

We are now in possession of practically all the data which can be obtained from the ancient historians with regard to the exact significance of the Asopos ravine, and the route through it. It must of course be borne in mind that the information of the historian Livy with regard to the topography of the Thermopylæ region was second-hand; but yet, in spite of that, there is a certain consistency about the evidence which enables us to form highly probable conclusions with regard to the exact value of this factor in the strategical geography of the region.

It seems to me to have been a recognized principle in later times that an effective defence of Œta included the occupation of Heraklea as well as of Thermopylæ, and the only conceivable reason for the existence of such a view is that Heraklea commanded the passage of the Asopos ravine.

The site of the Heraklea of this period is to be sought, I venture to think, at the place called Sideroporto, where there are large remains of a strongly fortified town. It is high on the slope of Œta, in an exceedingly inaccessible position, in the angle, as it were, between the Asopos ravine and the line of the Trachinian cliffs.

A local tradition, probably of recent date, and due, like so many traditions of modern Greece, to the visit of some inquirer whom the natives regarded as an authority, attaches the name of Heraklea to certain ruins which stand on the summit of a remarkable flat-topped mountain in the valley at the head of the Asopos ravine, to which reference has been already made, between the plains of Malis and Doris. It is infinitely more probable, however, that this was the stronghold of those Œteans whom Thucydides mentions.

The site is more than two hours distant from the nearest point of the Malian plain, at the outlet of the Asopos ravine.

113 Cp. H. vii. 175, ad fin.

114 It has been criticized in modern times on strategical principles (e.g. by Delbrück), for which a universality of truth has been claimed. It is said that, given two adversaries of equal strength, that one places himself at a disadvantage who attempts to defend the passage of a range of mountains. It is manifest that the assailant can concentrate his efforts on the forcing of one passage, whereas the defender has to distribute his defence among all the practicable passages of the chain. In the particular case of Mount Œta it is urged that there was, besides Thermopylæ, at least one practicable passage, and this is stated to have followed the modern road from Malis into Doris; which passes over the low part of the chain immediately east of the Asopos ravine.

Could it be proved that such a road ever existed the general criticism would be sound. As a fact, all but demonstrable proof exists that no such road, practicable from a military point of view, ever did exist in ancient times. Leave out of the calculation the Greek of 480 and the Gaul of 279—although in the case of the latter, if Pausanias’ evidence be worth anything, the Malians showed a very pardonable desire to expedite his departure from the region, and would have been most anxious to show Brennus such a road, had it existed—and merely take into consideration the Greek, the Macedonian, and the Roman of later times. For years and years these peoples were fighting in every part of North-East Greece. They knew its topography by heart. The land became the veriest strategic chessboard that ever existed in ancient warfare, on which every move could be calculated to a nicety. And yet Thermopylæ remained the same—that square on the board where king and consul could alike be checked. Could it have been so had such a path existed?

It has already been seen that the holding of Heraklea was regarded as a necessary factor in the defence of Thermopylæ. That Heraklea was almost certainly situated at Sideroporto, commanding the Asopos ravine indeed, but cut off by that very ravine—a mere crack several miles long and nine hundred feet deep—from the line taken by this imaginary road; that is to say, Heraklea would have been absolutely useless for its defence. If it existed, why then did neither Greek, Macedonian, nor Roman use it? Why did Philip of Macedon, the father of Alexander, shirk the attack on a pass which he could so easily have avoided?

It has already been said that the Greeks have made a new road along that line. It is an excellent piece of work, but so great is the climb to the summit of the pass that a two-horse carriage takes three hours to accomplish it. The gradient of the hill-side can best be imagined when it is stated that after climbing for an hour and a quarter along this road, the traveller finds himself within less than half a mile of the point from which the climb began, and the greater part of that half-mile is vertical.

In criticizing ancient warfare a tendency is but too frequently displayed to ignore the main factor of all warfare—the human element. In criticizing Greek warfare in particular, it is, moreover, too often the case that the critic is either unaware of, or has never realized, the nature of the country with which he is dealing. An ordinary Greek hill-side, though it appears easy of passage when viewed from even a short distance, presents difficulties which can hardly be paralleled in any other country in Europe. Thick, low, strong bush, much of it thorny, covers it just to a sufficient depth to hide the thickly sown, razor-edged rocks beneath. Human nature as represented by the Greek hoplite in his heavy armour could not face it, and progress over it even for a light-armed man is very slow and very exhausting. The strategy and tactics of war are bounded by the difficult rather than by the impossible. There can be no question that the passage over this part of the range of Œta can never have been practicable to anything more than the merest skeleton of a flying column, and could not possibly have been negotiated by any force sufficiently large to affect the defence of Thermopylæ by any turning movement, or sufficiently well provided with provisions to accomplish the long circuit which such a turning movement would have demanded.

There is one more striking proof that such a road did not exist in 480. Had it existed it must have crossed that path of the Anopæa by which Hydarnes and his men turned the pass. If so, why did he make the long circuit by the Asopos ravine, when a shorter way was practicable?

115 Cf. the mistake made as to the defensive nature of the position at Tempe; also, the ignorance of the existence of the path of the Anopæa at Thermopylæ.

116 The expedition to Thessaly was made while Xerxes was at Abydos, certainly not later than April, 480. The departure of Leonidas for Thermopylæ took place a little before the Carnean festival, about the beginning of the month of August.

117 Diodorus’ account of the circumstances preceding the battle is manifestly an imaginary tale of indeterminate origin concocted after the event.

118 This seems to indicate that the Persian camp was altogether outside the west gate, and not any part of it in the plain of Anthele.

119 This shows clearly that the wall was not, as some have supposed, on the low ground at the pass of the middle gate, but on the neck of the first mound (vide note on Topography of Thermopylæ). Had it been on the low ground, the scout would, from the comb of the mass of stream débris of the great ravine, have been able to see over it.

120 These last words are, I believe, the true translation of the expression in Herodotus. There would be little point in repeating the fact of the river flowing through the ravine as a sort of mark of the identity of a stream whose course the historian had recently described with considerable detail. There is much point in the indication of what investigation at the present day shows to have been the fact, that this path did start from the Asopos ravine.

My own impression is that it sprang into use originally as a means of communication between that upper valley which I have mentioned as existing in the range of Œta, and probably also the Dorian plain, and Thermopylæ, when a flood of the Asopos rendered the ravine impassable. It would also form a direct means of communication between the Œteans and Locrians without passing through Trachinia.

121 It is of course impossible to deduct the number of killed in the previous fighting, simply because we have no information as to what that number was.

122 Epialtes’ calculation that the circuit of the path would be completed about the middle of the morning must, judging from the details given of the actual march, have been singularly correct.

123 Leake says that the descent was not much less than the ascent in actual distance; but that as the ground was better, and the march performed by daylight, the time spent was shorter. Leake is certainly in error. The place were the Phocians were surprised is recognizable with certainty, I think, at the present day. It corresponds with what Herodotus tells us of the incident, and it is absolutely the only place along the whole path where the events narrated could have taken place. When the Persians reached that point, which is probably the highest altitude attained by the path, they would have traversed two-thirds of the whole distance. I must say that Leake’s attempt to reconcile his views with those of Herodotus by saying that the rest of the path is easier than that previously traversed is quite contrary to my own actual experience. From the summit to Drakospilia its character is that of a track winding amid rocks through a thick fir forest. Not until you get close to Drakospilia does the country really open up.

124 They advanced, that is to say, to a position somewhere near the modern baths.

125 They had fought, that is, on the low ground at the foot of Kallidromos immediately to the west of the mound.

126 The statement that they already knew that they must be taken in the rear is in accord with Herodotus’ idea of what took place. It is, however, probable that they heard early from Alpenoi, to which some of their sick had been sent, of the fact that the other division of their army had not succeeded in stopping, or, possibly, had not attempted to stop Hydarnes.

127 The pillar at Sparta, with their names inscribed upon it, remained standing in Pausanias’ time (iii. 14, 1).

128 The position was well designed for a last desperate stand. The rear was protected by the small but deep valley between the first and second mound. It is noticeable that they did not attempt to defend the wall. It may seem strange that they should not have done this. The position of the wall, however, running along the neck of land joining the hillock and the slope at Kallidromos, would expose its defenders to an attack from the rear. The Greeks evidently retreated from their position near the modern baths; through the narrows between that and the hillock; and up to the west slope of the latter, passing the wall at the summit of the slope on to the mound itself.

129 The parallel diary of events as it appears in Herodotus is as follows:⁠—

130 Herodotus does not give any indication as to the time at which the Greeks received news of the disaster. It is necessary therefore to make certain calculations as to the earliest possible moment at which the news can have reached them. As far as can be seen from the narrative, the ten Persian scouting vessels started from Therma on the same day as the main body of the fleet, but probably at an earlier hour. It must have been well on in the morning before they came upon the Greek vessels off the mouth of the Peneius, which is fifty miles from Therma. The only conceivable means by which news of the engagement could have reached Skiathos, some seventy miles south of this point, is by the appearance of those ten vessels with the captured Greek ships in their company. That being the case, the Greek fleet at Artemisium cannot have received the news before the evening of the day.

131 The ancient Mekistos.

132 It is more probable that it was under the shelter of the great cliffs of Mount Kandili, in the neighbourhood of the modern Limni. There is a sandy shore for many miles at the foot of those cliffs, upon which vessels might be conveniently drawn up.

133 A nine-knot steamer takes about seven hours from Chalkis to Stylida, which is about the same distance as from Chalkis to Artemisium. There is no question that a trireme could maintain a high rate of speed for hours together. Nor is there reason to doubt Herodotus’ statement that the voyage of the Persian fleet from Therma to the Sepiad strand took but one day, a distance, that is to say, of one hundred and twenty miles in fourteen hours of daylight, over eight miles an hour—even supposing that such a large number of vessels could put out and put in in the dark. The probability is, however, that the fleet never went to Chalkis at all; or, if it did, that it moved up the Euripus after receiving the news of the disaster to the Persian fleet, so as to be ready to go to Artemisium without delay so soon as the storm ceased.

134 The strong bias which Herodotus displays in his references to Themistocles is of itself sufficient to render the tale of bribery open to suspicion. Furthermore, the sum mentioned, thirty talents, is an extraordinarily large sum for the people of North Eubœa to raise at short notice.

135 Diod. xi. 12, mentions this, but gives the number of the squadron as three hundred.

136 Diodorus gives no exact indication of the time of despatch, though he mentions it immediately after describing the arrival at Aphetæ.

137 Same day as first engagement (vide note over page).

138 If any calculation can be made from this very defective chapter of Herodotus’ history, this day must have been the eighteenth day. H. viii 14, 15.The three combats at Artemisium are represented as having taken place on successive days. The last took place on the day of the disaster at Thermopylæ, i.e. the twentieth day. Therefore the first took place on the eighteenth, H. viii. 9. and it is represented as having taken place on the evening of the day on which the council of war was held.

139 This view is supported by Herodotus’ account of what took place next day. The storm in which the Persian flying squadron is wrecked takes place on the evening of the eighteenth day. When the storm ceased we do not know. But it is certain that the fifty-three Attic vessels must have ridden it out at Chalkis, and that they, after it was over, made the long voyage from Chalkis to Artemisium, where they found the Greek fleet. The storm must have been a brief one; and if, as Herodotus says, there had been a definite resolution on the part of the Greek commanders to move south in the early hours of the morning of the nineteenth day, no reason is apparent why it should not have been carried out. The real design of the Greeks was probably to make an attempt to beat the divided Persian fleet in detail.

140 Though Herodotus is aware of a connection between the positions at Thermopylæ and Artemisium, there is nothing whatever in his account which suggests that he understood how necessary the connection was for the maintenance of the pass. Had he appreciated this, he would hardly have treated as serious history such parts of the Artemisium tradition of his time as asserted that the responsible Greek commanders ever entertained the idea of such action as must have inevitably sacrificed the lives of the defenders of the pass. He has given the irresponsible gossip and criticism of the Peloponnesian section of the fleet the appearance of responsible and authoritative design, and has served up the whole with copious Attic sauce. There is, however, no reason for supposing that the historian was in any way guilty of historical dishonesty. He simply did not possess that knowledge of military affairs which would have enabled him to see the flaws in the evidence which came to his hand; and this negative defect was further complicated by what was, from the point of view of strict history, the positive one of accepting anything in the tradition of the war which would bring into relief the patriotic services of Athens. If we tone down the intensely Attic colouring in Herodotus’ account of Artemisium, that is to say, such passages as are designed to bring into relief the difficulty of keeping the fleet at its station, we have, in all probability, a good historical account, in so far as it goes, of this part of the campaign of 480.

141 The manœuvre of the διέκπλους seems one of the most simple things in the world when it has been discovered. Yet in modern times it took the English sailors more than a century of hard fighting to find out its effectiveness. Thucydides, who knows what he is talking about in naval matters, certainly conveys the impression that it was an invention of his own time, or, at any rate, that it had, as a manœuvre, been gradually evolved within the period of the Pentekontaëtia. And yet, here we have it at Artemisium! Nay, more than that, fourteen years earlier, according to Herodotus, Dionysios of Phokæa was trying to teach it to those unappreciative Ionians at Ladé. It is probable that both in this passage and in the one relating to Ladé, Herodotus is guilty of an anachronism in attributing that manœuvre to the naval warfare of the first quarter of the fifth century. The term was probably much in men’s mouths at the time which he wrote, and, in his ignorance of naval matters, he assumed that the ruling idea in the sea tactics of his own day might be safely attributed to the previous generation. Compare also H. viii. 11 with Thuc. ii. 83, ad fin.

142 There is an undesigned consistency between the two accounts of the effects of the storm in North and South Eubœa respectively. A glance at the map will show that, (1) in the North, the driving of the wreckage towards the shores of Aphetæ; (2) in the South, the driving of the 200 vessels upon the Hollows of Eubœa, both indicate a storm from the South or S.S.W.

143 It seems to have taken some thirty hours to round Skiathos, and voyage down the east coast of Eubœa.

144 The identity of these bays with Τά Κοίλα has been called in question in modern times. If this passage in Herodotus were the only evidence we possessed, the question of their position would manifestly be a very open one. All that Herodotus’ language seems to indicate is that they were a well-known feature in the geography of South Eubœa. Had they not been so, we should have expected so painstaking a topographer to have given some indications of their actual position. His silence, and the inference to be drawn from it, is not without significance. The Hollows would hardly have been a well-known feature had they been east of the South Cape, away from the line of sea traffic; whereas on the west shore they would be in full view of all vessels using the frequented passage of the Euripus. I think, too, that any one who has seen that coast of Eubœa, either from Attica, or when passing up the channel, cannot but have been struck with the depth of the colour which the retiring coast-line of these bays gives to the Eubœan landscape thus viewed. Their recesses give that appearance of “hollowness” from which the ancient name must have been derived. We are not, however, dependent on Herodotus alone for indications as to their locality. Vide Liv. 31, 47; Strabo, 445; Valer Max. 1, 8, 10.

145 It is exceedingly unlikely that the Persian squadron would have been able to force the narrows at Chalkis, if, as was almost certainly the case, the fifty-three Attic vessels were ready to defend it. But had they put in at Eretria and blocked the channel south, the position of the main Greek fleet, in case of anything resembling a reverse at Artemisium, would have been very precarious.

146 The fact that they were able to single out a special contingent for attack confirms, by implication, Diodorus’ statement as to the scattered nature of the anchorage at Aphetæ.

147 The effect of the engagement on the minds of the Greeks is mentioned in language which is almost, word for word, a repetition of that which he has used on a previous occasion. He says, Δρησμὸν δὴ ἐβούλευον ἔσω ἐς τὴν Ἑλλάδα.—H. viii. 18. Cf. the expression in viii. 4.

148 In speaking of Doris, Herodotus says: Ή δὲ χώρη αὕτη ἐστὶ μητρόπολις Δωριέων τῶν ἐν Πελοποννήσῳ. That the land had a Dorian population in the fifth century B.C. is undoubtedly the case; but its claim to be metropolis of the Dorians of the south was in all probability set up by the Spartan authorities, as affording a convenient pretext for interference in Greek affairs north of Isthmus. It is probable that this corner of Greece, of which the Malian plain was the centre, contained patches of various peoples which had in different ages traversed the peninsula, or which had been driven into its mountain fastnesses by the passage of invaders:⁠—Dorians, Œtæans, Trachinians, etc., were probably such remains of larger tribes.