² For the number cf. Hdt. ix. 27 and vii. 76 with Stein's note.

³ I gather from Hauvette's, M. remarks, Hérodote, p. 310Google Scholar, that M. de Gobineau makes the ἄρχοντες myriarchs, but I have not seen his Histoire des Perses.

⁴ Cf. Aeschylus, , Pers. 315Google Scholar, ἵππου μελαίνης ἡγεμὼν τρισμυρίας. It is true that in the preceding line this officer was only μυριόνταρχος, which is probably nearer the fact, but ‘30,000 horse’ may have implied to Aeschylus ‘all Xerxes' cavalry.’ What does μελαίνης mean? Were the horsemen black-capped and black-coated, like Circassian irregulars, or did the Persians share the Turkish preference for black chargers?

⁵ There is a flaw in the text of Herodotus, vii. 86. The word Κάσπιοι occurs twice. I would read Σάκαι for the first, because (1) the Sakae specially distinguished themselves among the Persian cavalry at Plataea (ix. 71), but are not mentioned here; (2) their infantry was brigaded with the Bactrians (vii. 64); (3) ΚΑΙΣΑΚΑΙ might easily be corrupted into ΚΑΣΠΙΟΙ through reduplication of ΚΑΙ and the proximity of ΚΑΣΠΙΟΙ two lines lower down.

⁶ It is unfortunate that Aeschylus is useless as evidence on this point. He appears to be content with any Persian name that will fit his verse.

⁷ Persae, 341–43:

Ξέρξη̢ δὲ, καὶ γὰρ οἰ̑δα, χιλιὰς μὲν ἠ̑ν

ὡ̑ν ἠ̑γε πλη̑θος, αἱ δ᾿ὑπέρκοποι τάχει

ἑκατὸν δὶς ἠ̑σαν ἑπτα θ᾿· ὡ̑δ᾿ ἔχει λόγος.

⁸ Strictly perhaps one ought to say he starts with 1327, for he has added 120 from the Greek cities of Thrace (vii. 185), but what few ships may really have come from them (e.g. from Samothrace, viii. 90) are probably already included in the Hellespontine contingent, and I do not believe that Herodotus pays any further heed to them in his reckonings. The total losses in the battles at Artemisium cannot be determined.

⁹ I observe that Ed. Meyer has made this point, Gesch. d. Alt. iii. §217. Cf. also Bauer, A. in vol. iv. of Jahreshefte des Österr. arch. Inst. pp. 93–4Google Scholar.

¹⁰ The line followed by the new railway from Xanthe to Drama by the garge of the Nestus is a fine piece of engineering only opened up by blasting. It was not a practicable road.

¹¹ Cf. Hdt. viii. 2, which may be coloured by afterthought so far as ‘hegemony’ is concerned v. Meyer, Ed., Forsch, ii. pp. 218–9)Google Scholar, but has probably some foundation in fact, and certainly is dramatically true to the situation. Cf. Plut., Them. 7Google Scholar.

¹² Cf. Ath. Pol. 22, Plut., Them. 11Google Scholar, and Arist. 8.

¹³ Meyer, Ed., Forsch, ii. pp. 210–17Google Scholar, has nearly expressed my view of the attitude of these states and Herodotus' treatment of them. As regards Argos Grote had already led the way, third ed. vol. v. pp. 88–90.

¹⁴ In logical parlance the opposition between Sparta and Argos, like the later opposition between Athens and Corinth, was contradictory, whereas the opposition between Sparta and Athens was merely contrary. Cf. Thuc. v. 91.

¹⁵ Cf. Hdt. vii. 206, φυλακὰς λιπόντες ἐν τη̢̑ Σπάρτη̢ viii. 26. αὐτόμολοι ἄνδρες ἀπ᾿ ᾿Αρκαδίης. Possibly the need of gathering the harvest may also have delayed the Peloponnesians. Cf. Meyer, Ed., Gesch. d. Alt. iii. § 234Google Scholar.

¹⁶ Cf. Hdt. vii. 175, ἀγχοτέρη τη̑ς ἑωυτω̑ν

¹⁷ The lessons of Marathon had not been lost on Themistocles; cf. Plut., Them. 3Google Scholar.

¹⁸ Συνετός, Diod. xi. 2, is an obvious clerical error, ΣΥ = ΣΥ and ΑΙ has been absorbed in the Ν Mnamias and his 500 Thebans may have joined the allies in Thessaly (Plut., de Her. malign. 31)Google Scholar, but the authority of the Boeotian Aristophanes is not above suspicion on such a point, as the context indeed might suggest.

¹⁹ See Hauvette's, arguments, Hérodote, p. 327Google Scholar, which do not, however, face the pluperfect ἐγεγόνεε at the beginning of Ch. 145.

²⁰ There is no cogent reason for rejecting a line of the two oracles. The alternative plans of defence were of course known at Delphi, and it was obvious that Salamis was the naval counterpart to the isthmus. Mr.Bury, (Class. Rev. x. (1896), p. 417)Google Scholar detects in the words ἔτι τοί ποτε κἀντίος ἔσση̢ a reference to Plataea, but so natural an idea needs no special explanation. ‘He that fights and runs away will live to fight another day. ᾿´Εσχατα γαίης could, if it proved convenient, be interpreted to mean the Peloponnese.

²¹ viii. 77; cf. 76, where Cynosura appears to be taken from the oracle, as Grote saw, and Munichia is mentioned for the sake of the temple of Artemis there, (cf. Stein ad loc). Κέος is quite unknown, but Κέον might conceivably be a corruption of Κέον and so transferred from a narrative of the Euboean operations. Grote was right, I believe, in suspecting the current explanation of the names (3rd. ed., vol. v. 176); but it is likely that the long eastern promontory of Salamis was called Κυνόσουρα (cf. Plut., Sol. 9Google Scholar, where χηλήν τινα πρὸς τὴν Εὔβοιαν ἀποβλέπουσαν suggests a confusion between two Κυνόσοιραι or a gloss upon the word), and that this coincidence helped the transfer. In the third line of the oracle I put no stop, and take Ἐλπίδι μαινομένη̢ to qualify πέρσαντες My interpretation of the oracle is supported by the lines on the monument of the Megarians, (Oesterr. Jahresheft. ii. pp. 238–9)Google Scholar—

μὲν ὑπ᾿ Εὐβοία καὶ Παλίω, ἔνθα καλει̑τε

ἁγνα̑ς ᾿Αρτέμιδος τοξοφόρου τέμενος

²² The epigram (Hdt. vii. 228) gives 4000 from the Peloponnese, and Herodotus, viii. 25, evidently took his 4000 dead from this source, although it speaks only of Peloponnesians and not of dead but of living. Herodotus vii, 202, enumerates only 3100 Peloponnesians. He elsewhere (viii. 25) mentions Helots, and seems to imply (vii. 229) that each Spartan was attended by one Helot. But it is not likely that the epigram includes the Helots any more than Herodotus does. Herodotus' list, therefore, appears to be incomplete. Isocrates (Paneg. 90, Archid. 99), Ctesias (Pers. 25), and Diodorus (xi. 4) speak of 1000 Lacedaemonians. This supplement may be thought to be either confirmed by Demaratus' words (Hdt. vii. 102) or derived from them. Diodorus gives 1000 Locrians and 1000 Malians. The Malians are improbable, but 1000 is nearer the mark for the Locrians than Pausanias' 6000 (x. 20, 2), and not many of them need have been hoplites.

²³ Griech. Gesch., 2nd ed., vol. ii. p. 681, note 3. Cf. Hauvette, , Hérodote, p. 372Google Scholar, Grundy, , Great, Persiem War, p. 319Google Scholar. Mr. Bury, on the other hand, accepts τριται̑ος and deducts two days from Xerxes', delay. (Ann. of Brit. Sch. at Athens, ii. pp. 95–7.)Google Scholar

²⁴ Herodotus (vii. 179–83) conceives that the Persian squadron made straight for Skiathus, and the three ships ran on beyond the rest as far as the rock, ἐπήλασαν περὶ τὸ ἕρμα But this conception does not fit his story.

²⁵ Mr.Grundy, , Great Persian War, p. 324Google Scholar, supposes that the whole Greek fleet had to run before the storm round Cape Kenaeum. It is more likely that the ships were drawn up for the night on the strand at Artemisium, but even if afloat they had fair shelter there, and safety hard by at Oreus; v. Lolling, in Ath. Mitt. viii. p. 16Google Scholar.

²⁶ Possibly Herodotus' information about the capture of the three ships came from Pytheas, (vii. 181, viii. 92), about the signal from some one with the fifty-three ships, and he has unskilfully combined them.

²⁷ It might help to explain Xerxes' delay before Thermopylae.

²⁸ The dead and wreckage of the first battle drift to shore at Aphetae—ἐξεφορέοντο ἐς τὰς ᾿Αφετάς. The exact position of Aphetae is unknown. In spite of Hdt. vii. 193–5, it is difficult to believe it lay inside the gulf of Pagasae, and he elsewhere estimates its distance from Artemisium at 80 stades (viii. 8). The ἐξ in ἐξεφορέοντο does not help us, for it has no reference to direction (cf. ἐκφέρεσθαι, viii. 49, 76, and ἐκπίπτειν). But one would expect to find Aphetae rather east than west of Artemisium, else the Greeks would have been in danger of being cut off from the Euripus. Near Olizon would be the natural site, perhaps on the narrow isthmus, and so practically on both the gulf and the outer strait. Possibly Herodotus imagined that the east and south coasts of Magnesia made an acute angle, and reckoned the latter to the gulf. The tide in the Euboean channel (Hdt. vii. 198) might account for the drift of the wreckage without any wind at all; but I do not pretend to know how it sets.

²⁹ If the maps are accurate there can be very little daylight, if any, between the west point of Skiathus and the east point of Magnesia, as viewed from Mount Dirphys.

³⁰ Mr. Grundy (p. 269 and elsewhere) seems to me to make too much of the difficulty of transport. He appears to imagine Xerxes' commissariat train entirely on wheels, whereas one might gather from Herodotus that it consisted entirely of pack-animals of various kinds, including camels (see, among numerous references, especially vii. 125, 187; and cf. ix, 39). Probably the truth lies between these extremes. But there are very few tracks too difficult for the light, narrow Asiatic ox-waggon, and an oriental army (pace Herodoti) requires extra-ordinarily little baggage or even food.

³¹ On Heraclea Trachis, v. Livy, xxxvi, 22–4, Thuc. iii. 92, Hdt. vii. 199, Strabo, 428, Paus. x. 22, Leake, , N. Greece, ii. pp. 24–31Google Scholar. Herodotus implies, I think (with Mr. Grundy, p. 282), that the lower town was on the Thermopylae road. It is clear from Thucydides that there was never any change of site, although in Roman times (Strabo, Paus.), when Heraclea had completely retreated up the hill (cf. Livy), the ruins of the lower town, six stades below, were exclusively known as Trachis.

³² It may be noted that the Locrians and Phocians mustered ἐς τὴν Τρηχι̑να and that Xerxes commands all northern Greece, μέχρ Τρηχι̑νος (Hdt. vii. 203, 201).

³³ So perhaps does the phrase ἐν δεξιη̢̑ μὲν ἔχοντες ὄρεα τὰ Οἰταίων, ἐν ἀριστερη̢̑ δὲ τὰ Τρηχινίων (v. Leake, pp. 54–5), and ἡ περίοδός τε καὶ ἀνάβασις in vii. 223.

³⁴ Herodotus's expression, ἐπ᾿ ἀκρωτηρίψ του̑ ὄρεος, (vii. 217), can no more be pressed than his κατὰ ῾ράχιντου̂ ὄρεος(216), or ἐπὶ του̂ ὄρεος τὸν κόρυμβον (218). It is likely enough that he travelled the coast road (as Mr. Grandy argues at large), but the upper path remained to him vaguely something ‘up there.’ Possibly his guide or his own imagination fastened upon the rocks which Mr. Grundy (p. 302) calls the Great Gable, and they may be either the ἀκρωτήριον or the κόρυμβος.

³⁵ Except perhaps the words καὶ τοι̑σι ἐτύγχανον παι̑δες ἐόντες in 205. Can the τε in the preceding clause ἄνδρας τε τοὺς κατεστεω̑τας τριηκορίους be a corrupt reduplication of τ´= τριηκοσίους and καὶ τοι̑σι ἐτύγχανον παι̑δες ἐόντες an afterthought added with the note on the Thebans which fills the rest of the chapter, (cf. the opening words of 206, which would fit on very well)?

³⁶ The view so often expressed (e.g. by Busolt, , Die Laked. p. 419seqq.)Google Scholar, that the Spartan government invented the oracle to excuse themselves, seems to me to make a false distinction between the Ephors and Leonidas, and between the Spartans and the general council of the allies.

³⁷ Herodotus (vii. 225) clearly means by the κολωνὸς ἐν τη̢̑ ἐσόδψ where the last stand was made, one of the two mounds in the middle ‘gate,’ cf. vii. 176. But it is perhaps possible that the tradition was attracted to the lion, and the last stand was really on the mound nea the east ‘gate.’ If so, the Greeks may have been overwhelmed in an attempt to withdraw, and the story of the sortie in front of the wall (Hdt. vii. 223, Diod. xi. 9, 10 Plut., de Her. mal. 32Google Scholar) may have arisen out of the confusion of the two mounds. Cf. Leake, , N. Greece, ii. pp. 36–7, 52Google Scholar; Grundy, Gt. Pers. War, pp. 288–90Google Scholar.

³⁸ It may be urged against the theory here stated that Leonidas must have had constant information of Hydarnes' progress. I quite admit the force of this objection, but every alternative is open to some objections, and this theory seems to me the least objectionable. It keeps nearer to Herodotus than that which I regard as the second-best, namely, that Leonidas sent the Peloponnesians to meet Hydarnes, whether in the pass near Bundonitza (as I should say), or near Alpeni (Bury, l.c. p. 102), or on the Anopaea (Grundy, p. 308).

³⁹ See Bury, in the Classical Review, x. (1896), pp. 416–7Google Scholar. His argument is perhaps open to criticism in details—the occupation of the acropolis is not quite in harmony with the decree recorded by Plutarch, (Them. 10)Google Scholar; the fortnight's siege is hard to reconcile with Hdt. viii. 66–70—but seems to me right in the main, and is quite borne out by a consideration of the position of parties.

⁴⁰ The question whether we ought to read τετρακοσίας in Thuc. i. 74, neither affects the problem here nor gains any elucidation from it.

⁴¹ See Busolt's excellent note, Griech. Gesch. 2nd ed. ii. p. 641, with the references there given.

⁴² ᾿Αθ. πολ. 22, ἄρχοντος ῾γφιχίδου, cf. Plut., Arist. 8Google Scholar. Probably on the return of the expedition from Thessaly.

⁴³ On the topography of. Lolling, in Hist. und Philol. Aufsätze E. Curtius gewidmet, pp. 1–10Google Scholar; Goodwin, in Papers of the Amer. Sch. at Athens, i. pp. 239–62Google Scholar; Milchhoefer, , Erläut. Text zu Karten von Attika, vii.–viii. pp. 26–35Google Scholar; Bauer, , Oesterr. Jahresh. iv. pp. 90–111Google Scholar.

⁴⁴ I am glad to see that Mr.Grundy, has adopted this suggestion (Gt. Pers. War, p. 405)Google Scholar. Plutarch, , de Her. mal. 39Google Scholar, justly appeals to the epitaphs on Adeimantus and the Corinthians to refute the story of their flight.

⁴⁵ The passage in Plutarch, , Sol. 9Google Scholar, which might throw some light on the point is unfortunately mutilated. Solon's sham Megarians may have approached Salamis from the side of Megara. Cf. De Her. mal. 39, περὶ τὰ λήγοντα τη̑ Σαλαμινίας.

⁴⁶ On this question Meyer, Ed. has some excellent remarks, Gesch. d. Alt., iii. § 224Google Scholar.

⁴⁷ Cf. Mr.Rose's, J. H. note in the Engl. Hist. Rev. July 1902, pp. 537–8CrossRefGoogle Scholar.